SEARCH

SEARCH BY CITATION

Keywords:

  • quality of life;
  • risk factors;
  • cardiovascular disease;
  • diabetes

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: Research has shown that risk factors for cardiovascular disease often cluster together, most notably overweight/obesity, diabetes, hyperlipidemia, and hypertension. The impact of cardiometabolic risk factor clusters on health-related quality of life (HRQL) is not well understood. The purpose of this study was to examine and quantify the impact of cardiometabolic risk factor clusters on HRQL as measured by the SF (Short Form)-12 Mental Component Scale (MCS-12), SF-12 Physical Component Scale (PCS-12), EQ-5D index (a generic quality of life index), and Visual Analogue Scale.

Research Methods and Procedures: The Medical Expenditure Panel Survey is a nationally representative survey of the U.S. population. From 2000 to 2002, detailed information on sociodemographic characteristics and health conditions were collected for 36,697 adults with complete responses. Controlling for comorbidity and sociodemographic characteristics, this study estimated the marginal impact of cardiometabolic risk factor clusters on MCS-12, PCS-12, EQ-5D index, and Visual Analogue Scale scores. Cardiometabolic risk factor clusters were defined as the presence of BMI ≥25 kg/m2 and at least two of the following: diabetes, hyperlipidemia, and hypertension. Using BMI ≥30 kg/m2 as the cut-off was also examined.

Results: The marginal impact of cardiometabolic risk factor clusters was highly statistically significant across all four HRQL measures and seemed to be clinically significant for all but the MCS-12. The PCS-12 showed a greater decrease in HRQL associated with physical function compared with mental function-related domains of the MCS-12.

Discussion: Common cardiometabolic risk factor clusters such as overweight/obesity, diabetes, hypertension, and hyperlipidemia have a significant and negative impact on HRQL in the United States.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Cardiovascular disease is the leading cause of death and morbidity in the United States, causing an estimated 40% of all deaths (1). In addition, more than one fourth of the population (70 million individuals) currently lives with cardiovascular disease (1). Cardiovascular disease has a profound deleterious impact on health-related quality of life (HRQL)1 (2, 3). It is largely preventable, and much research in the last century has focused on identifying key risk factors. As they were identified, it became apparent that risk factors for cardiovascular disease often cluster in the same individual, most notably overweight/obesity, diabetes, hyperlipidemia, and hypertension (4).

Although there is ample evidence of the deleterious impact of cardiovascular disease on HRQL, little is known about the impact of cardiometabolic risk factor clusters on HRQL. Wee et al. (5) found that diabetes and hypertension, as well as diabetes with other single chronic conditions, had a deleterious, additive impact on HRQL scores as measured by the short form (SF)-36 and the SF-6 days in a diverse Asian population in Singapore. Chambers et al. (6) examined the individual and combined impact of hypertension, hyperlipidemia, obesity, and smoking status on SF-36 scores in a small sample of adults from Southwestern Ohio. They found that SF-36 scores decreased as the number of cardiovascular risk factors increased. Oldridge et al. (7) showed that diabetes, hypertension, and obesity together had a deleterious impact on HRQL as measured by mobility difficulty, activities of daily living, and self-perceived health status (7).

The appraisal of HRQL is of paramount importance in the evaluation of disease and treatment (8). Measures such as the SF-12 provide an assessment of the generic health status of individuals by examining broad domains of HRQL such as physical and mental function (9). In addition, preference-based measures of HRQL, such as the EQ-5D index (10), facilitate an assessment of the value of a variety of health interventions through cost-effectiveness analysis and the calculation of quality-adjusted life years (11, 12). Rating scale methods, such as the visual analogue scale (VAS), provide important information on the subjective perception of health status (13). The purpose of this study was to examine and quantify the impact of cardiometabolic risk factor clusters on HRQL and preference-based HRQL as measured by the SF-12, EQ-5D index, and VAS in a nationally representative sample of U.S. adults.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Data Source

The Medical Expenditure Panel Survey (MEPS) is cosponsored by the Agency for Healthcare Research and Quality and the National Center for Health Statistics. The MEPS Household Component (HC), a nationally representative survey of the U.S. civilian non-institutionalized population, collects detailed information on demographic characteristics, health conditions, health status, and use of medical care services, income, and employment (14). The sampling frame for the MEPS HC is drawn from respondents to the National Health Interview Survey (NHIS). NHIS provides a nationally representative sample of the U.S. civilian non-institutionalized population, with over sampling of Hispanics and blacks. This research used the 2000 and 2002 MEPS public use data (14). The 2000 and 2002 data were used to provide a larger sample and because they each contain data on unique individuals that do not overlap (unlike MEPS 2001 data, which has panels that overlap with 2000 and 2002), but are equally nationally representative. The SF-12 and EQ-5D were administered through the self-administered questionnaire, a paper and pencil questionnaire administered to adults ≥18 years of age. If an individual was unable to respond to the self-administered questionnaire, the questionnaire was completed by a proxy. In the MEPS 2000 and 2002 data, there were 39,751 adults (>17 years of age) with non-zero person weights; 38,678 adults with valid EQ-5D scores; and 36,697 adults with complete data on all of the variables included in the multivariate regression analyses (see Tables 4 and 5). Examination of the population with missing observations showed that they were more likely to be older, have lower income and education levels, and to be black or Hispanic than individuals with complete data. Although these differences are a cause for concern, the small percentage of missing observations, the large number of individuals with complete data on a broad array of variables (36,697), and the MEPS sample weights incorporating adjustment for survey non-response lend confidence in the results. The sample design of the MEPS HC survey includes stratification, clustering, multiple stages of selection, and disproportionate sampling (15). MEPS sampling weights incorporate adjustment for the complex sample design and reflect survey non-response and population totals from the Current Population Survey.

