OBJECTIVES: Health technology assessment bodies are increasingly turning to cost comparison (cost-minimisation) analysis to cope with accelerating demand for appraisals in crowded treatment pathways. The NICE manual requires demonstration of similar effectiveness between technologies to forego a full cost-effectiveness analysis. Head-to-head comparison studies designed as equivalence or non-inferiority studies are usually accepted as providing such evidence. However, there is no guidance for when equivalence may be assumed from an indirect treatment comparison (ITC).

METHODS: We reviewed methods papers and previous NICE appraisals to (1) assess the methods available to determine whether two (or more) treatments are equivalent when no head-to-head trial exists; (2) assess the key considerations taken into account by NICE and its independent review groups to determine whether a cost-comparison route is appropriate; and (3) explore how information on the similarity of health effects can be best presented to inform decision making.

RESULTS: We reviewed 45 methods papers and identified 33 previous NICE appraisals that used a cost comparison approach relying on an ITC to show equivalence. All 33 appraisals were recommended. No appraisals used any methodological approaches described in included methods papers. Rather, companies narratively summarised ITC results and asserted similarity due to there being no (or few) significant differences. In eight appraisals the EAG expressed significant uncertainty regarding the assertion of similarity. The main reasons given were wide credible intervals, heterogeneity, generalisability or missing outcomes. Uncertainty did not, however, result in a delay to appraisal timelines (mean for uncertain appraisals 238 days vs 250 days overall [SD 109]) with submissions overcoming this by focussing on head-to-head comparison or presenting additional cost-utility analysis.

CONCLUSIONS: Methods to formally assess similarity for cost-comparison for ITC are emerging and have not been used in practice. These may help to resolve cases where there is uncertainty around the appropriateness of assuming equivalence.