PD17-10: Prostate specific antigen screening on a nationwide level: featuring the contribution of race and life expectancy in decision making

Nicholas J. Corsi*, Deepansh Dalela, Austin Piontkowski, Ivan Rakic, Sami E. Majdalany, Shravan Morisetty, Detroit, MI, Taylor Malchow, Dayton, OH, Marcus Jamil, Detroit, MI, Akshay Sood, Houston, TX, Sohrab Arora, Craig Rogers, Detroit, MI, Mara Schonberg, Boston, MA, Firas Abdollah, Detroit, MI

Introduction: The mortality benefit of prostate specific antigen (PSA) screening is centered on patient’s life expectancy (LE). While African American (AA) men represent a high-risk group for prostate cancer, there's limited evidence regarding the relationship between LE and PSA screening in AA men.  The aim of our study was twofold: first, to assess temporal trends in PSA screening in AA men, stratified by limited (i.e., <15 years) vs. extended (i.e., =15 years) LE using a nationally representative dataset, and second, to study the determinants and relative contribution of LE on PSA screening utilization in AA men

Methods: Utilizing the National Health Institution Survey (NHIS) from 2000-2018, we identified AA men (using self-reported ethnicity) aged =40 years without a history of prostate cancer, who underwent PSA testing in the last 12 months as ‘a part of routine exam.' To assess LE, we used the previously validated Schonberg index (based on 11 risk factors), which was developed using the NHIS dataset linked with the National Death Index to track mortality. LE-stratified temporal trends in PSA screening were analyzed separately for AA men. Weighted multivariable analyses and dominance analyses (DA) were performed to elucidate the predictors of PSA screening and their relative contribution

Results: Within the NHIS-study years, 33,715 (weighted n=35,897,435; 57.3%) men were estimated to have extended LE, while 29,212 (weighted n=26,7630,619; 42.7%) were included in the limited LE. PSA screening declined significantly in AA men with limited LE between 2008-2018, while increasing for those with extended LE. However, LE itself was not an independent predictor of receipt of PSA screening in MVA (OR 1.04, p=0.6), and accounted for only ~0.3% of variability in screening on DA. The two most important factors predicting PSA screening were prior receipt of colonoscopy within last 10 years (OR 3.56, p<0.001; 59.8% relative contribution on DA) and visit to healthcare provider in the year prior (OR 2.20, p<0.001; 24.9% relative contribution), together accounting for >80% of variation in screening even after adjusting for LE. Similar results were seen in sensitivity analyses using LE <10 years as a cut-off for limited LE

Conclusions: Screening rates declined for AA men with limited LE, and increased for those with extended LE over the last decade. However, these trends were seemingly driven more by social determinants of health (i.e. prior participation in screening practices like colonoscopy and regular access to health care provider) than LE considerations itself

Source of Funding: N/A