Racial differences in time from prostate cancer diagnosis to treatment initiation

A Population-Based Study

Authors

  • William A. Stokes BS,

    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Laura H. Hendrix MS,

    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Trevor J. Royce MS,

    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Ian M. Allen BS,

    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Paul A. Godley MD, PhD,

    1. Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    2. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    3. Division of Hematology/Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Andrew Z. Wang MD,

    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    2. Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    Search for more papers by this author
  • Ronald C. Chen MD, MPH

    Corresponding author
    1. Department of Radiation Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    2. Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    3. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina
    • Corresponding author: Ronald C. Chen, MD, MPH, Department of Radiation Oncology, University of North Carolina at Chapel Hill, 101 Manning Drive, CB #7512, Chapel Hill, NC 27516; Fax: (919) 966-7681; Ronald_chen@med.unc.edu

    Search for more papers by this author

  • Previously presented at a poster discussion session at the 2012 American Society for Clinical Oncology annual meeting.

  • Supported in part by a grant from the Howard Holderness Distinguished Medical Scholars Program at the University of North Carolina at Chapel Hill.

Abstract

BACKGROUND

Timely delivery of care has been identified by the Institute of Medicine as an indicator for quality health care, and treatment delay is a potentially modifiable obstacle that can contribute to the disparities among African American (AA) and Caucasian patients in prostate cancer recurrence and mortality. Using the Surveillance, Epidemiologic and End Results (SEER)-Medicare linked database, we compared time from diagnosis to treatment in AA and Caucasian prostate cancer patients.

METHODS

A total of 2506 AA and 21,454 Caucasian patients diagnosed with localized prostate cancer from 2004 through 2007 and treated within 12 months were included. Linear regression was used to assess potential differences in time to treatment between AA and Caucasian patients, after adjusting for sociodemographic and clinical covariates.

RESULTS

Time from diagnosis to definitive (prostatectomy and radiation) treatment was longer for AA patients in all risk groups, and most pronounced in high-risk cancer (96 versus 105 days, P < .001). On multivariate analysis, racial differences to any and definitive treatment persisted (β = 7.3 and 7.6, respectively, for AA patients). Delay to definitive treatment was longer in high-risk (versus low-risk) disease and in more recent years.

CONCLUSIONS

AA patients with prostate cancer experienced longer time from diagnosis to treatment than Caucasian patients with prostate cancer. AA patients appear to experience disparities across all aspects of this disease process, and together these factors in receipt of care plausibly contribute to the observed differences in rates of recurrence and mortality among AA and Caucasian patients with prostate cancer. Cancer 2013;119:2486-2493. © 2013 American Cancer Society.

INTRODUCTION

Prostate cancer is the most common malignancy in men, diagnosed in more than 240,000 US men per year, and the second leading cause of male cancer deaths, killing more than 28,000 annually.[1] Compared to Caucasians, AA men are more likely to develop prostate cancer and 2.4 times as likely to die from it.[2] The differential prostate cancer mortality rates in AA versus Caucasian men represent the largest disparity of any malignancy.[3] This disparity has been attributed to more aggressive disease,[4] less aggressive treatment,[5] worse access to health care,[6] and lower socioeconomic status[7] among AA versus Caucasian patients.

The Institute of Medicine has identified timely delivery of services as 1 of 6 major goals for improving the quality of health care in the United States.[8] Treatment delay in prostate cancer is associated with increased risk for biochemical recurrence,[9] especially for patients with aggressive (ie, high-risk) disease and is a potentially modifiable factor that may also contribute to racial disparities in prostate cancer. To our knowledge, the time interval from prostate cancer diagnosis to treatment initiation in AA and Caucasian patients has not been examined at the population level. In this study, we analyzed data from the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database to examine the interval from diagnosis of low-, intermediate-, and high-risk prostate cancer to treatment initiation in AA and Caucasian patients. We hypothesized that there is greater treatment delay in AA patients, including those with high-risk disease.