Table 4.  Results of multivariate analysis of four separate regressions, PCS-12, MCS-12, EQ-5D, and VAS, on CMRFCs (BMI ≥ 25 kg/m2), controlling for sociodemographic characteristics and comorbidity burden
 PCS-12SEMCS-12SEEQ-5D*SEVAS*SE
  • PCS, Physical Component Summary; MCS, Mental Component Summary; VAS, Visual Analogue Scale; CMRFC, cardiometabolic risk factor cluster; SE, standard error; NCC, number of chronic conditions. All coefficients p < 0.05 unless otherwise noted.

  • *

    Using censored least absolute deviations analysis.

  • Not statistically significant (p < 0.05).

  • CMRFC = BMI ≥ 25 kg/m2 and at least two of the following risk factors: diabetes, high blood pressure, and/or hyperlipidemia.

Age (years)        
 18 to 191.890.220.770.370.0180.003.810.65
 30 to 39−1.220.13−0.200.17−0.0160.00−2.610.35
 40 to 49−2.370.16−0.110.18−0.0300.00−3.220.35
 50 to 59−3.650.180.510.20−0.0670.00−3.610.37
 60 to 69−4.510.222.800.24−0.0690.00−2.940.43
 70 to 79−6.620.263.910.26−0.0730.00−4.760.48
 80 to 90−10.050.403.580.36−0.0860.00−7.550.62
Sex        
 Male−0.160.110.970.100.0000.00−0.390.21
Ethnicity        
 Hispanic−0.030.15−0.070.160.0000.000.000.30
Race        
 Black−0.880.160.020.190.0000.00−0.320.33
 Native American−0.780.69−0.960.85−0.0080.00−1.771.10
 Other−1.300.26−0.380.310.0000.00−3.100.52
Smoking        
 Smoking−1.100.12−2.220.16−0.038 −3.720.26
Income        
 Income21.240.380.210.420.0200.003.180.59
 Income31.900.251.750.250.0390.004.830.41
 Income42.850.233.140.220.0550.006.340.36
 Income53.810.243.820.230.0550.007.560.38
Education        
 High school degree1.820.170.940.170.0180.003.050.30
 Other degree2.560.241.470.220.0340.003.760.48
 Bachelor's degree3.500.190.860.230.0340.003.440.40
 Ph.D. or Master's degree4.400.251.000.250.0490.004.050.49
CMRFC        
 CMRFC (BMI ≥ 25 kg/m2)−3.480.26−0.870.26−0.0380.00−6.100.42
Comorbidity        
 NCC−1.980.04−1.260.03−0.0330.00−2.830.06
Intercept50.890.2648.880.290.9270.0083.010.48
Table 5.  Results of multivariate analysis of four separate regressions, PCS-12, MCS-12, EQ-5D, and VAS, on CMRFCs (BMI ≥ 30 kg/m2), controlling for sociodemographic characteristics and comorbidity burden
 PCS-12SEMCS-12SEEQ-5D*SEVAS*SE
  • PCS, Physical Component Summary; MCS, Mental Component Summary; VAS, Visual Analogue Scale; CMRFC, cardiometabolic risk factor cluster; SE, standard error; NCC, number of chronic conditions. All coefficients p < 0.05 unless otherwise noted.

  • *

    Using censored least absolute deviations analysis.

  • Not statistically significant (p < 0.05).

  • CMRFC = BMI ≥ 30 kg/m2 and at least two of the following risk factors: diabetes, high blood pressure, and/or hyperlipidemia.

Age (years)        
 18 to 191.880.220.770.370.0170.003.820.48
 30 to 39−1.220.13−0.200.17−0.0160.00−2.550.26
 40 to 49−2.390.16−0.120.18−0.0310.00−3.090.26
 50 to 59−3.710.180.490.20−0.0690.00−3.640.28
 60 to 69−4.670.222.750.24−0.0700.00−3.090.32
 70 to 79−6.880.253.830.26−0.0750.00−5.000.35
 80 to 90−10.240.403.530.37−0.0900.00−7.760.46
Sex        
 Male−0.210.110.960.100.0000.00−0.550.16
Ethnicity        
 Hispanic−0.050.15−0.080.160.0000.000.000.22
Race        
 Black−0.870.160.020.190.0000.00−0.270.24
 Native American−0.800.69−0.970.85−0.0060.00−1.730.82
 Other−1.330.26−0.390.310.000*0.00−3.240.39
Smoking        
 Smoking−1.100.12−2.220.16−0.038 −3.640.19
Income        
 Income21.230.380.210.420.0180.003.180.44
 Income31.890.251.740.250.0380.004.790.31
 Income42.840.233.140.220.0540.006.090.27
 Income53.800.243.820.230.0540.007.450.28
Education        
 High school degree1.810.170.940.170.0170.003.090.22
 Other degree2.560.241.480.220.0330.003.880.36
 Bachelor's degree3.480.190.860.230.0330.003.640.30
 Ph.D. or Master's degree4.380.251.000.250.0480.003.910.36
CMRFC        
 CMRFC (BMI ≥ 25 kg/m2)−4.870.33−1.030.37−0.0540.00−8.550.41
Comorbidity        
 NCC−1.990.04−1.260.03−0.0330.00−2.880.04
Intercept50.940.2648.890.290.9290.0083.090.36

Cardiometabolic Risk Factors and Sociodemographic Characteristics

This analysis used self-reported information resulting from the MEPS HC survey, in which respondents were asked the following: whether they had ever been diagnosed as having diabetes (excluding gestational diabetes) and whether they had been told on two or more different medical visits that they had high blood pressure. In addition, MEPS mapped medical conditions to three-digit International Classification of Diseases (ICD)-9 codes based on medical and pharmacy use and self-report. Then 259 mutually exclusive clinical classification categories (CCCs) were mapped from ICD-9 codes to provide clinically homogenous groupings (14, 16). The ICD-9 to CCC cross-walk is available at www.meps.ahrq (14). This research used CCC 053 “Disorders of Lipid Metabolism” to identify individuals with hyperlipidemia.

