11/
23/
05
i
AN
EVALUATION
OF
THE
EFFICACY
OF
THE
LEAD
HAZARD
REDUCTION
TREATMENTS
PRESCRIBED
IN
MARYLAND
ENVIRONMENTAL
ARTICLE
6­
8
FINAL
REPORT
SUBMITTED
TO:

BALTIMORE
CITY
HEALTHY
START,
INC.
BALTIMORE
CITY
HEALTH
DEPARTMENT­
LEAD
ABATEMENT
ACTION
PROGRAM
BALTIMORE,
MARYLAND
PREPARED
BY:

NATIONAL
CENTER
FOR
HEALTHY
HOUSING
(
formerly
the
National
Center
for
Lead­
Safe
Housing)
COLUMBIA,
MARYLAND
JANUARY
24,
2002
(
revised
April
30,
2002)

TABLE
OF
CONTENTS
11/
23/
05
ii
ABSTRACT
.....................................................................................................
1
EXECUTIVE
SUMMARY
.....................................................................................................
2
I.
INTRODUCTION
...................................................................................................
11
A.
Background
and
Purpose
of
this
Study...............................................................
11
B.
Specific
Research
Aims
and
Report
Objectives..................................................
12
II.
STUDY
DESIGN
AND
SAMPLE
COLLECTION
PROCEDURES
.............................
14
A.
Overview
of
the
Study
Design
...........................................................................
14
B.
Recruitment
and
Enrollment
Process..................................................................
15
C.
Selection
Criteria
for
Enrolled
Units
..................................................................
15
D.
Data
Collection
Forms
and
Protocols
.................................................................
16
E.
Dust
Lead
Sampling
Protocols
...........................................................................
18
F.
Visual
Assessment
Protocols..............................................................................
20
III.
LABORATORY
ANALYSIS
PROCEDURES
.............................................................
21
IV.
QUALITY
ASSURANCE
AND
QUALITY
CONTROL
..............................................
22
A.
Laboratory
Quality
Assurance
and
Quality
Control............................................
22
1.
Quality
Control
Spiked
Dust
Wipe
Sample
Results
................................
22
2.
Quality
Control
Field
Blank
Dust
Wipe
Sample
Results
.........................
22
3.
Procedures
for
Dust
Wipe
Sample
Results
that
Were
At
or
Below
Laboratory
Report
Limits
.......................................................................
22
B.
Performance
Audit
of
Field
Data
Collection
Activities.......................................
23
C.
Data
Audit
and
Data
Completeness
....................................................................
23
V.
DATA
PROCESSING
AND
STATISTICAL
ANALYSIS
PROCEDURES..................
24
A.
Data
Entry
and
Processing
.................................................................................
24
B.
Data
Summary
and
Statistical
Analysis
..............................................................
24
VI.
ENROLLMENT
RESULTS
BY
PHASE.......................................................................
24
VII.
PRE­
INTERVENTION
RESULTS
...............................................................................
25
A.
Results
of
Visual
Assessment
of
Baseline
Characteristics
and
Conditions
..........
25
1.
Baseline
Physical
Characteristics............................................................
25
2.
Baseline
Exterior
Building
and
Interior
Unit
Condition
..........................
27
B.
Results
of
Pre­
Intervention
Composite
and
Single
Surface
Dust
Sampling
.........
30
1.
Pre­
Intervention
Composite
Results........................................................
30
2.
Pre­
Intervention
Single
Surface
Results..................................................
32
3.
Comparison
of
Pre­
Intervention
Single
Surface
Results
with
Clearance
Standards/
Guidance................................................................................
34
11/
23/
05
iii
VIII.
IMMEDIATE
POST­
INTERVENTION
RESULTS              .
35
A.
Results
of
Immediate
Post­
Intervention
Composite
and
Single
Surface
Dust
Sampling      .
...................................................................................
35
1.
Immediate
Post­
Intervention
Composite
Results.....................................
35
2.
Reduction
in
Composite
Results
from
Pre­
to
Immediate
Post­
Intervention............................................................................................
37
3.
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention...................................................................
39
4.
Immediate
Post­
Intervention
Single
Surface
Results
...............................
40
5.
Reduction
in
Single
Surface
Results
from
Pre­
to
Immediate
Post­
Intervention............................................................................................
42
6.
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Preto
Immediate
Post­
Intervention                ..
45
7.
Comparison
of
Immediate
Post­
Intervention
Single
Surface
Results
with
Clearance
Standards/
Guidance
...............................................................
46
B.
Results
of
Immediate
Post­
Intervention
Visual
Assessments
..............................
48
1.
Types
of
Immediate
Post­
Intervention
"
Failures"
...................................
49
2.
Number
of
Rooms
with
"
Failures"
at
Phase
II
........................................
52
3.
Type
of
Contractor
and
Number
of
"
Failures"
at
Immediate
Post­
Intervention
...........................................................................................
52
C.
Description
of
and
Factors
Affecting
Costs
........................................................
54
D.
Description
of
Statistical
Modeling
Through
Immediate
Post­
Intervention.........
56
1.
Predicting
Dust
Lead
Loading
at
Immediate
Post­
Intervention
...............
56
2.
Predicting
Dust
Lead
Loading
"
Failures"
at
Immediate
Post­
Intervention57
3.
Predicting
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"...
58
IX.
ONE­
YEAR
POST­
INTERVENTION
(
PHASE
III)
RESULTS....................................
58
A.
Physical
Characteristics
and
Baseline
Condition
of
Enrolled
Units
Having
One­
Year
Post­
intervention
Data
...............................................................................
58
1.
Baseline
Physical
Characteristics
for
Enrolled
Units
Having
One­
Year
Post­
Intervention
Data............................................................................
58
2.
Baseline
Exterior
Building
and
Interior
Unit
Conditions
for
One­
Year
Post­
Intervention
Units...........................................................................
59
3.
Pre­
Intervention
Dust
Lead
Loadings
for
the
73
One­
Year
Post­
Intervention
Units...................................................................................
61
B.
One­
Year
Post­
Intervention
Composite
and
Single
Surface
Dust
Sampling
Results
...............................................................................................................
64
1.
Summary
of
One­
Year
Post­
Intervention
Composite
Dust
Sampling
Results
...................................................................................................
64
2.
Change
in
Composite
Results
from
Pre­
to
One­
Year
Post­
Intervention..
66
3.
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention....................................................................
68
4.
Summary
of
One­
Year
Post­
Intervention
Single
Surface
Results
............
69
11/
23/
05
iv
5.
Change
in
Single
Surface
Results
from
Pre­
to
One
Year
Post­
Intervention
...........................................................................................
69
6.
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Preto
One­
Year
Post
Intervention
................................................................
73
7.
Comparison
of
One
Year
Post
Intervention
Results
with
Clearance
Standards/
Guidance................................................................................
74
C.
Results
of
One­
Year
Post­
Intervention
Visual
Assessment        ...
74
1.
Types
of
One­
Year
Post­
Intervention
"
Failures".....................................
75
2.
Magnitude
of
One­
Year
Post­
Intervention
"
Failures"
.............................
76
3.
Rooms
with
One­
Year
Post­
Intervention
"
Failures"
...............................
78
4.
Type
of
Contractor
and
Number
of
One­
Year
Post­
Intervention
"
Failures"...............................................................................................
79
D.
Follow­
Up
Data
Summary
for
One­
Year
Post­
Intervention
(
Form
06)
...............
80
E.
Statistical
Modeling
Results
for
One­
Year
Post­
Intervention..............................
80
1.
Predicting
Dust
Lead
Loading
at
One­
Year
Post­
Intervention.................
80
2.
Predicting
One­
Year
Post­
Intervention
Dust
Lead
Loading
Exceedances
of
Standards
...............................................................................................
81
3.
Predicting
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
....
82
X.
SUMMARY
OF
TWO­
YEAR
POST­
INTERVENTION
RESULTS.............................
82
A.
Two­
Year
Post­
Intervention
Dust
Sampling
Results...........................................
82
1.
Summary
of
Two­
Year
Post­
Intervention
Composite
Results
.................
82
2.
Change
in
Composite
Results
from
Pre­
to
Two­
Years
Post­
Intervention............................................................................................
83
3.
Change
in
Composite
Results
from
One­
Year
to
Two­
Years
Post­
Intervention............................................................................................
84
B.
Results
of
Two­
Year
Post­
Intervention
Visual
Assessment
................................
84
XI.
SUMMARY
AND
DISCUSSION
OF
FINDINGS
........................................................
85
A.
Summary
and
Discussion
of
Findings
Through
Immediate
Post­
Intervention.....
85
1.
Comparison
of
Pre­
and
Immediate
Post­
Intervention
Dust
Lead
Loading
Results
...................................................................................................
85
2.
Immediate
Post­
Intervention
Visual
Assessment
Results
........................
87
3.
Cost
Data
Analysis
Results.....................................................................
88
B.
Summary
and
Discussion
of
Findings
Through
Two­
Years
Post­
Intervention
....
89
1.
Comparison
of
Pre­
Intervention
and
One­
Year
Post­
Intervention
Dust
Lead
Loading
Results.............................................................................
89
2.
One­
Year
Post­
Intervention
Visual
Assessment
Results
.........................
90
3.
Comparison
of
Pre­
Intervention
and
Two­
Year
Post­
Intervention
Dust
Lead
Loading
Results.............................................................................
91
C.
Comparing
the
Efficacy
of
the
Prescribed
Lead
Hazard
Reduction
Treatments
Both
With
and
Without
Window
Replacement...................................................
91
XII.
CONCLUSIONS
AND
RECOMMENDATIONS
.........................................................
92
11/
23/
05
v
A.
Conclusions
...................................................................................................
92
B.
Recommendations..............................................................................................
97
11/
23/
05
vi
TABLES
Table
1:
Data
Collection
Form
Type
and
Usage...............................................................
17
Table
2:
Dust
Wipe
Sample
Results
Below
Reporting
Limits
or
Reported
As
Zero..........
23
Table
3:
Dwelling
Unit
Baseline
Characteristics
­
Pre­
Intervention
Assessment...............
27
Table
4:
Units
with
Exterior
Building
Deterioration
 
Pre­
Intervention
Assessment.........
28
Table
5:
Number
of
Exterior
Baseline
Building
Deterioration
Items
 
Pre­
Intervention
Assessment
...................................................................................................
28
Table
6:
Units
with
Interior
Dwelling
Unit
Deterioration
 
Pre­
Intervention
Assessment..
29
Table
7:
Number
of
Interior
Dwelling
Unit
Deterioration
Items
 
Pre­
Intervention
Assessment
........................................................................................................
30
Table
8:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
....................................
31
Table
9:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loading...............................
33
Table
10:
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
.................
36
Table
11:
Percent
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Composite
Dust
Lead
Loadings
...........................................................................................
38
Table
12:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings..........................................................................
38
Table
13:
Number
and
Percent
of
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention...........................................
40
Table
14:
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loading
Results
............
41
Table
15:
Percent
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loadings                          
43
Table
16:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loadings                 ...
44
Table
17:
Number
and
Percent
of
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention           .
46
Table
18:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Floors........
47
Table
19:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Interior
Window
Sills
...................................................................................................
47
Table
20:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Window
Troughs                            ..
48
Table
21:
Number
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unit                              
49
Table
22:
Immediate
Post­
Intervention
Visual
Assessment
for
Lead
Hazard
Reduction
Treatment
"
Failures"
.........................................................................................
50
Table
23:
Reporting
Frequency
of
Specified
Types
of
Treatment
"
Failures"
at
Immediate
Post
Intervention................................................................................................
51
Table
24:
Mean
Number
of
Rooms
per
Unit
with
Specified
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
by
Treatment
Group............................................
53
Table
25:
Mean
Number
of
Immediate
Post­
Intervention
"
Failures"
per
Unit
by
Type
of
Contractor.                           .
53
Table
26:
Description
of
Lead
Hazard
Reduction
Costs
.....................................................
55
Table
27:
Mean
Lead
Hazard
Reduction
Costs
by
Contractor
Type
...................................
56
Table
28:
Dwelling
Unit
Baseline
Characteristics
for
One­
Year
Post­
Intervention
Units....
59
11/
23/
05
vii
Table
29:
Baseline
Exterior
Building
Deterioration
Summary
for
One­
Year
Post­
Intervention
Units
..............................................................................................
60
Table
30:
Number
of
Baseline
Interior
Dwelling
Unit
Deterioration
Items
for
One­
Year
Post­
Intervention
Units
......................................................................................
60
Table
31:
Summary
of
Baseline
Interior
Dwelling
Unit
Deterioration
Items
for
One­
Year
Post­
Intervention
Units
......................................................................................
61
Table
32:
Number
of
Baseline
Interior
Dwelling
Unit
Deterioration
Items
for
One­
Year
Post­
Intervention
Units
......................................................................................
61
Table
33:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
for
One­
Year
Post­
Intervention
Units
..............................................................................................
62
Table
34:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
for
One­
Year
Post­
Intervention
Units
..............................................................................................
63
Table
35:
One­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loadings
.................
65
Table
36:
Percent
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings............................................................................................................
67
Table
37:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Composition
Dust
Lead
Loadings
......................................................................
68
Table
38:
Number
and
Percent
of
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
.............................................
68
Table
39:
One­
Year
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Loadings
..........
70
Table
40:
Percent
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadings............................................................................................................
72
Table
41:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadings....................................................................
73
Table
42:
Number
and
Percent
of
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
............................................
73
Table
43:
Percentage
of
LHR
+
W
Units
with
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadings
Exceeding
Clearance
Standards
.........................................
74
Table
44:
Number
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unit
.
75
Table
45:
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
...............................
76
Table
46:
Reporting
Frequency
of
Specified
Types
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
......................................................................................
77
Table
47:
Mean
Number
of
Rooms
Per
Unit
with
Specified
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
by
Treatment
Group............................................
79
Table
48:
Mean
Number
of
One­
Year
Post­
Treatment
"
Failures"
Per
Unit
by
the
Type
of
Contractor..........................................................................................................
80
Table
49:
Two­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loadings.................
83
Table
50:
Median
Percent
Reductions
and
Absolute
Reduction
in
Pre­
to
Two­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings......................................................
84
Table
51:
Reporting
Frequency
of
Specified
Types
of
Two­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
......................................................................................
85
Table
52:
Summary
of
Pre­
and
Immediate
Post­
Intervention
Dust
Lead
Loading
Results
by
Assigned
Treatment
Group
and
Sampling
Method
............................................
86
Table
53:
Percentage
of
Units
with
One­
Year
Post­
Intervention
Single
Surface
Clearance
"
Failures"
.........................................................................................................
87
11/
23/
05
viii
Table
54:
Summary
of
Pre­
and
One­
Year
Post­
Intervention
Dust
Lead
Loadings
Results
by
Assigned
Treatment
Groups
and
Sampling
Method............................................
89
Table
55:
Percentage
of
Units
with
One­
Year
Post­
Intervention
Single
Surface
Clearance
"
Failures"
.........................................................................................................
90
Table
56:
Summary
of
Pre­
and
Two­
Years
Post­
Intervention
Composite
Dust
Lead
Loading
Results
by
Assigned
Treatment
Group
and
Sampling
Method
..............
91
Table
57:
Median
Percent
Reductions
in
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
for
LHR,
LHR
+
W,
and
Baltimore
Round
One
HUD
Evaluation
Units
.................................................................................................................
94
Table
58:
One­
Year
Post­
Intervention
"
Failure"
of
HUD
Clearance
Standards
for
LHR
+
W
and
Baltimore
Round
One
HUD
Evaluation
Units
Based
on
Single
Surface
Sampling
..........................................................................................................
95
Table
59:
Median
Percent
Reductions
in
Composite
Dust
Lead
Loadings
from
Pre­
to
Two­
Years
Post­
Intervention
for
LHR
and
LHR
+
W
Units
......................................
95
Table
60:
Median
Percent
Reduction
in
Lead
Loadings
from
Pre­
Intervention
to
Immediate
Post­
Intervention
for
LHR,
LHR
+
W,
and
Baltimore
Round
One
HUD
Evaluation
Units
................................................................................................
96
Table
61:
Immediate
Post­
Intervention
Clearance
"
Failure"
Rates
for
LHR,
LHR
+
W,
and
Baltimore
Round
One
HUD
Evaluation
Units
...................................................
96
11/
23/
05
ix
FIGURES
Figure
1:
Type
of
Building
 
Pre­
Intervention
Assessment
...............................................
26
Figure
2:
Year
Building
Constructed
 
Pre­
Intervention
Assessment.................................
26
Figure
3:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot......................
32
Figure
4:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
Box
Plot
................
34
Figure
5:
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot
...
37
Figure
6:
Pre­
and
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
for
LHR
and
LHR+
W
Units
(
Logarithmic
Scale)
....................................................
39
Figure
7:
Immediate
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Results
Loading
Box
Plot
...................................................................................................
42
Figure
8:
Pre­
and
Immediate
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
for
LHR
and
LHR+
W
Units
(
Logarithmic
Scale)
...............................................
45
Figure
9:
Percentage
of
Units
Having
Specified
Number
of
Rooms
Per
Unit
with
Immediate
Post­
Intervention
Visual
Assessment
"
Failures".................................................
52
Figure
10:
Pre­
and
One­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot
....................................................................................................................
66
Figure
11:
Pre­
and
One­
Year
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
Box
Plot.............................................................................................................
71
Figure
12:
Percentage
of
Units
Having
Specified
Number
of
Rooms
Per
Unit
with
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures".................................................
78
APPENDICES
Appendix
A:
Data
Collection
Forms
Appendix
B:
Laboratory
Analytical
Procedures
and
QA/
QC
Summary
Appendix
C:
Physical
Characteristics
of
Enrolled
Buildings/
Dwellings
 
Detailed
Listing
Appendix
D:
Baseline
Exterior
Building
and
Interior
Dwelling
Conditions
 
Detailed
Listing
Appendix
E:
Composite
Dust
Lead
Loading
(
µ
g/
ft2)
Results
 
Detailed
Listing
Appendix
F:
Single
Surface
Dust
Lead
Loading
(
µ
g/
ft2)
Results
 
Detailed
Listing
Appendix
G:
Complete
List
of
Visual
Assessment
"
Failures"­
by
Type
of
Failure
and
Assigned
Treatment
Group
Appendix
H:
Risk
Reduction
and
Window
Replacement
Costs
 
Detailed
Listing
Appendix
I:
Statistical
Data
for
Models
Predicting
Dust
Lead
Loading
at
Clearance,
Dust
Lead
Loading
"
Failures"
and
Post­
Intervention
Visual
Assessment
"
Failures"
Appendix
J:
Description
of
the
Relationship
Between
Single
Surface
and
Composite
Dust
Sample
Results
11/
23/
05
1
ABSTRACT
This
report
measured
the
degree
to
which
housing
is
made
lead­
safe
under
Maryland's
lead
law
(
Maryland
Environmental
Article
6­
8).
Baltimore
City
housing
units
were
enrolled
into
one
of
two
categories.
The
Lead
Hazard
Reduction
(
LHR)
category
included
57
units
that
underwent
the
minimum
set
of
treatments
required
by
the
statute,
and
the
Lead
Hazard
Reduction
plus
Windows
(
LHR+
W)
category
included
64
units
in
which
windows
were
replaced
as
well.
Upon
completion
of
the
work
and
as
required
by
the
law,
all
units
received
a
visual
inspection,
but
not
clearance
dust
testing,
by
Maryland
accredited
independent
Lead
Paint
Visual
Inspection
Contractors.
A
second
team
of
trained
and
certified
lead­
based
paint
risk
assessors
performed
a
visual
assessment
and
conducted
clearance
dust
testing
for
study
purposes.
Any
deficiencies
identified
by
the
study­
related
visual
assessment
were
corrected.

The
study
results
show
that
dust
lead
loadings
decreased
by
70
to
99
percent
between
pre­
and
immediate
post­
intervention.
However,
had
clearance
testing
been
required
by
the
statute,
46
percent,
8
percent
and
27
percent
of
units
would
have
"
failed"
clearance
based
on
floor,
window
sill
and
window
trough
clearance
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2
and
800
µ
g/
ft2,
respectively.
In
addition,
over
90
percent
of
units
had
one
or
more
treatment
"
failures"
based
on
the
studyrelated
confirmatory
visual
assessments
conducted
by
the
second
team
of
lead­
based
paint
risk
assessors.
In
order
to
ensure
that
units
are
safe
for
children,
this
study
recommends
clearance
dust
testing
and
more
rigorous
oversight
of
independent
Lead
Paint
Visual
Inspection
contractors
who
conduct
visual
inspections.
11/
23/
05
2
EXECUTIVE
SUMMARY
Working
with
the
U.
S.
Department
of
Housing
and
Urban
Development's
(
HUD's)
Office
of
Healthy
Homes
and
Lead
Hazard
Control
(
OHHLHC),
Baltimore
City's
Healthy
Start
Program
and
the
Baltimore
City
Health
Department's
Lead
Abatement
Action
Program
(
LAAP),
the
National
Center
for
Lead­
Safe
Housing
(
the
Center)
designed
this
study
to
evaluate
the
efficacy
of
a
prescribed
set
of
"
lead
hazard
reduction
treatments"
found
in
Maryland's
lead
law,
Maryland
Environmental
Article
6­
8
(
EA
6­
8)
in
reducing
dust
lead
loadings.

In
this
study,
pre­
and
immediate
post­
intervention
dust
sampling
and
visual
assessment
tools
were
utilized
to
determine
the
efficacy
of
the
law's
requirement
that,
upon
completion
of
the
prescribed
treatments,
only
an
independent
visual
inspection
is
needed
to
determine
whether
a
treated
rental
housing
unit
visually
"
passes"
or
"
fails"
the
prescribed
standard.
1
One­
year
postintervention
dust
sampling
and
visual
assessment
tools
were
then
utilized
to
evaluate
the
continued
effectiveness
of
treatments
prescribed
and
performed
under
EA
6­
8.
The
extent
to
which
lead
hazards
were
decreased
immediately
following
treatment
is
discussed,
primarily
by
comparing
pre­
with
immediate
post­
intervention
dust
lead
loadings
on
floors,
window
sills
and
window
troughs
in
a
set
of
rooms
where
young
children
were
likely
to
play,
eat
or
sleep.
This
report
also
documents
the
extent
to
which
lead
dust
re­
accumulated
on
various
surfaces
over
a
one­
year
period
after
completion
of
the
treatments,
primarily
by
comparing
one­
year
postintervention
dust
lead
loadings
on
floors,
window
sills
and
window
troughs
with
pre­
intervention
dust
lead
loadings.
Limited
data
for
a
two­
year
post­
intervention
period
are
also
discussed.

LAAP
inspectors,
who
were
trained
and
certified
as
lead­
based
paint
risk
assessors,
recruited
property
owners
or
managers
to
enroll
their
rental
housing
units
in
this
study.
To
be
enrolled,
a
unit
had
to
be
located
in
Baltimore
City,
constructed
prior
to
1950,
vacant
at
intervention
and
structurally
sound
as
determined
by
a
LAAP
screening
inspection.
A
structurally
deficient
unit
could
be
accepted
into
the
study
at
a
later
date
if
the
owner
corrected
observed
deficiencies.
(
Information
on
pre­
program
repairs
was
not
collected
as
part
of
the
study.)
Enrolled
units
were
assigned
into
two
treatment
categories:
(
1)
units
that
underwent
EA
6­
8'
s
prescribed
treatments
(
referred
to
as
"
LHR
units");
and
(
2)
units
that
underwent
these
treatments
plus
window
1
Ten
"
lead
hazard
reduction
treatments"
are
prescribed
by
the
standard:
 
Visual
review
of
all
exterior
and
interior
painted
surfaces;
 
Removal
and
repainting
of
chipping,
peeling
or
flaking
paint
on
exterior
and
interior
painted
surfaces;
 
Repair
of
any
structural
defect
that
is
causing
paint
to
chip,
peel
or
flake
that
the
owner
of
the
affected
property
has
knowledge
of
or,
with
the
exercise
of
reasonable
care,
should
have
knowledge
of;
 
Stripping
and
repainting,
replacing
or
encapsulating
all
interior
window
sills
with
vinyl,
metal
or
any
other
material;
 
Ensuring
that
caps
of
vinyl,
aluminum,
or
any
other
material
are
installed
in
all
window
wells
in
order
to
make
the
window
wells
smooth
and
cleanable;
 
Except
for
a
treated
or
replacement
window
that
is
free
of
lead­
based
paint
on
its
friction
surfaces,
fixing
the
top
sash
of
all
windows
in
place
in
order
to
eliminate
friction
caused
by
movement
of
the
top
sash;
 
Re­
hanging
all
doors
necessary
in
order
to
prevent
the
rubbing
together
of
a
lead­
painted
surface
with
another
surface;
 
Making
all
bare
floors
smooth
and
cleanable;
 
Ensuring
that
all
kitchen
and
bathroom
floors
are
overlaid
with
a
smooth,
water­
resistant
covering;
and
 
HEPA­
vacuuming
and
washing
the
interior
of
the
affected
property
with
high
phosphate
detergent
or
its
equivalent.
11/
23/
05
3
replacement
(
referred
to
as
"
LHR+
W
units").
The
owner
prepared
work
specifications
for
the
lead
hazard
intervention,
according
to
EA
6­
8'
s
prescribed
treatment
requirements.
LAAP
inspectors
then
conducted
a
pre­
intervention
walk­
through
inspection,
during
which,
in
accordance
with
study
protocols,
they
collected
composite
and
single
surface
dust
wipe
samples
from
floors,
window
sills
and
window
troughs
in
specific
rooms,
using
HUD
wipe
sampling
methods.
During
subsequent
phases,
in
accordance
with
study
protocols,
LAAP
inspectors
visited
units
to
collect
composite
and
single
surface
samples
from
the
same
surfaces,
rooms
and
locations
that
were
sampled
at
pre­
intervention.

All
samples
were
analyzed
for
total
lead
by
laboratories
participating
in
EPA's
National
Lead
Laboratory
Accreditation
Program
(
NLLAP)
and
proficient
in
the
Environmental
Lead
Proficiency
Analytical
Testing
Program
(
ELPAT).
Study
protocols
included
laboratory
and
field
quality
assurance/
quality
control
procedures.
Field
audits
and
data
audits
were
also
routinely
performed.

The
Center
was
responsible
for
all
data
entry
and
processing.
All
study
data
were
transmitted
into
and
maintained
in
Jetform's
FormFlow
software
program.
Statistical
analyses
were
performed
using
a
SAS
Institute
program,
which
also
generated
reports
and
tables.
A
Crystal
Reports
software
package
was
also
used
to
generate
certain
reports.

A.
Results
Through
Immediate
Post­
Intervention
A
total
of
177
units
(
91
LHR
and
86
LHR+
W)
underwent
complete
pre­
intervention
baseline
visual
assessments.
Pre­
intervention
composite
dust
samples
were
collected
in
148
of
the
177
units2
and
single
surface
dust
samples
were
collected
in
98
units.
The
EA
6­
8
prescribed
lead
hazard
reduction
treatments
were
completed
in
121
units
(
57
LHR
and
64
LHR+
W).
A
full
immediate
post­
intervention
study
assessment
was
performed
in
these
units,
including
a
visual
assessment
by
LAAP
inspectors,
collection
of
cost
and
concurrent
work
data
and
dust
testing.
Immediately
after
treatment,
both
the
LHR
and
the
LHR+
W
units
experienced
substantial
reductions
in
dust
lead
loadings
on
all
surfaces
(
see
Table
E­
1
on
page
4).
Using
either
of
the
dust
sampling
methods,
immediate
post­
intervention
dust
lead
loadings
on
bare
floors
of
LHR
units
were
generally
similar
to
those
on
bare
floors
of
LHR+
W
units.
However,
as
expected,
immediate
post­
intervention
dust
results
for
window
sills
and
troughs
in
LHR
units
were
significantly
higher
than
those
in
LHR+
W
units.

Despite
the
significant
median
decreases
in
dust
lead
loadings
between
pre­
and
immediate
postintervention
data
in
Table
E­
2
(
on
page
4)
do
show
that
some
units
experienced
increases
in
dust
lead
loadings
between
these
two
phases.
The
percentage
of
LHR+
W
units
that
underwent
an
increase
was
generally
lower
than
that
of
LHR
units
and
was
comparable
to
that
of
Baltimore
HUD
units.

2
Twenty­
nine
(
29)
units
were
not
included
primarily
because
the
owners'
grant
applications
were
not
approved.
11/
23/
05
4
Table
E­
1:
Median
Percent
and
µ
g/
ft2
Reductions
in
Pre­
to
Immediate
Post­
Intervention
Dust
Lead
Loadings
for
LHR,
LHR+
W
and
Baltimore
Round
One
HUD
Evaluation
Unitsb
Surface
Type
and
Sample
Type
LHR
Median
%
Reduction
(&
Median
µ
g/
ft2
Reduction)
LHR+
W
Median%
Reduction
(&
Median
µ
g/
ft2
Reduction)
Balt.
HUD
Median%
Reduction
(&
Median
µ
g/
ft2
Reduction)
Bare
Floors:
Composite
Single
Surface
85%
(
147
µ
g/
ft2)
70%
(
99
µ
g/
ft2)
89%
(
225
µ
g/
ft2)
88%
(
226
µ
g/
ft2)
NAa
95%
(
265
µ
g/
ft2)

Interior
window
sills:
Composite
Single
Surface
94%
(
1,686
µ
g/
ft2)
84%
(
497
µ
g/
ft2)
99%
(
2,596
µ
g/
ft2)
99%
(
2,516
µ
g/
ft2)
NAa
>
99%
(
2,492
µ
g/
ft2)

Window
Troughs:
Composite
Single
Surface
95%
(
9,708
µ
g/
ft2)
94%
(
4,111
µ
g/
ft2)
>
99%
(
11,284
µ
g/
ft2)
99%
(
12,022
µ
g/
ft2)
NAa
99%
(
3,591
µ
g/
ft2)

aNA=
not
applicable.
Composite
dust
samples
were
not
taken
in
the
Baltimore
Round
One
HUD
Evaluation
units.
bSingle
surface
dust
data
were
compared
to
data
obtained
from
Baltimore
City
housing
units
enrolled
in
Round
I
of
the
National
Evaluation
of
HUD's
OHHLHC's
Lead
Hazard
Control
Grant
Program.
These
units
underwent
interventions
similar
to
those
of
the
LHR+
W
units,
and
more
intensive
than
those
of
the
LHR
units.

Table
E­
2:
Number
and
Percent
of
Units
Having
an
Increase
in
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention
Surface
Type
and
Sample
Type
LHR
LHR+
W
Baltimore
HUD
Units
Bare
Floors:
Composite
Single
Surface
6/
49
(
12%)
4/
16
(
25%)
3/
56
(
5%)
4/
63
(
6%)
NA
12/
278
(
4%)

Interior
window
sills:
Composite
Single
Surface
4/
52
(
8%)
2/
16
(
12%)
2/
62
(
3%)
1/
62
(
2%)
NA
2/
277
(
1%)

Window
Troughs:
Composite
Single
Surface
3/
52
(
6%)
2/
16
(
12%)
1/
62
(
2%)
1/
62
(
2%)
NA
13/
273
(
5%)

Although
EA
6­
8
does
not
require
clearance
dust
testing
in
connection
with
the
prescribed
treatments,
post­
intervention
results
were
compared
with
clearance
standards
to
assess
the
extent
to
which
the
treatments
produced
dust
results
sufficient
for
safe
occupancy
of
the
treated
units
and
the
extent
to
which
the
prescribed
independent
visual
inspection
provided
a
sufficient
level
of
risk
reduction.
When
applying
these
clearance
standards
to
immediate
post­
intervention
individual
single
surface
dust
results,
an
appreciable
number
of
units
had
at
least
one
floor,
window
sill
or
window
trough
dust
lead
result
that
exceeded
the
Maryland
and
HUD/
EPA
guidance/
standards
at
the
time
of
the
study
(
see
Table
E­
3
on
page
5).
Assigned
treatment
group
was
significantly
associated
with
post­
intervention
window
sill
and
trough
results
"
passing"
or
11/
23/
05
5
"
failing"
their
respective
standards
but
was
not
significantly
associated
with
bare
floors
"
passing"
or
"
failing"
standards.
Had
clearance
testing
been
required,
46
percent
of
the
units
had
at
least
one
sample
that
would
have
"
failed"
a
floor
clearance
level
of
100
µ
g/
ft2,
nine
percent
had
at
least
one
sample
that
"
failed"
interior
window
sill
clearance
of
500
µ
g/
ft2
and
27
percent
had
at
least
one
sample
that
"
failed"
window
trough
clearance
of
800
µ
g/
ft2.
These
are
higher
"
failure"
rates
than
those
observed
for
vacant
Baltimore
HUD
units
enrolled
in
the
HUD
National
Evaluation
(
29
percent
for
floors,
one
percent
for
sills
and
six
percent
for
troughs).
3
It
is
apparent
that
dust
clearance
testing
is
critical
to
ensure
that
treated
dwellings
are
safe
for
reoccupancy

Table
E­
3:
Percentage
of
Units
with
Immediate
Post­
Intervention
Single
Surface
Clearance
"
Failures"
Percent
of
Units
within
Specified
Category
Surface
Type:
Floors
Interior
Window
Sills
Window
Troughs
Std/
Guidance:

100
µ
g/
ft2

200
µ
g/
ft2

500
µ
g/
ft2

800
µ
g/
ft2
LHR
Units
60%
35%
20%
50%
LHR+
W
Units
41%
25%
5%
19%
Total
All
Units
46%
28%
8%
27%

Although
statistically
significant
dust
lead
loading
reductions
were
found
and
each
unit
passed
the
independent
visual
inspection
prescribed
by
EA
6­
8,
immediate
post­
intervention
confirmatory
visual
assessments
conducted
by
LAAP
inspectors
yielded
a
high
percentage
of
units
(
93
percent)
with
visual
"
failures"
of
one
or
more
of
the
prescribed
lead
hazard
reduction
treatments
(
see
Table
E­
4
on
page
6).
4
The
three
most
common
"
failures"
were:
(
1)
not
all
paint
intact,
with
some
chipping,
flaking
and
peeling
paint
remaining
(
75
percent
of
all
units);
(
2)
one
or
more
painted
doors
continuing
to
rub
together
and/
or
bind
(
43
percent);
and
(
3)
visible
paint
chips
and/
or
debris
remaining
(
38
percent).
The
"
failures"
that
LAAP
inspectors
recorded
were
easily
observed,
with
a
dwelling
unit
geometric
mean
of
1
ft2
of
non­
intact
paint
reported
in
LHR
and
LHR+
W
units,
and
dwelling
unit
geometric
means
of
8
and
16
ft2
of
visible
paint
chips
and
debris
in
LHR
and
LHR+
W
units.
It
should
be
noted
that
the
areal
density
(
i.
e.,
depth)
and
lead
concentration
of
chips/
debris
were
not
measured.
The
owner
corrected
treatment
"
failures"
noted
during
the
immediate
post­
intervention
sampling
visit
before
tenants
moved
in.
These
corrections
were
"
validated"
by
LAAP
inspectors,
who
required
owners
and/
or
contractors
to
complete
their
work
prior
to
payment.

These
results
indicate
that
lead
hazard
reduction
treatments
did
not
completely
meet
the
EA
6­
8
risk
reduction
standard
and
more
strongly
suggest
that
the
independent
visual
inspections
failed
to
identify
all
lead
hazard
reduction
"
failures."
Because
only
five
or
six
independent
inspectors
conducted
inspections
for
these
study
units,
this
finding
does
not
necessarily
reflect
on
all
other
Maryland­
certified
inspectors.

3
Clearance
failure
rates
for
the
Baltimore
HUD
units
are
based
on
initial
clearance
testing
conducted
after
treatment.
4
For
the
purposes
of
this
study,
a
treatment
"
failure"
was
defined
as
an
observation
by
the
LAAP
inspector
that
one
or
more
of
the
ten
lead
hazard
reduction
treatments
prescribed
in
the
statute
had
not
been
fully
completed.
Note
that
the
Maryland
statute
did
not
specify
a
"
de
minimis"
level
above
which
the
treatment
was
considered
a
failure;
therefore,
any
observable
deficient
or
missing
treatment
was
classified
as
a
"
failure."
11/
23/
05
6
Table
E­
4:
Number
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unita
Number
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unitb
(
57
LHR
units,
64
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
lead
hazard
treatment
"
failures"
6
(
10%)
4
(
6%)
10
(
8%)
"
1"
lead
hazard
treatment
"
failure"
7
(
12%)
26
(
41%)
33
(
27%)
"
2"
lead
hazard
treatment
"
failures"
22
(
39%)
21
(
34%)
44
(
36%)
"
3"
lead
hazard
treatment
"
failures"
19
(
33%)
9
(
14%)
28
(
23%)
"
4"
lead
hazard
treatment
"
failures"
2
(
4%)
3
(
5%)
5
(
4%)
"
5"
lead
hazard
treatment
"
failures"
1
(
2%)
0
(
0%)
1
(
1%)
Source:
Form
05
aFor
the
purposes
of
this
study,
a
treatment
"
failure"
was
defined
as
an
observation
by
the
LAAP
inspector
that
one
or
more
of
the
ten
lead
hazard
reduction
treatments
prescribed
in
the
statute
had
not
been
fully
completed.
"
Failures"
were
counted
on
a
per
unit
basis.
Within
each
unit,
failures
were
listed
by
room
ID.
bThe
mean
number
of
"
failures"
per
unit
(
2.1
for
LHR
units;
1.7
for
LHR+
W
units)
was
significantly
associated
with
assigned
treatment
group
(
p=
0.0255).

