Considering minimal clinically important differences in assisted reproductive technology

Objective: To address the reporting of minimal clinically important differences (MCIDs) in the published literature and investigate the utility of absolute and relative differences when defining this parameter. Design: Expert opinion. Subjects: Not applicable. Exposure: Not applicable. Main Outcome Measures: It is essential to consider both statistical significance and clinical significance when interpreting study findings. Key to this is establishing the MCID that is beneficial for patients. In the context of assisted reproductive technology, several benchmarks are used to evaluate differences between treatments, including the mean number of retrieved oocytes, pregnancy rates, clinical pregnancies, live births, and cumulative live births. Results: Determining the MCID for assisted reproductive technology procedures is a subjective process that can be influenced by several factors, depending on whether an absolute or relative difference is selected. Furthermore, various and often overlapping MCIDs have been used in different studies, meaning that the same difference between treatment and comparator can be interpreted as evidence of superiority in some studies and as evidence of noninferiority in others. To address these inconsistencies, we recommend that comparative studies should, by design, include a clearly defined and justified MCID threshold, with differences expressed as relative measures. Conclusion: We recommend that the MCID should be defined in advance and expressed as a relative measure, which can be interpreted across various groups of women with diverse prognoses. However, absolute measures should also be reported for completeness.