Abstract

Purpose: The Medical Outcomes Study (MOS) 36-Item Short-Form Health Survey (SF-36) has been validated in many diverse samples. This measure of health-related quality of life, however, has not yet been examined among older Mexican Americans, a rapidly growing subset of the older population. Design and Methods: We address the validity of the SF-36 in a sample of older Mexican Americans (N = 621) selected from the Hispanic Established Populations for Epidemiologic Study of the Elderly. Using confirmatory factor analysis through structural equation modeling, we evaluate the construct validity of this scale. Results: The results indicate evidence for a model with eight first-order factors consistent with previous research on the SF-36 and two second-order factors representing mental and physical health. Implications: This, in addition to other evidence given here, leads us to the conclusion that the SF-36 is a valid measure of health-related quality of life in this sample of older Mexican Americans.

The Medical Outcomes Study (MOS) 36-Item Short-Form Health Survey (SF-36) was designed to be a generic health-status measure and address function in eight domains of health: physical function, role physical, bodily pain, general health, vitality, social function, role emotional, and mental health (McHorney, Ware, & Raczek, 1993; Simon, Revicki, Grothaus, & Vonkorff, 1998; Ware & Sherbourne, 1992). The SF-36 has gained acceptance for addressing changes in health status, evaluating medical treatments, and assessing health-related quality of life (HRQOL) in a wide variety of samples (Simon et al., 1998; Walters, Munro, & Brazier, 2001). The SF-36 was not created to be age, treatment, or disease specific (Benjamin-Coleman & Alexy, 1999; Ware et al., 1998; Ware & Sherbourne, 1993) and has been validated with various diverse samples, including patients (e.g., McHorney et al., 1993; Ware & Sherbourne, 1992), respondents from European countries (Keller et al., 1998; Ware et al., 1998), and Spanish-speaking adults (Arocho, McMillan, & Sutton-Wallace, 1998; Ayuso-Mateos, Vazquez-Barquero, Oviedo, & Diez-Manrique, 1999).

Although the SF-36 as a HRQOL measure has been used extensively in comparing the health of different samples, assessing health needs, and evaluating the effects of medical interventions, a significant and important part of the population has been understudied with respect to the use of the SF-36: older adults (Walters et al., 2001). Specifically, it is important to examine the use of the SF-36 in older adults because (a) outcomes of mortality and morbidity are insufficient to capture issues related to independence and quality of life in older adults (Stadnyk, Calder, & Rockwood, 1998); (b) information is needed on self-rated health because a decline could show early stages of more serious morbidity (Martikainen, Stansfeld, Hemingway, & Marmot, 1999); (c) population-based measures of health status are important for comparisons across subgroups of older adults (e.g., Lyons, Crome, Monaghan, Killalea, & Daley, 1997); and (d) as mentioned earlier, the SF-36 has become a widely used and validated measure of generic health status (McHorney, 1996). Though some research has examined the validity of the SF-36 among older British adults (e.g., Lyons, Perry, & Littlepage, 1994; Walters et al., 2001), there is no research that we are aware of that focuses on older minority adults. Although the SF-36 has been validated in general Spanish-speaking samples and in older adults, the validity has not yet been examined among older Mexican Americans.

Although only 5.5% of Hispanics are 65 or older, their numbers and proportions are expected to grow rapidly in the near future. By the year 2020, the Hispanic older population will grow by 76%, compared with a growth rate of 38% for non-Hispanic White older adults and 34% for African American older adults (Day, 1996). Of the Hispanic older population, approximately 66% are of Mexican origin (Therrien & Ramirez, 2000). Establishing the validity of the SF-36 in this population is essential to help monitor progress toward the Healthy People 2010 goals of improving health and independence among older minority adults.

Moreover, investigating HRQOL could potentially shed light on the connection between acculturation and health. There is a concept referred to as the “Hispanic paradox” or the “epidemiologic paradox” (Franzini, Ribble, & Keddie, 2001; Markides & Coreil, 1986), which posits that Mexican Americans have low socioeconomic status and yet have health and mortality profiles similar to or, in some instances, better than Whites (Franzini et al., 2001). Two explanations for why the Hispanic paradox exists include the healthy migrant hypothesis and a cultural or acculturation hypothesis. The healthy migrant hypothesis suggests that it is typically the strongest and healthiest members of a population who migrate (Franzini et al., 2001). Thus, older adults who migrated are more likely to be healthy later in life. The acculturation hypothesis argues that aspects of Mexican culture or being assimilated into larger American society influences health (Franzini et al., 2001). The evidence is mixed, showing health-protective effects of acculturation in some studies (e.g., Hazuda, Haffner, Stern, & Eifler, 1988) and detrimental effects in others (e.g., Scribner & Dwyer, 1989). Finding a reliable and valid measure of HRQOL could facilitate the investigation into the connection between acculturation, HRQOL, and health. Though the focus of this research is on the reliability and validity of the SF-36 among older Mexican Americans and not on the relationship between acculturation and HRQOL, we discuss potential acculturation explanations for our findings.

