Short-Term Economic Impact of Body Weight Change Among Patients With Type 2 Diabetes Treated With Antidiabetic Agents: Analysis Using Claims, Laboratory, and Medical Record Data

By Yu, Andrew P Wu, Eric Q; Birnbaum, Howard G; Emani, Srinivas; Fay, Madeleine; Pohl, Gerhardt; Wintle, Matthew; Yang, Elaine; Oglesby, Alan

Key words: Antidiabetic treatment – Economic benefit – Obesity – Type 2 diabetes – Weight change ABSTRACT

Background: Obesity is highly prevalent among patients with type 2 diabetes. Unfortunately, weight gain may also be a consequence of some antidiabetic medications. Although clinical benefits of weight loss have been established, the economic consequence of weight change among patients with type 2 diabetes is unclear.

Objectives: The objective was to measure 1-year total and diabetes-related health care costs associated with weight change during the preceding 6-month period among type 2 diabetic patients on antidiabetic therapy.

Methods: Administrative claims, electronic laboratory data and medical chart information were abstracted for continuously enrolled adults with type 2 diabetes from an health maintenance organization (HMO) for the period from July 1, 1997 through October 31, 2005. To assess the economic impact of weight change, three regression models were applied to estimate the following: (1) the effect of weight change in general (one-slope model); (2) the different effects of weight gain and no weight gain (two-slope model); and (3) the different effects of weight gain and no weight gain (i.e., no change or weight loss) among obese and non-obese patients (four-slope model). Patients included in the study had a baseline weight measurement and a second weight measurement approximately 6 months later. They were also required to be on at least one antidiabetic drug therapy within 1 month around the baseline weight measurement date (index date). Based on the measured weight change, patients were classified into two groups – weight gainers and non-weight gainer. Total health care cost and diabetes-related cost were measured during the 1-year period following the second weight measurement and were adjusted to 2004 dollars by the medical component of the Consumer Price Index (CPI). Generalized linear models with log link function and gamma distribution were applied to assess the impacts of weight change on the 1-year total health care cost as well as 1-year diabetes-related cost. All models controlled for patients’ baseline demographics, comorbidities, body mass index (BMI), glycosylated hemoglobin (HbA^sub 1c^), and prior resource utilization.

Results: The study included 458 patients, of whom 224 (48.9%) experienced minimum weight gain of 1 pound between the two weight measurements. The average 1-year total health care cost following the second weight measure was $6382 and the diabetes-related cost was $2002. The mean total health care cost was $7260 for the weight- gainers and $5541 for the non-weight gainers (p = 0.046), and the mean diabetes-related cost, respectively, was $2141 and $1869 (p = 0.006). Results from the models showed that one percentage point of weight change was positively associated with a 3.1% ($213, p < 0.01) change in total health care cost. When weight gain and no gain were modeled separately, one percentage point of weight loss was associated with a 3.6% ($256, p < 0.05) decrease in total health care cost and a 5.8% ($131, p < 0.01) decrease in diabetes-related cost. However, one percentage point of weight gain was not associated with significant increase in either total health care or diabetes-related cost. Further, results from the model with interactions between weight change and obesity status revealed that the economic benefit of weight loss was more pronounced in the obese group (BMI

Conclusions: Weight loss significantly reduced diabetes-related costs. Controlling for baseline factors in the regression model, the 1-year total health care cost following 1% weight loss (or gain) was $213 cost decrease (or increase). Diabetes-related cost did not appear to be associated with weight gain. Economic benefit of weight loss was evident among type 2 diabetic patients on antidiabetic therapy, especially among obese patients.

Introduction

Diabetes is one of the most common chronic conditions in the United States. In 2004, there were 20.8 million people (7.0% of the total population) living with diabetes1, about two-thirds of whom (14.7 million) were diagnosed2. Among the diagnosed cases, 90-95% have type 2 diabetes, also formerly known as non-insulindependent diabetes mellitus (NIDDM)12. The prevalence of diabetes has been increasing rapidly3″5. Based on statistics from the national surveys, the number of people with diagnosed diabetes has increased four- to eight-fold during the past half century, up from 1.6 million in 1958 to 12.1 million in 20003. This growth has affected people of all age, sex, racial and ethnic groups3,5. The rising prevalence of diabetes not only raises clinical concerns but also increases the economic burden of treating this condition in the United States.

In 2002, the total economic burden of type 2 diabetes in the US was estimated to be $132 billion, of which $91.8 billion was for direct medical cost6. Studies have shown that people with diabetes incur about 2.4 times higher medical cost compared to those without diabetes6,7. Direct medical cost among diabetic patients is positively associated with body weight and diabetes complications8. Cardiovascular disease accounts for greater than 70% of the mortality, morbidity, and economic cost associated with type 2 diabetes1.

