Integrating big data to assess the economic impact of the implantable Cardioverter Defibrillator Therapy

Abstract

Objectives: The main purpose of this study was to link clinical and administrative healthcare databases in order to assess the economic impact of implantable cardioverter defibrillator (ICD) with or without cardiac resynchronization pacing (CRT-D) in Lombardy, the most populated Italian region providing universal healthcare coverage for about 10 million inhabitants. Methods: Data were extracted from: i) data warehouse DENALI, that organizes healthcare administrative databases concerning all subjects covered by Lombardy Health System (HS); ii) national ICD database. After linking DENALI and clinical information extracted from the ICD database, we identified patients with ICD and followed them from the date of the first implant to 12/31/2010, recording hospitalizations, drugs and outpatient claims. Direct healthcare costs were analysed from the perspective of the HS. We estimated mean annual per-capita costs after the first ICD implant: overall and stratified by indication (primary and secondary prevention) or type of implanted device (single- chamber, dual-chamber and CRT-D). Results: During the period 2003-2010, 12,525 subjects underwent a first ICD implant. Mean annual per-capita cost during follow-up was 6,086 (95%CI: 5,970-6,211): 72.7% due to hospitalizations ( 4,422), 15.7% to pharmacological therapies ( 957) and 11.6% to outpatient services ( 706). No difference was observed between the mean annual expenditure for primary ( 6,179; 95%CI: 6,001-6,353) and secondary ( 5,996; 95%CI: 5,821-6,185) prevention. As for the type of implanted device, patients with CRT-D cost more than those with single-chamber or dual-chamber: respectively 6,592 (95%CI: 6,376-6,847), 5,728 (95%CI: 5,541-5,972) and 5,932(95%CI: 5,717-6,117). The difference was attributable to hospitalization expenses. Conclusions: ICD use is growing and it is important to assess the efficacy and the burden of this therapy, given its economic implications. The combined use of clinical information from national ICD database and big administrative data could overcome the limitation of both data sources, leading to an improvement in the monitoring of ICD therapy.