Adoption of shared decision-making and clinical decision support for reducing cardiovascular disease risk in community health centers

Abstract

Objective: Electronic health record (EHR)-based shared decision-making (SDM) and clinical decision support (CDS) systems can improve cardiovascular disease (CVD) care quality and risk factor management. Use of the CV Wizard system showed a beneficial effect on high-risk community health center (CHC) patients' CVD risk within an effectiveness trial, but system adoption was low overall. We assessed which multi-level characteristics were associated with system use.

Materials and methods: Analyses included 80 195 encounters with 17 931 patients with high CVD risk and/or uncontrolled risk factors at 42 clinics in September 2018-March 2020. Data came from the CV Wizard repository and EHR data, and a survey of 44 clinic providers. Adjusted, mixed-effects multivariate Poisson regression analyses assessed factors associated with system use. We included clinic- and provider-level clustering as random effects to account for nested data.

Results: Likelihood of system use was significantly higher in encounters with patients with higher CVD risk and at longer encounters, and lower when providers were >10 minutes behind schedule, among other factors. Survey participants reported generally high satisfaction with the system but were less likely to use it when there were time constraints or when rooming staff did not print the system output for the provider.

Discussion: CHC providers prioritize using this system for patients with the greatest CVD risk, when time permits, and when rooming staff make the information readily available. CHCs' financial constraints create substantial challenges to addressing barriers to improved system use, with health equity implications.

Conclusion: Research is needed on improving SDM and CDS adoption in CHCs.

Trial registration: ClinicalTrials.gov, NCT03001713, https://clinicaltrials.gov/.

Keywords: cardiovascular diseases; community health centers; decision making–shared; decision support systems–clinical; medical informatics.

Figure 1.

CV Wizard Alert. © 2021 Epic Systems Corporation.

Figure 2.

Example of CV Wizard provider view.

Figure 3.

Example of CV Wizard patient view.

Figure 4.

Multivariate regression results. Adjusted rate ratios and 95% confidence intervals of encounter, clinic, provider, and patient characteristics associated with system use (N encounters = 21,685) clustered at the clinic and provider level.