The current epidemic of type 2 diabetes is clearly associated with obesity, with a strong exponential relationship between body mass index (BMI) and risk of type 2 diabetes9,10. A recent study reported that 86% of patients with type 2 diabetes were at least overweight, and 52% were obese11. Obesity is a major independent risk factor in the development of type 2 diabetes12,13, among other risk factors including hypertension, dyslipidemia, and cardiovascular disease14. The high prevalence of obesity among patients with diabetes may also be related to the fact that weight gain is also a consequence of certain antidiabetic drugs, albeit in varying degrees depending on the drug, thiazolidinediones (TZDs) being associated with substantial weight gain while insulins may be associated with some weight gain15,16.

Traditional treatment for type 2 diabetes consists of a stepwise and progressive regimen from lifestyle modification to drug therapy. Pharmacological management of type 2 diabetes is generally divided into two groups – oral antidiabetic agents and insulin. The drugs under the umbrella of oral agents can further be classified based on their pharmacological mechanism17. The most commonly used oral agents are metformin (which decreases hepatic glucose production), sulfonylureas (which increase endogenous insulin secretion), and TZD (which decreases peripheral insulin resistance). Other antidiabetic classes include alpha-glucosidase inhibitors and meglitinides. The goal of antidiabetic treatment is to maintain a normal or near- normal glycemic level and control other important cardiovascular risk factors such as lipid level, blood pressure, and body weight. Tight glucose control can effectively prevent microvascular complications18,19. According to a recent treatment algorithm for type 2 diabetes, following metformin and lifestyle change, initiation of insulin is recommended as ‘a safer approach for individuals presenting with weight loss…’20, because intensive glucose control with insulin can increase the risk of weight gain16,19,20. Weight gain may also be associated with oral agents such as TZD21. Although the United Kingdom Prospective Diabetes Study (UKPDS) found that the positive effects of glucose control outweighed the negative impact of additional weight gain19, physicians may be reluctant to prescribe otherwise appropriate medication due to this concern22. In addition, weight gain can lead to patient non-compliance23,24.

The health benefits of even moderate weight loss in patients with type 2 diabetes have been demonstrated in both clinical trials and epidemiological studies25-28. The benefits of weight loss in type 2 diabetes include reduced insulin resistance, improved glycemic control, improved cardiovascular risk factors, such as blood lipids, blood pressure, and inflammatory markers, and decreased risk of diabetes-related complications, and, consequently, decreased morbidity and mortality29-31. Given its well recognized benefits, weight control is an essential component in the management of type 2 diabetes. The American Diabetes Association (ADA) treatment guidelines recommend weight loss for all overweight (BMI 25-30kg/ m^sup 2^) or obese (BMI > 30kg/m^sup 2^) adults with diabetes or at risk of diabetes20.

Although the clinical benefits of weight loss among patients with type 2 diabetes have been well established, the economic consequence of the weight change in type 2 diabetes is uncertain. The objective of this study is to investigate the short-term economic impact of weight change among patients with type 2 diabetes mellitus on antidiabetic therapy from a third party payer’s perspective.

Methods

DataAll study subjects had a primary care physician at Fallen Clinic, a multi-specialty medical group located in Central Massachusetts. The study subjects also were members of Fallon Community Health Plan (FCHP), a mixed model health maintenance organization (HMO) located in Central Massachusetts. In 2005, the Fallon Clinic number of capitated covered lives in FCHP was approximately 104000, and membership was stable during the study period. Data sources available from FCHP include enrollment records, medical utilization claims, pharmacy claims, electronic lab data covering care provided from July 1, 1997 to October 31, 2005. All electronic files (of medical, pharmacy and lab data) were linked by unique and de-identified patient identifiers. Complete medical charts on study subjects were available at Fallon Clinic and were manually reviewed by registered nurses in the Research Department at the Clinic to collect weight and height information of each patient, following protocol approval by the local Institutional Review Board. In this study, a claims database was used to obtain the information regarding patient demographics, comorbidities, medication use, costs, and other resource utilization. Electronic lab records and manual chart reviews were used to obtain patients baseline glycosylated hemoglobin (HbA^sub 1c^) values and to measure patients’ height and weight information.

Study design

The key measurement of this study was weight change as well as treatment costs and other covariates. Patient weights were measured over an approximate 6-month weight observation period from manual medical chart review. The weight observation period was defined by the dates of the two weight measurements, the first of which was designated as the index date. Accordingly, the pre-index period was defined as 6 months before the first weight measurement date (i.e., the index date), during which some baseline variables such as comorbidities and prior utilization were measured. Health care cost during the 12-month period following the second weight measurement (cost measurement period) was measured using an electronic medical and pharmacy claims database (Figure 1). The weight change was measured strictly prior to the 12-month cost measurement period. This distinct temporal relationship allows understanding of the impact of weight change during the weight observation period on subsequent cost measured during the 12-month cost measurement period. While patients may experience further weight change during the cost measurement period, the relationship between such weight change and cost measured in the same period would be ambiguous, because clinical events such as hospitalization due to macrovascular diseases (and associated cost) may also trigger weight change (Figure 1).

