.png){kind=icon}
.png){kind=icon}
Return
Estimating the probability of experiencing an improvement in randomized controlled trials based on anchor-based MIC studies-An alternative to responder analyses
.png){kind=icon}
OA
AI
W
F
DOI：10.1016/j.cct.2026.108243.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Background: When using a patient reported outcome as primary endpoint in a randomized contolled trial (RCT), responder analyses are widely used to describe the relevance of an observed effect attributable to the intervention of interest. The choice of thresholds in responder analyses is often based on minimal important change (MIC) values. Methods: As an alternative to simply computing MIC values, we suggest determining a translation function allowing to translate change scores values into the probability of experiencing an improvement. The application of such a translation function to the change scores observed in an RCT allows estimating arm-specific probabilities of experiencing an improvement. These values are conceptually similar to responder frequencies. The approach is illustrated using a pair of synthetic studies, both constructed mimicking existing studies. A simulation study investigates the gain in efficiency from avoiding the dichotomization. Results: The illustrative application of the methodology demonstrates that the approach is feasible and allows drawing conclusions in a similar way as in a responder analysis, but with greater precision. The simulation study confirms that this gain in efficiency holds more generally. Considering all response levels of the anchor variable may allow gaining more nuanced insights. Conclusions: Estimating the probability of experiencing an improvement provides an alternative to responder analyses. This approach makes more efficient use of information comprised in the data from anchor-based MIC studies than the translation into MIC values to inform the choice of thresholds in a responder analysis.
Keywords:
Responder analysis
Patient reported outcome measures
Randomized controlled trial
Power
Patient perscpetive
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
1.9
Papers:
197
Citations:
5.7K
Organization
Cited Papers