In addition, self-reported information from the MEPS HC survey was used for the assessment of BMI and other covariates. Respondents were asked to estimate their current body weight and height, from which BMI was calculated. The following formula (from the Centers for Disease Control and Prevention: http:www.cdc.gov) was used to calculate BMI for adults in MEPS based on reported height and weight: BMI = [weight in pounds/(height in inches)2)] × 703. Full documentation is provided on page 97 of the MEPS H60 documentation file (14). BMI ≥ 25 kg/m2 was considered a cardiometabolic risk factor.

The prevalence of cardiometabolic risk factor clusters has increased significantly for both men and women and for all race, ethnic, age, and education groups (17, 18). These risk factors often cluster in the same individual, most notably overweight/obesity, diabetes, hyperlipidemia, and hypertension (4). Among those with three risk factors, the most common clusters were hypertension, hyperlipidemia, and obesity, whereas among those with four risk factors, 40% had hypertension, hyperlipidemia, obesity, and diabetes (17, 18). While it is clear that diabetes, hypertension, hyperlipidemia, and overweight/obesity are prone to cluster and result in an elevated risk of cardiovascular disease, there is some controversy about whether these cardiometabolic risk factor clusters reflect an underlying syndrome thought to be related to insulin resistance (commonly referred to as the metabolic syndrome) or simply that they tend to cluster (4). This analysis examines the impact of these cardiometabolic risk factor clusters on HRQL, irrespective of the underlying pathophysiology.

Based on the aforementioned risk factor definitions, individuals were considered to have cardiometabolic risk factor clusters if they had BMI ≥25 kg/m2 and any two of the following three risk factors: hypertension, hyperlipidemia, and diabetes. Using BMI ≥30 kg/m2 as the cut-off was also examined. In other words, a dichotomous (yes or no) variable was created for cardiometabolic risk factor clusters based on this criterion.

To control for confounding in the statistical analysis, several comorbidity and sociodemographic characteristics were identified. A measure of comorbidity burden was constructed by adding the number of non-cardiovascular chronic conditions for each individual. The number of chronic conditions (NCC) was calculated by adding the total number of chronic ICD-9 codes reported for each individual. Age was measured in the following categories: 18 to 19; 20 to 29; 30 to 39; 40 to 49; 50 to 59; 60 to 69; 70 to 79; ≥80 years. Income was expressed as total family income as a percentage of the poverty line, Income1 (<100%), Income2 (100% to <125%), Income3 (125% to <200%), Income4 (200% to <400%), and Income5 (≥400%). Education was measured categorically: less than high school degree; high school degree; other degree; Bachelor's degree and Master's or PhD. Race was measured as white, black, Native American, or other. Ethnicity was measured as Hispanic or non-Hispanic. Sex was also captured.

HRQL Instruments

The EQ-5D consists of a five-item descriptive system that measures five dimensions of health status (mobility, self care, usual activities, pain/discomfort, and anxiety/depression) with three levels per dimension (no problem, some problems, and extreme problems) (19). There are 243 unique health states deriving from the combination of all possible dimensions and levels of the five questions. Preferences for these health states are mapped using a multiattribute value function. The scoring algorithm for the EQ-5D index descriptive system used in this research is based on U.S. community preferences (10). From this scoring algorithm, EQ-5D index scores are calculated based on responses to the five-item questionnaire. The EQ-5D has been used as a HRQL and preference measure in a wide variety of diseases and conditions in more than 600 publications, and its construct validity, reliability, and responsiveness have been documented extensively (20, 21). In addition, a single item question asking about self-perceived health status using a VAS was included. This question asked respondents to rate their current overall health on a scale that ranges from 0 through 100, where 0 represents “worst possible health” and 100 represents “best possible health.”

The SF-12 Health Survey is a general health status instrument with 12 questions producing two summary scores: the Physical Component Summary (PCS-12) and the Mental Component Summary (MCS-12). These summary scales are scored such that higher scores represent better physical and emotional function, respectively, and are standardized so that the mean score is 50 and standard deviation is 10 in the general population (22).

Data Analysis

To estimate the marginal impact of cardiometabolic risk factor clusters on HRQL, multivariate regression analysis was conducted for each of the following dependent variables: EQ-5D index, PCS-12, MCS-12, and VAS. In each of the four regressions, the dependent variable was regressed on the dichotomous cardiometabolic risk factor clusters variable, controlling for age, sex, ethnicity, race, income, education, smoking status, and comorbidity (NCC) as delineated above. To incorporate adjustment for the complex sample design, this research used MEPS person-level and variance adjustment weights using STATA 8.1 in all analyses to ensure nationally representative estimates. Ordinary least squares regression analysis was used for both the PCS-12 and MCS-12.

