Scott Alexander critiques a study claiming to have found biomarkers for suicide risk, arguing that the results are overstated and such a test would have limited clinical utility even if accurate.
Longer summary
Scott Alexander critiques a study claiming to have found biomarkers predicting suicide risk with 92% accuracy. He points out that most of the predictive power likely comes from clinical instruments rather than biomarkers, and that the biomarkers alone were insignificant in 7 out of 8 tests. Even if the 92% accuracy claim were true, Scott argues it wouldn't be very useful in practice due to the low base rate of suicide attempts. He concludes that while research on biological factors in suicide is valuable, a blood test for suicide is unlikely to ever be clinically useful due to the complex nature of suicide involving both biological and environmental factors.
Shorter summary