Finally, there are other measures of HRQOL among older adults, including the Index for Health-Related Quality of Life (IHQL; Livingston, Watkin, Manela, Rosser, & Katona, 1998), the Geriatric Quality of Life Questionnaire (GQLQ; Guyatt et al., 1993), and many measures assessing life satisfaction. There are two primary reasons for focusing on the SF-36 in this research. First, the SF-36 is the most widely used scale for assessing HRQOL. Second, the majority of the respondents in the survey that we use are Spanish speaking, and the SF-36 has been translated into Spanish.

To address the validity of the SF-36 in an older Mexican American sample, we (a) provide population scores for a sample of community-dwelling older Mexican American adults (i.e., 65 years of age or older); (b) assess internal consistency or reliability of each of the eight health-dimension subscale scores; and (c) assess the construct validity through an examination of the factor structure of the SF-36.

Methods

Data

The sample for the current study is a subsample from the Hispanic Established Populations for Epidemiologic Study of the Elderly (EPESE). The Hispanic EPESE is a population-based study of 3,050 noninstitutionalized Mexican American individuals aged 65 and older at baseline (1993–1994) from five Southwestern states (Texas, California, New Mexico, Colorado, and Arizona; see Markides et al., 1996; Rudkin, Markides, & Espino, 1997). Four waves of data have been collected (1993–1994, 1995–1996, 1998–1999, and 2000–2001). A multistage area probability sampling design was used and modeled after the earlier EPESE studies from New Haven, East Boston, North Carolina, and rural Iowa.

In Stage 1, counties were selected if at least 6.6% of the county population consisted of older Mexican Americans (using U.S. Census figures). Stage 2 included the selection of 300 randomly chosen census tracts, whereas in Stage 3 blocks were randomly selected. In this stage, one to two additional blocks were added to obtain at least 400 households within each sampling unit. Finally, in Stage 4, interviews were conducted with up to four members of the Mexican American household aged 65 and older (Markides et al., 1996). Approximately 85% of Mexican American older adults live in these five Southwestern states, and the sampling design ensured a sample generalizable to approximately 500,000 older adults (for further discussion of the sampling design, see Markides et al., 1996; Rudkin et al., 1997).

The older adults in the original sample have lower income and higher disability rates than comparable older non-Hispanic Whites (Markides et al., 1999). The response rate at baseline (1993–1994) was 83% (n = 2,873 interviews in person; n = 177 interviews by proxy). The participants were interviewed and screened in their own homes by trained interviewers. Interviewers were trained by Harris Interactive, Inc. staff and by Hispanic EPESE investigators, who provided training on measuring blood pressure and making performance-based assessments of physical functioning and physical and social characteristics.

Harris Interactive, Inc. interviewers followed up the original 3,050 participants at approximately 2-year intervals. Live interviews at Wave 2 were conducted with 2,439 participants (80%). Of these, 143 were proxy interviews (5.9%). At Wave 3, 1,979 respondents were reinterviewed, including 202 proxy interviews. Finally, at Wave 4, 1,687 respondents were reinterviewed with 143 proxies.

Subsample

After Wave 3 was collected, we created a list of respondents who were alive at the end of Wave 3 (n = 1,979) and who reported having Medicare coverage at either Wave 1 or Wave 2 (n = 1,598). This represents approximately 81% of the sample at Wave 3. From this group of respondents we randomly selected 800 to be the sample for a substudy focusing on the link between acculturation, disability, and HRQOL (under Grant RO1-AG17638). We chose respondents who had Medicare coverage (approximately 95%) because of the intent of the investigators to ultimately link the substudy data with Medicare claims data. The substudy will collect data on these respondents over two waves. The substudy piggybacked with the Hispanic EPESE on Wave 4, where the respondents selected for the substudy had additional interview questions.

Of the 800 respondents selected, 621 participants completed the interviews. The remaining 178 respondents included refusals to participate and proxy interviews. We did not allow proxy interviews because of the physical nature of some of the measurements in the substudy. The respondents in the substudy were 71 years of age and older.

Measures

The Appendix shows the eight domains of health in the SF-36. This table provides a brief description of each dimension, the number of items in each dimension, and sample items. The dimensions are physical function, role physical, bodily pain, general health, vitality, social function, role emotional, and mental health (McHorney et al., 1993).

The SF-36 was provided to the respondents in either Spanish or English. Approximately 83% chose to complete the interview in Spanish. The SF-36 has been translated into Spanish by use of the methodology of the International Quality of Life Assessment Project (IQOLA), which included forward and backward translation by qualified individuals. (For more information on the translation of the SF-36, see the SF-36 Health Survey Manual and Interpretation Guide, Ware, Snow, Kosinski, & Gandek, 1993.)

Analysis

To address the psychometric properties of the eight dimensions of the SF-36 in older Mexican Americans, first we provide the mean and standard deviation scores for two age-specific groups: 71–74 and 75 and older. We compare these scores with national norms of the same age groups taken from the SF-36 Health Survey Manual and Interpretation Guide (Ware et al., 1993). The U.S general population norms were estimated from a 1990 cross-sectional, nationally representative survey that included the SF-36 (for details, see Ware et al., 1993).

Second, we assess the internal consistency of the eight dimensions of the SF-36. The internal consistency refers to the extent of correlation of items with other items in the same dimension. We assess this by using Cronbach's alpha, which is a summary measure ranging between 0 and 1 for each dimension. Typically, values above 0.8 are considered acceptable (Carmines & Zeller, 1979).

