Print version ISSN 0102-311X
Cad. Saúde Pública vol.23 n.7 Rio de Janeiro Jul. 2007
NOTA RESEARCH NOTE
Importância e custo das hospitalizações associadas ao sobrepeso e obesidade no Brasil
Rosely SichieriI; Sileia do NascimentoI; Walmir CoutinhoII
IInstituto de Medicina Social, Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brasil
IIInstituto Estadual de Diabetes e Endocrinologia Luiz Capriglione, Rio de Janeiro, Brasil
Os custos de hospitalização associados ao sobrepeso/obesidade e às doenças associadas no Brasil foram estimados utilizando-se os dados das hospitalizações de homens e mulheres de 20 a 60 anos do Sistema de Informações Hospitalares do Sistema Único de Saúde (SIH-SUS) para o ano de 2001. O SUS cobre mais de 70% das hospitalizações. A fração atribuível a hospitalizações associadas com obesidade/sobrepeso ou doenças relacionadas foi estimada com base na combinação dos riscos relativos de coortes americanas e européias. O custo direto total do sobrepeso/obesidade foi estimado pela soma do risco atribuível à população para cada morbidade multiplicada pelo valor de reembolso de cada morbidade. O total de custos foi equivalente a 3,02% dos custos totais de hospitalização em homens e 5,83% em mulheres, correspondendo a 6,8 e 9,3% de todas as hospitalizações (excluindo gestantes). O excesso de peso no Brasil tem um impacto nas hospitalizações e nos custos similar ao observado em países desenvolvidos. Sendo a transição nutricional um processo em andamento no Brasil, o sobrepeso teve maior impacto nos custos do que a obesidade.
Obesidade; Sobrepeso; Hospitalização; Gastos em Saúde
This article estimates the burden of hospitalization associated with overweight and obesity in Brazil. The analysis of all hospitalizations for men and women from 20 to 60 years of age was based on the National Healthcare Expenditure Database (SIH-SUS), covering more than 70% of all hospital admissions. Data were for the year 2001. Attributable fraction of hospitalizations associated with diseases related to obesity and overweight was based on the combined risks of United States and European cohorts. The population-attributable fraction for each disease studied was multiplied by values reimbursed to the hospitals and summed to obtain total direct costs. Overall costs of overweight and obesity represent 3.02% of total hospitalization costs for men and 5.83% for women, corresponding to 6.8 and 9.3% of all hospitalization (excluding pregnancy). Diseases associated with overweight and obesity had a significant impact on hospitalizations and economic costs in Brazil, and overall percentages were similar to those from developed countries. Since the nutritional transition is still under way in Brazil, overweight had a higher impact than obesity on disease prevalence and costs.
Obesity; Overweight; Hospitalization; Health Expenditures
The increasing prevalence of obesity in three nationwide Brazilian surveys over the last 20 years 1 and the high prevalence of chronic diseases associated with obesity, such as stroke 2, diabetes 3, and cardiovascular diseases 4 suggest a significant burden of disease from overweight in the country. Monteiro et al. 5 identified clear changes from undernutrition to overnutrition between 1975 and 1996 in Brazil, but few studies have evaluated the implications of overweight for the national health systems of Latin America countries. A population-based survey conducted in the South of Brazil among individuals aged 20-69 years estimated that treatment for hypertension consumes approximately 23% of per capita income for hypertensive individuals 6. The total estimated annual cost associated with diabetes in Latin America in 2000 was US$65,216 million 7, but there are no overall figures for the burden of overweight.
Economic studies on the costs of obesity in developed countries suggest that preventing weight gain may help contain the rapidly rising healthcare costs 8. Reviews of cost-of-illness studies on obesity indicate that the annual burden represented 5.5 to 7.8% of healthcare expenditure in the 1990s 9, and a more recent analysis of 10 studies reported that the burden of obesity for national health systems accounts for 5.5-7.0% of national health expenditures in the United States and 2.0-3.5% in other developed countries 10. Indeed, Sturm & Wells 11 and Sturm 12 reported that obesity-related morbidity is greater than that associated with smoking, drinking, or poverty in the United States.
Although the burden of obesity and associated medical conditions varies across studies, the direct monetary implications of obesity for healthcare expenditures are based on the fact that overweight and obesity are known risk factors for highly prevalent conditions such as cardiovascular diseases, hypertension, and type 2 diabetes. Indirect costs of obesity to society come from losses in productivity due to morbidity and mortality, while intangible costs include poorer health-related quality of life such as social rejection, commuting difficulties, and psychological problems.