Patient selection

Patients were selected from the enrollees of FCHP during the period from July 1, 1997 to October 31, 2005. The sample selection process included adult patients (aged 20 years or over) with type 2 diabetes, who were treated with antidiabetic medications. Type 2 diabetic patients were identified as those who had at least two independent diagnoses of type 2 diabetes (International Classification of Disease, 9th Revision, Clinical Modification32 [ICD-9-CM]: 250.x0 or 250.x2) but without diagnosis of type 1 diabetes (ICD-9-CM: 250.xl and 250.x3) at any time in the medical claims database. Patients who met the above criteria but had the following diseases identified by the ICD-9-CM code were excluded: 256.4 (ovarian dysfunction: polycystic ovaries), 251.8 (other specified disorders of pancreatic internal secretion), 962.0 (poisoning by adrenal cortical steroids) or 648.8 (complicating pregnancy: abnormal glucose tolerance). Also, patients who underwent bariatric surgery were excluded (Current Procedural Terminology [CPT] codes: 43644, 43645, 43842, 43843, 43845, 43846, 43847, S2082, S2083, S2085). Furthermore, patients also had to meet the following criteria:

Figure 1. Illustration of the study design

* At least two weight measures, with the second weight measurement date approximately 6 months (+- 2 months) after the index date.

* Continuous enrollment in FCHP and eligible for both medical and pharmacy benefits during the pre-index period, weight change observation period, and 1 -year cost measurement period.

* At least one HbA^sub 1c^ measurement within 3 months around the index date.

* Use of at least one of the four major classes of antidiabetic drugs (i.e., TZD, sulfonylurea, metformin, or insulin) within 1 month around the index date.

During the sample selection and chart review process, patients treated with insulin and TZDs were sampled with preference in order to achieve a more balanced number of patients who were treated with the four major classes of antidiabetic medications.

Outcomes

Health care cost of patients with type 2 diabetes was calculated from a private third-party payer’s perspective. Because all patients were in an HMO, cost was defined as reimbursed amounts for medical services and was estimated using FCHP’s ratio of the paid amount to the billed amount for each type of services; patient co-payments are not included.

We examined two major economic outcomes in the analysis – total health care cost and diabetes-related cost – both of which were measured during the 1 -year cost measurement period. Total health care cost had two major components, total pharmacy cost, estimated from pharmacy claims, and total medical cost, estimated from medical claims. Similarly, diabetes-related cost had both pharmacy and medical components. The pharmacy component included antidiabetic agents and cardiovascular medications (e.g., antihypertensives, antianginals, lipid-lowering drugs, antiplatelet agents, etc.). The cardiovascular drugs were included because many type 2 diabetic patients have concurrent cardiovascular conditions, and these drugs have become important components of drug therapy regimens for diabetes patients who are at increased risk of cardiovascular diseases33. Diabetes-related medical cost included cost from the medical claims with a primary or secondary diagnosis of diabetes (ICD-9-CM: 250.xx). All costs were adjusted to 2004 dollars using the medical component of the Consumer Price Indexes (CPI). While cost data could have been inflated to a more recent year, we report the data for the most recent whole year within the study period.

Weight change and other covariates

Medical chart reviews extracted height information and weight change during the weight observation period. We measured weight change in both absolute value (number of pounds) and relative change (i.e., percentage change from baseline weight), with the latter used for regression analysis. For the regression modeling purpose, we also created a dichotomous indicator regarding whether or not a patient experienced weight gain, which allows for the differential economic effects estimation of weight change among patients with weight gain and those with no weight gain (i.e., defined as those with either no weight change or with weight loss).

To control for potential confounders, we included baseline patient characteristics in all regression models. Antidiabetic treatment information was obtained from the pharmacy claims data. Glycemic control status was measured by HbA^sub 1c^ extracted from the electronic lab data. Baseline BMI was calculated based on height and weight on the index date obtained from medical chart reviews. Following the recommendations by the World Health Organization and the National Institutes of Health34,35, obesity was defined as a BMI of 30kg/m^sup 2^ or over. Baseline comorbidities included coronary heart disease (ICD-9-CM: 410-414), congestive heart failure (ICD-9- CM: 398.91, 428), hypertension (ICD-9-CM: 401, 642), dyslipidemia (ICD-9-CM: 272), and depression (ICD-9-CM: 300.4, 301.12, 309.0, 309.1, and 311). In addition to the selected comorbidities, a Charlson comorbidity index36’37 was constructed to measure overall comorbidity. The Charlson comorbidity index reflects the cumulative likelihood of 1-year mortality (i.e., the higher the score, the more severe the burden of comorbidity), We used all available claims data including inpatient and outpatient claims to estimate patient comorbidity status38’39. Diabetes severity status was inferred from the fourth and fifth digit of ICD-9-CM diagnoses code in the medical claims data and was indicated as diabetes with complications (ICD-9- CM: 250.4, 250.5, 250.6) and as uncontrolled status of diabetes (ICD- 9-CM: 250.x2). Baseline medication use was measured by the four major classes of antidiabetic drugs received around the index date, total number of antidiabetic prescriptions during the preindex period, and use of any cardiovascular medication at baseline. Baseline resource utilization collected during the pre-index period was also controlled in the analysis, which included dichotomous variables, such as any inpatient hospitalization and any emergency visit, as well as continuous cost measures, including both total health care cost (medical cost and pharmacy cost) and diabetes- related cost.