MATERIALS AND METHODS

Data Source

Institutional review board waiver was obtained. The SEER-Medicare database combines clinical information from SEER, a National Cancer Institute-supported collection of cancer registries covering approximately 28% of the US population,[13] with health care utilization data from Medicare, the primary health insurance program for Americans ages 65 and older. The merged database enables a population-based analysis of prostate cancer diagnosis and treatment patterns.

Study Population

The cohort included patients diagnosed with prostate cancer between 2000 and 2007 (the most recent year with available data) who had no history of other cancers. We excluded patients diagnosed on autopsy or death certificate. To determine baseline comorbidity status using a validated modified Charlson score developed specifically for use with SEER-Medicare data,[14] we restricted our analysis to those aged 66 and older in order to have 1 year of Medicare claims data prior to prostate cancer diagnosis. Next, to ensure complete capture of health services for the duration of the study period, we excluded men who were enrolled in a health maintenance organization or who were not enrolled in both Medicare Part A and Part B for 12 months following their cancer diagnosis, which is the time period we used to capture initial treatment. Consistent with existing methodology, patients who did not receive treatment within 12 months of diagnosis were deemed to have pursued initial conservative management (surveillance)[15] and excluded. We restricted the cohort to those diagnosed after 2004 because Gleason score and prostate-specific antigen (PSA) data, which are necessary to assign patients to risk categories based on D'Amico criteria,[16] were not available in SEER-Medicare in prior years. Furthermore, because a primary focus of this study was to compare treatment delays in AA versus Caucasian patients, other races were excluded. We also excluded patients 80 years or older and those with metastatic (M1) or locally-advanced disease (T3b, T3NOS, T4 or node positive) because the optimal management for octogenarians and those with advanced disease is unclear.

Thus, the analytic cohort included 23,960 AA and Caucasian patients younger than 80 years old with potentially curable prostate cancer, who received treatment within 12 months of diagnosis.

Outcome Variables

The primary outcome was time from diagnosis to initiation of treatment, defined as the interval in days between diagnosis and the beginning of prostate cancer–directed therapy.

SEER collects the month and year of diagnosis; for this analysis, we assigned the date of diagnosis as the first day of the indicated month. The beginning of treatment was defined as the first occurrence in Medicare claims data of prostate cancer treatment according to Healthcare Common Procedure Coding System codes (Table 1). We examined time-to-treatment in 2 ways: first, we considered any therapy for prostate cancer, including radical prostatectomy, external beam radiation therapy, brachytherapy, and hormonal therapy. Second, we examined time-to-definitive treatment, which included all of the aforementioned modalities except hormonal therapy.

Table 1. Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD9) Indicators for Prostate Cancer Treatment Modalities
Treatment TypeCPT and ICD9 Codes
Brachytherapy77326, 77327, 77328, 77776, 77777, 77778, 77781, 77782, 77783, 77784, 77790, 77799, Q3001, A9527, C1715, C1716, C1717, C1719, C1728,C2616, C2634, C2635, C2636, C2637, C2638, C2639, C2640, C2641, C2642, C2643, C2698, C2699, C9725, 55875, 55876, 76873, 77761, 77762, 77763, 77785, 77786, 77787, 77789
External beam radiation therapy77371, 77372, 77373, 77402, 77403, 77404, 77406, 77407, 77408, 77409, 77411, 77412, 77413, 77414, 77416, 92.26, 77418, 0073T, 77380, 77381, 77520, 77522, 77523, 77525, G0174, 77401, 77431, 0082T, 0083T, 77432, 77435, G0251, G0339, G0340
Hormonal therapy54520, 54522, 54530, 54535, 54690, J1050, J1051, J1950, J3315, J9202, J9217, J9218, J9219, J9165, S0175, C9216, C9430, G0356, J0128, S0165, S9560, G9132, 4164F, 62.3, 62.4, 62.41, 62.42
Prostatectomy55866, 55810, 55812, 55815, 55840, 55842, 55845, 60.5

