New EHR-Based Model PRIME Achieves 0.75 AUC in Predicting Pancreatic Cancer Risk Across 11 Million Patients

New EHR-Based Model PRIME Achieves 0.75 AUC in Predicting Pancreatic Cancer Risk Across 11 Million Patients

Researchers developed PRIME, a parsimonious risk stratification model for pancreatic ductal adenocarcinoma using electronic health record data. Validated across US health systems and the UK Biobank, PRIME retained 19 predictors and achieved strong discrimination, enabling targeted early detection in high-risk populations.
Behavioral Science-Integrated EHR Prompts Significantly Increase Deprescribing Rates for Older Adults: A Randomized Clinical Trial

Behavioral Science-Integrated EHR Prompts Significantly Increase Deprescribing Rates for Older Adults: A Randomized Clinical Trial

This randomized clinical trial demonstrates that EHR-based behavioral interventions, such as precommitment and boostering prompts, significantly improve deprescribing rates for potentially inappropriate medications in older adults, providing a scalable strategy to mitigate polypharmacy risks in primary care.
Multicomponent Nudges Significantly Boost Influenza Vaccination Rates in Primary Care: Insights from the BE IMMUNE Trial

Multicomponent Nudges Significantly Boost Influenza Vaccination Rates in Primary Care: Insights from the BE IMMUNE Trial

The BE IMMUNE randomized trial demonstrates that combining patient text reminders, clinician EHR pended orders, and peer comparisons increases flu vaccination by 5.1 percentage points among adults aged 50 and older, highlighting the efficacy of multilevel behavioral interventions in clinical settings.