However, previous research has noted that EQ-5D index scores exhibit a ceiling effect, with a significant number of respondents rating themselves in full health (11). For example, in this MEPS sample, 46% of respondents rated themselves in full health (i.e., “no problem” on each of the five items of the EQ-5D scale, or 11111), resulting in an EQ-5D index score of 1.0. Given the bounded nature of preference-based scores, failure to account for censoring using ordinary least squares regression will result in biased and inconsistent estimates (23). For these cases, the Tobit model provides an unbiased and consistent alternative (24). However, the Tobit model is only appropriate when errors are normally distributed with constant variance (23, 24). When these assumptions are violated, the censored least absolute deviations estimator (CLAD) is a robust alternative to maximum likelihood estimation for the Tobit model (25). Previous research has shown that these assumptions are violated when analyzing the EQ-5D index in MEPS (11). Hence, CLAD was used to estimate EQ-5D index scores. CLAD was performed in STATA using an add-in that was adjusted to incorporate sample weights (26). The Tobit model was used to estimate VAS scores.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Men, individuals >50 years of age, blacks, individuals with lower levels of educational attainment and income, and Hispanics had a higher prevalence of cardiometabolic risk factor clusters (Table 1). The incidence of cardiometabolic risk factor clusters, BMI ≥25 kg/m2, diabetes, hypertension, and hyperlipidemia in the nationally representative adult MEPS sample with valid BMI was 6.29%, 58.95%, 6.29%, 20.18%, and 8.62%, respectively. Individuals with cardiometabolic risk factor clusters were older, had greater comorbidity burden, and had lower HRQL than those without cardiometabolic risk factor clusters and the U.S. average; these relationships were worse using BMI ≥30 kg/m2 as the cut-off than BMI ≥25 kg/m2 (Table 2).

Table 1.  Prevalence of CMRFCs by selected characteristics (adult MEPS 2000 to 2002 with valid BMI)
PopulationPrevalence of CMRFC (BMI ≥ 25 kg/m2)*Prevalence of CMRFC (BMI ≥ 30 kg/m2)*Prevalence of BMI ≥ 25 kg/m2Prevalence of diabetesPrevalence of hypertensionPrevalence of hyperlipidemia
  • MEPS, Medical Expenditure Panel Survey; CMRFC, cardiometabolic risk factor cluster.

  • *

    CMRFC = BMI ≥ 25 kg/m2 (or BMI ≥ 30 kg/m2) and at least two of the following risk factors: diabetes, high blood pressure, and/or hyperlipidemia.

General MEPS (total N = 43,221)6.29%3.26%58.95%6.29%20.18%8.62%
Sex      
 Female5.76%3.22%52.08%6.29%21.12%7.74%
 Male6.86%3.30%66.25%6.29%19.19%9.57%
Age categories (years)      
 18 to 190.00%0.00%31.36%0.66%0.44%0.06%
 20 to 290.25%0.19%48.17%0.87%2.21%0.36%
 30 to 391.00%0.72%59.92%1.96%6.49%2.26%
 40 to 494.11%2.52%64.97%4.15%15.61%6.71%
 50 to 599.81%5.72%65.96%9.39%29.77%13.28%
 60 to 6917.14%8.72%67.41%14.83%43.74%22.26%
 70 to 7918.61%7.90%60.85%17.30%51.89%22.95%
 80+12.68%4.53%46.96%14.30%55.47%15.80%
Race      
 White6.23%3.13%58.66%5.90%19.80%9.15%
 Black7.64%4.83%68.78%9.11%25.30%5.48%
 Native American6.26%3.39%70.00%8.66%21.63%5.08%
 Other4.05%1.61%37.63%5.90%13.97%7.65%
Education      
 No degree8.08%4.30%59.44%9.40%24.96%7.89%
 High school degree6.46%3.43%60.96%6.10%20.59%8.68%
 Other degree5.13%3.04%60.51%4.92%18.35%8.59%
 Bachelor's degree4.58%2.09%53.86%4.68%14.99%8.25%
 MA or Ph.D.5.23%2.24%52.88%4.05%18.13%10.88%
Income      
 Poor/negative6.76%3.96%59.30%7.85%23.43%6.94%
 Near poor9.73%5.48%61.31%11.06%26.33%8.02%
 Low income7.84%3.88%60.43%8.84%23.81%8.16%
 Middle income5.98%3.06%60.12%6.09%19.51%8.23%
 High income5.62%2.84%57.31%4.84%18.22%9.52%
Ethnicity      
 Non-Hispanic6.58%3.43%58.09%6.23%21.20%9.22%
 Hispanic4.05%1.90%65.68%6.79%12.17%3.95%
Table 2.  Unadjusted mean age, EQ-5D, SF-12, VAS, and NCC with and without CMRFCs compared with national average (adult MEPS 2000 to 2002 with valid BMI)
 Without CMRFC (BMI ≥ 25 kg/m2)*With CMRFC (BMI ≥ 25 kg/m2)*With CMRFC (BMI ≥ 30 kg/m2)*U.S. average
  • SF-12, Short Form-12; VAS, Visual Analogue Scale; NCC, number of chronic conditions; CMRFC, cardiometabolic risk factor cluster; MEPS, Medical Expenditure Panel Survey; PCS, Physical Component Summary; MCS, Mental Component Summary.

  • *

    CMRFC = BMI ≥ 25 kg/m2 (or BMI ≥ 30 kg/m2) and at least two of the following risk factors: diabetes, high blood pressure, and/or hyperlipidemia.