Various
cost
data
were
collected
during
this
study,
including
LAAP­
approved
costs
for
the
prescribed
lead
hazard
reduction
treatments,
plus
the
additional
window
replacement
costs
for
the
LHR+
W
units.
Owner­
estimated
turnover
costs
that
would
have
been
incurred
even
if
the
enrolled
units
had
not
been
subject
to
the
EA
6­
8
treatments
were
also
collected.
However,
these
data
were
highly
subjective
and
were
excluded
from
the
study.
Finally,
the
type
of
contractor
performing
the
work
and
information
on
concurrent
work
was
also
collected.
The
median
cost
for
completing
the
prescribed
treatments
in
LHR
units
and
LHR+
W
units
was
$
2,154
and
$
1,649,
respectively.
Costs
in
LHR
units
were
likely
higher
because
the
expense
of
the
prescribed
window
treatments
was
included,
while
window
replacement
costs
in
the
LHR+
W
units
were
calculated
separately.
The
median
cost
of
treatment
plus
window
replacement
in
the
LHR+
W
units
was
$
4,348,
with
an
average
of
9
to
10
windows
being
replaced
per
unit.
Treatment
costs
were
higher
than
expected
for
a
typical
Baltimore
City
dwelling
and
should
not
be
considered
average
or
representative
of
all
Baltimore
City
rental
unit
costs.

For
LHR
units,
the
mean
costs
for
for­
profit
contractors
were
slightly
higher
than
those
for
property
owners/
employees,
while
the
reverse
was
observed
for
LHR+
W
units.
Concurrent
work
(
e.
g.,
roofing,
plumbing,
heating
and
electrical
repairs
and
replacement
of
other
fixtures
and
components)
was
performed
in
17
of
the
57
LHR
units
and
in
27
of
the
64
LHR+
W
units.

For
each
surface
type,
a
backward
elimination
multiple
regression
model
was
run
to
identify
factors
that
were
significant
predictors
of
immediate
post­
intervention
dust
lead
loadings.
No
significant
predictors
were
found
for
bare
floors.
However,
pre­
intervention
sill
dust
lead
loadings
and
the
assigned
treatment
group
were
found
to
be
significant
predictors
of
immediate
post­
intervention
sill
results.
For
window
troughs,
pre­
intervention
trough
dust
lead
loadings
and
the
assigned
treatment
group
were
significant
predictors
of
immediate
post­
intervention
trough
results.

Clearance
"
failures"
on
bare
floors
was
not
modeled
since
no
relationship
was
found
when
attempting
to
predict
dust
lead
loading
at
clearance.
For
sills
and
troughs,
a
logistic
regression
model
was
run
to
identify
factors
that
are
significant
predictors
of
whether
units
would
have
11/
23/
05
7
clearance
"
failures"
based
on
floor,
sill
and
trough
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2
and
800
µ
g/
ft2,
respectively.
For
sills,
the
only
variable
found
to
significantly
predict
immediate
postintervention
clearance
"
failures"
was
the
percent
of
windows
replaced
out
of
the
total
number
of
windows
in
a
unit.
For
troughs,
the
number
of
items
of
exterior
building
deterioration
(
yielding
more
clearance
"
failures")
and
the
percent
of
windows
replaced
(
yielding
fewer
clearance
"
failures")
were
found
to
be
significant
predictors
of
clearance
"
failure"
at
immediate
postintervention

A
logistic
model
was
employed
to
identify
factors
that
are
significant
predictors
of
immediate
post­
intervention
visual
assessment
"
failures."
The
only
variable
found
to
significantly
influence
visual
assessment
"
failure"
was
the
number
of
items
of
exterior
building
deterioration
(
increasing
items
yielding
more
visual
assessment
"
failures").

B.
Results
Through
One­
Year
Post­
Intervention
A
total
of
73
units
(
36
LHR
and
37
LHR+
W)
were
sampled
approximately
one
year
after
treatments
were
completed.
LAAP
inspectors
also
performed
a
full
visual
assessment
at
one
year
post­
intervention,
noting
any
"
failures"
that
had
occurred
since
their
previous
visit.

At
one­
year
post­
intervention,
composite
and
single
surface
dust
lead
loadings
on
all
surfaces
remained
well
below
pre­
intervention
levels
(
see
Table
E­
5).
Substantial
median
percent
decreases
and
median
µ
g/
ft2
decreases
in
dust
lead
loadings
were
observed
for
floors,
sills
and
window
troughs
between
pre­
and
one­
year
post­
intervention.
Dust
lead
loadings
on
window
sills
and
troughs
in
LHR
units
were
significantly
higher
than
those
in
LHR+
W
units.

Table
E­
5:
Median
Percent
(
µ
g/
ft2)
Reductions
in
Dust
Lead
Loadings
for
LHR,
LHR+
W
and
Baltimore
Round
One
HUD
Evaluation
Units
Between
Pre­
Intervention
and
One
Year
Post­
Intervention
Surface
Type
and
Sample
Type
LHR
Median
%
Reduction
(&
Median
µ
g/
ft2
Reduction)
Pre
to
1
Year
Post
LHR+
W
Median
%
Reduction
(&
Median
µ
g/
ft2
Reduction)
Pre
to
1
Year
Post
Balt.
HUD
Median
Reduction
(&
Median
µ
g/
ft2
Reduction)
Pre
to
1
Year
Post
Bare
Floors:
Composite
Single
Surface
82%
(
87
µ
g/
ft2)
NAb
94%
(
335
µ
g/
ft2)
96%
(
420
µ
g/
ft2)
NAa
84%
(
220
µ
g/
ft2)

Interior
window
sills:
Composite
Single
Surface
82%
(
713
µ
g/
ft2)
NAb
97%
(
2,803
µ
g/
ft2)
97%
(
1,875
µ
g/
ft2)
NAa
98%
(
2,617
µ
g/
ft2)

Window
Troughs:
Composite
Single
Surface
89%
(
4,507
µ
g/
ft2)
NAb
98%
(
12,014
µ
g/
ft2)
99%
(
16,338
µ
g/
ft2)
NAa
90%
(
3,308
µ
g/
ft2)

aNA=
not
applicable.
Composite
dust
samples
were
not
taken
in
Baltimore
Round
I
HUD
Eval.
units.
bNA=
not
applicable.
Single
surface
samples
were
collected
in
only
2
LHR
unit
at
one
year
post­
intervention.
11/
23/
05
8
One­
year
post­
intervention
single
surface
results
for
LHR+
W
units
were
compared
with
clearance
standards
to
assess
the
extent
to
which
the
treatments
continued
to
produce
dust
results
sufficient
for
safe
occupancy
of
the
treated
units.
Insufficient
single
surface
data
were
available
for
LHR
units.
Less
than
20
percent
of
units
had
at
least
one
bare
floor,
sill,
or
trough
result
that
exceeded
HUD
standards
(
see
Table
E­
6).
The
percentages
for
floors
and
troughs
in
LHR+
W
units
were
significantly
lower
than
those
found
for
Baltimore
Round
I
units
in
the
HUD
National
Evaluation.

Table
E­
6:
Percentage
of
Units
with
One­
Year
Post­
Intervention
Single
Surface
Clearance
"
Failures"
Percent
of
Units
within
Specified
Category
Surface
Type:
Floors
Interior
Window
Sills
Window
Troughs
Std/
Guidance:

100
µ
g/
ft2

200
µ
g/
ft2

500
µ
g/
ft2

800
µ
g/
ft2
LHR+
W
Units
14
11
17
14
Balt.
HUD
Units
57
NA
11
35
Although
dust
lead
loadings
remained
less
than
pre­
intervention
levels
and
a
large
percentage
of
units
had
dust
lead
loadings
less
than
HUD
clearance
standards
at
one
year
post­
intervention,
inspections
conducted
by
LAAP
inspectors
one
year
after
treatments
were
first
implemented
found
that
96%
of
the
73
units
had
at
least
one
visual
assessment
"
failure"
(
see
Table
E­
7
on
page
9).
The
most
common
"
failures"
were
the
same
as
those
identified
at
immediate
postintervention
not
all
paint
intact
(
93%),
doors
continuing
to
rub
(
41%)
and
visible
chips
or
debris
(
29%).
It
should
be
noted
that,
for
the
visible
paint
chips
or
debris,
the
areal
density
(
i.
e.,
depth)
of
the
chips/
debris
and
the
lead
concentration
in
those
chips/
debris
was
not
measured.
A
geometric
mean5
of
1
ft2
of
paint
was
not
intact,
1
ft2
of
paint
chips
or
debris
remained,
and
4
ft2
of
flooring
was
not
smooth
and
cleanable.
While
these
types
of
problems
were
reportedly
treated
after
immediate
post­
intervention
dust
samples
were
collected
and
visual
assessments
were
performed,
these
results
are
not
entirely
surprising
given
that
few
units
appeared
to
experience
turnover
or
had
further
treatments
during
the
one­
year
post­
intervention
period.

For
each
surface
type,
a
multiple
regression
model
with
backward
elimination6
was
run
to
identify
factors
that
were
significant
predictors
of
one­
year
post­
intervention
dust
lead
loadings.
The
only
significant
predictor
of
bare
floor
dust
lead
loadings
was
the
percent
of
rooms
with
visual
assessment
"
failures"
at
one
year
post­
intervention.
Pre­
intervention
sill
dust
lead
loadings
and
assigned
treatment
group
were
significant
predictors
of
one­
year
post­
intervention
5
Note
that
these
geometric
mean
values
were
calculated
using
only
values
reported
for
units
that
had
the
specified
type
of
"
failure"
reported.
Units
that
did
not
have
the
specified
type
of
"
failure"
reported
were
not
included
in
the
calculations.
6
In
a
multiple
regression
model
with
backward
elimination,
all
possible
predictors
of
the
outcome
are
initially
entered
into
the
model.
Then
hypothesis
tests
are
run
to
determine
if
any
factors
can
be
removed
from
the
predictive
equation
when
the
other
factors
are
retained.
The
least
significant
factor
(
i.
e.,
the
factor
with
the
largetst
observed
significance
level)
is
removed
and
the
process
is
repeated
to
determine
if
more
factors
can
be
dropped.
In
a
logistic
model
with
backward
elimination,
the
outcome
of
interest
is
a
binary
response
variable
(
e.
g.,
"
pass/
fail").
In
a
Poisson
regression
model
with
backward
elimination,
the
outcome
of
interest
is
a
count
variable
(
e.
g.,
number
of
failures).
The
same
backward
elimination
procedure
described
above
for
regression
modeling
was
followed
for
both
the
logistic
and
the
Poisson
models.
11/
23/
05
9
sill
dust
lead
loadings.
Assigned
treatment
group
was
the
only
significant
predictor
of
one
year
post­
intervention
trough
dust
lead
loadings.

Table
E­
7:
Number
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unita
Number
of
Visual
Assessment
"
Failures"
Per
Unit
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
One­
yr
post­
int
visual
assessment
"
failures"
0
3
(
8%)
3
(
4%)
"
1"
One­
yr
post­
int
visual
assessment
"
failure"
8
(
22%)
13
(
35%)
21
(
29%)
"
2"
One­
yr
post­
int
visual
assessment
"
failures"
12
(
33%)
9
(
24%)
21
(
29%)
"
3"
One­
yr
post­
int
visual
assessment
"
failures"
11
(
31%)
9
(
24%)
20
(
27%)
"
4"
One­
yr
post­
int
visual
assessment
"
failures"
3
(
8%)
3
(
8%)
6
(
8%)
"
5"
One­
yr
post­
int
visual
assessment
"
failures"
2
(
6%)
0
2
(
3%)
Source:
Form
05
aFor
the
purposes
of
this
study,
a
treatment
"
failure"
was
defined
as
an
observation
by
the
LAAP
inspector
that
one
or
more
of
the
ten
lead
hazard
reduction
treatments
prescribed
in
the
statute
had
not
been
fully
completed.
"
Failure"
was
counted
on
a
per
unit
basis.
Within
each
unit,
"
failures"
were
listed
by
room
ID.

For
floors,
sills
and
troughs,
a
logistic
regression
model
with
backward
elimination6
was
run
to
identify
factors
that
were
significant
predictors
of
whether
units
would
have
clearance
"
failures"
at
immediate
post­
intervention
based
on
floor,
sill
and
trough
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2
and
800
µ
g/
ft2,
respectively.
For
bare
floors,
the
only
significant
predictor
of
one­
year
postintervention
composite
results
exceeding
100
µ
g/
ft2
was
the
year
of
construction.
For
sills,
preintervention
sill
dust
lead
loadings
and
assigned
treatment
group
were
significant
predictors
of
one­
year
post­
intervention
composite
dust
lead
loadings
exceeding
500
µ
g/
ft2.
For
troughs,
the
only
significant
predictor
of
one­
year
post­
intervention
composite
dust
lead
loadings
being
above
800
µ
g/
ft2
was
assigned
treatment
group.

Finally,
a
Poisson
regression
model
with
backward
elimination6
was
employed
to
identify
factors
that
are
significant
predictors
of
visual
assessment
"
failures"
at
one­
year
post­
intervention.
Significant
predictors
were
lead
hazard
reduction
cost
at
Phase
II,
building
type,
whether
work
had
been
performed
during
the
past
year
and
the
estimated
market
value
of
the
dwelling.

C.
Summary
of
Study
Findings
The
main
findings
of
this
study
are
that
dust
lead
loadings
declined
substantially
immediately
after
EA
6­
8
prescribed
lead
hazard
reduction
treatments
were
implemented.
However,
many
units
would
not
have
"
passed"
clearance
dust
tests
immediately
following
treatment
had
such
testing
been
required.
In
addition,
the
prescribed
independent
visual
inspections
conducted
in
these
units
immediately
following
treatment
generally
missed
many
treatment
"
failures."
At
one
year
post­
intervention,
dust
lead
loadings
generally
remained
very
low,
with
over
80
percent
of
units
having
floor,
sill
and
trough
dust
lead
loadings
that
were
below
clearance
standards.
However,
almost
every
unit
had
at
least
one
visual
assessment
failure
one
year
after
treatment.
11/
23/
05
10
Based
on
study
findings,
it
is
recommended
that:
 
Appropriate
state
agencies
should
increase
oversight
of
independent
visual
inspectors
to
ensure
that
such
inspectors
are
performing
visual
inspections
in
accordance
with
approved
protocols
and
inspector
training;
 
Clearance
dust
testing
should
be
added
to
the
independent
visual
inspection
as
part
of
the
lead
hazard
reduction
treatment
requirements
of
EA
6­
8,
either
by
regulation
or
by
amendment
to
the
statute;
 
Rental
property
owners
and
their
crews
and
independent
contractors
should
perform
a
more
intensive
final
cleaning
upon
completion
of
the
prescribed
treatments;
 
More
comprehensive
window
treatments
may
be
needed
to
ensure
that
interior
window
sills
and
troughs
do
not
continue
to
be
an
exposure
source
for
lead;
and
 
Proper
adjustment
and
re­
hanging
of
doors
to
eliminate
friction
points
should
be
an
emphasis
of
the
prescribed
treatments.
11/
23/
05
11
I.
INTRODUCTION
A.
Background
and
Purpose
of
this
Study
In
May
1994,
then
Maryland
Governor
William
Donald
Schaefer
signed
comprehensive
legislation
into
law
that
was
designed
to
further
two
primary
goals:
the
prevention
of
childhood
lead
poisoning
resulting
from
deteriorating
lead­
based
paint
and
the
preservation
of
the
state's
affordable
rental
housing
stock.
"
The
Lead
Poisoning
Prevention
Program,"
codified
variously
in
the
Environmental
Article
as
Section
6­
801
et.
seq.
(
EA
6­
8),
in
the
Insurance
Article
as
Section
734
et.
seq.
and
in
the
Real
Property
Article
as
Section
8­
208.2
et.
seq.
of
the
Maryland
Annotated
Code,
was
to
have
become
effective
on
October
1,
1994.
However,
due
to
inevitable
delays
in
promulgating
and
implementing
regulations
by
the
Maryland
Department
of
the
Environment
(
MDE),
the
law
did
not
become
fully
effective
until
February
24,
1996.

EA
6­
8
applies
to
all
privately
owned
rental
housing
constructed
prior
to
1950
(
referred
to
as
"
target
housing"
in
the
law)
and,
at
a
property
owner's
option,
to
any
residential
rental
property
constructed
after
1949
but
prior
to
1978.
The
heart
of
the
law
is
a
system
and
schedule
for
reducing
lead­
based
paint
hazards
at
rental
unit
turnover
by
meeting
a
prescribed
"
risk
reduction
standard,"
as
found
in
Section
6­
815.
The
standard
can
be
met
by
performing
tests
to
verify
that
surface
dust
lead
loadings
in
affected
property
are
within
acceptable
limits
or,
alternatively,
by
performing
prescribed
"
lead
hazard
reduction
treatments."
These
treatments
include
the
following:

 
A
visual
review
of
all
exterior
and
interior
painted
surfaces;
 
The
removal
and
repainting
of
chipping,
peeling
or
flaking
paint
on
exterior
and
interior
painted
surfaces;
 
The
repair
of
any
structural
defect
that
is
causing
the
paint
to
chip,
peel
or
flake
that
the
owner
of
the
affected
property
has
knowledge
of
or,
with
the
exercise
of
reasonable
care,
should
have
knowledge
of;
 
Stripping
and
repainting,
replacing
or
encapsulating
all
interior
window
sills
with
vinyl,
metal
or
any
other
material;
 
Ensuring
that
caps
of
vinyl,
aluminum
or
any
other
material
are
installed
in
all
window
wells
in
order
to
make
the
window
wells
smooth
and
cleanable;
 
Except
for
a
treated
or
replacement
window
that
is
free
of
lead­
based
paint
on
its
friction
surfaces,
fixing
the
top
sash
of
all
windows
in
place
in
order
to
eliminate
the
friction
caused
by
movement
of
the
top
sash;
 
Re­
hanging
all
doors
necessary
in
order
to
prevent
the
rubbing
together
of
a
lead­
painted
surface
with
another
surface;
 
Making
all
bare
floors
smooth
and
cleanable;
 
Ensuring
that
all
kitchen
and
bathroom
floors
are
overlaid
with
a
smooth,
water­
resistant
covering;
and
 
HEPA­
vacuuming
and
washing
of
the
interior
of
the
affected
property
with
high
phosphate
detergent
or
its
equivalent.
11/
23/
05
12
Lead
hazard
reduction
treatments
could
be
performed
by
a
for­
profit
contractor,
the
property
owner
and/
or
his/
her
employees,
or
by
a
non­
profit
contractor.
Regardless
of
their
affiliation,
all
individuals
performing
risk
reduction
measures
had
to
meet
state
and
federal
lead
training,
certification,
and
licensing
requirements.

At
each
change
in
occupancy,
an
owner
is
required
to
have
the
housing
unit
inspected
to
verify
that
the
lead
hazard
reduction
treatments
have
been
implemented.
This
is
accomplished
by
having
a
visual
inspection
performed
by
an
independent
inspector
certified
by
MDE.

In
essence,
under
EA
6­
8,
children
and
their
representatives
cannot
sue
for
traditional
tort
damages
for
lead
poisoning
if
the
rental
property
owner
has
met
the
risk
reduction
standard
and
certain
housing
unit
registration
requirements.
Additionally,
the
owner
and/
or
his
or
her
property
insurer
must
make
a
"
qualified
offer"
to
a
resident
child
and
his
or
her
legal
representative
when
the
child's
blood
lead
level
has
exceeded
25
micrograms
per
deciliter
(
25
µ
g/
dL)
of
blood.
The
qualified
offer
is,
in
effect,
a
remedial
compensation
settlement
of
the
child's
potential
lead
poisoning
claim,
which
is
designed
to
pay
for
necessary
out­
of­
pocket
medical
treatment
and
relocation
costs
for
the
child's
family
to
move
to
"
lead­
safe"
housing,
up
to
a
combined
total
of
$
17,000.

Because
of
this
limited
tort
immunity
and
concerns
over
whether
the
prescribed
lead
hazard
reduction
treatments
are
protective
of
children's
health,
some
critics
of
EA
6­
8
have
charged
that
the
failure
to
require
clearance
dust
testing
upon
completion
of
the
treatments
constitutes
a
major
deficiency
in
the
law.
Many
have
questioned
whether
these
treatments
would,
in
fact,
result
in
safer
home
environments
in
the
absence
of
dust
testing.
For
this
reason,
there
was
a
need
to
evaluate
the
efficacy
of
both
the
prescribed
treatments
and
the
independent
visual
inspection
protocol.

Although
the
legislature
directed
the
Maryland
Lead
Poisoning
Prevention
Commission
to
study
and
collect
information
on
the
effectiveness
of
the
bill's
prescribed
lead
hazard
reduction
treatments,
the
legislature
did
not
specifically
earmark
funding
for
such
an
evaluation.
Therefore,
the
National
Center
for
Lead­
Safe
Housing
(
Center)
and
the
Baltimore
City
Healthy
Start,
Inc.
(
Healthy
Start)
elected
to
independently
evaluate
the
aforementioned
measures
through
a
US
Department
of
Housing
and
Urban
Development,
Office
of
Healthy
Homes
and
Lead
Hazard
Control
(
HUD
OHHLHC)
funded
research
project.
Specifically,
the
Center
was
funded
under
the
HUD
OHHLHC
Lead
Hazard
Control
Round
Three
Grant
to
the
Baltimore
City
Health
Department
(
Grant
No.
MDLAG0045­
95).
(
Initial
funding
received
from
the
Abell
Foundation
also
aided
in
the
development
of
this
study.)
Working
with
the
HUD
OHHLHC,
Healthy
Start
and
the
Baltimore
City
Health
Department's
Lead
Abatement
Action
Program
(
LAAP),
the
Center
designed
this
study
and
has
analyzed
all
data
collected.

B.
Specific
Research
Aims
and
Objectives
The
primary
goal
of
this
study
was
to
evaluate
the
efficacy
of
the
EA
6­
8
lead
hazard
reduction
treatments,
both
in
units
that
undergo
these
treatments
alone
and
in
units
in
which
the
treatments
plus
window
replacement
have
been
carried
out.
This
report
documents
the
extent
to
which
lead
hazards
were
reduced
immediately
following
treatment,
primarily
by
comparing
the
level
of
lead
11/
23/
05
13
loading
in
household
dust
collected
shortly
before
the
lead
hazard
reduction
intervention
with
the
level
of
dust
collected
immediately
after
the
intervention.
This
report
also
includes
a
further
evaluation
of
lead
treatment
efficacy
using
the
results
of
dust
lead
sampling
and
visual
assessments
conducted
12
months
post­
intervention.
Limited
data
collected
at
approximately
24
months
post­
intervention
are
also
summarized.

Key
questions
for
the
study
are:

1.
How
effective
are
the
prescribed
treatments
in
reducing
dust
lead
loading
on
floors,
interior
window
sills
and
window
troughs
(
i.
e.,
window
wells)
to
acceptable
levels?
2.
To
what
extent
does
lead
dust
re­
accumulate
on
these
surfaces
over
a
one­
and
two­
year
period
after
completion
of
the
treatments?
3.
Does
the
replacement
of
windows
in
units
that
also
undergo
the
prescribed
treatments
result
in
substantially
lower
dust
lead
loadings
immediately
after
completion
of
the
treatments?
4.
Do
single
surface
dust
wipe
sample
and
composite
dust
wipe
sample
results
correlate
well
in
measuring
dust
lead
loadings
both
before
and
after
completion
of
the
prescribed
treatments?
(
Note:
Since
this
question
does
not
specifically
apply
to
an
evaluation
of
EA
6­
8,
related
findings
are
presented
in
Appendix
J
and
are
not
discussed
in
the
body
of
this
report.)
5.
Aside
from
lead
dust
sample
results,
do
independent
visual
inspections
accurately
assess
whether
the
prescribed
treatments
have
been
completed?

Specific
objectives
for
this
report
are
to:

1.
Describe
the
physical
characteristics
and
baseline
(
i.
e.,
pre­
intervention)
condition
of
enrolled
housing
units,
including
dust
lead
loading
on
floors,
interior
window
sills
and
window
troughs;
2.
Describe
the
immediate
post­
intervention
condition
of
the
enrolled
units,
including
their
conformance
with
the
prescribed
lead
hazard
reduction
treatments
and
whether
they
would
pass
an
independent
visual
inspection,
the
dust
lead
loading
and
the
number
of
windows
replaced
in
the
window
replacement
subset;
3.
Describe
the
relationship
between
baseline
building
characteristics
and
dwelling
unit
conditions
and
the
type
of
lead
hazard
reduction
intervention
performed
(
i.
e.,
lead
hazard
reduction
treatments
only
and
lead
hazard
reduction
treatments
plus
window
replacement);
4.
Compare
the
efficacy
of
the
prescribed
lead
hazard
reduction
treatments
and
the
prescribed
treatments
plus
window
replacement
in
reducing
dust
lead
loadings
and
bringing
post­
intervention
dust
lead
loadings
below
"
clearance"
standards
established
by
MDE
and
by
the
HUD
OHHLHC
and
the
US
Environmental
Protection
Agency
(
EPA)
as
"
guidance;"
5.
Describe
the
dust
lead
loading
data
and
post­
treatment/
visual
inspection
data
for
the
enrolled
dwelling
one
and
two
years
after
completion
of
the
treatments;
6.
Describe
the
relationship
between
such
factors
as
type
and
age
of
building,
baseline
condition,
type
of
contractor,
amount
and
character
of
any
concurrent
work,
turnover
treatment,
any
notice
of
newly
identified
lead
hazard
or
of
a
child
with
an
elevated
blood
11/
23/
05
14
lead
level
and
the
effectiveness
of
the
treatments,
including
dust
lead
loading
and
continued
conformance
with
EA
6­
8
standards;
and
7.
Quantify
any
additional
effectiveness
of
window
treatment
in
maintaining
low
dust
lead
loadings.

Originally,
there
were
plans
to
compare
immediate
post­
intervention
dust
wipe
results
with
those
collected
at
one
year
post­
intervention;
however,
visual
assessment
"
failures"
found
at
immediate
post­
intervention
were
corrected
upon
discovery.
Therefore,
comparison
of
immediate
postintervention
dust
wipe
data
with
one
year
post­
intervention
data
is
not
appropriate.

II.
STUDY
DESIGN
AND
SAMPLE
COLLECTION
PROCEDURES
A.
Overview
of
the
Study
Design
LAAP
originally
planned
to
enroll
200
housing
units
in
the
study,
primarily
three­
bedroom,
twostory
single­
family
rowhouses
of
approximately
900
to
1200
square
feet
located
in
Baltimore
City.
One
hundred
(
100)
of
these
units
were
to
receive
the
prescribed
EA
6­
8
lead
hazard
reduction
treatments
(
LHR)
while
the
other
100
were
to
receive
the
prescribed
lead
hazard
reduction
treatments
plus
full
window
replacement
(
LHR+
W).

The
original
goal
of
200
study
units
was
increased
to
250
units
to
allow
for
an
anticipated
20
percent
attrition
rate
per
sampling/
inspection
phase.
However,
as
a
result
of
difficulties
with
attracting
rental
property
owners
to
the
program,
suitability
of
certain
units
because
of
preexisting
conditions
(
e.
g.,
structural
defects),
inability
of
some
property
owners
to
self­
finance
20
percent
of
the
lead
hazard
reduction
measures
costs,
grant
approval
delays
and
other
administrative
factors,
that
goal
was
reduced
to
180
units
and
then,
finally,
to
150
units.
In
the
end,
177
units
underwent
pre­
intervention
inspections
and
121
units
completed
post­
intervention
inspections,
prior
to
the
cut­
off
date
of
October
1,
1999
for
completing
lead
hazard
reduction
measures
in
enrolled
units.

LAAP
did
not
specify
how
the
treatments
were
to
be
carried
out,
with
the
exception
of
a
model
window
replacement
specification.
Instead,
property
owners
provided
LAAP
with
treatment
specifications
and
cost
estimates;
LAAP
inspectors
then
reviewed
the
specifications
and
costs,
and
made
adjustments
with
the
owner's
knowledge
and
consent.
All
units
were
vacant
at
the
time
of
developing
the
specifications
and
costs,
and
during
performance
of
the
intervention.

LAAP
inspectors
collected
data
from
enrolled
units
during
four
(
4)
sampling/
inspection
phases:

 
Pre­
Intervention
(
Phase
I):
within
four
(
4)
weeks
prior
to
the
property
owner
and
his/
her
contractor
performing
the
lead
hazard
control
intervention;
 
Immediate
Post­
Intervention
(
Phase
II):
immediately
after
completion
of
the
treatments
and
the
visual
inspection
(
i.
e.,
no
more
than
three
days),
as
performed
by
an
independent
inspector
hired
by
the
owner;
 
One­
year
post­
intervention
(
Phase
III):
12
months
post­
intervention
(
generally
within
6
weeks
of
the
12­
month
anniversary);
and
11/
23/
05
15
 
Two­
years
post­
intervention
(
Phase
IV):
24
months
post­
intervention
(
generally
with
6
weeks
of
the
24­
month
anniversary).

At
each
phase,
LAAP
inspectors
took
composite
dust
samples
in
all
enrolled
housing
units.
Additionally,
single
surface
samples
were
taken
in
the
LHR+
W
units.
LAAP
inspectors
also
performed
visual
assessments
for
dwelling
unit
conditions
at
each
phase.
Finally,
cost
and
concurrent
work
information
was
collected
immediately
after
the
intervention.

B.
Recruitment
and
Enrollment
Process
LAAP
was
responsible
for
the
recruitment
and
enrollment
of
housing
units
from
within
the
city
of
Baltimore.
As
an
incentive
to
enroll
units
in
the
study,
rental
property
owners
would
receive
80
percent
of
the
lead
hazard
reduction
treatment
costs
in
the
form
of
a
grant,
up
to
a
cap
of
$
10,000;
the
remaining
20
percent
would
have
to
be
self­
funded.
This
funding
was
provided
through
LAAP's
HUD
OHHLHC
Round
III
Lead
Hazard
Control
Grant.
A
LAAP
grant
coordinator
kept
track
of
owners'
contributions,
which
had
to
be
received
before
the
unit
could
go
to
settlement.

Participating
rental
property
owners
were
asked
to
enroll
two
(
2)
units
in
the
study:
a
LHR
unit
and
a
LHR+
W
unit.
A
initial
limit
of
four
(
4)
enrolled
units
per
owner
was
eventually
lifted,
although
the
goal
remained
to
keep
the
ratio
of
LHR
to
LHR+
W
units
at
one
to
one
for
each
participating
owner
and
for
the
study
overall.

In
order
to
qualify
for
the
grant,
the
property
owner
had
to
substantiate
that
the
unit
was
registered
with
MDE
(
as
per
EA
6­
8
requirements),
provide
evidence
of
fire
insurance
on
the
unit,
produce
documentation
of
acceptable
lead
treatment
specifications
and
costs
prior
to
construction,
agree
to
rent
the
enrolled
unit
to
low­
income
families
for
a
period
of
five
years,
maintain
rents
affordable
to
low­
income
families
for
the
unit,
agree
to
dust
testing
and
visual
assessment
of
the
unit
until
24
months
post­
intervention,
and
sign
a
grant
contract.
Costs
were
first
estimated
by
owners
or
their
representatives,
then
reviewed
by
LAAP
and
revised
if
spending
caps
were
exceeded
or
other
deficiencies
were
noted.

Recruitment
efforts
did
not
begin
until
late
1996.
At
that
point,
LAAP
and
Center
personnel
made
a
presentation
to
several
representatives
of
the
Property
Owners
Association
of
Greater
Baltimore.
Additionally,
a
"
notice
of
funding
availability"
was
published
and
LAAP
asked
MDE
and
a
local
childhood
lead
poisoning
prevention
advocacy
organization
for
assistance
in
directing
property
owners
to
the
program.
In
order
to
pilot
test
the
data
collection
forms
and
protocols,
four
owners
enrolled
two
(
2)
units
each,
one
LHR
and
one
LHR+
W,
for
a
total
of
eight
(
8)
pilot
study
units.
These
owners
received
full
reimbursement
of
their
treatment
costs
as
an
incentive
to
participate
in
the
pilot.

C.
Selection
Criteria
for
Enrolled
Units
In
addition
to
the
aforementioned
recruitment
and
enrollment
criteria,
a
housing
unit
had
to
be
located
in
Baltimore
City,
constructed
prior
to
1950,
vacant
at
the
time
of
the
intervention
and
structurally
sound,
as
determined
by
a
LAAP
screening
inspection,
to
qualify
for
this
study.
11/
23/
05
16
Only
pre­
1950
rental
units
were
enrolled
since
these
units
were
automatically
subject
to
the
law's
prescribed
risk
reduction
standard.

Conditions
that
could
exclude
a
unit
from
the
study
included,
but
were
not
limited
to:

 
Roof
and
plumbing
leaks;
 
Unsound
floor
or
wall
structures;
 
Non­
functional
baths
or
kitchens;
 
Non­
working
utilities,
including
plumbing,
electrical
and
heating
systems;
 
Pest
and
vermin
problems;
 
Excessive
surface
problems,
such
as
holes,
cracks,
excessive
peeling
paint
or
plaster
deterioration;
 
A
property
with
building
code
violations;
 
A
unit
with
an
existing
lead
violation
when
the
prescribed
lead
hazard
reduction
treatments
would
not
satisfy
the
violation
notice;
 
More
than
14
windows
in
the
unit;
 
Size
of
the
dwelling
unit
(
e.
g.,
greater
than
1200
square
feet);
 
Units
deemed
to
be
a
threat
to
the
safety
of
LAAP
inspectors;
 
Rents
to
be
charged
post­
intervention
that
would
exceed
$
450
per
month;
and
 
Failure
to
register
the
unit
with
MDE,
in
accordance
with
the
EA
6­
8
requirements.

Eligibility
of
owners
for
grant
funds
and
units
for
enrollment
in
the
study
was
sometimes
conditional
based
on
the
screening
inspection.
On
occasion,
LAAP
was
able
to
assist
owners
with
structural
issues
by
allowing
up
to
$
500
towards
increasing
structural
integrity
or
fixing
leaks
(
e.
g.,
toward
a
new
roof).
If
an
owner
corrected
one
or
more
observed
deficiencies,
the
unit
might
be
accepted
and
the
owner
would
then
be
required
to
prepare
detailed
work
specifications
for
the
intervention,
closely
following
the
prescribed
lead
hazard
reduction
treatments.
After
a
final
pre­
intervention
walk­
through
inspection
by
LAAP,
the
unit
would
be
approved
to
proceed
with
construction.
Information
on
pre­
program
repairs
was
not
collected
as
part
of
this
research
study.

D.
Data
Collection
Forms
and
Protocols
LAAP
inspectors
were
trained
by
Center
personnel
to
follow
the
data
collection
protocols
and
complete
the
eight
(
8)
data
collection
forms
used
for
this
study
(
see
Table
1
on
page
17
and
Appendix
A).
11/
23/
05
17
Table
1:
Data
Collection
Form
Type
and
Usage
Form
Name
Phase
I:
Pre
Intervention
Phase
II:
Immediate
Post
Intervention
Phase
III:
12
Mo.
Post
Intervention
Phase
IV:
24
Mo.
Post
Intervention
Form
01
 
Building
and
Dwelling
Unit
Information
X
Form
02
 
Baseline
Dwelling
Condition
X
Form
03
 
Sketch
of
Sampling
Locations
X
Form
04A
 
Dust
Sample
Collection
(
Composite)
X
X
X
X
Form
04B
 
Dust
Sample
Collection
(
Single
Surface)
X
X
X
X
Form
05
 
Post
Treatment
Assessment/
Visual
Inspect.
X
X
X
Form
06
 
Follow­
Up
Questions
X
X
Form
07
 
De­
Enrollment
Form
*
a
*
a
*
a
*
a
aForm
07
was
only
completed
if
a
unit
was
de­
enrolled
from
the
study.

Form
01
 
Building
and
Dwelling
Unit
Information
was
designed
to
collect
basic
information
about
the
type
of
building
(
e.
g.,
single
detached,
row
house,
duplex,
etc.),
year
of
construction,
estimated
market
value
(
obtained
directly
from
the
property
owner),
number
of
rooms,
bedrooms,
windows
and
approximate
square
feet
of
living
space.
This
form
was
completed
once:
during
the
pre­
intervention
phase
(
i.
e.,
Phase
I).

Form
02
 
Baseline
Dwelling
Conditions
was
designed
to
collect
information
about
obvious
exterior
building
deterioration
on
roofs,
gutters,
downspouts,
chimneys,
walls
and
siding,
windows
and
doors,
porches
and
steps
and
the
foundation,
and
whether
the
exterior
ground
was
contaminated
with
paint
chips.
It
also
was
used
to
collect
information
about
interior
dwelling
unit
deterioration
on
walls,
ceilings,
doors
and
trim,
floors,
heating/
cooling
and
plumbing
systems,
interior
damage
from
roof
leakage
and
paint
condition.
This
form
was
completed
once:
during
the
pre­
intervention
phase
(
i.
e.,
Phase
I).

Form
03
 
Sketch
of
Sampling
Locations
was
designed
to
record
room
and
dust
sampling
locations
on
a
sketch
of
each
interior
floor,
which
included
room
layouts.
It
was
also
used
to
code
room
numbers
for
the
purpose
of
room
identification
on
other
forms.
Even
though
this
form
was
a
reference
source
for
room
and
sample
locations
during
subsequent
inspection/
sampling
phases,
it
was
completed
only
once:
during
the
pre­
intervention
phase
(
i.
e.,
Phase
I).

