Poster #10

Category: Junior Faculty Social Science/Clinical/Behavioral

TESTING THE RELIABILITY AND VALIDITY OF THE HEALTH-RELATED QUALITY OF LIFE MEASURE FOR STROKE SUVIVORS

Eun-Shim Nahm, PhD, RN¹; Barbara Resnick, PhD, CRNP, FAAN¹; Kathleen M. Michael, PhD, RN, CRRN²; Marianne Shaughnessy, PhD, CRNP²; Sue Kopunek, BSN, RN²; John Sorkin, PhD, MD²

¹Organizational Systems and Adult Health, School of Nursing, University of Maryland, Baltimore
²The UMB Claude D. Pepper Older Americans Independence Center, BVAMC

Introduction/Purpose:  Health-related quality of life (HRQoL) has become an important outcome in intervention studies for stroke survivors and has been assessed using various measures.  The primary aims of this study were (1) to examine the psychometric aspects of the HRQoL measures included in two stroke exercise intervention studies (the SF-12, Stroke Impact Scale [SIS], and Visual Analog Scale [VAS]; and (2) to explore the most reliable and valid measure using a second-order confirmatory factor analysis measurement model.  (HRQoL as a second-order construct and general, disease-specific, and global HRQoL as first order constructs).

Methods:  This was a single group descriptive study using baseline data.  A total of 164 adults (mean age: 63 + 11.99) with residual hemiparetic motor dysfunction participated in the study.  Reliability testing was conducted estimating the internal consistency and stability and calculating squared multiple correlations (R2) employing Structural Equation Modeling (SEM).  Validity of the measures was assessed using the model fit and structural validity.

Findings: The VAS and all subscales of the SF-12 (PCS-12, MCS-12) and SIS (eight subscales) were internally consistent (alpha coefficients ranging .70 to .87) and stable (ICC ranging .75 to .89, p<.001).  The R² values for most items of the PCS-12, MCS-12 and six subscales of the SIS were less than .5.  The validity coefficients for most items in the SF-12 and SIS were within an acceptable range (»>.5). The validity of the VAS was evidenced by significant correlations with the SF-12 and SIS.  Testing of the second order HRQoL measurement model yielded moderate model fit (Ç2/df = 1.55; GFI = .76; RMSEA = .08; CFI = .82).  The R² values for the SIS, MCS, PCS, and VAS were .93, .81, .65, and 26 respectively.  The validity coefficients were .97, 90, .81, and .51, respectively.

Discussion:  Using the traditional measurement theories, the SIS, SF-12, and VAS were reliable and valid.  Low R² values, however, seem to be associated with potential measurement errors and need further study.  The findings from the second-order measurement model suggest that the SIS is the most reliable and valid measure for assessing HRQoL in the selected samples.  The model, however, must be replicated using different samples.