30-Second Takeaway
- Patient decision aids can improve risk perception and knowledge but may not increase informed prevention choices.
- Brief parenting interventions may reduce total abuse reports and improve child mental health over 2.5 years.
Latest - Week ending July 4, 2026
Five recent studies with direct implications for preventive medicine practice
Decision support improved knowledge but not informed chemoprevention choice in high-risk women
In a 31-site cluster randomized trial of 412 enrolled high-risk women, a patient decision aid plus provider tool did not increase informed chemoprevention choice at 6 months (35% vs 27%, p=0.20) or 12 months (37% vs 25%, p=0.05). At 6 months the intervention improved accurate breast cancer risk perception (31% vs 21%, p=0.03) and chemoprevention knowledge (33% vs 23%, p=0.04). Self-reported chemoprevention initiation at 12 months was similar between arms (50% intervention, 52% control). Interpretation: decision support yields modest cognitive benefits but did not translate into greater informed choice in this trial.
Families value recommendations and report high involvement in pediatric precision-medicine trials
In the PRISM precision-medicine trial, most parents and adolescent patients expected benefit (87%) and a treatment recommendation (68%) at enrollment. Seventy percent of parents received a recommendation but only about half recalled it afterwards. Parents reported high involvement (93/100) and high satisfaction (95/100) with treatment decisions, and receipt of a recommendation was not linked to regret. Interpretation: families derive hope and perceived value from recommendations, but recall is limited and expectation management is important.
Protocol: LLM-driven text coaching trial embedded in a large digital cardiovascular biobank
This protocol describes a prospective digital cohort aiming to enroll 15,000 adults with an embedded randomized crossover trial of LLM-generated versus generic text coaching to increase daily steps. Primary outcome is change in daily step count, with secondary outcomes including active minutes, calorie burn, fitness tests, and sensor-derived biomarkers linked to EHR outcomes. The design emphasizes cross-platform scalability, passive sensor data capture, and integration with HL7 FHIR for longitudinal outcomes. Interpretation: this is a scalable prevention model but currently a protocol—effectiveness and safety of LLM coaching remain to be shown.
References
Numbered in order of appearance. Click any reference to view details.
Additional Reads
Optional additional studies from this edition.