Racial Inequities in the Quality of Surgical Care Among Medicare Beneficiaries with Localized Prostate Cancer

Abstract

Background:

US Black men are twice as likely to die from prostate cancer as men of other races. Lower quality care may contribute to this higher death rate.

Methods:

Sociodemographic and clinical data were obtained for men in SEER-Medicare diagnosed with clinically localized prostate cancer (cT1–4N0/xM0/x) and managed primarily by radical prostatectomy (2005–2015). Surgical volume was determined for facility and surgeon. Relationships between race, surgeon/facility volume, and characteristics of treating facility with survival (all-cause and cancer-specific) were assessed using multivariable Cox regression and competing risk analysis.

Results:

Black men represented 6.7% (n=2,123) of 31,478 cohort. They were younger at diagnosis, had longer time from diagnosis to surgery, lower socioeconomic status, higher PSA, and higher comorbid status compared with men of other races (p<0.001). They were less likely to receive care from a surgeon or facility in the top volume percentile (p<0.001); less likely to receive surgical care at an NCI-designated cancer center and more likely seen at a minority-serving hospital; and less likely to travel ≥50 miles for surgical care. On multivariable analysis stratified by surgical volume, Black men receiving care from a surgeon or facility with lower volumes demonstrated increased risk of prostate cancer mortality (HR1.61, 95%CI 1.01, 2.69) adjusting for age, clinical stage, PSA, and comorbidity index.

Conclusions:

Black Medicare beneficiaries with prostate cancer more commonly receive care from surgeons and facilities with lower volumes, likely affecting surgical quality and outcomes. Access to high-quality prostate cancer care may reduce racial inequities in disease outcomes, even among insured men.

Lay summary

• Black men are twice as likely to die of prostate cancer than other US men.

• Lower quality care may contribute to higher rates of prostate cancer death.

• We used surgical volume to evaluate the relationship between race and quality of care. Black Medicare beneficiaries with prostate cancer more commonly received care from surgeons and facilities with lower volumes, resulting in a higher risk of prostate cancer death and indicating scarce resources for care.

• Access to high-quality prostate cancer care eases disparities in disease outcomes. Patient-centered interventions that increase access to high-quality care for Black men with prostate cancer are needed.

Precis

Lack of access to high quality care for Black men with prostate cancer may contribute to their substantially higher rate of death when compared with men of other races in the US. Surgical volume—a surrogate for quality—and race are assessed in localized prostate cancer care.

INTRODUCTION

Black men have a twofold higher risk of prostate cancer death compared with men of other races, representing the widest racial inequity in cancer-related death in the US.¹ Various studies have demonstrated that Black and White men with localized prostate cancer have similar rates of prostate cancer-related death when utilization of standard of care definitive treatments are equitable, matched by grade and stage.^2,3 Clinically significant localized prostate cancer confers high survival rates when treated.^4–7 However, Black men are more likely to receive primary androgen deprivation therapy (ADT) or observation for localized prostate cancer compared with White men,^8,9 and among those who undergo surgery, time from diagnosis to treatment is longer and they may be more likely to see surgeons with lower volumes.^10,11

Inequities in prostate cancer care and outcomes reflect a complex relationship between structural determinants of equity, social determinants of health, and healthcare utilization.¹² These interconnected factors influence prostate cancer care and outcomes, and present an opportunity for modifying the wide inequity in prostate cancer death observed among Black men. Higher surgical volume has been associated with improved quality care and lower mortality, ^13,14 with radical prostatectomy volume serving as a reliable proxy for quality for perioperative complications,^15–17 patient-reported outcomes,^18,19 and mortality.²⁰ We hypothesized that Black men were less likely to be treated by surgeons noted for higher-volumes of prostatectomy. The aims of this study were (1) to assess the relationship between surgical volume and race, (2) to evaluate the association between race and treatment facility type; and (3) to evaluate the impact of surgical volume and race on prostate cancer mortality following treatment.