EQ-5D0.880.760.730.87
SF-12 PCS49.9140.2838.6149.29
SF-12 MCS51.3249.4848.7451.20
VAS80.968.065.0080.1
Age44626045
NCC1.615.295.531.84

When interpreting results, it is helpful to consider the minimum clinically important difference in conjunction with statistical significance and effect size (27). While the EQ-5D index minimal clinically important difference (MCID) has not been established, previous research has estimated an MCID of 0.033 to 0.074. The former is based on an effect size of 0.2 in MEPS (11); the latter is based on changes using the SF-36 as an anchor in 11 different patient populations (28). For comparison purposes, previous research has found the MCID of other use instruments to be 0.03 to 0.041 for the SF-6D (28, 29) and 0.03 for the Health Utilities Index (30). The SF-36 MCID has been estimated to be ∼3 to 5 points (31).

After controlling for comorbidity burden, age, sex, race, ethnicity, education, smoking status, and income, the marginal impact of cardiometabolic risk factor clusters was highly statistically significant across all four HRQL measures and seemed to be clinically significant for all but the MCS-12 (Table 3). The PCS-12 showed a greater decrease in HRQL associated with physical function–related domains of the SF-12 compared with mental function–related domains measured by the MCS-12. Using BMI ≥30 kg/m2 as the cut-off resulted in greater negative magnitude of the estimates compared with BMI ≥25 kg/m2 for all four instruments. As expected, older age, lower income and education, and greater comorbidity burden were associated with lower HRQL across all four instruments (Tables 4 and 5). Hispanic ethnicity did not have a statistically significant impact on HRQL. Black, Native American, and other race were not statistically significant for most instruments and were small in magnitude where statistically significant. Male sex was not statistically significant for most instruments, and where statistically significant, the magnitude was very small.

Table 3.  Summary results* of multivariate analysis of four separate regressions, PCS-12, MCS-12, EQ-5D, and VAS, on CMRFCs, controlling for sociodemographic characteristics and comorbidity burden
 PCS-12SEMCS-12SEEQ-5DSEVASSE
  • PCS, Physical Component Summary; MCS, Mental Component Summary; VAS, Visual Analogue Scale; CMRFC, cardiometabolic risk factor cluster; SE, standard error. All coefficients are p < 0.01.

  • *

    Excerpts from the full regression results in Tables 4 and 5.

  • CMRFC = BMI ≥ 25 kg/m2 (or BMI ≥ 30 kg/m2) and at least two of the following risk factors: diabetes, high blood pressure, and/or hyperlipidemia.

CMRFC (BMI ≥ 25 kg/m2)−3.480.26−0.8660.26−0.0380.00−6.100.42
CMRFC (BMI ≥ 30 kg/m2)−4.870.33−1.030.37−0.0540.00−8.550.41

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This study shows that cardiometabolic risk factor clusters seem to have a significant deleterious impact on HRQL in a nationally representative sample of U.S. adults, after controlling for comorbidity and sociodemographic characteristics. In addition, cardiometabolic risk factor clusters seem to have a more deleterious impact when BMI ≥30 kg/m2 is used as the cut-off compared with BMI ≥25 kg/m2.

There have been previous examinations of the individual impact of cardiovascular risk factors on HRQL. Sullivan et al. (11) estimated the impact of chronic conditions on EQ-5D index scores in the nationally representative MEPS sample and found that hypertension, hyperlipidemia, and diabetes, among many other chronic conditions, had a significant deleterious impact on EQ-5D scores in the United States. They also found that comorbidity had both a statistically and clinically significant impact on EQ-5D scores but did not assess specific cardiometabolic risk factor clusters. When assessing the impact of individual chronic conditions, other research has also shown the negative impact of diabetes (32), obesity (33), and hypertension (34) on EQ-5D index scores (33). In addition, the individual impact of obesity, hypertension, and diabetes have been shown to negatively affect HRQL as measured by the SF-12 and VAS in the nationally representative MEPS sample (33). Using MEPS, Jia and Lubetkin (33) found overweight, Class I, and Class II obesity to have the following negative magnitudes on PCS-12 (−0.73, −1.86, −4.0); EQ-5D (−0.013, −0.033, −0.073), and VAS (−0.52, −3.23, −4.84) scores, respectively. The estimates of Jia and Lubetkin for Class II obesity (BMI ≥ 35 kg/m2) are similar (PCS-12; VAS) or larger (EQ-5D) than the estimates of cardiometabolic risk factor clusters (CMRFCs) using BMI ≥25 kg/m2 found in this research but smaller than the estimates of CMRFC using BMI ≥30 kg/m2 as the criterion. However, the estimates may not be comparable because Jia and Lubetkin did not adequately control for chronic comorbidities, resulting in a likely increase in the magnitude of the estimates. There is a clear link between chronic disease and obesity, and not controlling for comorbidities may confound the magnitude of the estimates for obesity. In contrast, this research controlled for all chronic comorbidities. Nonetheless, it is interesting to note that the magnitude of the estimates for Class II obesity reported in Jia and Lubetkin is comparable with our estimates of CMRFC using the criteria of BMI ≥25 kg/m2.

Hypertension has also been shown to negatively impact HRQL as measured by the SF-36 and the time trade-off measure (35). Examinations of the impact of hyperlipidemia have had mixed results, however. Schlenk et al. (36) found no difference in SF-36 scores between normal adults and those with hyperlipidemia. Lalonde et al. (37) examined the impact of hypertension and/or hyperlipidemia on HRQL as measured by preference-based and non–preference-based generic HRQL instruments. They found that the existence of hypertension and/or hyperlipidemia was associated with lower General Health Perception on the SF-36 but not associated with lower HRQL on the SF-36 PCS or MCS or preference-based HRQL as measured by the time trade-off or standard gamble techniques. In a separate study, Lalonde et al. (38) found a negative association between SF-36 scores and hypertension but a positive association with hyperlipidemia. Sullivan et al. (11) found a statistically significant and negative, but very small magnitude (−0.0047), effect of hyperlipidemia on EQ-5D scores after controlling for other chronic conditions and sociodemographic characteristics in a nationally representative sample. In summary, the individual impact of hyperlipidemia on HRQL is unclear and may be confounded by comorbidity and other sociodemographic factors. However, previous research has consistently shown the deleterious individual impact of hypertension, diabetes, and overweight/obesity on HRQL.

