Page 1
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Paul A. Godley, MD, PhD
Prostate Cancer
Prostate Cancer
Anatomy
Cancer Incidence, Males
Cancer Incidence, Males
2004 Estimated New Cancer Cases
Prostate 230,110
Lung 93,110
Colorectal 73,620
Bladder 44,640
Prostate Cancer Epidemiology
Prostate Cancer Epidemiology
Prostate Cancer Incidence per 100,000 Males
Prostate Cancer Epidemiology
Prostate Cancer Epidemiology
Incidence Trends by Race 1973-1999
0
50
100
150
200
250
300
Age-adjusted
Incidence
per
100,000
Males
Year of Diagnosis
Cancer Mortality, Males
Cancer Mortality, Males
2003 Estimated Cancer Deaths
Lung 88,400
Prostate 28,900
Colorectal 28,300
Page 2
Prostate Cancer Epidemiology
Prostate Cancer Epidemiology
Age-adjusted Prostate Cancer Death Rates by State and Race
1984-1988
U.S. Cancer Mortality, Males
U.S. Cancer Mortality, Males
1991-1995
All Sites
Lung
Prostate
Colorectal
Oral
Lymphatic
-15
-10
-5
0
5
10
Percent
Change
-4.3
-8.7
-2.1
-6.7
-13.8
-2.6
-6.3
-7.4
-3.8
-7
-5.2
-7.6
-10.4
-15
-5.9
4.1
0.4
6.3
All Ages
<65
65+
U.S. Cancer Mortality, Prostate Cancer
U.S. Cancer Mortality, Prostate Cancer
1971-90 vs. 1991-95
White Men
Black Men
-10
-5
0
5
10
15
20
25
30
35
40
Percent
Change
19.6
-6.7
34.7
-4.5
1971-90
1991-95
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Rationale
Stage-standardized variations in the treatment of 251,416 prostate cancer
patients diagnosed in the U.S. between 1992 and 1994 by race.
Percentage
of Patients
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Rationale
00Previous studies had assumed that patients receiving identical
treatments had similar outcomes. Our goal was to investigate racial
disparities in prostate cancer mortality by treatment choice.
00About 75% of the estimated 220,900 prostate cancers diagnosed in
2003 will be clinically localized at the time of diagnosis
00Several studies of localized prostate cancer, some based in health
care systems that lacked financial barriers to treatment, found no
differences in survival between black and white patients
Page 3
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Previous Studies
00Biochemical recurrence rates of 148 white (dashed line) and 125 black (solid line) men who
underwent radical prostatectomy for prostate cancer as determined by the Kaplan-Meier
method. No differences were found in PSA recurrence between the two ethnic groups ( P =
0.651).
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
SEER
00Surveillance Epidemiology and End Results (SEER)
00SEER Program collects data on 120,000 new cancer cases yearly
00States of Connecticut*, Iowa, New Mexico, Utah, Hawaii and the
metropolitan areas of Detroit*, San Francisco*, Seattle-Puget
Sound*, and Atlanta*
00Hospitals, Laboratories, clinics, physician offices
00Report incidence rates adjusted to the U.S. 1970 standard
population
00Monitor incidence patterns and changes in cancer therapy and
patient survival
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Medicare
00Provides health insurance coverage for 97% of American age 65
and older
00Information available through the Medicare program includes health
service claims for care provided by
00Physicians
00Inpatient hospital stays
00Hospital outpatient clinics
00Home health care agencies
00Skilled nursing facilities
00Hospice programs
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Methods
00Surveillance, Epidemiology and End Results (SEER) tumor registry
data were used
00to identify prostate cancer patients
00determine their initial treatment
00follow them for outcomes over time
00SEER-Medicare linked files
00Facilitate following cancer patients from their diagnosis beyond
their initial treatment window
00Allow examination of care provided throughout the course of the
disease.
00Provide information about other diseases and conditions that
might confound the relationship between race and survival.
