The importance of minimal important differences to inform systematic reviews and clinical practice guidelines: an example




Short oral session 2: Considerations for meta-analyses


Wednesday 13 September 2017 - 11:00 to 12:30


All authors in correct order:

Devji T1, Guyatt G1, Lytvyn L2, Brignardello-Petersen R1, Foroutan F1, Sadeghirad B1, Buchbinder R3, Harris I4, Carrasco-Labra A1, Siemieniuk R1, Vandvik P5
1 McMaster University, Canada
2 Oslo University Hospital, Canada
3 Monash University, Canada
4 Ingham Institute for Applied Medical Research, Canada
5 University of Oslo, Canada
Presenting author and contact person

Presenting author:

Tahira Devji

Contact person:

Abstract text
Background: Investigators increasingly rely on patient-reported outcome measures (PROMs) as key end points in clinical trials. However, interpretation of the magnitude of treatment effects on PROMs presents challenges. The smallest change that patients perceive as important – the minimal important difference (MID) – can enhance the interpretation of PROMs. We present an example in which we identified credible MIDs to facilitate understanding of the importance of intervention effects in a meta-analysis.

Objectives: To identify credible anchor-based MIDs for the PROMs used in trials comparing arthroscopic surgery to conservative management; describe our approach to gathering and interpreting the credibility of MID estimates; and, show how our results informed the linked systematic review (SR) and subsequent development of the BMJ Rapid Recommendation (RapidRec).

Methods: We searched MEDLINE, EMBASE and PsycINFO for studies documenting the development of anchor-based MIDs for PROMs reported in trials included in the linked SR and judged by the parallel BMJ RapidRec panel as critically important for informing their guideline: measures of pain, function and health-related quality of life. We assessed the credibility of MIDs by focusing on the correlation between change in the PROM under consideration and the anchor. The SR and guideline authors used the credible MIDs for each PROM to interpret their results.

Results: We were able to distinguish between more and less trustworthy MIDs and provide best estimates for key instruments that informed evidence presentation in the associated meta-analysis of treatment effects, and judgments by the BMJ RapidRec panel. Using the MIDs, the panel judged that arthroscopy had only a trivial-to-very small impact on short-term knee pain and function. The MIDs allowed the panel to weigh the magnitude of benefit against the harms of arthroscopy. In doing so, the panel was confident making a strong recommendation against knee arthroscopy.

Conclusions: Our study provides a model for applying the MID concept to aid in the interpretation of evidence, and the formulation of recommendations for clinical practice guidelines.