Cost-effectiveness of different strategies for treating invasive aspergillosis in hematopoietic stem cell transplant recipients using a Markov model

Objectives: Hematopoietic stem cell transplant recipients have a 10% probability of developing invasive aspergillosis, which is a serious fungal infection with a high mortality. Several cost-effectiveness studies have been performed regarding the treatment of invasive aspergillosis in this particular patient population. Those studies use a decision tree as their model. However, an important disadvantage is that time is not explicitly incorporated in those models. Markov models, in contrary, do have this property which should make them better suited to represent the real-life situation. The objective of this study is twofold. First, it will assess the feasibility and additional value of using a Markov model (and thereby explicitly incorporating time) over a decision tree. Second, it will assess the cost-effectiveness of the different strategies in treating hematopoietic stem cell transplant recipients with invasive aspergillosis, explicitly considering that patients can switch at any time between health states. Methods: This study is based on the study of Ament et al. (2007) and since no individual patient data was available, data from their study had to be used. The study population consists of adult hematopoietic stem cell transplant recipients with proven or probable invasive aspergillosis. Five treatment strategies are compared which included Voriconazole, Caspofungin, Desoxycholate amphotericin B and Liposomal amphotericin B. The study is performed from the perspective of the health care purchaser. The Markov model consists of cycles of 1 day, with a time horizon of 5 years. At any time, patients in the study were in either of the following five health states: first-line treatment, second-line treatment, other licensed antifungal treatment (OLAT), finished antifungal treatment, death. The main outcomes are incremental cost-effectiveness ratio’s and net monetary benefits with a threshold of 50.000 euro per life year gained. Results: It is feasible to build a Markov model for this particular patient population. However, along the process, some problems emerged, which involved negative patient values, incorporating costs, and choosing the right distribution for the survival analysis. In our model Voriconazole + Desoxycholate, amphotericin B was dominant over all strategies except the Voriconazole + Caspofungin strategy, which was both costlier and more effective. The associated ICER was 39.899,14 euro per extra life year gained, which is deemed cost-effective when a threshold of 50.000 euro is used. Results were similar to the results in the decision tree of Ament et al. (2007). However, differences with their model include among others higher costs and less life years gained for all strategies and the treatment strategy Voriconazole + Desoxycholate amphotericin B was not dominated by Voriconazole + Caspofungin. Conclusions: This paper shows that it is possible to build a Markov model for evaluating cost-effectiveness of different treatment strategies for invasive aspergillosis in hematopoietic stem cell transplant recipients, thereby explicitly incorporating the time aspect. Voriconazole + Caspofungin appears to be the preferred treatment strategy based on cost-effectiveness, although evidence is low due to the imperfect data used.