A primary care Web-based Intervention Modeling Experiment replicated behavior changes seen in earlier paper-based experiment.

OBJECTIVES Intervention Modeling Experiments (IMEs) are a way of developing and testing behavior change interventions before a trial. We aimed to test this methodology in a Web-based IME that replicated the trial component of an earlier, paper-based IME. STUDY DESIGN AND SETTING Three-arm, Web-based randomized evaluation of two interventions (persuasive communication and action plan) and a "no intervention" comparator. The interventions were designed to reduce the number of antibiotic prescriptions in the management of uncomplicated upper respiratory tract infection. General practitioners (GPs) were invited to complete an online questionnaire and eight clinical scenarios where an antibiotic might be considered. RESULTS One hundred twenty-nine GPs completed the questionnaire. GPs receiving the persuasive communication did not prescribe an antibiotic in 0.70 more scenarios (95% confidence interval [CI] = 0.17-1.24) than those in the control arm. For the action plan, GPs did not prescribe an antibiotic in 0.63 (95% CI = 0.11-1.15) more scenarios than those in the control arm. Unlike the earlier IME, behavioral intention was unaffected by the interventions; this may be due to a smaller sample size than intended. CONCLUSIONS A Web-based IME largely replicated the findings of an earlier paper-based study, providing some grounds for confidence in the IME methodology.