Brazil's National Healthcare Expenditure System (SIH-SUS) 13 covers more than 70% of all hospitalizations, but there are no longitudinal data on the risk of hospitalization or overall illness associated with obesity in the country. We thus estimated the burden of hospitalization associated with overweight and obesity, assuming that the relative risk of developing medical conditions associated with obesity and overweight in large cohorts of developed countries was similar to the risk of being hospitalized for these conditions in Brazil.
Material and methods
Design and subjects
The SIH-SUS covers 75.5% of all hospitalizations in the country according to the National Household Sample Survey (PNAD 1998) 14. Individual records in the SIH-SUS include the main cause of hospitalization, length of stay, age, and sex. The database also includes the values reimbursed to the hospital for each cause of hospitalization or procedures. Throughout the paper these will be referred to jointly as cost of hospitalization.
Diseases used in our analysis were those with well-established epidemiological associations with obesity 1,2,3,4. For the year 2001 we counted all hospitalizations and all days of hospitalization with the following main three-digits diagnoses 15: obesity (E66), diabetes (includes 1 and 2 as in the database) (E10 and E11), hypertension (I10), cholelithiasis (K80) and cholecystitis (K81), myocardial infarction (I21), other ischemic heart diseases (I24), stroke (I64), and colon cancer (C18), for both men and women 20 to 60 years of age. Three-digit diagnoses have greater reliability than those with four digits 16.
Individuals more than 60 years of age were excluded from the analysis for three reasons: (1) Calculation of relative risk was based on middle-aged individuals from large cohorts 17,18,19; (2) Since obesity is a recent public health problem in Brazil, older people may still not have suffered the chronic consequences of obesity 1; and (3) With aging, overweight tends to become relatively less important as a cause of mortality 20.
Prevalence of obesity and overweight
Data on prevalence of obesity and overweight were based on the World Health Organization (WHO) 21 definition and obtained from a Brazilian nationwide survey on living standards (PPV) in 1996-1997 22. Body mass index (BMI) of 25-29.9kg/m2 was classified as overweight and > 30kg/m2 was classified as obese. The PPV is a probabilistic multistage household survey that was based on the World Bank design for the international survey on living standards. Weight and height were measured at the subject's home using standard methods.
The attributable fraction of hospitalization associated with obesity and overweight-related diseases was based on the incidence observed in large cohorts from the United States and Europe, which presented data broken down by BMI. The nurses' cohort and health professional cohort 18 estimated risks were chosen preferentially when they were not far from other cohorts, because the reports from these two cohorts show a 10-year risk of developing all of the diseases we were investigating. Also, for both cohorts, risk values were presented by sex and BMI category, which allowed weighted calculations of the sex-specific relative risk of illness associated with BMI, further explained in this section. Gallstone relative risk was estimated based on NHANES I follow-up data, which is a population-based cohort from the United States 17; for hypertension, the relative risks of two large studies were combined 19,23.
The odds ratio of the nurses' cohort and health professional cohort for diabetes, cholelithiasis and cholecystitis, colon cancer, heart disease, and stroke for both cohorts were adjusted for age, smoking, and race 18, and our study transformed them into relative risks 24. Weighted sex-specific relative risks were averaged to estimate the overall RR for BMI > 30, using the RR for BMI 30-34.9 and BMI > 35. Weighted RR was chosen because the prevalence and relative risks showed large differences with increasing BMI. Obesity and overweight prevalence rates (P) were obtained from the latest Brazilian survey (PPV) 22. The population-attributable fraction (PAF) for obesity was computed for each disease condition, using the formula: (P) (RR - 1)/ [P (RR - 1) + 1] (Table 1). The PAF for each disease studied was multiplied by the total cost of the disease.
Overweight prevalence was approximately 30% in both sexes, and obesity was twice as frequent in women (13%) than in men (7%). Diabetes showed the greatest population-attributable risk (PAR): 38% for overweight and 28% for obesity among men, and about 40% for both BMI categories among women. For cholelithiasis and cholecystitis, PAR was more than twice as great for women than for men, and for both sexes the attributable risks for most diseases were greater for overweight than for obesity (Table 1).
