Economic Outcomes of the Obese Patient

Obesity is a well-established risk factor for many major chronic diseases such as coronary heart disease, hypertension, type 2 diabetes, cerebrovascular disease, and osteoarthritis. It has also been associated as a risk factor for certain types of cancer (1, 2, 3, 4, 5). It stands to reason then that obesity has a sizeable economic burden on the health-care system. This report discusses the literature on the economic impact of obesity, bringing to light the economic outcomes that are primarily impacted by obesity. Also, this report proposes a set of economic outcomes that could be measured in clinical practice or future research studies so that there will be continued assessment of the economic impact of obesity. Economic outcomes are the measures to assess the economic burden that disease places on the nation. They were classically categorized into two components: 1) direct medical outcomes and 2) indirect morbidity and mortality outcomes. Direct medical outcomes, now referred to as “medical-care outcomes,” are the measures of preventive, diagnosis, and treatment services used in the intervention or treatment regimen (e.g., hospital care, physicians services, and medications). Indirect outcomes refer to productivity changes related to illness and death. Morbidity outcomes are days of work lost by people because of illness, restricted activity, and disability. Mortality outcomes are premature death due to disease, and a mortality cost would be the value of future earnings lost by people who die prematurely. Because the definition indirect outcomes varies substantially by the purpose of the analysis, this report refers only to lost-productivity indirect outcomes. This report and proposal focuses on medical care and lost-productivity outcomes and costs (Table 1) because they are more relevant for shorter-length studies (<5 years) and more relevant to manage care organizations and employers—the primary decision makers and payers of health care in the United States. To date, there have been many published reports of studies and models illustrating the association between obesity and medical-care cost and use (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). However, there are fewer studies that evaluate the impact of weight loss on medical-care costs. Overall, these studies tend to be short-term, small, or uncontrolled; hence, caution is warranted when interpreting their results. In addition, most of the weight-loss studies measured only one aspect of medical-care cost (e.g., medication changes) (17, 18, 19, 20, 21, 22, 23). Although the evidence of the impact of weight loss on medical-care use and cost is mounting, the evidence of the relationship between obesity and medical-care use and cost has been consistent and generally dose-responsive [i.e., as body mass index (BMI) increases, health-care use increases]. Overall, obesity (BMI ≥30 kg/m2) is associated with greater health-care use and cost, particularly pharmaceutical and laboratory costs. There is no difference in use and costs among obese people by age. Gender differences were explored in one study (6), and obese women contributed 67% of the excess physician visits even though they comprised 51% of the population. In general, women do use more health care than men regardless of their weight (15). Another study that evaluated health-care claims and costs in an employer setting (n = 843) reported that employees with “at-risk” BMIs (defined as ≥27.8 kg/m2 for men and 27.3 kg/m2 for women) had, on average, health-care costs that were $2326 greater over a 3-year period than employees with a lower BMI, supporting evidence that obesity is associated with increased health-care use (24). The association between obesity and medical-care costs in this study was found to be greater in obese women and older obese (≥45 years) employees. In addition, Pronk et al. (15), evaluated health-care use and costs among 5689 managed-care members and found that for every one-unit increase in BMI, there was a 1.9% increase in median costs. Although this does not translate into large cost increases at the individual level ($11.26 per unit increase in BMI per year), the cost does become considerable at the population level. The dominant diagnoses increasing medical-care costs among the obese population are coronary heart disease, type 2 diabetes, hypertension and osteoarthritis (6, 8), suggesting that these comorbidities be preferentially targeted when completing cost evaluations among the obese population. In general, a relatively small percentage of people use the majority of health-care services. This leads to a greater skewed distribution with greater variance than other types of outcomes (i.e., clinical or humanistic). Therefore, when assessing health-care use, the sample size may need to be greater. Additionally, more in-depth evaluation of the outliers is required to explore their legitimacy within the results. This is illustrated in two studies (8, 12) that have similar populations and medical-care outcomes. There is a dose-response trend in both studies, however, only the study with a greater sample size achieved statistical significance in many of the economic outcomes presented. With respect to the study by Thompson et al., there were two outliers (person with above average medical costs) in the lean group. One had end-stage renal disease and the other had a motor vehicle accident with complications. These two individuals in the lean group created an overall nonsignificant difference in inpatient and outpatient use between the groups (personal communication with D. Thompson). Hence, when evaluating health-care use and economic data, large sample size is typically needed to show statistical significance and outliers in both lean and obese people need to be explored to assess their impact on statistical trends. Almost all of the studies evaluating the impact of weight loss on economic outcomes have focused on pharmaceutical outcomes. Modest weight loss induced by pharmacotherapy, diet, or behavior modification has been shown to reduce diabetes (18, 19, 20, 25), hypertensive (18, 19, 26, 27) and hyperlipidemia medications (19). Hollander et al. (20) reported that after 1-year treatment with Orlistat and an average 6% weight loss, 43% of the patients in the treated group decreased their dosage of sulfonylurea as compared with only 29% in the control group. Collins et al. evaluated pharmaceutical expenses for 32 obese patients who had hypertension and diabetes and were on medications to treat these conditions before and after a 12-week weight-loss program (18). They reported that on average, these patients had $63.30/month out-of-pocket pharmaceutical expenses, which decreased to $20.40/month at the completion of the diet period. After 1-year (40 weeks after the completion of the intervention) the patients still had lower expenses than before weight loss (1 year: $32.40/month). Greenway et al. determined the pharmaceutical cost savings of a modest weight loss (6% to 10%) for patients who were obese and on medications for diabetes, hypertension, and hyperlipidemia. Subjects were treated with anti-obesity medications to promote