Third, we address the construct validity by examining the factor structure of the SF-36 in older Mexican Americans, consistent with research on the construct validity of this scale conducted for the IQOLA, which focused on standardizing comparisons of health across different countries (e.g., Keller et al., 1998; Ware et al., 1998). We use structural equation modeling, AMOS 4.01, to assess the construct validity, which assesses the extent to which a measure performs in a manner consistent with theoretical expectations (Carmines & Zeller, 1979). Thus, we fit a measurement model that includes eight first-order factors representing each of the eight SF-36 dimensions described in the Appendix. The model will also simultaneously fit two second-order factors representing mental and physical factors, consistent with research conducted by Keller and colleagues (1998) and Ware and colleagues (1998) and originally proposed by McHorney and colleagues (1993). There will be positive evidence for construct validity if the fit of the model is acceptable.

We evaluate the goodness of fit of the model to the data by using the normed fit index (NFI), the comparative fit index (CFI), and the root mean square error of approximation (RMSEA; see, e.g., Bentler, 1990; Kline, 1998). We define model fit as (a) the NFI and CFI are greater than.90; (b) the RMSEA is less than.10; and (c) the parameter estimates or the factor loadings for each of the dimensions. In other words, there will be evidence for the model fitting if the variables load on the factors that they are hypothesized to comprise.

Results

Table 1 shows selected demographic and health characteristics of the sample. The mean age of the members of the sample is 78.3 years; approximately 59% are female. The mean years of education are low, similar to the larger Hispanic EPESE sample, at about 5 years of education completed. The distribution by state is also similar to the Hispanic EPESE, with predominant proportions being in Texas and California. Approximately 19% of the members of the sample report at least one activity of daily living (ADL) limitation, such as needing help with bathing, dressing, or eating because of a health problem. Finally, the incidence rates of heart attack, stroke, and hip fracture in the past 2 years are relatively low, ranging from around 3% to 4%.

Score Comparisons

Table 2 shows the mean and standard deviation scores for two age-specific groups, 65–74 years and 75 years and older, in our sample of older Mexican Americans. Table 2 also includes scores from the national norms of the same age groups taken from the SF-36 Health Survey Manual and Interpretation Guide (Ware et al., 1993). The age group of 65–74 years is mismatched because the respondents from the subsample of the Hispanic EPESE are 71 years and older. Thus, the age group of 65–74 years represents those respondents ages 71–74 years for the Mexican Americans.

In comparing the younger age groups, we see that the members of the Mexican American sample have higher scores on all dimensions of the SF-36 except for physical functioning (M = 65.42, Mexican Americans; M = 69.38, national) and general health perceptions, where the scores are almost identical. Thus, at first glance, Mexican Americans in this age group have generally higher HRQOL as measured by the SF-36 than those respondents who are nationally representative in the same age groups. This is especially striking because the Mexican Americans in this age group are, on average, older than the national sample. This finding extends to the older age groups as well. Among those aged 75 and older, the respondents from the Mexican American sample scored higher on every dimension of the SF-36. The Mexican Americans in this sample scored approximately 3 to 20 points higher across the categories. For example, in the general health-perception dimension, the score from the sample of Mexican Americans was about 3 points higher (M = 59.36, Mexican American; M = 56.66, national), whereas for the dimension of physical role limitations, the score from the Mexican American sample was approximately 20 points higher (M = 65.77, Mexican American; M = 45.28, national).

Examining the scores across age groups, we see that there is decline in all dimension scores from younger to older ages. This is true for both samples. However, among the Mexican American sample, this decline appears to be less dramatic. Focusing on the dimensions of physical role limitations and vitality, for instance, we find that there is an approximate 19- and 9-point decline from the 65–74 age group to the 75 and older age group, respectively, for the national sample. However, for the Mexican American sample, the corresponding declines are 5 and 3 points. One reason for this may be that the Mexican Americans in the 65–74 age group are actually 71–74. Thus, the expectation for decline from a 71–74 age group to a 75 and older age group would be diminished. In general, though, the sample of older Mexican Americans appears to have higher HRQOL than the older national sample.

Internal Consistency

Table 3 shows the reliability or internal consistency found with Cronbach's alpha. Other researchers who have examined the psychometric properties of the dimensions of the SF-36 across many different samples have found the SF-36 to be very reliable (e.g., McHorney, 1996; Lyons et al., 1994; Ware et al., 1993). The results from Table 3 are consistent with previous research in that each dimension of the SF-36 has an alpha greater than.80, which is generally the acceptable value, except for social functioning (α =.76). The reliability estimates for the SF-36 in this Mexican American sample closely resemble those from other studies (range.43 to.96, based on 15 studies; Ware et al., 1993, p. 7:5).

Construct Validity

To address construct validity of the SF-36, we used a confirmatory factor analysis to fit a model that included eight first-order factors representing each of the eight SF-36 dimensions described in the Appendix as well as two second-order factors representing mental and physical factors consistent with research conducted by others (Keller et al., 1998; McHorney et al., 1993; Ware et al., 1998). The mental component scale (MCS) is hypothesized to be composed of four dimensions: mental health (MH), emotional role limitations (RE), vitality (VT), and social functioning (SF). The physical component scale (PCS) is hypothesized to be composed of the other four dimensions: physical functioning (PF), physical role limitations (RP), bodily pain (BP), and general health (GH).