Statistical analysis

We first descriptively investigated the bivariate relationship between weight change and health care cost. We divided patients into weight gainers (n = 234) versus non-weight gainers (including those with actual weight loss [n = 193] and with no weight change [n = 31]), and then compared health care cost between the two groups using Wilcoxon tests to evaluate statistical significance.

Because the cost distribution is highly skewed, generalized linear models (GLM) with a log link function and gamma error distribution were applied to estimate the total health care cost and diabetes-related cost. In contrast to traditional log transformed ordinary least square models, this model allows direct estimation of the mean cost and enables the coefficients to be directly transformed to a dollar value40. To assess the economic impact of weight change with different levels of refinement, we estimated three regression models. Model 1 (one-slope model) included percentage weight change as a continuous variable and estimated the change in dollar per one percentage point change in weight. Model 2 (two-slope model) allowed differential effects of weight change among weight gainers versus non-weight gainers by including an interaction term between weight change and weight gain indicator. This resulted in separate slope estimates for weight gainers and non- weight gainers. Based on Model 2, Model 3 (four-slope model) further added an obesity indicator to the interaction term, allowing separate estimates for the effect of weight gain and weight loss by baseline obesity status. In order to select an appropriate model, we used a log likelihood ratio test to decide which model fit the data more parsimoniously, or whether the additional interaction terms could explain significantly more variations.

Table 1. Sample selection

The regression coefficients were on a log scale. For ease of interpretation, they were converted to percentage of cost differences through exponential transformation. Because the log link function is nonlinear, we estimated transformed cost at the individual patient level in order to calculate the cost change in dollar amount.

SAS version 9.1 was used for all analyses. (SAS Institute Inc., Cary, North Carolina, USA)

Results

Sample characteristics

A total of 458 adult type 2 diabetic patients (mean age: 57.7 years, 59% male) were included in the study sample (Table 1). There were 224 (48.9%) patients who gained weight during the weight observation period (Table 2). On the index date, the average weight was 96.8kg and the average BMI was 33.7kg/m^sup 2^. The baseline weight was similar between the weight gain and no weight gain group, and 65.5% of all patients were obese. The average baseline HbA^sub 1c^ was 8.5% for the weight gainer group and 8.1% for the no weight gain group (p = 0.026). The majority of the sample was on combination therapy of the four drug classes at baseline. Sulfonylureas were the most commonly used, followed by metformin, TZD, and insulin. The weight gainer group had significantly higher use of TZD than the no weight gain group (p < 0.001). In addition, cardiovascular comorbidities were common for the study population; 7.4% of patients were diagnosed with coronary heart disease, 19.0% had hypertension, and 26.9% had dyslipidemia. The Charlson comorbidity index showed that the weight gainer group had higher overall comorbidity than the no weight gain group (p = 0.008). The majority (76.6%) of these patients was also on cardiovascular drugs. Most patients did not have episodes of hospitalization or emergency visit during the pre-index period. The average total health care cost during the 6-month preindex period were $3167 and $1852 for the weight gainer group and non-weight gainer group, respectively (p = 0.003). Diabetes-related costs were also high in the weight gainer group vs. the non-weight gainer group (p = 0.004).

Table 2. Baseline characteristics of the study sample for the 6- month pre-index period

Weight change and health care cost

On average, weight gainers (n = 224) increased 3.9% in weight and non-weight gainers (n = 234) lost 3.3% weight during the 6-month weight observational period. During the 1 -year follow up period, the average total health care cost for the entire sample was $6382 (including $4424 medical cost and $1958 pharmacy cost), with a median of $2721 during the 12-month cost measurement period. The average diabetes-related cost during the same period was $2002 (median of $1289). Health care cost of the weight gainers was significantly higher than that of the non-weight gainers (Table 2). The mean total health care cost were $7260 and $5541, respectively, for the two groups (p = 0.046) and the mean diabetes-related cost were $2141 and $1869, respectively (p = 0.006) (Table 3).

Economic impact of weight change

Realizing the comorbidity, drug and resource utilization differences associated with weight change at baseline, we applied generalized linear model with gamma distribution and log link function to control baseline confounders for economic costs outcomes.

