A Newsletter for Clinicians

AI systems in medicine
are population-level
interventions.
We should treat them that way.

Every clinical AI tool you use today is changing how you think, what you order, and what you miss — at scale, across thousands of patients. The Epidemiology of Algorithms is the discipline — and the newsletter — built to watch those systems the way we watch drugs, devices, and disease.

700+ FDA-cleared AI/ML devices
~0 Post-market surveillance systems
Bi-
weekly Rigorous. Free. For clinicians.
Free for Clinicians
Subscribe & get two free clinical AI resources
Every new subscriber receives both immediately in their inbox.
Free — Both Included
10 Questions Every Clinician Should Ask Before Using an AI Tool
  • What population was this algorithm trained on?
  • Has it been validated in your clinical environment?
  • Who is notified when the model is updated?
  • What is the post-deployment surveillance plan?
  • And six more questions your vendor cannot answer
Clinical AI Pre-Deployment Checklist
  • Population representativeness of training data
  • Algorithm version control & change notification
  • Bias and equity pre-screening criteria
  • Post-deployment surveillance plan requirements
  • Governance and accountability structure

Biweekly. Rigorous. Free. No spam — ever.
Unsubscribe anytime.

✓
You're on the list.
Check your inbox — your two clinical AI resources
(10 Questions + Pre-Deployment Checklist)
are on their way. Issue 01 publishes Thursday March 19.

Welcome to the discipline.
Why this matters
01
AI shifts and drifts — silently
Like influenza, algorithms degrade gradually or change suddenly. You won't see it at the bedside. Population surveillance will.
02
Clinician behavior is the causal pathway
AI doesn't treat patients directly. It changes how you think. Alert fatigue, automation bias, and deskilling are epidemiological variables — not just annoyances.
03
Harm is invisible at the bedside
A 2% systematic error across 50,000 encounters is 1,000 harmed patients — none visible to any individual clinician. Epidemiology closes that gap.
"We do not approve a drug and then stop watching. We should not deploy an algorithm and then stop watching. The science that watches algorithms is the Epidemiology of Algorithms in Medicine."
— The Epidemiology of Algorithms, Issue 01
Join the Discipline

The algorithm you used today — is it the same one from six months ago?

If you can't answer that, you're practicing in a surveillance gap.

Biweekly newsletter for clinicians. Free. No spam.