Internal consistency reliability is a poor predictor of responsiveness.

Area of Science:

• Health Outcomes Research
• Psychometrics
• Clinical Trial Methodology

Background:

• Responsiveness is a key measurement property for health-related quality of life (HRQL) instruments.
• Debate exists on whether responsiveness is distinct from reliability and validity.
• A recent study suggested internal consistency reliability could predict instrument responsiveness.

Purpose of the Study:

• To evaluate the claim that internal consistency reliability can predict instrument responsiveness.
• To assess the relationship between internal consistency reliability, instrument type, and responsiveness in HRQL instruments.

Main Methods:

• 516 patients with COPD or knee injury were included in four longitudinal studies.
• Patients completed generic and disease-specific HRQL questionnaires pre- and post-intervention.
• Pearson correlation coefficients and linear regression analyzed the relationship between Cronbach's alpha, instrument type, and standardized response mean (SRM).

Main Results:

• The mean Cronbach's alpha was 0.83 and the mean SRM was 0.59.
• Correlation between Cronbach's alpha and SRM was weak (0.10) and not statistically significant.
• Internal consistency reliability (Cronbach's alpha) did not predict responsiveness (SRM); instrument type was a significant predictor.

Conclusions:

• Internal consistency reliability is not a sufficient indicator of HRQL instrument responsiveness.
• Investigators should consider other measurement properties beyond Cronbach's alpha when selecting instruments.
• Instrument type is a crucial factor influencing responsiveness in HRQL measures.