Testing the SF-36 in Parkinson's disease. Implications for reporting rating scale data

Area of Science:

• Clinical Epidemiology
• Health Outcomes Research
• Psychometrics

Background:

• Rating scales are primary outcomes in clinical trials, but their interpretation requires validated scaling assumptions.
• The SF-36 is a widely used patient-reported outcome measure, but its psychometric properties in Parkinson's disease (PD) are underexplored.

Purpose of the Study:

• To assess the scaling assumptions, reliability, and validity of the SF-36 in patients with Parkinson's disease (PD).
• To provide evidence-based guidelines for the appropriate use of the SF-36 in PD clinical research.

Main Methods:

• Analysis of SF-36 data from 202 Parkinson's disease (PD) patients.
• Evaluation of scaling assumptions, reliability (Cronbach's alpha), and validity of the eight SF-36 scales and two summary measures.
• Assessment of floor/ceiling effects for detecting change.

Main Results:

• SF-36 scale scores were generally reliable (alpha >= 0.74), supporting score generation.
• Limitations were found in four scales regarding meaningfulness and in four scales regarding change detection (19.6-46.2% floor/ceiling effects).
• The two SF-36 summary measures were not valid indicators of physical and mental health in this PD cohort.

Conclusions:

• The SF-36 demonstrates significant limitations for use in Parkinson's disease (PD) research, particularly concerning its summary measures.
• Findings suggest the need for routine psychometric evaluation of rating scales within specific patient populations in clinical trials.
• Investigators should report scale evaluations to ensure the meaningful interpretation of rating scale endpoints in research.