Form
04A
 
Dust
Sample
Collection
(
Composite)
was
designed
to
record
a
variety
of
data
for
composite
dust
samples,
including
type
of
surface
sampled
(
e.
g.,
smooth
floors,
interior
window
11/
23/
05
18
sills,
window
troughs),
room
and
sample
locations,
dimensions
of
the
surfaces
sampled,
total
area
sampled,
sample
number,
total
micrograms
of
lead
per
composite
sample
and
average
lead
loading
results
measured
as
micrograms
of
lead
per
square
foot
(
µ
g/
ft2).
The
form
also
recorded
observations
regarding
unit
cleanliness
and
occupancy
status.
It
was
completed
during
each
of
the
four
sampling
phases
in
all
units.

Form
04B
 
Dust
Sample
Collection
(
Single
Surface)
was
designed
to
record
data
similar
to
Form
04A,
except
that
it
applied
solely
to
single
surface
samples.
Originally
intended
strictly
for
LHR+
W
units,
Form
04B
was
also
used
for
a
number
of
LHR
units
as
well.
The
form
was
completed
during
each
of
the
four
sampling
phases.

Form
05
 
Post
Treatment
Assessment/
Visual
Inspection,
a
two­
page
form,
was
designed
to
collect
immediate
post­
intervention
information
on
whether
the
unit
was
LHR
or
LHR+
W
and,
if
the
latter,
how
many
windows
were
replaced.
It
also
captured
information
on
the
type
of
contractor
performing
the
work
(
for­
profit,
nonprofit
or
property
owner),
the
cost
of
the
LHR,
the
cost
of
window
replacement
(
if
applicable),
the
estimated
turnover
cost
of
the
unit
had
it
not
been
subject
to
EA
6­
8
lead
hazard
reduction
requirements,
concurrent
work
(
if
any)
performed
by
the
owner,
and
dates
the
work
started
and
was
completed.
Form
05
also
contained
a
list
of
10
questions
designed
to
determine
whether
there
was
visual
evidence
that
the
unit
had
met
the
prescribed
lead
hazard
reduction
treatments.
For
any
observed
"
failure"
of
a
prescribed
treatment,
page
2
of
the
form
was
used
to
record
the
type
of
"
failure,"
room
and
location,
quantity
and
units
of
measurement,
and
LAAP
inspector's
notes
on
the
"
failure."
This
form
was
completed
three
times:
immediate
post­
intervention,
12
months
post­
intervention
and
24
months
post­
intervention
(
i.
e.,
Phases
II,
III
and
IV).

Form
06
 
Follow­
Up
Questions
was
designed
to
collect
information
from
the
property
owner
about
any
activities
that
occurred
in
the
unit
in
the
12
and
24­
month
periods
after
the
intervention.
This
information
included
whether
any
additional
lead
hazard
reduction
treatments
were
conducted
in
the
unit,
whether
the
unit
had
turned
over
and
how
many
times,
whether
a
tenant
called
to
report
a
newly
identified
lead
hazard,
whether
the
owner
was
notified
of
a
resident
child
with
an
elevated
blood
lead
level,
and
whether
a
"
qualified
offer"
was
made
to
any
resident
child
with
a
blood
lead
level
equal
to
or
in
excess
of
25
µ
g/
dL.

E.
Dust
Lead
Sampling
Protocols
LAAP
inspectors
collected
all
dust
wipe
samples
for
this
study.
These
inspectors
were
trained
and
certified
as
both
lead­
based
paint
inspectors
and
lead­
based
paint
risk
assessors
under
applicable
MDE
regulations.

The
dust
sampling
protocol
was
designed
to
yield
information
on
dust
lead
loadings
from
floors
(
including
bare
and
carpeted
floors),
interior
window
sills
and
window
troughs
(
also
known
as
window
wells
or
exterior
window
sills)
using
both
composite
and
single
surface
sampling.
Samples
were
collected
during
all
four
sampling/
inspection
phases.
Originally,
composite
dust
samples
were
to
have
been
taken
in
all
enrolled
units,
with
single
surface
samples
being
collected
only
in
the
LHR+
W
units.
However,
because
of
problems
with
the
initial
analytical
laboratory
selected
for
this
study,
single
surface
samples
were
also
taken
as
a
precautionary
measure
in
11/
23/
05
19
LHR
units
first
inspected
during
the
period
from
approximately
November
1,
1997
until
February
1,
1998,
when
a
new
laboratory
was
retained
to
replace
the
original
laboratory.

The
HUD
wipe
sampling
method,
utilizing
moist
towellettes
placed
in
50­
milliter
(
ml)
hard
shell
plastic
centrifuge
tubes,
was
initially
employed.
Later,
after
the
new
analytical
laboratory
was
retained,
125
ml
glass
jars
with
plastic
lids
were
used
to
hold
the
towellettes.
Plastic
templates
measuring
8
inches
by
9
inches
(
72
square
inches
or
one­
half
square
foot
total)
were
secured
to
bare
or
carpeted
floor
sampling
areas.
Additionally,
plastic
templates
measuring
2
inches
by
18
inches
(
36
square
inches
or
one­
quarter
square
foot
total)
were
secured
to
interior
window
sills,
when
possible.
All
plastic
templates
were
carefully
cleaned
between
samples.
Masking
tape
was
used
to
define
wipe
sampling
areas
within
window
troughs
and
on
interior
window
sills
if
templates
were
impractical.
When
masking
tape
was
used
to
define
composite
sub­
sample
areas,
care
was
taken
so
that
each
sub­
sample
area
was
within
plus
or
minus
15
percent
of
other
subsample
areas
for
a
given
composite
sample
type.

Composite
dust
samples
contained
no
less
than
two
and
no
more
than
four
sub­
samples
and
were
collected
from
the
following
rooms,
according
to
surface
type:

Bare
floor
sub­
sample
locations:

1.
Entryway
(
just
inside
the
exterior
doorway,
slightly
left
of
center);
2.
Living
room/
principal
play
area
of
children
(
slightly
left
of
center
of
the
exterior
doorway);
3.
Kitchen
(
slightly
left
of
center
of
the
interior
doorway);
and
4.
Bedroom
#
1/
smallest
bedroom
(
slightly
left
of
center
of
the
doorway).

If
any
of
the
above
rooms
had
wall­
to­
wall
carpeting,
a
separate
composite
sample
was
created,
with
sub­
samples
being
collected
from
these
carpeted
rooms.
(
As
discussed
later,
very
few
units
had
carpeted
floors.
Therefore,
sample
results
for
carpeted
floors
are
not
further
discussed
in
this
report.)

Window
sub­
sample
locations
(
sill
and
troughs
as
separate
composite
samples):

1.
Living
room/
principal
play
area;
2.
Kitchen;
3.
Bedroom
#
1/
smallest
bedroom;
and
4.
Bedroom
#
2/
next
smallest
bedroom.

Window
sub­
samples
were
collected
from
the
left
half
(
looking
out)
of
each
window
unit.
Only
one
window
per
room
was
sampled,
with
the
selected
window
being
sampled
during
subsequent
sampling/
inspection
phases.
11/
23/
05
20
Single
surface
samples
were
collected
from
the
rooms
listed
below,
according
to
surface
type.
All
single
surface
samples
were
collected
to
the
right
of
each
defined
area
as
one
looks
into
a
given
room
or
out
of
a
given
window.

Floor
sample
locations
(
bare
or
carpeted):

1.
Entryway
(
just
inside
the
exterior
doorway);
2.
Living
room/
principal
play
area
(
just
inside
the
exterior
doorway);
3.
Kitchen
(
just
inside
the
interior
doorway);
4.
Bedroom
#
1/
smallest
bedroom
(
just
inside
the
interior
doorway);
and
5.
Bedroom
#
2/
next
smallest
bedroom
(
just
inside
the
interior
doorway).

Interior
window
sill
sample
locations:

1.
Kitchen;
and
2.
Bedroom
#
1/
smallest
bedroom.

Window
trough
locations:

1.
Living
room/
principal
play
area;
and
2.
Bedroom
#
2/
next
smallest
bedroom.

Using
Forms
04A
and
04B,
LAAP
inspectors
completed
laboratory
submittal
forms,
sent
the
forms
and
samples
out
to
the
analytical
laboratory
and
then
transcribed
dust
lead
loading
results
onto
the
forms
after
receiving
this
information
from
the
lab.

LAAP
inspectors
also
collected
field
blank
samples
of
both
the
composite
and
single
surface
type
after
sampling
the
final
dwelling
unit
of
the
day.
Additionally,
dust­
spiked
samples
prepared
by
the
University
of
Cincinnati,
using
the
National
Institute
of
Standards
and
Technology
(
NIST)
Standard
Lead
Paint
Dust,
were
inserted
into
the
sampling
stream
by
the
inspectors
at
a
rate
of
one
for
every
50
single
surface
samples
and
one
for
every
20
composite
samples
collected.
Each
composite
field
blank
sample
and
spike
sample
was
comprised
of
four
wipes.
The
laboratory
was
blinded
to
all
spike
samples.

F.
Visual
Assessment
Protocols
In
addition
to
collecting
composite
and
single
surface
dust
wipe
samples,
LAAP
inspectors
performed
visual
assessments
during
all
four
sampling/
inspection
phases.

Pre­
intervention
visual
assessments
were
accomplished
by
observing
existing
building
and
dwelling
unit
conditions
for
signs
of
deterioration
followed
by
completing
the
questions
on
Forms
01
and
02.
Also,
diagrams
of
the
dwelling
unit
floor
plan
were
prepared
using
Form
03.
Immediate
post­
intervention
visual
assessments
required
LAAP
inspectors
to
closely
inspect
the
building
and
dwelling
unit
for
completion
of
the
prescribed
lead
hazard
reduction
treatments.
Utilizing
Form
05,
the
inspectors
would
visually
assess
each
room
in
the
unit
and
certain
11/
23/
05
21
portions
of
the
building
exterior.
Any
observed
treatment
"
failures"
would
be
recorded
on
the
form
along
with
the
room
in
which
the
"
failure"
was
found
and
the
amount
of
and
unit
of
measure
for
the
"
failure"
(
e.
g.,
10
square
inches
of
chipping,
flaking
or
peeling
paint.

Once
lead
hazard
reduction
interventions
were
complete,
and
in
anticipation
of
having
an
independent
inspector
perform
the
prescribed
visual
inspection,
participating
property
owners
and/
or
their
contractors
would
contact
LAAP
inspectors
and
advise
them
that
a
unit
was
ready
for
dust
sampling,
visual
assessment
and
a
final
construction
walk­
through.
The
LAAP
inspectors
would
then
collect
samples
and
perform
the
visual
assessment,
usually
immediately
after
the
independent
inspector
had
inspected
the
unit
for
compliance
with
the
prescribed
treatments.

Because
the
program
paid
for
up
to
80
percent
of
all
pre­
approved
intervention
costs,
LAAP
inspectors
had
to
perform
the
walk­
through
after
the
dust
sampling
and
visual
assessment
to
determine
if
the
original
intervention
specifications
had
been
completed.
If
any
specification
was
incomplete
or
improperly
performed,
the
owner
and/
or
contractor
was
instructed
to
finish
the
work
before
LAAP
would
release
its
final
payment
for
the
intervention.

By
following
this
sequence
and
protocol,
LAAP
inspectors
sought
to
avoid
influencing
unit
conditions
that
might
improve
dust
sampling
and
visual
assessment
results.
The
overarching
objective
was
to
collect
samples
and
to
visually
assess
the
same
conditions
that
the
independent
visual
inspector
would
have
observed
at
the
time
of
his/
her
compliance
inspection.
However,
in
a
handful
of
units,
the
LAAP
inspectors
did
not
proceed
with
the
sampling
and
visual
assessment
when
an
obvious
treatment
had
been
missed
(
e.
g.,
stripping
and
repainting
or
replacing
a
window
sill).
In
those
cases,
the
inspectors
would
return
after
such
missed
treatments
were
completed,
obtain
dust
samples
and
visually
assess
the
unit.

As
noted
above,
if
the
LAAP
inspectors
found
any
specification
that
was
incomplete
or
improperly
performed,
the
owner
and/
or
contractor
was
instructed
to
finish
the
work
before
LAAP
would
release
its
final
payment
for
the
intervention.
Therefore,
the
one­
year
and
twoyears
post­
intervention
inspections
were
checks
of
how
well
treatments
remained
in
place
12
and
24
months,
respectively,
following
the
completion
of
all
immediate
post­
intervention
work.

III.
LABORATORY
ANALYSIS
PROCEDURES
As
a
prerequisite,
the
analytical
laboratory
selected
to
analyze
dust
samples
for
this
study
had
to
be
recognized
by
EPA
as
participating
in
the
National
Lead
Laboratory
Accreditation
Program
(
NLLAP).
The
lab
had
to
show
evidence
of
being
proficient
in
lead
analysis
under
the
Environmental
Lead
Proficiency
Analytical
Testing
Program
(
ELPAT).
All
dust
samples
had
to
be
analyzed
for
total
lead
as
required
in
the
HUD
Guidelines
for
the
Evaluation
and
Control
of
Lead­
Based
Paint
Hazards
in
Housing.
Because
of
problems
detected
in
the
analysis
of
the
composite
samples
by
the
first
laboratory
(
designated
"
Lab
A"),
a
second
laboratory
("
Lab
B")
was
hired
to
replace
Lab
A
as
the
study
laboratory
(
see
Appendix
B
for
details).
Analytical
procedures
for
each
laboratory
are
provided
in
Appendix
B.
Lab
B
was
used
for
the
analysis
of
all
one­
year
and
two­
years
post­
intervention
data.
11/
23/
05
22
IV.
QUALITY
ASSURANCE
AND
QUALITY
CONTROL
A.
Laboratory
Quality
Assurance
and
Quality
Control
1.
Quality
Control
Spiked
Dust
Wipe
Sample
Results
As
noted
above,
LAAP
inspectors
periodically
submitted
single
surface
and
composite
spiked
dust
wipe
samples
(
using
the
NIST
standard)
to
the
study
laboratory.
Early
in
the
project,
LAAP
and
the
Center
noted
deficiencies
in
composite
spike
recoveries
and
other
observations
that
led
LAAP
to
replace
Lab
A
with
Lab
B.
Samples
potentially
associated
with
Lab
A's
unacceptable
composite
spike
results
were
excluded
from
all
analyses
presented
in
this
report.
No
other
samples
were
excluded
on
the
basis
of
QC
spike
sample
results.
Spike
sample
submittal
procedures,
as
well
as
details
concerning
the
spike
results,
are
provided
in
Appendix
B.

2.
Quality
Control
Field
Blank
Dust
Wipe
Sample
Results
As
noted
in
Section
III.
E,
composite
and
single
surface
field
blank
dust
wipe
samples
were
prepared
by
LAAP
inspectors
and
submitted
to
the
laboratory
after
sampling
the
final
dwelling
unit
of
the
day.

The
presence
of
detectable
lead
in
a
field
blank
sample
could
suggest
several
potential
problems:
sample
collection
may
have
been
poorly
conducted
(
e.
g.,
lead
may
in
fact
have
been
introduced
into
a
sample
container
from
an
inspector's
glove,
a
contaminated
sample
container,
or
from
other
sources
not
under
experimental
control);
contamination
and
poor
sample
handling
may
have
occurred
at
the
laboratory;
or
there
may
have
been
errors
in
laboratory
measurement
procedures.
For
both
Lab
A
and
Lab
B,
composite
and
single
surface
field
blank
results
were
considered
acceptable.
Therefore,
no
composite
or
single
surface
samples
were
excluded
from
this
report
based
on
field
blank
results.
Details
concerning
field
blank
results
are
provided
in
Appendix
B.

3.
Procedures
for
Dust
Wipe
Sample
Results
that
Were
At
or
Below
Laboratory
Report
Limits
A
high
percentage
of
both
composite
and
single
surface
dust
wipe
sample
results
were
below
the
reporting
limits
of
both
Lab
A
and
Lab
B.
Lab
A's
reporting
limits
varied
from
5
µ
g/
sample
to
24
µ
g/
sample,
Lab
B's
single
surface
sample
reporting
limit
was
10
µ
g/
sample,
and
Lab
B's
composite
sample
reporting
limit
was
20
µ
g/
sample.

Because
one
of
the
most
important
measures
of
intervention
outcome
 
the
change
in
dust
lead
loadings
over
time
 
could
be
at
or
below
the
laboratory's
reporting
limits,
Lab
B
supplied
the
actual
machine
values
for
composite
and
single
surface
samples
with
lead
content
below
reporting
limits.
For
the
few
cases
in
which
Lab
B
reported
a
machine
value
of
zero
(
i.
e.,
the
sample's
lead
content
could
not
be
distinguished
from
the
machine's
"
background
noise"),
the
zero
values
were
replaced
with
the
laboratory's
reporting
limit
divided
by
the
square
root
of
two.
Lab
A
did
not
supply
actual
machine
values;
sample
results
reported
by
this
laboratory
to
be
at
or
below
reporting
limits
were
also
replaced
with
a
value
calculated
by
taking
Lab
A's
reporting
11/
23/
05
23
limit
and
dividing
it
by
the
square
root
of
two.
Based
on
data
shown
in
Table
2,
a
total
of
133
results
underwent
this
conversion.

Table
2:
Dust
Wipe
Sample
Results
Below
Reporting
Limits
or
Reported
As
Zeroa
Surface
Typeb
Pre­
Intervention
Samples
#
Samples/
Total
(%
Samples)
Immediate
Post­
Intervention
Samples
#
Samples/
Total
(%
Samples)
One­
Year
Post­
Intervention
Samples
#
Samples/
Total
(%
Samples)
Two­
years
postintervention
Samples
#
Samples/
Total
(%
Samples)
Bare
Floor:
Composite
Single
Surface
0/
148
(
0.0%)
1/
467
(
0.2%)
3/
108
(
3%)
49/
373
(
13%)
3/
68
(
4%)
11/
168
(
6%)
1/
24
(
4%)
3/
64
(
5%)

Window
Sill:
Composite
Single
Surface
1/
152
(
0.7%)
1/
196
(
0.5%)
4/
116
(
3%)
28/
164
(
17%)
0/
73
(
0%)
8/
76
(
10%)
0/
24
(
0%)
0/
26
(
0%)

Window
Trough:
Composite
Single
Surface
1/
152
(
0.7%)
0/
191
(
0.0%)
4/
116
(
3%)
13/
157
(
8%)
0/
72
(
0%)
1/
75
(
1%)
0/
24
(
0%)
0/
26
(
0%)

Total
All
Surface
Types
and
Sample
Types
4/
1306
(
0.3%)
101/
1034
(
10%)
23/
532
(
4%)
4/
188
(
2%)
aLab
A
recorded
results
below
reporting
limits
and
Lab
B
recorded
results
as
zero.
bBecause
very
few
units
had
carpeted
floors,
sample
results
for
carpeted
floors
were
excluded
from
this
study.

B.
Performance
Audit
of
Field
Data
Collection
Activities
On
six
different
occasions,
Center
staff
observed
LAAP
inspectors
in
the
field
collecting
single
surface
and
composite
dust
wipe
samples,
performing
visual
assessments
and
completing
the
data
collection
forms.
LAAP
inspectors
were
critiqued
for
their
adherence
to
the
data
collection
protocols
and
their
dust
wipe
sampling
technique.
Although
the
LAAP
inspectors
generally
performed
to
expectations,
the
Center's
representatives
provided
on­
site
guidance
and
written
recommendations
to
improve
sampling
and
data
collection.

During
several
visits
to
the
LAAP
office,
Center
staff
reviewed
unit
files
and
other
project
documentation
for
completeness
and
adherence
to
the
study
protocols.
Center
representatives
also
made
on­
site
and
written
recommendations
for
improvements
during
these
visits,
particularly
regarding
tracking
of
follow­
up
inspection
assignments.

C.
Data
Audit
and
Data
Completeness
The
Center
was
responsible
for
auditing
all
data
submitted
by
LAAP
for
correctness,
completeness
and
adherence
to
the
data
collection
protocols.
After
completion
by
LAAP
inspectors,
data
collection
forms
(
and
laboratory
analysis
reports)
were
submitted
to
the
Center
for
initial
review
and,
if
appropriate,
data
entry
and
processing.
Center
staff
reviewed
each
form
for
missing
and/
or
unclear
answers
and
data,
use
of
proper
answer
codes
and
values,
insertion
of
correct
laboratory
results
and
calculation
of
dust
lead
loading,
and
consistency
of
room
number
assignments
and
sampling
locations
across
the
different
sampling/
inspection
phases.
When
11/
23/
05
24
errors
and/
or
missing
information
were
detected,
the
Center
would
return
forms
to
LAAP,
with
detailed
instructions
for
correction
or
clarification
unless
questions
could
be
resolved
easily
by
telephone.
LAAP
would
then
make
revisions
and
return
the
corrected
forms
to
the
Center,
which
then
performed
a
final
audit
before
preparing
the
forms
for
data
entry.

V.
DATA
PROCESSING
AND
STATISTICAL
ANALYSIS
PROCEDURES
A.
Data
Entry
and
Processing
The
Center
was
responsible
for
all
data
entry
and
data
processing.
An
electronic
facsimile,
as
well
as
the
actual
hard­
copy
version,
of
each
data
collection
form
was
created
in
Jetform's
FormFlow
software
program.
After
receiving
the
completed
field
forms
from
LAAP,
the
Center's
Data
Coordinator
would
enter
data
from
the
field
forms
directly
into
the
electronic
version
of
the
forms
in
FormFlow.
A
hard­
copy
of
the
completed
form
was
then
printed
out
and
reviewed
for
errors
and/
or
missing
information
by
the
Data
Coordinator
and
other
Center
staff.
After
corrections
were
made
and
entered
into
the
electronic
version
of
the
form,
a
revised
hard
copy
of
the
form,
along
with
the
LAAP
field
form,
was
placed
in
a
file
folder
created
for
each
enrolled
unit
and
stored
at
the
Center.

B.
Data
Summary
and
Statistical
Analyses
As
noted,
all
study
data
were
transmitted
into
and
maintained
in
FormFlow.
Statistical
analyses
of
these
data
were
performed
using
a
SAS
Institute
program
(
Versions
6.12
and
8.0),
which
also
generated
reports
and
tables.
Finally,
aside
from
SAS
reports,
the
Crystal
Reports
software
package
was
used
to
generate
other
reports
by
directly
downloading
dBase
data
from
FormFlow.

VI.
ENROLLMENT
RESULTS
BY
PHASE
A
total
of
177
units
underwent
pre­
intervention
inspections
and
121
units
completed
immediate
post­
intervention
inspections
prior
to
the
cut­
off
date
of
October
1,
1999
for
completing
lead
hazard
reduction
measures
in
enrolled
units.
These
121
units
were
followed
for
the
12
and
24
month
post­
intervention
inspections,
with
a
total
of
73
having
complete
one­
year
postintervention
data
collected
and
a
total
of
24
having
complete
two­
years
post­
intervention
data.
The
decrease
in
units
between
immediate­
and
one­
year
post­
intervention
was
primarily
due
to
the
fact
that
tenants
did
not
move
into
units
until
after
Phase
II,
meaning
that
tenants
were
likely
unaware
of
the
study
until
LAAP
inspectors
visited
the
occupied
unit
at
one­
year
postintervention
Tenants'
lack
of
familiarity
with
the
project
often
meant
that
they
were
reluctant
to
allow
access
for
one­
year
and
two­
years
post­
intervention
sampling.
Other
problems
included
the
growing
number
of
conflicting
duties
for
the
LAAP
inspectors
over
the
lengthy
period
of
this
study
and
a
temporary
halt
in
the
data
collection
while
immediate
post­
intervention
results
underwent
review.

Because
so
few
units
had
complete
two­
years
post­
intervention
data,
the
major
part
of
this
report
necessarily
focuses
on
the
results
of
the
first
three
phases
to
meet
the
study
objectives.
A
brief
description
of
the
two­
years
post­
intervention
results
for
the
24
units
that
had
24
month
postintervention
inspections
is
presented
in
Section
X.
11/
23/
05
25
VII.
PRE­
INTERVENTION
RESULTS
A.
Results
of
Visual
Assessments
of
Baseline
Characteristics
and
Conditions
A
total
of
177
units
(
91
LHR
and
86
LHR+
W)
underwent
complete
pre­
intervention
visual
assessments
using
Forms
01,
02
and
03.
However,
pre­
intervention
dust
testing
was
often
delayed
until
the
property
owner's
project
specifications
and
grant
application
had
been
approved
and
a
final
pre­
construction
"
walk­
through"
was
completed.
Because
some
applications
were
not
approved,
this
delay
resulted
in
pre­
intervention
composite
dust
samples
being
obtained
from
148
of
the
177
units
(
74
LHR
and
74
LHR+
W)
and
single
surface
dust
samples
from
98
units
(
25
LHR
and
73
LHR+
W).
The
number
of
LHR
units
having
single
surface
samples
collected
was
much
less
than
that
of
the
LHR+
W
units
because,
as
discussed
in
Section
II.
E,
single
surface
samples
were
originally
to
be
collected
only
from
LHR+
W
units.
Single
surface
samples
were
collected
from
LHR
units
for
only
a
short
period
of
time
when
laboratory
analysis
of
composite
samples
was
being
investigated
by
the
Center
and
LAAP
(
see
Appendix
B).

In
order
to
more
fully
characterize
initial
building
and
dwelling
unit
conditions,
this
section
summarizes
physical
characteristics
and
baseline
condition
data
for
the
177
units
for
which
such
data
were
collected.
Appendix
C
contains
a
detailed
presentation
of
physical
characteristics
for
each
enrolled
unit.
Appendix
D
presents
a
complete
list
of
baseline
exterior
conditions
for
each
building
and
of
baseline
interior
conditions
for
each
unit.
Complete
pre­
intervention
dust
lead
loading
results
for
composite
and
single
surface
samples
are
provided
in
Appendices
E
and
F,
respectively.

1.
Baseline
Physical
Characteristics
A
Chi­
square
test
was
used
to
determine
if
there
was
a
relationship
between
pre­
intervention
housing
characteristics
and
conditions
(
i.
e.,
variables)
and
assigned
treatment
groups
(
i.
e.,
LHR
versus
LHR+
W
units).
Such
associations
may
indicate
differences
in
the
baseline
characteristics
and
conditions
of
LHR
versus
LHR+
W
units;
these
differences
had
to
be
taken
into
account
before
the
effectiveness
of
the
lead
hazard
reduction
treatments
could
be
fully
evaluated.
Variables
having
a
significant
association
with
treatment
group
are
summarized
below.

Building
Type.
The
majority
(
71
percent)
of
the
residential
buildings
initially
enrolled
in
the
study
were
single­
family
rowhouses,
which
are
common
in
Baltimore
City.
Building
type
was
not
significantly
associated
with
the
assigned
treatment
group.
Other
enrolled
building
types
and
their
relative
percentages
as
part
of
total
enrollment
are
displayed
in
Figure
1.
11/
23/
05
26
Figure
1:
Type
of
Building
 
Pre­
Intervention
Assessment
Source:
Form
01
Note:
Based
on
Chi­
square
testing,
building
type
not
significantly
associated
with
assigned
treatment
group.

Building
Construction
Period.
Many
of
the
residential
buildings
in
Baltimore
City,
including
84
percent
of
those
initially
enrolled
in
the
study,
were
constructed
70
or
more
years
ago.
The
largest
percentage
of
enrolled
buildings
(
45
percent)
was
those
built
between
1920
and
1929.
Construction
period
was
not
significantly
associated
with
the
assigned
treatment
group.
The
age/
period
of
construction
for
all
enrolled
buildings
is
displayed
in
Figure
2.

Figure
2:
Year
Building
Constructed
 
Pre­
Intervention
Assessment
Note:
Based
on
Chi­
square
testing,
construction
period
not
significantly
associated
with
assigned
treatment
group.
70
8
10
3
3
3
2
72
8
6
5
6
2
1
71
8
8
4
5
3
2
0
10
20
30
40
50
60
70
80
Rowhouse
More
than
4
units
in
bld.
Triplex
Duplex
Four­
plex
Two
flats
Single
detached
Percentage
of
Units
LHR
Units
LHR+
W
Units
All
Units
LHR
Units
16%

22%

43%
11%
8%

Pre
1910
1910­
1919
1920­
1929
1930­
1939
1940­
1949
LHR+
W
Units
15%

23%

48%
7%
7%

Pre
1910
1910­
1919
1920­
1929
1930­
1939
1940­
1949
11/
23/
05
27
Other
Dwelling
Unit
Baseline
Conditions.
In
order
to
identify
important
characteristics
about
initially
enrolled
dwelling
units,
LAAP
inspectors
collected
information
about
estimated
real
estate
market
value,
total
number
of
rooms
and
bedrooms,
number
of
windows,
including
unpainted
ones
(
e.
g.,
vinyl
and
aluminum
replacement
windows),
and
the
size
of
the
interior
living
space.
A
t­
test
was
used
to
test
for
significant
differences
in
the
mean
values
for
LHR
and
LHR+
W
units.
Generally,
the
LHR
and
LHR+
W
units
possessed
similar
characteristics,
the
mean
values
for
which
are
displayed
in
Table
3.
However,
the
total
number
of
windows
in
a
dwelling
unit
was
marginally
associated
with
the
assigned
treatment
group
(
p­
value
=
0.0641),
with
LHR+
W
units
having
slightly
more
windows.
The
number
of
unpainted
windows
was
also
significantly
associated
with
the
treatment
group
assignment
(
p­
value=
0.0088),
the
mean
being
higher
for
LHR
units.

Table
3:
Dwelling
Unit
Characteristics
 
Pre­
Intervention
Assessment
Dwelling
Unit
Characteristic
(
91
LHR
units,
86
LHR+
W
units)
Mean
(
std.
deviat.)
for
LHR
Units
Mean
(
std.
deviat.)
for
LHR+
W
Units
Estimated
market
value
of
dwelling
unita
(
mean
rounded
to
nearest
$
100)
$
22,500
($
10,900)
$
23,000
($
12,000)

#
of
rooms
in
dwelling
unita
6
(
2)
7
(
2)

#
of
bedrooms
in
dwelling
unita
3
(
1)
3
(
1)

#
of
windows
in
dwelling
unitb
9
(
3)
10
(
4)

#
of
unpainted
windows
in
dwelling
unitc
(
vinyl,
aluminum)
1.5
(
3)
0.5
(
2)

Approximate
square
feet
of
living
spacea
1,100
ft2
(
340)
1,100
ft2
(
380)

Source:
Form
01
aThese
building
condition
items
not
associated
with
assigned
treatment
group
based
on
t­
test.
bTotal
number
of
windows
in
unit
marginally
associated
with
assigned
treatment
group
based
on
t­
test
(
p=
0.0641).
cNumber
of
unpainted
windows
significantly
associated
with
assigned
treatment
group
based
on
t­
test
(
p=
0.0088)

2.
Baseline
Exterior
Building
and
Interior
Unit
Condition
Baseline
Exterior
Building
Deterioration.
In
order
to
assess
the
level
of
pre­
existing
building
deterioration,
LAAP
inspectors
performed
an
exterior
visual
assessment
of
enrolled
buildings,
checking
for
seven
(
7)
items
considered
to
serve
as
obvious
signs
of
deterioration
(
see
Table
4
on
page
28).
In
general,
there
were
minimal
observed
differences
between
the
LHR
and
LHR+
W
units
with
respect
to
most
of
these
exterior
deterioration
variables.
For
example,
26
percent
of
the
LHR
units
had
roofs,
gutters
and
downspouts
that
were
either
missing,
broken,
cracked
or
had
holes
and
21
percent
of
the
LHR+
W
units
had
similar
deterioration.
LHR+
W
units
did,
however,
differ
from
LHR
units
for
one
particular
item
of
exterior
deterioration:
30
percent
of
the
LHR+
W
units
versus
13
percent
of
the
LHR
units
had
two
or
more
windows
or
doors
that
were
broken,
boarded­
up
or
missing.
This
condition
was
highly
associated
with
the
assigned
11/
23/
05
28
treatment
group
(
p­
value
=
0.006).
Given
that
the
LHR+
W
units
were
slated
to
receive
new
windows
as
part
of
the
study,
this
could
explain
the
difference
between
the
two
treatment
groups.
Units
not
randomly
assigned
to
treatment
groups
could
affect
costs.

Table
4:
Units
with
Exterior
Building
Deterioration
 
Pre­
Intervention
Assessment
Types
of
Exterior
Building
Deterioration
(
91
LHR
units,
86
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

Roofs,
gutters,
downspouts
 
missing,
broken,
holes,
cracks
24
(
26%)
18
(
21%)
42
(
24%)

Chimney
 
cracked,
bricks
loose
or
missing,
unstable,
out
of
plumb
2
(
2%)
2
(
2%)
4
(
2%)

Walls
and
siding
 
large
cracks
or
holes,
boards
or
shingles
broken
or
missing
16
(
18%)
14
(
16%)
30
(
17%)

Windows
and
doors
 
 
two
windows
or
doors
broken,
missing,
boarded
upa
12
(
13%)
26
(
30%)
38
(
21%)

Porch
or
steps
 
major
elements
broken,
missing,
out
of
plumb
12
(
13%)
11
(
13%)
23
(
13%)

Foundation
 
major
visible
cracks,
missing
materials,
unsound
2
(
2%)
2
(
2%)
4
(
2%)

Evidence
of
exterior
ground
contaminated
with
paint
chips
33
(
36%)
30
(
35%)
63
(
36%)

Source:
Form
02
aBroken/
missing
windows
&
doors
significantly
associated
with
assigned
treatment
group
based
on
Chi­
square
test
(
p=
0.006).
No
other
exterior
building
deterioration
item
was
associated
with
treatment
group.

The
number
of
observed
items
of
exterior
building
deterioration
in
LHR
units
was
not
appreciably
different
from
the
number
observed
in
LHR+
W
units
(
see
Table
5).
There
was
no
significant
association
between
the
number
of
exterior
deteriorations
and
the
assigned
treatment
group
(
p­
value=
0.95).

Table
5:
Number
of
Exterior
Building
Deterioration
Items
 
Pre­
Intervention
Assessment
#
Items
of
Exterior
Deterioration
per
Building
(
91
LHR
units,
86
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)
"
0"
items
of
exterior
deterioration
36
(
40)
29
(
34)
65
(
37)
"
1"
item
of
exterior
deterioration
28
(
31)
30
(
35)
58
(
33)
"
2"
items
of
exterior
deterioration
14
(
15)
15
(
17)
29
(
16)
"
3"
items
of
exterior
deterioration
9
(
10)
7
(
8)
16
(
9)
"
4"
items
of
exterior
deterioration
3
(
3)
4
(
5)
7
(
4)
"
6"
items
of
exterior
deterioration
1
(
1)
1
(
1)
2
(
1)
Source:
Form
02
Baseline
Interior
Dwelling
Unit
Deterioration.
LAAP
inspectors
also
visually
assessed
the
level
of
pre­
existing
interior
dwelling
unit
deterioration
by
checking
against
five
(
5)
evaluation
items.
Here,
LHR+
W
units
had
higher
percentages
of
deterioration
than
the
LHR
units
for
four
11/
23/
05
29
(
4)
of
the
five
(
5)
items
(
see
Table
6).
Most
notably,
LHR+
W
units
were
more
likely
than
LHR
units
to
have
walls,
ceilings,
doors
and
trim
that
were
in
need
of
repair,
replacement
or
repainting,
and
floors
that
were
loose,
cracked,
had
missing
boards
or
a
worn
finish,
or
had
deteriorated
carpeting.
Condition
of
the
floors
was
significantly
associated
with
the
assigned
treatment
group
(
p­
value
=
0.010),
while
condition
of
the
walls,
ceiling,
doors
and
trim
was
marginally
associated
with
assigned
treatment
group
(
p­
value
=
0.088).
A
slightly
higher
percentage
of
LHR+
W
units
had
interior
deterioration
caused
at
least
in
part
by
disrepair
of
heating/
cooling
or
plumbing
systems,
while
slightly
more
LHR
units
had
interior
deterioration
caused
by
roof
leaks.
However,
neither
disrepair
of
heating/
cooling
or
plumbing
systems
nor
roof
leaks
were
significantly
associated
with
assigned
treatment
group.
Both
LHR
and
LHR+
W
units
had
a
very
high
percentage
of
units
(
83.5
percent
LHR
and
88.4
percent
LHR+
W)
with
at
least
one
room
having
more
than
2
ft2
of
paint
deterioration
on
walls,
woodwork,
doors,
windows
and
other
painted
surfaces.
The
mean
number
of
rooms
with
such
paint
deterioration
was
2.3
rooms
in
LHR
units
and
2.6
rooms
in
LHR+
W
units.

Table
6:
Units
with
Interior
Dwelling
Unit
Deterioration
 
Pre­
Intervention
Assessment
Type
of
Interior
Dwelling
Unit
Deterioration
(
91
LHR
units,
86
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

Walls,
ceiling,
doors
and
trim
 
cracks,
need
for
repair,
replace
or
major
repaintinga
36
(
40%)
45
(
52%)
81
(
46%)

Floors
 
loose,
missing
or
cracked,
finish
worn,
deteriorated
carpetingb
45
(
49%)
59
(
69%)
104
(
59%)

Heating/
cooling
and
plumbing
 
need
for
extensive
repairc
8
(
9%)
11
(
13%)
19
(
11%)

Interior
damage
due
to
roof
leak
 
need
for
extensive
repairc
17
(
19%)
14
(
16%)
31
(
18%)

Rooms
with
>
2
ft2
of
paint
deterioration
on
walls,
woodwork,
doors,
windows,
etc.
c
76
(
84%)
76
(
88%)
152
(
86%)

Source:
Form
02
aCondition
of
walls,
ceilings,
doors,
&
trim
marginally
associated
with
assigned
treatment
group
based
on
Chisquare
test
(
p=
0.088).
bFloor
condition
significantly
associated
with
assigned
treatment
group
based
on
Chi­
square
test
(
p=
0.010).
cThese
interior
deterioration
items
not
associated
with
assigned
treatment
group
based
on
Chi­
square
test.