METHODS

This is a retrospective cohort study using the Surveillance, Epidemiology, and End Results cancer registry data linked to Medicare claims (SEER-Medicare) among patients with prostate cancer diagnosed from 2004 to 2015. Men in this cohort are Medicare beneficiaries, age >66 years. The 12-month period prior to diagnosis was necessary to determine comorbidity status. Patients with managed care coverage in the 12 months before the diagnosis date were excluded. Additional demographic and surgeon data were obtained from the Patient Entitlement and Diagnosis Summary File, which contains demographic information and Medicare entitlement and utilization data; the Medicare Surgeon Analysis and Review file, which contains diagnosis and procedure codes for each hospital admission; and the National Claims History (NCH) records, which contain surgeon billing information. Using the SEER registry data, men with localized prostate cancer (cT1-3N0/xM0/x) who underwent radical prostatectomy within 12-months of initial diagnosis were identified using the International Classification of Diseases, Ninth Revision (ICD-9), and Current Procedural Terminology (CPT) codes in contiguous years over the study period. Medicare claims data were available through 2016. Patients had to have claims for prostatectomy from both parts A and B to determine facility and physician information. The study was designated an exemption status by the Institutional Review Board at the University of Washington.

All available and relevant patient data were extracted from the registry; data included demographics, medical comorbidities, tumor characteristics, and geographic data. Race—as defined by the US Office of Management and Budget (i.e., White non-Hispanic, Hispanic, Black, other)—was the primary independent variable for this analysis. Additional covariates analyzed included age at diagnosis, marital status, National Cancer Institute (NCI) comorbidity score, tumor characteristics (i.e., Gleason score, tumor stage [TNM classification], and prostate-specific antigen [PSA] level at diagnosis), socioeconomic measures based on census tract (i.e., census-tract poverty levels, income, and education), urban/rural status, and SEER registry site.

Surgeon volume was defined by cross-referencing surgeon identifiers (i.e., encrypted unique physician identification number [UPIN]/national provider identifier [NPI]) with NCH claims data. The SEER/Medicare facility hospital provider number for a prostatectomy admission was linked to the hospital information file. Annual prostatectomy volume was determined for each surgeon and facility. Both individual surgeon and facility were then ordered by this volume count and categorized into percentiles for each year (<50%, 50–75%, 75.1–90%, 90.1–95%, 95.1–99%, 99.1–100%). These correlated with surgeon volumes of 1, 2–3, 4–7, 8–11, 12–20, and 21+ and facility volumes of 1–2, 3–7, 8–14, 15–22, 23–41, and 42+. These volumes are representative of Medicare claims only. To account for surgeons with higher volumes operating in lower-volume facilities and surgeons with lower volumes operating in high-volume facilities, a composite volume score was created. Surgeon and facility volumes were combined to create a high, medium, and low volume ‘provider’ category defined in the following manner: high (surgeon volume of 8+ with facility volume of 23+ or surgeon volume of 12+), medium (surgeon volume 1–7 with facility volume of 23+ or facility volume 4–11), or low (surgeon volume of 1–3 and facility volume of 1–7). Note that these volume counts are derived from percentile categories. Nearly 15% of claims have missing surgeon information, which we treated as its own category in our analysis to avoid inclusion bias. For our multivariable analyses, the dependent variable of prostatectomy volume was dichotomized to a high-volume provider yes/no which was defined as above and below the 95^th percentile.

We evaluated a wide range of provider characteristics that could serve as proxies for quality of care. These characteristics included distance travelled, travel outside of Hospital Referral Region or Hospital Service Area for prostatectomy, number of hospital beds, number of resident physicians, NCI cancer center designation status, and minority-serving facility status. Distance traveled for prostatectomy was examined as both a continuous and categorical (cut point of 50 miles) variable. A minority-serving hospital was defined as having over 33% of the patients admitted indicating Hispanic or Black race/ethnicity.

Prostate cancer death was determined from the SEER registry Patient Entitlement and Diagnosis Summary File. The follow-up for this data point extends to time of censorship from death or loss to follow-up through the ends of 2016. In addition to survival time, this data point categorizes vital status into alive at last follow-up, died from prostate cancer, and died from other causes.

