Point of care randomization (POCR) engine: demonstration of a transportable clinical informatics technology that improves health systems' ability to conduct embedded clinical trials

Why clinicians should know about it

• Picked for Health Informatics (paper of the day, 1 November 2026): EHR-embedded randomization deployed with measurable practice impact

Abstract

OBJECTIVE: We describe a new point of care randomization (POCR) engine developed at UCSF and implemented at UCLA that supports randomization of patients, encounters, clinics, or any other "object" in Epic, with programmable ratios, strata, and randomly varying block sizes. MATERIALS AND METHODS: The feature, programmed in Caché, uses Epic rule infrastructure to implement inclusion criteria and generate and use blocks of randomization assignments. Configuration details, randomization assignments, and context information (including stratification variables) are stored in custom master files in Chronicles and exported to Clarity (Epic's reporting database). As a demonstration use-case, we conducted a pragmatic RCT of a provider-facing alert for short-interval (30-day) repeat HbA1c orders at UCLA Health. At the instant providers initiated a repeat order, the POCR engine silently enrolled and randomized patients, stratified by age and previous HbA1c, with randomly varying block sizes. RESULTS: During the 3-month study, 333 patients were enrolled. The total number of orders signed in the intervention arm (primary outcome) was 18/172 (10%) vs 104/161 (65%) in the control arm (Incidence Rate Ratio 0.16, 0.95 CI, 0.10-0.27). DISCUSSION: We successfully transported and deployed a new technology with the capability to instantaneously perform automated stratified block randomization. By performing eligibility assessment, enrollment, and assignment using programmable ratios, strata, and randomly varying block sizes, the POCR engine preserves internal validity while minimizing disruption to care, representing a substantive advance in informatics infrastructure for embedded randomized trials. CONCLUSION: The POCR engine is a transportable clinical informatics technology that transforms health systems' ability to conduct pragmatic clinical trials within the EHR.

Abstract as published, via PubMed.