Covariates

Demographic characteristics including race, age and year of diagnosis, and marital status were obtained from SEER. Prostate cancer risk category was defined using the D'Amico criteria: low-risk (clinical stage T1c or T2a, and PSA ≤ 10 ng/mL, and Gleason score ≤ 6), intermediate-risk (T2b, or PSA > 10 and ≤ 20 ng/mL, or Gleason score 7), or high-risk (T2c or T3a, or PSA > 20 ng/mL, or Gleason score ≥ 8).[16] Baseline comorbidity score was determined using Medicare claims data from the 12 months preceding prostate cancer diagnosis.[14] Socioeconomic characteristics from SEER included census-tract-level measures of high school graduation and population density. In addition, Medicaid dual-eligibility data from Medicare provides a person-level socioeconomic measure. SEER regions were grouped into Northeast (Connecticut, New Jersey), South (Atlanta, rural Georgia, Kentucky, Louisiana), Central (Detroit, Iowa, New Mexico, Utah), and West (San Francisco, Hawaii, Seattle, San Jose, Los Angeles, greater California).

Statistical Analysis

Descriptive statistics were used to report the median number of days from diagnosis to treatment. Chi-square and Wilcoxon rank sum tests were used to assess potential differences among groups for categorical and continuous variables, respectively. Multivariate linear regression analysis was used to assess whether race was significantly associated with time-to-treatment, while adjusting for age category (66-69, 70-74, 75-79), diagnosis year, marital status, risk group, comorbidity score (stratified by median value of 0),[17] census-tract educational attainment, Medicaid dual eligibility, population density (urban or rural), SEER region, and treatment modality. We examined potential interaction terms race*region, race*risk group, race*diagnosis year, race*age category, and race*comorbidity. For sensitivity analysis, we explored models with log-transformed outcomes (days to treatment), as well as hierarchical models to account for potential clustering by region. These models resulted in qualitatively similar findings and data are not presented.

Statistical significance was set at 0.05; all tests were 2-tailed. Analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).

RESULTS

Table 2 summarizes the characteristics of AA and Caucasian patients in each risk category. In all risk groups, AA patients were younger, had higher baseline comorbidity scores, and were more likely to live in areas with lower educational attainment. AA patients were also less likely to receive prostatectomy and more likely radiation therapy or hormonal therapy.