Statistical Analysis

Descriptive statistics are presented for our cohort with testing between groups using the Pearson chi-square test for categorical variables and Wilcoxon rank sum test for continuous variables. Data are presented as proportions or medians with interquartile ratios. The primary aim of this analysis is to assess the relationship between race and the outcome of surgeon/facility volume of radical prostatectomy. The secondary aim of this analysis is to assess the relationship between race and other quality-of-care outcome metrics. For both aims, Pearson-chi square tests will demonstrate univariate associations between dependent outcome variables and race.

Multivariable logistic regression models determined risk estimates for association of race adjusting for age at diagnosis, socioeconomic status (SES), urban status, clinical stage, PSA level, NCI comorbidity index, and SEER region with volume measures. The secondary independent variables in this model were selected based on established associations observed with the primary independent variable of race and hypothesized associations with the dependent variable of treatment at a high-volume center. All variables were significant independent predictors of treatment at a high-volume center and were retained in the model. Clustering on facility identifier using random fixed-effect models showed a negligible effect compared with multivariable logistic regression, thus only base model risk estimates were reported.

The third aim was to explore the rate of all cause and prostate cancer death as a function of race and surgical volume. Cox proportional hazards models were constructed to assess disease-specific survival. Variables shown to be associated with mortality were considered as secondary adjustment variables. These included age at diagnosis, marital status, SES, urban status, clinical stage, Gleason grade, PSA, and the NCI comorbidity index. The final models retained age, marital status, SES, clinical stage, PSA levels and NCI comorbidity index, as these variables independently predicated both all-cause and cancer-specific mortality, in addition to appreciably altering the risk estimates of the primary independent variable. A tiered approach was taken in model construction. First, all-cause and cancer-specific survival models included both race and the composite volume measure, along with secondary adjustment covariates. Second, an interaction term between race and volume was considered. Lastly, the survival risk estimates of race were estimates across composite volume strata. The proportional hazards assumption was assessed with Schoenfeld residuals for each covariate in the analysis and was found not to be violated. Competing risk analysis was performed to evaluate disease-specific survival accounting for other-cause mortality. Statistical significance was defined as p < 0.05. All testing is 2-sided. Analyses were performed with R 3.4.0 and with SAS (version 9.4; SAS, Cary, NC).

RESULTS

Baseline demographic data are presented in Table 1 for the 31,478 men identified for this analysis. Black men represented 6.7% (n = 2,123) of included men. Black men were younger at diagnosis, had longer time from diagnosis to surgery, were of lower SES by income and education, and were more likely to be treated in the Southern US compared with men of other races. Black men demonstrated higher PSA, had higher comorbid status, and had lower-stage cancer at surgery.

Table 1.

Baseline demographic of men with cT1-3N0 prostate cancer treated by radical prostatectomy in SEER-Medicare (2004–2016)

