Ford ME, Havstad SL; Association for Health Services Research. Meeting.
Abstr Book Assoc Health Serv Res Meet. 1999; 16: 344.
Henry Ford Health System, RCMAR, Detroit, MI 48202, USA.
RESEARCH OBJECTIVE: Managed care organizations rely on practice guidelines and public policies. These guidelines and policies are based upon research using exisiting measurement instruments to assess physical health and mental health outcomes. However, racial/ethnic group differences may exist in the structure and measurement of these outcomes. This paper examines the psychometric properties of the SF-12 Scale and the National Chronic Care Consortium Health Risk Appraisal Measure (NCCCHRAM) among older (aged 50+ years) African American and Caucasian adults in a large, vertically integrated health system with a managed care component. STUDY DESIGN: This was a cross-sectional study of 400 African American and Caucasian adults aged 50+ years randomly selected from among patients at Henry Ford Health System (HFHS) in Detroit, Michigan who made at least one visit in 1997. The SF-12 Scale was developed by Ware et al. and consists of two components measuring physical health (PCS-12) and mental health (MCS-12). The reliabilities of the PCS-12 and MCS-12 are comparable to the reliabilities of the eight SF-36 scales, which range from 0.63 to 0.89 (median, 0.80). The NCCCHRAM was developed by members of the National Chronic Care Consortium. This instrument includes measures of instrumental activities of daily living (IADLs) and activities of daily living (ADLs). There are no published reliability and validity data for this instrument. PRINCIPAL FINDINGS: No statistically significant differences were found in comparisons of mean age or gender among the respondents in the two racial/ethnic groups (n=231, 58% response rate). Differences in age, race, and gender among respondents and non-respondents were also founds to be nonsignificant. In the combined racial/ethnic group, confirmatory factor analysis of the SF-12 (n=212) showed that Bentler's Comparative Fit Index was 0.83. The Bentler & Bonett Non-normed Index was 0.79. For the African Americans, confirmatory factor analysis of the SF-12 (n=106) showed that Bentler's Comparative Fit Index was 0.81, and the Bentler & Bonett Non-normed Index was 0.76. Among African Americans, the reliability of the PCS-12 and MCS-12 were, respectively, 0.86 and 0.87, as measured by Cronbach's alpha. For the Caucasians, confirmatory factor analysis of the SF-12 (n=108) showed that Bentler's Comparative Fit Analysis was 0.76, and the Bentler & Bonett Non-normed Index was 0.70. Among the Caucasians, the reliabilities of the PCS-12 and the MCS-12 were, respectively, 0.87 and 0.81, as measured by Cronbach's alpha. In all of the SF-12 models, the chi-square goodness of fit test was significant (p<0.001). Exploratory factor analysis of the IADL scale in the combined racial/ethnic group (n=212) showed that a unifactorial model could be constructed by combining into one factor those items who communality estimates were greater than 0.45. This model would include shopping and errands, light housekeeping, doing laundry, preparing meals, and using transportation. In subgroup analysis of the IADL scale by racial/ethnic group, exploratory factor analysis showed the same results as were found in the combined racial/ethnic group analysis; the reliability if the IADL was 0.81 in the combined racial/ethnic group, among the African Americans only, and among the Caucasians only. Exploratory factor analysis of the ADL scale in the combined racial/ethnic group (n=214) showed that a unifactorial model could be constructed by combining into one factor those items whose communality estimates were greater than 0.45. This model would include bathing, dressing, getting out of bed, and getting out of a chair. For the African Americans, exploratory factor analysis (n=106) showed that only two items had communality estimates greater than 0.45. In contrast, for the Caucasians, exploratory Factor analysis (n=108) showed that a unifactorial ADL model could be constructed. This model would include bathing, dressing, getting out of bed, getting out of a chair, walking inside the home, and using the toilet. The reliability of the ADL was 0.78 in the combined racial/ethnic group, and 0.81 for the Caucasians. CONCLUSIONS: These findings suggest that the SF-12 and IADL scale of the National Chronic Care Consortium Health Risk Appraisal Measure show the same factor structure when used with older African Americans or with older Caucasians. Thus, these instruments appear to measure the same constructs in each racial/ethnic group. In addition, both instruments show fairly high levels of internal consistency when used with either group. However, the ADL scale appeared to have a different factor structure for the African Americans and Caucasians. IMPLICATIONS FOR POLICY, DELIVERY OR PRACTICE: Clinical decision-making algorithms and public policy are typically based on the results of outcomes research using measurement instruments. (ABSTRACT TRUNCATED)
Publication Types:
Keywords:
- Activities of Daily Living
- Adult
- African Americans
- Continental Population Groups
- Cross-Sectional Studies
- Cultural Diversity
- Ethnic Groups
- European Continental Ancestry Group
- Factor Analysis, Statistical
- Health Status Indicators
- Humans
- Michigan
- Research Design
- rehabilitation
- hsrmtgs
Other ID:
UI: 102194145
From Meeting Abstracts