Table 2. Demographics and Clinical Characteristics by Risk Group and Race
 Low Risk (n = 7317)Intermediate Risk (n = 8086)High Risk (n = 8557)
CharacteristicCA (n = 6651)AA (n = 666)PCA (n = 7226)AA (n = 860)PCA (n = 7577)AA (n = 980)P
  1. Abbreviations: AA, African American; CA, Caucasian; EBRT, external beam radiation therapy; HT, hormone therapy; NCI, National Cancer Institute; RP, radical prostatectomy.
Age at diagnosis, y  .002  .002  <.001
66-692579 (38.8)298 (44.7) 2524 (34.9)329 (38.3) 2197 (29.0)376 (38.4) 
70-742637 (39.7)258 (38.7) 2770 (38.3)350 (40.7) 2930 (38.7)357 (36.4) 
75-791435 (21.6)110 (16.5) 1932 (26.7)181 (21.1) 2450 (32.3)247 (25.2) 
Year of diagnosis  .358  .037  .106
20041697 (25.5)180 (27.0) 1656 (22.9)179 (20.8) 1952 (25.8)264 (26.9) 
20051598 (24.0)171 (25.7) 1639 (22.7)217 (25.2) 1837 (24.2)264 (26.9) 
20061699 (25.6)168 (25.2) 1839 (25.5)242 (28.1) 1977 (26.1)228 (23.3) 
20071657 (24.9)147 (22.1) 2092 (29.0)222 (25.8) 1811 (23.9)224 (22.9) 
Marital status  <.001  <.001  <.001
Married5219 (78.5)390 (58.6) 5555 (76.9)522 (60.7) 5471 (72.2)526 (53.7) 
Not married/unknown1432 (21.5)276 (41.4) 1671 (23.1)338 (39.3) 2106 (27.8)454 (46.3) 
NCI combined comorbidity index score<.001  <.001  <.001
04664 (70.1)394 (59.2) 5049 (69.9)479 (55.7) 4994 (65.9)527 (53.8) 
>01987 (29.9)272 (40.8) 2177 (30.1)381 (44.3) 2583 (34.1)453 (46.2) 
Medicaid dual-eligibility <.001  <.001  <.001
No6270 (94.3)575 (86.3) 6771 (93.7)709 (82.4) 6945 (91.7)743 (75.8) 
Yes381 (5.7)91 (13.7) 455 (6.3)151 (17.6) 632 (8.3)237 (24.2) 
% Non–high school graduate in census tract<.001  <.001  <.001
0%-25%1908 (28.7)58 (8.7) 2139 (29.6)63 (7.3) 1888 (24.9)57 (5.8) 
26%-50%1503 (22.6)87 (13.1) 1560 (21.6)102 (11.9) 1644 (21.7)80 (8.2) 
51%-75%1539 (23.1)133 (20.0) 1674 (23.2)154 (17.9) 1860 (24.6)161 (16.4) 
76%-100%1701 (25.6)388 (58.3) 1853 (25.6)541 (62.9) 2185 (28.8)682 (69.6) 
Population density  <.001  <.001  <.001
Urban5649 (84.9)620 (93.1) 6158 (85.2)786 (91.4) 6230 (82.2)865 (88.3) 
Rural1002 (15.1)46 (6.9) 1068 (14.8)74 (8.6) 1347 (17.8)115 (11.7) 
Geographic region  <.001  <.001  <.001
Northeast1811 (27.2)229 (34.4) 1731 (24.0)255 (29.7) 1768 (23.3)246 (25.1) 
South1371 (20.6)213 (32.0) 1348 (18.7)302 (35.1) 1373 (18.1)374 (38.2) 
Central1035 (15.6)121 (18.2) 1302 (18.0)142 (16.5) 1544 (20.4)188 (19.2) 
West2434 (36.6)103 (15.5) 2845 (39.4)161 (18.7) 2892 (38.2)172 (17.6) 
Any treatment modality .001  <.001  <.001
EBRT1590 (23.9)180 (27.0) 1377 (19.1)216 (25.1) 908 (12.0)156 (15.9) 
Brachytherapy2129 (32.0)201 (30.2) 1017 (14.1)123 (14.3) 633 (8.4)66 (6.7) 
RP1284 (19.3)88 (13.2) 2176 (30.1)160 (18.6) 1643 (21.7)137 (14.0) 
HT1648 (24.8)197 (29.6) 2656 (36.8)361 (42.0) 4393 (58.0)621 (63.4) 
Definitive treatment modality <.001  <.001  <.001
EBRT2552 (38.4)302 (45.4) 3366 (46.6)504 (58.6) 4480 (59.1)706 (72.0) 
Brachytherapy2769 (41.6)276 (41.4) 1569 (21.7)188 (21.9) 1278 (16.9)130 (13.3) 
RP1330 (20.0)88 (13.2) 2291 (31.7)168 (19.5) 1819 (24.0)144 (14.7) 

Time From Diagnosis to Treatment

Time from diagnosis to a definitive treatment (prostatectomy, external beam radiation therapy, or brachytherapy) (Table 3) was longer for AA patients by a median of 4 (intermediate risk, 92 versus 96 days) to 9 days (high-risk, 96 versus 105 days). However, examining the interquartile ranges (eg, 75th percentile for high-risk cancer: 135 days for Caucasian and 155 days for AA patients) demonstrates that the disparity for some patients is 20 days or greater. In high-risk disease, older age, more recent years, being unmarried, and Medicaid dual-eligibility were all associated with longer time from diagnosis to treatment.