	White (n = 25,745)	Black (n = 2,123)	Hispanic (n = 2,024)	Other (n = 1,586)
Continuous, median (IQR)
Age, years	69 (67, 72)	68 (67, 71)	69 (67, 72)	70 (68, 73)*
Percent of residents living below poverty	8.2 (4.6, 13.7)	18.2 (9.3, 29.4)*	15.0 (7.7, 23.8)*	7.8 (4.4, 13.6)
Census Tract, Median Income, $	63723 (46708, 88395)	42692 (30301, 61263)*	50718 (37386, 69816)*	72009 (53422, 95013)*
Census Tract, College Education, %	31.4 (18.4, 49.3)	17.3 (10.3, 30.5)*	17.9 (9.4, 31.1)*	33.2 (21.0, 48.9)*
Time to surgery, days	78 (56, 106)	84 (59, 115)*	79 (54, 113)	79 (58, 107)*
Follow-up time, months	71 (38, 104)	71 (36, 103)	71 (38, 103)	68 (36, 100)
Categorical, %
Married/partnered, yes Poverty level	80.7	64.1*	74.4*	80.2
0%-<5% poverty	31.5	11.8*	15.4*	33.4*
5% to <10% poverty	31.4	15.6	21.5	30.1
10% to <20% poverty	25.1	28.7	32.8	25.5
20% to 100% poverty Clinical Stage	11.1	43.6	30.0	10.8
T1	55.1	64.9*	56.0*	58.9*
T2	39.7	31.8	38.8	35.4
T3	2.7	1.7	3.0	3.3
Tx Gleason Scores	2.5	1.7	2.2	2.3
2–6	26.6	27.7	30.3	21.3*
7–10	73.0	72.1	69.1	78.3
PSA Categories 0.1–10.0	70.6	64.5*	63.4*	66.9*
10.1–20.0	11.5	13.3	15.6	15.5
>20.0	3.6	7.0	5.9	5.0
NCI Comorbidity Index Healthy	72.8	58.5*	61.9*	62.8*
Moderate	20.8	30.1	29.0	27.4
Severe	6.4	11.5	9.1	9.8
Region West	50.8	23.1*	84.4	83.9
Northeast	14.4	15.0	10.8	9.7
Midwest	12.8	18.3	1.1	3.6
South	22.0	43.5	3.8	2.8
Urban, yes Volume	71.4	80.5	84.2	89.2
All	25745	2123	2024	1586
Surgeon volume percentile 1st–50th	7.1	9.2	7.3	6.1
50th–74th	15.5	20.8	19.9	12.2
75th–89th	22.0	22.7	23.4	20.6
90th–94th	13.2	14.3	9.8	12.4
95th–98th	20.0	13.1	16.0	20.4
99th–100th	9.8	5.5	9.6	13.2
Surgeon not known	12.4	14.4	14.0	15.3
Facility volume percentile 1st–50th	8.2	11.6	13.3	7.4
50th–74th	17.4	23.5	23.3	13.5
75th–89th	26.0	32.8	25.1	18.3
90th–94th	16.2	12.8	14.6	10.6
95th–98th	20.0	11.5	10.7	29.6
99th–100th	12.3	7.8	13.0	20.7

^*p < 0.05 on univariable analysis compared with White men. SEER = Surveillance, Epidemiology, and End Results, IQR = interquartile ratio, NCI = National Cancer Institute

Surgical Volume and Race

The top 1^st and 5^th percentile of surgical volume by facility performed a median of 75 [50–100] and 32 [26–36] procedures per year, respectively. The top 1^st and 5^th percentile of surgical volume by surgeon was 29 [23–35] and 14 [12–18] procedures per year, respectively. On univariable analysis, Black men were less likely to be treated by a surgeon with higher volumes or at a high-volume facility (Figure 1, p<0.01). The proportion of Black men receiving care from a facility in the top percentile of facility volume was 7.8% compared with 12.3%, 13.0%, and 20.7% for White, Hispanic, and other races, respectively (p < 0.001; Supplemental Table 1). Black men (5.5%) were less likely to receive their care from a surgeon in the top percentile compared with White (9.8%), Hispanic (9.6%), and other (13.2%) races (p < 0.001). On multivariable analysis adjusted for age and other demographic characteristics and prostate cancer grade/stage, Black men were at lower odds of receiving care from a surgeon or facility with higher volumes (Table 2). In the adjusted model, men of lower SES, rural residence, and higher clinical stage and PSA level were less likely to receive treatment at a facility or surgeon with higher volumes (p< 0.05).

Figure 1 –

Distribution of Surgical Volume among Medicare Beneficiaries (2004–2016) by facility (panel a) and by surgeon (panel b)

Table 2.