The all-cause hospitalization rate was 4.1% in men and 4.2% in women. These values were lower than the overall hospitalization rate found in a population-based survey in the city of Rio de Janeiro in 1996 25 (5% for men and 5.8% for women). The main causes of hospitalization in Brazil were respiratory tract (32%), all cardiovascular (20%), digestive tract (18%), and infectious diseases (16%) (data not shown).
The estimated hospitalization rate attributable to overweight (per 100,000) was 60.7 for men and 107.2 for women; for obesity, the figures were 37.5 for men and 130.8 for women (Table 2). Total hospitalizations in Table 2 correspond to 6.8% of all admissions for men and 9.3% for women. Annual workdays lost due to hospitalization attributable to overweight/obesity was estimated by multiplying average length of stay by the number of cases of disease attributable to overweight/obesity. The greatest loss was due to diabetes, and the workdays lost in 2001 totaled 289,625 for men and 795,324 for women (Table 2).
More than half of hospitalization costs were due to myocardial infarction and other ischemic heart diseases (Table 3). The largest proportion of costs was attributable to overweight for both men and women, and costs for women were greater than for men. Overall costs of overweight and obesity accounted for 3.02% (1.94 + 1.08) of total hospitalization costs in men 20-60 years of age and 5.83 % in women (Table 3).
Although obesity is now recognized as a major public health problem in developing countries, the impact of overweight and obesity on healthcare costs has not been evaluated.
Our attempt to quantify the burden of overweight/obesity in Brazil used national data on hospitalizations. Other studies have used hospitalization rates to estimate disease burden. For example, the United States National Hospital Discharge Survey data (1979-1999) were used to analyze changes in obesity-associated diseases and economic costs in youth. Diabetes, obesity, sleep apnea, and gallbladder disease were examined to explore the trend in burden of the disease listed as primary or secondary diagnosis. During this period, the percentage of discharges with diabetes nearly doubled, obesity and gallbladder diseases tripled, and sleep apnea increased fivefold, with annual hospital costs increasing more than threefold 26.
Hospitalizations were also analyzed in a population-based study in the United States, where five classifications of overweight, based on BMI, were compared for their ability to predict subsequent all-cause hospitalization and mortality. In this analysis, the WHO classification of overweight was a better predictor of 12-year hospitalization than a death predictor, with a population-attributable risk of hospitalization of 3.4% in men and 3.9% in women 27.
The present study used a prevalence-based burden-of-obesity approach, assuming relative risks associated with overweight/obesity from secondary sources. A similar but hypothetical analysis for a large managed care plan in the Pacific Northwest region of the United States estimated the economic burden of diseases for which obesity is an established risk factor, such as coronary heart disease, hypertension, hypercholesterolemia, gallbladder disease, stroke, type 2 diabetes, osteoarthritis of the knee, and endometrial cancer. Results of the study, which included individuals 35 to 84 years of age, showed that obesity accounted for approximately 45% of all cases of hypertension, 85% of type 2 diabetes, 18% of hypercholesterolemia, and 35% of coronary heart disease, with costs attributable to obesity reaching 41% of total healthcare costs 28. Our results are less striking than the estimated burden of obesity in this managed care setting, but are still substantial considering that the nutritional transition in Brazil is still under way. Obesity prevalence in Brazil is less than one-fourth that of the United States, but Brazilian surveys have shown a greater and increasing prevalence of overweight and increasing prevalence of obesity 1. Our results thus indicate that the burden and costs associated with overweight are greater than the burden of obesity, mainly for men. Among women, costs associated with obesity and overweight were nearly the same.
Hospitalization was used as a proxy for morbidity in the present analysis, because it has the advantage of using an inclusive national database in Brazil, although also displaying important limitations. First, as shown by data on overall hospitalization in Rio de Janeiro, men were at lower risk than women 25, but overall mortality and obesity-related mortality were higher among men 29. Therefore, hospitalization rates associated with overweight/obesity among men may underestimate the burden of overweight/obesity.
Other possible limitations of the study involve the quality of the SIH-SUS data due to: (1) incorrect information registered on medical records; (2) incomplete data; (3) lack of training and unfamiliarity with coding rules among hospital employees; and (4) the fact that principal diagnoses may have been picked based on the amount of reimbursement for them 16,30,31. In addition, since no data are available on relative risks based on Brazilian cohorts, attributable risks may not reflect the burden of the diseases in the country. We chose the relative risk associated with overweight and obesity from United Stares cohorts 18, and these values are quite similar to the Europe figures 19,23. However, in the United States, obesity is less associated with mortality among blacks 20 as compared to whites, an important consideration given the extensive racial admixture in Brazil.