Positive evidence for construct validity exists if the fit of the model is acceptable. Again, model fit is defined as (a) the NFI and CFI greater than.90; (b) the RMSEA less than.10; and (c) the parameter estimates or the factor loadings for each of the dimensions.

Table 4 shows the parameter estimates or factor loadings for the eight first-order factors and the two second-order factors. The results indicate that each of the items for each of the hypothesized first-order factors loads relatively highly on its respective factors. The items all load at.60 or higher, with the exception of one GH item (estimate =.59). Furthermore, a majority of the items load at.80 or higher. This is evidence for construct validity. The measures are performing in a manner consistent with theoretical expectations (Carmines & Zeller, 1979).

The model also simultaneously addresses the two hypothesized second-order factors of physical (PCS) and mental (MCS) health. Apparent in the last two columns of Table 4, each of the first-order factors loads highly on its hypothesized second-order factors. Each dimension loads at.80 or higher with the exception of the role limitations caused by emotional problems (RE) for the MCS. This dimension loads at.56 onto the MH second-order factor. Because each first-order factor loads well onto its hypothesized second-order factors, this is further evidence for the construct validity of the SF-36 in this older Mexican American sample.

As mentioned previously, we address model fit through the NFI, the CFI, and the RMSEA. From Table 4, each of these estimates shows reasonable fit. Again, the NFI and CFI indicate a well-fitting model if their values exceed.90. In this model, the NFI =.97, and the CFI =.96. In addition, for a model that has acceptable fit, the RMSEA should fall below.10. In this model the RMSEA is.08. Thus, the model fits reasonably well, suggesting that the specified model fits the data and that the measures are performing in a manner consistent with the theoretical expectations.

Furthermore, we present the second-order factor model only. The first- and second-order models are nested, and because of this, their chi-squares can be contrasted to test the hypothesis that one model fits the data better than the other. In this case, the restrictive model (second-order model) significantly fits the data better than the general model (first-order model; Δχ2 = 1912, Δdf = 8, p <.001). Further, we can compare the chi-square to the degrees of freedom ratio. As the ratio decreases and approaches zero, the fit of the given model improves (restrictive model: χ2/df = 5.05; general model: χ2/df = 8.38). Thus, we have evidence that the SF-36 second-order factor model is more appropriate than the first-order model using data from the Hispanic EPESE subsample.

Discussion

Our intent in this research was to examine the validity of the SF-36 in a sample of older Mexican Americans. Although the psychometric properties of the SF-36 have been established in other samples, including respondents from European countries (e.g., Keller et al., 1998), various patient groups (e.g., Ahroni & Boyko, 2000; Andresen, Gravitt, Aydelotte, & Podgorski, 1999), and national U.S. samples (e.g., Ware et al., 1993), older ethnic minorities have been largely overlooked. Older Mexican Americans, in particular, are a rapidly growing ethnic minority with a multitude of health problems. Examining the SF-36 among older Mexican Americans, if deemed to be a valid measure, could facilitate research into their HRQOL.

We examined the validity of the SF-36 in three ways. First, we compared scores from our sample of older Mexican Americans with scores from older respondents in a nationally representative sample. Second, we examined the reliability of the SF-36 dimensions within our sample. Finally, we evaluated the construct validity of the SF-36 by using confirmatory factor analysis. Each of these steps revealed evidence for the validity of the SF-36 in this sample of older Mexican Americans. The comparison of scores suggested that this sample of older Mexican Americans had higher scores than the older respondents from the national sample across most of the dimensions of the SF-36 (all of the dimensions for the older age group). The results from the Cronbach's alpha for each of the dimensions indicated high internal consistency or reliability for the dimensions. Finally, the estimates from the confirmatory factor analysis using structural equation modeling suggested evidence for construct validity for the SF-36 in this sample.

One of the interesting findings from these analyses is that this sample of older Mexican Americans scored higher on almost every dimension of the SF-36 than the national sample comparisons. This may be relevant to the concept of the Hispanic paradox or the epidemiologic paradox (Franzini et al., 2001; Markides & Coreil, 1986) discussed earlier. In summary, the Hispanic paradox suggests that Mexican Americans have low socioeconomic status and yet have health and mortality profiles similar to or better than Whites (Franzini et al., 2001). This speaks, to a certain extent, to the resiliency of older Mexican Americans.

For HRQOL, higher levels of acculturation (especially English-language usage) may theoretically provide the skills necessary to maintain an active social role in the community that could benefit the MCS (Patel, Eschbach, Rudkin, Peek, & Markides, 2004). Alternatively, higher levels of acculturation may lead to the adoption of unhealthy behaviors that are detrimental to the PCS. Though examining these arguments with respect to HRQOL goes beyond the scope of this research, future research should address the potential influence of acculturation on HRQOL as well as examine the Hispanic paradox when focusing on the SF-36 among older Mexican Americans.

