[Over- and Underestimation of Success Rates]

Area of Science:

• Clinical research methodology
• Psychometrics
• Biostatistics

Background:

• Patient-reported outcomes (PROs) in interventional studies are crucial for assessing treatment effects.
• Responder analysis, classifying individuals based on minimal important difference (MID), is widely used but susceptible to misclassification errors.
• Low reliability of change scores often leads to substantial over- or underestimation of responder proportions.

Purpose of the Study:

• To introduce a novel, simple method for estimating the true proportion of responders in interventional studies.
• To address and correct for misclassifications arising from unreliable outcome differences.
• To provide a more accurate assessment of treatment efficacy based on patient-reported outcomes.

Main Methods:

• The proposed method is grounded in the principles of classical test theory.
• It involves estimating the true proportion of responders, accounting for measurement error.
• The method's application is demonstrated using empirical data from interventional studies.

Main Results:

• The study highlights that misclassifications due to unreliable differences can be substantial and do not typically cancel out.
• The new method provides a more accurate estimation of the true responder proportion compared to traditional approaches.
• Empirical data confirmed the significant impact of misclassification and the utility of the proposed estimation technique.

Conclusions:

• It is recommended to adopt the new method for reporting the true proportion of responders in clinical research.
• This approach enhances the accuracy of treatment effect evaluation in both single-group and control-group study designs.
• Accurate responder estimation is vital for reliable interpretation of patient-reported outcomes and intervention effectiveness.