Related Experiment Videos
The standard gamble demonstrated lower reliability than the feeling thermometer.
Milo A Puhan1, Gordon H Guyatt, Victor M Montori
1Horten Center, University of Zurich, Zurich, Switzerland.
Journal of Clinical Epidemiology
|April 23, 2005
Summary
Clinical marker states (CMS) showed moderate reliability with the feeling thermometer (FT) but poor reliability with the standard gamble (SG). These findings highlight the need for careful selection of reliability statistics in health state valuation.
Area of Science:
- Health Outcomes Research
- Patient-Reported Outcomes
- Psychometrics
Background:
- Clinical marker states (CMS) are used to familiarize participants with preference instruments and ground health state ratings.
- Reliability of CMS ratings is crucial for accurate interpretation of patient-reported outcomes.
Purpose of the Study:
- To assess the reliability of CMS ratings using established statistical methods.
- To evaluate the performance of the feeling thermometer (FT) and standard gamble (SG) in capturing CMS reliability.
Main Methods:
- Eighty-one patients with chronic respiratory disease rated CMSs for mild, moderate, and severe states.
- Ratings were collected using the FT and SG before and after a 3-month rehabilitation program.
- Reliability was assessed using intraclass correlation coefficients (ICC), scatter plots, and Bland-Altman plots.
Main Results:
- Intraclass correlation coefficients (ICC) for FT were 0.47 and for SG were 0.37.
- Scatter and Bland-Altman plots indicated significant between- and within-person variability in ratings.
- Correct ordering of CMS ratings occurred in 64.2% of cases for FT and 11.3% for SG.
Conclusions:
- CMS ratings demonstrate moderate reliability with the FT and poor reliability with the SG.
- The poor reliability of the SG may limit its ability to improve measurement properties in health state valuation.
- Investigators should employ appropriate reliability statistics to ensure the validity of health state assessments.