Though the evidence suggests that the SF-36 is a valid and reliable instrument for this sample of older Mexican Americans, there are several limitations to this study. First, in the comparison of scores, we compared different age groups for the Mexican American sample and the national sample. Though this is due to the restricted ages of the Mexican American sample, the comparison lacks accuracy. Second, by selecting respondents for the substudy who had Medicare, we may have picked respondents who are not representative of the larger original sample. We may have oversampled the healthier older adults by focusing on those with Medicare coverage. Related to the representativeness, the scores on the dimensions of the SF-36 for the Mexican Americans were generally higher than the national sample, and this may be due to a survivor effect. These respondents are essentially the ones who are healthy enough to have participated in four waves of a study. Moreover, the respondents in this substudy are not representative of Mexican Americans in the Southwest, as the sample of the parent study was. Therefore, we can only make limited conclusions about the generalizability of these findings to the current population of older Mexican Americans residing in the Southwest. Future research should address these issues in larger population-based studies.

Nonetheless, the evidence cited in this article indicates that the SF-36 could be useful as one way of evaluating HRQOL among older Mexican Americans. Future research on the SF-36 among older Mexican Americans should examine changes in the scores on SF-36 dimensions over time. This will be useful research to evaluate how HRQOL fluctuates as Mexican Americans age and experience poor health.

This research was supported by Grants R01 AG17638, F31 AG21872-01, T32 AG00270-02, and R01 AG17231 from the National Institute on Aging.

1

Department of Preventive Medicine and Community Health, University of Texas Medical Branch, Galveston.

2

Sealy Center on Aging, University of Texas Medical Branch, Galveston.

3

School of Allied Health Sciences, University of Texas Medical Branch, Galveston.

Decision Editor: Linda S. Noelker, PhD

Table 1.

Selected Demographic and Health Characteristics of the Sample.

Selected Variables n (M ± SD, or %) 
Age 621 (78.3 ± 5.2) 
Years of education 613 (5.1 ± 3.8) 
Female 369 (59.3%) 
Distribution by state  
    Arizona 33 (5.3%) 
    California 147 (23.6%) 
    Colorado 17 (2.7%) 
    New Mexico 33 (5.3%) 
    Texas 392 (63.0%) 
Reporting any ADL disability 116 (18.7%) 
Chronic conditions  
    Heart attack since Wave 3 23 (3.7%) 
    Stroke since Wave 3 20 (3.2%) 
    Hip fracture since Wave 3 16 (2.6%) 
    Ever told had hypertension 328 (52.7%) 
    Ever told had cancer 34 (5.5%) 
    Ever told had diabetes 174 (28.0%) yes; 20 (3.2%) borderline 
Selected Variables n (M ± SD, or %) 
Age 621 (78.3 ± 5.2) 
Years of education 613 (5.1 ± 3.8) 
Female 369 (59.3%) 
Distribution by state  
    Arizona 33 (5.3%) 
    California 147 (23.6%) 
    Colorado 17 (2.7%) 
    New Mexico 33 (5.3%) 
    Texas 392 (63.0%) 
Reporting any ADL disability 116 (18.7%) 
Chronic conditions  
    Heart attack since Wave 3 23 (3.7%) 
    Stroke since Wave 3 20 (3.2%) 
    Hip fracture since Wave 3 16 (2.6%) 
    Ever told had hypertension 328 (52.7%) 
    Ever told had cancer 34 (5.5%) 
    Ever told had diabetes 174 (28.0%) yes; 20 (3.2%) borderline 

Notes: Sample is N = 621. ADL = activities of daily living. There are 2 years in between the third and fourth wave.

Table 2.

Score Comparisons for Dimensions of the SF-36 Across the Subsample From Hispanic EPESE and National Scores.

Dimension Age 65–74
 
   Age 75+
 
   
Mexican Americans
 
 National
 
 Mexican Americans
 
 National
 
 
M SD M SD M SD M SD 
Physical functioning 65.42 (32.42) 69.38 (26.26) 58.92 (33.43) 53.20 (29.98) 
Role limitations—physical 70.57 (42.11) 64.54 (41.30) 65.77 (42.94) 45.28 (41.95) 
Social functioning 82.62 (23.58) 80.61 (25.63) 81.22 (23.95) 73.89 (28.75) 
Bodily pain 72.21 (25.48) 68.49 (26.42) 68.19 (25.86) 60.88 (26.01) 
General mental health 81.01 (16.76) 76.87 (18.08) 80.74 (16.52) 73.99 (20.23) 
Role limitations—emotional 82.55 (35.35) 81.44 (34.56) 80.76 (35.71) 63.18 (42.96) 
Vitality 67.06 (19.73) 59.94 (22.12) 64.23 (20.32) 50.41 (23.62) 
General health perceptions 62.50 (22.47) 62.56 (22.42) 59.36 (21.74) 56.66 (21.21) 
n 192  442  430  264  
Dimension Age 65–74
 
   Age 75+
 
   
Mexican Americans
 
 National
 
 Mexican Americans
 
 National
 
 
M SD M SD M SD M SD 
Physical functioning 65.42 (32.42) 69.38 (26.26) 58.92 (33.43) 53.20 (29.98) 
Role limitations—physical 70.57 (42.11) 64.54 (41.30) 65.77 (42.94) 45.28 (41.95) 
Social functioning 82.62 (23.58) 80.61 (25.63) 81.22 (23.95) 73.89 (28.75) 
Bodily pain 72.21 (25.48) 68.49 (26.42) 68.19 (25.86) 60.88 (26.01) 
General mental health 81.01 (16.76) 76.87 (18.08) 80.74 (16.52) 73.99 (20.23) 
Role limitations—emotional 82.55 (35.35) 81.44 (34.56) 80.76 (35.71) 63.18 (42.96) 
Vitality 67.06 (19.73) 59.94 (22.12) 64.23 (20.32) 50.41 (23.62) 
General health perceptions 62.50 (22.47) 62.56 (22.42) 59.36 (21.74) 56.66 (21.21) 
n 192  442  430  264  

