ABSTRACT
OBJECTIVE: The cost-effectiveness threshold (CET) is a key parameter to guide objective reimbursement decisions, yet very few countries have defined a reference CET, and there is no reference method for defining it. Our objective was to determine the factors explaining the author-reported CETs in the literature. METHODS: Our systematic review targeted original articles referenced in EMBASE and published between 2010 and 2021. Selected studies had to use Quality-Adjusted Life-Year (QALY), and being conducted in high-income countries. Our explanatory variables were: estimated cost-effectiveness ratio (ICER), region of the world, source of funding, type of intervention, disease, year of publication, justification of the author-reported Cost-Effectiveness Threshold (ar-CET), economic perspective, and declaration of interest. Multivariable linear regression models implemented on R software were used, guided by a Directed Acyclic Graph. RESULTS: Two hundred and fifty four studies were included. The mean ar-CET was 63,338/QALY (standard deviation (SD) 34,965) overall, and 37,748/QALY (SD 20,750) in studies conducted in the British Commonwealth. The ar-CET increased slightly with the ICER (+ 66/QALY for each additional 10,000/QALY in the ICER, 95% confidence interval (IC) [31-102], p < 0.001), was higher in the United States (+ 36,225/QALY; IC [25,582; 46,869]) and Europe (+ 10,352/QALY; IC [72; 20,631]) compared to the British Commonwealth (p < 0.001), and was higher when the ar-CET was not defined a priori (+ 22,393/QALY; [5809; 38,876]) compared to state recommendations defined ar-CET (p < 0.001). CONCLUSIONS: Our results underline the virtuous role of state recommendations in the choice of a low and homogeneous CET. We also highlight the need to integrate the a priori justification of the CET into good publishing guidelines.
