Family Medicine

LLM-assisted EMR in Kenyan primary care was safe but did not lower 14-day treatment failure

In a pragmatic cluster-randomized trial across 16 Kenyan primary care clinics, LLM-assisted EMR did not significantly reduce 14-day treatment failure versus EMR alone. Treatment failure occurred in 2.2% (102/4,693) with LLM support and 2.0% (94/4,654) in controls; adjusted OR 0.77 (95% CI 0.55–1.08; P=0.13). Independent review found no safety signal attributable to the intervention. Any clinical benefit, if present, is likely modest and not proven for short-term treatment failure in this setting. 1

Families report high involvement and satisfaction when precision trials provide treatment recommendations

Among parents and adolescents in a precision medicine trial for poor-prognosis childhood cancer, most expected and valued treatment recommendations. Seventy percent of parents received a recommendation, though only half later recalled it; recommendations provided hope and perceived options. Parents reported high involvement (93/100) and satisfaction (95/100) with treatment decisions. Receiving a recommendation was not associated with increased regret about trial participation. 2

Modified Zelen (delayed-consent) design was acceptable to most CAPS-THA participants

In the CAPS-THA cohort (n=235 respondents), 76% agreed the delayed-consent design was justified after disclosure. Ninety-one percent judged the study ethical and 94% reported unchanged or increased trust in research after debriefing. Seventy-one percent expressed strong willingness to join similar studies in future, supporting acceptability with robust debriefing. Findings endorse careful use of modified Zelen designs when preference effects threaten trial validity and post-disclosure choice is meaningful. 3