The
total
number
of
observed
items
of
interior
deterioration
was
significantly
associated
with
the
assigned
treatment
group
(
p­
value
=
0.0212).
The
number
of
LHR
units
with
at
least
one
observed
interior
dwelling
unit
deterioration
item
(
approximately
76
percent)
was
lower
than
that
of
LHR+
W
units
(
approximately
84
percent)
(
see
Table
7
on
page
30).
Interestingly,
more
LHR
units
(
45
percent)
than
LHR+
W
units
(
35
percent)
had
one
item
of
interior
deterioration,
but
the
LHR+
W
units
were
more
likely
to
have
two
(
2)
or
more
items
of
deterioration.
11/
23/
05
30
Table
7:
Number
of
Interior
Dwelling
Unit
Deterioration
Items
 
Pre­
Intervention
Assessment
Number
of
Items
of
Interior
Deterioration
per
dwelling
unit
(
91
LHR
units,
86
LHR+
W
units)
a
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
items
of
interior
deterioration
22
(
24%)
14
(
16%)
36
(
20%)
"
1"
item
of
interior
deterioration
41
(
45%)
30
(
35%)
71
(
40%)
"
2"
items
of
interior
deterioration
20
(
22%)
29
(
34%)
49
(
28%)
"
3"
items
of
interior
deterioration
7
(
8%)
11
(
13%)
18
(
10%)
"
4"
items
of
interior
deterioration
1
(
1%)
2
(
2%)
3
(
2%)
Source:
Form
02
aTotal
number
of
items
of
interior
deterioration
significantly
associated
with
assigned
treatment
group
based
on
Chisquare
test
(
p=
0.0212).

B.
Results
of
Pre­
Intervention
Composite
and
Single
Surface
Dust
Sampling
As
noted
above,
for
most
of
the
enrolled
units,
LAAP
inspectors
did
not
take
dust
samples
when
pre­
intervention
visual
assessments
were
performed
because
a
number
of
initially
assessed
units
ultimately
were
de­
enrolled
by
the
property
owners
and/
or
by
LAAP
before
lead
hazard
reduction
specifications
and
grant
applications
were
submitted
and
approved.
Therefore,
some
units
were
de­
enrolled
prior
to
the
collection
of
dust
samples
but
after
pre­
intervention
characterizations
were
done.
To
ensure
that
dust
samples
were
collected
no
more
than
four
(
4)
weeks
prior
to
the
intervention,
LAAP
inspectors
typically
obtained
composite
and
single
surface
samples
during
a
pre­
construction
walk­
through
of
the
unit
with
the
property
owner
and,
if
applicable,
the
contractor
performing
the
lead
hazard
reduction
work.
For
this
reason,
there
are
fewer
units
having
pre­
intervention
dust
sample
results
(
148
units)
than
those
with
visual
assessment
results
(
177
units).

1.
Pre­
Intervention
Composite
Results
Composite
dust
samples
were
taken
from
bare
floors,
interior
window
sills,
window
troughs
and,
when
necessary,
carpeted
floors
in
all
enrolled
LHR
and
LHR+
W
units.
Because
very
few
units
had
carpeted
floors,
sample
results
for
carpeted
floors
were
excluded
from
this
study.
A
Wilcoxon
rank
sum
test
was
used
to
test
the
equality
of
the
median
dust
lead
loadings
for
the
assigned
treatment
group
(
i.
e.,
LHR
versus
LHR+
W
units).
Differences
in
median
preintervention
dust
lead
loadings
between
the
two
groups
had
to
be
accounted
for
before
the
effectiveness
of
the
lead
hazard
control
treatments
could
be
fully
evaluated.

For
the
three
(
3)
surface
types,
the
median
pre­
intervention
dust
lead
loadings
in
the
LHR+
W
units
exceeded
median
loadings
for
the
LHR
units
(
see
Table
8
on
page
31
and
Figure
3
on
page
32).
These
differences
were
significant
for
bare
floors
(
p­
value=
0.0374),
marginally
significant
for
window
troughs
(
p­
value=
0.1077),
but
not
significant
for
interior
window
sills
(
pvalue
0.2803).
The
5th,
25th,
75th,
and
95th
percentile
dust
lead
loadings
in
the
LHR+
W
units
were
all
higher
than
those
in
the
LHR
units,
in
some
cases
dramatically
so.
11/
23/
05
31
Table
8:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
Surface
Type
LHR
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsa,
b
(
73
LHR
units,
71
LHR+
W
units)
22
77
273
634
5,199
258
173;
385
39
141
333
1,635
23,960
500
325;
768
Interior
window
sillsa,
b
(
74
LHR
units,
74
LHR+
W
units)
31
631
1,986
8,362
42,619
1,854
1,148;
2,995
286
954
2,780
9,545
101,200
3,153
2,098;
4
737
Window
troughsa,
b
(
74
LHR
units,
74
LHR+
W
units)
67
2,050
9,427
38,308
260,182
7,680
4,568;
12,911
740
3,287
10,998
90,480
447,429
16,659
10,142;
27,365
Source:
Form
04A
aApproximately
95%
of
bare
floor
composites,
interior
window
sill
composites,
and
window
trough
composites
were
comprised
of
4
sub­
samples;
the
rest
were
comprised
of
either
2
or
3
sub­
samples.
b
Based
on
Wilcoxon
rank
sum
tests,
pre­
intervention
dust
lead
loadings
significantly
associated
with
assigned
treatment
group
for
floors
(
p=
0.0374),
marginally
associated
for
troughs
(
p=
0.1077),
&
not
associated
for
sills
(
p=
0.2803).
11/
23/
05
32
Figure
3:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot
1
10
100
1000
10000
100000
1000000
LHR
Bare
Floors
LHR+
W
Bare
Floors
LHR
Window
Sills
LHR+
W
Window
Sills
LHR
Window
Troughs
LHR+
W
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

2.
Pre­
Intervention
Single
Surface
Results
As
noted
in
Section
II.
E,
although
single
surface
dust
wipe
samples
were
to
be
collected
only
in
LHR+
W
units,
both
composite
and
single
surface
samples
were
collected
in
certain
LHR
units
from
November
1997
until
February
1,
1998
due
to
laboratory
analysis
concerns.
(
All
such
LHR
units
dust
sampled
during
that
period
had
the
two
sample
types
collected
through
Phase
II.)

Similar
to
composites,
single
surface
samples
were
taken
from
bare
floors,
interior
window
sills,
window
troughs
in
designated
rooms.
(
There
were
too
few
carpeted
floor
samples
to
allow
full
characterization
of
this
surface
type.)
There
were
more
LHR+
W
units
(
73)
with
single
surface
samples
than
LHR
units
(
25),
but
the
median
dust
lead
loading
values
in
the
LHR+
W
units
again
exceeded
those
of
the
LHR
units
by
a
factor
of
two
to
three
(
see
Table
9
on
page
33
and
Figure
4
on
page
34).
Based
on
Chi­
square
testing,
these
differences
were
not
significant
for
bare
floors
(
p­
value=
0.1176)
or
window
troughs
(
p­
value=
0.2818),
but
were
highly
significant
for
interior
window
sills
(
p­
value=
0.0036).
The
5th,
25th,
75th,
and
95th
percentile
dust
lead
loadings
were
also
much
higher
in
the
LHR+
W
units.

The
pre­
intervention
single
surface
dust
sample
results
for
both
LHR
and
LHR+
W
units
were
also
compared
to
pre­
intervention
single
surface
dust
data
obtained
from
Baltimore
City
housing
units
enrolled
in
Round
I
of
the
National
Evaluation
of
HUD's
OHHLHC's
Lead
Hazard
Control
11/
23/
05
33
Grant
program.
These
units
underwent
lead
hazard
reduction
interventions
that
were
similar
to
those
of
the
LHR+
W
units,
and
more
intensive
than
those
of
LHR
units.
No
consistent
differences
in
pre­
intervention
dust
lead
levels
were
found
to
exist
between
the
LHR
and
LHR+
W
treatment
groups
and
the
Baltimore
Round
One
units,
suggesting
that
the
Round
One
units
could
be
used
to
contrast
the
effectiveness
of
higher
level
lead
hazard
reduction
treatments
versus
the
somewhat
lower
level
EA
6­
8
prescribed
treatments.
Pre­
intervention
bare
floor
dust
lead
levels
for
LHR
and
LHR+
W
units
were
not
significantly
different
than
for
the
Round
One
units.
But
pre­
intervention
interior
window
sill
lead
levels
were
lower
for
LHR
units
than
the
Round
One
units,
while
LHR+
W
units
had
levels
similar
to
the
Round
One
units.
For
window
troughs,
the
opposite
was
observed.
Here,
the
LHR+
W
units
had
significantly
higher
preintervention
dust
lead
levels
than
the
Round
One
units.

Table
9:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
Surface
Type
LHR
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Baltimore
HUD
Unitsa
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsb
(
25
LHR
units,
73
LHR+
W
units)
29
83
192
881
7,225
293
140;
611
58
154
338
986
29,258
511
335;
779
37
130
314
681
1,955
292
254;
337
Interior
window
sillsb
(
25
LHR
units,
73
LHR+
W
units)
41
186
844
1,580
45,179
705
283;
1,755
146
850
2,350
7,303
229,108
2,830
1,790;
4,475
223
832
2,510
7,228
27,470
2,509
2,133;
2,950
Window
troughsb
(
25
LHR
units,
72
LHR+
W
units)
329
2,253
6,135
39,954
98,100
7,795
3,372;
18,020
359
1,949
12,227
92,595
547,652
13,462
7,910;
22,911
136
1,070
3,630
13,000
50,000
3,451
2,807;
4,243
Source:
Form
04B
aBaltimore
HUD
unit
data
are
from
Baltimore
Round
One
housing
units
enrolled
in
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
bBased
on
Wilcoxon
rank
sum
tests,
pre­
intervention
single
surface
dust
lead
loadings
not
associated
with
assigned
treatment
group
for
bare
floors
(
p=
0.1176)
or
troughs
(
p=
0.2818),
but
significantly
associated
for
sills
(
p=
0.0036).
11/
23/
05
34
Figure
4:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
Box
Plot
Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

3.
Comparison
of
Pre­
Intervention
Single
Surface
Results
with
Clearance
Standards/
Guidance
Comparison
of
pre­
intervention
single
surface
dust
sample
results
for
interior
window
sills
with
the
Maryland
statutory
abatement
clearance
standard
and
the
joint
HUD/
EPA
guidance
clearance
standard
of
500
µ
g/
ft2
yields
useful
information
in
terms
of
statistical
association
with
higher
dust
lead
loading.
Assigned
treatment
group
was
found
to
be
significantly
associated
with
preintervention
interior
window
sill
dust
lead
loading
being
above
or
below
a
standard
of
500
µ
g/
ft2
(
p­
value
=
0.009).
(
Note:
Similar
analyses
were
performed
using
clearance
standards
for
floors
and
window
troughs,
but
no
statistically
significant
associations
were
observed
for
the
assigned
treatment
group.)
In
summary,
the
results
of
pre­
intervention
exterior
and
interior
visual
assessment
and
composite
and
single
surface
dust
sampling
all
indicate
that
LHR+
W
units
initially
enrolled
in
the
study
were
in
poorer
condition
than
enrolled
LHR
units,
suggesting
that
assignment
of
units
by
property
owners
into
these
two
categories
was
not
random.
Paint
lead
content
was
not
measured
during
this
study.
However,
given
historical
residential
lead
paint
usage
patterns
in
Baltimore
City
for
housing
units
of
similar
age
to
those
enrolled
in
the
study,
it
is
unlikely
that
the
lead
content
of
paint
in
the
LHR
units
was
substantially
different
from
that
in
the
LHR+
W
units.
1
10
100
1000
10000
100000
1000000
LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)
11/
23/
05
35
VIII.
IMMEDIATE
POST­
INTERVENTION
RESULTS
Of
the
148
units
that
had
pre­
intervention
dust
results
available,
a
total
of
121
units
(
57
LHR
and
64
LHR+
W)
completed
lead
hazard
reduction
treatment
interventions
and
a
full
immediate
postintervention
assessment,
including
a
visual
assessment,
collection
of
cost
and
concurrent
work
information
and
post­
intervention
dust
testing.
Samples
potentially
associated
with
Lab
A's
unacceptable
composite
spike
sample
results
were
excluded
from
all
analyses
(
see
Appendix
B).
The
relationship
between
these
post­
intervention
factors
and
the
assigned
treatment
group
(
i.
e.,
LHR
versus
LHR+
W
units)
was
examined
using
a
Chi­
square
test.
A
Wilcoxon
rank
sum
test
was
used
to
test
the
equality
of
the
post­
intervention
median
dust
lead
loadings
for
the
assigned
treatment
group.

A.
Results
of
Immediate
Post­
Intervention
Composite
and
Single
Surface
Dust
Sampling
1.
Immediate
Post­
Intervention
Composite
Results
Immediate
post­
intervention
composite
dust
samples
were
taken
from
the
same
surface
types
and
locations
as
those
sampled
at
pre­
intervention
(
see
Table
10
on
page
36
and
Figure
5
on
page
37).
For
bare
floors,
immediate
composite
post­
intervention
median
dust
lead
loadings
for
LHR
and
LHR+
W
units
were
not
significantly
different
(
p­
value=
0.6422).
However,
as
might
be
expected,
immediate
post­
intervention
median
dust
lead
loadings
for
sills
and
troughs
were
highly
associated
with
assigned
treatment
group
(
p­
values=
0.0014
and
0.0001,
respectively).
The
LHR+
W
units
had
much
lower
median
dust
lead
loadings
for
interior
window
sills
and
window
troughs
than
those
in
LHR
units.
11/
23/
05
36
Table
10:
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
Surface
Type
LHR
Composite
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Composite
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsa,
b
(
51
LHR
units,
57
LHR+
W
units)
4
15
27
71
195
30
22;
42
4
11
29
58
727
31
21;
46
Interior
window
sillsa,
b
(
54
LHR
units,
62
LHR+
W
units)
12
27
83
265
1,038
92
62;
135
5
14
36
103
508
36
25;
53
Window
troughsa,
b
(
54
LHR
units,
62
LHR+
W
units)
12
174
344
1,100
15,516
416
251;
690
14
30
100
312
1,282
106
72;
156
Source:
Form
04A
aApproximately
95%
of
bare
floor
composites,
interior
window
sill
composites,
and
window
trough
composites
were
comprised
of
4
sub­
samples;
the
rest
were
comprised
of
either
2
or
3
sub­
samples.
bBased
on
Wilcoxon
rank
sum
tests,
post­
intervention
composite
dust
lead
loadings
were
not
significantly
associated
with
assigned
treatment
group
for
bare
floors
(
p=
0.6422),
but
highly
associated
for
sills
and
troughs
(
p=
0.0014
and
0.0001,
respectively).
11/
23/
05
37
Figure
5:
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot
1
10
100
1000
10000
100000
LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

2.
Reduction
in
Composite
Results
from
Pre­
to
Immediate
Post­
Intervention
Large
median
percent
decreases
and
median
µ
g/
ft2
decreases
were
observed
when
measuring
the
change
from
pre
to
post­
intervention
dust
lead
loading
results
for
composite
samples
taken
from
bare
floors,
interior
window
sills
and
window
troughs,
with
the
most
dramatic
decreases
being
observed
for
sills
and
troughs
(
see
Tables
11
and
12
on
page
38,
and
Figure
6
on
page
39).
The
association
between
the
percent
reduction
in
composite
dust
lead
loadings
and
the
assigned
treatment
group
was
highly
significant
for
interior
window
sills
(
p­
value=
0.0001)
and
troughs
(
pvalue
0.0001)
but
was
not
significant
for
bare
floors
(
p­
value=
0.2459).
11/
23/
05
38
Table
11:
Percent
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Composite
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Composite
Dust
Samples
LHR
Units
Pre­
to
Post­
Intervention
Percent
Reduction
LHR+
W
Units
Pre­
to
Post­
Intervention
Percent
Reduction
Bare
floors
(
49
LHR,
56
LHR+
W)
a:
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
70
85
93
74
89
97
Interior
window
sills
(
52
LHR,
62
LHR+
Wa
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
78
94
98
96
99
>
99
Window
troughs
(
52
LHR,
62
LHR+
W)
a
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
74
95
98
97
>
99
>
99
Source:
Form
04A
aBased
on
Wilcoxon
rank
sum
tests,
percent
reduction
in
composite
dust
lead
loadings
was
significantly
associated
with
assigned
treatment
group
for
sills
(
p=
0.0001)
and
troughs
(
p=
0.0001)
but
not
for
bare
floors
(
p=
0.2459).

Table
12:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Composite
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Composite
Dust
Samples
LHR
Units
Pre
to
Post­
Intervention
µ
g/
ft2
Reduction
LHR+
W
Units
Pre
to
Post­
Intervention
µ
g/
ft2
Reduction
Bare
floors
(
49
LHR,
56
LHR+
W):
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
40
147
506
71
225
807
Interior
window
sills
(
52
LHR,
62
LHR+
W:
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
415
1,686
8,272
716
2,596
7,433
Window
troughs
(
52
LHR,
62
LHR+
W):
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
1,230
9,708
25,175
4,254
11,284
90,424
Source:
Form
04A
11/
23/
05
39
Figure
6:
Pre­
and
Immediate
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
for
LHR
and
LHR+
W
Units
(
Logarithmic
Scale)

1
10
100
1000
10000
100000
1000000
Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

3.
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention
Despite
the
significant
median
decreases
in
dust
lead
loadings
between
pre­
and
immediate
postintervention
data
in
Table
13
on
page
40
show
that
some
units
experienced
increases
in
dust
lead
loadings
between
these
two
phases.
The
percentage
of
LHR+
W
units
that
underwent
an
increase
was
generally
lower
than
that
of
LHR
units.
11/
23/
05
40
Table
13:
Number
and
Percent
of
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention
Surface
Type
and
Sample
Type
LHR
LHR+
W
Bare
Floors:
6/
49
(
12%)
3/
56
(
5%)

Interior
window
sills:
4/
52
(
8%)
2/
62
(
3%)

Window
Troughs:
3/
52
(
6%)
1/
62
(
2%)

4.
Immediate
Post­
Intervention
Single
Surface
Results
As
previously
noted,
single
surface
dust
samples
taken
in
all
62
LHR+
W
units
and
in
20
LHR
units
were
also
included
in
this
sampling
subset
(
see
Table
14
on
page
41).
With
this
in
mind,
although
the
median
immediate
post­
intervention
single
surface
dust
lead
loading
for
bare
floors
in
LHR
units
was
higher
than
that
in
LHR+
W
units,
this
difference
was
not
significant
(
pvalue
0.2168).
However,
as
might
be
expected,
single
surface
immediate
post­
intervention
median
dust
lead
loadings
for
sills
and
troughs
were
highly
associated
with
assigned
treatment
group
(
p­
values=
0.0012
and
0.0035,
respectively).
The
LHR+
W
units
had
much
lower
median
dust
lead
loadings
for
interior
window
sills
and
window
troughs
than
those
in
LHR
units.

Immediate
post­
intervention
single
surface
dust
sample
results
for
both
LHR
and
LHR+
W
units
were
compared
with
immediate
post­
intervention
single
surface
dust
data
obtained
from
Baltimore
city
housing
units
enrolled
in
Round
One
of
the
National
Evaluation
of
HUD
OHHLHC's
Lead
Hazard
Control
Grant
Program.
These
units
underwent
interventions
that
were
similar
to
those
of
LHR+
W
units
and
generally
more
intensive
than
those
of
LHR
units.
(
see
Table
14
on
page
41
and
Figure
7
on
page
42).
Median
dust
lead
loading
results
for
bare
floors,
interior
window
sills
and
window
troughs
were
all
much
lower
in
the
Baltimore
Round
One
units
than
in
both
the
LHR
and
LHR+
W
units.
Each
surface
type
comparison
was
found
to
be
highly
significant
(
p­
value=
0.0001
for
each).
11/
23/
05
41
Table
14:
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loading
Results
Surface
Type
LHR
Single
Surface
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Single
Surface
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Baltimore
HUD
Units
Single
Surface
Dust
Lead
Loading
(
µ
g/
ft2)
a
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsb,
c
(
20
LHR
units,
63
LHR+
W
units)
10
26
66
112
1217
60
34;
109
6
20
39
89
307
43
31;
59
0.3
5
15
35
96
14
12;
16
Interior
window
sillsb,
c
(
20
LHR
units,
62
LHR+
W
units)
14
33
103
178
961
96
51;
178
7
15
29
51
257
32
24;
42
0
2
10
31
103
17
15;
20
Window
troughsb,
c
(
20
LHR
units,
62
LHR+
W
units)
29
217
610
1,139
5,071
460
230;
921
9
32
115
347
2,262
128
79;
207
0
6
38
125
607
57
47;
68
Source:
Form
04B
aBaltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program
and
were
compiled
using
initial
Phase
2
data
(
immediate
post­
intervention).
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
bBased
on
Wilcoxon
rank
sum
tests,
median
post­
intervention
single
surface
dust
lead
loadings
with
significantly
associated
with
assigned
treatment
group
for
sills
(
p=
0.0012)
and
troughs
(
p=
0.0035),
but
not
associated
for
bare
floors
(
0.2168).
cBased
on
Wilcoxon
rank
sum
tests,
median
post­
intervention
single
surface
dust
lead
loadings
for
Baltimore
Round
1
units
were
significantly
lower
than
those
for
study
units
(
p=
0.0001
for
each
surface
type).
11/
23/
05
42
Figure
7:
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loading
Results
Box
Plot
1
10
100
1000
10000
LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

5.
Reduction
in
Single
Surface
Results
from
Pre­
to
Immediate
Post­
Intervention
Similar
to
the
composite
dust
sample
results,
when
single
surface
sample
pre­
to
immediate
postintervention
dust
lead
loading
results
were
compared,
substantial
median
percent
decreases
and
median
µ
g/
ft2
decreases
were
observed
for
bare
floors,
interior
window
sills
and
window
troughs.
These
reductions
were
most
dramatic
for
window
sills
and
troughs
and
were
particularly
large
in
the
LHR+
W
units
(
see
Tables
15
and
16
on
pages
43
and
44,
respectively,
and
Figure
8
on
page
45).
For
each
of
these
surface
types,
the
median
percent
decrease
for
LHR+
W
units
was
greater
than
that
for
LHR
units.
The
association
between
the
median
percent
reduction
in
single
surface
dust
lead
loadings
and
the
assigned
treatment
group
was
highly
significant
for
interior
window
sills
and
troughs
(
p­
values=
0.0001
and
0.0057,
respectively)
but
was
not
significant
for
bare
floors
(
p­
value=
0.1655).

When
reductions
for
LHR
and
LHR+
W
units
were
compared
with
reductions
for
Baltimore
Round
One
HUD
Evaluation
units,
the
Round
One
units
had
statistically
significant
larger
decreases
for
all
three
surface
types,
as
measured
by
median
percentage
and
actual
µ
g/
ft2,
than
did
both
LHR
and
LHR+
W
units
(
p­
value=
0.0001
for
all
three
surface
types).
11/
23/
05
43
Table
15:
Percent
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Single
Surface
Dust
Samples
LHR
Units
Pre
to
Immediate
Post­
Intervention
Percent
Reduction
LHR+
W
Units
Pre
to
Immediate
Post­
Intervention
Percent
Reduction
Baltimore
HUD
Unitsa
Pre
to
Immediate
Post­
Intervention
%
Reduction
Bare
floors
(
16
LHR,
63
LHR+
W)
b:
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
24
70
97
62
88
98
86
95
98
Interior
window
sills
(
16
LHR,
62
LHR+
W)
b:
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
55
84
95
96
99
>
99
98
>
99
>
99
Window
troughs
(
16
LHR,
62
LHR+
W)
b:
25th
Percentile
Percent
Decrease
Median
Percent
Decrease
75th
Percentile
Percent
Decrease
82
94
98
96
99
>
99
95
99
>
99
Source:
Form
04B
aBaltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program
and
were
compiled
using
initial
Phase
2
data
(
immediate
post­
intervention)
for
units
that
were
vacant
at
pre­
intervention
and
at
immediate
post­
intervention.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
bBased
on
Wilcoxon
rank
sum
tests,
the
association
between
the
median
percent
reduction
in
single
surface
dust
lead
loadings
and
assigned
treatment
group
was
significant
for
sills
(
p=
0.0001)
and
troughs
(
p=
0.0057),
but
not
significant
for
bare
floors
(
p=
0.1655).
cBased
on
Wilcoxon
rank
sum
tests,
Baltimore
Round
1
units
had
statistically
significant
larger
median
percent
decreases
in
dust
lead
loadings
than
did
study
units
in
either
treatment
group
(
p=
0.0001
for
all
3
surface
types).
11/
23/
05
44
Table
16:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Single
Surface
Dust
Samples
LHR
Units
Pre
to
Immediate
Post­
Intervention
µ
g/
ft2
Reduction
LHR+
W
Units
Pre
to
Immediate
Post­
Intervention
µ
g/
ft2
Reduction
Baltimore
HUD
Units
Pre
to
Immediate
Post­
Intervention
µ
g/
ft2
Reduction
Bare
floors
(
16
LHR,
63
LHR+
W):
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
5
99
3,027
97
226
971
98
265
618
Interior
window
sills
(
16
LHR,
62
LHR+
W):
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
70
497
9,102
752
2,516
8,598
800
2,492
7,124
Window
troughs
(
16
LHR,
62
LHR+
W):
25th
Percentile
µ
g/
ft2
Decrease
Median
µ
g/
ft2
Decrease
75th
Percentile
µ
g/
ft2
Decrease
998
4,111
36,239
1,640
12,022
131,245
1,020
3,591
12,994
Source:
Form
04B
1Baltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program
and
were
compiled
using
initial
Phase
2
data
(
immediate
post­
intervention).
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
11/
23/
05
45
Figure
8:
Pre­
and
Immediate
Post­
Intervention
Single
Surface
Dust
Lead
Loading
for
LHR
and
LHR+
W
Units
(
Logarithmic
Scale)

1
10
100
1000
10000
100000
1000000
Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

Pre­
Int.

Post­
Int.

LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

6.
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention
Despite
the
significant
median
decreases
in
dust
lead
loadings
between
pre­
and
immediate
postintervention
data
in
Table
17
on
page
46
show
that
some
units
experienced
increases
in
single
surface
dust
lead
loadings
between
these
two
phases.
The
percentage
of
LHR+
W
units
that
underwent
an
increase
was
generally
lower
than
that
of
LHR
units
and
was
comparable
to
that
of
Baltimore
HUD
units.
11/
23/
05
46
Table
17:
Number
and
Percent
of
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
Immediate
Post­
Intervention
Surface
Type
and
Sample
Type
LHR
LHR+
W
Baltimore
HUD
Units
Bare
Floors:
4/
16
(
25%)
4/
63
(
6%)
12/
278
(
4%)

Interior
window
sills:
2/
16
(
12%)
1/
62
(
2%)
2/
277
(
1%)

Window
Troughs:
2/
16
(
12%)
1/
62
(
2%)
13/
273
(
5%)

7.
Comparison
of
Immediate
Post­
Intervention
Single
Surface
Results
with
Clearance
Standards/
Guidance
Under
the
lead
hazard
reduction
treatment
option
of
EA
6­
8,
enrolled
LHR
and
LHR+
W
units
were
only
required
to
pass
a
visual
inspection
performed
by
an
independent
visual
inspector;
collection
of
post­
treatment
dust
samples
and
comparison
of
such
results
to
clearance
standards
were
not
required.
However,
this
study
considered
clearance
dust
standards
set
by
Maryland
and
by
HUD/
EPA
to
assess
the
extent
to
which
prescribed
treatments
produced
dust
results
that
were
sufficient
for
safe
occupancy
of
treated
units
and
to
help
determine
whether
the
prescribed
independent
visual
inspection
procedure
provided
a
sufficient
level
of
lead
hazard
reduction.
This
comparison
to
clearance
standards
is
not
required
by
EA
6­
8
and
was
performed
in
this
study
for
research
purposes
only.

If
such
testing
had
been
required,
for
single
surface
floor
dust
samples,
35
percent
of
the
LHR
units
and
25
percent
of
the
LHR+
W
units
had
at
least
one
floor
sample
that
would
have
"
failed"
the
Maryland
standard
of
200
µ
g/
ft2
(
see
Table
18
on
page
47).
Sixty
(
60)
percent
of
the
LHR
units
and
41
percent
of
the
LHR+
W
units
had
at
least
one
floor
sample
that
would
have
"
failed"
the
HUD/
EPA
1995
guidance7
of
100
µ
g/
ft2.
Assigned
treatment
group
was
not
significantly
associated
with
a
unit
"
passing"
or
"
failing"
either
of
these
two
clearance
standards
(
pvalues
0.4
and
0.14
for
clearance
standards
of
200
and
100
µ
g/
ft2,
respectively).
These
are
generally
higher
"
failure"
rates
than
the
29
percent
initial
floor
failure
rate
observed
for
vacant
Baltimore
Round
I
units
enrolled
in
the
National
Evaluation.
8
7
HUD/
EPA
guidance
values
of
100
µ
g/
ft2
for
floors,
500
µ
gft2
for
window
sills,
and
800
µ
g/
ft2
for
window
troughs
were
in
effect
during
the
study
period.
New
standards
that
became
effective
in
September
2000
(
40,
250,
and
400
µ
g/
ft2
were
not
in
effect
when
immediate
post­
intervention
data
were
collected
and
are
therefore
not
discussed
in
this
report.
8
Clearance
failure
rates
for
Baltimore
HUD
units
are
based
on
initial
clearance
testing
conducted
after
treatment,
comparing
results
with
floor,
sill
and
trough
standards
of
100,
500
and
800
µ
g/
ft2,
respectively.
11/
23/
05
47
Table
18:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Floors
Immediate
Post­
Intervention
Single
Surface
Dust
Sample
Results
 
FLOOR
"
Clearance
Failures"
a
LHR
(
20
units
total)
LHR+
W
(
63
units
total)
All
Units
(
83
units
total)

#
units
(%)

100
µ
g/
ft2
12
(
60%)
26
(
41%)
38
(
46%)

#
units
(%)

200
µ
g/
ft2
7
(
35%)
16
(
25%)
23
(
28%)

aClearance
"
failure"
was
not
significantly
associated
with
assigned
treatment
group
(
p=
0.403
and
p=
0.143
for
clearance
stds
of
200
and
100
µ
g/
ft2,
respectively)

For
single
surface
interior
window
sill
dust
samples,
replacement
of
windows
in
the
LHR+
W
units
resulted
in
fewer
"
clearance
failures"
(
see
Table
19).
If
clearance
testing
had
been
required,
20
percent
of
the
LHR
units
and
five
percent
of
the
LHR+
W
units
had
at
least
one
sill
sample
that
would
have
"
failed"
the
Maryland
and
HUD/
EPA
guidance
of
500
µ
g/
ft2.
Assigned
treatment
group
was
significantly
associated
with
post­
intervention
interior
window
sill
dust
lead
loadings
"
passing"
or
"
failing"
at
500
µ
g/
ft2
(
p­
value=
0.035).
These
rates
are
higher
than
the
one
percent
sill
initial
failure
rate
observed
for
vacant
Baltimore
Round
I
units
enrolled
in
the
National
Evaluation.
8
Table
19:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Interior
Window
Sills
Immediate
Post­
Intervention
Single
Surface
Dust
Sample
Results
 
INTERIOR
WINDOW
SILL
"
Clearance
Failures"
a
LHR
(
20
units
total)
LHR+
W
(
62
units
total)
All
Units
(
82
units
total)

#
units
(%)

500
µ
g/
ft2
4
(
20%)
3
(
5%)
7
(
8%)

aClearance
"
failure"
was
significantly
associated
with
assigned
treatment
group
(
p=
0.035).

Again,
for
single
surface
window
trough
dust
samples,
replacement
of
windows
in
the
LHR+
W
units
resulted
in
a
lower
percentage
of
"
clearance
failures"
(
see
Table
20
on
page
48).
Had
clearance
testing
been
required,
50
percent
of
the
LHR
units
and
19
percent
of
the
LHR+
W
units
had
at
least
one
trough
sample
that
would
have
"
failed"
the
Maryland
and
HUD/
EPA
guidance
standard
of
800
µ
g/
ft2.
Assigned
treatment
group
was
significantly
associated
with
postintervention
window
trough
dust
lead
loadings
"
passing"
or
"
failing"
this
standard
(
pvalue
0.007).
These
rates
are
higher
than
the
six
percent
initial
trough
failure
rate
noted
for
Baltimore
Round
I
units.
8
11/
23/
05
48
Table
20:
Immediate
Post­
Intervention
Single
Surface
"
Clearance
Failures"
 
Window
Troughs
Immediate
Post­
Intervention
Single
Surface
Dust
Sample
Results
 
WINDOW
TROUGH
"
Clearance
Failures"
a
LHR
(
20
units
total)
LHR+
W
(
62
units
total)
All
Units
(
82
units
total)

#
units
(%)

800
µ
g/
ft2
10
(
50%)
12
(
19%)
22
(
27%)

aClearance
"
failure"
was
significantly
associated
with
assigned
treatment
group
(
p=
0.007).

B.
Results
of
Immediate
Post­
Intervention
Visual
Assessments
As
required
in
EA
6­
8,
each
of
the
121
units
was
independently
inspected
and
passed
the
required
visual
inspection
after
the
prescribed
treatments
were
completed.
As
one
of
the
evaluation
measures
of
this
study,
LAAP
inspectors
performed
a
separate,
confirmatory
visual
assessment
(
not
related
to
the
property
owner's
statutory
requirement
for
an
independent
visual
inspection)
in
each
unit
to
determine
if
any
visible
treatment
"
failures"
could
be
detected.
For
the
purposes
of
this
study,
a
treatment
"
failure"
was
defined
as
an
observation
by
the
LAAP
inspector
that
one
or
more
of
the
ten
lead
hazard
reduction
treatments
prescribed
in
the
statute
had
not
been
fully
completed.
The
Maryland
statute
did
not
specify
a
"
de
minimis"
level
above
which
the
treatment
was
considered
a
"
failure;"
therefore,
any
observable
deficient
or
missing
treatment
was
classified
as
a
"
failure."

A
complete
list
of
immediate
post­
intervention
visual
assessment
"
failures"
is
presented
in
Appendix
G.
Table
21
on
page
49
presents
the
distribution
of
treatment
"
failures"
per
unit.
There
were
no
units
having
more
than
5
"
failures."
Notably,
the
LAAP
inspectors
found
that
51
of
the
57
LHR
units
(
89
percent)
and
60
of
the
64
(
94
percent)
of
the
LHR+
W
units
"
failed"
the
confirmatory
visual
assessment
for
at
least
one
item.
This
equates
to
an
overall
visual
assessment
"
failure"
rate
of
92
percent
for
the
121
study
units.
The
owner
corrected
treatment
"
failures"
noted
during
the
immediate
post­
intervention
sampling
visit
before
tenants
moved
in.
These
corrections
were
"
validated"
by
LAAP
inspectors,
who
required
owners
and/
or
contractors
to
complete
their
work
prior
to
payment.

Overall,
the
likelihood
of
LHR
units
having
one
or
more
"
failures"
was
almost
the
same
as
that
of
the
LHR+
W
units.
The
mean
number
of
"
failure"
types
per
unit
for
LHR
units
(
2.1)
was
significantly
different
and
higher
than
that
for
LHR+
W
units
(
1.7)
(
p­
value=
0.0255).
If
only
"
failure"
types
applicable
to
both
treatment
groups
are
considered,
LHR
units
have
slightly
more
mean
"
failures"
(
1.8)
than
LHR+
W
units
(
1.6).
However,
this
difference
is
not
significant
(
pvalue
0.1389).
11/
23/
05
49
Table
21:
Number
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unit
Number
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unita
(
57
LHR
units,
64
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
lead
hazard
treatment
"
failures"
6
(
10%)
4
(
6%)
10
(
8%)
"
1"
lead
hazard
treatment
"
failure"
7
(
12%)
26
(
41%)
33
(
27%)
"
2"
lead
hazard
treatment
"
failures"
22
(
39%)
21
(
34%)
44
(
36%)
"
3"
lead
hazard
treatment
"
failures"
19
(
33%)
9
(
14%)
28
(
23%)
"
4"
lead
hazard
treatment
"
failures"
2
(
4%)
3
(
5%)
5
(
4%)
"
5"
lead
hazard
treatment
"
failures"
1
(
2%)
0
(
0)
1
(
1)
Source:
Form
05
aThe
mean
number
of
"
failures"
per
unit
(
2.1
for
LHR
units;
1.7
for
LHR+
W
units)
was
significantly
associated
with
assigned
treatment
group
(
p=
0.0255).