Notes: The respondents from the Hispanic Established Populations for Epidemiologic Study of the Elderly (EPESE) subsample are 71 and older. Thus, the sample size for the 65–74 group is small because it represents ages 71–74 for the Mexican Americans. SF-36 = 36-item Short-Form Health Survey. Scores range from 0 to 100 in each dimension, with higher scores reflecting better functioning. The source for the national scores is the SF-36 Health Survey: Manual and Interpretation Guide (Ware et al., 1993, pp. 10:16–10:17).

Table 3.

SF-36 Dimension Reliability for Older Mexican Americans.

Dimension α 
Physical functioning .96 
Role limitations—physical .93 
Social functioning .76 
Bodily pain .86 
General mental health .82 
Role limitations—emotional .90 
Vitality .83 
General health perceptions .82 
Dimension α 
Physical functioning .96 
Role limitations—physical .93 
Social functioning .76 
Bodily pain .86 
General mental health .82 
Role limitations—emotional .90 
Vitality .83 
General health perceptions .82 

Notes: SF-36 = 36-Item Short-Form Health Survey. For the table, n = 621.

Table 4.

First- and Second-Order Factor Loadings for the SF-36 Among Mexican Americans.

SF-36 Item PF RP SF BP MH RE VT GH PCS MCS 
First order           
    PF1 .69          
    PF2 .86          
    PF3 .84          
    PF4 .88          
    PF5 .90          
    PF6 .82          
    PF7 .87          
    PF8 .91          
    PF9 .85          
    PF10 .66          
    RP1  .88         
    RP2  .83         
    RP3  .91         
    RP4  .90         
    SF1   .78        
    SF2   .78        
    BP1    .81       
    BP2    .93       
    MH1     .67      
    MH2     .68      
    MH3     .60      
    MH4     .75      
    MH5     .73      
    RE1      .88     
    RE2      .87     
    RE3      .86     
    VT1       .79    
    VT2       .83    
    VT3       .69    
    VT4       .67    
    GH1        .69   
    GH2        .62   
    GH3        .82   
    GH4        .59   
    GH5        .73   
Second order           
    PF         .81  
    RP         .81  
    GH         .83  
    BP         .86  
    RE          .56 
    MH          .81 
    SF          .89 
    VT          .86 
CFI .97          
NFI .96          
RMSEA .08          
SF-36 Item PF RP SF BP MH RE VT GH PCS MCS 
First order           
    PF1 .69          
    PF2 .86          
    PF3 .84          
    PF4 .88          
    PF5 .90          
    PF6 .82          
    PF7 .87          
    PF8 .91          
    PF9 .85          
    PF10 .66          
    RP1  .88         
    RP2  .83         
    RP3  .91         
    RP4  .90         
    SF1   .78        
    SF2   .78        
    BP1    .81       
    BP2    .93       
    MH1     .67      
    MH2     .68      
    MH3     .60      
    MH4     .75      
    MH5     .73      
    RE1      .88     
    RE2      .87     
    RE3      .86     
    VT1       .79    
    VT2       .83    
    VT3       .69    
    VT4       .67    
    GH1        .69   
    GH2        .62   
    GH3        .82   
    GH4        .59   
    GH5        .73   
Second order           
    PF         .81  
    RP         .81  
    GH         .83  
    BP         .86  
    RE          .56 
    MH          .81 
    SF          .89 
    VT          .86 
CFI .97          
NFI .96          
RMSEA .08          

Notes: SF-36 = 36-Item Short-Form Health Survey; PF = physical functioning; RP = role limitations—physical; SF = social functioning; BP = bodily pain; MH = mental health; RE = role limitations—emotional; VT = vitality; GH = general health; PCS = physical component scale; MCS = mental component scale; CFI = comparative fit index; NFI = normed fit index; RMSEA = root mean square error of approximation. For the table, n = 615.

The Dimensions of the SF-36.

