COIOSDeep risk insights through multi-disciplinary research across the globe
a close up of a network with wires connected to it
Item new Preprint · not peer-reviewed Added

ITE-Transformer estimates individualised treatment effects from electronic health records

A transformer trained on about 350,000 All of Us records estimated per-patient treatment effects in diabetes and cancer drug comparisons.

The authors built a transformer to estimate individualised treatment effects from records of about 350,000 All of Us participants. It learned diagnosis and medication histories without outcome labels, then fed doubly robust estimation of treatment choice and outcome risk. Tests compared GLP-1 receptor agonists with DPP-4 inhibitors on cardiovascular events in diabetes, and immunotherapy with chemotherapy on cancer mortality. Treatment-choice and outcome models discriminated and calibrated better than tabular methods and the same transformer without pre-training. Predicted effects varied substantially between patients, with patterns linked to diagnoses, medication histories and comorbidities. The estimates rest on observational records, so unmeasured confounding remains possible, and no external cohort was used for validation.

Why it is interesting: Most health-record foundation models predict risk rather than estimate treatment effects. This one targets which patients may gain more from one drug than another, using observational data.

Source
medRxiv, 20 November 2026
Paper
The ITE-Transformer: A pre-trained foundation model for predicting individualized treatment effects from electronic health records
Authors
Arceri C, Angelis E, Heesen P, Melnychuk V, Bendszus M, Bauer S et al.
DOI
10.64898/2026.11.19.26368479
Type
Preprint
Design
Methods development and internal evaluation: self-supervised transformer plus doubly robust causal ML on structured EHRs of ~350,000 All of Us participants, two comparative-effectiveness tasks
Verdict
New finding
Driver
AI in medicine

Other drivers in Data, models and systems