Gene therapy in hemophilia A: a cost-effectiveness analysis

Abstract

Gene therapy provides a potential phenotypic cure for hemophilia, yet the cost of this novel treatment is high, tempering enthusiasm and raising questions regarding cost vs benefit. To evaluate the cost-effectiveness of gene therapy treatment of severe hemophilia A compared with prophylaxis with factor VIII (FVIII), we developed a Markov state-transition model to estimate the costs and effectiveness of severe hemophilia A treatment strategies from a United States health care system perspective. Quality-adjusted life-years (QALYs) were the effectiveness measure. In the base case, hypothetical cohorts of 30-year-old patients received gene therapy or FVIII prophylaxis. We obtained model probabilities and utilities from the literature and costs from Medicare reimbursement data. One-way and probabilistic sensitivity analyses were performed to test the robustness of results. Over a 10-year time horizon, total per-person gene therapy strategy costs were $1.0M and resulted in 8.33 QALYs, whereas prophylaxis cost $1.7M and resulted in 6.62 QALYs. Thus, gene therapy dominated prophylaxis (costs less and was more effective). Gene therapy remained dominant unless initial costs exceeded $1.6M and were <$100 000 per 1 QALY gained compared with prophylaxis if initial costs were <$1.7M. Results were not sensitive to variation of all other parameters over clinically plausible ranges. In a probabilistic sensitivity analysis simultaneously varying all parameters 3000 times over parameter distributions, gene therapy was dominant in 92% of model iterations. Treatment of severe hemophilia A with gene therapy is likely to be cost-saving or cost-effective compared with FVIII prophylaxis.

Graphical abstract

Figure 1.

Markov model. Patients remain in a health state (curved arrows) or move from 1 health state to another (straight arrows) based on transition probabilities. As patients cycle through the model, they accumulate costs and utilities expressed as QALYs. The Markov cycle length is 1 month and repeats over 10 years. During each cycle, patients are at risk for death from any cause, a bleeding event with or without joint damage. Gene therapy patients have an initial 1-year risk for failed therapy, resulting in transfer to the prophylaxis state.

Figure 2.

One-way sensitivity analysis. The results of 1-way sensitivity analysis are shown. Only variables whose variation caused a significant change in incremental cost-effectiveness (x-axis) are shown. Ranges of 1-way sensitivity analyses are shown next to each bar.