Objective
To examine prostate-specific antigen (PSA) as a screening test in randomly selected primary care practices.
Design
Chart abstractions of a random selection of 552 men (age, 
50 years) and physician and practice-level surveys.
Setting
Fifty-eight randomly selected, nonteaching, nongovernmental, primary care practices in a 43-county area in Virginia.
Main Outcome Measures
Documented evidence of PSA screening for asymptomatic male patients and physician self-report on regular (annual or biannual) use of PSA screening.
Results
Of 496 asymptomatic men, 123 (25%) seen in a 1-year period had documentation of PSA screening. Sixty (50%) of 120 asymptomatic men with documentation of a health maintenance examination had a PSA screening, whereas only 63 (17%) of 376 men without health maintenance examination had a PSA screening (X2, P<.001), after adjusting for race, insurance status, and age. Comparison of PSA screening use before and after the November 1992 publication of the American Cancer Society guidelines on PSA screening revealed that PSA screening after the publication date was 26%, compared with 19% before the date (P=.045), adjusting for health maintenance examination and the length of time for which the patient was eligible for screening. Logistic regression on physician self-report of regular vs not regular PSA use found an odds ratio of 6.12 (95% confidence interval, 1.28 to 29.30) for influence of the guidelines and 0.96 (95% confidence interval, 0.93 to 0.99) for the proportion receiving Medicaid or uninsured in the practice.
Conclusions
Despite controversy over PSA as a screening test, PSA screening has now spread substantially in primary care practice.
