Potential selection bias in a community-based study of PSA levels in African-American men

Abstract

The first phase of the Flint Men's Health Study (FMHS) in Michigan was a community-based epidemiologic study of prostate cancer and benign prostatic hyperplasia (BPH) in African-Americans aged 40 to 79 years. An objective of the FMHS was to determine age-specified prostate specific antigen (PSA) reference ranges in a random population sample of African-American men without clinically evident prostate cancer. The FMHS study protocol included an initial in-home epidemiologic interview followed by PSA testing and a urologic examination of eligible subjects. Since the participation rate in the PSA phase of the study was under 60%, it was important to determine whether selectivity in participation biased the FMHS results for age-specific PSA distributions. Logistic regression analyses were used to investigate selectivity in the sample of subjects who participated in the PSA testing and urologic examination. Younger men, with current urologic symptoms, and with a family history of prostate cancer were more likely to participate in the PSA testing and urologic examination. Linear regression analysis indicated that greater participation by African-American men without clinically evident prostate cancer but with obstructive or irritative lower urinary tract symptoms or a family history of prostate cancer did not bias the estimated age-specific reference ranges for total PSA concentrations and free-to-total PSA ratios.