Additionally, for several reasons, this study represents a conservative estimate of the burden associated with obesity. First, some important obesity-related diseases were not included in the database search, including several types of cancer other than colon cancer. Second, if the age range were extended beyond 60 years, a significant increase in hospitalization would be expected.
Despite these methodological limitations, our study indicates that overweight and obesity have an important impact on hospitalization in Brazil, representing a major share of total healthcare costs and more than a million workdays lost in 2001.
The percentage of costs, as previously discussed, tends to be underestimated, but the Brazilian costs are still within the range observed in other studies 8. The estimated direct cost attributable to obesity (BMI > 30kg/m2) in a French national survey was 0.7 to 1.5% of total health expenditures 32, and data from developed countries other than the United States indicate expenditure in the 2.0-3.5% range 10, while our hospitalization data indicated that obesity-related costs accounted for at least 3% of total hospitalization costs in men and 5% in women.
In conclusion, among all hospital discharges in 2001 in Brazil, the proportion of discharges with overweight/obesity-associated diseases had a significant impact, with overweight greater than obesity. These findings may reflect the nutritional transition still under way in Latin American countries and indicate that the burden of these diseases tends to increase due to the growing prevalence and severity of obesity. Also, since the greatest increase in obesity prevalence rates has been reported among lower-income groups in the Brazilian population, the results point to an alarming scenario for the country's healthcare system in the near future.
R. Sichieri contributed to the project design, collecting, analyzing, and interpreting the data, and drafting the manuscript. S. Nascimento conducted the literature search and data collection, analysis, and interpretation. W. Coutinho participated in designing the project and interpreting the data.
Funding included a scholarship from the Carlos Chagas Filho Rio de Janeiro State Research Foundation and a grant from Abbott do Brazil.
1. Monteiro CA, Benício MHD'A, Conde WL, Popkin BM. Shifting obesity trends in Brazil. Eur J Clin Nutr 2000; 54:342-6. [ Links ]
2. Guimarães AC. Hypertension in Brazil. J Hum Hypertens 2002; 16 Suppl 1:S7-10. [ Links ]
3. Barbosa RB, Barcelo A, Machado CA. National campaign to detect suspected diabetes cases in Brazil: a preliminary report. Rev Panam Salud Pública 2001; 10:324-7. [ Links ]
4. Lotufo PA. Increasing obesity in Brazil: predicting a new peak of cardiovascular mortality. São Paulo Med J 2000; 118:161-2. [ Links ]
5. Monteiro CA, Conde WL, Popkin BM. The burden of disease from undernutrition and overnutrition in countries undergoing rapid nutrition transition: a view from Brazil. Am J Public Health 2004; 94:433-4. [ Links ]
6. Costa JSD, Fuchs SC, Olinto MTA, Gigante DP, Menezes AMB, Macedo S, et al. Cost-effectiveness of hypertension treatment: a population-based study. São Paulo Med J 2002; 120:100-4. [ Links ]
7. Barceló A, Aedo C, Rajpathak S, Robles S. The cost of diabetes in Latin America and the Caribbean. Bull World Health Organ 2003; 81:19-27. [ Links ]
8. Wolf AM, Colditz GA. The cost of obesity: the US perspective. Pharmacoeconomics 1994; 5 Suppl 1:34-7. [ Links ]
9. Kortt MA, Langley PC, Cox ER. A review of cost-of-illness studies on obesity. Clin Ther 1998; 20:772-9. [ Links ]
10. Thompson D, Wolf AM. The medical-care cost burden of obesity. Obes Rev 2001; 2:189-97 [ Links ]
11. Sturm R, Wells KB. Does obesity contribute as much to morbidity as poverty or smoking? Public Health 2001; 115:229-35. [ Links ]
12. Sturm R. The effects of obesity, smoking, and drinking on medical problems and costs. Health Aff (Millwood) 2002; 21:245-53. [ Links ]