1.
Types
of
Immediate
Post­
Intervention
"
Failures"

The
most
common
"
failure"
was
that
not
all
paint
was
intact
after
completion
of
the
prescribed
lead
hazard
reduction
treatments
(
i.
e.,
chipping,
flaking
and/
or
peeling
paint
remained)
(
see
Table
22
on
page
50).
Almost
75
percent
of
all
units
(
77
percent
of
the
LHR
units
and
73
percent
of
the
LHR+
W
units)
exhibited
evidence
of
this
"
failure"
type.
Almost
40
percent
of
all
units
(
44
percent
of
the
LHR
units
and
33
percent
of
the
LHR+
W
units)
had
evidence
of
visible
paint
chips
and/
or
debris
remaining.
It
should
be
noted
that
the
areal
density
(
i.
e.,
depth)
and
the
lead
concentration
of
the
chips/
debris
were
not
measured.
In
43
percent
of
all
units
(
49
percent
of
the
LHR
units
and
38
percent
of
the
LHR+
W
units),
the
treatment
"
failure"
of
doors
with
painted
surfaces
continuing
to
rub
together
or
bind
was
observed.
In
26
percent
of
the
LHR
units,
the
window
trough
material
(
e.
g.,
vinyl
or
metal)
was
not
properly
attached.
Other
treatment
"
failure"
types
were
much
less
prevalent.
None
of
the
ten
types
of
treatment
"
failures"
were
significantly
associated
with
treatment
group
based
on
chi­
square
testing.
11/
23/
05
50
Table
22:
Immediate
Post­
Intervention
Visual
Assessment
for
Lead
Hazard
Reduction
Treatment
"
Failures"
Type
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
a
(
57
LHR
units,
64
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

Not
all
paint
intact
(
chipping,
peeling
or
flaking
paint
remains)
44
(
77%)
47
(
73%)
91
(
75%)

Visible
structural
defect
that
could
cause
paint
deterioration
remains
1
(
2%)
2
(
3%)
3
(
2%)

If
sills
stripped/
repainted
or
replaced,
not
all
treatments
in
place
2
(
4%)
6
(
9%)
8
(
7%)

If
sills
encapsulated
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attachedb
0
(
0%)
1
(
2%)
1
(
1%)

If
troughs
capped
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attachedb
15
(
26%)
N/
Ac
15
(
12%)
d
For
windows,
not
all
top
sashes
are
fixed
in
place
0
(
0%)
N/
Ac
0
(
0)
d
For
doors,
painted
surfaces
continue
to
rub
together
28
(
49%)
24
(
38%)
52
(
43%)

Some
bare
floors
are
not
smooth
and
cleanable
2
(
4%)
4
(
6%)
6
(
5%)

Not
all
kitchen
and
bathroom
floors
are
overlaid
with
water­
resistant
coverings
4
(
7%)
4
(
6%)
8
(
7%)

Visible
paint
chips
or
debris
remains
25
(
44%)
21
(
33%)
46
(
38%)

Source:
Form
05
aNone
of
these
"
failure"
types
were
significantly
associated
with
the
assigned
treatment
group
(
p­
values
vary
from
0.1949
to
0.633).
bProperly
attached
is
defined
as
correctly
installed,
with
caps
held
firmly
in
place
with
a
sufficient
number
of
fittings
(
e.
g.,
nails)
and
with
caulking
completed.
cIn
LHR+
W
units,
these
treatments
were
not
necessary
by
virtue
of
the
installation
of
new
vinyl
replacement
windows
with
unpainted
top
sashes
and
troughs.
dBoth
values
are
understated
due
to
the
inapplicability
of
these
treatments
for
LHR+
W
units.

The
geometric
means
of
"
failures"
presented
in
Table
23
on
page
51
were
calculated
by
first
summing,
within
a
unit,
all
reports
of
a
particular
"
failure"
type
having
the
same
measurement
units
(
e.
g.,
ft2),
then
calculating
the
geometric
mean
across
dwelling
units.
These
data
indicate
that
problems
recorded
by
LAAP
inspectors
were
easily
observed.
For
example,
LAAP
inspectors
reported
a
dwelling
unit
geometric
mean
of
1
ft2
of
non­
intact
paint
in
both
LHR
and
LHR+
W
units
at
immediate
post­
intervention.
This
type
of
"
failure"
was
reported
63
times
in
31
LHR
units,
and
reported
78
times
in
34
LHR+
W
units,
indicating
that
this
was
a
widespread
and
large
problem.
"
Visible
paint
chips
and
debris
remaining"
at
Phase
2
was
also
a
extensive
problem;
a
geometric
mean
of
8
ft2
of
visible
chips
remained
in
LHR
units
and
16
ft2
remained
in
LHR+
W
units,
with
this
"
failure"
being
reported
29
times
in
19
LHR
units,
and
48
times
in
21
LHR+
W
units.
As
previously
noted,
the
areal
density
(
i.
e.,
depth)
and
the
lead
concentration
of
these
chips/
debris
were
not
measured.
"
Painted
door
surfaces
rubbing
together"
was
reported
40
times
in
28
LHR
units
and
38
times
in
24
LHR+
W
units,
with
at
least
one
door
noted
as
a
problem
in
each
report.
The
status
of
floors
at
Phase
2
was
generally
less
widely
reported
but
11/
23/
05
51
nonetheless
noteworthy,
with
a
geometric
mean
of
4
ft2
of
bare
floors
in
LHR
units
and
2
ft2
in
LHR+
W
units
remaining
not
smooth
and
cleanable.
A
geometric
mean
of
2
ft2
of
kitchen
or
bathroom
floors
in
LHR
units
and
1
ft2
in
LHR+
W
units
were
reportedly
not
overlaid
with
waterresistant
coverings
at
immediate
post­
intervention.

Table
23:
Reporting
Frequency
of
Specified
Types
of
Treatment
"
Failures"
at
Immediate
Post­
Intervention
LHR
LHR+
W
"
Failure"
Type
(
57
LHR,
64
LHR+
W)
Unit
of
Measure
#
Units
with
Specified
"
Failure
"
Rpt'd
Reporting
Freq.
of
"
Failure"
Type
at
Ph
2
"
Failure"
Magnit.
(
geom.
mean
&
95%
Conf
Limits)
#
Units
with
Specified
"
Failure
"
Rpt'd
Reporting
Freq.
of
"
Failure"
Type
at
Ph
2
"
Failure"
Magnit.
(
geom.
mean
&
95%
Conf
Limits)
EAa
8
19
2
(
1,
6)
1
1
1c
LF
16
30
2
(
1,
4)
20
27
1
(
1,
2)
Not
all
paint
intact
(
44
LHR,
47
LHR+
W)
SF
31
63
1
(
0.2,
1)
34
78
1
(
0.4,
3)

For
doors,
painted
surfaces
continued
to
rub
together
(
28
LHR,
24
LHR+
W)
EAa
28
40
1
(
1,
2)
24
38
1
(
1,
2)

Some
bare
floors
not
smooth
&
cleanable
(
2
LHR,
4
LHR+
W)
SF
2
5
4c
4
8
2c
Not
all
kitchen
&
bathroom
floors
overlaid
with
water­
resistant
coverings
(
4
LHR,
4
LHR+
W)
SF
4
4
2
(
1,
4)
4
4
1c
EAb
6
25
4
(
2,
13)
1
1
2c
LF
4
10
21c
0
0
­­
Visible
paint
chips
or
debris
remains
(
25
LHR,
21
LHR+
W)
SF
19
29
8
(
3,
19)
21
48
16
(
6,
42)

EA=
each;
LF=
linear
foot;
SF=
square
foot
aFor
these
"
failure"
types,
"
each"
denotes
window
components
(
e.
g.,
upper
sash,
lower
sash),
door
casings,
or
handrails
in
need
of
re­
painting.
LF/
SF
values
are
not
available
for
these
occurrences.
bFor
these
"
failure"
types,
"
each"
denotes
a
window
in
need
of
re­
cleaning.
cNo
95%
confidence
limits
were
calculated
due
to
the
small
number
of
units
having
this
failure
reported.
11/
23/
05
52
2.
Number
of
Rooms
with
"
Failures"
at
Phase
II
In
general,
two
to
three
rooms
per
unit
had
treatment
"
failures;"
few
units
had
"
failures"
in
only
one
room
or
in
more
than
four
rooms
(
see
Figure
9).

Figure
9:
Percentage
of
Units
Having
Specified
Number
of
Rooms
Per
Unit
with
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"

10%

9%
16%
21%
21%

11%

4%

2%
4%

0%
4%
6%
17%
28%

23%

12%

3%
3%
3%
3%

0%
0%

0%
5%
10%
15%
20%
25%
30%

0
1
2
3
4
5
6
7
8
9
10
Percentage
of
Units
Having
Specified
Number
of
Rooms
with
Visual
Assessment
Treatment
"
Failures"
at
Immediate
Post­
Intervention
Percentage
of
Units
LHR
Units
LHR+
W
Units
In
the
57
LHR
units,
considering
only
those
rooms
with
"
failures"
that
are
applicable
to
both
LHR
and
LHR+
W
units,
the
mean
number
of
rooms
with
lead
hazard
control
treatment
"
failures"
was
3.1
rooms
per
unit,
against
a
mean
total
number
of
6.6
rooms
per
unit.
In
the
64
LHR+
W
units,
the
mean
number
of
rooms
with
"
failures"
was
2.6
rooms,
against
a
mean
total
number
of
6.7
rooms
per
unit.
This
mean
number
of
rooms
with
prescribed
treatment
"
failures"
was
not
significantly
associated
with
the
assigned
treatment
group
(
p­
value=
0.206).
Table
24
on
page
53
presents
data
on
the
mean
number
of
rooms
per
unit
with
a
given
"
failure"
type.

3.
Type
of
Contractor
and
Number
of
"
Failures"
at
Immediate
Post­
Intervention
For­
profit
contractors,
the
property
owner/
employees
or
non­
profit
contractors
performed
lead
hazard
reduction
treatments.
(
Only
five
units
were
treated
by
non­
profit
contractors;
therefore,
these
data
are
excluded
from
the
study.)
For
LHR
units,
the
mean
number
of
"
failures"
per
unit
for
for­
profit
contractor­
treated
units
and
property
owner/
employee­
treated
units
were
3.3
and
5.5,
respectively
(
see
Table
25
on
page
53).
For
LHR+
W
units,
the
mean
numbers
of
"
failures"
per
unit
for
for­
profit
contractors
and
property
owners/
employees­
treated
units
were
3.1
and
3.0,
respectively.
There
was
no
significant
association
between
the
mean
number
of
"
failures"
per
unit
and
the
contractor
type
based
on
a
Fisher's
Exact
Test
(
p­
value=
0.4823).
11/
23/
05
53
Table
24:
Mean
Number
of
Rooms
Per
Unit
with
Specified
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
by
Treatment
Groupa
Type
of
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"
(
57
LHR
units,
64
LHR+
W
units)
LHR
Mean
#
Rooms
Per
Unit
(
95%
Conf.
Int.)
LHR+
W
Mean
#
Rooms
per
Unit
(
95%
Conf.
Int.)
Not
all
paint
intact
(
chipping,
peeling
or
flaking
paint
remains)
1.7
(
1.4;
2.1)
1.4
(
1.1;
1.7)
Visible
structural
defect
that
could
cause
paint
deterioration
remains
0.02
(
0;
0.05)
0
(
0;
0.08)
If
sills
stripped/
repainted
or
replaced,
not
all
treatments
in
place
0.05
(
0;
0.11)
0.2
(
0.07;
0.3)
If
sills
encapsulated
with
vinyl,
metal,
etc.,
not
all
material
is
property
attached
0
0
If
troughs
capped
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attached
0.5
(
0.35;
0.7)
N/
Ab
For
windows,
not
all
top
sashes
are
fixed
in
place
0
N/
Ab
For
doors,
painted
surfaces
continue
to
rub
together
0.7
(
0.5;
0.9)
0.6
(
0.4;
0.8)
Some
bare
floors
are
not
smooth
and
cleanable
0.09
(
0.01;
0.2)
0.1
(
0;
0.2)
Not
all
kitchen
and
bathroom
floors
are
overlaid
with
water­
resistant
coverings
0.07
(
0;
0.1)
0.1
(
0;
0.1)
Visible
paint
chips
or
debris
remains
1.1
(
0.8;
1.4)
0.7
(
0.5;
1.0)
Source:
Form
05
aMean
number
of
rooms
with
prescribed
treatment
"
failures"
was
not
significantly
associated
with
the
assigned
treatment
group
(
p=
0.206).
bIn
LHR+
W
units,
these
treatments
were
not
necessary
by
virtue
of
the
installation
of
new
vinyl
replacement
windows
with
unpainted
top
sashes
and
troughs.

Table
25:
Mean
Number
of
Immediate
Post­
Intervention
"
Failures"
per
Unit
by
the
Type
of
Contractora,
b
Mean
Number
of
Immediate
Post­
Intervention
"
Failures"
per
Unit
Contractor
Type
LHR
Units
LHR+
W
Units
For­
profit
contractor
3.3
3.1
Property
owner/
employees
5.5
3.0
Source:
Form
05
a
Because
only
a
limited
number
of
LHR
and
LHR+
W
units
(
i.
e.,
five
combined)
had
nonprofit
contractors
performing
the
work,
the
number
of
"
failures"
is
not
easily
compared
against
for­
profit
contractors
and
property
owner/
employees
work.
Therefore,
nonprofit
contractor
data
are
not
presented
in
this
table.
bThe
mean
number
of
"
failures"
per
unit
and
contractor
type
were
not
significantly
associated
(
p=
0.4823).
11/
23/
05
54
C.
Description
of
and
Factors
Affecting
Costs
Various
cost
data
were
collected
during
the
study,
including
the
LAAP­
approved
costs
for
the
prescribed
lead
hazard
reduction
treatments,
plus
the
additional
window
replacement
costs
for
the
LHR+
W
units.
The
study
also
collected
owner­
estimated
turnover
costs
that
would
have
been
incurred
even
if
the
enrolled
units
had
not
been
subject
to
the
EA
6­
8
risk
reduction
requirements.
However,
because
these
latter
data
were
based
on
interviews
with
property
owners
and/
or
their
representatives,
they
are
considered
highly
subjective
and
were
excluded
from
this
study.
Finally,
the
study
determined
what
type
of
contractor
(
e.
g.,
for
profit,
nonprofit
or
property
owner/
employee)
performed
the
work
and
whether
any
concurrent
work
(
e.
g.,
roof,
plumbing
and
electrical
repairs)
was
completed
in
addition
to
the
lead­
related
treatments.

The
costs
presented
in
this
section
are
those
considered
to
be
solely
associated
with
EA
6­
8
prescribed
treatments.
In
other
words,
measures
that
are
not
otherwise
contained
in
existing
housing
code
provisions
and/
or
part
of
a
normal
rental
unit
turnover
regimen
would
be
included
when
determining
the
incremental
cost
of
complying
with
the
statute.
The
measures
that
are
considered
unique
to
EA
6­
8
include:

 
Stripping
and
repainting,
replacing
or
encapsulating
all
interior
window
sills
with
vinyl,
metal
or
any
other
material;
 
Ensuring
that
caps
of
vinyl,
aluminum
or
other
material
are
installed
in
all
window
troughs
in
order
to
make
the
window
troughs
smooth
and
cleanable;
 
Except
for
a
treated
or
replacement
window
that
is
free
of
lead­
based
paint
on
its
friction
surfaces,
fixing
the
top
sash
of
all
windows
in
place
in
order
to
eliminate
the
friction
caused
by
movement
of
the
top
sash;
 
Re­
hanging
all
doors
necessary
in
order
to
prevent
the
rubbing
together
of
a
lead­
painted
surface
with
another
surface;
 
Making
all
bare
floors
smooth
and
cleanable;
and
 
HEPA­
vacuuming
and
washing
of
the
interior
of
the
affected
property
with
high
phosphate
detergent
or
its
equivalent.

Depending
on
the
nature
of
certain
bathroom
and
kitchen
floor
repairs,
another
EA
6­
8
prescribed
treatment
that
might
or
might
not
be
considered
extra
cost
is
ensuring
that
all
kitchen
and
bathroom
floors
are
overlaid
with
a
smooth,
water­
resistant
covering.

As
discussed
in
Section
II.
B,
participating
property
owners
received
up
to
80
percent
of
their
full
lead
hazard
control
costs
from
LAAP
in
the
form
of
a
loan/
grant,
subject
to
a
spending
cap
of
$
10,000.
The
remaining
20
percent
was
self­
funded.
LAAP
funding
was
provided
through
the
HUD
Round
III
Lead
Hazard
Control
Grant.
However,
as
part
of
this
study
and
in
order
to
determine
the
incremental
cost
of
complying
with
EA
6­
8,
LAAP
inspectors
reviewed
and
revised
the
owners'
original
lead
hazard
reduction
costs
by
subtracting
cost
items
that
could
be
considered
code­
related
and/
or
part
of
normal
unit
turnover.
The
most
notable
of
these
is
painting,
which
was
often
the
largest
single
cost
element.
(
See
Appendix
H
for
the
gross
lead
hazard
reduction
costs,
80
percent
of
which
were
reimbursed
by
LAAP,
and
the
estimated
EA
6­
8
lead
hazard
reduction
treatment
costs
for
each
of
the
enrolled
units.)
11/
23/
05
55
Based
on
the
LAAP
inspectors'
estimated
EA
6­
8
lead
hazard
reduction
treatment
cost
calculations,
the
median
cost
for
completing
the
prescribed
treatments
in
LHR
units
was
$
2,154,
while
for
LHR+
W
the
median
cost
was
$
1,649.
The
LHR
units'
median
cost
is
likely
higher
than
the
LHR+
W
units'
median
cost
because
LHR
unit
costs
included
the
expense
of
the
prescribed
window
treatments,
while
window
replacement
expenses
for
LHR+
W
units
were
calculated
separately
(
see
Table
26).
However,
the
median
total
lead
cost
for
LHR+
W
units
($
4,348)
was
over
twice
that
for
LHR
units
($
2,154),
reflecting
the
added
expense
of
window
replacement.

T
able
26:
Description
of
Lead
Hazard
Reduction
Costs
Lead
Hazard
Reduction
Intervention
Costs
and
Window
Replacement
Costs
(
57
LHR
units,
64
LHR+
W
units)
LHR
5th
Percentile
25th
Percent.
Median
75th
Percent.
95th
Percent.
LHR+
W
5th
Percentile
25th
Percent.
Median
75th
Percent.
95th
Percent.

Cost
of
lead
hazard
reduction
treatments
only
$
820
$
1,462
$
2,154
$
3,015
$
3,920
$
773
$
1,168
$
1,649
$
2,212
$
3,036
Cost
of
window
replacement
only
N/
A
N/
A
N/
A
N/
A
N/
A
$
1,625
$
2,200
$
2,600
$
3,185
$
4,200
Total
Lead
Costs
(
lead
hazard
reduction
treatments
and,
if
applicable,
window
replacement)
$
820
$
1,462
$
2,154
$
3,015
$
3,920
$
2,589
$
3,750
$
4,348
$
5,173
$
6,458
Source:
Form
05
Interestingly,
when
analyzing
costs
against
the
type
of
contractor
performing
the
lead
hazard
reduction
treatments,
for­
profit
contractors'
mean
costs
were
slightly
higher
than
property
owner/
employees'
mean
costs
in
LHR
units,
but
lower
in
LHR+
W
units
(
see
Table
27
on
page
56).
11/
23/
05
56
Table
27:
Mean
Lead
Hazard
Reduction
Costs
by
Contractor
Typea
Contractor
Type
(
56
LHR,
64
LHR+
W)
LHR
Mean
Cost
(
std.
deviat.)
b
LHR+
W
Mean
Cost
(
std.
deviat.
)
b
For­
profit
contractor
(
27
LHR,
32
LHR+
W)
$
2,405
(
1,196
)
$
4,277
(
1,309
)
Property
owner/
employees
(
28
LHR,
28
LHR+
W)
$
2,214
(
770
)
$
4,639
(
1,106
)
Source:
Form
05
aBecause
only
a
limited
number
of
LHR
and
LHR+
W
units
(
i.
e.,
five
combined)
had
nonprofit
contractors
performing
the
work,
the
costs
are
not
easily
compared
against
for­
profit
contractor
and
property
owner/
employee
work.
Therefore,
nonprofit
contractor
data
are
not
presented
in
this
table.
b
Mean
costs
for
LHR
units
were
calculated
using
only
the
cost
of
prescribed
lead
hazard
reduction
treatments,
while
those
for
LHR+
W
units
also
included
additional
costs
associated
with
window
replacement.

In
the
LHR+
W
units,
an
average
of
nine
(
9)
to
10
windows
were
replaced.
(
As
previously
noted,
this
replacement
was
not
part
of
the
lead
hazard
reduction
treatments
required
by
the
statute.)
Replacement
windows
were
typically
double­
hung,
thermo­
pane
vinyl
units;
existing
window
jambs
and
exterior
casing
were
usually
wrapped
with
aluminum
coil
stock
and
any
remaining
painted
components
were
stabilized
and
repainted.
The
median
cost
per
window
replacement
was
$
275,
with
a
minimum
recorded
cost
per
window
of
$
220
and
a
maximum
recorded
cost
per
window
of
$
357.
Additional
(
i.
e.,
concurrent)
work
was
performed
in
17
of
57
LHR
units
(
29.8
percent)
and
27
of
64
LHR+
W
units
(
42.2
percent).
This
work
included
roofing,
plumbing,
heating
and
electrical
repairs
and
replacement
of
other
fixtures
and
components.

For
the
57
LHR
units,
the
mean
number
of
days
to
complete
the
intervention
was
57.8
days.
For
the
64
LHR+
W
units,
the
mean
number
of
days
to
complete
the
intervention
was
56.5
days.
These
relatively
long
periods
reflect
the
fact
that
many
property
owners/
employees
who
performed
their
own
work,
and
some
for­
profit
and
nonprofit
contractors,
worked
intermittently
over
time
for
various
reasons.
Therefore,
these
mean
numbers
are
not
representative
of
the
actual
uninterrupted
time
that
would
have
been
needed
to
complete
the
prescribed
lead
hazard
reduction
treatments.

D.
Description
of
Statistical
Modeling
Through
Immediate
Post­
Intervention
1.
Predicting
Dust
Lead
Loading
at
Immediate
Post­
Intervention
A
multiple
regression
model
with
backward
elimination9
was
run
to
identify
which
factors
would
be
significant
predictors
of
immediate
post­
intervention
dust
lead
loadings.
A
separate
model
was
run
for
each
surface
type
(
i.
e.,
bare
floors,
interior
window
sills
and
window
troughs).
The
dwelling
unit
average
of
single
surface
samples
for
each
surface
type
was
the
dependent
variable
in
each
model.
Dust
lead
loading
measurements
were
log
transformed.
(
Tables
of
statistical
results
from
these
models
are
provided
in
Appendix
I.)

9
In
a
multiple
regression
model
with
backward
elimination,
all
possible
predictors
of
the
outcome
are
initially
entered
into
the
model.
Then
hypothesis
tests
are
run
to
determine
if
any
factors
can
be
removed
from
the
predictive
equation
when
the
other
factors
are
retained.
The
least
significant
factor
(
i.
e.,
the
factor
with
the
largest
observed
significance
level)
is
removed,
and
the
process
is
repeated
to
determine
if
more
factors
can
be
dropped.
11/
23/
05
57
The
set
of
possible
predictors
considered
were:

 
Type
of
building
(
e.
g.,
rowhouse,
more
than
four
units
in
a
building,
single
detached,
etc.);
 
Year
of
construction
(
e.
g.,
pre­
1910,
1910­
1919,
1920­
1929,
etc.);
 
Contractor
type
(
i.
e.,
for­
profit,
nonprofit,
property
owner/
employee);
 
Concurrent
work
(
i.
e.,
yes
or
no);
 
Treatment
group
(
i.
e.,
LHR
or
LHR+
W);
 
Pre­
intervention
dwelling
unit
average
of
single
surface
samples
dust
lead
loading
results
for
a
given
surface
type
(
i.
e.,
bare
floors,
interior
window
sills
or
window
troughs);
 
Number
of
items
of
interior
dwelling
unit
deterioration
(
0­
4);
 
Number
of
items
of
exterior
building
deterioration
(
0­
6);
 
Percent
of
windows
replaced
out
of
the
total
number
of
windows
in
a
dwelling
unit;
 
Market
value
of
a
dwelling
unit
($);
 
Total
lead
hazard
reduction
cost
for
a
dwelling
unit,
including
window
replacement
costs,
if
applicable
($);
and
 
Interactions
of
(
1)
the
number
of
interior
dwelling
unit
deterioration
items,
(
2)
the
number
of
exterior
building
deterioration
items,
and
(
3)
pre­
intervention
dust
lead
loadings
with
each
of
the
following:
concurrent
work,
percent
of
windows
replaced,
total
lead
hazard
reduction
cost,
and
assigned
treatment
group.

(
Note:
The
relationship
between
components
of
the
visual
assessment
and
immediate
postintervention
dust
lead
loading
results
was
assessed
and
found
to
be
non­
significant
in
every
case.
Hence,
these
variables
were
not
included
in
the
models.)

On
bare
floors,
the
entire
set
of
possible
predictors
did
not
significantly
predict
immediate
postintervention
floor
dust
lead
loading
(
p­
value=
0.4687).
On
interior
window
sills,
the
variables
found
to
predict
immediate
post­
intervention
dust
loading
were
pre­
intervention
interior
window
sill
dust
lead
loading
and
treatment
group.
The
percentage
of
variation
accounted
for
in
immediate
post­
intervention
dust
lead
loading
by
these
two
variables
(
i.
e.,
R2)
is
27
percent.
On
window
troughs,
the
variables
found
to
predict
immediate
post­
intervention
dust
lead
loading
were
pre­
intervention
window
trough
dust
lead
loading
and
treatment
group.
The
percentage
of
variation
accounted
for
in
immediate
post­
intervention
dust
lead
loading
by
these
two
variables
(
i.
e.,
R2)
is
26
percent.

2.
Predicting
Dust
Lead
Loading
"
Failures"
at
Immediate
Post­
Intervention
Separate
logistic
regression
models
with
backward
elimination
were
employed
to
predict
whether
dwelling
units
would
have
clearance
"
failures"
on
bare
floors,
interior
window
sills
and
window
troughs
based
on
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2,
and
800
µ
g/
ft2,
respectively
(
see
Appendix
I).
In
a
logistic
model,
the
outcome
of
interest
is
a
binary
response
variable
such
as
"
pass/
fail."
The
same
set
of
possible
predictors
that
were
used
in
predicting
dust
lead
loading
at
immediate
post­
intervention
were
also
employed
for
this
model,
except
that
interactions
were
not
included
(
not
appropriate
for
logistic
models).

Clearance
"
failure"
on
bare
floors
was
not
modeled
since
no
relationship
was
found
when
attempting
to
predict
floor
dust
lead
loading
at
immediate
post­
intervention.
On
interior
window
11/
23/
05
58
sills,
the
only
variable
found
to
significantly
predict
immediate
post­
intervention
clearance
"
failures"
was
the
percent
of
windows
replaced
out
of
the
total
number
of
windows
in
a
dwelling
unit.
The
higher
the
percentage
of
windows
replaced,
the
lower
the
clearance
"
failures."
On
window
troughs,
the
variables
found
to
predict
immediate
post­
intervention
clearance
"
failures"
were
number
of
items
of
exterior
building
deterioration
(
increasing
number
would
yield
more
clearance
"
failures")
and
percent
of
windows
replaced
(
increasing
percent
would
yield
fewer
clearance
"
failures").

3.
Predicting
Immediate
Post­
Intervention
Visual
Assessment
"
Failures"

A
logistic
model
with
backward
elimination
was
also
employed
to
predict
whether
dwelling
units
would
have
visual
assessment
"
failures,"
again
using
the
same
set
of
possible
predictors
employed
for
predicting
dust
lead
loading
at
immediate
post­
intervention
and
dust
lead
loading
"
failures."
The
only
variable
found
to
significantly
influence
visual
assessment
"
failure"
was
the
number
of
items
of
exterior
building
deterioration;
increasing
numbers
would
yield
more
visual
assessment
"
failures."

IX.
ONE­
YEAR
POST­
INTERVENTION
(
PHASE
III)
RESULTS
On
the
average,
one­
year
post­
intervention
visits
occurred
55
weeks
after
the
immediate
postintervention
sampling
visit,
with
a
minimum
of
40
weeks
and
a
maximum
of
74
weeks
between
the
two
visits.
Of
the
121
units
that
had
the
prescribed
lead
hazard
reduction
treatment
interventions
and
a
full
immediate
post­
intervention
assessment,
a
total
of
73
units
(
36
LHR
and
37
LHR+
W)
completed
one­
year
post­
intervention
dust
lead
sampling
and
a
full
one­
year
postintervention
visual
assessment.
The
relationship
between
various
one­
year
post­
intervention
factors
and
the
assigned
treatment
group
(
i.
e.,
LHR
versus
LHR+
W
units)
was
examined
using
various
statistical
tests.

A.
Physical
Characteristics
and
Baseline
Condition
of
Enrolled
Units
Having
One­
Year
Post­
Intervention
Data
Section
V.
A
and
Appendices
C
and
D
contain
detailed
discussions
and
lists
of
the
physical
characteristics
and
baseline
conditions
for
the
177
units
that
underwent
complete
preintervention
visual
assessments,
including
those
having
one­
year
post­
intervention
data.
Because
these
variables
may
have
influenced
one­
year
post­
intervention
results,
and
because
these
variables
may
be
different
for
the
smaller
one­
year
post­
intervention
dataset
of
73
units,
this
section
summarizes
baseline
characteristics
and
conditions,
including
pre­
intervention
dust
lead
loadings,
for
the
73
units
for
which
one­
year
post­
intervention
data
were
collected.

1.
Baseline
Physical
Characteristics
for
Enrolled
Units
Having
One­
Year
Post­
Intervention
Data
The
majority
(
81%)
of
the
73
one­
year
post­
intervention
units
were
single­
family
rowhouses,
two
were
single
detached
residences,
and
the
remaining
units
were
present
in
multifamily
buildings
(
two
in
duplexes,
one
in
a
two­
flat,
two
in
a
triplex,
and
seven
in
building
with
more
11/
23/
05
59
than
4
units).
This
grouping
is
not
significantly
different
from
that
of
the
larger
dataset
based
on
Fisher's
exact
test
(
p=
0.39
and
0.14
in
LHR
and
LHR+
W
units,
respectively).

Eighty­
five
percent
of
the
one­
year
post­
intervention
units
were
constructed
70
or
more
years
ago,
with
the
largest
percentage
(
42%)
built
between
1920
and
1929.
Controlling
for
treatment
group,
this
grouping
is
not
significantly
different
from
that
of
the
larger
dataset
discussed
in
Section
VII.
A
based
on
Mantel­
Haenszel's
mean
score
statistic
(
p=
0.28).
Other
baseline
dwelling
unit
characteristics
are
summarized
in
Table
28.

Table
28:
Dwelling
Unit
Baseline
Characteristics
for
One­
Year
Post­
Intervention
Units
Dwelling
Unit
Characteristic
(
36
LHR,
37
LHR+
W
Units)
Mean
(
std.
deviat.)
LHR
Units
Mean
(
std.
deviat.)
LHR+
W
Units
Mean
(
std.
deviat.)
All
Units
Estimated
market
value
of
dwelling
unit
(
mean
rounded
to
nearest
$
100)
$
23,400
(
12,400)
$
24,000
(
14,800)
$
23,700
(
13,600)
#
of
rooms
in
dwelling
unit
7
(
1)
7
(
1)
7
(
1)
#
of
bedrooms
in
dwelling
unit
3
(
1)
3
(
1)
3
(
1)
#
of
windows
in
dwelling
unit
10
(
3)
11
(
4)
10
(
3)
#
of
unpainted
windows
in
dwelling
unit
(
e.
g.,
vinyl,
aluminum)
2
(
4)
1
(
2)
1
(
3)

Approximate
square
feet
of
living
space
1,100
(
280)
1,100
(
330)
1,100
(
305)

2.
Baseline
Exterior
Building
and
Interior
Unit
Conditions
for
One­
Year
Post­
Intervention
Units
Baseline
exterior
building
conditions
for
the
73
one­
year
post­
intervention
units
are
summarized
in
Table
29
on
page
60,
which
lists
the
seven
items
checked
by
the
LAAP
inspectors
on
their
initial
visit
to
each
unit.
Table
30
on
page
60
lists
the
number
of
observed
items
of
exterior
building
deterioration
in
the
73
one­
year
post­
intervention
units.
In
general,
there
were
minimal
observed
differences
in
exterior
deterioration
between
the
LHR
and
LHR+
W
units,
with
none
of
the
types
of
exterior
building
deteriorations
being
significantly
associated
with
assigned
treatment
group
based
on
Fisher's
exact
tests.

Based
on
Cochran­
Mantel­
Haenszel
statistical
testing,
the
number
of
observed
items
of
exterior
building
deterioration
was
also
not
significantly
associated
with
assigned
treatment
group
(
p=
0.37;
see
Table
30
on
page
60).
Controlling
for
treatment
group,
the
73­
unit
one­
year
postintervention
dataset
tended
to
have
more
exterior
deteriorations
than
the
larger
dataset
discussed
in
Section
VII.
A.
2
based
Mantel­
Haenszel
mean
score
statistic
(
p<
0.001).
11/
23/
05
60
Table
29:
Baseline
Exterior
Building
Deterioration
Summary
for
One­
Year
Post­
Intervention
Units
Types
of
Exterior
Building
Deteriorationa
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

Roofs,
gutters,
downspouts
 
missing,
broken,
holes,
cracks
3
(
8%)
7
(
19%)
10
(
14%)

Chimney
 
cracked,
bricks
loose
or
missing,
unstable,
out
of
plumb
1
(
3%)
1
(
3%)
2
(
3%)

Walls
and
siding
 
large
cracks
or
holes,
boards
or
shingles
broken
or
missing
4
(
11%)
6
(
16%)
10
(
14%)

Windows
and
doors
 
 
two
windows
or
doors
broken,
missing,
boarded
up
7
(
19%)
9
(
24%)
16
(
22%)

Porch
or
steps
 
major
elements
broken,
missing,
out
of
plumb
2
(
6%)
4
(
11%)
6
(
8%)

Foundation
 
major
visible
cracks,
missing
materials,
unsound
0
1
(
3%)
1
(
1%)

Evidence
of
exterior
ground
contaminated
with
paint
chips
7
(
19%)
10
(
27%)
17
(
23%)

Source:
Form
02
a
Based
on
Fisher's
Exact
test,
none
of
the
7
types
of
exterior
building
deterioration
were
significantly
associated
with
assigned
treatment
group
(
p
varied
from
0.31
to
1).

Table
30:
Number
of
Baseline
Exterior
Building
Deterioration
Items
for
One­
Year
Post­
Intervention
Units
#
Items
of
Exterior
Deterioration
per
Buildinga
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)
"
0"
items
of
exterior
deterioration
19
(
53%)
15
(
40%)
34
(
47%)
"
1"
item
of
exterior
deterioration
12
(
33%)
13
(
35%)
25
(
34%)
"
2"
items
of
exterior
deterioration
3
(
8%)
5
(
14%)
8
(
11%)
"
3"
items
of
exterior
deterioration
2
(
6%)
1
(
3%)
3
(
4%)
"
4"
items
of
exterior
deterioration
0
3
(
8%)
3
(
4%)
Source:
Form
02
aBased
on
Cochran­
Mantel­
Haenszel
test,
the
number
of
exterior
deterioration
items
was
not
significantly
associated
with
assigned
treatment
group
(
p=
0.37).

Baseline
interior
dwelling
unit
conditions
for
the
73
one­
year
post­
intervention
units
are
summarized
in
Table
31
on
page
61,
which
lists
the
five
items
checked
by
LAAP
inspectors
during
their
initial
visit
to
each
unit,
and
in
Table
32
on
page
61,
which
lists
the
number
of
observed
items
of
interior
unit
deterioration
in
the
73
units.
LHR+
W
units
had
higher
percentages
of
interior
deterioration
than
LHR
units
for
three
of
the
five
items.
Most
notably,
LHR+
W
units
were
more
likely
than
LHR
units
to
have
walls,
ceilings,
doors,
and
trim
in
need
of
repair,
replacement
or
repainting.
This
type
of
interior
deterioration
was
significantly
associated
with
assigned
treatment
group
based
on
Fisher's
exact
testing
(
p=
0.0559).
Controlling
for
treatment
group,
the
one­
year
post­
intervention
dataset
did
not
significantly
differ
11/
23/
05
61
from
the
dataset
discussed
in
Section
VII.
A.
2
in
terms
of
the
number
of
interior
deteriorations
based
on
a
Mantel­
Haenszel
mean
score
statistic
(
p=
0.28).

Table
31:
Summary
of
Baseline
Interior
Dwelling
Unit
Deterioration
for
One­
Year
Post­
Intervention
Units
Type
of
Interior
Dwelling
Unit
Deterioration
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)
Walls,
ceiling,
doors
and
trim
 
cracks,
need
for
repair,
replace
or
major
repaintinga
10
(
28%)
19
(
51%)
29
(
40%)

Floors
 
loose,
missing
or
cracked,
finish
worn,
deteriorated
carpeting
21
(
58%)
25
(
68%)
46
(
63%)

Heating/
cooling
and
plumbing
 
need
for
extensive
repair
5
(
14%)
5
(
14%)
10
(
14%)

Interior
damage
due
to
roof
leak
 
need
for
extensive
repair
5
(
14%)
5
(
14%)
10
(
14%)

Rooms
with
>
2
ft2
of
paint
deterioration
on
walls,
woodwork,
doors,
windows,
etc.
27
(
75%)
32
(
89%)
59
(
81%)

Source:
Form
02
aBased
on
Chi­
square
testing,
deterioration
of
walls,
ceilings,
doors
and
trim
was
significantly
associated
with
assigned
treatment
group
(
p=
0.0559).
None
of
the
other
four
types
of
interior
deterioration
were
associated
with
treatment
group
(
p
varied
from
0.47
to
1).