Dimension No. of Items Description Example Items 
First-order factors    
    PF 10 Addresses PF through the number of physical limitations Does your health limit you in vigorous activities, such as running or lifting heavy objects; moderate activities, such as moving a table or pushing a vacuum cleaner? 
    RP Represents role limitations in work or usual activities, amount of time spent in work or typical activities, and performance in work or usual activities During the past 4 weeks, have you had any of the following problems with your work or regular daily activities as a result of your physical health? Cut down on the amount of time you spent on work or other activities? Accomplished less than you would like? 
    BP Addresses intensity of BP or discomfort How much BP have you had in the past 4 weeks? 
    GH Represents GH, including health outlook and perception of resistance to illness How true or false is each of the following statements for you? I seem to get sick a little easier than other people; I expect my health to get worse. 
    VT Addresses energy level and fatigue How much of the time during the past 4 weeks did you feel full of pep? Did you have a lot of energy? 
    SF Measures quantity and quality of social activities with others During the past 4 weeks, to what extent has your physical health or emotional problems interfered with your normal social activities with family, friends, neighbors, or groups? 
    RE Similar to RP but addresses limitations in usual activities caused by emotional problems During the past 4 weeks, have you had any of the following problems with your work or regular daily activities as a result of any emotional problems? Cut down on the amount of time you spent on work or other activities? Accomplished less than you would like? 
    MH Represents general MH from four major mental health dimensions (anxiety, depression, loss of behavioral or emotional control, and well-being) How much of the time during the past 4 weeks have you been a very nervous person? Have you been a happy person? 
Second-order factors    
    Mental Health Summary Scale 4 scales Represents MCS Includes VT, RE, MH, and SF 
    Physical Health Summary Scale 4 scales Represents PCS Includes BP, GH, RP, and PF 
Dimension No. of Items Description Example Items 
First-order factors    
    PF 10 Addresses PF through the number of physical limitations Does your health limit you in vigorous activities, such as running or lifting heavy objects; moderate activities, such as moving a table or pushing a vacuum cleaner? 
    RP Represents role limitations in work or usual activities, amount of time spent in work or typical activities, and performance in work or usual activities During the past 4 weeks, have you had any of the following problems with your work or regular daily activities as a result of your physical health? Cut down on the amount of time you spent on work or other activities? Accomplished less than you would like? 
    BP Addresses intensity of BP or discomfort How much BP have you had in the past 4 weeks? 
    GH Represents GH, including health outlook and perception of resistance to illness How true or false is each of the following statements for you? I seem to get sick a little easier than other people; I expect my health to get worse. 
    VT Addresses energy level and fatigue How much of the time during the past 4 weeks did you feel full of pep? Did you have a lot of energy? 
    SF Measures quantity and quality of social activities with others During the past 4 weeks, to what extent has your physical health or emotional problems interfered with your normal social activities with family, friends, neighbors, or groups? 
    RE Similar to RP but addresses limitations in usual activities caused by emotional problems During the past 4 weeks, have you had any of the following problems with your work or regular daily activities as a result of any emotional problems? Cut down on the amount of time you spent on work or other activities? Accomplished less than you would like? 
    MH Represents general MH from four major mental health dimensions (anxiety, depression, loss of behavioral or emotional control, and well-being) How much of the time during the past 4 weeks have you been a very nervous person? Have you been a happy person? 
Second-order factors    
    Mental Health Summary Scale 4 scales Represents MCS Includes VT, RE, MH, and SF 
    Physical Health Summary Scale 4 scales Represents PCS Includes BP, GH, RP, and PF 

Notes: This table is based on information from the SF-36 Health Survey: Manual and Interpretation Guide (Ware et al., 1993) and the SF-36 Physical and Mental Health Summary Scales: A Users Manual (Ware, Kosinksi, & Keller, 1994). The original eight subscales have a range from 0 to 100. SF-36 = 36-Item Short-Form Health Survey; PF = physical functioning; RP = role limitations—physical; SF = social functioning; BP = bodily pain; MH = mental health; RE = role limitations—emotional; VT = vitality; GH = general health; PCS = physical component scale; and MCS = mental component scale.

References

Ahroni, J. H., & Boyko, E. J. (
2000
). Responsiveness of the SF-36 among veterans with diabetes mellitus.
Journal of Diabetes and Its Complications,
 
14,
31
-39.
Andresen, E. M., Gravitt, G. W., Aydelotte, M. E., & Podgorski, C. A. (
1999
). Limitations of the SF-36 in a sample of nursing home residents.
Age and Ageing,
 
28,
562
-566.
Arocho, R., McMillan, C. A., & Sutton-Wallace, P. (
1998
). Construct validation of the USA-Spanish version of the SF-36 health survey in a Cuban-American population with benign hyperplasia.
Quality of Life Research,
 
7,
121
-126.
Ayuso-Mateos, J. L., Vazquez-Barquero, J. L., Oviedo, A., & Diez-Manrique, J. F. (
1999
). Measuring health status in psychiatric community surveys: Internal and external validity of the Spanish version of the SF-36.
Acta Psychiatrica,
 
99,
26
-32.
Benjamin-Coleman, R., & Alexy, B. (
1999
). Use of the SF-36 to identify community dwelling rural elderly at risk for hospitalization.
Public Health Nursing,
 
16,
223
-227.
Bentler, P. M. (
1990
). Comparative fit indexes in structural models.
Psychological Bulletin,
 
107,
238
-246.
Carmines, E., & Zeller R. (
1979
). Reliability and validity assessment: Quantitative application in the social sciences. Beverley Hills, CA: Sage.
Day, J. C. (
1996
). Population projections of the United States by age, sex, race and Hispanic origin: 1995 to 2050 (Current Population Reports, P-25). Washington, DC: U.S. Department of Commerce, Bureau of the Census.
Franzini, L., Ribble, J. C., & Keddie, A. M. (
2001
). Understanding the Hispanic paradox.
Ethnicity and Disease,
 
11,
496
-518.
Guyatt, G. H., Eagle, D. J., Sackett, B., Willan, A., Griffith, L., & McIlroy, W., et al (
1993
). Measuring quality of life in the elderly.
Journal of Clinical Epidemiology,
 