Table 3. Days from Diagnosis to Definitive Treatment by Risk Group
 Low RiskIntermediate RiskHigh Risk
 Median (IQR)PMedian (IQR)PMedian (IQR)P
  1. Abbreviations: AA, African American; CA, Caucasian; EBRT, external beam radiation therapy; IQR, interquartile range; NCI, National Cancer Institute; RP, radical prostatectomy.
Race.002 .003 <.001
CA89 (61-133) 92 (62-131) 96 (65-135) 
AA97 (62-152) 96 (67-143) 105 (68-155) 
Age at diagnosis, y.156 .140 <.001
66-6991 (63-133) 91 (63-128) 92 (63-131) 
70-7490 (61-136) 83 (63-134) 97 (66-137) 
75-7987 (58-136) 84 (62-133) 102 (69-147) 
Year of diagnosis<.001 <.001 .001
200487 (60-132) 89 (61-130) 96 (64-138) 
200585 (59-128) 92 (61-129) 95 (64-133) 
200691 (62-139) 95 (65-137) 96 (66-137) 
200792 (64-138) 94 (65-132) 100 (69-141) 
Marital status.011 <.001 <.001
Married89 (61-132) 92 (63-129) 95 (64-133) 
Not married/unknown92 (62-141) 96 (63-141) 102 (68-152) 
NCI combined comorbidity index score.078 .135 .235
089 (61-133) 92 (63-131) 96 (65-136) 
>091 (62-138) 93 (63-134) 96 (66-139) 
% Non–high school graduate in census tract.131 .192 .321
0%-25%90 (62-136) 93 (63-133) 95 (65-135) 
26%-50%91 (62-133) 91 (63-127) 96 (68-137) 
51%-75%90 (61-140) 94 (63-132) 96 (67-135) 
76%-100%86 (61-129) 93 (62-134) 97 (64-140) 
State Medicaid dual-eligibility.793 .007 <.001
No90 (61-134) 92 (63-131) 96 (65-136) 
Yes88 (61-133) 98 (67-142) 104 (69-154) 
Geographic region<.001 <.001 <.001
Northeast102 (68-156) 108 (74-155) 111 (75-155) 
South82 (57-123) 82 (57-116) 84 (59-123) 
Central76 (52-121) 80 (58-123) 87 (62-126) 
West87 (60-130) 92 (62-131) 98 (67-140) 
Population density.002 .001 <.001
Urban91 (61-135) 93 (63-132) 98 (66-138) 
Rural83 (57-128) 87 (59-127) 92 (61-131) 
Definitive treatment modality<.001 <.001 .018
EBRT93 (63-138) 98 (66-145) 107 (72-153) 
Brachytherapy87 (57-138) 93 (59-138) 104 (65-146) 
RP85 (60-120) 82 (60-115) 75 (55-103) 

Results for time to any treatment (which includes hormonal therapy) also demonstrated longer intervals for AA versus Caucasians men (P < .02 for all risk groups) (data not shown).

Similar results were found by examining the proportion of AA versus Caucasian patients who received any or definitive treatment within 6 months of diagnosis (data not shown). Overall, greater than 85% of patients who underwent treatment received it within 6 months. AA patients in different subgroups were 2% to 4% less likely than Caucasian patients to initiate treatment within 6 months of diagnosis.

Multivariate Analysis

After adjusting for covariates, AA race remained significantly associated with longer time from diagnosis to any or definitive treatment (β = 7.3 and 7.6, respectively) (Table 4). Patients diagnosed in more recent years and those who were unmarried also had significantly longer time to treatment. Furthermore, patients with high-risk cancer compared to low-risk cancer had a longer time to definitive treatment (β = 4.3, P < .001). When considering any treatment modality, the time to hormonal therapy was significantly shorter compared to external beam radiation therapy (β = −30.6, P < .001). There was regional variation in time to treatment. Of the tested interaction terms, only race*region was significantly associated with the outcomes and therefore included in the models.

