Unadjusted effect modifiers (EM) in indirect treatment comparisons (ITC) can produce biased and uncertain results in health technology assessments (HTA). Even though research on advanced ITC methods to adjust for EMs has attracted much attention, less emphasis has been placed on how these EMs are identified and assessed in the first place. This is surprising given the evidence-based approach underpinning all analyses supporting HTAs. To address this gap, our aim was to identify what HTA guidance exists on the selection process for EMs, and how the selection of EMs has been justified in the context of published ITCs.

A pragmatic review of HTA guidance documents was conducted in 2021 to describe current requirements for the selection of EMs for ITCs. A supplementary Embase and Medline search was conducted to identify primary research on ITCs published between 2015-2021 presenting information on how EMs were selected to inform these analyses.

Our review found that guidance on this topic focusses on developing and testing different methodologies to adjust for EMs. No detailed guidance was identified in any of the reviewed HTAs, although the National Institute for Health and Care Excellence (NICE) briefly mentioned that companies should identify EMs through a topical review or expert discussions. Similar findings were also revealed through the database search; few published ITCs included information on the EM selection process which was either based on evidence highlighted in the literature or findings from prior trial subgroup analyses. No reference to a systematic identification of EMs was found.

To fill the guidance gap identified in our review an extension of current ITC guidelines (including those from HTA bodies) is needed, including (i) indication on how EMs should be identified through systematic reviews, (ii) a quantitative assessment of the EM distribution and (iii) formal expert elicitation prior to the selection of ITC methods. Without these additional steps, ITC results may be biased, potentially negatively impacting decision-making and ultimately patient care.