The
total
number
of
baseline
interior
unit
deterioration
items
for
the
one­
year
post­
intervention
units
was
not
significantly
associated
with
assigned
treatment
group
(
p=
0.5956;
see
Table
32).

Table
32:
Number
of
Baseline
Interior
Dwelling
Unit
Deterioration
Items
for
One­
Year
Post­
Intervention
Units
Number
of
Items
of
Interior
Deterioration
per
Dwelling
Unit
(
36
LHR
Units,
37
LHR+
W
Units)
a
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
items
of
interior
deterioration
11
(
31%)
7
(
19%)
18
(
25%)
"
1"
item
of
interior
deterioration
13
(
36%)
13
(
35%)
26
(
35%)
"
2"
items
of
interior
deterioration
8
(
22%)
10
(
27%)
18
(
25%)
"
3"
items
of
interior
deterioration
4
(
11%)
7
(
19%)
11
(
15%)
Source:
Form
02
aTotal
number
of
items
of
interior
deterioration
was
not
significantly
associated
with
assigned
treatment
group
based
on
Cochran­
Mantel­
Haenszel
testing
(
p=
0.5956).

3.
Pre­
Intervention
Dust
Lead
Loadings
for
the
73
One­
Year
Post­
Intervention
Units
Pre­
Intervention
Composite
Dust
Lead
Loadings.
Composite
dust
samples
were
collected
from
bare
floors,
interior
window
sills,
window
troughs
and
when
necessary
from
carpeted
floors
in
all
one­
year
post­
intervention
LHR
and
LHR+
W
units.
Because
very
few
one­
year
postintervention
units
had
carpeted
floors,
sample
results
for
carpeted
floors
were
excluded
from
this
report.
For
the
three
surface
types,
the
median
pre­
intervention
dust
lead
loadings
in
the
LHR+
W
units
were
greater
than
those
in
the
LHR
units
(
Table
33
on
page
62).
Pre­
intervention
11/
23/
05
62
composite
dust
lead
loadings
were
marginally
significantly
associated
with
assigned
treatment
group
for
bare
floors
and
window
troughs
(
p=
0.0676
and
0.0547,
respectively)
but
not
for
window
sills
(
p=
0.1109).
Controlling
for
treatment
group,
the
73­
unit
one­
year
post­
intervention
dataset
was
not
significantly
different
from
the
larger
pre­
intervention
dataset
(
see
Section
VII.
B.
1)
in
terms
of
pre­
intervention
composite
dust
lead
loadings
on
bare
floors,
sills
and
troughs
(
p=
0.16,
0.57
and
0.40,
respectively).

Table
33:
Pre­
Intervention
Composite
Sample
Dust
Lead
Loading
for
One­
Year
Post­
Intervention
Units
Surface
Type
LHR
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsa,
b
(
34
LHR
units,
34
LHR+
W
units)
30
64
120
639
36,400
251
(
127;
498)
38
128
454
1,635
16,800
495
(
263;
933)
Interior
window
sillsa,
b
(
35
LHR
units,
37
LHR+
W
units)
31
384
1,739
7,520
32,366
1,283
(
592;
2,780)
287
979
2,930
9,920
234,305
3,850
(
1,982,
7,480)
Window
troughsa,
b
(
35
LHR
units,
37
LHR+
W
units)
67
2,050
10,219
34,648
132,166
7,070
(
3,318;
15,065)
1,774
4,610
12,524
128,571
447,429
21,888
(
11,578;
41,378)
Source:
Form
04A,
Phase
III
aApproximately
95%
of
bare
floor
composites,
interior
window
sill
composites,
and
window
trough
composites
were
comprised
of
4
sub­
samples;
the
rest
were
comprised
of
either
2
or
3
sub­
samples.
b
Based
on
a
Kruskal­
Wallis
test,
median
pre­
intervention
composite
dust
lead
loadings
were
marginally
significantly
associated
with
assigned
treatment
group
for
bare
floors
and
window
troughs
(
p=
0.0676
and
0.0547,
respectively)
but
not
for
window
sills
(
p=
0.1109).

Pre­
Intervention
Single
Surface
Dust
Lead
Loadings.
Single
surface
dust
samples
were
collected
from
bare
floors,
interior
window
sills,
window
troughs
and,
when
necessary,
from
carpeted
floors
in
all
one­
year
post­
intervention
LHR
and
LHR+
W
units.
Because
very
few
one
11/
23/
05
63
year
post­
intervention
units
had
carpeted
floors,
sample
results
for
carpeted
floors
were
excluded
from
this
report.
As
previously
discussed,
single
surface
samples
were
collected
in
a
subset
of
LHR
units.
However,
due
to
budget
constraints,
LAAP
ceased
collecting
such
samples
in
LHR
units
during
one­
year
and
two­
years
post­
intervention,
but
continued
to
collect
single
surface
samples
in
LHR+
W
units.
Therefore,
pre­
intervention
single
surface
results
are
summarized
only
for
LHR+
W
units
(
see
Table
34).

Table
34:
Pre­
Intervention
Single
Surface
Sample
Dust
Lead
Loadings
for
One­
Year
Post­
Intervention
Unitsa
Surface
Type
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Baltimore
HUD
Unitsb
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsb
(
36
LHR+
W
units;
249
Balt
HUD
units)
52
154
440
2,296
72,761
646
(
319;
1,305)
36
114
289
629
1,928
270
(
231;
317)
Interior
window
sillsb
(
36
LHR+
W
units)
122
772
1,987
8,797
363,045
3,498
(
1,605;
7,622)
254
909
2,732
7,142
30,498
2,670
(
2,222;
3,209)
Window
troughsb
(
36
LHR+
W
units)
381
2,039
17,083
131,591
547,652
17,107
(
8,079;
36,222)
136
1,300
3,909
14,263
75,911
4,114
(
3,253;
5,204)
Source:
Form
04B
aOnly
two
LHR
units
had
Phase
III
single
surface
samples
collected;
therefore,
no
pre­
intervention
single
surface
results
are
presented
for
LHR
units.
b
Baltimore
HUD
unit
data
are
from
Baltimore
Round
One
housing
units
enrolled
in
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program.
Baltimore
HUD
unit
data
are
for
units
that
were
vacant
at
Phase
I.
Both
interior
floor
and
entry
floor
data
are
included.
cBased
on
Wilcoxon
rank
sum
testing,
pre­
intervention
median
single
surface
dust
lead
loadings
for
floors
were
marginally
significant
(
p=
0.045),
sills
were
not
significantly
different(
p=
0.969)
and
troughs
were
significantly
different
(
p=
0.001)
from
those
for
Baltimore
HUD
units.
11/
23/
05
64
Median
pre­
intervention
single
surface
dust
lead
loadings
on
bare
floors
and
troughs
in
LHR+
W
units
were
greater
than
those
in
Baltimore
HUD
units,
while
those
on
sills
in
LHR+
W
units
were
less
than
those
in
Baltimore
HUD
units.
Looking
at
other
percentile
values,
it
generally
appears
that
LHR+
W
units
had
greater
baseline
dust
lead
loadings
than
did
Baltimore
HUD
units,
implying
that
LHR+
W
units
were
in
worse
condition
before
treatment.
However,
Wilcoxon
rank
sum
testing
indicates
that
pre­
intervention
median
single
surface
dust
lead
loadings
for
floors
were
marginally
significantly
different
(
p­
value=
0.045),
those
for
sills
were
not
significantly
different
(
p­
value=
0.969),
and
those
for
troughs
were
significantly
different
(
p­
value=
0.001)
from
those
in
Baltimore
HUD
units.

B.
One­
Year
Post­
Intervention
Composite
and
Single
Surface
Dust
Sampling
Results
A
complete
listing
of
one­
year
post­
intervention
composite
and
single
surface
dust
sample
results
is
provided
in
Appendices
E
and
F,
respectively.

1.
Summary
of
One­
Year
Post­
Intervention
Composite
Dust
Sampling
Results
During
one­
year
post­
intervention,
composite
dust
samples
were
taken
from
bare
floors,
interior
window
sills,
window
troughs
and,
when
necessary,
carpeted
floors
in
all
enrolled
LHR
and
LHR+
W
units.
Because
very
few
one­
year
post­
intervention
units
had
carpeted
floors,
sample
results
for
carpeted
floors
were
excluded
from
this
report.
A
Wilcoxon
rank
sum
test
was
used
to
test
the
equality
of
the
median
dust
lead
loadings
for
the
assigned
treatment
group
(
i.
e.,
LHR
versus
LHR+
W
units).
Differences
in
median
one­
year
post­
intervention
dust
lead
loadings
between
the
two
groups
had
to
be
accounted
for
before
the
continued
effectiveness
of
the
prescribed
lead
hazard
reduction
treatments
could
be
fully
evaluated.

In
both
LHR
and
LHR+
W
units,
one­
year
post­
intervention
median
composite
dust
lead
loadings
for
bare
floors
were
substantially
less
than
the
floor
clearance
standard
of
100
µ
g/
ft2;
however,
11%
of
LHR
units
and
6%
of
LHR+
W
units
had
at
least
one
composite
bare
floor
result
above
100
µ
g/
ft2.
One­
year
post­
intervention
median
window
sill
composite
results
for
both
treatment
groups
were
less
than
500
µ
g/
ft2;
however,
47%
of
LHR
units
and
11%
of
LHR+
W
units
had
composite
sill
results
above
this
level.
Median
loadings
for
LHR
units
were
three
times
higher
than
those
for
LHR+
W
units.
For
troughs,
one­
year
post­
intervention
median
composite
dust
lead
loadings
in
LHR+
W
units
were
also
less
than
the
clearance
standard
of
800
µ
g/
ft2;
however,
the
LHR
median
trough
result
of
1,074
µ
g/
ft2
exceeded
this
clearance
standard.
Over
half
(
58%)
of
LHR
units
but
only
16%
of
LHR+
W
units
had
one­
year
post­
intervention
trough
composite
results
above
800
µ
g/
ft2.
For
the
window
surface
types,
the
median
one­
year
post­
intervention
composite
dust
lead
loadings
in
the
LHR
units
substantially
exceeded
those
for
the
LHR+
W
units.
Based
on
Wilcoxon
rank
sum
tests,
these
differences
were
significant
for
sills
and
troughs
11/
23/
05
65
(
p=
0.0052
and
0.0025,
respectively)
but
not
for
bare
floors
(
p=
1.0;
see
Table
35
and
Figure
10
on
page
66).

Table
35:
One­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loadings
Surface
Type
LHR
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsa,
b
(
35
LHR
units,
33
LHR+
W
units)
2
7
22
75
203
23
(
14;
38)
2
11
20
38
250
23
(
15;
36)
Interior
window
sillsa,
b
(
36
LHR
units,
37
LHR+
W
units)
1
85
449
1,045
8,466
252
(
99;
640)
9
29
121
239
1,666
90
(
56;
145)
Window
troughsa,
b
(
35
LHR
units,
37
LHR+
W
units)
2
170
1,074
11,550
56,707
1,061
(
403;
2,795)
22
118
247
510
1,731
242
(
160;
367)
Source:
Form
04A,
Phase
III
aApproximately
85%
of
bare
floor
composites,
interior
window
sill
composites,
and
window
trough
composites
were
comprised
of
4
sub­
samples;
the
rest
were
comprised
of
either
2
or
3
sub­
samples.
b
Based
on
Wilcoxon
rank
sum
tests,
median
one
year
post­
intervention
dust
lead
loadings
for
sills
and
troughs
were
significantly
associated
with
assigned
treatment
group
(
p=
0.0052
and
0.0025,
respectively)
but
not
for
bare
floors
(
p=
1.0).
11/
23/
05
66
Figure
10:
Pre­
and
One­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loading
Box
Plot
1
10
100
1000
10000
100000
1000000
Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

LHR
LHR+
W
LHR
LHR+
W
LHR
LHR+
W
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.

2.
Change
in
Composite
Results
from
Pre
to
One­
Year
Post­
Intervention
Given
that
all
visual
assessment
"
failures"
found
at
immediate
post­
intervention
were
corrected
upon
discovery
and
that
all
immediate
post­
intervention
samples
were
collected
prior
to
these
corrections,
it
is
not
appropriate
to
compare
immediate
post­
intervention
dust
lead
loadings
with
those
measured
at
one­
year
post­
intervention.
However,
comparison
of
one­
year
postintervention
results
with
pre­
intervention
results
aids
in
evaluating
treatment
effectiveness
one
year
after
intervention.
11/
23/
05
67
Large
median
percent
decreases
and
median
µ
g/
ft2
decreases
were
observed
when
measuring
the
change
from
pre
to
one­
year
post­
intervention
for
composite
samples
taken
from
bare
floors,
window
sills
and
window
troughs
(
see
Tables
36
and
37,
and
Figure
10).
Results
indicate
that
at
one­
year
post­
intervention,
composite
dust
lead
loadings
on
bare
floors,
sills
and
troughs
in
both
treatment
groups
generally
remained
well
below
pre­
intervention
levels
in
both
treatment
groups.
Based
on
Wilcoxon
rank
sum
testing,
the
association
between
the
median
percent
reduction
in
median
composite
dust
lead
loadings
and
the
assigned
treatment
group
was
significant
for
window
sills
(
p<
0.0001)
and
troughs
(
p=
0.0001)
but
not
for
bare
floors
(
p=
0.2270).

Table
36:
Percent
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Composite
Dust
Samples
LHR
Units
Pre
to
1­
Yr
Post­
Intervention
Percent
Change
LHR+
W
Units
Pre
to
1­
Yr
Post­
Intervention
Percent
Change
Bare
floors
(
34
LHR,
32
LHR+
W)
a:
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
53
82
95
76
94
98
Interior
window
sills
(
35
LHR,
37
LHR+
Wa
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
­
45
(
increase)
82
96
94
97
99
Window
troughs
(
34
LHR,
37
LHR+
W)
a
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
4
89
96
93
98
>
99
Source:
Form
04A
aBased
on
Wilcoxon
rank
sum
tests,
although
changes
in
median
bare
floor
dust
lead
loadings
were
not
significantly
associated
with
assigned
treatment
group
(
p=
0.2246),
changes
in
median
sill
and
trough
loadings
were
(
p<
0.0001
and
p=
0.0001,
respectively).

Based
on
Wilcoxon
rank
sum
testing,
the
association
between
the
absolute
reduction
in
median
composite
dust
lead
loadings
and
the
assigned
treatment
group
was
significant
for
window
sills
(
p=
0.0185)
and
troughs
(
p=
0.0047)
but
not
for
bare
floors
(
p=
0.1600)
(
see
Table
37).
11/
23/
05
68
Table
37:
Micrograms
Per
Square
Foot
Reduction
in
Pre­
to
One­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Composite
Dust
Samples
LHR
Units
Pre
to
1­
Yr
Post­
Intervention
µ
g/
ft2
Reduction
LHR+
W
Units
Pre
to
1­
Yr
Post­
Intervention
µ
g/
ft2
Reduction
Bare
floors
(
34
LHR,
32
LHR+
W)
a:
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
35
87
486
71
335
804
Interior
window
sills
(
35
LHR,
37
LHR+
W)
a:
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
­
25
(
increase)
713
6,458
693
2,803
9,771
Window
troughs
(
34
LHR,
37
LHR+
W)
a:
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
2
4,507
23,385
4,356
12,014
128,550
Source:
Form
04A
aBased
on
Wilcoxon
rank
sum
tests,
although
changes
in
median
bare
floor
dust
lead
loadings
were
not
significantly
associated
with
assigned
treatment
group
(
p=
0.1600),
changes
in
sill
and
trough
loadings
were
(
p=
0.0185
and
p=
0.0047,
respectively).

3.
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
Despite
the
significant
median
decreases
in
composite
dust
lead
loadings
between
pre­
and
oneyear
post­
intervention,
data
in
Table
38
show
that
some
units
experienced
increases
in
dust
lead
loadings
between
these
two
phases.
Indeed,
the
findings
show
that
approximately
a
quarter
of
LHR
units
had
an
increase
in
composite
dust
lead
loadings
on
sills
and
troughs
during
this
time
period.
The
percentages
of
LHR+
W
units
that
underwent
an
increase
were
lower
than
those
of
LHR
units.

Table
38:
Number
and
Percent
of
Units
Having
an
Increase
in
Composite
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
Surface
Type
and
Sample
Type
LHR
LHR+
W
Bare
Floors:
3/
34
(
9%)
1/
32
(
3%)

Interior
window
sills:
10/
35
(
28%)
1/
37
(
3%)

Window
Troughs:
8/
34
(
24%)
0/
37
(
0%)
11/
23/
05
69
4.
Summary
of
One­
Year
Post­
Intervention
Single
Surface
Results
As
discussed
in
Section
II.
E,
single
surface
samples
were
collected
in
a
subset
of
LHR
units.
However,
due
to
budget
constraints,
LAAP
ceased
collecting
such
samples
in
LHR
units
during
one­
year
and
two­
years
post­
intervention
but
continued
to
collect
them
in
LHR+
W
units.
Samples
were
collected
in
the
same
locations
of
LHR+
W
units
that
were
sampled
during
the
first
two
study
phases
(
see
Table
39
on
page
70
and
Figure
11
on
page
71).
There
were
too
few
carpeted
floor
samples
to
allow
full
characterization
of
this
surface
type.
The
one­
year
postintervention
single
surface
dust
sample
results
for
LHR+
W
units
were
compared
with
one­
year
post­
intervention
single
surface
dust
data
obtained
from
Baltimore
City
housing
units
enrolled
in
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
program.
These
units
underwent
lead
reduction
control
interventions
that
were
generally
similar
to
those
received
by
the
LHR+
W
units
during
this
study.

Median
single
surface
one­
year
post­
intervention
dust
lead
loading
results
for
bare
floors
and
window
troughs
in
LHR+
W
units
were
less
than
those
in
Baltimore
HUD
Evaluation
units,
while
the
median
single
surface
dust
lead
loading
for
sills
in
LHR+
W
units
was
slightly
higher
than
that
of
Baltimore
HUD
units.
Based
on
Wilcoxon
rank
sum
tests,
these
differences
were
significant
for
bare
floors
(
p<
0.0001)
and
troughs
(
p=
0.0017)
but
not
for
sills
(
p=
0.545).
These
results
are
somewhat
surprising
given
that
the
HUD
Round
I
units
generally
underwent
similar
lead
hazard
control
treatments
as
the
LHR+
W
units
and
given
that
the
LHR+
W
units
generally
had
higher
pre­
intervention
dust
lead
loadings
than
the
Baltimore
Round
I
units.

5.
Change
in
Single
Surface
Results
from
Pre­
to
One­
Year
Post­
Intervention
Similar
to
the
composite
dust
lead
loading
results,
substantial
median
percent
decreases
and
median
µ
g/
ft2
decreases
in
single
surface
dust
lead
loadings
were
observed
for
floors,
sills
and
window
troughs
between
pre­
and
one­
year
post­
intervention
(
see
Tables
40
and
41
on
pages
72
and
73).
Based
on
Wilcoxon
rank
sum
tests,
median
percent
reductions
for
bare
floors
and
troughs
in
LHR+
W
units
were
significantly
different
from
those
for
Baltimore
Round
I
HUD
Evaluation
units
(
p<
0.0001
for
both
surface
types)
but
were
not
significantly
different
for
window
sills
(
p=
0.813).
Again,
these
results
are
surprising
given
that
LHR+
W
units
appeared
to
have
larger
pre­
intervention
dust
lead
loadings
and
were
treated
in
a
similar
manner
to
HUD
Round
I
units.
11/
23/
05
70
Table
39:
One­
Year
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Loadingsa
Surface
Type
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Baltimore
HUD
Unitsb
(
µ
g/
ft2)
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Geometric
Mean
95%
Confid.
Limits
Bare
floorsc
(
36
LHR+
W
units)
(
249
Balt
Rd
I
units)
4
11
20
37
141
20
(
15;
28)
5
21
44
90
319
42
(
36;
48)
Interior
window
sillsc
(
36
LHR+
W
units)
(
237
Balt
Rd
I)
10
23
64
152
3,628
69
(
41;
118)
4
18
58
139
505
50
(
42;
61)
Window
troughsc
(
36
LHR+
W
units)
(
242
Balt
Rd
I)
18
78
211
391
1,058
185
(
122;
282)
51
143
421
917
2,233
376
(
321;
441)
Source:
Form
04B
aOnly
two
LHR
units
had
Phase
III
single
surface
samples
collected;
therefore,
no
comparison
with
pre­
intervention
single
surface
results
is
possible.
bBaltimore
HUD
unit
data
are
from
Baltimore
Round
One
housing
units
enrolled
in
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program.
Units
were
vacant
at
pre­
intervention.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
cBased
on
Wilcoxon
rank
sum
tests,
LHR+
W
median
dust
lead
loadings
for
floors
and
troughs
were
significantly
different
from
those
of
Baltimore
Round
I
units
(
p<
0.0001
for
floors
and
0.0017
for
troughs)
but
not
for
sills
(
p=
0.545).
11/
23/
05
71
Figure
11:
Pre­
and
One­
Year
Post­
Intervention
Single
Surface
Sample
Dust
Lead
Loading
Box
Plot
1
10
100
1000
10000
100000
1000000
Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

Pre­
Int.

1­
Yr
Post­
Int.

LHR+
W
Balt.
HUD
LHR+
W
Balt.
HUD
LHR+
W
Balt.
HUD
Bare
Floors
Window
Sills
Window
Troughs
Dust
Lead
Loading
(
ug/
ft^
2)

Note:
Bottom
whisker=
5th
percentile;
top
whisker=
95th
percentile;
box=
25th
&
75th
percentile;
circle=
median;
triangle
=
geometric
mean.
11/
23/
05
72
Table
40:
Percent
Reduction
in
Pre
to
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadingsa
Surface
Type
and
Statistic­
Single
Surface
Dust
Samples
LHR+
W
Units
Pre
to
1­
Yr
Post­
Intervention
Percent
Reduction
Baltimore
HUD
Unitsa
Pre
to
1­
Yr
Post­
Intervention
Percent
Reduction
Bare
floors
(
36
LHR+
W;
249
Balt
HUD)
b:
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
88%
96%
99%
63%
84%
94%

Interior
window
sills
(
36
LHR+
W;
237
Balt
HUD)
b:
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
75%
97%
>
99%
93%
98%
99%

Window
troughs
(
36
LHR+
W;
242
Balt
HUD)
b:
25th
Percentile
Percent
Reduction
Median
Percent
Reduction
75th
Percentile
Percent
Reduction
93%
99%
>
99%
60%
90%
98%

Source:
Form
04B
aOnly
two
LHR
units
had
Phase
III
single
surface
samples
collected;
therefore,
no
comparison
with
pre­
intervention
single
surface
results
is
possible.
bBaltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead­
Based
Paint
Grant
Program
and
were
compiled
using
pre­
intervention
and
one­
year
post­
intervention
data
and
were
from
units
that
were
vacant
at
pre­
intervention.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.
cBased
on
Wilcoxon
rank
sum
tests,
LHR+
W
units
had
significantly
larger
median
percent
decreases
in
single
surface
dust
lead
loadings
for
bare
floors
(
p<
0.0001)
and
troughs
(
p<
0.0001)
than
did
Baltimore
Round
I
units;
however,
median
percent
decreases
in
sill
loadings
were
not
significantly
different
(
p=
0.813).
11/
23/
05
73
Table
41:
Micrograms
Per
Square
Foot
Reductions
in
Pre­
to
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadingsa
Surface
Type
and
Statistic­
Single
Surface
Dust
Samples
LHR+
W
Units
Pre
to
1­
Yr
Post­
Intervention
µ
g/
ft2
Reduction
Baltimore
HUD
Units
Pre
to
1­
Yr
Post­
Intervention
µ
g/
ft2
Reductionb
Bare
floors
(
36
LHR+
W;
249
Balt
HUD):
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
148
µ
g/
ft2
420
µ
g/
ft2
2,396
µ
g/
ft2
57
µ
g/
ft2
220
µ
g/
ft2
557
µ
g/
ft2
Interior
window
sills
(
36
LHR+
W;
237
Balt
HUD):
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
540
µ
g/
ft2
1,875
µ
g/
ft2
10,322
µ
g/
ft2
860
µ
g/
ft2
2,617
µ
g/
ft2
7,114
µ
g/
ft2
Window
troughs
(
36
LHR+
W;
242
Balt
HUD)):
25th
Percentile
µ
g/
ft2
Reduction
Median
µ
g/
ft2
Reduction
75th
Percentile
µ
g/
ft2
Reduction
1,758
µ
g/
ft2
16,338
µ
g/
ft2
135,512
µ
g/
ft2
800
µ
g/
ft2
3,308
µ
g/
ft2
13,283
µ
g/
ft2
Source:
Form
04B
aOnly
two
LHR
units
had
Phase
III
single
surface
samples
collected;
therefore,
no
comparison
with
pre­
intervention
single
surface
results
is
possible.
bBaltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program
and
were
compiled
using
initial
and
Phase
4
data
(
one­
year
post­
intervention)
for
units
that
were
vacant
at
pre­
intervention.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.

6.
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
Despite
the
substantial
median
decreases
in
dust
lead
loadings
between
pre
and
one­
year
postintervention
data
in
Table
42
show
that
a
few
units
experienced
increases
in
single
surface
dust
lead
loadings
between
these
two
phases.
The
percentages
of
LHR+
W
units
that
underwent
an
increase
on
floors
and
troughs
were
lower
than
those
of
Baltimore
HUD
units,
while
the
sill
percentage
was
slightly
higher
than
that
of
Baltimore
HUD
units.

Table
42:
Number
and
Percent
of
Units
Having
an
Increase
in
Single
Surface
Dust
Lead
Loadings
from
Pre­
to
One­
Year
Post­
Intervention
Surface
Type
and
Sample
Type
LHR+
W
Baltimore
HUD
Units
Bare
Floors:
1/
36
(
3%)
25/
249
(
10%)

Interior
window
sills:
3/
36
(
8%)
2/
237
(
1%)

Window
Troughs:
2/
36
(
6%)
28/
242
(
12%)
11/
23/
05
74
7.
Comparison
of
One­
Year
Post­
Intervention
Results
with
Clearance
Standards/
Guidance
Under
the
risk
reduction
treatment
option
of
EA
6­
8,
enrolled
LHR
and
LHR+
W
units
were
required
only
to
pass
a
visual
inspection
performed
by
an
independent
visual
inspector
immediately
after
intervention
was
complete.
Collection
of
dust
samples
and
comparison
of
such
results
to
clearance
standards
were
not
required
during
any
phase.
However,
this
study
considered
clearance
dust
standards
set
by
Maryland
and
by
HUD/
EPA
to
assess
the
extent
to
which
prescribed
treatments
continued
to
produce
dust
results
that
were
sufficient
for
safe
occupancy
of
treated
units
one
year
after
treatments
were
complete.
This
comparison
to
clearance
standards
is
not
required
by
EA
6­
8
and
was
performed
in
this
study
for
research
purposes
only.
Again,
only
LHR+
W
units
are
presented
here
because
one­
year
post­
intervention
single
surface
data
were
not
available
for
the
vast
majority
of
LHR
units.

One
year
after
interventions
were
completed,
less
than
20%
of
LHR+
W
units
had
single
surface
dust
lead
loadings
that
exceeded
clearance
standards
for
any
surface
type
(
see
Table
43).
At
oneyear
post­
intervention,
Baltimore
HUD
units
had
a
higher
percentage
of
bare
floors
and
window
troughs
that
exceeded
HUD
clearance
standards
of
100
and
500
µ
g/
ft2,
respectively.
Based
on
Chi­
square
testing,
there
is
a
significant
difference
between
the
clearance
exceedance
rate
for
LHR+
W
units
versus
Baltimore
HUD
units
(
p<
0.0001
for
bare
floors,
p=
0.016
for
troughs).
LHR+
W
units
had
a
higher
percentage
of
units
with
window
sills
above
500
µ
g/
ft2;
however,
there
was
only
a
marginally
significant
difference
based
on
a
Fisher's
Exact
Test
(
p=
0.096).
This
is
surprising
given
the
fact
that
LHR+
W
units
generally
had
larger
pre­
intervention
dust
lead
loadings
and
treatments
that
were
similar
to
those
of
Baltimore
HUD
round
I
units.
Visual
assessment
"
failures"
reported
at
immediate
post­
intervention
that
were
subsequently
repaired
may
have
contributed
to
the
relatively
lower
rates
for
LHR+
W
units.

Table
43:
Percentage
of
LHR+
W
Units
with
One­
Year
Post­
Intervention
Single
Surface
Dust
Lead
Loadings
Exceeding
Clearance
Standardsa
Percent
of
Units
within
Specified
Category
Surface
Type:
Bare
Floors
Interior
Window
Sills
Window
Troughs
Std/
Guidance:

100
µ
g/
ft2

200
µ
g/
ft2

500
µ
g/
ft2

800
µ
g/
ft2
LHR+
W
Units
(
36
units)
5
(
14%)
4
(
11%)
6
(
17%)
5
(
14%)

Baltimore
HUD
Unitsb
(
249
units)
143
(
57%)
NA
25
(
11%)
85
(
35%)

aOnly
two
LHR
units
had
Phase
III
single
surface
samples
collected;
therefore,
no
comparison
with
clearance
standards/
guidance
is
possible.
bBaltimore
HUD
unit
data
are
Round
I
data
from
the
National
Evaluation
of
HUD's
Lead
Hazard
Control
Grant
Program
and
were
compiled
using
initial
and
Phase
4
data
(
one­
year
post­
intervention)
for
units
that
were
vacant
at
pre­
intervention.
For
floors,
Baltimore
HUD
unit
data
include
both
interior
floor
and
entry
floor
data.

C.
Results
of
One­
Year
Post­
Intervention
Visual
Assessment
At
one­
year
post­
intervention,
LAAP
inspectors
performed
another
visual
assessment
to
evaluate
the
continued
effectiveness
of
treatments
one
year
after
intervention.
A
complete
listing
of
oneyear
post­
intervention
visual
assessment
"
failures"
is
provided
in
Appendix
G.
Table
44
presents
11/
23/
05
75
the
distribution
of
one­
year
post­
intervention
"
failures"
per
unit.
There
were
no
units
having
more
than
five
"
failures."
LAAP
inspectors
found
that
all
36
LHR
units
and
34
of
the
37
(
92%)
LHR+
W
units
had
at
least
one
visual
assessment
"
failure"
one
year
after
interventions
had
been
completed.
This
equates
to
an
overall
one­
year
post­
intervention
visual
assessment
"
failure"
rate
of
96%
for
the
73
one­
year
post­
intervention
units.
Landlords
were
informed
of
these
"
failures"
and
were
provided
with
recommendations
to
fix
the
problems.

Overall,
the
likelihood
of
LHR
units
having
one
or
more
one­
year
post­
intervention
"
failures"
was
almost
the
same
as
that
of
the
LHR+
W
units.
However,
t­
tests
indicate
that
the
mean
number
of
"
failures"
per
unit
for
LHR
units
(
2.4)
was
marginally
significantly
different
from
and
higher
than
that
for
LHR+
W
units
(
1.9)
(
p=
0.0482).
If
only
"
failure"
types
applicable
to
both
treatment
groups
are
considered,
LHR
units
have
a
slightly
lower
mean
number
of
one­
year
postintervention
"
failures"
per
unit
(
2.2),
while
the
mean
for
LHR+
W
units
remains
almost
the
same
(
1.8);
these
two
means
are
not
significantly
different
from
each
other
based
on
t­
tests
(
p=
0.1246).

Table
44:
Number
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unit
Number
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
Per
Unita
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

"
0"
1­
yr
post­
int
visual
assessment
"
failures"
0
3
(
8%)
3
(
4%)
"
1"
1­
yr
post­
int
visual
assessment
"
failure"
8
(
22%)
13
(
35%)
21
(
29%)
"
2"
1­
yr
post­
int
visual
assessment
"
failures"
12
(
33%)
9
(
24%)
21
(
29%)
"
3"
1­
yr
post­
int
visual
assessment
"
failures"
11
(
31%)
9
(
24%)
20
(
27%)
"
4"
1­
yr
post­
int
visual
assessment
"
failures"
3
(
8%)
3
(
8%)
6
(
8%)
"
5"
1­
yr
post­
int
visual
assessment
"
failures"
2
(
6%)
0
2
(
3%)
Source:
Form
05
aThe
mean
number
of
"
failures"
per
unit
for
LHR
units
(
2.4)
was
marginally
significantly
different
from
that
for
LHR+
W
units
(
1.9)
(
p=
0.0482).

1.
Types
of
One­
Year
Post­
Intervention
"
Failures"

The
most
common
visual
assessment
"
failure"
at
one­
year
post­
intervention
was
that
not
all
paint
was
intact
after
completion
of
the
lead
hazard
reduction
treatments
(
i.
e.,
chipping,
flaking
and/
or
peeling
paint
remained)
(
see
Table
45
on
page
76),
with
over
90%
of
all
units
(
94%
of
LHR
units
and
92%
of
LHR+
W
units)
exhibiting
this
type
of
"
failure."
Over
40%
of
units
(
50%
of
LHR
units
and
32%
of
LHR+
W
units)
had
doors
with
painted
surfaces
that
continued
to
rub
together
or
bind.
Approximately
one­
third
of
all
units
(
33%
of
LHR
units
and
24%
of
LHR+
W
units)
had
visible
paint
chips
and
debris
in
various
locations,
and
over
25%
(
31%
of
LHR
units
and
24%
of
LHR+
W
units)
had
some
kitchen
and
bathroom
floors
whose
water­
resistant
coverings
were
not
intact.
Other
types
of
one­
year
post­
intervention
visual
assessment
"
failures"
were
much
less
prevalent.
Although
LHR
units
had
a
higher
percentage
of
units
with
each
of
the
abovementioned
types
of
"
failures,"
none
of
the
"
failure"
types
was
significantly
associated
with
assigned
treatment
group
based
on
Chi­
square
tests.
11/
23/
05
76
Table
45:
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
Type
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
a
(
36
LHR
units,
37
LHR+
W
units)
LHR
#
Units
(%
Units)
LHR+
W
#
Units
(%
Units)
All
Units
#
Units
(%
Units)

Not
all
paint
intact
(
chipping,
peeling
or
flaking
paint
remains)
34
(
94%)
34
(
92%)
68
(
93%)

Visible
structural
defect
that
could
cause
paint
deterioration
remains
1
(
3%)
2
(
5%)
3
(
4%)

If
sills
stripped/
repainted
or
replaced,
not
all
treatments
in
place
1
(
3%)
0
1
(
1%)

If
sills
encapsulated
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attachedb
0
0
0
If
troughs
capped
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attachedb
5
(
14%)
NAb
5
(
7)
c
For
windows,
not
all
top
sashes
are
fixed
in
place
0
NAb
0
For
doors,
painted
surfaces
rub
together
18
(
50%)
12
(
32%)
30
(
41%)

Some
bare
floors
are
not
smooth
and
cleanable
5
(
14%)
4
(
11%)
9
(
12%)

Not
all
kitchen
and
bathroom
floors
are
overlaid
with
water­
resistant
coverings
11
(
31%)
9
(
24%)
20
(
27%)

Visible
paint
chips
or
debris
remains
12
(
33%)
9
(
24%)
21
(
29%)

Source:
Form
05
aBased
on
Chi­
square
tests,
none
of
the
listed
types
of
visual
assessment
"
failures"
were
significantly
associated
with
assigned
treatment
group
(
p
varied
from
0.127
to
0.689).
bIn
LHR+
W
units,
these
treatments
were
not
necessary
by
virtue
of
the
installation
of
new
vinyl
replacement
windows
with
unpainted
top
sashes
and
troughs.
cValue
is
understated
due
to
the
inapplicability
of
this
treatment
for
LHR+
W
units.

2.
Magnitude
of
One­
Year
Post­
Intervention
"
Failures"

The
magnitude
of
the
above
"
failures"
was
evaluated
by
reviewing
notes
that
LAAP
inspectors
recorded
about
the
observed
"
failures"
by
room
location
and
"
failure"
type,
the
specific
component
or
surface
type
that
"
failed"
in
a
given
room,
and
their
estimated
measurement
of
each
"
failure"
in
units
of
linear
inches,
square
inches,
linear
feet,
square
feet
or
"
each."

The
geometric
means
of
"
failures"
presented
in
Table
46
on
page
77
were
calculated
by
first
summing,
within
a
unit,
all
reports
of
a
particular
"
failure"
type
having
the
same
measurement
units
(
e.
g.,
square
feet),
then
calculating
the
geometric
mean
across
dwelling
units.
These
data
indicate
that
problems
recorded
by
LAAP
inspectors
one
year
after
treatment
were
easily
observed
(
see
Table
46).
For
example,
LAAP
inspectors
reported
that
a
geometric
mean
of
1
ft2
of
paint
in
both
LHR
and
LHR+
W
units
was
not
intact
at
one­
year
post­
intervention.
This
type
of
"
failure"
was
reported
61
times
in
23
LHR
units
and
61
times
in
25
LHR+
W
units,
indicating
a
widespread
occurrence.
Painted
door
surfaces
rubbing
together
was
also
noted
often,
reported
22
times
in
18
LHR
units
and
16
times
in
12
LHR+
W
units.
The
status
of
floors
at
one­
year
11/
23/
05
77
post­
intervention
is
also
noteworthy,
with
a
geometric
mean
of
4
ft2
of
flooring
in
both
LHR
and
LHR+
W
units
being
not
smooth
and
cleanable
one
year
after
treatments.
A
geometric
mean
of
2
ft2
of
kitchen
and
bathroom
floors
in
LHR
and
LHR+
W
units
had
tile
or
coverings
that
were
damaged,
cracked,
or
missing.
These
results
are
somewhat
surprising
considering
that
these
types
of
problems
were
reportedly
corrected
during
the
lead
hazard
reduction
intervention
or
were
corrected
when
LAAP
inspectors
discovered
such
problems
during
the
visual
inspection
conducted
immediately
after
treatments
were
completed.

