Obesity Is Associated with Reduced Self-Rated General Health Status: Evidence from a Representative Sample of White, Black, and Hispanic Americans

• Published in: Preventive medicine 2001-05, Vol.32 (5), p.429-436
• Main Authors: Okosun, Ike S, Choi, Simon, Matamoros, Tara, Dever, G.E.Alan
• Format: Article
• Language: English
• Subjects: African Americans - psychology; Alcohol Drinking; BMI; European Continental Ancestry Group - psychology; Female; Health Status; Hispanic Americans - psychology; Humans; Logistic Models; Male; NHANES III; obesity; Obesity - classification; Obesity - complications; Obesity - psychology; Perception; Self Concept; Self-Assessment; self-perceived health; self-rated health; self-reported health; Smoking; United States
• ISSN: 0091-7435; 1096-0260
• DOI: 10.1006/pmed.2001.0840

Background. Although the evidence linking obesity with many chronic diseases is well established, the relationship with self-rated health is not clear. Self-rating of health is a broad summary measure of different domains of health that include psychosocial domain. The aims of this study were to examine the relationship between obesity and self-rated health and the degree of agreement between patients' self-rated health status and physicians' impression of patients' health in a representative sample of healthy noninstitutionalized American adults. Methods. Data (n = 10, 298) used for this analysis were obtained from the Third National Health and Nutrition Examination Survey. Normal weight was defined as BMI 18.5–24.9 kg/m2 and preobesity was defined as BMI 25–29.9 kg/m2. Moderate (class I) obesity was defined as BMI 30–34.9 kg/m2 and severe (class II) obesity as BMI ≥35 kg/m2. Self-rated health conditions were categorized as excellent, very good, good, fair, or poor. Agreement between physicians' impression and self-rated health and a linear relationship between obesity and individual perceived health were quantified using κ and γ statistics, respectively. Using the logistic regression analysis, odds of reporting reduced health in preobese, class I obese, and class II obese individuals were compared with those for normal-weight individuals, adjusting for age, current smoking, and alcohol intake. The contribution of obesity to ethnic differences in reduced self-rated health was determined by comparing blacks with whites and Hispanics with whites fitted in multiple logistic regression models. Results. Among both men and women, there was a statistically significant linear association between obesity and self-rated health (P < 0.05). The proportion of subjects reporting excellent health tended to decrease with increasing level of obesity in the three ethnic groups. In the three ethnic groups, the degrees of concordance between self-rated health and physicians' impressions were poor and decreased with increasing obesity level. In each ethnic group, class II obesity was associated with approximately twofold increased odds of reporting reduced health compared with normal-weight individuals. Compared to whites, black and Hispanic races/ethnicities were respectively associated with 23 and 175% increased odds of reduced self-rated health among men. The corresponding values for women were 45 and 177%, respectively. Conclusion. The results of this study provide eviden