weight loss (19). They report a net cost savings of $122.64/month for insulin treated diabetics; $42.92/month for diabetics treated with sulfonylureas; $61.07/month for people treated for hyperlipidemia; and $0.20/month for people treated for hypertension. The impact of weight loss on economic outcomes by surgical intervention has also been studied but the results may not be generalized to nonsurgical weight-loss interventions because the invasive nature of this intervention increases medical-care use and costs significantly more than behavioral or pharmacological interventions. The Swedish Obese Subjects Study (SOS) is the largest, longitudinal intervention trial evaluating the impact of weight loss by surgical treatment; to date there is 6-year follow-up data. Six years after intervention, average weight loss was 15% in the surgically treated group and +1% in the conventionally treated group. The total medication cost increased in both groups over the 6-year period and there was no significant difference in total pharmaceutical cost between groups. Average drug-specific costs were different between groups though. The surgically treated group had lower diabetes and cardiovascular-medication costs but higher anemia and gastrointestinal-disorder-medication cost as compared with the conventionally treated group (28). Other uncontrolled surgical studies have also reported a decreased use of hypertensive (23, 29) and diabetes medications (22). The SOS also looked at changes in inpatient costs comparing the surgically treated group to the conventionally treated group. They reported that the aggregated cost of hospital care for 6 years of follow-up was $10,204 in the surgically treated group as compared with $2827 in the conventionally treated group. When hospital-care costs for surgical intervention and conditions common to bariatric surgery were excluded from both groups, there was no significant difference in the average cost over the 6-year period (30). Hence, weight loss by bariatric surgery had little or a negative impact on inpatient costs. Long-term studies that can evaluate the impact of long-term weight loss on a fuller spectrum of economic outcomes are in progress but results will not be available for many years. Therefore, models that project likely scenarios have been developed to help estimate the impact and determine when is it cost effective to treat obesity. A dynamic model was developed by Oster et al. (31) that projected the lifetime economic benefits of a sustained, modest weight loss (10%) considering five obesity-related diseases (hypertension, hypercholesterolemia, type 2 diabetes, coronary heart disease, and stroke). They project that a modest weight loss, if sustained, would increase life expectancy by 2 to 7 months and reduce lifetime medical-care costs of these five diseases by $2200 to $5300. Medical-care savings were increased as BMI increased, and age had a positive association with savings among women. A recently published model, with reasonable assumptions, evaluated the cost-effectiveness of treating obese patients with type 2 diabetes as compared with diet and exercise alone. They reported that Orlistat is a cost-effective therapeutic [CE ratio range: 3462 euros/life-year gained (LYG) when patients had type 2 diabetes, hypercholesterolemia, and hypertension; 7407 euros/LYG with hypercholesterolemia and type 2 diabetes; 7388 euros/LYG with hypertension and type 2 diabetes; 19,986 euros/LYG with type 2 diabetes without cardiovascular events] (32). Although the current evidence seems to be positive, we do not have enough long-term data to truly support this optimism. The economic impact of weight loss on future medical-care costs and productivity is a complicated issue. It is highly likely that obesity treatment will not be cost-saving and we simply do not know if obesity treatment will be cost-effective until the Look AHEAD trial results are published. Costs of some obesity-related diseases can be considered “sunk costs” in that weight loss will be too late to effect any change in the disease process (e.g. cancer). The SOS results are not particularly favorable regarding the impact of long-term weight loss on intermediate cardiovascular outcomes (33). Although many clinicians and researchers emphasize medical-care outcomes in trying to obtain reimbursement for weight management, indirect morbidity health-care outcomes (productivity, restricted activity, disability) are equally important because employers are the major payers and decision-makers of health care and obesity impacts productivity so greatly. Approximately 27% of the U.S. labor force has a BMI ≥29 kg/m2 (34). Eight studies have evaluated the impact of obesity on indirect morbidity outcomes and costs (6, 9, 24, 34, 35, 36, 37, 38). All the studies have found that obesity impacts productivity (measured by days missed at work due to illness). In addition, one study reported the impact of obesity on workers’ productiveness (36); three studies evaluated increased restricted activity due to obesity (6, 10, 34), and one study found that obesity impacts long-term disability (24). Hence, the evidence is quite strong that obesity (BMI ≥30 kg/m2) impacts productivity and potentially all indirect morbidity outcomes. Overall, women seem to be driving the association between obesity and productivity. Only the SOS trial has evaluated the impact of weight loss on productivity (35); this trial compared surgically to conventionally treated patients. After 5-years of follow-up, patients in the surgically treated group lost, on average, 20% of their body weight as compared with +1% in the conventionally treated group. Compared with the controls, the surgically treated group had 35% more days of sickness during the first year after treatment, but 10% to 14% fewer days during years 2 to 3. During the years 4 to 5 year, the difference in sick leave between groups narrowed (p = 0.07) but was statistically significant in older subjects, ages 47 to 60 years. It is recommended that researchers and clinicians consider collecting economic outcomes (Table 1) and integrating this information into their weight-management studies. We have adequate data to show that obesity impacts both medical-care and lost-productivity outcomes, but we have little data to show that weight loss has an impact on these outcomes. For short-term and smaller studies, it would be practical to collect pharmaceutical and laboratory cost measures. Inpatient and outpatient use and costs may require larger studies (>2000 subjects) due to skewed data and the likelihood of outliers. Furthermore, it is strongly recommended that lost-productivity outcomes and costs be collected because they are relatively easy to collect and the impact of obesity is well documented. If sample size is limited, it is recommended that obese patients with coronary heart disease, type 2 diabetes or hypertension be preferentially evaluated. Additionally, gender differences should be evaluated among lost-productivity outcomes. Table 2 lists economic outcome data sources. When researchers and clinicians undertake cost-effectiveness evaluations, it is recommended that they follow the standardized guidelines (39).