46,
1433
-1444.
Hazuda, H. P., Haffner, S. M., Stern, M. P., & Eifler, C. W. (
1988
). Effects of acculturation and socioeconomic status on obesity and diabetes in Mexican Americans.
American Journal of Epidemiology,
 
128,
1289
-1301.
Keller, S. D., Ware, J. E., Jr., Bentler, P. M., Aaronson, N. K., Alonso, J., & Apolone, G., et al (
1998
). Use of structural equation modeling to test the construct validity of the SF-36 health survey in 10 countries: Results from the IQOLA Project.
Journal of Clinical Epidemiology,
 
51,
1179
-1188.
Kline, R. B. (
1998
). Principles and practice of structural equation modeling. New York: Guilford Press.
Livingston, G., Watkin, V., Manela, M., Rosser, R., & Katona, C. (
1998
). Quality of life in older people.
Aging and Mental Health,
 
2,
20
-23.
Lyons, R. A., Crome, P., Monaghan, S., Killalea, D., & Daley, J. A. (
1997
). Health status and disability among elderly people in three UK districts.
Age and Ageing,
 
26,
203
-209.
Lyons, R. A., Perry, H. M., & Littlepage, B. N. C. (
1994
). Evidence for the validity of the Short-form 36 Questionnaire (SF-36) in an elderly population.
Age and Ageing,
 
23,
182
-184.
Markides, K. S., & Coreil, J. (
1986
). The health of Hispanics in the southwestern United States: An epidemiologic paradox.
Public Health Reports,
 
101,
253
-265.
Markides, K. S., Stroup-Benham, C. A., Black, S. A., Satish, S., Perkowski, L. C., & Ostir, G. (
1999
). The health of Mexican American elderly: Selected findings from the Hispanic EPESE. In M. Wykle and A. Ford (Eds.), Planning services for minority elderly in the 21st century. New York: Springer.
Markides, K. S., Stroup-Benham, C. A., Goodwin, J. S., Perkowski, L. C., Lichtenstein, M., & Ray, L. A. (
1996
). The effect of medical conditions on the functional limitations of Mexican American elderly.
Annals of Epidemiology,
 
6,
386
-391.
Martikainen, P., Stansfeld, S., Hemingway, H., & Marmot, M. (
1999
). Determinants of socioeconomic differences in change in physical and mental functioning.
Social Science & Medicine,
 
49,
499
-507.
McHorney, C. A. (
1996
). Measuring and monitoring general health status in elderly persons: Practical and methodological issues in using the SF-36 Health Survey.
The Gerontologist,
 
36,
571
-583.
McHorney, C. A., Ware, J. E., Jr., & Raczek, A. E. (
1993
). The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs.
Medical Care,
 
31,
247
-263.
Patel, K. V., Eschbach, K., Rudkin, L. L., Peek, M. K., & Markides, K. S. (
2003
). Neighborhood context and self-rated health in older Mexican Americans.
Annals in Epidemiology,
 
13,
620
-628.
Rudkin, L., Markides, K. S., & Espino, D. V. (
1997
). Functional disability in older Mexican Americans.
Topics in Geriatric Rehabilitation,
 
12,
38
-46.
Scribner, R., & Dwyer, J. H. (
1989
). Acculturation and low birth weight among Latinos in the Hispanic HANES.
American Journal of Public Health,
 
79,
1263
-1267.
Simon, G. E., Revicki, D. A., Grothaus, L., & Vonkorff, M. (
1998
). SF-36 summary scores: Are physical and mental health truly distinct?
Medical Care,
 
36,
567
-572.
Stadnyk, K., Calder, J., & Rockwood, K. (
1998
). Testing the measurement properties of the Short Form-36 Health Survey in a frail elderly population.
Journal of Clinical Epidemiology,
 
51,
827
-835.
Therrien, M., & Ramirez, R. R. (
2000
). The Hispanic population in the United States: March 2000 (Current Population Reports, P20-535). Washington, DC: U.S. Department of Commerce, Bureau of the Census.
Walters, S. J., Munro, J. F., & Brazier, J. E. (
2001
). Using the SF-36 with older adults: A cross-sectional community-based survey.
Age and Ageing,
 
30,
337
-343.
Ware, J. E., Jr., Kosinski, M., Gandek, B., Aaronson, N. K., Apolone, G., & Bech, P., et al (
1998
). The factor structure of the SF-36 Health Survey in 10 countries: Results from the IQOLA Project.
Journal of Clinical Epidemiology,
 
51,
1159
-1165.
Ware, J. E., Jr., Kosinski, M., & Keller, S. D. (
2001
). SF-36 Physical and Mental Health Summary Scale: A manual for users of Version 1. (2nd ed.). Lincoln, RI: QualityMetric Inc.
Ware, J. E., Jr., & Sherbourne, C. D. (
1992
). The MOS 36-Item Short-Form Health Survey (SF-36) I. Conceptual framework and item selection.
Medical Care,
 
30,
473
-483.
Ware, J. E., Jr., Snow, K. K., Kosinski, M., & Gandek, B. (
1993
). SF-36 Health Survey manual and interpretation guide. Boston: The Health Institute, New England Medical Center.