Making Evidence-Based Clinical Decision Support Implementable in Different Electronic Health Records
Clinical decision support that can be implemented in different types of electronic health records has the potential to scale evidence-based practice across healthcare systems.
Scaling evidence-based clinical decision support for serious conditions is difficult
Venous thromboembolism (VTE)—a condition that occurs when a blood clot forms in a vein—includes both deep venous thrombosis (DVT) and pulmonary embolism (PE) and can lead to death or disability if not caught early and treated. While there are effective evidence-based clinical prediction rules for assessment of VTE that have been integrated into clinical decision support (CDS) tools, it is often difficult to scale these tools across organizations and electronic health record (EHR) platforms.
EvidencePoint is an EHR-independent CDS platform
To support early detection and treatment of VTE, Drs. Alex Spyropoulos and Thomas McGinn and their team at the Northwell Health Center for Health Innovations and Outcomes Research developed an EHR-independent CDS software platform called EvidencePoint. This platform provides a suite of individual CDS solutions capable of being integrated into clinical workflows within various EHRs at various clinical sites without requiring the solutions to be “rebuilt” for each deployment. As Dr. Spyropoulos describes, “The beauty of this platform is that it was developed as EHR agnostic. If it works for one EHR, it should work for any EHR.”
The research team deployed and tested three CDS applications using the platform: 1) the Northwell COVID-19 Survival (NOCOS) CDS application, which was tested among patients with and without COVID-19; 2) the International Medical Prevention Registry on Venous Thromboembolism-D-Dimer (IMPROVE-DD) CDS application for VTE risk assessment; and 3) the Wells’ Criteria CDS application for PE diagnosis risk stratification.
When the COVID-19 pandemic hit in spring 2020, they were able to quickly deploy the NOCOS CDS across all Northwell hospitals, where it was available for easy access to providers in the ED directly through the EHR. In December 2020, they launched a clustered randomized trial and deployed the IMPROVE-DD CDS application for VTE risk assessment to two of Northwell’s largest tertiary hospitals, where it was configured as a mandatory component of the VTE prophylaxis process for all hospitalized patients. Two similarly sized hospitals were randomized to serve as controls (i.e., no CDS application). Over the course of a 1-year study, the application was used with 5,249 unique patients at the intervention sites.
In September 2021, they deployed the Wells’ Criteria CDS application for PE diagnostic risk assessment to the same two tertiary hospitals that were using the IMPROVE-DD application. One hospital used the Wells’ Criteria application with a standard user interface. The other hospital added a feature in the user interface that offered a nudge designed to increase user adoption of the tool. Over the course of a 6-month pilot study, the application was used during 1,735 patient encounters.
CDS use improved across the platform
Appropriate thromboprophylaxis rates were higher at intervention sites using IMPROVE-DD during inpatient stays and after discharge, and had fewer venous, arterial, and total thromboembolic events. The study also found that the IMPROVE-DD VTE risk assessment model demonstrated very good discrimination to identify hospitalized COVID-19 patients at low, moderate, and high VTE risk.
“What we saw was a 50% relative increase in appropriate in-hospital prophylaxis when comparing the intervention group that used the tool with the control group that did not use the tool.”- Dr. Alex Spyropoulos
At the ED site where the Wells’ Criteria PE diagnostic risk assessment with nudges was deployed, providers followed the tool’s CDS recommendations 46.3% of the time, compared with a 23.2% at the ED that used the standard, no-nudge version of the tool.
This virtual doubling of provider adoption is an indication that nudges have the potential to dramatically increase CDS adoption rates. However, the use of the Well’s Criteria was found to not differ significantly in ruling out PE in COVID-19-positive versus -negative patients.
Ultimately, the research accomplished an important goal of furthering the dissemination of evidence-based practices at the point of care. By demonstrating the feasibility of a workflow-integrated CDS platform that can be used in different EHRs, the team has shown the value of these CDS applications in the short term (the NOCOS app, the IMPROVE-DD app, and the Wells’ Criteria app). The team also developed a system that is capable of bringing those applications, and others, to health systems beyond Northwell Health.