Clinical decision support as an implementation strategy to expand identification and administration of treatment of opioid use disorder in the emergency department

Abstract

Introduction

US opioid overdoses and deaths continue to increase, despite historic national investment to mitigate risk and improve access to evidence-based treatment. Unfortunately, implementation of emergency department (ED) buprenorphine – an effective medical treatment for opioid use disorder (OUD) – has been limited. Our objective was to assess the effectiveness of an electronic health record (EHR)-integrated, interruptive clinical decision support (CDS) tool to improve rates of ED initiated OUD treatment.

Methods

This is an observational, pre-post study of a CDS tool designed to identify and facilitate treatment of patients with OUD using electronic health record data. Patients were included if treated at our urban, academic ED between May 1, 2022, and November 8, 2023. The CDS triggered based on a rules-based algorithm using routinely collected EHR data which were identified from a previously validated EHR OUD phenotype. Outcomes are organized under a modified RE-AIM framework, with the primary outcome, Effectiveness, measured by the proportion of OUD patients receiving buprenorphine (administered/prescribed; filled prescriptions). Secondary outcomes include patient Reach, clinician Adoption, and fidelity to Implementation. Chi Square tests and Bayesian structural time-series models evaluate differences in outcomes before and after CDS implementation (CausalImpact package v1.3.0 in R v4.4.0).

Results

There were 171,221 total ED visits during the study period. Patient characteristics before and after CDS implementation were similar. CDS triggered in 4.7 % (2754/58,173) of encounters after initiation of intervention, reaching 116 unique emergency medicine providers and 2566 ED patients. Clinicians adopted the CDS, accessing the OUD treatment pathway link or ordering a social work consult for substance use, in 27 % (1266/4746) of CDS alerts. When compared to the pre-implementation period, CDS implementation was associated with increased buprenorphine administration in the ED by 31 % (95 % CI: 16–47 %, p = 0.001), buprenorphine prescribing from the ED by 20 % (95 % CI: 5–38 %, p = 0.007), and the buprenorphine fill rate at an affiliated ED pharmacy by 17 % (95 % CI: 1–36 %, p = 0.017).

Conclusions

Implementation of an EHR-integrated, CDS was associated with increased ED buprenorphine administration, prescribing, and prescription fills among ED patients with OUD. Further efforts are needed to assess maintenance strategies that improve adoption, minimize interruptiveness, and optimize workflow congruence.