Economic impact of generic antiretrovirals in France for HIV patients’ care: a simulation between 2019 and 2023

Abstract

Background In a context where the economic burden of HIV is increasing as HIV patients now have a close to normal lifespan, the availability of generic antiretrovirals commonly prescribed in 2017 and the imminence of patent expiration are expected to provide substantial savings in the coming years. This article aims to assess the economic impact of these generic antiretrovirals in France and specifically over a five-year period. Methods An agent-based model was developed to simulate patient trajectories and treatment use over a five-year period. By comparing the results of costs for trajectories simulated under different predefined scenarios, a budget impact model can be created and sensitivity analyses performed on several parameters of importance. Results The potential economic savings from 2019 to 2023 generated by generic antiretrovirals range from €309 million when the penetration rate of generics is set at 10% to €1.5 billion at 70%. These savings range from €984 million to €993 million as the delay between patent and generic marketing authorisation varies from 10 to 15 years, and from €965 million to €993 million as the Negotiated Price per Unit (NPU) of generics at market-entry varies from 40 to 50% of the NPU for patents. Discussion This economic savings simulation could help decision makers to anticipate resource allocations for further innovation in antiretrovirals therapies as well as prevention, especially by funding the Pre-Exposure Prophylaxis (PrEP) or HIV screening.

Agent-Based modeling in Medical Research. Example in Health Economics

“…The full results are out of the scope of this paper which focuses on the methodological aspects of the agent-based simulation approach especially a statistical point of view focusing on the variability issue. Completed results together with their health economical interpretation are presented in [12]. These results include, for example, sensitivity analyses on the marketing authorization date, on the decrease in tariffs.…”

“…Whatever the objective of the study is, the common line is to the use of exogeneous datasets to learn / calibrate a predictive model which is used to feed an agent-based model of a (some) pre-specified outcome(s) of interest. As examples of execution models one finds Baseline's parameters evolution in time (an example is given in [12] and in Section 6), Virtual outcome generator, Disease progression model [10], Side effect model, drop-out model, Patients Recruitment model [30,31]. There is a wide variety of models available as candidate for execution model: parametric models such as Markov Process, Cox process, regression model, Bayesian network,... and non-parametric models such as machine learning techniques (Random forest, XGBoost, Decision trees, SVM, deep learning,...).…”

“…These population-based approaches yields to punctual predictions of the variations of costs embedded by the switch to the generics. In [12], of which Section 6 is a summary, authors have set Figure 8: Simulation of 100 datasets of 500 observations each were performed using the discrete, the continuous and the Copula approaches. For each dataset, the p-value using Student test between Prognosis variable and Blood pressure were computed and reported using boxplot and jitter plot.…”

“…The full results are out of the scope of this paper which focuses on the methodological aspects. Complete results together with their health economical interpretation are discussed in [12]. However, as an illustration of the relevancy of this approach, results on the impact of penetration rate on the differential costs are presented in Section 6.6.…”

“…Section 5 focuses on the issue of the error of prediction in execution models and investigate the importance of accounting for this error on the toy example. Section 6 give the main lines of a study [12] on VIH treatments switching to generics involving an ABM. Finally, Section 7 states some conclusions and recommendations on the use of ABM in medical research.…”