Multivariable logistic regression evaluating the relationship between race, clinicodemographic data, and treatment by high volume surgeon/facility

	Treatment at Facility in top 5th percentile OR (95% CI)	Treatment by Surgeon in top 5th percentile OR (95% CI)	Treatment at Facility or by Provider in top 5th percentile OR (95% CI)
Race
White	ref	ref	ref
Black	0.68 (0.6–0.77)*	0.68 (0.6–0.76)*	0.66 (0.59–0.73)*
Other	1.53 (1.37–1.7)*	0.99 (0.88–1.11)*	1.32 (1.19–1.47)*
Hispanic	0.52 (0.46–0.58)*	0.69 (0.62–0.77)*	0.59 (0.53–0.65)*
Age	1.02 (1.01–1.03)*	1.02 (1.01–1.03)*	1.02 (1.01–1.02)*
Socioeconomic status
0%-<5% poverty	ref	ref	ref
5% to <10% poverty	0.9 (0.85–0.96)*	0.99 (0.93–1.06)	0.93 (0.87–0.99)*
10% to <20% poverty	0.77 (0.72–0.83)*	0.86 (0.80–0.93)*	0.8 (0.75–0.85)
20% to 100% poverty	0.7 (0.64–0.77)*	0.83 (0.75–0.91)*	0.74 (0.68–0.81)*
SES unknown	0.61 (0.42–0.89)*	0.8 (0.57–1.11)	0.66 (0.49–0.88)*
Rural, yes	0.55 (0.52–0.59)*	0.77 (0.72–0.82)*	0.69 (0.65–0.73)*
Clinical stage
cT1	ref	ref	ref
cT2	0.79 (0.75–0.83)*	0.82 (0.78–0.87)*	0.79 (0.75–0.83)*
cT3	0.87 (0.74–1.02)*	0.75 (0.63–0.88)*	0.78 (0.67–0.9)*
cTx	1.41 (1.2–1.67)*	0.61 (0.5–0.76)*	1.04 (0.88–1.21)
PSA categories
≤10.0	ref	ref	ref
10.1 – 20.0	0.92 (0.85–0.99)*	0.88 (0.81–0.95)*	0.9 (0.84–0.97)*
≥ 20.1	0.86 (0.75–0.97)*	0.78 (0.68–0.9)*	0.77 (0.68–0.87)*
missing	0.57 (0.53–0.62)*	0.64 (0.59–0.7)*	0.62 (0.58–0.67)*
NCI Comorbidity Index
Healthy	ref	ref	ref
Moderate	0.89 (0.84–0.95)*	0.95 (0.89–1.01)	0.93 (0.88–0.98)*
Severe	0.92 (0.83–1.02)	0.89 (0.8–0.99)*	0.94 (0.85–1.03)
SEER Region
Northeast	ref	ref	ref
Midwest	0.91 (0.82–1.01)	0.81 (0.72–0.9)	0.86 (0.78–0.95)
South	0.64 (0.58–0.71)*	0.88 (0.8–0.97)*	0.74 (0.68–0.81)*
West	1.78 (1.65–1.92)*	1.67 (1.54–1.81)*	1.6 (1.49–1.72)*

^*p < 0.05. OR = odds ratio; CI = confidence interval; PSA = prostate-specific antigen; NCI = National Cancer Institute; SEER = Surveillance, Epidemiology, and End Results

Treatment Facility Type and Race

Black men were less likely to travel to receive care, less likely to receive care at an NCI-designated comprehensive cancer center, and more likely to be treated at a minority-serving hospital (Table 3). White men were more likely to travel outside their Hospital Referral Region to receive their surgical care compared with Black and Hispanic men (p < 0.001) and were more likely to travel outside their Hospital Service Area than Black men (p < 0.001). Lastly, Black men (34.6%) were more likely to receive their surgical care at facilities with 601 or more hospital beds compared with White (30.1%) men, Hispanic (20.6%) men, and men of other (22.2%) races.

Table 3.