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Study Population
N=11,353
with prior cancer
Dropped
N=665
2nd Cancer diagnosed
in same month dropped
N=3,386
in situ cancer dropped
N=720
No month of diagnosis
Dropped
N=14,989
age <65 dropped
N=73,424
*
age 65 included
N=88,413
with valid diagnosis month
Included
N=89,133
Invasive Cancer included
N=92,519
Only prostate Cancer
diagnosed that month
N=93,184
prostate cancer
is first cancer included
Prostate Cancer Cases From 5 Regions
Diagnosed 1986-1996, N=104,537
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Study Population
N=4,428
Other races
Dropped
N=814
Diagnosed at death
N=40
No Medicare
during study period
N=22,003
Non-surgery, regional
or advanced stage
N=2,150
*
Age 85
Dropped
N=43,989
Age <85
Study Population
N=46,139
Met staging
criteria
N=68,142
with Medicare
coverage included
N=68,182
Diagnosed alive
N=68,996
Caucasian or
African American
N=73,424
*
with age 65
Page 4
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Staging Issues
00Staging information in SEER is based on clinical information except
for surgical patients, whose staging information is updated with
pathological findings
00If analyses of locally staged surgical patients are conducted, the
surgery group will exclude some patients who were initially
staged as local but after surgery were reclassified to a more
advanced stage
00To avoid this potential misclassification, we staged all radical
prostatectomy and lymphadenectomypatients as local
00Retained SEER pathological staging information for the
multivariate analysis.
00Patients receiving other treatments or watchful waiting were
staged as local using SEER information
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Follow up
00Date of death was available in both SEER and Medicare files
00Medicare dates of death provided the longest follow-up (through
1998) and were used for analyses of all cause mortality.
00Only SEER data provided the cause of death, SEER death dates
were used for analyses of prostate cancer-specific mortality
00Survival was measured in months from the date of diagnosis to the
end of the study period.
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Covariates
00No individual measures of socioeconomic status were available
00SEER data included a number of census-based measures linked to the residence
of the patient
00Three measures were divided into quartiles to adjust for socioeconomic status:
00The percent of persons of the same race residing in the patient's census tract
that had incomes below the poverty level
00The median income of the census tract
00The percent of persons of the same race in the census tract thathad less than
high school education
00Other Variables
00SEER site, age at diagnosis, PSA period (1986-1988, pre-PSA period;
1989-1991, early PSA period; 1992-1996, recent PSA period), Gleason
grade, and stage
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Comorbidity
00A modified Charlsonindex comorbidityscore was calculated from
hospital discharge diagnoses in Medicare claims
00Inpatient claims for the year prior to diagnosis were used
00Only 17,659 subjects had comorbidityinformation
Characteristics of Study Population
Characteristics of Study Population
No. of white
No. of black
patients
patients
Characteristic
(% of total)
(% of total)
All patients
38 242 (86.9) 5747 (13.1)
Age group, y00/span>
65-69
12 839 (33.6) 2149 (37.4)
70-74
13 057 (34.1) 1902 (33.1)
75-79
8490 (22.2) 1164 (20.3)
80-84
3856 (10.1)
532 (9.3)
Era of Diagnosis00/span>
Pre-PSA (1986-1988)
7489 (19.6) 1019 (17.7)
Early PSA (1989-1991)
11 122 (29.1) 1356 (23.6)
Recent PSA (1992-1996)
19 631 (51.3) 3372 (58.7)
Marital status00/span>
Single
2064 (5.4) 673 (11.7)
Married
29 442 (77.0) 3426 (59.6)
Divorced/separated/widowed
4521 (11.8) 1154 (20.1)
Not reported/missing
2215 (5.8)
494 (8.6)
SEER Site00/span>
San Francisco
6904 (18.1) 1114 (19.4)
Connecticut
8363 (21.9)
516 (9.0)
Detroit
9553 (25.0) 2977 (51.8)
Seattle
10 532 (27.5) 222 (3.9)
Atlanta
2890 (7.6) 918 (16.0)
*Not all percentages add to 100 because of rounding.
00wo-sided P <.001 ( X2 test).
