Personalizing Clinical Decision Support for Heart Failure Treatment to Clinicians' Needs

点击翻译按钮获取中文摘要

PROJECT SUMMARY Clinical decision support (CDS) tools are pervasive and can “nudge” clinicians to make the best decisions easy, yet currently lead to minimal improvements in patient outcomes. Although often ignored, consideration of contextual factors and minimizing irrelevant information improves CDS outcomes. To minimize irrelevance, currently existing, ‘traditional CDS’ are often designed to be patient-specific, but are not tailored to clinicians. For example, traditional CDS address common prescribing misconceptions that are not relevant for all clinicians. However, prescribing patterns could be used to determine whether prescribing misconceptions might exist and then conditionally present information within a ‘personalized CDS’ to address a specific clinician’s misconceptions; thereby minimizing irrelevance and alert fatigue. A ‘personalized CDS’ could substantially improve guideline-directed management and therapy (GDMT) for the many suboptimally treated patients with heart failure and reduced ejection fraction (HFrEF). Aim 1: Design and build prototypes of traditional and personalized CDS to address common misconceptions of GDMT for HFrEF. We will create a personalized and traditional CDS prototype for 4 categories of GDMT: beta blockers, sacubitril/valsartan, mineralocorticoid receptor antagonists and sodium/glucose cotransport 2 inhibitors. Clinicians will prioritize the misconceptions to address. The traditional CDS will address all prioritized misconceptions, while the personalized CDS will conditionally address the misconceptions based on clinician-specific prescribing patterns. To account for contextual factors, we will use the Practical Robust Implementation and Sustainability Model (PRISM) to guide design and usability testing. Aim 2: Pilot the traditional and personalized CDS tools in real-world care settings. Aim 3: Compare the traditional and personalized CDS in a pragmatic randomized controlled trial. Cardiology and primary care clinics at one health system will be cluster-randomized. We will use sequential mixed methods and PRISM evaluation metrics to compare the two CDS tools. Quantitative outcomes include reach, adoption and effectiveness of prescribing. We will interview 15 frontline clinicians and 5 leaders to 1) identify PRISM factors influencing implementation outcomes, and 2) plan for external dissemination. This proposal was designed to address my training gaps: 1) EHR architecture, 2) behavioral economics/nudges, and 3) pragmatic trial design. Completion of this proposal will ensure my development into an independent investigator that leverages implementation science to create innovative CDS solutions that consistently and effectively optimize GDMT for HFrEF across health systems. This research is significant because it has the potential to substantially improve GDMT and outcomes for high-risk patients with HFrEF. Our innovative, personalized CDS challenges the status quo of “one size fits all” CDS by individualizing CDS to both patients and clinicians; a paradigm-shift that will have far-reaching influence on CDS development and GDMT.