Changes in body mass index and short-term healthcare expenditures in a Medicare population.

摘要

Purpose. The main objective of the study was to estimate the short-term changes in healthcare expenditures that result from weight gain or loss in a population of Medicare patients using longitudinal data. Changes in total healthcare expenditures and its components (inpatient, outpatient, prescription, dental and other) by changes in body mass index (BMI) categories were analyzed.;Methods. Design. The study design was longitudinal and retrospective.;Data. The Medicare Current Beneficiary Survey, a nationally representative survey of Medicare beneficiaries for the years 2000 through 2005 was used. Four panels of Medicare beneficiaries were followed for three years each.;Sample. Analytical sample included individuals: (a) who were community dwelling, (b) age 65 and older; (c) with full-year enrollment in Medicare parts A and B; and (d) with no missing BMI values for the years BMI changes were measured. After exclusions, the analytical sample consisted of 10,698 Medicare beneficiaries, with a similar distribution across panels: 2000-02 (n= 2,757); 2001-03 (n = 2,745); 2002-04 (n= 2,562); and 2003-05 (n= 2,634).;Measures -- Changes in BMI categories. Changes in BMI were measured between Year 1 and Year 2 and individuals were classified into 6 categories: (1) Stayed normal (BMI 18.5-24.9 kg/m2); (2) BMI loss (decrease in BMI in individuals with baseline BMI greater than or equal to 25 kg/m2, where the resulting BMI belongs to a category lower than that of baseline); (3) Stayed overweight (BMI between 25 and 29.9 kg/m2 in baseline and follow-up years); (4) Stayed obese (BMI greater than or equal to 30 kg/m2 in baseline and follow-up years); (5) BMI gain (increase in BMI category in individuals with baseline BMI > 18.5 kg/m2, where the resulting BMI belongs to a category higher than that of baseline) and (6) Stayed Underweight/Other (BMI 18.5 kg/m2 in baseline and follow-up years/weight changes not defined by previous categories). The group "BMI loss" included individuals who moved to a lower category of BMI between Year 1 and Year 2. For example, an individual moving from obese to overweight or from obese to normal BMI would have experienced BMI loss. Similarly, "BMI gain" included individuals who moved to a higher BMI category. For example, an individual moving from normal weight to overweight or overweight to obese would be classified as having BMI gain.;To examine the relationship between BMI changes and expenditures, three specifications of expenditures were used. The primary measure of expenditures was relative changes in expenditures transformed into log-ratio. Under the log-ratio approach positive values indicate increased expenditures, and negative values indicate decreased expenditures. Other measures included logged expenditures (Year 3) derived after BMI changes were measured and relative changes in healthcare expenditures (i.e. percent change in expenditures) grouped into (1) no change; (2) greater than 10% decline; (3) greater than 10% increase and (4) minimal variations.;Bivariate. The Chi-square tests of independence were used to determine significance between BMI change categories and subject characteristics. F-tests were used to test significant differences in average expenditures and BMI changes.;Multivariate. Multinomial Logistic Regression was used to analyze the relationships between BMI change categories and independent variables. Ordinary Least Squares (OLS) regressions were used to model logged expenditures and log-ratio of expenditures by changes in BMI, while controlling for other independent variables (age, gender, race/ethnicity, marital status, education, poverty level, metro status, supplemental insurance, general health, functional limitations, smoking and physical and mental illness). Multinomial logistic regressions were used to analyze the relationship between relative changes in healthcare expenditures by groups and BMI changes, after controlling for other independent variables measured in the study.;Results. Only a small fraction of the elderly experienced weight loss as measured by changes in BMI: Of 10,698 individuals, 9% experienced BMI loss (N=982) and 8.2% experienced BMI gain (N=867). Subgroup differences in BMI changes were noted: females were less likely than males to experience BMI loss or BMI gain [AOR 0.81, 95% CI 0.68-0.97; AOR 0.73 95%CI 0.60-0.90]. Increased age (80 + years) decreased the likelihood of BMI loss or BMI gain as compared to individuals 65-69 years [AOR 0.57, 95%CI 0.45-0.73; AOR 0.40 95%CI 0.31-0.52] ; individuals older than 80 years were the only age group with increased risk of staying underweight/other.;After controlling for all the independent variables measured in this study, results from OLS regression for short-term logged expenditures revealed that compared to persons who stayed in normal BMI group, individuals with BMI loss had outpatient expenditures that were 23% higher. Compared to individuals who belonged to stayed normal BMI group, individuals with BMI gain had total expenditures that were 11% higher, outpatient expenditures that were 25% higher, and other expenditures that were 20% higher (p0.05).;OLS regressions on log-ratio of expenditures suggested that when compared to individuals who stayed in normal BMI group, individuals who experienced BMI loss had significantly lower inpatient expenditures ( beta = -0.54). When compared to individuals who stayed in normal BMI group, individuals with BMI gain significantly higher outpatient expenditures (beta = 0.172).;Comorbid mental illness did not substantially alter the magnitude or direction of association between BMI and changes in expenditures in the population studied.;Conclusions/implications. Obesity has become an epidemic affecting all ages including the elderly, and is associated with increased morbidity, mortality and healthcare expenditures. Previous studies have focused primarily on the effects of weight loss on healthcare expenditures an d have several limitations: . The current study explored how changes in BMI affect a variety of healthcare expenditures within a Medicare population and additionally how comorbid mental illness impacts this relationship.;The current findings highlight the importance of maintaining normal weight and avoiding weight fluctuations. In light of the increased expenditures due to overweight and obesity, and the positive association between increasing BMI values and healthcare expenditures in the elderly, having a stable BMI is necessary to reduce healthcare costs. Collectively, findings from the current study emphasize the need for preventing overweight and obesity rather than treating these conditions and their negative effects. In this context, weight management and wellness programs that include nutrition and physical activity need to be an integral part of health promotion efforts for all individuals including the elderly.;Future research needs to distinguish between intentional versus unintentional weight loss and its effect on healthcare expenditures to understand the dynamics of weight change. Patterns of service utilization among persons with BMI changes also deserve closer examination. The link between obesity and mental illness remains unclear. Research in this area is ongoing, and a better understanding of the causal pathways of the relationship between obesity, mental illness and healthcare expenditures is needed.