Treatment facility data by race among men with cT1-3N0 prostate cancer treated by radical prostatectomy in SEER-Medicare (2004–2016)

	White (n = 25,745)	Black (n = 2,123)	Hispanic (n = 2,024)	Other (n = 1,586)
Categorical, %
Travel Distance > 50 miles, yes	19.9	8.4*	13.6*	15.3*
Hospital Referral Range Change, yes	31.5	21.2*	21.3*	26.1*
Health Services Area Change, yes	61.5	47.7	60.3*	59.5
Number of Hospital Beds
< 150; 10th percentile	8.5	10.7*	10.1*	8.6*
151–400; 11–50th%	36.7	31.9*	44.5*	32.9*
401–600; 51–75th%	24.8	22.7*	24.8*	36.3*
601+; 76th+%	30.1	34.6*	20.6*	22.2*
Number of Resident Physicians
none	35.1	30.1*	42.5*	30.1*
1–99; 1–49%	31.9	30.2*	33.8*	39.2*
100–499; 50–89%	22.4	30.4*	18.0*	21.9*
500+; 90 percentile	10.6	9.4*	5.6*	8.7*
NCI Comprehensive Cancer Center, yes	15.5	11.1*	11.7*	21.0*
Minority Serving Hospital, yes	3.3	24.7*	5.8*	3.1

^*p < 0.05 on univariable analysis compared to White men. Seer = Surveillance, Epidemiology, and End Results; NCI = National Cancer Institute

Race, Surgical Volume, and Mortality

On univariable analysis, Black race (HR 1.40, 95% CI 1.25, 1.56), treatment at low- (HR 1.29, 95% CI 1.17, 1.43) and medium- (HR 1.19, 95% CI 1.10, 1.30) volume provider/facility were associated with an increased risk of death from any cause. On multivariable analysis adjusting for surgeon/facility volume, Black race remained associated with an increased risk of all-cause mortality (HR 1.18, 95% CI 1.05, 1.32, Figure 2). On adjusted analysis stratified by volume, there was no association between Black race and death from any cause (Figure 2). On univariate analysis, Black race and treatment by a low-volume provider/facility was associated with an increased risk of death from prostate cancer. On multivariable analysis stratified by volume, Black men treated at low-volume facilities demonstrated an increased risk of death from prostate cancer (HR 1.61, 95% CI 1.01, 2.69), adjusting for age, clinical stage, PSA, and comorbidity index. This association was not observed among Black men treated by surgeons with medium or higher volumes or facilities in separate volume-specific models.

Figure 2 –

Adjusted Hazard Ratios and Confidence Intervals of Mortality (panel a) and Prostate Cancer Specific

DISCUSSION

Black men are less likely to receive definitive treatment and much more likely to die when diagnosed with localized prostate cancer, findings that highlight the complex relationship between structural determinants of equity, social determinants of health, and healthcare delivery.¹² Although inequities in both treatment and outcomes have been well described in the prostate cancer literature, few studies have disentangled sources of these inequities. In this analysis, we evaluated the relationship between race and quality of prostate cancer care using surgical volume as a surrogate for quality. Black men in this population-based cohort of Medicare beneficiaries had 34% lower odds of receiving surgical care from a surgeon or facility with higher volumes—defined as the top 5^th percentile—even when adjusted for a wide range of clinical and demographic variables. Receipt of care by surgeons and facilities with lower volumes was associated with a 61% increased risk of death from prostate cancer among Black men.

Black men were less likely to receive treatment from a surgeon or facility with higher volumes in our analysis. In this cohort of Medicare beneficiaries, 33% and 20% of Black men received a prostatectomy by surgeon or facility in the top 10^th percentile of volume, respectively. In comparison, 43% and 32% of White men in the cohort received a prostatectomy by a surgeon or facility in the top 10^th percentile of volume, respectively. The difference in surgical care between Black men and White men persisted on multivariable analysis. These findings are consistent with those from a SEER-Medicare analysis of men treated from 1991–2002, which found Black men were more likely to receive treatment from a surgeon with lower volumes, regardless of hospital volume status.¹¹ Various studies have demonstrated that radical prostatectomy volume is a strong surrogate for quality as measured by perioperative complications,^15–17 patient-reported outcomes,^18,19 and mortality.²⁰ Surgeons and facilities with lower volumes have been associated with lower rates of board certification and less access to cancer-specific clinical resources, respectively.¹¹ Lack of access to care from surgeons and facilities characterized by higher quality care reflects the impact of structural factors (e.g., structural racism, economic and public policy) and social determinants of health (e.g., lack of insurance) on care delivery.¹² Recognizing that these factors drive this inequity is important in developing interventions to create parity in the delivery of high quality prostate cancer care.

