OBJECTIVES

: In Australia, a limited number of patients who are eligible and waiting for a heart transplant (HTx) can be supported by a ventricular assist device (VAD) as “bridge to transplant” (BTT). Occasionally, a patient receives a VAD as “bridge to candidacy” (BTC) prior to wait list activation. This economic evaluation compares the real-world, restricted use of VADs as BTT/BTC with two hypothetical policies of “Unlimited VADs” and “Unlimited heart transplants”.

METHODS

: A cohort state‑transition model was developed. A retrospective cohort study of 77 waitlisted patients with linked administrative data on hospital admissions and emergency visits provided costs. Transition probabilities for survival and time on the wait list were estimated from published registries. Registry data was utilised as there are no randomised controlled trials for VAD and HTx. Model time horizon was 5 years with an Australian health care perspective and 5% discount rate for costs and outcomes. Outcomes include life years and quality‑adjusted life‑years (QALYs).

RESULTS

: The retrospective cohort included 19 VAD-HTx, 6 VAD only, 42 HTx only and 10 patients who received neither during the study period. VAD recipients were generally sicker than those unsupported by VADs. The cost of the initial hospitalisation were larger for VAD recipients ($260,654) than HTx only ($126,333), whilst subsequent hospitalisations were similar. The “Unlimited VAD” policy was dominated, being the most costly with patients still awaiting and receiving HTx, however, produced greater QALYs than the real-world situation. The “Unlimited heart transplant” policy delivered greater QALYs compared to real-world but was not cost-effective under current acceptable thresholds in Australia.

CONCLUSIONS

: The current restricted use situation was the most cost‑effective compared to the unlimited supply policies. Understanding the interplay between time on wait list and the availability of limited resources may require more complex modelling to account for resource constraints.