Title: 87: Randomized trial of behavioral change for cytomegalovirus risk reduction
Abstract: To estimate the impact of a brief prenatal behavioral intervention on risk behaviors for CMV infection and assess whether routine serostatus knowledge influences risk behaviors. Randomized clinical trial of women screened with CMV serology during prenatal care before 20 weeks'. Participants were apprised of their serostatus before randomization to either a brief behavioral intervention during their prenatal care visit or to standard care (a brochure) in a 2:1 ratio. The 5-minute intervention included an in-office video with hygiene teaching, a take-home calendar, and weekly text message reminders. Randomization was stratified by serostatus (positive or negative). Those with primary CMV infection were excluded. The primary outcome was change in behavioral compliance score at least 10 weeks after randomization. From 2013-2015, 195 women were randomized and 187 (96%) had outcome data available. Mean gestational age at screening and randomization was 9.4 and 13.6 weeks' respectively, 51.3% were seronegative. Follow-up visits occurred at a mean of 28 weeks'. The intervention and control groups were similar in site of prenatal care, maternal age, occupation, race, ethnicity, children under age 5 in the home, or use of daycare. Baseline behavioral compliance scores were higher than anticipated at baseline, 54.6 of possible 68 points (80%), yet increased in the intervention group compared to controls, mean difference after controlling for baseline score: 2.0 (95% confidence interval: 0.6, 3.5), p=0.007. Seronegative women were no more likely to change from baseline behavioral compliance score than were seropositive women. The domain with most absolute score change was in perception of the severity of infection (Table). Women in this randomized trial reported high baseline compliance with recommended CMV prevention measures. Despite this, a brief behavioral intervention delivered in the prenatal care setting was modestly effective at improving compliance. Knowledge of serostatus did not influence behavioral change.