In this study, the weight change was measured as a continuous variable. The association between weight change and economic costs was expressed using estimated regression coefficient from generalized linear model, and transformed back to the dollar value. For interpretation of the coefficient, we expressed the dollar amount change in treatment costs for each percentage of weight increase or decrease. However, this interpretation does not imply a clinical significance change of one percentage of weight change.

For 1-year total health care costs, results from the GLM models revealed that one percentage point of weight change was statistically significantly associated with a 3.1% (or $213) change (p < 0.01) (Table 4).

When assessing the effects of weight gain and weight loss separately, one percentage point of weight loss was associated with a 3.6% decrease in total health care cost, which translated into $256 on average in the study sample; the corresponding total health care cost increase of weight gain was 2.3% or $165. To test whether the Model 2 would be a better fitted model, the log likelihood ratio test revealed that additional interaction terms did not significantly improve the model overall goodness of fit (p = 0.622). Thus, the one-slope model is a better fit for total health care cost.

Further examination of the effects among different weight groups in Model 3 showed that the economic benefits from weight loss were more pronounced in obese patients. Compared to Model 2, the additional interaction term did not improve the goodness of fit of the model either (p = 0.172).

For diabetes-related costs outcomes, the findings from the three models for diabetes-related cost showed a similar pattern to the one identified in total health care cost (Table 5). First, when weight change was examined as a continuous variable, one percentage point of weight increase/decrease was associated with a 3.2% (or $65) increase/decrease in diabetes-related cost (p < 0.01).

In Model 2 (two-slope model), when weight loss and weight gain were modeled separately, the slopes were different. Every one percentage point weight loss was associated with a 5.8% or $131 cost decrease (p < 0.01), while one percentage point weight gain was insignificantly related to a 0.40% or $9 cost increase. Log likelihood ratio tests showed a significant improvement in the model fitness by modeling weight gain and weight loss separately (p = 0.035). Thus, the two-slope model is a better fit for diabetes- related cost.

Similarly, results from Model 3 also revealed that the diabetes- related economic benefit of weight loss was more pronounced in the obese group, but this additional interaction term did not improve the model fitness (p = 0.522). In this group, one percentage point weight loss was associated with a 6.7% or $150 decrease in diabetes- related cost. Comparing the results between total health care cost and diabetes-related cost demonstrated that reduction in diabetes- related cost from weight loss contributed to 51 % of the decrease in total health care cost in the overall sample ($131 diabetes out of $256 total cost) and 37% among the obese patients ($150 diabetes out of $408 total cost).

Table 3. Comparison of weight, weight change and health care costs* between weight gainers and non-weight gainers

Discussion

Diabetes is one of the most common chronic diseases threatening the health of the US population. The economic impact of diabetes on the US is substantial. Of $865 billion total health care expenditure in the United States, $160 billion was attributed to diabetes- related expenses6. Diabetes-related health care expenditure is likely to continue increasing as studies have warned of an unchecked diabetes epidemic3-5. Therefore, it is a critical issue to improve understanding of factors associated with health care cost among diabetic population.

This study examined the impact of weight change on 1-year health care cost among type 2 diabetic patients. The results demonstrated that weight change was positively associated with 1 year total health care cost and diabetes-related cost from a third-party payer’s perspective. Weight loss among type 2 diabetic patients, particularly if obese, may significantly reduce their total health care and diabetes-related cost. Such findings have practical implications for cost containment strategy as third-party payers continue to face growing health care cost. Our study also suggests that the economic benefit associated with weight loss is mainly attributable to reduced diabetes-related resource utilization.

In this study, no antidiabetic agent washout period was required, and weight change could be measured at any time during the course of antidiabetic therapy, and this result implies that this study has more generalizability for long term effect. Health benefits of weight control among type 2 diabetic patients have been established by numerous clinical and epidemiological studies. Moderate weight loss, especially loss of intraabdomen fat, has been shown to improve glycemic control and improve dyslipidemia, blood pressure, and biomarkers of inflammation26””. Intentional weight loss is also associated with reduced mortality in overweight diabetic patients27. The economic benefits of weight loss are also suggested by several previous studies. Lamotte et al.2S evaluated the cost-effectiveness of using weight control medications among obese patients with type 2 diabetes. They found that weight loss interventions were cost- effective among this group, especially in those with hypertension and hypercholesterolemia. Another cost-effectiveness analysis based on UKPDS suggested that metformin, a glucose management drug that is not associated with weight gain, was cost saving among overweight patients with type 2 diabetes compared with conventional diet therapy42. Additionally, a comprehensive research synthesis on the long-term effects of obesity treatment also suggested that weight loss through lifestyle changes, medication or surgery is likely to be cost-effective among patients with impaired glucose tolerance31. Notably, the Diabetes Prevention Program Research Group reported that from both a health care system and societal perspectives, a lifestyle-modification program with the goal of at least a 7% weight loss was more effective and cost-effective than metformin treatment43-45. Table 4. Relative charge in 1-year health care costs associated with factors from GLM regressions

Table 4. Relative change in diabetes-related health care costs associated with factors from GLM regressions

This study went beyond clinical trial settings and tracked the utilization and cost of type 2 diabetic patients from an HMO. To the authors’ knowledge, this study is among the first to examine the short-term economic impact of weight loss from a third party payer’s perspective in the US. From the economic perspective, these findings further strengthen the importance of weight control among type 2 diabetic patients.