Black men were also less likely to receive care at an NCI-designated comprehensive cancer center, and they were more likely to receive their care at minority-serving hospitals—which more commonly are safety net health facilities/systems. Treatment receipt at minority-serving hospitals among Black men has also been described in a study evaluating definitive therapy utilization among men treated in the national cancer database.¹⁰ However, our findings of lower treatment rates among NCI-designated cancer centers contradict a SEER-Medicare analysis from 1991–2002 that found that Black men were more likely to receive care at NCI-designated cancer centers. However, that study included seven SEER registry sites (Atlanta, Connecticut, Detroit, San Francisco, Seattle, San Jose, and Los Angeles) and represents an older cohort.¹¹ Black men in our cohort were more likely to be treated at a facility with 400 or more beds. Black Medicare beneficiaries in our cohort were also less likely to travel to receive a radical prostatectomy. Where patients receive their care again reflects a complex relationship between structural and social factors that impact healthcare utilization.¹² Unfortunately, our dataset lacks the granularity to understand how travel to receive care, facility designation/status and size, and utilization of treatment outside a hospital referral region reflects patient preferences, access to care and second opinions, and patient access to choice for their treatment. These are affected by social status, and they influence access to high quality care among Black men and other minoritized populations.

Various studies have demonstrated that Black men with localized prostate cancer demonstrate mortality outcomes similar to those of White men when receiving definitive treatment.^2,3 However, in the SEER registry, Black men receiving treatment demonstrate higher likelihood of prostate cancer death despite adjusting for cancer risk and socioeconomic status. This differs from data from the Veterans Affairs Health System and pooled clinical trial data, both of which demonstrate parity in survival between Black and White men.³ One explanation for this observation is that the quality of care that Black and White men receive is more similar in clinical trials and the Veterans Affairs Health System compared with men treated outside these settings. We found that Black men treated in low-volume facilities had an increased risk of death from prostate cancer. Low volume for surgeons and facilities is likely a surrogate for quality of care, access to care, and adverse structural and social factors among Black men. It is challenging to assess how these factors impact critical components of high-quality care such as post-treatment PSA surveillance and timely utilization of salvage therapies in SEER-Medicare.

This study is limited by the utilization of a cohort of Medicare beneficiaries, which limits the cohort to an insured population of men over the age of 65 years. The utilization of Medicare data also means that volumes for both surgeon and facility reflect the number of procedures for Medicare beneficiaries. For this reason, we chose to evaluate both surgeon and facility volume, given that a particular surgeon may be high volume if they operated on younger men. The utilization of Medicare claims data has an inherent selection bias, which must be considered in the interpretation and generalization of these results. Second, our survival analyses are limited by lack of longer-term follow-up, which is significant given the long natural history of prostate cancer and the high rate of successful treatment in the localized setting. Lastly, there are many structural, social, healthcare, and patient-specific factors that impact treatment decisions that cannot be accounted for in a registry-based analysis. Despite these limitations, our analysis compellingly suggests that Black men generally have less access to surgeons who may offer higher quality care.

CONCLUSION

In conclusion, Black Medicare beneficiaries demonstrated clear patterns of surgical care with a higher likelihood of undergoing surgery from surgeons and facilities with lower volumes. Black men were less likely to travel to seek care, and they were less likely to receive care from an NCI-designated comprehensive cancer center. Receipt of care from a surgeon or facility with lower volumes conferred a higher risk of prostate cancer death for Black men, likely reflecting the lack of resources to support care for Black men across their entire continuum of prostate cancer care. Future studies should evaluate patient perspectives and patient-centered interventions that increase both access and utilization of high-quality care among Black men with localized prostate cancer.