Remote Follow-up in a Heart Failure Pragmatic Trial: Insights From the CONNECT-HF

•Using one of the largest heart failure pragmatic trials (CONNECT-HF; (Care Optimization Through Patient and Hospital Engagement For HF), we assessed the completeness and validity of remote follow-up, integrating participant-reported via a centralized call center and site-reported approaches.•Both participant-reported via centralized call center- and site-reported approaches substantially missed hospitalization events. In contrast, both approaches demonstrated high and comparable adjudication rates for hospitalizations among the reported hospitalization events.•Our study highlighted that integrating both approaches remains essential to capture and validate effective clinical outcomes in the context of current pragmatic trials. Randomized controlled trials typically require study-specific visits, which can burden participants and sites. Remote follow-up, such as centralized call centers for participant-reported or site-reported, holds promise for reducing costs and enhancing the pragmatism of trials. In this secondary analysis of the CONNECT-HF (Care Optimization Through Patient and Hospital Engagement For HF) trial, we aimed to evaluate the completeness and validity of the remote follow-up process. The CONNECT-HF trial evaluated the effect of a post-discharge quality-improvement intervention for heart failure compared to usual care for up to 1 year. Suspected events were reported either by participants or by health care proxies through a centralized call center or by sites through medical-record queries. When potential hospitalization events were suspected, additional medical records were collected and adjudicated. Among 5942 potential hospitalizations, 18% were only participant-reported, 28% were reported by both participants and sites, and 50% were only site-reported. Concordance rates between the participant/site reports and adjudication for hospitalization were high: 87% participant-reported, 86% both, and 86% site-reported. Rates of adjudicated heart failure hospitalization events among adjudicated all-cause hospitalization were lower but also consistent: 45% participant-reported, 50% both, and 50% site-reported. Participant-only and site-only reports missed a substantial number of hospitalization events. We observed similar concordance between participant/site reports and adjudication for hospitalizations. Combining participant-reported and site-reported outcomes data is important to capture and validate hospitalizations effectively in pragmatic hear