CHECK YOURSELF: A RANDOMIZED TRIAL OF ELECTRONIC SCREENING AND PERSONALIZED FEEDBACK FOR ADOLESCENT HEALTH RISK BEHAVIORS IN PRIMARY CARE

Purpose: Health risk behaviors increase during adolescence and are associated with significant morbidity. Screening and counseling are recommended as part of routine care but are infrequently performed. The goal of this study was to test the efficacy of an electronic multiple risk behavior screening and personalized motivational feedback tool, "Check Yourself", on the delivery of provider counseling during well visits and subsequent teen behaviors. Methods: We conducted a randomized controlled trial comparing electronic screening and personalized feedback with usual care at five pediatric clinics. Three hundred 13-18 years old participants who were scheduled for a well visit between February, 2015 and August, 2017 were recruited. Age and sex stratified randomization was conducted prior to baseline assessment. Of the original sample, 293 (98%) and 292 (97%) completed follow-up surveys at 1 day and 3 months, respectively. Intervention youth received the Check Yourself tool including screening questions and personalized feedback based on a HEADSS framework. The providers received a summary report of risk behaviors rated as high, moderate or low risk based on pre-determined criteria. Control youth received usual care. Two main outcomes, determined a priori, included receipt of counseling during the well-visit and change in a summary score of all risk behaviors from baseline to 3 months. Results: Among participating youth, 52% were female and the mean age was 14.5 (SD 1.42) years. 95% of the study sample had at least one moderate or high health risk behavior at baseline. The mean number of moderate-risk behaviors at baseline was 2.18 (SD 1.27, IQR 1-3) and high-risk behaviors was 0.68 (SD 0.95, IQR: 0-1). There were no significant differences in demographics or baseline rates of risk behaviors between control and intervention youth. After adjusting for age, sex, and random effect of clinic, intervention youth were 1.32 times more likely to receive counseling for each of their endorsed moderate and high risk behaviors than control youth (RR=1.32, 95% CI: 1.07,1.63). Intervention youth had a significantly greater reduction (b=-0.48, p=0.02) in their risk behavior scores at 3 months when compared to control youth, with sub-analyses showing that the intervention had a greater effect on high risk behaviors than on moderate risk behaviors. Conclusions: The Check Yourself tool increased provider counseling for health risk behaviors and was associated with a reduction