Several approaches have been recommended to control weight in the diabetic population46,47. As an integral part of diabetes treatment, lifestyle modification therapy (including both nutrition care and physical activity) helps to prevent and treat obesity. Increased physical activity is not only important to maintain weight loss among diabetic patients but also instrumental in optimizing glucose level and reducing other cardiovascular risk factors48. Moreover, a recent study suggests that the decline of physical activities and increased sedentary behaviors, such as watching TV, are associated with an elevated risk of obesity and type 2 diabetes49. Many clinical guidelines have been developed for weight managements, including a combination of diet, exercise, behavioral interventions, pharmacotherapy, and surgical procedures20,32,50,51. Nonetheless, a major challenge of lifestyle modification is poor adherence, especially among obese patients. This is primarily due to the fact that obese patients are more likely to have multiple chronic conditions, which lower their endurance and increase their discomfort during physical activity52. In patients with type 2 diabetes, weight management is even more difficult as many agents that improve glucose control are associated with significant weight gain. In addition to clinical efficacy of antidiabetic therapy, the impact on weight change may also be an important factor for consideration when selecting antidiabetic regimens, as well as their potential impact on multiple clinical endpoints in type 2 diabetes, such as HbAlc, post-prandial glucose, lipids, and blood pressure, among others.

Limitations

This analysis was conducted from a third-party payer’s perspective so the cost only included reimbursed amount. Direct and indirect cost incurred to other parties, such as patients, employers, and caregivers, were not collected in the current study but are also important for a comprehensive assessment of the economic impact of weight change among type 2 diabetic patients. In addition, the study focused on relatively short term (i.e., 1 year) economic benefits and thus does not provide empirical evidence on long-term economic benefits. However, this study did not impose an antidiabetic washout period, and the weight change was measured at any possible time during the course of antidiabetic therapy in this study. This suggests that the economic impact observed over the 1 – year period can be extended to a longer period of time.

Another caveat is generalizability. This study is based on a sample selected from the capitated group component of a single HMO in Massachusetts. Because patients’ characteristics vary across geographic location and health care plans, the economic benefits of weight loss identified in this study may not be applicable to other settings. To the extent that the economic benefit of weight loss is homogeneous across type 2 diabetic populations in the US, the findings in this study can still provide guidance for diabetes management in general. Future studies should be conducted to verify the economic benefit of weight loss in type 2 diabetes in other patient populations and to determine if differences exist between treatment regimens.

Further, although the study results indicated an economic benefit of weight loss, we were unable to identify the mechanism through which weight loss was achieved. Many factors could have caused weight change such as comorbidities and antidiabetic regimens. In addition, other factors can also affect weight change, such as concomitant non-antidiabetic medications, enrollment in a weight loss program or counseling by dietician. Therefore, this study did not compare the economic benefits of alternative weight management approaches.

Finally, this study also suffers from some of the common limitations of economic studies, such as confounding and measurement errors. For example, a potential confounder in this study could be patient’s adherence to treatment, which is associated with health services utilization.

The discussion of limitations also sheds light on future possible studies in this area. To better inform policies, the economic benefits of weight loss could be further studied from a societal point of view, which include direct medical cost, caregiver cost, and indirect cost (productivity loss and premature death). Moreover, future research may extend the study period to empirically address the long-term economic benefits, and investigate the mechanisms of cost consequences as a result of weight change. In addition, cost- effectiveness analyses of alternative interventions in weight management among the diabetic population by different weight groups are recommended to better guide clinical practice and resource allocation.

Conclusions

Although the health benefits of weight loss in patients with type 2 diabetes have been well established, the economic consequence of weight loss is unclear. To the authors’ knowledge, this study is among the first published research on the economic impact of weight loss among diabetic patients. Based on a local HMO population, we estimated that every 1 % of weight loss (or gain) were associated with $213 total health care cost decrease (or increase). Particularly, weight loss significantly reduced diabetes-related costs. In general, the study found that in addition to the established clinical benefits of weight loss for patients with type 2 diabetes, there may be also an economic benefit of weight loss, especially among those who are obese. In the era of increasing prevalence of diabetes and its economic impact, consideration of weight control is a relevant factor in the selection of therapy for type 2 diabetes.

Acknowledgments

Declaration of interest: This research was funded by an unrestricted research grant from Eli Lilly and Company, and Amylin Pharmaceuticals to Analysis Group Inc., and its subcontractor Fallon Clinic, Inc. GP and AO are employees of Eli Lilly and Company, and MW is an employee of Amylin Pharmaceuticals.

