The prescribing of inappropriate medications to older adults is extremely common in the United States. Older adults are twice as likely as younger patients to be hospitalized because of drug-related problems. Electronic health record (EHR)-based tools are commonly used but have had limited application in facilitating deprescribing in older adults. The Novel Uses of adaptive Designs to Guide provider Engagement in Electronic Health Records (NUDGE-EHR) study was a pragmatic randomized controlled trial conducted at Atrius Health using an adaptive study design. This study tested whether designing EHR tools using behavioral science principles such as default bias, timing of tools, simplification, and pre-commitment, reduced inappropriate prescribing and clinical outcomes in older adults. Incorporating most of these behavioral principles into EHR alerts did not significantly reduce prescribing of benzodiazepines and sedative hypnotics. However, those timed to trigger specifically during patient encounters or incorporating pre-commitment (a strategy that encouraged PCPs to first review risks with their patient) significantly increased deprescribing.