Abstract
Background/Aims Implantable cardioverter defibrillators (ICDs) result in striking mortality benefits among randomized trial populations, but elucidation of outcomes in real-world populations is needed to optimize care and coverage decisions. The Cardiovascular Research Network (CVRN) Longitudinal ICD Study is a new 7-site 3.5-year project evaluating the rate of appropriate device therapies, device complications, hospitalization, mortality, and utilization/cost among a cohort of primary prevention ICD patients.
Methods This project links baseline data from the National Cardiovascular Data Registry (NCDR) for ICDs, the virtual data warehouses (VDW) of participating CVRN sites, and novel collection of post-implant ICD device activity from CVRN care delivery systems. NCDR ICD data from 14 implanting hospitals from 2006–2009 were matched to health plan membership and uploaded to the study following necessary approvals. VDW tables have been constructed at the CVRN sites to capture longitudinal clinical and administrative detail. Forms and procedures for abstracting ICD device interrogations and treated arrhythmic episodes have been pilot tested, and procedures have been established both for central review of source documentation and adjudication by an external panel of expert electrophysiologists.
Results A cohort of 2500 primary prevention ICD recipients has been assembled. The pilot study evaluated collection and review of device interrogations and treated arrhythmic episodes among a sample of 43 subjects. Lessons from the first study year include:
-
the need to allow additional time and effort to establish/modify contractual agreements between external partners;
-
the benefits of support from an engaged group of stakeholders;
-
the substantial differences across and within study sites as to how ICD device interrogation records are archived and tracked, with local data sources including centralized medical records, stand alone pacemaker/ ICD clinic files, electronic interrogation archives, and remote interrogation websites; and,
-
the importance of caution when planning for the review and adjudication of real world medical record data based on published approaches from prospective trials.
Conclusions Following establishment of the study cohort and piloting abstraction/adjudication procedures, the data collection phase is underway. Further lessons from initial abstraction will be available at the time of the HMORN conference. The study period ends March 2013.