Although there are many studies examining the individual impact of separate cardiometabolic risk factors on HRQL, there are very few published studies examining the combined impact of cardiometabolic risk factors on HRQL, and no studies to the authors’ knowledge of the impact of the most common cardiometabolic risk factor clusters (diabetes, overweight/obesity, hyperlipidemia, and hypertension) on generic and preference-based HRQL as conducted in this research. The use of three different instruments provides important information on the impact of cardiometabolic risk factor clusters on HRQL. The SF-12 is a non–preference-based measure of generic quality of life and provides a broad picture of subjects’ generic HRQL. In contrast, the EQ-5D index is a preference-based measure of generic HRQL and provides a preference-based score that can be used to calculate quality-adjusted life years for cost-effectiveness analyses of treatment or prevention of cardiometabolic risk factor clusters. In addition, the VAS provides an important assessment of self-perceived health status and a subjective measure of preferences for current health status. This research quantifies the marginal impact of cardiometabolic risk factor clusters on each of these instruments after controlling for comorbidity and sociodemographic characteristics.

The finding that cardiometabolic risk factors have a relatively larger impact on physical function–related HRQL is consistent with other studies (39). Although the magnitude was small, it is interesting to note that the impact of cardiometabolic risk factor clusters on mental function–related HRQL was highly statistically significant in this analysis.

While it is clear that diabetes, hypertension, hyperlipidemia, and overweight/obesity are prone to cluster and result in an elevated risk of cardiovascular disease, there is some controversy about whether these cardiometabolic risk factor clusters reflect an underlying syndrome thought to be related to insulin resistance (commonly referred to as the metabolic syndrome) or simply that they tend to cluster (4). This analysis examines the impact of these cardiometabolic risk factor clusters on HRQL, irrespective of the underlying pathophysiology.

This analysis is not without limitations. Although higher than those estimated by Oldridge et al. (7), the diabetes, overweight/obesity, hypertension, and hyperlipidemia prevalence estimates presented here are consistent with other survey-based national-level estimates in the United States (40, 41, 42, 43, 44). However, it is likely that the prevalence rates in this study are underestimates of national prevalence. First, similar to the National Health Interview Survey and the Behavioral Risk Factor Surveillance System, MEPS is based on self-report. There is evidence that self-reported conditions may be under-reported in general (45) and that blacks, whites, and Hispanics differ in reporting of disease labeling and levels of illness and disability (46, 47, 48). Second, previous studies have shown that overweight respondents tend to underestimate their weight and overestimate their height (49, 50). Third, unlike the National Health and Nutrition Examination Survey, MEPS does not contain information on undiagnosed diabetes, hypertension, or hyperlipidemia, which contributes to the likelihood that the estimates from this analysis are biased downward. Recent estimates suggest that ∼35% of individuals with diabetes have not been diagnosed (51). As a result, it is likely that the national prevalence of cardiometabolic risk factor clusters estimated in this analysis is an underestimate. Hence, the impact of cardiometabolic risk factor clusters on HRQL is likely underestimated to a similar degree. In addition, for individuals who cannot complete the self-administered questionnaire, MEPS allows the use of proxy respondents (most commonly the wife, daughter, or mother of the individual). Proxy respondents may rate health differently than the individuals themselves, and this may introduce a source of bias. Approximately 12% of all MEPS responses to the self-administered questionnaire were completed by proxies, whereas ∼13% of responses for individuals with CMRFC were completed by proxies. This slightly higher percentage is not surprising given the greater age and morbidity level associated with CMRFC: individuals with CMRFC are more likely to have health limitations that would preclude them from filling in the self-administered questionnaire than the general population. Nonetheless, this may be a limitation to our research. Previous studies have found Hispanics to have a higher prevalence of diabetes, hyperlipidemia, and overweight/obesity (52, 53, 54, 55, 56). These results are inconsistent with these findings. There is evidence that blacks, whites, and Hispanics differ in self-reporting of disease labeling and levels of illness and disability (46, 47, 48). This study provides further evidence of this phenomenon. Other sources that may contain more specific diagnosis and laboratory data (such as fasting glucose levels) do not contain the rich array of HRQL data available in MEPS. Another potential limitation is the necessary exclusion of individuals without complete data in the multivariate analyses. Despite the limitations, this study provides important evidence of the impact of cardiometabolic risk factor clusters on HRQL in a nationally representative population and shows that common cardiometabolic risk factor clusters such as overweight/obesity, diabetes, hypertension, and hyperlipidemia have a significant and negative impact on HRQL in the United States.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This research was supported by an unrestricted grant from Sanofi-Aventis.