13. Departamento de Informação e Informática do SUS. Sistema de informação hospitalar. Brasília: Ministério da Saúde; 2002. [ Links ]
14. Instituto Brasileiro de Geografia e Estatística. Pesquisa Nacional por Amostra de Domicílios acesso e utilização de serviços de saúde 1998. Rio de Janeiro: Instituto Brasileiro de Geografia e Estatística; 1998. [ Links ]
15. Organização Mundial da Saúde. Classificação estatística internacional de doenças e problemas relacionados à saúde, 10ª revisão. v. 1. São Paulo: Centro Colaborador da OMS para a Classificação de Doenças em Português; 1995. [ Links ]
16. Veras CMT, Martins MS. A confiabilidade dos dados nos formulários de Autorização de Internação Hospitalar (AIH), Rio de Janeiro, Brasil. Cad Saúde Pública 1994; 10:339-55. [ Links ]
17. Sichieri R, Everhart JE, Roth HP. Low incidence of hospitalization with gallbladder disease among blacks in the United States. Am J Epidemiol 1990; 131:826-35. [ Links ]
18. Field AE, Coakley EH, Must A, Spadano JL, Laird N, Dietz W, et al. Impact of overweight on the risk of developing common chronic diseases during a 10-year period. Arch Intern Med 2001; 160:1581-6. [ Links ]
19. Sharabi Y, Grotto I, Huerta M, Grossman E. Susceptibility of the influence of weight on blood pressure in men versus women: lessons from a large-scale study of young adults. Am J Hypertens 2004; 17 (5 Pt 1):404-8. [ Links ]
20. Fontaine KR, Redden DT, Wang C, Westfall AO, Allison DB. Years of life lost due to obesity. JAMA 2003; 289:187-93. [ Links ]
21. World Health Organization. Obesity: preventing and managing the global epidemic. Geneva: World Health Organization; 1998. [ Links ](Report of a WHO Consultation on Obesity).
22. Instituto Brasileiro de Geografia e Estatística. Pesquisa Nacional por Amostra de Domicílios - pesquisa padrões de vida (PPV) 1996-1997 [CD-ROM]. Rio de Janeiro: Instituto Brasileiro de Geografia e Estatística; 1998. [ Links ]
23. Huang Z, Willett WC, Manson JE, Rosner B, Stampfer M, Speizer FE, et al. Body weight, weight change, and risk for hypertension in women. Ann Intern Med 1998; 128:81-8. [ Links ]
24. Zhang J, Yu KF. What's the relative risk? A method for correcting the odds ratio in cohort studies of common outcomes. JAMA 1998; 280:1690-1. [ Links ]
25. Afonso FM, Sichieri R. Associação do índice de massa corporal e da relação cintura/quadril com hospitalizações em adultos do município do Rio de Janeiro - RJ. Rev Bras Epidemiol 2002; 5:153-63. [ Links ]
26. Wang G, Dietz WH. Economic burden of obesity in youths aged 6 to 17 years: 1979-1999. Pediatrics 2002; 109:E81. [ Links ]
27. Sichieri R, Everhart JE, Hubbard VS. Relative weight classifications in the assessment of underweight and overweight in the United States. Int J Obes Relat Metab Disord 1992; 16:303-12. [ Links ]
28. Oster G, Edelsberg J, O'Sullivan AK, Thompson D. The clinical and economic burden of obesity in a managed care setting. Am J Manag Care 2000; 6:681-9. [ Links ]
29. Fundação Nacional de Saúde. Sistema de informações sobre mortalidade (SIM). Brasília: Ministério da Saúde; 2002. [ Links ]
30. Mathias TAF, Soboll MLMS. Confiabilidade de diagnósticos nos formulários de autorização de internação hospitalar. Rev Saúde Pública 1998; 32:526-32. [ Links ]
31. Bittencourt SA, Camacho LAB, Leal MC. O Sistema de Informação Hospitalar e sua aplicação na saúde coletiva. Cad Saúde Pública 2006; 22:19-30. [ Links ]
32. Detournay B, Fagnani F, Phillippo M, Pribil C, Charles MA, Sermet C, et al. Obesity morbidity and health care costs in France: an analysis of the 1991-1992 Medical Care Household Survey. Int J Obes Relat Metab Disord 2000; 24:151-5. [ Links ]
Instituto de Medicina Social
Universidade do Estado do Rio de Janeiro
Rua São Francisco Xavier 524, 7º andar, Bloco E, sala 7002, Rio de Janeiro, RJ
Submitted on 09/Aug/2005
Final version resubmitted on 27/Oct/2006
Approved on 09/Jan/2007