Testing measurement equivalence in a patient satisfaction instrument

Area of Science:

• Psychometrics
• Health Services Research
• Statistical Modeling

Background:

• Patient satisfaction is a key metric in healthcare quality assessment.
• Ensuring measurement consistency across diverse patient groups is crucial for valid comparisons.
• Previous research has highlighted potential biases in patient-reported outcomes.

Purpose of the Study:

• To evaluate measurement equivalence of a patient satisfaction scale across different patient demographics and time points.
• To examine configural, metric, scalar invariance, invariant uniquenesses, and invariant factor variances.
• To determine the scale's suitability for cross-group comparisons in gender and race.

Main Methods:

• Employed confirmatory factor analysis (CFA) and structural equation modeling (SEM).
• Assessed five sources of measurement equivalence: configural, metric, scalar invariance, invariant uniquenesses, and invariant factor variances.
• Utilized a sample of 1,897 patients, with subgroup analyses for gender and race.

Main Results:

• The 10-item Likert-type scale demonstrated good performance and reliability across different measurement times.
• Measurement equivalence was largely supported over time.
• Significant differences were found when comparing satisfaction between genders and racial groups, indicating a need for caution.

Conclusions:

• The patient satisfaction scale is reliable for longitudinal measurement.
• Adjustments are necessary to ensure valid comparisons of patient satisfaction between genders and racial groups.
• Further methodological research is needed to refine patient-reported outcome measures for diverse populations.