Clinical Decision Support Software May Improve Atrial Fibrillation Outcomes

Researchers developed a clinical decision support (CDS) tool that may help primary care providers (PCPs) deliver individualized guideline-based atrial fibrillation (AF) care. The findings were published in BMC.

In this trial, which took place in Nova Scotia, called the Integrated Management Program Advancing Community Treatment of Atrial Fibrillation (IMPACT-AF), software design began in 2013. “The goal was to create a fully integrated application that computerized national clinical guidelines into decision rules to support AF patient management at point-of-care. A key CDS feature was prioritized automated alerts signaling material changes in patient clinical or biochemical profiles requiring expedited treatment modifications. Users, both providers and patients, could interact with the software by entering and receiving health-related information and or alerts. Where possible, relevant patient data could be captured in real time,” the researchers wrote of the software.

The study identified, enrolled, and randomized 1:1 to CDC access 104 PCPs and 597 AF patients, and 99 PCPs and 548 AF patients to usual care. Throughout the study duration, the researchers noted, the team documented key challenges regarding CDS development and implementation (2013–2018), data integration (2013 to 2020), and study recruitment (2014–2016).

Keys to Developing Successful CDS Software

The study identified three key findings. Firstly, adequate resources should be allocated for software development, updates and feasibility testing, as the software development took longer than projected. According to the researchers, end-user feedback suggested network access and broadband speeds impeded uptake; they felt further that the CDS was not sufficiently user-friendly or efficient in supporting AF care (i.e., repetitive alerts).

Secondly, integration across e-platforms is crucial. Challenges such as intellectual property prohibited CDS integration within electronic medical records and provincial e-health platforms, the researchers noted. Moreover, double login and data entry were impediments to participation or reasons for provider withdrawal. They further noted that data integration challenges prevented easy and timely data access, analysis, and reporting.

Lastly, primary care study recruitment is resource intensive, according to the researchers; with 25% of eligible providers and 12% of AF patients in Nova Scotia enrolled. They noted that the most effective provider recruitment strategy was in-office, small group lunch-and-learns.

“A rapid growth in healthcare data is leading to widespread development of CDS software to analyze it, better support evidence-informed patient management and thereby improve health outcomes,” the researchers wrote in conclusion.

“Our experience of developing and implementing a pragmatic, cluster randomized controlled trial of a CDS in the real-world setting found a variety of practical issues to address if such applications are to succeed. There is a need for the allocation of adequate resources for CDS software development and updates, robust feasibility testing, as well as for primary care study recruitment. Most critical is a need for the integration of applications across e-health platforms. We hope that by sharing our experiences others can learn and achieve greater success in their future pragmatic clinical trials of CDS tools.”