Footnotes
  • 1

    Nonstandard abbreviations: HRQL, health-related quality of life; SF, short form; VAS, visual analogue scale; MEPS, Medical Expenditure Panel Survey; HC, household component; NHIS, National Health Interview Survey; ICD-9, International Classification of Diseases-9; CCC, clinical classification category; NCC, number of chronic conditions; PCS-12, Physical Component Summary; MCS-12, Mental Component Summary; CLAD, censored least absolute deviations estimator; MCID, minimal clinically important difference; CMRFC, cardiometabolic risk factor cluster.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    U.S. Department of Health and Human Services Centers for Disease Control and Prevention (2005) Preventing Heart Disease and Stroke: Addressing the Nation's Leading Killers Centers for Disease Control and Prevention Washington, DC.
  • 2
    Smith, H. J., Taylor, R., Mitchell, A. (2000) A comparison of four quality of life instruments in cardiac patients: SF-36, QLI, QLMI, and SEIQoL. Heart 84: 390394.
  • 3
    Dempster, M., Donnelly, M. (2000) Measuring the health related quality of life of people with ischaemic heart disease. Heart 83: 641644.
  • 4
    Kahn, R., Buse, J., Ferrannini, E., Stern, M. (2005) The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 28: 22892304.
  • 5
    Wee, H. L., Cheung, Y. B., Li, S. C., Fong, K. Y., Thumboo, J. (2005) The impact of diabetes mellitus and other chronic medical conditions on health-related Quality of Life: is the whole greater than the sum of its parts? Health Qual Life Outcomes 3: 2.
  • 6
    Chambers, B. A., Guo, S. S., Siervogel, R., Hall, G., Chumlea, WC (2002) Cumulative effects of cardiovascular disease risk factors on quality of life. J Nutr Health Aging 6: 179184.
  • 7
    Oldridge, N. B., Stump, T. E., Nothwehr, F. K., Clark, DO (2001) Prevalence and outcomes of comorbid metabolic and cardiovascular conditions in middle- and older-age adults. J Clin Epidemiol. 54: 928934.
  • 8
    Guyatt, G. H., Feeny, D. H., Patrick, DL (1993) Measuring health-related quality of life. Ann Intern Med. 118: 622629.
  • 9
    Ware, J., Jr, Kosinski, M., Keller, SD (1996) A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care 34: 220233.
  • 10
    Shaw, J. W., Johnson, J. A., Coons, SJ (2005) US valuation of the EQ-5D health states: development and testing of the D1 valuation model. Med Care 43: 203220.
  • 11
    Sullivan, P. W., Lawrence, W. F., Jr, Ghushchyan, V. (2005) A national catalogue of preference-based scores for chronic conditions in the U.S. Med Care 43: 736749.
  • 12
    Gold, M. R., Siegel, J. E., Russell, L. B., Weinstein, M. C., eds. Cost-Effectiveness in Health and Medicine. 1996 Oxford University Press New York.
  • 13
    Froberg, D. G., Kane, RL (1989) Methodology for measuring health-state preferences–II: scaling methods. J Clin Epidemiol. 42: 459471.
  • 14
    Agency for Healthcare Research and Quality (AHRQ). PUF Documentation Files. http:www.meps.ahrq.govPUFFilesH60H60doc.htm (Accessed April 28, 2006).
  • 15
    Agency for Healthcare Research and Quality. Computing Standard Errors For MEPS Estimates. http:www.meps.ahrq.govmepswebsurvey_compstandard_errors.jsp (Accessed December 21, 2006).
  • 16
    Elixhauser, A., Andrews, R. M., Fox, S. (1993) Clinical Classifications for Health Policy Research: Discharge Statistics by Principal Diagnosis and Procedure Agency for Health Policy and Research Rockville, MD.
  • 17
    Greenlund, K. J., Zheng, Z. J., Keenan, N. L., et al (2004) Trends in self-reported multiple cardiovascular disease risk factors among adults in the United States, 1991–1999. Arch Intern Med. 164: 181188.
  • 18
    American Heart Association (2005) Heart Disease and Stroke Statistics—2005 Update American Heart Association Dallas, TX.
  • 19
    The EuroQol Group (1990) EuroQol—a new facility for the measurement of health-related quality of life. Health Policy 16: 199208.
  • 20
    McDowell, I., Newell, C. (1996) Measuring Health: A Guide to Rating Scales and Questionnaires 2nd ed. Oxford University Press New York.
  • 21
    The EuroQol Group EQ-5D. http:www.euroqol.orgdownloadref.pdf (Accessed April 28, 2006).
  • 22
    Ware, J., Jr, Kosinski, M., Keller, SD (1998) How to Score the SF-12 (r) Physical and Mental Health Summary Scales 3rd ed. Quality Metric Lincoln, RI.
  • 23
    Green, WH (2000) Econometric Analysis 4th ed. Prentice Hall Upper Saddle River, NJ.
  • 24
    Maddala, GS (1999) Limited-Dependent and Qualitative Variables in Econometrics Cambridge University Press Cambridge.
  • 25
    Powell, J. (1984) Least absolute deviations estimation for the censored regression model. J Econometrics 25: 303325.
  • 26
    Jolliffe, D., Krushelnytskyy, B., Semykina, A. Censored Least Absolute Deviations Estimator: CLAD. http:www.stata.comstbstb58 (Accessed January 18, 2007).
  • 27
    Guyatt, G. H., Osoba, D., Wu, A. W., Wyrwich, K. W., Norman, GR (2002) Methods to explain the clinical significance of health status measures. Mayo Clin Proc. 77: 371383.
  • 28
    Walters, S. J., Brazier, JE (2005) Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Qual Life Res. 14: 15231532.
  • 29
    Walters, S. J., Brazier, JE (2003) What is the relationship between the minimally important difference and health state utility values? The case of the SF-6D. Health Qual Life Outcomes 1: 4.
