OBJECTIVES: The UK's National Institute for Health and Care Excellence has proposed to align the health economic methods utilized for both pharmaceuticals and medtech in evidence submissions, suggesting a requirement for cost-utility analyses (CUAs) to attain reimbursement. It is unclear whether there is sufficient volume of HSUV data for medtech to robustly populate CUAs. The objective of this research was to conduct a TLR of HSUVs in T2DM, as a case study, to understand what data exist for medtech and DH relative to pharmaceutical interventions.

METHODS: MEDLINE, Embase and the International HTA Database were searched in October 2023 for data published since 2020, using terms for T2DM and HSUVs. Records were reviewed against prespecified eligibility criteria wherein all studies presenting novel HSUV data were included for analysis. CUAs were also retained to compare the volume of models published on medtech and pharmaceuticals.

RESULTS: 1,999 records identified from database searches were screened, yielding a total of 41 studies reporting on novel HSUVs, and 173 CUAs. Of the HSUV studies, 8 (20%) reported on medtech, comprising DH (n=6) and metabolic surgery (n=2) relative to 33 (80%) reporting on pharmaceuticals. EQ-5D was predominant, utilized in 37/41 (90%) studies. DH included smartphone apps (n=3), telemedicine (n=2) and a website (n=1), and 5/6 (83%) collected HSUV data through randomized controlled trials, while 24/33 (73%) studies of pharmaceuticals employed prospective observational designs. The identified CUAs were overwhelmingly conducted in pharmaceuticals (n=141), followed by DH (n=13), diagnostics (n=12) and surgery (n=7).

CONCLUSIONS: Whilst HSUVs have been elicited for DH specifically, the volume of data is small relative to pharmaceutical interventions in T2DM. The large disparity in the number of CUAs may suggest that inputs available for economic models of DH are lacking, and that primary data collection will be required of manufacturers to elicit HSUVs suitable for use in economic modelling.