References

1. Centers for Disease Control and Prevention. Data and trends, National Diabetes Surveillance System, 2005. Available from http:// www.cdc.gov/diabetes/statistics/prev/national/ figpersons.htm [Last accessed 19 May 2006]

2. Centers for Disease Control and Prevention. National diabetes fact sheet, 2003. Available from http://www.cdc.gov/diabetes/ pubs/ estimates05.htm#prev [Last accessed 19 May 2006]

3. Engelgau MM, Geiss LS, Saaddine JB, et al. The evolving diabetes burden in the United States. Ann Intern Med 2004; 140: 945- 50

4. Mokdad AH, Ford ES, Bowman BA, et al. Diabetes trends in the US: 1990-1998. Diabetes Care 2000;23:1278-83

5. Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA 2003;289:76-9

6. Hogan P, DaIl T, Nikolov P; American Diabetes Association. Economic costs of diabetes in the US in 2002. Diabetes Care 2003)26:917-32

7. Ramsey S, Summers KH, Leong SA, et al. Productivity and medical costs of diabetes in a large employer population. Diabetes Care 2002;25:23-9

8. Brandie M, Zhou H, Smith BR, et al. The direct medical cost of type 2 diabetes. Diabetes Care 2003;26:2300-4

9. Must A, Spadano J, Coakley EH, et al. The disease burden associated with overweight and obesity. JAMA 1999;282:1523-9

10. Smyth S, Heron A. Diabetes and obesity: the twin epidemics. Nat Med 2006;12:75-80

11. Daousi C, Casson IF, Gill GV, et al. Prevalence of obesity in type 2 diabetes in secondary care: association with cardiovascular risk factors. Postgrad Med J 2006;82:280-4

12. Ford ES, Williamson DF, Liu S. Weight change and diabetes incidence: findings from a national cohort of US adults. Am J Epidemiol 1997;146:214-22

13. Resnick HE, Valsania P, Halter JB, et al. Relation of weight gain and weight loss on subsequent diabetes risk in overweight adults. J Epidemiol Community Health 2000;54:596-602

14. Eckel RH, Kahn R, Robertson RM, et al. Preventing cardiovascular disease and diabetes: a call to action from the American Diabetes Association and the American Heart Association. Diabetes Care 2006)29:1697-9

15. Serrano Rios M. Relationship between obesity and the increased risk of major complications in non-insulin-dependent diabetes mellitus. Eur J Clin Invest 1998;28:14-7

16. Zinman B. Translating the Diabetes Control and Complications Trial (DCCT) into clinical practice: overcoming the barriers. Diabetologia 1997;40:S88-90

17. Chan JL, Abrahamson MJ. Pharmacological management of type 2 diabetes mellitus: rationale for rational use of insulin. Mayo Clin Proc 2003)78:459-67 18. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin- dependent diabetes mellitus. N Engl J Med 1993)329:977-86

19. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood- glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998)352:837-53

20. American Diabetes Association. Standards of medical care in diabetes – 2007. Diabetes Care 2007)30:54-41

21. Fonseca V. Effect of thiazolidinediones on body weight in patients with diabetes mellitus. Am J Med 2003;115:42S-8

22. Wallace TM, Matthews DR. Poor glycaemic control in type 2 diabetes: a conspiracy of disease, suboptimal therapy and attitude. QJM 2000)93:369-74

23. Douek IF, Gale EAM. The problem of weight gain on insulin. In: Barnett AH, editor. Insulin made easy. London: Medical Education Partnership, 2001. pp.80-9

24. United-Kingdom PDSG. United Kingdom Prospective Diabetes Study 24: a 6-year, randomized, controlled trial comparing sulfonylurea, insulin, and metformin therapy in patients with newly diagnosed type 2 diabetes that could not be controlled with diet therapy. Ann Intern Med 1998)128:165-75

25. Lamotte M, Annemans L, Lefever A, et al. A health economic model to assess the long-term effects and cost-effectiveness of orlistat in obese type 2 diabetic patients. Diabetes Care 2002;25:303-8

26. Markovic TP, Campbell LV, Balasubramanian S, et al. Beneficial effect on average lipid levels from energy restriction and fat loss in obese individuals with or without type 2 diabetes. Diabetes Care 1998;21:695-700

27. Williamson DF, Thompson TJ, Thun M, et al. Intentional weight loss and mortality among overweight individuals with diabetes. Diabetes Care 2000;23:1499-504

28. Wolf AM, Conaway MR, Crowther JQ, et al. Translating lifestyle intervention to practice in obese patients with type 2 diabetes: Improving Control with Activity and Nutrition (ICAN) study. Diabetes Care 2004;27:1570-6

