Cost-effectiveness evaluation of the PROPPR trial transfusion protocols

Abstract

Background: There have been no prior investigations of the cost effectiveness of transfusion strategies for trauma resuscitation. The Pragmatic, Randomized, Optimal Platelet and Plasma Ratios (PROPPR) study was a Phase III multisite, randomized trial in 680 subjects comparing the efficacy of 1:1:1 transfusion ratios of plasma and platelets to red blood cells with the 1:1:2 ratio. We hypothesized that 1:1:1 transfusion results in an acceptable incremental cost-effectiveness ratio, when estimated using patients' age-specific life expectancy and cost of care during the 30-day PROPPR trial period.

Study design and methods: International Classification of Diseases, Ninth Revision codes were prospectively collected, and subjects were matched 1:2 to subjects in the Healthcare Utilization Program State Inpatient Data to estimate cost weights. We used a decision tree analysis, combined with standard costs and estimated years of expected survival to determine the cost effectiveness of the two treatments.

Results: The 1:1:1 group had higher overall costs for the blood products but were more likely to achieve hemostasis and decreased hemorrhagic death by 24 hours (p = 0.006). For every 100 patients treated in the 1:1:1 group, eight more achieved hemostasis than in the 1:1:2 group. At 30 days, the total hospital cost per 100 patients was $5.6 million in the 1:1:1 group compared with $5.0 million in the 1:1:2 group. For each 100 patients, the 1:1:1 group had 218.5 more years of life expectancy. This was at a cost of $2994 per year gained.

Conclusion: The 1:1:1 transfusion ratio in severely injured hemorrhaging trauma patients is a very cost-effective strategy for increasing hemostasis and decreasing trauma deaths.

Fig. 1.

Decision tree used to estimate trial cost-effectiveness ratios. Tran 1:1:1 denotes treatment group. No HS is proportion who did not reach hemostasis. Live 30D are patients surviving at end of study. Die are patients who died during the study period. Population are the estimated number of patients on each tree branch expected if 100 patients were enrolled in each trial arm. Numbers under each branch in the tree are the proportion of patients in the PROPPR trial with the outcome.

Fig. 2.

Sensitivity analysis of cost and life expectancy differences. Sensitivity analysis represents cost and life expectancy differences between treatments for 1000 Monte Carlo simulations of ICERs for model estimates for total costs per additional year of life expectancy with the percentage of simulations per quadrant.