The scans and tests that promise more than they deliver

The edition in brief

A meta-analysis of 18,043 people who had whole-body MRI screening without an evidence-based indication found 78% had some finding reported and about 15% had an actionable non-malignant finding — most of whom underwent further tests and one in five a procedure — while cancer was found in 1.4% and the clinical value of any of it remains unknown. For the clinician asked about full-body health scans, the honest answer is that they mostly generate work, not benefit. A matched cohort of 100,140 people found cannabis use disorder associated with higher all-cause mortality (hazard ratio 1.52), driven by injury, drug overdose (2.59), accidents and cancer rather than cardiovascular or respiratory death — an association that marks risk rather than proving cannabis caused these deaths. Among food-insecure older adults discharged home, a meal-delivery pilot was associated with lower 14-day (adjusted odds ratio 0.31) and 21-day readmissions, attenuating by 30 days — observational, but a reminder to address food insecurity at discharge. A clinic pearl: reconcile the medication list against the admitting problem on every admission, because the error that harms is usually one already on the list. Finally, a 1.41-million-veteran comparison found that, against fee-for-service Medicare, VA primary care tested less for prostate cancer in men over 80 (incidence rate ratio 0.70) and those with under five years' life expectancy, while testing slightly more in men aged 65–69 — aligning testing with who stands to benefit. The thread across the edition: match the intensity of testing and screening to the patient in front of you.