What fixes PSA screening guidelines gone wrong?

Abstract

TPS159 Background: Guidelines provide guidance but patient (PT) factors alter recommendations. We studied PT factors that alter the correct recommendation (CR) and physician (MD) advice in real-time cancer screening discussions in a 2x2 cluster-randomized trial of MD educational supports. Methods: We enrolled an age-(30-89 years) and sex-stratified sample of 18 patients (PTs) who underwent a physical examination at 2 urban hospitals by each of 12 primary care MDs. We assigned screening guideline formatting (color-coding [CC]) and academic detailing (AD) to MDs randomly. Immediate post-encounter surveys recorded PT and MD recall of screening discussions. Because PSA screening, the focus here, is controversial, we defined correct recommendations (CR) only when against screening, for men too old ( > 75 yrs) or too young (African American [AA] < 40 or non-AA < 50). For other men, either recommendation was considered correct. MD-reported serious comorbidity or PT choice against screening were also valid reasons against. Results: Of the first 174 participants, 92 were male. They were diverse (69% white) and highly educated (87% attended, 61% graduated college). Using age and race alone, 49% should not be screened, but increased to 65% by adding comorbidity (1%) and PT choice (15%). By these criteria, MDs advised correctly in 91% of encounters. In univariate analysis, CC (p = 0.08) but not AD, PT education, shared decision making and age tended to increase CR. Conclusions: In a highly educated, diverse patient population, including comorbidity and PT choice increased CR against screening to 16% more men. An educational support (CC) tended to increase CR. General guidelines can produce incorrect recommendations. Simple supports may improve them. Clinical trial information: NCT02430948.