OBJECTIVES: Randomised controlled trials (RCTs) using comparator arms in which treatment is determined by physician- or investigator- choice (IC) are increasingly common. This study sought to review current practice in the utilisation and assessment of such trials in health technology assessment (HTA).

METHODS: NICE technology appraisals with a final draft guidance between 2022-2024 were searched for variations of the terms “investigator’s choice” and “physician’s choice”. Appraisals were screened to determine whether RCTs with IC arms were included. For NICE appraisals identified, equivalent assessments were identified in Canada, Germany and France. Data on how the IC trial data were utilised and assessed in each assessment were extracted.

RESULTS: Five relevant NICE assessments were identified. Equivalent assessments were identified in all but one case in which there was no assessment identified in Canada. In 11 of these 19 assessments, HTA bodies stated that the IC arm did not capture all relevant comparators. However, this was not considered a major issue as most comparators were noted to have low prevalence of use, off-label use only, or were relatively new to the market. In 6/19 assessments, the IC arm was deemed to include some comparators with little/no relevance to clinical practice. For both of the German assessments in which this was the case, a post-hoc analysis excluding the irrelevant comparators was requested and carried out. In one assessment in England, the manufacturer chose to pro-actively present such a post-hoc analysis. In the other 3 cases (2x England, 1x Canada) a similar post-hoc analysis was not requested and it was assumed the irrelevant comparators had no impact on efficacy.

CONCLUSIONS: HTA bodies appear to be generally accepting of trials in which IC arms don’t fully reflect clinical practice. Where feasible, analyses excluding comparators deemed irrelevant to practice should be planned, particularly in Germany.