Background
Over-testing and overtreatment in older adults may lead to harm. Clinician decision support redirecting attention to possible harms and engaging social and reputational concerns might reduce overuse.
Methods
We performed an 18-month pragmatic cluster-randomized trial in 60 primary care practices with 371 clinicians, comparing behavioral-science-informed point-of-care clinical decision support tools plus clinician education to education alone. Co-primary outcomes were: (1) prostate-specific antigen (PSA) screening in men 76 and older, (2) urine testing for non-specific reasons in women 65 and older, and (3) overtreatment of diabetes in patients 75 and older.
Results
At randomization, mean clinic rates were 24.9% for PSA screening (n = 8,275 eligible men), 23.9% for urine testing for non-specific reasons (n = 11,051 eligible women), and 16.8% for diabetes overtreatment (n = 5,553 eligible patients). After 18 months of intervention, the adjusted difference-in-differences (aDID) for two measures in the intervention vs. education only groups were lower: -8.6% (Bonferroni-corrected 95% CI, -10.1 to -7.1) for PSA screening, and -5.5% (-7.1 to -3.3) for urine testing. For diabetes overtreatment the difference was -1.4% (-2.9 to 0.1). Safety measures did not show increased emergency care related to urinary tract infections or hyperglycemia. Subsequent HbA1C >9.0% was more common among previously overtreated diabetes patients in the intervention group.
Conclusions
Decision support informed by behavioral science designed to increase clinicians’ attention to possible harms, social norms, and reputational concerns produced moderate reductions in diagnostic testing overuse in older adults.