Health-Related Quality of Life Varies among Obese Subgroups

Introduction

See Editorial by Fontaine on

In recent years, there has been extensive interest in the health-related quality of life (HRQOL) of obese persons (1,2,3,4). However, most of what is known about the HRQOL of obese persons has been derived from studying individuals in treatment settings (5,6,7,8), and much of the early research was conducted on individuals undergoing surgical intervention (9,10,11,12,13,14,15,16). Many obese persons do not seek treatment, and others seek treatment in commercial settings that do not conduct quality of life or outcome research. Therefore, it may be unwise to draw conclusions about the HRQOL of obese persons in general from the current reports in the literature. Research is needed on the HRQOL differences that exist between treatment-seeking and nontreatment-seeking populations of obese persons and among obese persons who seek treatment modalities that vary in type or intensity.

There is some research to date suggesting that there are differences in HRQOL among obese persons depending on how they are sampled. Obese persons who seek treatment are more likely than obese persons who do not seek treatment to experience psychological disturbance, eating disorders (17,18) and impaired HRQOL (19). Furthermore, obese persons seeking treatment in a hospital setting are more distressed than those seeking treatment in a community program (20), and obese persons who choose surgical treatment have poorer HRQOL than matched obese controls (matched on 18 variables) who choose conventional treatment (12). However, one study found that HRQOL factors were not predictive of treatment-seeking for obesity, but that demographic factors predicted which obese persons sought treatment (21).

This study was undertaken to compare the HRQOL of obese persons from different subgroups that varied in treatment modality and treatment intensity. The specific research questions were as follows: 1) are overweight/obese treatment-seekers more impaired on obesity-specific HRQOL than overweight/obese nontreatment-seekers; 2) does obesity-specific HRQOL vary among subgroups of obese persons seeking treatments that vary in modality and intensity; and 3) does obesity-specific HRQOL vary depending on variables such as body mass index (BMI), gender, or race?

Research Methods and Procedures

Results

BMI and Treatment Modality

Figure 1 displays the relationship between BMI category and treatment modality for the IWQOL-Lite total score. Two clear patterns emerge. First, as BMI increases, greater impairment (i.e., lower IWQOL-Lite scores) in obesity-specific HRQOL is observed. Second, within each BMI category, greater treatment intensity is associated with greater impairment in HRQOL. Figure 1 also shows that among the severely obese (BMI of 40 kg/m² and over), even nontreatment-seekers had greatly impaired HRQOL. The correlation between BMI and IWQOL-Lite total score was −0.527 (p < 0.001), indicating that almost 28% of the variance in IWQOL-Lite total score can be accounted for by BMI.

Discussion

This is a report of obesity-specific HRQOL in overweight/obese individuals from five distinct groups that differ in treatment modality and treatment intensity: overweight/obese nontreatment-seeking community volunteers, participants in clinical trials, participants in outpatient weight-loss programs/studies, day treatment program participants, and gastric bypass patients. Results of this study indicate that HRQOL differs across overweight/obese subgroups that vary by treatment-seeking status, treatment modality, gender, race, and BMI. Two notable strengths of this study are the large sample size (n = 3353 subjects) and the geographic and demographical diversity (subjects were from 13 different states in the U.S.; age range, 18–90 years; at least 14% African Americans; 32.6% men). A large, diverse sample enhances the potential generalization of our findings. Furthermore, we have used a well-established and validated measure of obesity-specific quality of life.

This study found statistically significant differences across treatment modalities on obesity-specific quality of life. As treatment modality intensified (from nontreatment-seeking Community Volunteers to Clinical Trials Participants to Outpatient Weight-Loss Programs/Studies Participants to Day Treatment Participants to Gastric Bypass Patients), all five scales and total score of the IWQOL-Lite became more impaired. Although the order of the Clinical Trial and Outpatient groups could be argued, the ordering we give here is supported by the data. To our knowledge, only two other studies have investigated HRQOL differences by treatment modality. In one study (20), obese persons seeking treatment in a hospital setting (gastric stapling, dietary counseling, or appetite suppressants) were more distressed on the Symptom Checklist–90 Revised (26) than those seeking treatment in a community program (Weight Watchers). In the other study (12), obese persons who chose surgical treatment had poorer HRQOL on a variety of measures than matched obese controls who chose conventional treatment. The results of these studies taken as a whole suggest that health-care providers and other individuals who provide treatment services or conduct research on obese persons need to be especially aware of the differences in HRQOL that exist across treatment modalities, with an eye toward managing the greater levels of distress and impairment that are apt to be found in those seeking the more intensive treatments.

Results of this study indicate that overweight/obese-treatment seekers were generally more impaired than overweight/obese nontreatment-seekers on obesity-specific quality of life. This finding is consistent with earlier research comparing obese treatment-seekers with obese nontreatment-seekers. Two previous studies found that treatment-seekers were more likely to experience psychological disturbance and eating disorders (17,18). Furthermore, obese treatment-seekers were found to have more impaired HRQOL than nontreatment-seekers on the SF-36, a generic measure of HRQOL (27), particularly on scales measuring bodily pain, general health perception, and vitality (19). Taken together, these results remind us to avoid making generalizations about obese persons as a group and to pay close attention to treatment-seeking status when drawing conclusions about obese individuals.

