OBJECTIVES : The Pediatric Interagency Registry for Mechanical Circulatory Support (PediMACS) is a national registry for U.S. FDA-approved ventricular assist devices (VADs) in patients ≤18 years old. Between the launch in September 2012 and June 2015, 37 hospitals in the USA have enrolled patients. We sought to develop a comprehensive clinical and financial database in order to perform health economics and outcomes research. Specifically, we merged data from enrolled PediMACS and the Pediatric Health Information Systems (PHIS), a health care administrative database from 49 freestanding children’s hospitals in the US, to estimate resource utilization and costs, and to address important health economics and outcomes questions.

METHODS : We merged the data from PediMACS with PHIS using a stepwise matching algorithm with indirect identifiers including hospital ID, gender, birth date, implant date. Health costs were then estimated for the patients from hospitalization of VAD implant through death, transplant, or VAD explant based on PHIS data.

RESULTS : Twenty of 37 PediMACS hospitals contributed data to PHIS. Of 228 patients enrolled on the PediMACS at PHIS hospitals, 202 (89%) were successfully matched. All of these patients had inpatient PHIS data available beginning at the date of study enrollment. Total inpatient costs varied by VAD type with temporary to durable VAD (median cost $0.97 million, IQR $0.78-$1.23 Million) being most expensive, followed by durable VAD (median cost $0.69 million, IQR $0.51-$1.04 Million) and temporary VAD (median cost $0.52 million, IQR $0.25-$1.06 Million). Room/board and clinical services were the largest contributors to inpatient treatment cost, representing 52% of the total cost.

CONCLUSIONS : These data demonstrate that patients enrolled the PediMACS can be successfully identified in PHIS administrative database and that health care resource utilization and costs can be estimated. The merged data can be used as a platform for comparative effectiveness and cost-effectiveness studies.