Table
46:
Reporting
Frequency
of
Specified
Types
of
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
LHR
LHR+
W
"
Failure"
Type
(
36
LHR,
37
LHR+
W)
Unit
of
Measure
#
Units
with
Specified
"
Failure"
Rpt'd
Reporting
Freq.
of
"
Failure"
Type
at
Ph
2
"
Failure"
Magnit.
(
geom.
Mean
&
95%
Conf
Limits)
#
Units
with
Specified
"
Failure"
Rpt'd
Reporting
Freq.
of
"
Failure"
Type
at
Ph
2
"
Failure"
Magnit.
(
geom.
mean
&
95%
Conf
Limits)
EAa
3
3
1
(
1,
3)
3
3
2
(
NA)
c
LF
32
94
2
(
1,
3)
26
82
2
(
1,
3)
Not
all
paint
intact
(
34
LHR,
34
LHR+
W)
SF
23
61
1
(
0.2,
2)
25
61
1
(
0.2,
2)

For
doors,
painted
surfaces
continued
to
rub
together
(
18
LHR,
12
LHR+
W)
EAa
18
22
1
(
1,
1)
12
16
1
(
1,
2)

Some
bare
floors
not
smooth
&
cleanable
(
5
LHR,
4
LHR+
W)
SF
5
10
4
(
1,
21)
4
6
4
(
NA)
c
Not
all
kitchen
&
bathroom
floors
overlaid
with
water­
resistant
coverings
(
11
LHR,
9
LHR+
W)
LF
SF
0
11
13
0
2
(
2,
3)
2
8
3
9
2
(
NA)
2
(
1,
5)

EAb
4
19
6
(
4,
10)
3
10
5
(
2,
13)
LF
1
2
8
(
NA)
0
0
NA
Visible
paint
chips
or
debris
remains
(
12
LHR,
9
LHR+
W)
SF
7
23
1
(
0.1,
9)
6
10
1
(
0.2,
4)

EA=
each;
LF=
linear
foot;
SF=
square
foot
aFor
these
"
failure"
types,
"
each"
denotes
window
components
(
e.
g.,
upper
sash,
lower
sash),
door
casings,
or
handrails
in
need
of
re­
painting.
LF/
SF
values
are
not
available
for
these
occurrences.
bFor
this
"
failure"
type,
"
each"
usually
denotes
a
window
with
debris
or
chips
in
the
well.
cNo
95%
confidence
limits
were
calculated
due
to
the
small
number
of
units
having
this
failure
reported.
11/
23/
05
78
3.
Rooms
with
One­
Year
Post­
Intervention
"
Failures"

In
general,
three
to
six
rooms
per
unit
(
mean
5
rooms
for
LHR
units
and
4
rooms
for
LHR+
W
units)
had
visual
assessment
"
failures"
at
one­
year
post­
intervention
(
see
Figure
12
on
page
74).
All
LHR
units
had
at
least
one
room
with
"
failures,"
while
8%
of
LHR+
W
units
had
no
rooms
with
"
failures."
Although
no
LHR+
W
units
had
more
than
9
rooms
with
"
failures,"
2%
of
LHR
units
had
10
or
more
rooms
with
"
failures."

Figure
12:
Percentage
of
Units
Having
Specified
Number
of
Rooms
Per
Unit
with
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"

0%
3%
8%
14%
22%

17%
17%

6%

3%
6%

3%
3%
8%

5%
5%
19%
19%
19%

11%

8%

0%
5%

0%
0%
4%
4%
7%
16%
20%

18%

14%

7%

1%
5%

1%
1%

0%
5%
10%
15%
20%
25%

0
1
2
3
4
5
6
7
8
9
10
11
Percentage
of
Units
Having
Specified
Number
of
Room
with
Visual
Assessment
"
Failures"
at
One­
Year
Post­
Intervention
Percentage
of
Units
LHR
Units
LHR+
W
Units
All
Units
In
both
LHR
and
LHR+
W
units,
paint
was
not
intact
in
an
average
of
3
rooms
per
unit
(
see
Table
47
on
page
79).
For
other
types
of
"
failures"
(
e.
g.,
painted
doors
rubbing,
visible
paint
chips
and
debris
present),
a
given
"
failure"
was
generally
reported
for
1
room
per
unit.
11/
23/
05
79
Table
47:
Mean
Number
of
Rooms
Per
Unit
with
Specified
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"
by
Treatment
Groupa
Type
of
Post­
Intervention
Visual
Assessment
"
Failures"
(
36
LHR
units,
37
LHR+
W
units)
LHR
Mean
#
Rooms
Per
Unit
(
95%
Conf.
Int.)
LHR+
W
Mean
#
Rooms
per
Unit
(
95%
Conf.
Int.)
Not
all
paint
intact
(
chipping,
peeling
or
flaking
paint
remains)
3.4
(
2.8,
4.0)
3.2
(
2.6,
3.8)
Visible
structural
defect
that
could
cause
paint
deterioration
remains
0
(
0,
0.1)
0.1
(
0,
0.2)
If
sills
stripped/
repainted
or
replaced,
not
all
treatments
in
place
0
(
0,
0.1)
0
If
sills
encapsulated
with
vinyl,
metal,
etc.,
not
all
material
is
property
attached
0
0
If
troughs
capped
with
vinyl,
metal,
etc.,
not
all
material
is
properly
attached
0.2
(
0,
0.3)
NAa
For
windows,
not
all
top
sashes
are
fixed
in
place
0
NAa
For
doors,
painted
surfaces
continue
to
rub
together
0.6
(
0.4,
0.9)
0.4
(
0.2,
0.6)
Some
bare
floors
are
not
smooth
and
cleanable
0.3
(
0.1,
0.5)
0.2
(
0,
0.3)
Not
all
kitchen
and
bathroom
floors
are
overlaid
with
water­
resistant
coverings
0.4
(
0.2,
0.6)
0.3
(
0.1,
0.5)
Visible
paint
chips
or
debris
remain
1.2
(
0.8,
1.6)
0.5
(
0.3,
0.8)
Source:
Form
05
aIn
LHR+
W
units,
these
treatments
were
not
necessary
by
virtue
of
the
immediate
post­
intervention
installation
of
new
vinyl
replacement
windows
with
unpainted
top
sashes
and
troughs.

4.
Type
of
Contractor
and
Number
of
One­
Year
Post­
Intervention
"
Failures"

At
immediate
post­
intervention,
lead
hazard
reduction
treatments
were
generally
performed
by
for­
profit
contractors
or
the
property
owner/
employees.
Only
four
of
the
73
one­
year
postintervention
units
had
been
treated
by
non­
profit
contractors;
therefore,
these
data
are
excluded
from
this
section's
discussion.
An
average
of
5.3
"
failures"
per
unit
were
observed
at
one­
year
post­
intervention
in
LHR
units
treated
by
for­
profit
contractors,
and
6.7
in
LHR
units
treated
by
the
property
owner/
employees
(
see
Table
48
on
page
80).
The
mean
number
of
one­
year
postintervention
"
failures"
per
unit
in
LHR+
W
units
treated
by
for­
profit
contractors
and
property
owner/
employees
was
5.1
and
3.9,
respectively.
A
Poisson
regression
model
was
run
to
predict
the
number
of
"
failures"
per
unit
at
one­
year
post­
intervention
based
on
assigned
treatment
group,
contractor
type
(
for­
profit
or
property
owner/
employees),
and
their
interaction.
None
of
the
three
variables
were
significant
predictors
of
the
number
of
visual
assessment
"
failures"
per
unit.
11/
23/
05
80
Table
48:
Mean
Number
of
One­
Year
Post­
Treatment
"
Failures"
per
Unit
by
the
Type
of
Contractora,
b
Mean
Number
of
"
Failures"
per
Unit
Contractor
Type
LHR
Units
LHR+
W
Units
For­
profit
contractor
5.3
5.1
Property
owner/
employees
6.7
3.9
Source:
Form
05
a
Because
only
four
of
the
73
had
nonprofit
contractors
performing
the
work,
the
number
of
"
failures"
is
not
easily
compared
against
for­
profit
contractors
and
property
owner/
employees
work.
Therefore,
nonprofit
contractor
data
are
not
presented
in
this
table.
bBased
on
a
Poisson
regression
model,
neither
the
assigned
treatment
group
nor
the
type
of
contractor
nor
the
interaction
of
these
two
variables
were
significant
predictors
of
the
number
of
"
failures"
per
unit
at
Phase
III.

D.
Follow­
Up
Data
Summary
for
One­
Year
Post­
Intervention
(
Form
06)

Form
06
(
Follow­
Up
Questions)
was
designed
to
collect
information
from
the
property
owner
or
his/
her
designee
about
any
activities
that
occurred
in
the
unit
in
the
12­
month
period
after
intervention.
Only
four
units
reportedly
had
a
change
in
tenancy.
And,
in
general,
few
units
appear
to
have
had
further
lead
hazard
reduction
treatments
done
during
the
one­
year
postintervention
period.
During
the
12
months
prior
to
LAAP's
one­
year
post­
intervention
visit,
lead
hazard
reduction
work
was
reportedly
done
in
six
units,
four
of
which
had
work
done
at
turnover.
Costs
for
this
lead
hazard
reduction
work
ranged
from
$
100
to
$
2,000
(
mean
of
approximately
$
600),
with
work
done
one
to
eight
months
before
the
one­
year
post­
intervention
visit.
In
two
units,
tenants
reported
to
the
owner
that
new
lead
hazards
existed;
these
were
not
the
same
two
units
that
had
lead
hazard
reduction
work
done
but
no
work
was
completed
at
turnover.
There
was
one
unit
for
which
the
owner
was
notified
that
a
tenant
child
had
a
blood
lead
level
between
15
and
24
µ
g/
dl.
A
qualified
offer
was
made
for
this
case.
No
owner
was
notified
that
a
tenant
child
had
a
blood
lead
level
above
25
µ
g/
dl.

E.
Statistical
Modeling
Results
for
One­
Year
Post­
Intervention
1.
Predicting
Dust
Lead
Loading
at
One­
Year
Post­
Intervention
A
multiple
regression
model
with
backward
elimination
was
used
with
multiple
regression
to
identify
factors
that
are
significant
predictors
of
one­
year
post­
intervention
composite
dust
lead
loadings.
Composite
dust
lead
loadings
were
used
instead
of
single
surface
results
because
composite
data
were
available
for
all
73
one­
year
post­
intervention
units.
A
separate
model
was
run
for
each
surface
type
(
i.
e.,
bare
floors,
interior
window
sills
and
window
troughs).
Dust
lead
loading
measurements
were
log
transformed.
(
Statistical
tables
resulting
from
these
models
are
provided
in
Appendix
I.)

The
set
of
possible
predictors
considered
were:

 
Type
of
building
(
e.
g.,
single
family
or
multifamily
dwelling);
 
Year
of
construction
(
e.
g.,
pre­
1920
or
1920­
1949);
 
Contractor
type
(
i.
e.,
for­
profit,
nonprofit,
property
owner/
employee);
11/
23/
05
81
 
Concurrent
work
(
i.
e.,
yes
or
no);
 
Treatment
group
(
i.
e.,
LHR
or
LHR+
W);
 
Pre­
intervention
composite
dust
lead
loading
results
for
a
given
surface
type
(
i.
e.,
bare
floors,
interior
window
sills
or
window
troughs);
 
Number
of
items
of
baseline
interior
dwelling
unit
deterioration;
 
Number
of
items
of
baseline
exterior
building
deterioration;
 
Percent
of
windows
replaced
out
of
the
total
number
of
windows
in
a
dwelling
unit;
 
Estimated
market
value
of
a
dwelling
unit;
 
Total
lead
hazard
control
cost
for
a
dwelling
unit,
including
window
replacement
costs,
if
applicable;
 
Was
work
performed
during
12­
month
post­
intervention
period
(
yes/
no);
 
Number
of
rooms
with
any
visual
assessment
failures
at
one­
year
post­
intervention;
and
 
Interactions
of
(
1)
the
number
of
interior
dwelling
unit
deterioration
items,
(
2)
the
number
of
exterior
building
deterioration
items,
and
(
3)
pre­
intervention
dust
lead
loadings
with
each
of
the
following:
concurrent
work,
percent
of
windows
replaced,
total
lead
hazard
reduction
cost,
and
assigned
treatment
group.

The
only
significant
predictor
of
one­
year
post­
intervention
bare
floor
dust
lead
loadings
was
the
percent
of
rooms
with
visual
assessment
"
failures"
at
one­
year
post­
intervention
(
i.
e.,
units
with
more
one­
year
post­
intervention
visual
assessment
failures
were
more
likely
to
have
higher
oneyear
post­
intervention
composite
floor
dust
lead
loadings).
The
percentage
of
variation
(
i.
e.,
R2)
accounted
for
in
one­
year
post­
intervention
composite
dust
lead
loadings
by
this
variable
is
13%.
For
interior
window
sills,
the
variables
found
to
predict
one­
year
post­
intervention
dust
lead
loadings
were
pre­
intervention
sill
dust
lead
loadings
and
the
assigned
treatment
group
(
R2=
24%)
(
i.
e.,
units
with
higher
pre­
intervention
sill
dust
lead
loadings
were
more
likely
to
have
higher
one­
year
post­
intervention
dust
lead
loadings,
and
LHR
units
were
more
likely
to
have
higher
sill
dust
lead
loadings
than
LHR+
W
units).
The
only
variable
that
was
predictive
of
one­
year
postintervention
composite
dust
lead
loadings
for
window
troughs
was
the
assigned
treatment
group
(
R2=
10%)
(
i.
e.,
LHR
units
were
more
likely
to
have
higher
one­
year
post­
intervention
composite
trough
dust
lead
loadings
than
LHR+
W
units).

2.
Predicting
One­
Year
Post­
Intervention
Dust
Lead
Loading
Exceedances
of
Standards
Separate
logistic
regression
models
with
backward
elimination
were
employed
to
predict
whether
dwelling
units
would
have
one­
year
post­
intervention
dust
lead
loadings
on
bare
floors,
interior
window
sills
and
window
troughs
that
exceeded
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2,
and
800
µ
g/
ft2,
respectively
(
see
Appendix
I).
The
same
set
of
possible
predictors
that
were
used
in
predicting
dust
lead
loading
at
one­
year
post­
intervention
were
also
employed
for
this
model,
except
that
interactions
were
not
included
because
they
are
not
appropriate
for
logistic
models.

For
bare
floors,
the
only
significant
predictor
of
one­
year
post­
intervention
composite
results
being
above
a
standard
of
100
µ
g/
ft2
was
the
year
of
construction,
with
older
units
more
likely
to
yield
one­
year
post­
intervention
floor
results
above
100
µ
g/
ft2.
For
window
sills,
variables
found
to
significantly
predict
one­
year
post­
intervention
composite
dust
lead
loadings
above
a
standard
of
500
µ
g/
ft2
were
pre­
intervention
window
sill
dust
lead
loadings
and
assigned
treatment
group,
11/
23/
05
82
with
units
having
higher
pre­
intervention
composite
sill
results
more
likely
to
have
one­
year
post­
intervention
sill
results
above
500
µ
g/
ft2,
and
LHR
units
more
likely
to
have
one­
year
postintervention
sill
results
above
500
µ
g/
ft2
than
LHR+
W
units.
For
window
troughs,
the
only
predictor
of
one­
year
post­
intervention
composite
dust
lead
loadings
for
window
troughs
was
the
assigned
treatment
group
(
R2=
10%),
with
LHR+
W
units
less
likely
to
have
one­
year
postintervention
composite
trough
results
above
a
standard
of
800
µ
g/
ft2.

3.
Predicting
One­
Year
Post­
Intervention
Visual
Assessment
"
Failures"

A
Poisson
regression
model
with
backward
elimination10
was
employed
to
predict
the
number
of
rooms
having
a
visual
assessment
"
failure"
reported
at
one­
year
post­
intervention.
Again,
the
same
set
of
possible
predictors
employed
for
predicting
one­
year
post­
intervention
composite
dust
lead
loading
and
one­
year
post­
intervention
composite
dust
lead
loading
exceedances
of
standards
were
used
for
this
model.
Significant
predictors
were
the
total
lead
hazard
reduction
cost
at
immediate
post­
intervention
(
higher
costs
would
yield
fewer
rooms
with
"
failures"),
building
type
(
single
and
rowhouses
would
have
more
rooms
with
"
failures"
than
multifamily
buildings),
whether
work
had
been
performed
in
the
unit
during
the
previous
12
months
(
yes
would
yield
fewer
rooms
with
"
failures"),
and
the
estimated
market
value
of
the
dwelling
(
higher
market
values
would
yield
fewer
rooms
with
"
failures").

X.
SUMMARY
OF
TWO­
YEAR
POST­
INTERVENTION
RESULTS
As
discussed
in
the
introduction
to
Section
VI,
only
24
units
(
11
LHR,
13
LHR+
W)
had
complete
two­
years
post­
intervention
data.
Given
this
small
number
of
units,
rigorous
statistical
testing
is
not
considered
appropriate.
However,
this
section
provides
a
general
summary
of
the
two­
years
post­
intervention
findings.

A.
Two­
Year
Post­
Intervention
Dust
Sampling
Results
No
single
surface
samples
were
collected
from
any
of
the
24
units
during
two­
years
postintervention
therefore,
only
composite
results
are
discussed.
A
complete
listing
of
two­
years
post­
intervention
composite
results
is
provided
in
Appendix
E.

1.
Summary
of
Two­
Year
Post­
Intervention
Composite
Results
For
bare
floors,
median
composite
dust
lead
loadings
for
LHR
units
were
slightly
lower
than
those
for
LHR+
W
units.
However,
for
sills
and
troughs,
LHR+
W
units
had
much
lower
median
dust
lead
loadings
than
LHR
units
(
See
Table
49
on
page
83).
Based
on
Kruskal­
Wallis
testing,
two­
years
post­
intervention
median
composite
dust
lead
loadings
for
sills
and
troughs
were
significantly
associated
with
assigned
treatment
group
(
p=
0.0064
and
0.0300,
respectively)
but
not
for
bare
floors
(
p=
0.7098).

10
In
a
Poisson
regression
model
with
backward
elimination,
the
outcome
of
interest
is
a
count
variable
(
e.
g.,
number
of
failures).
11/
23/
05
83
Table
49:
Two­
Year
Post­
Intervention
Composite
Sample
Dust
Lead
Loadings
Surface
Type
LHR
Dust
Lead
Loading
(
µ
g/
ft2)
Median
Geometric
Mean
95%
Confid.
Limits
LHR+
W
Dust
Lead
Loading
(
µ
g/
ft2)
Median
Geometric
Mean
95%
Confid.
Limits
Bare
floorsa,
b
(
10
LHR
units,
13
LHR+
W
units)
19
27
(
9,
85)
24
18
(
10,
30)

Interior
window
sillsa,
b
(
10
LHR
units,
13
LHR+
W
units)
568
485
(
199;
1,186)
46
99
(
45;
221)

Window
troughsa,
b
(
10
LHR
units,
13
LHR+
W
units)
1,710
3,119
(
779;
12,492)
452
558
(
235;
1,329)

Source:
Form
04A,
Phase
IV
aApproximately
96%
of
bare
floor
composites,
interior
window
sill
composites,
and
window
trough
composites
were
comprised
of
4
sub­
samples;
the
rest
were
comprised
of
2
sub­
samples.
b
Kruskal­
Wallis
test
indicates
that
median
sill
and
trough
values
were
significantly
associated
with
assigned
treatment
group
(
p=
0.0064
and
0.0300,
respectively)
but
not
the
median
bare
floor
value
(
p=
0.7098).

2.
Change
in
Composite
Results
from
Pre­
to
Two­
Years
Post­
Intervention
Large
median
percent
decreases
were
observed
when
measuring
the
change
in
dust
lead
loadings
from
pre
to
two­
years
post­
intervention
for
composite
samples
taken
from
bare
floors,
sills
and
troughs
(
Table
50
on
page
84).
The
association
between
the
percent
reduction
in
preintervention
to
two­
years
post­
intervention
dust
lead
loadings
and
assigned
treatment
group
was
significant
for
window
sills
(
p=
0.0008)
and
troughs
(
p=
0.0019)
but
not
for
bare
floors
(
p=
0.2917).
Although
pre­
intervention
window
component
dust
lead
loadings
for
the
LHR+
W
group
were
significantly
higher
than
those
for
LHR
units,
at
two
years
post­
intervention,
the
LHR+
W
group
had
lower
window
component
dust
lead
loadings.
Although
two­
year
postintervention
window
component
dust
lead
loadings
remained
lower
than
pre­
intervention
loadings
for
LHR
units,
the
wide
confidence
intervals
shown
in
Table
49
indicate
that
some
dwellings
have
rather
high
loadings
two
years
after
treatment.
11/
23/
05
84
Table
50:
Median
Percent
Reductions
and
Absolute
Reductions
in
Pre­
to
Two­
Year
Post­
Intervention
Composite
Dust
Lead
Loadings
Surface
Type
and
Statistic­
Composite
Dust
Samples
LHR
Units
Pre
to
2­
Yr
Post­
Intervention
Reduction
LHR+
W
Units
Pre
to
2­
Yr
Post­
Intervention
Reduction
Median
Percent
Reduction
(
10
LHR,
13
LHR+
W)
a:
Bare
Floors
Window
Sills
Window
Troughs
78%
79%
39%
88%
97%
96%

Median
Absolute
Reduction
(
10
LHR,
13
LHR+
Wa
Bare
Floors
Window
Sills
Window
Troughs
45
µ
g/
ft2
1,960
µ
g/
ft2
1,154
µ
g/
ft2
170
µ
g/
ft2
2,563
µ
g/
ft2
10,104
µ
g/
ft2
Source:
Form
04A
a
The
association
between
the
percent
reduction
in
pre­
intervention
to
Phase
IV
dust
lead
loadings
and
assigned
treatment
group
was
significant
for
window
sills
(
p=
0.0008)
and
troughs
(
p=
0.0019)
but
not
for
bare
floors
(
p=
0.2917).

3.
Change
in
Composite
Results
from
One­
Year
to
Two­
Years
Post­
Intervention
Of
the
24
units
having
two­
years
post­
intervention
data,
only
15
(
7
LHR
and
8
LHR+
W)
units
had
complete
one­
year
post­
intervention
data.
These
numbers
were
too
small
to
permit
a
reliable
comparison
of
one­
year
post­
intervention
and
two­
years
post­
intervention
data
to
be
conducted.
Therefore,
these
results
are
excluded
from
this
report.

B.
Results
of
Two­
Year
Post­
Intervention
Visual
Assessment
Too
few
visual
assessment
"
failures"
were
observed
in
each
treatment
group
to
allow
comparison
of
two­
years
post­
intervention
visual
assessment
"
failures"
in
LHR
versus
LHR+
W
units.
Therefore,
this
discussion
focuses
on
the
24
two­
years
post­
intervention
units
as
a
whole.
A
complete
listing
of
two­
years
post­
intervention
visual
assessment
"
failures"
is
provided
in
Appendix
G.
Overall,
90
percent
(
22)
of
the
24
units
having
two­
years
post­
intervention
data
had
at
least
one
visible
assessment
"
failure"
two
years
after
treatment
(
Table
51
on
page
85).
LAAP
informed
landlords
of
these
"
failures"
and
recommended
that
they
be
repaired.

The
most
common
visual
assessment
"
failure"
at
two­
years
post­
intervention
was
that
not
all
paint
was
intact,
with
22
of
the
24
units
exhibiting
this
type
of
"
failure".
The
magnitude
of
this
"
failure"
was
similar
to
that
reported
at
one­
year
post­
intervention:
a
geometric
mean
of
2
ft2
of
non­
intact
paint
per
unit
was
noted.
Other
"
failure"
types
were
less
prevalent.
For
example,
10
of
the
24
units
reportedly
had
at
least
one
door
with
painted
surfaces
rubbing
together,
while
9
units
had
some
non­
intact
or
damaged
kitchen
and/
or
bathroom
floor
coverings,
with
a
geometric
mean
of
2
ft2
of
such
damage
reported.
Three
to
seven
rooms
per
unit
had
some
type
of
visual
assessment
"
failure"
at
two­
years
post­
intervention
(
mean=
5
rooms).
11/
23/
05
85
Table
51:
Reporting
Frequency
of
Specified
Types
of
Two­
Year
Post­
Intervention
Visual
Assessment
"
Failures"

"
Failure"
Type
(
24
units)
Unit
of
Measure
#
Units
with
Specified
"
Failure"
Rpt'd
Reporting
Freq.
of
"
Failure"
Type
at
Ph
IV
"
Failure"
Magnit.
(
Geom.
Mean
&
95%
Conf.
Limits)
EAa
2
2
2
(
NA)
c
LF
20
58
2
(
1,
3)
Not
all
paint
intact
(
22
units)

SF
15
47
2
(
0.4,
5)
For
doors,
painted
surfaces
continued
to
rub
together
(
10
units)
EAa
10
12
1
(
1,
1)

Some
bare
floors
not
smooth
&
cleanable
(
4
units)
LF
SF
1
3
1
3
4
(
NA)
c
2
(
0.4,
14)
Not
all
kitchen
&
bathroom
floors
overlaid
with
water­
resistant
coverings
(
9
units)
LF
SF
2
7
3
9
4
(
NA)
c
2
(
1,
4)

EAb
1
4
7
(
NA)
c
Visible
paint
chips
or
debris
remains
(
6
units)
SF
5
10
1
(
0,
82)
EA=
each;
LF=
linear
foot;
SF=
square
foot
aFor
these
"
failure"
types,
"
each"
denotes
window
components
(
e.
g.,
upper
sash,
lower
sash),
door
casings,
or
handrails
in
need
of
re­
painting.
LF/
SF
values
are
not
available
for
these
occurrences.
bFor
this
"
failure"
type,
"
each"
usually
denotes
a
window
with
debris
or
chips
in
the
well.
c
No
95%
confidence
limits
were
calculated
due
to
the
small
number
of
units
having
this
failure
reported.

XI.
SUMMARY
AND
DISCUSSION
OF
FINDINGS
The
main
findings
of
this
study
are
that
dust
lead
loadings
declined
substantially
immediately
after
EA
6­
8
prescribed
lead
hazard
reduction
treatments
were
implemented
and
remained
well
below
pre­
intervention
levels
one
year
later.
However,
many
study
units
would
not
have
"
passed"
clearance
dust
samples
immediately
post­
intervention
had
such
testing
been
required.
The
results
also
suggest
that
the
independent
visual
inspections
conducted
in
these
units
missed
many
lead
hazard
reduction
treatment
"
failures."
Although
these
"
failures"
were
corrected
upon
discovery,
several
visual
assessment
"
failures"
were
observed
one
year
after
treatment.
Details
concerning
these
findings
are
summarized
below.

A.
Summary
and
Discussion
of
Findings
Through
Immediate
Post­
Intervention
1.
Comparison
of
Pre­
and
Immediate
Post­
Intervention
Dust
Lead
Loading
Results
Both
the
LHR
and
LHR+
W
units
were
generally
in
poor
condition
and
exhibited
high
dust
lead
loadings
prior
to
the
lead
hazard
reduction
treatments.
LHR+
W
units
had
a
greater
number
of
items
of
interior
deterioration
and
their
pre­
intervention
composite
and
single
surface
dust
lead
loadings
were
generally
higher
than
those
for
the
LHR
units,
indicating
that
LHR+
W
units
may
have
been
in
poorer
condition
than
LHR
units
at
baseline.
However,
after
completion
of
the
EA
6­
8
prescribed
treatments,
both
the
LHR
and
LHR+
W
units
experienced
substantial
reductions
in
dust
lead
loading
on
floors,
interior
window
sills
and
window
troughs,
as
measured
by
either
composite
or
single
surface
sampling
methods.
11/
23/
05
86
Pre
and
immediate
post­
intervention
dust
lead
loading
results
for
composite
and
single
surface
sampling
methods
are
summarized
in
Table
52.

Table
52:
Summary
of
Pre
and
Immediate
Post­
Intervention
Dust
Lead
Loading
Results
By
Assigned
Treatment
Group
and
Sampling
Method
LHR
LHR+
W
Surface
Type
and
Statistic
Composite
Single
Surface
Composite
Single
Surface
Bare
Floors:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
µ
g/
ft2
Decrease
Median
Immed
Post­
Int
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
Percent
Decrease
273
147
27
85
192
99
66
70
333
225
29
89
338
226
39
88
Interior
window
sills:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
µ
g/
ft2
Decrease
Median
Immed
Post­
Int
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
Percent
Decrease
1,986
1,686
83
94
844
497
103
84
2,780
2,596
36
99
2,350
2,516
29
99
Window
Troughs:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
µ
g/
ft2
Decrease
Median
Immed
Post­
Int
Loading
(
µ
g/
ft2)
Median
Pre­
to
Immed
Post­
Int
Percent
Decrease
9,427
9,708
344
95
6,135
4,111
610
94
10,998
11,284
100
>
99
12,227
12,022
115
99
Note:
Because
a
smaller
number
of
LHR
units
were
single
surface
sampled
both
pre
and
immediate
postintervention
these
results
are
possibly
more
uncertain
even
though
they
are
consistent
with
single
surface
sample
results
observed
for
the
LHR+
W
units.

While
none
of
the
study
units
were
legally
required
(
under
EA
6­
8)
to
undergo
and
pass
clearance
dust
testing,
all
immediate
post­
intervention
median
composite
and
dwelling
unit
average
of
single
surface
sample
dust
lead
loading
values
for
bare
floors,
interior
window
sills
and
window
troughs
were
below
the
Maryland
and
HUD/
EPA
guidance
abatement
"
clearance"
standards
for
such
surfaces.
However,
when
applying
these
clearance
standards
to
individual
single
surface
dust
sample
results,
an
appreciable
number
of
LHR
and
LHR+
W
units
had
at
least
one
immediate
post­
intervention
floor,
interior
window
sill
and/
or
window
trough
dust
lead
loading
value
that
exceeded
the
Maryland
and
HUD/
EPA
guidance
standards
(
see
Table
53
on
page
87).
Forty­
six
(
46)
and
28
percent
of
units
exceeded
floor
clearance
standards
of
100
µ
g/
ft2
and
200
µ
g/
ft2,
respectively.
Eight
(
8)
percent
of
units
exceeded
the
interior
window
sill
clearance
standard
500
µ
g/
ft2.
Twenty­
seven
(
27)
percent
of
units
exceeded
the
window
trough
clearance
standard
of
800
µ
g/
ft2.
Assigned
treatment
group
was
significantly
associated
with
immediate
post­
intervention
interior
window
sill
and
window
trough
results
"
passing"
or
"
failing"
their
standards
(
LHR+
W
units
generally
did
better)
but
was
not
significantly
associated
with
bare
floors
"
passing"
or
"
failing"
their
respective
standards.
It
should
be
noted
that,
had
clearance
testing
been
required,
single
surface
dust
wipe
results
would
yield
more
clearance
"
failures"
than
would
composite
dust
wipe
results
(
see
Appendix
J
for
more
details).
11/
23/
05
87
Table
53:
Percentage
of
Units
with
Immediate
Post­
Intervention
Single
Surface
Clearance
"
Failures"
Percent
of
Units
within
Specified
Category
Surface
Type:
Floors
Interior
Window
Sills
Window
Troughs
Std/
Guidance:

100
µ
g/
ft2

200
µ
g/
ft2

500
µ
g/
ft2

800
µ
g/
ft2
LHR
Units
60%
35%
20%
50%
LHR+
W
Units
41%
25%
5%
19%
Total
All
Units
46%
28%
8%
27%

2.
Immediate
Post­
Intervention
Visual
Assessment
Results
Although
statistically
significant
dust
lead
loading
reductions
from
pre
to
immediate
postintervention
were
observed,
immediate
post­
intervention
confirmatory
visual
assessments
performed
by
the
LAAP
inspectors
suggest
that
a
high
percentage
of
study
units
(
93
percent
of
LHR
and
LHR+
W
units
combined)
had
"
failures"
of
one
or
more
of
the
prescribed
lead
hazard
reduction
treatments
and,
as
such,
would
have
"
failed"
to
meet
the
EA
6­
8
risk
reduction
standard.
This
is
noteworthy
since
each
one
of
the
121
study
units
passed
the
independent
visual
inspection
conducted
shortly
before
the
LAAP
confirmatory
visual
assessments.

In
addition,
the
reported
"
failures"
were
large
enough
to
be
easily
observed.
For
the
most
commonly
reported
"
failure" 
not
all
paint
intact 
a
dwelling
unit
geometric
mean
of
1
ft2
of
non­
intact
paint
was
reported
in
LHR
and
in
LHR+
W
units.
Visible
paint
chips
and
debris
remaining
at
clearance
was
also
quite
frequently
reported,
with
a
dwelling
unit
geometric
mean
of
8
ft2
noted
in
LHR
units
and
16
ft2
in
LHR+
W
units.
Door
surfaces
continuing
to
rub
together
were
also
frequently
reported
at
an
average
of
1
door
per
unit.

Of
these
121
units,
27
percent
would
have
"
failed"
one
(
1)
of
the
prescribed
lead
hazard
reduction
treatment
types,
another
36
percent
would
have
"
failed"
two
(
2)
of
the
prescribed
treatment
types
and
an
additional
23
percent
would
have
"
failed"
three
(
3)
of
the
prescribed
treatment
types.
For
both
LHR
and
LHR+
W
units,
the
most
prevalent
"
failure"
types
were:
not
all
paint
intact,
with
some
chipping,
flaking
and
peeling
paint
remaining
(
75
percent
of
all
units);
one
or
more
painted
doors
continuing
to
rub
together
and/
or
bind
(
43
percent);
and
visible
paint
chips
and/
or
debris
remaining
(
38
percent).
Additionally,
in
26
percent
of
the
LHR
units,
not
all
window
trough
cap
material
was
properly
attached.

In
those
units
that
"
failed"
the
LAAP
inspectors'
visual
assessment
at
immediate
postintervention
considering
only
those
rooms
with
"
failures"
applicable
to
both
LHR
and
LHR+
W
units,
an
average
of
3.1
rooms
in
the
LHR
units
and
2.6
rooms
in
the
LHR+
W
units
had
"
failures."

These
results
suggest
that,
at
least
for
this
study
group,
lead
hazard
reduction
treatments
did
not
completely
meet
the
prescribed
risk
reduction
standard
and
more
strongly
suggest
that
the
independent
visual
inspections
conducted
in
these
units
failed
to
identify
all
lead
hazard
reduction
treatment
"
failures,"
even
ones
that
should
have
been
easily
observed
(
e.
g.,
71
ft2
of
visible
chips/
debris).
Because
only
five
or
six
independent
inspectors
conducted
visual
inspections
in
the
group
of
121
study
units,
this
finding
does
not
necessarily
reflect
on
all
other
11/
23/
05
88
Maryland­
certified
inspectors
but
does
indicate
that
these
five
or
six
inspectors
missed
important
items.

3.
Cost
Data
Analysis
Results
Two
factors
potentially
impacted
the
lead
hazard
reduction
treatment
costs
recorded
for
the
121
study
units:

 
"
Adverse
selection"
of
enrolled
units.
The
majority
of
the
study
units
were
in
poor
condition
to
begin
with
(
e.
g.,
extensive
deferred
maintenance
resulted
in
a
high
degree
of
interior
and
exterior
deterioration),
which
likely
resulted
in
higher
costs.
Because
property
owners
could
receive
a
grant
for
up
to
80
percent
of
the
total
lead
hazard
control
costs,
they
may
have
deliberately
enrolled
units
that
needed
more
extensive
work,
either
due
to
deferred
maintenance
or
the
length
of
time
since
last
rental
turnover.

 
Effect
of
participating
in
the
study.
Since
property
owners
and
contractors
were
aware
that
the
enrolled
units
were
being
evaluated
for
the
study,
it
is
possible
that
extra
effort
may
have
gone
into
the
lead
hazard
reduction
treatments,
thereby
potentially
inflating
reported
treatment
costs.
(
Interestingly,
even
if
this
was
the
case,
it
was
not
borne
out
by
the
LAAP
inspectors'
visual
assessment
results.)

Window
replacement
also
added
appreciably
to
the
total
lead
hazard
reduction
treatment
costs
recorded
for
the
LHR+
W
units.
While
median
costs
for
only
the
prescribed
lead
hazard
reduction
treatments
were
$
2,154
and
$
1,649
for
LHR
and
LHR+
W
units,
respectively,
the
median
lead
treatment
plus
window
replacement
cost
for
LHR+
W
units
was
$
4,348;
this
represents
an
approximate
two­
fold
increase
in
the
total
LHR+
W
lead
costs.

