POSC143 Comparison of Model Structures Used in NICE and ICER Cost Effectiveness Evaluations

Abstract

The Institute for Clinical and Economic Review (ICER) in the United States and the National Institute for Health and Care Excellence (NICE) in United Kingdom produce evidence-based reports on cost-effectiveness of health technologies to inform policymaking and reimbursement decisions. We aimed to compare the degree to which the structure of cost-effectiveness models submitted to NICE and developed by ICER differ. In 2020–2021, ICER published reports in nine disease areas: multiple myeloma, hereditary angioedema, high cholesterol, atopic dermatitis, sickle cell disease, ulcerative colitis, migraine, non-alcoholic steatohepatitis, and beta thalassemia. We identified 17 NICE technology appraisals in the same disease areas published in 2016–2021. Full-text screening of committee papers and ICER reports was conducted by one investigator and abstracted data was validated by a second. Information on treatment comparators, model structure, and approach were extracted. Evidence review group commentaries were considered. Comparisons were based on the adopted model structure and approach to elicit inputs. Data sources and results were not relevant. Among the disease areas, ICER and NICE used different model structures in three, broadly similar structures in four, and similar structures in two evaluations. Among the models with different structures, NICE were more complex with larger numbers of health states (2/3). Among models with broadly similar structures, NICE implemented approaches that more closely captured the effect of new treatments (2/4) with more complex data elicitation approaches (3/4). Among the models with very similar model structures, the key difference was the means of treatment effect measure. Despite NICE and ICER having similar guidelines for conducting cost-effectiveness assessments, there are differences among model structures within the same disease areas. Availability of trial patient-level data to inform submissions to NICE and manufacturers’ intent to capture treatment effect to the best possible extent could be reasons behind the discordance.