A Measurement Model of the Medical Outcomes Study 36-Item Short-Form Health Survey in a Clinical Sample of Disadvantaged, Older, Black, and White Men and Women

Datos Bibliográficos

ID9102675
AutoresFredric D Wolinsky (0000-0002-0916-4955, Regenstrief Institute), Timothy E Stump (0000-0001-9593-3011, Regenstrief Institute)
Año1996
Volumen34
Número6
Páginas537-548
Fecha de publicación1996-06-01
Peer ReviewedSí
Open AccessNo
TipoARTICLE
RevistaMedical Care (JOURNAL)
Identificadores de la revistaISSN: 0025-7079 • E-ISSN: 1537-1948
EditorialOvid Technologies (Wolters Kluwer Health) (PUBLISHER)
DOI10.1097/00005650-199606000-00004
PMID8656720
OpenAlexW2322902774
IdiomaEN
Citas recibidas17
Referencias citadas13

The authors assess the factorial validity of the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) for use in a clinical sample of disadvantaged, older adults with significant comorbidities. Confirmatory factor analysis was performed using LISREL VIII on data obtained from baseline face-to-face interviews with a clinical sample of 1,051 patients who were at risk for acute deterioration of their clinical condition due either to their age alone (75 years or older), or to their age (50 to 74 years old) and major comorbid conditions. An acceptable eight-factor measurement model reflecting the original specification (ie, subscales) of the SF-36 was obtained (chi-square to degrees of freedom ratio = 2.14; root mean squared residual = .055; adjusted goodness of fit index = .90). That model, however, required relaxing the assumptions associated with seven correlated error terms. Moreover, an alternative nine-factor model that allowed the ¿getting sick¿ and ¿getting worse¿ items to form their own factor, labeled ¿health optimism,¿ fit the data significantly better (8 degrees of freedom chi-square improvement = 61; P< 0.0001). Although continued use of the SF-36 in older, disadvantaged, clinical samples is appropriate, further assessment of the underlying measurement model in other samples using confirmatory factor analytic techniques is needed to resolve the issue of correlated error structures and the existence of the health optimism factor

Confirmatory factor analysis · Disadvantaged · Goodness of fit · LISREL · Optimism · Statistics · Structural equation modeling · Clinical Psychology · Health and Wellbeing Research · Health disparities and outcomes · Healthcare Policy and Management · Mathematics · Medicine · Psychology · Social Psychology · Gerontology