Detection of Prostate Cancer Screening Disparities Using Geospatial Analysis of Laboratory Results

Abstract

Abstract Black men experience higher prevalence and mortality from prostate cancer (PCa) but are less likely to undergo PCa screening with prostate-specific antigen (PSA). The reasons for this disparity are multifactorial and include lack of access, PCa education, and mistrust in the healthcare system. Identifying geographic regions with low PSA screening and high positivity rates could potentially help inform targeted interventions. This study aimed to examine racial, socioeconomic, and geospatial disparities in PSA screening and positivity rates in the Siteman Cancer Center (St. Louis, MO) catchment region. The results of all PSA screening tests ordered in the Barnes-Jewish Healthcare Corporation (St. Louis, MO) and performed from 9/2019-12/2022 were retrieved from the laboratory information system with patient demographics. Patient addresses were geocoded using ArcGIS Pro (ESRI, Redlands, CA) and assigned to a census tract of residence. The PSA screening-eligible population was defined as the number of persons ≥50 years of age based on U.S. Census data. Social Vulnerability Index (SVI) values were retrieved from the Centers for Disease Control. The PSA screening rate was calculated by census tract as the total number of PSA tests performed/PSA screening-eligible population. The PSA positivity rate was calculated by census tract as the number of PSA positive tests/total PSA tests. All census tracts with fewer than 10 PSA tests performed were omitted from PSA positivity analysis. Hotspot analysis was performed by calculating Local Moran’s I statistic per census tract and comparing with 999 randomly-generated maps. A total of 90,994 screening tests for 60,005 patients (48,024 White [80%], 10,425 Black [17%], and 1,556 Other [3%]) were performed for patients in the catchment region. Black patients were 78% more likely to receive a positive PSA screen than White patients (p<0.001). Of those screening positive, Black patients had a significantly higher median[IQR] (7.94[6.2-13.7 ng/mL]) PSA result compared to White patients (7.4[6.2-10.1 ng/mL]) (p<0.001) and a median[IQR] age of 67[61-74 years] vs. 69[63-76 years] (p<0.001). In a multivariate logistic regression using age, race, and census tract of residence SVI to predict PSA positivity, only census tract SVI was not statistically significant. A model using only age and race demonstrated that every 10 years of age and Black race were associated with 11% and 7% increased odds of receiving a PSA positive test, respectively (p<0.001). 75 and 191 of 789 census tracts in the catchment region were identified as PSA positivity hotspots and PSA screening cold spots, respectively. In conclusion, this study confirms previous reports of higher PSA values among Black patients and suggests that geospatial analysis of PSA testing can identify geographic regions with low testing and high positivity rates.