29. Fabricatore AN, Wadden TA. Treatment of obesity: an overview. Clinical Diabetes 2003;21:67-72

30. Klein S, Burke LE, Bray GA, et al. Clinical implications of obesity with specific focus on cardiovascular disease: a statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism: endorsed by the American College of Cardiology Foundation. Circulation 2004; 110:2952-67

31. Avenell A, Broom J, Brown TJ, et al. Systematic review of the long-term effects and economic consequences of treatments for obesity and implications for health improvement. Health Technol Asses 2004;8: iii-iv,l-182

32. The U.S. Department of Health and Human Services. The international classification of diseases, 9th revision, clinical modification. 4th edn. Washington, DC, 1991

33. Buse JB, Ginsberg HN, Bakris GL, et al. Primary prevention of cardiovascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Circulation 2007; 115: 114-26

34. National Heart, Lung and Blood Institute. Aim for a healthy weight. Available from http://www.nhlbi.nih.gov [Last accessed 31 May 2006]

35. World Health Organization Expert Committee on Physical Status. The use and interpretation of anthropometry. Report of a World Health Organization Expert Committee. Technical support, series 854, 1995. Geneva: World Health Organization

36. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:3 73-83

37. Deyo RA, Gherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613-9

38. Klabunde CN. Does the outpatient experience matter? Incorporating outpatient care in a comorbidity index for use with large administrative claims datasets. Abstr Book Assoc Health Serv Res Meet 1997;14:330-9

39. Ash AS, Posner MA, Speckman J, et al. Using claims data to examine mortality trends following hospitalization for heart attack in Medicare. Health Serv Res 2003;38:1253-62

40. Basu A, Manning WG, Mullahy J. Comparing alternative models: log vs. Cox proportional hazard? Health Econ 2004;13:749-65

41. Pi-Sunyer FX. Weight and non-insulin-dependent diabetes mellitus. Am J Clin Nutr 1996;63:426S-29

42. Clarke P, Gray A, Adler A, et al. Cost-effectiveness analysis of intensive blood-glucose control with metformin in overweight patients with Type II diabetes (UKPDS No. 51). Diabetologia 2001;44:298-304

43. Diabetes Prevention Program Research Group. Within-trial cost- effectiveness of lifestyle intervention or metformin for the primary prevention of type 2 diabetes. Diabetes Care 2003;26:2518-23

44. Herman WH, Hoerger TJ, Brandie M, et al. The costeffectiveness of lifestyle modification or metformin in preventing type 2 diabetes in adults with impaired glucose tolerance. Ann Intern Med 2005;142:323-32

45. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393-403

46. Bantle JP, Wylie-Rosett J, Albright AL, et al. Nutrition recommendations and interventions for diabetes – 2006: a position statement of the American Diabetes Association. Diabetes Care 2006;29:2140-57

47. Franz MJ, Bande JP, Beebe CA, et al. Nutrition principles and recommendations in diabetes. Diabetes Care 2004;27: S36-46

48. American Diabetes Association task force for writing nutrition principles and recommendations for the management of diabetes and related complications. American Diabetes Association position statement: evidence-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. J Am Diet Assoc 2002; 102: 109-18

49. Hu FB, Li TY, Colditz GA, et al. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 2003;289:1785-91

50. Gandjour A, Westenhofer J, Wirth A, et al. Development process of an evidence-based guideline for the treatment of obesity. Int J Qual Health Care 2001;13:325-32

51. Nawaz H, Katz DL. American College of Preventive Medicine practice policy statement. Weight management counseling of overweight adults. Am J Prev Med 2001;21:73-8

52. Foreyt JP, Posten WS 2nd. The challenge of diet, exercise and lifestyle modification in the management of the obese diabetic patient. Int J Obes Relat Metab Disord 1999;23:S5-11

CrossRef links are available in the online published version of this paper: http://www.cmrojournal.com

Paper CMRO-3992_4, Accepted for publication: 21 June 2007

Published Online: 31 July 2007

doi: 10.1185/0300799007X219544

Andrew P. Yu(a), Eric Q. Wu(a), Howard G. Birnbaum(a), Srinivas Emanib, Madeleine Fay(b), Gerhardt Pohl(c), Matthew Wintle(d), Elaine Yang(a) and Alan Oglesby(c)

a Analysis Group, Inc., Boston, MA, USA

b Fallon Clinic, Inc., Worchester, MA, USA

c Eli Lilly and Company, Indianapolis, IN, USA

d Arnylin Pharmaceuticals, San Diego, CA, USA

Address for correspondence: Eric Q. Wu, PhD, Analysis Group, Inc., 111 Huntington Ave, 10th Floor, Boston, MA 02199, USA. Tel.: +1 617 425 8254; Fax: +1 617 425 8001; ewu@analysisgroup.com

Copyright Librapharm Sep 2007

(c) 2007 Current Medical Research and Opinion. Provided by ProQuest Information and Learning. All rights Reserved.