NEIGHBORHOOD DISADVANTAGE, RACE, AND
PROSTATE CANCER PRESENTATION, TREATMENT, AND
MORTALITY
Judith A. Long, MD
Nicole Lurie, MD
Jose Escarce, MD, PhD
Chantal M. Montagnet
Katrina Armstrong, MD
NCI P50-CA105641-05 CPHHD
Prostate Cancer and Neighborhood
• Burden of prostate cancer higher in African
Americans (AA) (incidence, stage, and mortality)
• Individuals are nested within social and physical
environments
• Neighborhood characteristics differ by race and may
influence disease development as well as access to
and receipt of care
• Which may contribute to racial disparities in prostate
cancer presentation, treatment and survival
Objective
Investigate the relationship between neighborhood SES,
race, and prostate cancer
• Presentation
• Treatment
• Mortality
SEER-Medicare Data (1991-1999)
111,640
104,876
92,783
59,388
53,591
• Incident Cases of Prostate Cancer
• AA and Whites (94%)
• Census Tract Assignable (83%)
• Block Group Income Assignable (53%)
• Staged Disease (48%)
Outcomes
Entire Sample
• Odds of metastatic DX vs non-metastatic DX
Localized DX (Non-Metastatic)
• Odds of receiving active TX (Radiation or
Prostatectomy) vs no TX
Localized DX (Non-Metastatic)
• Hazard of prostate CA mortality by Dec. 31st
2003
Neighborhood
SES
• Composite index
developed by RAND
• Uses interpolated
census data
• Variables chosen using
factor analysis
• Index range 0-100
• Higher scores indicate
higher neighborhood
SES
% Adults < HS
% Male
Unemployment
% Households in
Poverty
% Households w/
Public Assistance
% of Households
Headed by a
Female
Mean Household
Income
Analysis
• Stage at presentation: logistic regression
• TX received: multinomial logit models
• Prostate CA mortality: cox regression
• Main predictors: neighborhood index quartile, race
• Covariates: age, marital status, comorbidities, block
group income, SEER site, grade at presentation (1 or
2 versus 3 or 4), and TX
• Stepwise models: to determine the influence of
covariates on the coefficient for race
Sample Characteristics
SES Index Quartile
N
Age, m ± sd
Block Group Inc., m ± sd
African American, %
Married,%
Comorbidities, %
0
1
2
3
4 or more
Metastatic, %
Total
53,591
74 ± 6.0
56 ± 30K
11
73
35
29
16
9
11
10
Lowest
13,024 12,663 13,464
74 ± 6.2 74 ± 6.0 73 ± 5.9
33 ± 12K 44 ± 12K 55 ± 14K
38
4
2
63
75
76
28
26
18
10
17
13
36
29
16
9
10
11
36
30
16
8
10
9
Highest
14,440
73 ± 5.6
90 ± 33K
1
79
p value
<0.0001
<0.0001
<0.0001
<0.0001
40
29
15
8
7
9
<0.0001
<0.0001
* Sample: All Blacks and Whites, Metastatic and non-Metastatic. Other Race and Unstaged disease
excluded
Neighborhood, Race, and Outcomes
SES Index Quartile
Highest SES
Lowest SES
African American Race
Metastatic DX
vs Local DX
OR (95% CI)
RX vs
No TX*
OR (95% CI)
1.0
0.99
(0.90-1.09)
1.08
(0.97-1.20)
1.18
(1.05-1.33)
1.45
(1.31-1.40)
1.0
0.95
(0.87-1.02)
0.85
(0.76-0.93)
0.74
(0.65-0.83)
0.79
(0.70-0.88)
Prostatectomy
Prostate
vs No TX*
CA Mortality**
OR (95% CI)
HR (95% CI)
1.0
1.03
(0.93-1.12)
0.87
(0.77-0.98)
0.70
(0.58-0.82)
0.61
(0.49-0.73)
1.0
1.08
(1.02-1.37)
1.24
(1.05-1.46)
1.29
(1.07-1.54)
1.18
(1.01-1.37)
All models adjusts for age, marital status, comorbidities, block group income, and SEER Site
* Models also adjusts for grade at presentation (1,2 versus 3,4)
**Models also adjusts for grade at presentation (1,2 versus 3,4) and treatment
Degree to Which Neighborhood SES
Explains Racial Disparities
African American Race
(race only)
African American Race
(+ demographics and clinical)*
African American Race
(+ treatment)
African American Race
(+ neighborhood SES)
Metastatic DX
vs Local DX
RX vs
No TX
Prostatectomy
Prostate
vs No TX
CA Mortality
OR (95% CI)
1.51
(1.39-1.63)
1.54
(1.41-1.69)
OR (95% CI)
0.85
(0.79-0.91)
0.72
(0.64-0.80)
OR (95% CI)
0.54
(0.46-0.62)
0.52
(0.41-0.63)
1.45
(1.31-1.40)
0.79
(0.70-0.88)
0.61
(0.49-0.73)
HR (95% CI)
1.37
(1.22-1.54)
1.22
(1.06-1.40)
1.21
(1.06-1.40)
1.18
(1.01-1.37)
*All models adjusts for age, marital status, comorbidities, block group income, and SEER Site, models
evaluating mortality and treatment also include grade at presentation (1,2 versus 3,4)
Conclusions
• Neighborhood SES associated with prostate cancer
stage of presentation, TX, and mortality
• Neighborhood SES did little to explain racial
disparities in mortality
• Neighborhood SES did explain to some extent racial
disparities in presentation and treatment
Limitations
• We lacked specific individual level SES data (possible
ecological fallacy)
• We lacked block group income for about 30% of the
identified sample
• We do not know how patient driven decisions may be
influencing our results
Implications
• Patients from low SES neighborhood and AA men
may be receiving cancer care from providers or sites
where less aggressive screening and treatment is
pursued
• Targeting care within these sites may be a means to
ameliorating racial disparities in prostate cancer
presentation and care