This study also found that overweight/obese treatment-seekers differed from their nontreatment-seeking counterparts on demographic variables. Treatment-seekers tended to be women, older, white, and heavier than nontreatment-seekers. These findings on demographic variables are identical to those reported by Fontaine (21), who compared obese individuals in an outpatient weight-loss program with obese volunteers who were not in treatment. However, in both of these studies (the present one and the one by Fontaine), the nontreatment-seeking subjects may not have been representative of overweight/obese persons in general in that the nontreatment-seekers were both samples of convenience. Thus, one weakness of the present study is that the nontreatment-seeking subjects were not a randomly derived sample from the population of overweight/obese persons in general; therefore, we cannot be certain whether their responses are representative of that group.

The relationship between BMI and generic HRQOL has been examined in population studies in the United States (28,29,30), Sweden (4), England (31), Australia (32), and France (33). A consistent finding in these studies is that increasing BMI is associated with impaired generic HRQOL, particularly for the physical aspects of quality of life. Similarly, obesity-specific quality of life (as assessed by the Obesity Specific Quality of Life Scale) was found to decrease as BMI increased in a population study in France (33). Furthermore, studies using subjects in treatment for overweight/obesity have also reported that increasing BMI is associated with greater HRQOL impairments (8,23,24). In the present study of treatment-seeking and nontreatment-seeking overweight/obese persons, impaired obesity-specific HRQOL was associated with increasing BMI, with BMI accounting for almost 28% of the variance in IWQOL-Lite score. Future research is needed to determine which factors account for the remaining 72% of the variance. For example, physical activity may improve HRQOL even at high BMI levels (29), and chronic illness may reduce HRQOL in obese persons (31). This study also found, as one might expect, that individuals with higher BMIs tended to seek treatments of greater intensities.

Studies of self-reported health status have shown that women tend to report more impaired health than men (34,35,36). Similarly, women have more impaired body image/body satisfaction than men (37,38,39), and obese women have more impaired HRQOL than obese men (13,40,41,42). Previous research on the IWQOL-Lite suggested that women have more impaired HRQOL than men, especially in individuals with BMIs of 35 kg/m² and below (23,24). In the present study, we found that women were more impaired than men on three of the five scales of the IWQOL-Lite and on total score. Furthermore, women in the Community group and in the Clinical Trials group had more impaired quality of life than men, but in the other treatment modalities, women's quality of life was comparable with men's. Because the Community group and the Clinical Trials group had the lowest BMIs, these results are consistent with previous research on the IWQOL-Lite. Clinicians treating overweight and mildly obese individuals need to be aware of women's poorer HRQOL relative to men, perhaps in part because of their tendency toward greater body dissatisfaction than men (43). Furthermore, researchers studying overweight/obese persons who are not in treatment need to be aware of the greater impact of weight on quality of life for women vs. men. Of note, we also found that men in the Gastric Bypass group had poorer HRQOL than women in that group. This unexpected finding may be related to the small sample size of men in this group, or it may be that men seek the most extreme treatment only when their quality of life is considerably impaired, compared with women. Differences in health care use by men and women are quite common, with men using the health care system less often than women (34,35,36,44).

This study found that white women and men were significantly more impaired on obesity-specific quality of life than African-American women and men, after controlling for age and BMI. In other studies reporting on race and quality of life in obesity, findings were mixed. Katz et al. (41) found that overweight/obese African Americans had more impaired HRQOL than whites on the SF-36, whereas Clarkson et al. (45) found no difference between overweight/obese African Americans and whites. Similarly, investigations of body satisfaction by racial group have yielded mixed results, with some studies finding greater body satisfaction for African Americans (38,46) and some finding no differences between African Americans and whites (39,43,47).

One weakness of this study is the overrepresentation of white subjects and the lack of data on race for 17% of the participants. Of those reporting race, only 17% of the sample was African American and only 6% were from other racial backgrounds. Therefore, we are limited in our conclusions regarding the HRQOL of persons from non-white groups. Another limitation of this study is that we did not collect data on the presence or absence of pain. The presence or absence of pain has been determined to be an important variable influencing HRQOL in obesity. Barofsky et al. (48) found that obese persons reporting pain were more impaired on HRQOL than obese persons not reporting pain.

We are planning research studies that examine HRQOL in obese persons with accompanying comorbidities such as diabetes or sleep apnea to determine the extent to which the presence of comorbidities in obese persons is responsible for HRQOL impairments. We are also interested in the role of physical activity in HRQOL and obesity. Future work with these data might include trying to predict treatment modality from demographic information and BMI using multinomial logit or trying to predict treatment- vs. nontreatment-seeking using ordinary logistic regression.

In conclusion, obesity-specific quality of life differs across overweight/obese subgroups that vary by treatment-seeking status and treatment modality. It seems that the degree of impairment in quality of life may have a strong influence on the type of treatment sought and whether treatment is sought. Furthermore, obesity-specific quality of life also varies with gender, race, and BMI, with women, whites, and persons with higher BMIs having the greatest impairments.

Acknowledgments

Financial support for this project was provided by Abbott Laboratories. Data from this study were presented in oral sessions at NAASO, 2001, Quebec City, Quebec Canada, and Nutrition Week, 2002, in San Diego, California.