Speculating
that
the
recorded
costs
were
higher
than
expected
for
a
typical
Baltimore
City
rental
housing
unit,
Center
and
LAAP
representatives
met
with
a
local
rental
property
owner
to
discuss
his
firm's
experience
in
complying
with
the
EA
6­
8
risk
reduction
standard.
On
average,
this
owner's
total
expense
of
performing
the
prescribed
treatments
at
first
unit
turnover
costs
from
$
700
to
$
900
per
unit.
Subsequent
turnover
prescribed
treatment
costs
typically
run
from
$
300
to
$
400
per
unit.
The
initial
expenses
that
are
added
to
this
owner's
normal
unit
turnover
regimen
include
HEPA
vacuuming
the
entire
unit
($
80),
window
treatments
($
35
to
$
40
per
window
unit),
re­
hanging
and
re­
working
doors
($
20
to
$
25
per
door),
and
the
cost
of
a
visual
inspector
($
35
to
$
40
per
inspection).
According
to
this
owner,
normal
unit
turnover
expenses,
exclusive
of
the
required
lead­
related
costs,
can
run
as
high
as
$
5,000
to
$
6,000
per
unit,
depending
on
the
length
of
time
the
unit
has
been
occupied
by
one
tenant
and/
or
the
amount
of
tenant­
related
damage.
(
The
owner
counts
interior
repainting
as
a
normal
turnover
expense.)

Considering
the
factors
that
potentially
inflated
recorded
costs
and
reflecting
on
the
typical
experience
of
a
Baltimore
City
rental
property
owner
in
complying
with
the
statute,
the
lead
hazard
reduction
treatment
costs
recorded
by
this
study
should
not
be
considered
average
and/
or
representative
of
all
Baltimore
City
rental
housing
units.
11/
23/
05
89
B.
Summary
and
Discussion
of
Findings
Through
Two­
Years
Post­
Intervention
1.
Comparison
of
Pre­
Intervention
and
One­
Year
Post­
Intervention
Dust
Lead
Loading
Results
Prior
to
prescribed
lead
hazard
reduction
treatments,
both
the
LHR
and
the
LHR+
W
units
in
the
73­
unit
dataset
(
followed
through
one
year)
were
generally
in
poor
condition
and
exhibited
high
dust
lead
loadings.
During
the
one­
year
post­
intervention
period,
only
six
(
6)
of
the
73
units
reportedly
had
further
lead
hazard
reduction
work
done.
At
one­
year
post­
intervention,
composite
dust
lead
loadings
on
bare
floors,
window
sills
and
window
troughs
remained
far
below
pre­
intervention
levels
in
both
treatment
groups.
Median
percent
reductions
in
composite
dust
lead
loadings
from
pre­
intervention
to
one­
year
post­
intervention
on
sills
and
troughs
in
LHR+
W
units
were
significantly
different
from
and
lower
than
those
in
LHR
units.
The
median
percent
reduction
in
composite
dust
lead
loadings
for
bare
floors
in
LHR+
W
units
was
about
the
same
as
in
LHR
units.

At
one­
year
post­
intervention,
single
surface
dust
lead
loadings
on
floors,
sills
and
troughs
remained
far
below
pre­
intervention
levels
for
LHR+
W
units,
the
only
treatment
group
having
single
surface
data
at
one­
year
post­
intervention.

Pre
and
one­
year
post­
intervention
dust
lead
loading
results
for
composite
and
single
surface
sampling
methods
are
summarized
in
Table
54.

Table
54:
Summary
of
Pre­
and
One­
Year
Post­
Intervention
Dust
Lead
Loading
Results
By
Assigned
Treatment
Group
and
Sampling
Method
LHR
LHR+
W
Surface
Type
and
Statistic
Composite
Single
Surface
Composite
Single
Surface
Bare
Floors:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
µ
g/
ft2
Reduction
Median
1­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Percent
Reduction
120
87
22
82%
NA
NA
NA
NA
454
335
20
94%
440
420
20
96%
Interior
window
sills:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
µ
g/
ft2
Reduction
Median
1­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Percent
Reduction
1,739
713
449
82%
NA
NA
NA
NA
2,930
2,803
121
97%
1,987
1,875
64
97%
Window
Troughs:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
µ
g/
ft2
Reduction
Median
1­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Percent
Reduction
10,219
4,507
1,074
89%
NA
NA
NA
NA
12,524
12,014
247
98%
17,083
16,338
211
99%
NA=
Not
applicable.
Single
surface
samples
were
collected
in
only
two
LHR
units
at
Phase
III;
therefore,
no
single
surface
results
are
presented
for
LHR
units.
11/
23/
05
90
For
bare
floor
and
sills,
one­
year
post­
intervention
median
composite
dust
lead
loadings
were
less
than
HUD
clearance
standards
of
100
and
500
µ
g/
ft2,
respectively.
Median
composite
dust
lead
loadings
for
troughs
in
LHR+
W
units
were
also
less
than
the
HUD
standard
of
800
µ
g/
ft2,
but
the
median
composite
trough
result
for
LHR
units
exceeded
this
standard.
At
one­
year
postintervention
14%
of
LHR+
W
units
had
at
least
one
single
surface
floor
dust
lead
result
exceeding
a
clearance
standard
of
100
µ
g/
ft2,
17%
had
at
least
one
single
surface
sill
result
greater
than
500
µ
g/
ft2
and
14%
had
at
least
one
single
surface
trough
result
above
800
µ
g/
ft2
(
see
Table
55).

Comparison
of
LHR+
W
clearance
exceedance
rates
with
those
of
Baltimore
HUD
units
studied
during
the
National
Evaluation
yielded
some
surprising
results.
Although
LHR+
W
units
appeared
to
be
in
worse
condition
at
baseline
and
Baltimore
HUD
units
had
similar
treatments
and
clearance
testing
immediately
after
intervention,
LHR+
W
units
appeared
to
be
in
better
condition
one
year
later
based
on
single
surface
dust
wipe
results.
This
finding
is
possibly
due
to
the
correction
of
treatment
"
failures"
in
LHR+
W
units
after
the
immediate
post­
intervention
visit.
Median
percent
reductions
from
pre
to
one­
year
post­
intervention
were
generally
greater
for
LHR+
W
units
than
for
Baltimore
HUD
units,
possibly
due
to
the
worse
initial
condition
of
floors
and
troughs.
LAAP
reported
that
the
correction
of
Phase
II
treatment
"
failures"
sometimes
entailed
re­
cleaning
if
the
correction
potentially
generated
dust
(
e.
g.,
if
painted
doors
had
to
be
planed).

Table
55:
Percentage
of
Units
with
One­
Year
Post­
Intervention
Single
Surface
Clearance
"
Failures"
Percent
of
Units
within
Specified
Category
Surface
Type:
Floors
Interior
Window
Sills
Window
Troughs
Std/
Guidance:

100
µ
g/
ft2

200
µ
g/
ft2

500
µ
g/
ft2

800
µ
g/
ft2
LHR+
W
Units
(
36
units)
14%
11%
17%
14%

Balt.
HUD
Units
(
249
units)
57%
NA
11%
35%

2.
One­
Year
Post­
Intervention
Visual
Assessment
Results
Although
immediate
post­
intervention
treatment
"
failures"
were
corrected
and
one­
year
postintervention
median
composite
dust
lead
loadings
in
a
majority
of
units
were
less
than
clearance
standards,
a
high
percentage
of
units
(
96%)
had
at
least
one
visual
assessment
"
failure"
reported
one
year
after
treatments
were
completed.
These
"
failures"
were
widespread,
with
non­
intact
paint
reported
in
68
units,
painted
door
surfaces
rubbing
together
in
30
units,
visible
paint
chips
and
debris
in
21
units,
deterioration
of
kitchen
and
bathroom
floor
water­
resistant
coverings
in
20
units,
and
bare
floors
not
smooth
and
cleanable
in
nine
(
9)
units.
Such
"
failures"
tended
to
be
of
small
magnitude
but
were
large
enough
to
be
easily
observed,
with
geometric
means
of
1
ft2
of
non­
intact
paint
reported
per
unit,
one
rubbing
painted
door
problem
per
unit,
4
ft2
of
bare
floors
that
were
not
smooth
and
cleanable,
2
ft2
of
deteriorated
kitchen/
bathroom
floor
water­
resistant
coverings,
and
1
ft2
of
visible
paint
chips
and
debris
present
one
year
after
interventions.
As
previously
stated,
the
areal
density
(
i.
e.,
depth)
and
lead
concentration
of
chips/
debris
were
not
measured.
These
visual
assessment
results
suggest
that,
at
least
for
this
study
group,
lead
hazard
11/
23/
05
91
reduction
treatments
did
not
remain
completely
intact
one
year
after
lead
hazard
reduction
treatments.

3.
Comparison
of
Pre­
Intervention
and
Two­
Years
Post­
Intervention
Dust
Lead
Loading
Results
Two
years
after
lead
hazard
reduction
treatments
were
first
performed,
composite
dust
lead
loadings
on
all
surfaces
remained
well
below
pre­
intervention
levels
(
Table
56).
Of
the
24
units
having
two­
years
post­
intervention
data,
only
15
had
complete
one­
year
post­
intervention
data;
therefore,
no
comparison
of
one­
year
post­
intervention
and
two­
years
post­
intervention
data
could
be
made.

Table
56:
Summary
of
Pre­
and
Two­
Years
Post­
Intervention
Compositea
Dust
Lead
Loading
Results
By
Assigned
Treatment
Group
and
Sampling
Method
Surface
Type
and
Statistic
LHR
LHR+
W
Bare
Floors:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
µ
g/
ft2
Reduction
Median
2­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
Percent
Reduction
139
45
19
78%
177
170
24
88%
Interior
window
sills:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
µ
g/
ft2
Reduction
Median
2­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
Percent
Reduction
2,201
1,960
568
79%
2,930
2,563
46
97%
Window
Troughs:
Median
Pre­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
µ
g/
ft2
Reduction
Median
2­
Yr
Post­
Intervention
Loading
(
µ
g/
ft2)
Median
Pre­
to
2­
Yr
Post­
Int
Percent
Reduction
3,408
1,154
1,710
39%
10,900
10,104
452
96%
aSingle
surface
samples
were
not
collected
in
any
units
at
Phase
IV;
therefore,
no
single
surface
results
are
presented.

C.
Comparing
the
Efficacy
of
the
Prescribed
Lead
Hazard
Reduction
Treatments
Both
With
and
Without
Window
Replacement
When
measured
by
the
immediate
post­
intervention
visual
assessments
and
composite
and
single
surface
dust
lead
loading
results,
pre­
intervention
lead
hazards
in
the
LHR+
W
units
were
more
effectively
reduced
than
in
the
LHR
units.
Although
LHR+
W
units
had
more
baseline
building
and
dwelling
unit
deterioration
and
higher
dust
lead
loadings
prior
to
the
intervention,
these
trends
were
reversed
slightly
immediately
after
treatment.
Replacement
of
windows
may
account
for
these
observed
immediate
post­
intervention
results,
particularly
with
respect
to
interior
window
sill
and
window
trough
dust
lead
loadings.
This
is
supported
by
the
results
of
predictive
modeling,
which
showed
that
pre­
intervention
dust
lead
loadings
on
sills
and
troughs,
as
well
as
treatment
group
assignment,
significantly
predicted
immediate
post­
intervention
dust
lead
loadings
on
sills
and
troughs.
11/
23/
05
92
At
immediate
post­
intervention,
the
LHR+
W
units
had
fewer
lead
hazard
reduction
treatment
visual
assessment
"
failure"
types
per
unit
and
fewer
rooms
per
unit
with
"
failures."
Also,
the
mean
number
of
"
failure"
types
per
LHR
unit,
which
was
greater
than
the
LHR+
W
units,
was
found
to
be
statistically
significant
for
treatment
group
assignment
(
p­
value
=
0.0255).

For
both
composite
and
single
surface
sampling
methods,
immediate
post­
intervention
median
bare
floor
dust
lead
loading
results
were
similar
for
LHR
and
LHR+
W
units.
However,
for
interior
window
sills
and
window
troughs,
both
the
composite
and
single
surface
immediate
postintervention
sample
results
were
significantly
associated
with
the
assigned
treatment
group,
and
were
substantially
lower
in
the
LHR+
W
units.
This
finding
indicates
that
window
replacement
is
much
more
effective
in
reducing
lead
dust
hazards
on
window
surfaces
within
the
unit
immediately
after
treatment.

Although
LHR+
W
units
had
more
baseline
building
and
dwelling
unit
deterioration
and
higher
dust
lead
loadings
than
LHR
units
prior
to
intervention,
at
one­
year
post­
intervention,
LHR+
W
units
tended
to
have
lower
dust
lead
loadings
and
greater
percent
reductions
in
dust
lead
loadings
on
window
sills
and
troughs
between
pre­
and
one­
year
post­
intervention
than
did
LHR
units.
Little
difference
in
one­
year
post­
intervention
bare
floor
dust
lead
loadings,
however,
was
discernible
between
the
two
treatment
groups.
Replacement
of
windows
likely
accounts
for
these
observed
one­
year
post­
intervention
results,
as
is
supported
by
the
results
of
predictive
modeling,
which
showed
that
pre­
intervention
sill
dust
lead
loadings,
as
well
as
treatment
group
assignment,
significantly
predicted
one­
year
post­
intervention
dust
lead
loadings
on
sills
and
troughs.
Assigned
treatment
group
did
not
significantly
predict
one­
year
post­
intervention
bare
floor
dust
lead
loadings;
the
only
significant
predictor
of
one­
year
bare
floor
dust
lead
loadings
was
the
percent
of
rooms
with
visual
assessment
"
failures."

Although
differences
in
dust
lead
loadings
at
one­
year
post­
intervention
were
observed,
visual
assessment
"
failures"
for
LHR
units
were
generally
similar
to
those
observed
for
LHR+
W
units.
While
there
were
marginally
significant
differences
in
the
mean
number
of
visual
assessment
"
failures"
per
unit
for
LHR
(
2.4)
versus
LHR+
W
units
(
1.9)
(
p=
0.0482),
there
was
no
significant
association
between
the
types
of
visual
assessment
"
failures"
and
the
assigned
treatment
group
and
the
magnitude
of
"
failures"
was
generally
very
similar
between
the
two
groups.

These
findings
indicate
that
window
replacement
is
more
effective
in
maintaining
lower
dust
lead
loadings
on
window
surfaces
within
the
unit
one
year
after
treatment.

XII.
CONCLUSIONS
AND
RECOMMENDATIONS
A.
Conclusions
The
primary
research
goals
of
this
study
were
to
evaluate
the
efficacy
of
the
EA
6­
8
lead
hazard
reduction
treatments
in
reducing
dust
lead
loading,
both
alone
and
in
combination
with
window
replacement.
Overall,
results
through
immediate
post­
intervention
showed
that
dust
lead
loadings
were
significantly
reduced
on
floors,
interior
window
sills
and
window
troughs.
Additionally,
results
through
one­
year
post­
intervention
showed
that
the
levels
remained
low
one
year
after
treatment,
with
over
80
percent
of
units
having
floor,
sill
and
trough
dust
lead
loadings
11/
23/
05
93
that
were
below
clearance
standards.
However,
immediately
after
the
treatment
intervention,
many
units
would
have
"
failed"
clearance
dust
testing
had
it
been
required
in
the
statute.
More
importantly,
a
substantial
number
of
units
"
failed"
the
LAAP
inspectors'
visual
assessments
at
immediate
post­
intervention,
despite
the
fact
that
each
of
the
121
study
units
passed
the
prescribed
independent
inspector's
visual
inspection.

By
addressing
the
five
(
5)
key
evaluation
questions
found
in
Section
II.
B,
additional
conclusions
can
be
made:

1.
How
effective
are
the
EA
6­
8­
prescribed
treatments
in
reducing
dust
lead
loading
on
floors,
interior
window
sills
and
window
troughs
to
acceptable
levels?

Although
dramatic
decreases
in
dust
lead
loadings
on
all
surface
types
were
found
immediately
post­
intervention,
remaining
dust
lead
loadings
were
high.
As
discussed
in
Section
I.
A,
although
EA
6­
8
does
not
require
clearance
dust
testing
upon
completion
of
the
prescribed
treatments,
study
results
strongly
suggest
that
visual
inspections
alone,
as
performed
by
independent
visual
inspectors
under
current
oversight
conditions,
are
not
sufficient
to
assure
that
housing
units
will
be
safe
for
re­
occupancy
by
families
with
young
children.

Applying
the
same
HUD/
EPA
guidance
clearance
standards
of
100
µ
g/
ft2
for
floors,
500
µ
g/
ft2
for
interior
window
sills
and
800
µ
g/
ft2
for
window
troughs
that
were
used
to
determine
dust
clearance
"
failures"
in
the
HUD
Lead
Hazard
Control
Grant
Program
National
Evaluation,
immediate
post­
intervention
single
surface
dust
sample
"
failure"
rates
were
high
in
the
study
units.
Forty­
six
(
46)
percent
of
the
units
had
at
least
one
sample
that
"
failed"
floor
clearance,
eight
(
8)
percent
had
at
least
one
sample
that
"
failed"
interior
window
sill
clearance
and
27
percent
had
at
least
one
sample
that
"
failed"
window
trough
clearance.
This
is
a
higher
"
failure"
rate
than
that
observed
for
vacant
Baltimore
Round
One
units
enrolled
in
the
HUD
Evaluation
(
i.
e.,
29
percent
for
floors,
one
(
1)
percent
for
interior
window
sills
and
six
(
6)
percent
for
window
troughs).

HUD
recently
promulgated
a
new
lead­
based
paint
regulation
governing
all
pre­
1978
housing
receiving
federal
assistance
as
well
as
federally
owned
housing.
The
regulation
includes
new
clearance
standards
applicable
to
such
housing.
Applying
these
clearance
standards
of
40
µ
g/
ft2
for
floors,
250
µ
g/
ft2
for
interior
window
sills
and
400
µ
g/
ft2
for
window
troughs
to
the
immediate
post­
intervention
single
surface
dust
sample
results
recorded
during
this
study,
73
percent
of
the
study
units
would
have
had
at
least
one
sample
that
"
failed"
floor
clearance,
15
percent
would
have
had
at
least
one
sample
that
"
failed"
interior
window
sill
clearance
and
38
percent
would
have
had
at
least
one
sample
that
"
failed"
window
trough
clearance.

Considering
that
the
study
units
were
not
randomly
selected
and
that
property
owners,
contractors
and
independent
visual
inspectors
were
all
presumably
aware
that
the
units
would
undergo
dust
testing
as
part
of
this
study,
these
high
clearance
"
failure"
rates
support
the
need
for
clearance
dust
testing
in
conjunction
with
the
prescribed
treatments.
11/
23/
05
94
2.
To
what
extent
does
lead
dust
re­
accumulate
on
these
surfaces
over
a
one
and
two­
year
period
after
completion
of
the
treatments?

Overall,
one­
year
post­
intervention
results
showed
that
dust
lead
loadings
on
floors,
sills
and
troughs
remained
far
below
pre­
intervention
levels.
Although
one­
year
post­
intervention
dust
lead
loadings
were
dramatically
less
than
pre­
intervention
levels
for
all
three
surface
types,
several
units
had
floor,
sill
and
trough
composite
dust
sample
results
that
exceeded
HUD
clearance
standards
of
100,
500,
and
800
µ
g/
ft2,
respectively,
and
many
had
visual
assessment
"
failures"
one
year
after
interventions
were
complete.

Replacement
of
windows
in
units
that
also
underwent
the
prescribed
treatments
resulted
in
substantially
lower
dust
lead
loadings
on
sills
and
troughs
one
year
after
completion
of
treatments,
but
no
significant
differences
in
bare
floor
dust
lead
loadings
were
observed
between
LHR+
W
and
LHR
units.
These
results
indicate
that
lead
hazard
reduction
interventions
that
include
window
replacement
are
more
effective
than
those
that
simply
treat
windows
(
Table
57).

Table
57:
Median
Percent
Reductions
in
Dust
Lead
Loadings
from
Pre
to
One­
Year
Post­
Intervention
for
LHR,
LHR+
W
and
Baltimore
Round
One
HUD
Evaluation
Units
Surface
Type
and
Sample
Type
LHR
Median
%
Decline,
Pre
to
1­
Yr
Post
LHR+
W
Median
%
Decline,
Pre
to
1­
Yr
Post
Balt.
HUDc
Median
%
Decline,
Pre
to
1­
Yr
Post
Bare
Floors:
Composite
Single
Surface
82
NAb
94
96
NAa
84
Interior
window
sills:
Composite
Single
Surface
82
NAb
97
97
NAa
98
Window
Troughs:
Composite
Single
Surface
89
NAb
98
99
NAa
90
aNA=
not
applicable.
Composite
dust
samples
were
not
taken
in
the
Baltimore
Round
One
HUD
Evaluation
units.
bNA=
not
applicable.
No
single
surface
results
are
presented
for
LHR
units
because
only
2
LHR
units
had
Phase
III
single
surface
samples
collected.
cBaltimore
HUD
unit
data
are
from
Round
I
of
the
National
Evaluation
of
HUD's
Lead­
Based
Paint
Grant
Program
and
were
compiled
using
pre­
intervention
and
one­
year
post­
intervention
data
for
units
that
were
vacant
before
treatment.
For
floors,
both
interior
and
entry
data
were
included.

At
one­
year
post­
intervention,
Baltimore
Round
One
HUD
Evaluation
units
had
a
significantly
higher
percentage
of
bare
floor
and
window
trough
results
that
exceeded
HUD
clearance
standards
of
100
and
500
µ
g/
ft2,
respectively,
indicating
that
treatments
in
LHR+
W
units
may
have
been
"
performing
better"
than
those
in
Baltimore
Round
One
units.
This
finding
is
surprising
given
the
similarity
of
treatments
performed
in
the
two
sets
of
units
(
Table
58
on
page
95).
11/
23/
05
95
Table
58:
One­
Year
Post­
Intervention
"
Failures"
of
HUD
Clearance
Standards
for
LHR+
W
and
Baltimore
Round
One
HUD
Evaluation
Units
Based
on
Single
Surface
Sampling
Surface
Type
LHR
Units
%
"
Failed"
b
LHR+
W
Units
%
"
Failed"
b
Baltimore
HUD
Units
%
"
Failed"
b
Bare
Floorsa
NA
14
57%

Interior
window
sills
NA
17
11%

Window
Troughs
NA
14
35%

aBaltimore
Round
One
HUD
evaluation
unit
floor
samples
include
all
floor
surfaces
(
i.
e.,
bare
and
carpeted);
LHR
and
LHR+
W
unit
floor
samples
are
bare
floor
surfaces
only.
b
Based
on
initial
clearance
testing
using
clearance
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2
and
800
µ
g/
ft2,
for
floors,
sill
and
troughs,
respectively.

Two
years
after
treatments
were
first
performed,
composite
dust
lead
loadings
on
all
surfaces
were
less
than
pre­
intervention
levels
(
Table
59).

Table
59:
Median
Percent
Reductions
in
Compositea
Dust
Lead
Loadings
from
Pre­
to
Two­
Years
Post­
Intervention
for
LHR
and
LHR+
W
Units
Surface
Type
and
Sample
Type
LHR
Median
%
Decline,
Pre
to
2­
Yrs
Post
LHR+
W
Median
%
Decline,
Pre
to
2­
Yrs
Post
Bare
Floors
Interior
Window
Sills
Window
Troughs
78
79
39
88
97
96
aNo
single
surface
results
are
presented
because
only
composite
samples
were
collected
at
Phase
IV.

3.
Does
the
replacement
of
windows
in
units
that
also
undergo
the
prescribed
treatments
result
in
substantially
lower
lead
dust
loadings
immediately
after
completion
of
the
treatments?

As
discussed
in
Section
VII.
E
and
as
summarized
in
Table
60
on
page
96,
although
the
percent
reduction
in
pre
to
immediate
post­
intervention
composite
dust
lead
loadings
was
similar
for
the
LHR
and
LHR+
W
treatment
groups,
the
median
percent
reductions
in
dust
lead
loadings
as
measured
by
single
surface
dust
sampling
were
significantly
associated
with
assigned
treatment
group
for
sills
and
troughs
but
not
for
floors.
These
reductions
imply
that
lead
hazard
reduction
interventions
that
include
window
replacement
are
more
effective
than
those
that
simply
treat
windows
in
lowering
window­
related
dust
lead
loadings.
This
conclusion
is
further
supported
by
a
comparison
of
study
unit
data
with
those
of
the
Baltimore
Round
One
HUD
Evaluation
units.
In
general,
Baltimore
Round
One
units
had
all
windows
replaced
and
the
magnitude
in
dust
lead
loading
reduction
for
these
units
closely
matched
that
of
the
LHR+
W
units.
11/
23/
05
96
Table
60:
Median
Percent
Reductions
in
Dust
Lead
Loadings
from
Pre­
Intervention
to
Immediate
Post­
Intervention
for
LHR,
LHR+
W
and
Baltimore
Round
One
HUD
Evaluation
Units
Surface
Type
and
Sample
Type
LHR
Median
%
Decline
LHR+
W
Median%
Decline
Balt.
HUD
Median%
Decline
Bare
Floors:
Composite
Single
Surface
85
(
49
units)
70
(
16
units)
89
(
56
units)
88
(
63
units)
NAa
95
(
278
units)

Interior
window
sills:
Composite
Single
Surface
94
(
52
units)
84
(
16
units)
99
(
62
units)
99
(
62
units)
NAa
>
99
(
277
units)

Window
Troughs:
Composite
Single
Surface
95
(
52
units)
94
(
16
units)
>
99
(
62
units)
99
(
62
units)
NAa
99
(
273
units)

aNA=
not
applicable.
Composite
dust
samples
were
not
taken
in
the
Baltimore
Round
One
HUD
Evaluation
units.

When
applying
the
HUD
Evaluation
clearance
standards
for
floors,
interior
window
sills
and
window
troughs
(
100
µ
g/
ft2,
500
µ
g/
ft2,
and
800
µ
g/
ft2,
respectively),
LHR+
W
units
had
fewer
single
surface
sample
clearance
"
failures"
than
LHR
units.
Upon
initial
clearance
testing,
Baltimore
Round
One
HUD
Evaluation
units
had
even
fewer
"
failures"
than
either
LHR+
W
or
LHR
study
units
(
see
Table
61).
(
All
Baltimore
Round
One
units
that
"
failed"
clearance
were
recleaned
and
eventually
met
clearance
standards
when
re­
tested.)
Assigned
treatment
group
was
significantly
associated
with
immediate
post­
intervention
interior
window
sill
and
window
trough
results
"
passing"
or
"
failing"
their
respective
standards,
but
was
not
significantly
associated
with
bare
floors
"
passing"
or
"
failing"
their
standard.

Table
61:
Immediate
Post­
Intervention
Clearance
"
Failure"
Rates
for
LHR,
LHR+
W,
and
Baltimore
Round
One
HUD
Evaluation
Units
Surface
Type
LHR
Units
%
"
Failed"
b
LHR+
W
Units
%
"
Failed"
b
Baltimore
HUD
Units
%
"
Failed"
b
Bare
Floorsa
60
41
29
Interior
window
sills
20
5
1
Window
Troughs
50
19
6
aBaltimore
Round
One
HUD
evaluation
unit
floor
samples
include
all
floor
surfaces
(
i.
e.,
bare
and
carpeted);
LHR
and
LHR+
W
unit
floor
samples
are
bare
floor
surfaces
only.
b
Based
on
initial
clearance
testing
using
clearance
standards
of
100
µ
g/
ft2,
500
µ
g/
ft2
and
800
µ
g/
ft2,
for
floors,
sill
and
troughs,
respectively.
All
Baltimore
HUD
units
that
initially
failed
were
re­
cleaned
and
eventually
met
clearance
standards
upon
re­
testing.

These
results
clearly
suggest
that,
as
anticipated,
the
prescribed
risk
reduction
treatments,
when
coupled
with
full
window
replacement,
achieve
greater
dust
lead
loading
reductions
for
interior
window
sills
and
window
troughs,
as
measured
immediately
after
treatment.
While
the
11/
23/
05
97
prescribed
EA
6­
8
risk
reduction
window
treatments
are
less
costly
and
do
reduce
preintervention
dust
lead
levels
on
interior
window
sills
and
window
troughs,
such
treatments
by
themselves
do
not
appear
to
consistently
remedy
all
lead
hazards
associated
with
windows,
as
measured
by
immediate
post­
intervention
dust
lead
testing.
For
example,
unless
the
lower
sash
is
removed,
it
cannot
be
completely
scraped
and
repainted
before
being
reinstalled.
Therefore,
many
layers
of
built­
up
paint
might
remain,
which
could
result
in
paint
chips
being
easily
dislodged
as
soon
as
the
sash
is
opened
or
closed,
even
after
the
prescribed
window
treatments
have
been
performed.

4.
Do
single
surface
dust
wipe
sample
and
composite
dust
wipe
sample
results
correlate
well
in
measuring
the
amount
of
lead
dust
both
before
and
after
completion
of
the
prescribed
treatments?

This
evaluation
question
is
addressed
in
Appendix
J.

5.
Do
independent
visual
inspections
accurately
assess
whether
the
prescribed
treatments
have
been
completed?

The
most
significant
finding
of
this
study
is
that,
based
on
the
LAAP
inspectors'
confirmatory
visual
assessments,
92
percent
of
the
121
study
units
should
have
"
failed"
the
EA
6­
8
immediate
post­
intervention
independent
visual
inspection.
These
"
failures"
were
driven
by
one
or
more
of
the
prescribed
lead
hazard
reduction
treatments
not
being
performed
or
properly
completed,
as
assessed
by
the
LAAP
inspectors.
In
addition,
the
magnitude
of
these
identified
"
failures"
was
large,
indicating
that
most
reported
"
failures"
should
have
been
easily
observable
to
the
independent
visual
inspectors.
Considering
that
each
of
the
units
passed
a
visual
inspection
performed
by
a
certified
independent
inspector,
this
result
suggests
that
such
independent
visual
inspections
did
not
accurately
determine
whether
the
prescribed
treatments
were
met.

While
performing
their
dust
sampling
and
visual
assessments,
LAAP
inspectors
also
noted
anecdotally
that
the
five
or
six
independent
visual
inspectors
they
observed
spent
very
little
time
(
e.
g.,
five
to
10
minutes
in
many
cases)
conducting
the
visual
inspections,
suggesting
either
carelessness
or
a
lack
of
training
in
the
level
of
detail
needed
to
completely
visually
inspect
a
unit
for
compliance
with
the
prescribed
requirements.

B.
Recommendations
Based
on
the
study
findings,
the
Center
makes
the
following
recommendations:

1.
Increased
oversight
of
independent
visual
inspectors
by
appropriate
state
agency
officials
is
needed
to
assure
that
such
inspectors
are
performing
visual
inspections
in
accordance
with
approved
protocols
and
inspector
training.

The
LAAP
inspectors'
observations
suggest
that
in­
field
compliance
checks
of
independent
visual
inspectors
are
critical
to
assuring
that
these
inspections
are
properly
performed.
Frequent
oversight
evaluations
of
the
inspectors
may
be
necessary
to
emphasize
regulatory
commitment
to
proper
performance
of
the
visual
inspection
protocol
and
to
ensure
that
fatigue
and/
or
routine
11/
23/
05
98
does
not
result
in
incomplete
inspections.
Additionally,
state
agency
officials
may
want
to
review
and,
as
necessary,
revise
the
protocol
and
training
for
visual
inspectors.

Neither
EA
6­
8
nor
the
Code
of
Maryland
Regulations
(
COMAR)
specifies
objective
standards
defining
a
visual
inspection
"
success"
or
"
failure."
Further
research
(
outside
the
scope
of
this
study)
is
needed
to
provide
data
that
would
help
identify
"
de
minimis"
standards
for
visual
inspections,
i.
e.,
in
identifying
thresholds
at
which
certain
observable
conditions
(
e.
g.,
binding
doors)
contribute
to
elevated
dust
lead
loadings.
Until
such
data
are
available,
visual
inspectors
should
be
trained
to
consider
any
easily
observed
"
defect"
on
the
surface
types
and
components
listed
in
Section
6­
815(
A)(
2)
to
constitute
a
"
failure,"
especially
given
that
many
large
and
easily
observed
"
failures"
remained
in
the
study
units.
In
particular,
attention
should
be
paid
to
treatment
"
failures"
of
non­
intact
paint,
visible
paint
chips
and
debris,
and
rubbing
of
doors.

2.
Clearance
dust
testing,
upon
completion
of
the
prescribed
lead
hazard
reduction
treatments,
should
be
added
to
the
independent
visual
inspection
as
part
of
the
lead
hazard
reduction
treatment
requirements
of
EA
6­
8,
,
either
by
regulation
or
amendment
to
the
statute.

Notwithstanding
the
"
failure"
of
92
percent
of
the
independent
visual
inspections
to
accurately
determine
whether
the
study
units
were
in
compliance
with
the
risk
reduction
standard,
lead
dust
sample
results
obtained
immediately
after
intervention
revealed
that
a
large
percentage
of
units
also
had
at
least
one
bare
floor,
interior
window
sill
or
window
trough
lead
dust
result
that
would
have
"
failed"
clearance
testing
had
such
a
test
been
required
by
EA
6­
8.
This
finding
confirms
that
visual
inspections
alone
do
not
reveal
all
lead
hazards.
Although
dust
clearance
sampling
evaluates
dust
lead
loading
only
at
a
single
point
in
time
and
may
not
represent
conditions
existing
throughout
a
dwelling
unit,
such
sampling,
when
performed
in
conjunction
with
a
visual
inspection,
would
likely
be
more
effective
in
determining
the
immediate
post­
intervention
status
of
the
unit
at
the
most
critical
point:
immediately
before
new
tenants
assume
residency.
For
these
reasons,
it
is
recommended
that,
for
units
in
which
the
owner
chooses
the
Lead
Hazard
Reduction
treatment
option
of
EA
6­
8,
a
clearance
dust
testing
requirement
be
added
to
the
immediate
post­
intervention
independent
visual
inspection
requirement,
either
by
regulation
or
by
amendment
to
the
statute.
In
other
words,
a
"
clearance
examination,"
including
dust
testing
and
a
visual
inspection,
should
be
performed
whenever
an
owner
carries
out
prescribed
lead
hazard
reduction
treatments.

3.
Rental
property
owners
and
their
crews
and
independent
contractors
must
perform
a
more
intensive
final
cleaning
upon
completion
of
the
prescribed
lead
hazard
reduction
treatments.

Dust
sample
results
suggest
that
final
cleaning
after
completion
of
the
prescribed
treatments
was
not,
in
many
cases,
sufficient
to
meet
clearance
dust
standards,
had
they
existed
as
a
requirement
in
EA
6­
8.
It
is
presumed
that
those
who
meet
such
clearance
standards
and
who
pass
the
visual
inspection
are
performing
adequate
cleaning.
In
the
absence
of
clearance
testing,
owners
and/
or
contractors
cannot
know
if
dust
lead
loading
has
been
reduced
below
such
standards.
However,
these
results
point
out
that
a
thorough
final
cleaning
is
critical
to
reducing
all
lead
hazards.
For
example,
a
substantial
number
of
treatment
visual
assessment
"
failures"
observed
in
the
study
11/
23/
05
99
units
were
due
to
paint
chips
and
debris
being
present
even
after
the
lead
hazard
reduction
treatments
had
been
completed.
Owners'
and
contractors'
training
must
highlight
and
emphasize
proper
cleaning
techniques
to
increase
the
likelihood
that
dust
lead
loading
can
be
reduced
even
when
dust
samples
are
not
taken
to
confirm
the
results.

4.
More
comprehensive
window
treatments
may
be
necessary
to
assure
that
interior
window
sills
and
window
troughs
do
not
continue
to
be
an
exposure
source
to
lead.

Study
results
suggest
that
the
window
treatments
prescribed
in
the
statute
produce
reductions
in
lead
dust
and
debris
that
may
not
be
sufficiently
protective
based
on
relative
percent
changes
and
clearance
"
failures,"
particularly
in
housing
units
that
are
in
poor
condition.
Study
units
that
underwent
full
window
replacement
experienced
a
larger
reduction
in
dust
lead
loading
on
interior
window
sills
and
window
troughs,
as
measured
by
both
composite
and
single
surface
sampling
results,
than
the
units
undergoing
just
the
prescribed
window
treatments
and
associated
final
cleaning.
Also,
by
applying
clearance
lead
dust
standards
for
sills
and
troughs
to
single
surface
sample
results,
a
higher
percentage
of
study
units
that
received
just
the
prescribed
window
treatments
would
have
"
failed"
clearance
as
opposed
to
those
units
that
underwent
window
replacement.
As
older
window
sashes
coated
with
many
layers
of
lead­
based
paint
continue
to
rub
against
the
window
jamb,
parting
bead
and
stops,
paint
can
chip
off
and
fall
onto
the
sill
and/
or
trough.
Even
when
paint
on
window
surfaces
has
been
stabilized,
a
sill
has
been
stripped
and
repainted
or
replaced
and
a
trough
cap
has
been
installed,
paint
chips
and
debris
can
still
be
generated
due
to
ongoing
friction
of
the
window
component
surfaces.
For
these
reasons,
appropriate
state
agency
officials
should
consider
thoroughly
reviewing
prescribed
window
treatment
specifications
for
effectiveness.

The
two­
year
results
suggesting
that
window
component
dust
lead
loadings
remain
much
lower
in
window
replacement
units
indicates
that
linking
retreatment
to
tenant
turnover
makes
more
sense
than
conducting
annual
re­
inspections.

5.
More
emphasis
needs
to
be
placed
on
properly
adjusting
and
re­
hanging
doors
to
eliminate
friction
points.

Many
doors
in
the
study
units
were
still
rubbing
or
binding
after
the
intervention
had
been
completed.
Furthermore,
the
independent
visual
inspectors
typically
"
passed"
units
in
which
one
or
more
doors
still
required
adjusting
and
re­
hanging.
Owners
and
contractors
need
more
thorough
training
on
how
to
properly
address
this
prescribed
treatment.
Additionally,
visual
inspectors
must
be
trained
to
check
each
door
throughout
a
unit
to
make
certain
that
it
no
longer
rubs
or
binds,
or
otherwise
causes
friction.
