Racial and Ethnic Disparities in the Use of Prostate Magnetic Resonance Imaging Following an Elevated Prostate-Specific Antigen Test

Key Points Question Are there racial and ethnic disparities in patients undergoing prostate magnetic resonance imaging after receiving an elevated prostate-specific antigen test result? Findings In this cohort study of 794 809 commercially insured men from 2011 to 2017 undergoing a prostate-specific antigen test, Black and Hispanic patients were significantly less likely than White patients to undergo prostate magnetic resonance imaging after receiving an elevated test score. Meaning These results suggest that racial and ethnic disparities exist in the use of subsequent prostate magnetic resonance imaging, highlighting a need to better understand and mitigate physician decision-making biases and other potential sources of the disparities in the identification and management of prostate cancer.


Introduction
In the US, racial and ethnic disparities have been identified in multiple aspects of prostate cancer diagnosis and treatment. [1][2][3][4] Specifically, Black patients are less likely to undergo appropriate prostate-specific antigen (PSA) screening, 5 less likely to undergo intensive follow-up while on close monitoring for low-risk prostate cancer (ie, active surveillance), 6,7 and less likely to undergo PSA surveillance after radical prostatectomy. 8 Such variation in care contributes to increased mortality in Black patients diagnosed with prostate cancer. 4,9 Accurate prostate cancer diagnoses can help reduce disparities in health outcomes.
Traditionally, men are screened for prostate cancer by PSA, and those with elevated PSA results then undergo prostate biopsy. 10 However, systematic nontargeted biopsies both miss clinically meaningful prostate cancer and overdiagnose insignificant prostate cancer. 11,12 Recently, prostate magnetic resonance imaging (MRI) has emerged as a means to evaluate men with an elevated PSA before possible prostate biopsy [13][14][15][16] and its use is rapidly increasing nationally. 17 Prostate MRI may identify suspicious regions in the prostate to target for subsequent biopsy, significantly increasing the likelihood of identifying clinically meaningful prostate cancer, 18 and prostate cancer detection by MRI-targeted biopsy is not significantly different between Black vs White patients. 19 Additionally, a prostate MRI may obviate the need for biopsy, thus decreasing the overdiagnosis of insignificant prostate cancer. 20 The American Urological Society (AUA) and National Comprehensive Cancer Network have both recognized the important role of prostate MRI for prostate cancer diagnosis. 21,22 However, consensus is lacking regarding the use of MRI in the initial detection of prostate cancer, and accepted guidelines are not yet available. This lack of standardization likely fosters both variation and disparities in prostate MRI utilization. A 2018 study, 23 for example, showed variation in the utilization of prostate MRI among Medicare beneficiaries. For this study, we used a large commercial and Medicare Advantage health insurance claims database combined with data on PSA test results to examine racial and ethnic disparities in the use of prostate MRI following receipt of elevated PSA values.

Methods
This cohort study used deidentified data from an administrative database containing health insurance claims for members of large commercial and Medicare Advantage health plans (Optum).
Claims data covered a geographically diverse population, including all 50 states, and additionally contained laboratory test results for approximately 30% of the patients in our sample. The Georgia Institute of Technology's institutional review board deemed this retrospective study of deidentified administrative claims data as not constituting human subjects research and thus not subject to review. The data analysis was performed in January 2021, and the study followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for reporting cohort studies.
Our study cohort was constructed by first restricting plan members to male enrollees between 2011 and 2017. In May 2013, the AUA published an updated guideline that generally recommends starting PSA screening at age 55. The AUA additionally recommends an individualized approach to PSA screening after age 40 years for higher risk men, such as Black patients, who have been reported to be diagnosed with prostate cancer at younger ages than patients of other races or ethnicities, and recommends against routine PSA screening in men younger than 40 years. 24 29 We combined responses of unknown race and ethnicity and missing responses into a single other category. We estimated heteroskedasticityrobust standard errors that were clustered at the state level. Additional regression results were estimated separately for each age group and PSA category.
In addition, Cox proportional hazard models were used to explore associations between PSA test results and the number of days until subsequent prostate MRI. For these models, the outcome variable was the number of days between the PSA test date and a subsequent prostate MRI date (bounded to 180 days), and the models included the same patient-level control variables described above. Cox regressions were also separately estimated for each age group and PSA threshold.
Cohort construction was performed using SAS version 9.4 (SAS Institute). All descriptive and regression analyses were performed using Stata 16 (StataCorp). Statistical tests of significance were assessed with 2-sided tests at the P < .05 significance level.

Results
Between 2011 and 2017, we identified 1 563 534 PSA tests that met our study inclusion criteria. PSA laboratory results were available for 794 809 (50.8%) unique participants of these test events. The Of all available PSA results, 51 500 (6.5%) were above 4 ng/mL and 1524 (3.0%) of these patients underwent a prostate MRI in the subsequent 180 days. The ORs for separate regression analyses for each age group and race and ethnicity are displayed in Figure 2. Black patients between 40 and 54 years of age and with a PSA result above 4 ng/mL were 39.8% less likely (OR, 0.60; 95% CI, 0.38-0.95) to undergo prostate MRI than White  decision-making biases and other potential sources of these disparities, such as physician characteristics or biases in the health care system. 30,31 Interestingly, the racial and ethnic disparities observed were insignificant among patients over 75 years of age-a population for which the US Preventive Services Task Force recommends against screening for prostate cancer. 32 This indicates the important role that clearly defined guidelines can    ng/mL because PSA testing would be used more selectively-for example, only on patients with higher risk (family history or abnormal rectal exam).

JAMA Network Open | Health Policy
In addition to a lack of clear guidelines on the use of prostate MRI following a PSA test, both unconscious and conscious biases may play a role in this health care inequity. 34 For example, prior research has shown that physicians are less likely to discuss treatment options and potential side effects with Black vs White patients. 35 Further research is needed to assess the role of these decisionmaking biases among physicians relative to other potential sources in the health care system for the observed racial and ethnic disparities in the use of prostate MRI, as well as examining whether these disparities extend to the use of prostate biopsy.

Limitations
This study had several limitations. Although the analyzed data set covers all 50 states, the data came from a single source, which may limit the generalizability of our results to other insurance

Conclusions
Significant racial and ethnic disparities exist in the use of prostate MRI following elevated PSA test results, and these disparities widen with higher PSA results. These disparities mirror those for other aspects of prostate cancer care 4 and may contribute to known differences among racial/ethnic populations in prostate cancer outcomes. 36