Cost-utility analysis or equivalent income approach to evaluate healthcare programs: what impact on allocation recommendations? A tutorial

Context : Since the late 2000s, various approaches have been explored to adapt health technology assessment methods to take into account distributive principles. This article deals with one of these innovative methods, the Equivalent Income (EI) approach. The EI approach requires microsimulation, as it directly takes into account individual preferences for health by valuing health outcomes in monetary terms and estimate how health states can affect individual incomes. It also allows for varying degrees of societal aversion to health and income inequalities. Objective : The article proposes a generic Markov microsimulation model to illustrate in a didactic way how the EI approach can be implemented and to what extent it can lead to recommendations that differ from those derived from Cost-Utility Analysis (CUA). Methods : The model is an adaptation of Krijkamp et al. (2018), which simulates and compares two strategies for treating a hypothetical disease. The first step is to estimate, for each individual in a representative sample of the French population, the evolution of their health status and health care consumption over a 30 year time horizon. The second step is to simulate changes in individual income, willingness to pay for perfect health, health insurance contributions and, finally, individual EI. The individual EIs are then aggregated into a social welfare function (SW) that incorporates different degrees of inequality aversion. Allocation recommendations based on the EI approach and the ACU are compared according to the different scenarios developed to highlight differences between the two approaches. The first scenario (Scenario A) describes a situation where there is no health inequality in the disease prevalence. Conversely, scenario B deals with a situation where the disease is more common among disadvantaged people. In addition, two sensitivity analyses were carried out: the most disadvantaged (the last deciles two of income) do not contribute to compulsory health insurance (sensitivity analysis 1); the probability of death is doubled (sensitivity analysis 2). Results : The context of health inequality in terms of disease prevalence and health financing system may lead to different recommendations under the EI approach, whereas the CUA is insensitive to such information.