Seeing What's Important: Eliciting Expert Judgements to Estimate the Minimally Clinically Important Difference for Six-Minute Walking Distance in PH-ILD

OBJECTIVES: A challenge of examining patient outcomes in clinical research is determining if the observed differences constitute a clinically meaningful change. There is currently no estimate for minimally clinically important difference (MCID) for the six-minute walking distance (6MWD) for populations with pulmonary hypertension associated with interstitial lung disease (PH-ILD). This study sought to employ a novel approach to MCID estimation, employing validated techniques to elicit clinical experts’ judgement.

METHODS: After implementing the feedback generated during a pilot exercise with two clinical experts, five experts from the UK and Germany participated in the workshop. A fixed interval method was used to encode judgements. The first elicitation task estimated the mean baseline 6MWD for PH-ILD for orientation to ensure task comprehension, followed by two MCID elicitation exercises. The MCID tasks were interspersed with a facilitated group discussion to examine between-expert variability. Responses were fit to a suitable parametric distribution via least squares estimation, and an aggregate distribution was generated via linear opinion pooling. The MCID was computed as the expected value of the final pooled distribution.

RESULTS: Responses to the first MCID elicitation task indicated estimates ranging between 17.5m and 47.1m (mean=31.9m). Four experts altered their responses during the second task, displaying estimates ranging between 24.8m and 40.6m (mean=31.0m). An ad hoc scenario analysis was conducted, excluding one expert, who reported contradictory comments during the group discussion and in their rationale statements, such as failure to adjust their plausible limits, suggesting insufficient task comprehension. The expected value for the pooled distribution in the scenario analysis was 28.6m (SD=6.9m).

CONCLUSIONS: The elicited values indicate an approximate MCID value lies between 28.6m and 31.0m, falling below estimates for PAH populations. The findings provide a foundation for interpreting clinical trials for PH-ILD-specific therapies.