Quality of health economic evaluations submitted to the Italian Medicines Agency: current state and future actions.

OBJECTIVES

The purpose of this study was to evaluate the current state of health economic evaluations (HEEs) submitted by pharmaceutical companies to the Italian Medicines Agency (AIFA) as part of their pricing and reimbursement (P&R) dossiers, and to explore potential future actions in order to enhance their quality.

METHODS

All company dossiers submitted from October 2016 to December 2018 were reviewed to select those containing pharmacoeconomic studies. The general characteristics of HEEs were described and their quality assessed based on a checklist adapted from Philips et al. (Review of guidelines for good practice in decision-analytic modelling in health technology assessment. Health Technol Assess. 2004;8: 1-158).

RESULTS

Of the 299 dossiers submitted to AIFA, 105 included one or more pharmacoeconomic studies, of which fifty-three were cost-effectiveness analyses. Overall, the compliance of the HEEs with the quality checklist was highly variable: some studies reached high methodological standards whereas others had serious flaws (mean 59.22 percent, range 19.35-90.32 percent). The main weaknesses were the unjustified exclusion of relevant alternatives, poor description and justification of model data and assumptions, and insufficient exploration of uncertainty and study validity. Non-homogeneity across studies was found in study perspectives, discount rates, methods for costing, estimating quality-adjusted life-years and conducting sensitivity analyses.

CONCLUSIONS

Based on the results of this study, the recommended actions for increasing the quality of HEEs within reimbursement submissions in Italy are twofold: first, to set methodological standards for conducting and reporting HEEs; second, to strengthen the internal assessment process, also through the acquisition of companies' models and re-evaluation of results. These actions will hopefully provide greater contribution to the evidence-based P&R decision making.