Graded appraisals of the tests and interventions sold in preventive medicine. Each piece states what the literature supports, what it does not, and what would change our grade.
PATHFINDER 2 put Galleri through 35,878 people and produced the strongest evidence any multi-cancer blood test has. It also showed the test misses about six in ten cancers, and both facts are compatible with it being useful.
Bone density scanning carries a guideline recommendation and moderate net benefit in older women. Premium programmes sell it as a body composition readout, which is a different test with a different evidence base.
Direct-to-consumer gut testing has a reproducibility problem large enough to invalidate the product. Variation between providers analysing identical samples is as large as the variation between different people.
Cardiorespiratory fitness predicts all-cause mortality more powerfully than almost anything else a clinic can measure, with no observed upper limit of benefit. It is also cheap, unglamorous, and buried on page four of most programme brochures.
A coronary calcium score costs a fraction of a whole-body MRI, appears by name in the ACC/AHA cholesterol guideline, and reclassifies risk downward in about 40% of intermediate-risk people who score zero. In Germany it is also, as a pure screening test, not legally available.
The sensor is accurate, the data is real, and the evidence that wearing one changes anything for a metabolically healthy adult rests on seven small studies, only one of which measured an outcome.
Epigenetic clocks are the most technically reproducible biomarkers in consumer longevity testing and among the least biologically stable. Repeated-measures work shows the two properties are unrelated, and that the most trusted clocks are the worst offenders.
Screening MRI sits at the centre of nearly every premium health programme in Europe and North America. No randomised trial has shown it reduces mortality in asymptomatic adults, and the evidence on what it costs the people it does not help is unusually specific.