What this covers

Philosopher Jacob Stegenga of the University of Cambridge makes a case for medical nihilism in conversation with Russ Roberts on EconTalk. The discussion centers on why we should hold low confidence in most modern medical interventions: the evidence base systematically inflates benefits while downplaying harms, and the biological reality—most diseases have multiple interlocking causes, most drugs bind to numerous receptors with cascading effects—makes magic bullet specificity rare. Stegenga draws on his book to walk through how the regulatory and publication apparatus creates this distortion, then turns to concrete cases (rosiglitazone, ACL surgery, flibanserin) where the gap between marketed promise and actual evidence has proven stark.

The conversation spans the mechanics of bias at multiple levels: how randomized control trial populations systematically exclude the elderly and polypharmacy patients—precisely those who bear the real harms—so trials underestimate dangers when drugs reach actual patients; how publication bias and outcome switching allow researchers to hunt for benefits while publishing a skewed subset; and how the FDA's approval threshold (two positive trials of any effect size, no specified magnitude) is permissive enough to let through drugs with negative benefit-harm profiles. Stegenga also argues there is an inverse correlation of trial quality and effect size—as methodology improves over time, measured effectiveness shrinks—suggesting future evidence will reveal even smaller gains and larger harms. The discussion acknowledges disagreement on whether stricter FDA standards would cost lives, whether personalized medicine might dissolve the complexity problem, and which diseases (like diseases of deficiency versus diseases of infection) remain tractable for intervention.

Sharpest takeaway

Stegenga argues for 'medical nihilism'—that we should have low confidence in most modern medical interventions because the evidence base systematically overestimates benefits and underestimates harms, and because most diseases and drug mechanisms are too complex for 'magic bullet' specificity.

  • Better-quality trials consistently reveal smaller effect sizes, while harm detection remains weak so harms are underestimated
  • Most diseases have radically complex causal bases and most drugs bind to multiple receptors, making magic-bullet specificity rare
  • The FDA's evidential standard (two positive RCTs of any effect size) is too lenient and is distorted by financial conflicts of interest

The argument · threads4 threads · 32 claims
0.85

Natural recovery over time makes it nearly impossible to determine whether treatments caused improvement.

1 pointcentrality 5/5
  • Many things get better by themselves with the passage of time, which makes it very hard to evaluate cures—people who report being cured or helped by a procedure may have improved anyway.

    “many things get better by themselves with the passage of time is very challenging for most of us to remember”

0.71

Professional bias, not just financial incentive, drives unnecessary medical recommendations.

1 pointcentrality 4/5
  • When you have a hammer everything looks like a nail; it is shocking how often surgery is recommended by surgeons, so non-financial professional biases can be as important as pharmaceutical financial incentives in distorting medical decisions.

    “it's shocking to me how often surgery is recommended by surgeons. Strangely enough.”

0.57

Third-party payment reduces patients' financial incentive to scrutinize drug use and safety.

1 pointcentrality 3/5
  • Because many drugs are not paid for directly by patients—a phenomenon occurring worldwide, not just in the US—patients have weaker financial incentive to be careful about taking them, even though they retain the incentive to avoid death and side effects.

    “many, many of these drugs are not paid for by the patients. Their incentive to be careful about taking these drugs certainly is there, because they don't want to die... But there's financial incentive, often, for not paying for them.”

0.44

Historical medical interventions often failed and harmed patients despite widespread belief in their efficacy.

1 pointcentrality 2/5
  • Through most of history, cures and interventions such as ingesting mercury and bloodletting did not work, did not improve patients, and were often dangerous and harmful on net.

    “Many of the cures and interventions of the past--ingesting mercury, bloodletting, and other things--didn't work; didn't improve the patient; in fact often were dangerous and harmful on net.”