  • 30
    Grootendorst, P., Feeny, D., Furlong, W. (2000) Health Utilities Index Mark 3: evidence of construct validity for stroke and arthritis in a population health survey. Med Care 38: 290299.
  • 31
    Hays, R. D., Morales, LS (2001) The RAND-36 measure of health-related quality of life. Ann Med. 33: 350357.
  • 32
    Clarke, P., Gray, A., Holman, R. (2002) Estimating utility values for health states of type 2 diabetic patients using the EQ-5D (UKPDS 62). Med Decis Making 22: 340349.
  • 33
    Jia, H., Lubetkin, EI (2005) The impact of obesity on health-related quality-of-life in the general adult US population. J Public Health (Oxf). 27: 156164.
  • 34
    Roca-Cusachs, A., Dalfo, A., Badia, X., Aristegui, I., Roset, M. (2001) Relation between clinical and therapeutic variables and quality of life in hypertension. J Hypertens. 19: 19131919.
  • 35
    Lawrence, W. F., Fryback, D. G., Martin, P. A., Klein, R., Klein, BE (1996) Health status and hypertension: a population-based study. J Clin Epidemiol. 49: 12391245.
  • 36
    Schlenk, E. A., Erlen, J. A., Dunbar-Jacob, J., et al (1998) Health-related quality of life in chronic disorders: a comparison across studies using the MOS SF-36. Qual Life Res. 7: 5765.
  • 37
    Lalonde, L., Clarke, A. E., Joseph, L., Mackenzie, T., Grover, SA (1999) Comparing the psychometric properties of preference-based and non-preference-based health-related quality of life in coronary heart disease. Canadian Collaborative Cardiac Assessment Group. Qual Life Res. 8: 399409.
  • 38
    Lalonde, L., O'Connor, A., Joseph, L., Grover, SA (2004) Health-related quality of life in cardiac patients with dyslipidemia and hypertension. Qual Life Res. 13: 793804.
  • 39
    Rijken, M., van Kerkhof, M., Dekker, J., Schellevis, FG (2005) Comorbidity of chronic diseases: effects of disease pairs on physical and mental functioning. Qual Life Res. 14: 4555.
  • 40
    Druss, B. G., Marcus, S. C., Olfson, M., Tanielian, T., Elinson, L., Pincus, HA (2001) Comparing the national economic burden of five chronic conditions. Health Aff (Millwood). 20: 233241.
  • 41
    Harris, M. I., Flegal, K. M., Cowie, C. C., et al (1998) Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S. adults. Diabetes Care 21: 518524.
  • 42
    Mokdad, A. H., Ford, E. S., Bowman, B. A., et al (2003) Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA 289: 7679.
  • 43
    Engelgau, M. M., Geiss, L. S., Saaddine, J. B., et al (2004) The evolving diabetes burden in the United States. Ann Intern Med. 140: 945950.
  • 44
    U.S. Department of Health and Human Services Healthy People 2010 Progress Review: Physical Activity and Fitness. http:www.healthypeople.govdata2010progfocus22 (Accessed April 28, 2006.
  • 45
    Edwards, D. S., Winn, D. M., Kuriantzik, V. (1994) Evaluation of the National Health Interview Survey Diagnostic Reporting National Center for Health Statistics Hyattsville, MD.
  • 46
    Berkanovic, E., Telesky, C. (1985) Mexican-American, black-American and white-American differences in reporting illnesses, disability and physician visits for illnesses. Soc Sci Med. 20: 567577.
  • 47
    Stewart, A. L., Napoles-Springer, A. (2000) Health-related quality-of-life assessments in diverse population groups in the United States. Med Care 38 (9 Suppl): II10224.
  • 48
    Johnson, T. P., O'Rourke, D., Chavez, N., Sudman, S. (1996) Cultural variations in the interpretation of health survey questions Warnecke, RB eds. Health Survey Research Methods, 5762 National Center for Health Statistics Hyattsville, MD.
  • 49
    Palta, M., Prineas, R. J., Berman, R., Hannan, P. (1982) Comparison of self-reported and measured height and weight. Am J Epidemiol. 115: 223230.
  • 50
    Rowland, ML (1990) Self-reported weight and height. Am J Clin Nutr. 52: 11251133.
  • 51
    Centers for Disease Control and Prevention (2002) National Diabetes Fact Sheet: General Information and National Estimates on Diabetes in the US, 2000 Centers for Disease Control and Prevention Atlanta, GA.
  • 52
    Winkleby, M. A., Fortmann, S. P., Rockhill, B. (1993) Health-related risk factors in a sample of Hispanics and whites matched on sociodemographic characteristics. The Stanford Five-City Project. Am J Epidemiol. 137: 13651375.
  • 53
    Gaskill, S. P., Allen, C. R., Garza, V., Gonzales, J. L., Waldrop, RH (1981) Cardiovascular risk factors in Mexican Americans in Laredo, Texas. I. Prevalence of overweight and diabetes and distributions of serum lipids. Am J Epidemiol. 113: 546555.
  • 54
    Flegal, K. M., Ezzati, T. M., Harris, M. I., et al (1991) Prevalence of diabetes in Mexican Americans, Cubans, and Puerto Ricans from the Hispanic Health and Nutrition Examination Survey, 1982–1984. Diabetes Care 14: 628638.
  • 55
    Gardner, L. I., Jr, Stern, M. P., Haffner, S. M., et al (1984) Prevalence of diabetes in Mexican Americans. Relationship to percent of gene pool derived from native American sources. Diabetes 33: 8692.
  • 56
    Stern, M. P., Gaskill, S. P., Hazuda, H. P., Gardner, L. I., Haffner, SM (1983) Does obesity explain excess prevalence of diabetes among Mexican Americans? Results of the San Antonio Heart Study. Diabetologia 24: 272277.