Characteristics of Study Population
Characteristics of Study Population
Page 5
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Median Survival time and Survival Rates by Race
Median Survival time, y (95% CI)
10-year survival rate (95% CI)
Mortality endpoint
White Patients Black Patients
White Patients Black Patients
All-cause mortality
9.8 (9.6 to 9.9) 8.1 (7.8 to 8.3)
0.49 (0.48 to 0.50) 0.38 (0.36 to 0.41)
Nonaggressive
6.8 (6.6 to 6.9) 5.8 (5.7 to 6.2)
9.32 (0.31 to 0.34) 0.26 (0.23 to 0.28)
treatment
Radiation therapy
9.4 (9.2 to 9.7) 8.7 (8.4 to 9.3)
0.47 (0.46 to 0.49) 0.42 (0.38 to 0.46)
Surgery
12.6 (12.4 to NA00 10.8 (10.0 to 11.7) 0.64 (0.62 to 0.65) 0.55 (0.51 to 0.60)
Prostate cancer mortality
-- --
0.90 (0.89 to 0.90) 0.84 (0.82 to 0.86)
Nonaggressive
-- --
0.87 (0.86 to 0.88) 0.79 (0.76 to 0.83)
treatment
Radiation therapy
-- --
0.89 (0.88 to 0.90) 0.86 (0.83 to 0.90)
Surgery
-- --
0.92 (0.92 to 0.93) 0.88 (0.84 to 0.91)
Non-prostate cancer
10.8 (10.61 to 10.9) 9.1 (8.8 to 9.6) 0.55 (0.54 to 0.55) 0.45 (0.43 to 0.48)
mortality
-- --
*Survival times and rates were obtained from Kaplan-Meier survival estimates. CI = confidence interval; - = median survival time not
yet reached.
00pper limit to confidence interval is not available because of censoring.
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Survival Curves by Race and Treatment
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Overall Survival by Race
Overall Survival After Prostate Cancer Treatment
Overall Survival After Prostate Cancer Treatment
Survival Curves for Surgery Cases by Time Period
Survival After Prostate Cancer Treatment
Survival After Prostate Cancer Treatment
Survival Curves by Race
Page 6
Adjusted Survival After Prostate Cancer Treatment
Adjusted Survival After Prostate Cancer Treatment
Median Survival time, y
10-year survival rate (95% CI)
Estimate type
White Patients Black Patients White Patients Black Patients
All patients
Unadjusted
9.8 8.0
0.49 (0.48 to 0.50) 0.38 (0.37 to 0.40)
Adjusted00/span>
10.4 7.9
0.52 (0.50 to 0.54) 0.38 (0.35 to 0.41)
Surgery patients
Unadjusted
12.6 10.4
0.64 (0.63 to 0.65) 0.53 (0.50 to 0.56)
Adjusted00/span>
-- 11.3
0.70 (0.68 to 0.71) 0.56 (0.52 to 0.61)
*Unadjusted estimates were obtained from Cox proportional hazards survival estimates; adjusted estimates were obtained by
the mean covariate level method. CI = confidence interval; - = median survival was not reached.
Adjusted for treatment; age; Surveillance, Epidemilogy, and End Results (SEER) Proram site; census tract educational level
and median household income; marital stats; cancer grade; PSA era; and comorbidity score.
0Adjusted for pathologic stage, age, SEER site, census tract educational level and median household income, marital status,
cancer grade, PSA era, and comorbidity score.
Adjusted Survival After Prostate Cancer Treatment
Adjusted Survival After Prostate Cancer Treatment
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Conclusions
00African Americans suffer disparities after all treatment modalities, but most
prominently after surgery for localized prostate cancer
00Racial differences in treatment outcomes become less pronounced over time
00The consistency of the racial disparity across prostate cancer treatment, PSA era,
and cause-specific mortality may suggest a global social or environmental factor
that disproportionately diminishes survival among blacks regardless of how
mortality is measured
00Another possible mechanism for these results is a selection process systematically
choosing poorer prognosis black patients than white patients forprostate cancer
surgery
00Objective PSA criteria may explain the apparent decline over time in the racial
differences in treatment outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Next Steps
00SEER/Medicare analyses including hospital prostatectomy
experience
00SEER/Medicare analyses of prostate cancer relapse-free
survival
00Survey prostate cancer patients
00Treatment decisions
00Access to cancer care
Racial Differences in Prostate Cancer Treatment
Outcomes
Racial Differences in Prostate Cancer Treatment
Outcomes
Investigators
00Anna Schenck, MSPH, PhD
00Medical Review of North Carolina, Inc.
00Victor J. Schoenbach, PhD
00M. AhineeAmamoo, MS
00Michael Symons, PhD
00Michelle Manning, MPH
00James A. Talcott, MD, SM
00Center for Outcomes Research, Massachusetts General Hospital Cancer
Center and Harvard Medical School
00UNC Center for Health Promotion and Disease Prevention
00Association of Teachers of Preventive Medicine/ CDC
download Racial Differences in Prostate Cancer Treatment Outcomes
