Utilizing the index for mortality prediction after cardiac transplantation risk score to predict hospital resource consumption.

Abstract

The index for mortality prediction after cardiac transplantation (IMPACT) risk score incorporates 12 preoperative recipient-specific variables, and has been validated as an accurate predictor of short- and long-term mortality after orthotopic heart transplantation (OHTx). We believe it can also be used to predict hospital costs, and we hypothesize that higher preoperative IMPACT risk scores are associated with increased hospital resource consumption. All OHTx patients ≥18 years of age at our institution were reviewed from 1 January 2000 to 31 December 2014. Total index hospitalization costs post-transplant were extracted and presented in 2014 consumer price index inflation-adjusted US dollars. Patients were stratified into quartiles (Q) according to IMPACT risk scores. Logarithmic transformation normalized cost data, and linear regression assessed for correlation. A comparison of cost between Q of IMPACT risk score was performed using rank-sum and Kruskal-Wallis tests. Survival was estimated using the Kaplan-Meier method. Three hundred fifty-six (n = 356) OHTx were performed during the study period. The median IMPACT score for the cohort was five (interquartile range [IQR] 3-6). Eight (2.2%) patients died within 30-days and 1-year Kaplan-Meier survival was 88.3%. The median length of stay (LOS) was 16 (IQR 14-24) days. The median hospital cost for index admission was $222 200 (IQR:$169 200-$313 700). Median LOS was longer in Q4 vs Q1 (18 days vs 15 days, P = .01) and index hospital costs in Q4 were significantly higher compared to Q1 patients ($280 400 vs $205 000, P < .01). There was a significant positive correlation between IMPACT risk score and cost (regression coefficient .04, P < .01). This is the first study in adult cardiac transplantation to identify a positive correlation between hospital cost and recipient risk using the IMPACT risk score. Cost and resource consumption for the index admission after OHTx were significantly higher in the highest IMPACT risk Q compared with patients in the lowest Q.