FamilyComorbid AI: a Web-Based Clinical Decision Support Tool for Cardiovascular Risk Prediction in Patients with Type 2 Diabetes and Multimorbidity

Abstract

Objective: To develop and validate a web-based clinical decision support tool (FamilyComorbid AI) that integrates evidence-based risk multipliers and protective factors to provide personalized cardiovascular risk estimates and therapeutic recommendations for primary care. Methods: Risk multipliers and protective factors were derived from pivotal randomized clinical trials and meta-analyses (EMPA-REG OUTCOME, CREDENCE, LEADER, SUSTAIN-6, REWIND, FIDELIO-DKD, FIGARO-DKD, CTT Collaboration). A multiplicative model was implemented using a progressive web app with a React.js frontend and a Node.js backend. Content validity was assessed by an expert panel (n=7), and usability was evaluated using the System Usability Scale (SUS) with primary care physicians (n=24). Results: The tool incorporates six risk multipliers and eight protective interventions. The expert panel reached a 100% consensus on content validity. The mean SUS score was 82.4 (SD 8.7), indicating excellent usability. The average calculation time was 47 seconds. Users reported high satisfaction with the visual communication of risk (4.6/5) and the therapeutic recommendations (4.5/5). Conclusions: FamilyComorbid AI provides an accessible, evidence-based tool for cardiovascular risk stratification in patients with diabetes and multimorbidity.