Sensible Medicine This Fortnight in Medicine XXII
13 snips
Mar 24, 2026 A lively debate about GLP‑1 drugs and whether observational VA data can really link them to substance use outcomes. A deep dive into timing and scope of PCI for heart attacks, questioning complete revascularization versus treating only the culprit lesion. Discussion about trial design, confounding, and risks of extra procedures. Clinical caution and calls for randomized trials run through the conversation.
AI Snips
Chapters
Transcript
Episode notes
Avoid Changing Practice From This Observational GLP-1 Study
- Don't change clinical practice based on this observational GLP-1 study.
- If treating substance use disorder, prioritize established treatments and consider GLP-1s only within randomized trials or for approved indications.
VA Population And Treatment Selection Weaken Generalizability
- Generalizability and confounding limit the VA GLP-1 findings.
- Study population was veterans with diabetes, treatment choice by clinicians introduces baseline differences that likely explain small absolute effects.
Death Signal Larger Than Surrogates Suggests Selection Bias
- Stronger mortality reduction than intermediary outcomes suggests selection bias.
- John Mandrola notes death reduced more than hospitalizations, which is unlikely if drug truly reduced substance use and signals healthier patient selection.