Development of the international cardiac rehabilitation registry data dictionary using a Delphi process

Abstract

Background: Cardiac rehabilitation (CR) registries are a powerful tool for improvement of care quality, and consequently patients’ outcomes. However, there are few CR registries worldwide, with Canada’s currently being re-initiated. The International Council of Cardiovascular Prevention and Rehabilitation (ICCPR) aims to establish a registry specifically for low-resource settings, where the burden of cardiovascular disease is greatest and the need for program development highest. The aim of this study was to develop the data dictionary for the International Cardiac Rehabilitation Registry (ICRR). Methods: The ICRR executive contacted all available CR registries identified in a literature review (Poffley et al., 2017, EJPC) to request they share their data dictionaries; 36 unique variables (including patient-reported outcomes) were extracted. Representatives from existing registries, authors of CR quality indicators (identified in Moghei et al., 2019 review in JCRP), and ICCPR members/friends from low-resource settings were invited to serve on a Delphi panel; in accordance with best practices, 20 experts were sought. For the first step, an online survey was sent through REDCap in June 2020, where panelists rated the variables on a scale from 1-7 in terms of feasibility, importance/relevance, actionability, and evidence, as well as overall rating. Panelists were also asked to specify potential additional variables. A webcall is being held to reach consensus. For stage 2, panelists will be asked to finalize the definition of each included variable, and to rate which variables should be used for benchmarking against other programs on a registry dashboard; responses will be discussed on a 2nd call, until consensus is reached. Results: 20/26 panelists completed the 1st survey. Variables with an overall rating <5/7 were to be excluded, however the minimum for any variable was 5.3 (processed food consumption; followed by economic security, marital status and medication coverage); the panel is considering a short and long list. Variables with the highest ratings included (in descending order): sex, tobacco use, year of birth, referral diagnosis, medications, referral intervention, and premature program termination/completion. Additional variables considered included: ethnocultural background, barrier assessment, risk assessment/ stratification, vaccination, sleep, blood glucose, alcohol consumption, anxiety, and support. The ICRR dictionary will be posted for public comment prior to finalization. Conclusion: The final data dictionary will be built into the ICRR, along with dashboards for benchmarking, and piloted-tested in low-resource settings before launch. It is hoped this will serve to harmonize CR assessment internationally and enable quality improvement in CR delivery.