Table 4. Multivariate Linear Regression Models for Time From Diagnosis to Any or Definitive Treatment
 Any TreatmentaDefinitive Treatmenta
 β (days)P95% CIβ (days)P95% CI
  1. aThe intercept for the “any treatment” model is 95.0 days, and for the “definitive treatment” model is 120.6 days.
  2. Abbreviations: AA, African American; CA, Caucasian; EBRT, external beam radiation therapy; HT, hormone therapy; NCI, National Cancer Institute; RP, radical prostatectomy.
  3. Sample calculation (days from diagnosis to definitive treatment): patients aged 66 years, diagnosed in 2007, married, high-risk cancer, no comorbidity, census tract with 0%-25% non–high school graduates, not Medicaid dual-eligible, living in urban area, in the Northeast, and receiving external beam radiation therapy:
    • Caucasian: 120.6 + 4.2 + 4.3 = 129.1 days
    • African American: 120.6 + 7.6 + 4.2 + 4.3 = 136.7 days
Race      
CAreference reference
AA7.3<.0013.3, 11.27.6.0042.5, 12.6
Age,y      
66-69referencereference-
70-74–2.0.01–3.6, –0.5–1.4.15–3.3, 0.5
75-79–4.2<.001–6.0, –2.5–2.3.05–4.5, 0.0
Year of diagnosis      
2004reference--reference--
20052.0.030.2, 3.9-0.9.47-3.2, 1.5
20066.7<.0014.9, 8.53.8.0011.5, 6.1
20078.0<.0016.2, 9.84.2<.0011.9, 6.5
Marital status      
Marriedreferencereference-
Not married/unknown3.1<.0011.7, 4.65.0<.0013.0, 6.9
Risk status      
Low-riskreferencereference--
Intermediate-risk-3.8<.001–5.4, –2.21.3.22-0.8, 3.3
High-risk-6.6<.001–8.3, –5.04.3<.0012.2, 6.4
NCI combined comorbidity score    
0referencereference--
>0–1.5.03–2.9, -0.1-0.2.79-2.0, 1.5
% Non-high school graduate in census tract    
0-25%reference--reference--
26-50%1.8.050.2, 2.61.7.15-0.6, 4.0
51-75%-0.1.88-2.0, 1.71.3.261.0, 3.7
76-100%0.7.46-1.2, 2.71.5.25-1.0, 3.9
Medicaid dual-eligibility     
Noreference--reference--
Yes3.8.0021.4, 6.32.7.12-0.7, 6.1
Population density      
Urbanreference--reference--
Rural1.8.07-0.2, 3.7-1.0.46-3.5, 1.6
Geographic region      
Northeastreference--reference-
South–14.8<.001–16.9, –12.6–23.3<.001-26.1, -20.6
Central–12.4<.001–14.6, –10.3–20.1<.001-22.8, -17.4
West–4.6<.001-6.4, -2.8–9.9<.001-12.1, -7.6
Any treatment      
EBRTreference-- 
RP–1.6.14–3.7, 0.5 
Brachytherapy–4.7<.001–6.8, –2.5 
HT–30.6<.001–32.4, –28.8 
Definitive Treatment      
EBRT-- reference 
RP-- –20.6<.001–22.8, –18.4
Brachytherapy-- –4.8<.001–6.8, –2.8
Race×Region      
CA, Westreference-reference-
AA, West0.8.80–5.4, 7.0–5.7.17–13.6, 2.4
CA, Centralreference--reference
AA, Central12.8<.0016.6, 19.17.0.09–1.1, 15.2
CA, Southreference--reference-
AA, South0.7.81–4.7, 6.0–4.4.23–11.4, 2.7
CA, Northeastreferencereference
AA, Northeastreferencereference

DISCUSSION

In this study using data from SEER-Medicare, AA men with low-, intermediate-, and high-risk prostate cancers experienced longer intervals from diagnosis to initiation of any or definitive treatment compared with Caucasian patients, even after adjusting for socioeconomic, clinical, and other demographic covariates. In multivariate analysis, this difference was a median of 7 days. There was regional variation with the disparity being largest in the Central SEER regions. Comparing treatment modalities, time-to-treatment was shortest for hormonal therapy and longest for external beam radiation therapy. This finding is likely attributable to the common use of neoadjuvant hormonal therapy with radiation, with hormonal therapy usually starting 1 to 2 months prior to initiation of radiation.

