A systematic survey on reporting and methods for handling missing participant data for continuous outcomes in...

Area of Science:

• Clinical Research Methodology
• Biostatistics

Background:

• Missing participant data (MPD) is common in randomized controlled trials (RCTs).
• Patient-important continuous outcomes are frequently analyzed in clinical research.

Purpose of the Study:

• To evaluate the analytical strategies employed by RCT authors for handling missing participant data (MPD).
• To assess the methods used for continuous patient-important outcomes in RCTs.

Main Methods:

• Systematic survey of RCTs published in core clinical journals in 2014.
• Included trials reported at least one patient-important outcome analyzed as a continuous variable.

Main Results:

• Over 93% of RCTs reported on the occurrence of MPD, with a median of 11.4% missing data.
• Commonly used methods included available data analysis (67%), while optimal methods like multiple imputation were rare (4.5%).
• Only 9.8% performed sensitivity analyses for MPD impact, and 11.1% discussed bias risks.

Conclusions:

• RCTs with continuous outcomes frequently exceed 10% MPD.
• Suboptimal analytical methods for MPD are prevalent in published RCTs.
• There is a lack of sensitivity analyses and bias risk discussions related to MPD.