Few prior studies have examined racial disparities in delay of treatment for prostate cancer. In a study of 1532 patients who underwent prostatectomy for localized prostate cancer at 4 Veterans Affairs (VA) hospitals,[18] Bañez et al. found no difference in time from diagnosis to surgery between AA and Caucasian patients. However, the generalizability of these findings is limited by the single treatment modality and provision of care in the VA setting, which may not be applicable to findings in the general community. Because timely initiation of treatment after cancer diagnosis may be important to maximize long-term disease control and survival outcomes, especially for patients with aggressive cancers, an examination of racial differences in time-to-treatment may allow a better understanding of potentially modifiable causes of the mortality differences seen among AA and Caucasian patients with prostate cancer. This was the underlying motivation for this study. We were interested in potential treatment delays in patients in all risk categories, but especially high-risk cancer, which is the most aggressive form of prostate cancer.

Treatment delay has been associated with worse outcomes in many cancers. A review of 46 studies found an increased odds of local recurrence when initiating adjuvant radiotherapy more than 6 weeks after surgery for head and neck cancer (OR 2.89), and more than 8 weeks after surgery for breast cancer (OR 1.62).[19] In muscle-invasive bladder cancer, delay of greater than 12 weeks from diagnosis to cystectomy was associated with worse overall survival (HR of death 1.96).[20] Other studies have demonstrated similar adverse recurrence or survival outcomes associated with treatment delays in high-grade glioma[21] and lung cancer.[22]

Although prostate cancer is often indolent, several prior studies have identified an association between delay in surgical or radiation treatments and increased risk of biochemical failure. In a single-institution study of 1111 men with low-risk disease treated with prostatectomy, patients treated more than 6 months after diagnosis had a 12% rate of developing biochemical recurrence, compared to 5% in those treated within 6 months.[12] A study of 895 VA patients likewise found an increased risk of biochemical failure when surgery was performed more than 6 months from diagnosis.[11] Similar results were found with surgical delays exceeding 3 months in another study of 3324 men, particularly those with high-risk disease.[9] Delaying radiotherapy has also been associated with increased risk of biochemical recurrence in patients with high-risk disease.[10]

In the context of these prior studies, the clinical significance of our findings alone (median 7-day difference between AA and Caucasian patients) is unclear. However, by examining the interquartile ranges of time from diagnosis to treatment in AA and CA patients, we found that a proportion of patients experience a disparity of 20 days or more, including those with high-risk cancers. This difference could plausibly be of clinical importance. Furthermore, this study contributes to a growing body of knowledge surrounding racial disparities in prostate cancer throughout the disease process. AA men are less likely than Caucasians to undergo screening,[23, 24] are more likely to be diagnosed with advanced disease,[25, 26] have longer delays from diagnosis to treatment (demonstrated by this study), and are less likely to receive aggressive treatment.[6, 27, 28] All of these factors together can contribute to an increased rate of recurrence[29] and prostate cancer mortality[2] in AA compared with Caucasian prostate cancer patients.

There are several limitations of this study. As an analysis of SEER-Medicare data, this study did not include individuals under the age of 65, and therefore results may not be generalizable to younger populations. Nevertheless, our findings are relevant to a majority of prostate cancer patients, who are diagnosed at a median age of 67.[2] Furthermore, SEER-Medicare has limited data on person-level socioeconomic variables, which can affect access to care. In addition, despite our efforts to account for demographic, clinical, and socioeconomic covariates, unmeasured differences between AA and Caucasian patients may have contributed to our findings.

In summary, time from diagnosis to treatment initiation was 7 days longer for AA versus Caucasian patients with localized prostate cancer. This difference was seen across all risk groups. This study enhances our understanding of the racial disparities in prostate cancer care in all aspects of this disease process. To our knowledge, this is the first population-based study examining time-to-treatment in prostate cancer. Further research into the underlying causes of treatment delay among AA patients is needed and may lead to strategies for mitigating these disparities.

Ancillary