The medical costs for a type 2 diabetes patient are two to four times greater than the costs for a patient without diabetes. Bariatric surgery is the most effective weight-loss therapy and has marked therapeutic effects on diabetes. We estimate the economic effect of the clinical benefits of bariatric surgery for diabetes patients with BMI ≥35 kg/m2. Using an administrative claims database of privately insured patients covering 8.5 million lives 1999–2007, we identify obese patients with diabetes, aged 18–65 years, who were treated with bariatric surgery identified using Healthcare Common Procedure Coding System codes. These patients were matched with nonsurgery control patients on demographic factors, comorbidities, and health-care costs. The overall return on investment (RoI) associated with bariatric surgery was calculated using multivariate analysis. Surgery and control patients were compared postindex with respect to diagnostic claims for diabetes, diabetes medication claims, and adjusted diabetes medication and supply costs. Surgery costs were fully recovered after 26 months for laparoscopic surgery. At month 6, 28% of surgery patients had a diabetes diagnosis, compared to 74% of control patients (P < 0.001). Among preindex insulin users, insulin use dropped to 43% by month 3 for surgery patients, vs. 84% for controls (P < 0.001). By month 1, medication and supply costs were significantly lower for surgery patients (P < 0.001). The therapeutic benefits of bariatric surgery on diabetes translate into considerable economic benefits. These data suggest that surgical therapy is clinically more effective and ultimately less expensive than standard therapy for diabetes patients with BMI ≥35 kg/m2.
Diabetes mellitus is a major public health concern in the United States and elsewhere because of its prevalence, considerable morbidity and mortality, and economic burden. In 2007, the prevalence rate of diabetes in the United States was 7.8%, affecting 12 million men and 11.5 million women (1). Diabetes is associated with serious complications, including coronary heart disease, kidney failure, neuropathy, blindness, and amputation, and was the seventh leading cause of death in 2006, accounting for over 72,000 deaths (1). Type 2 diabetes accounts for 90–95% of all diagnosed cases (1).
Estimated yearly costs of managing a diabetes patient ($13,243) are more than five times that of a patient without diabetes ($2,560) (2). The estimated annual total economic cost of diabetes in the United States was $174 billion in 2007—$116 billion in medical expenditures and $58 billion in reduced productivity (2). The largest components of costs are hospital in-patient care (50%), medication and supplies (12%), retail prescriptions to treat complications (11%), and physician office visits (9%) (2). The annual cost has been projected to reach over $350 billion by 2025 and a cumulative $2.6 trillion over the next 30 years (2).
Obesity is a major risk factor for type 2 diabetes (3), and the risk of diabetes increases directly with BMI (4). Thus, increased prevalence of diabetes is related to the increased prevalence of obesity. In 2007–2008, one-third of US adults had BMI ≥30 kg/m2, and 5.7% were Class III (BMI ≥40 kg/m2) (5). Diabetes-related costs represent a disproportionate share of health-care costs among the obese (6). Weight loss is an important therapeutic goal in obese patients with type 2 diabetes, because even moderate weight loss (5%) improves insulin sensitivity (7). Bariatric surgery is the most effective weight-loss therapy and has considerable beneficial effects on diabetes and other obesity-related comorbidities (8,9,10). However, bariatric surgery requires a significant up-front expenditure.
The purpose of this study was to estimate the economic impact of the clinical benefits of bariatric surgery on medical costs and return on investment (RoI) of the surgery in diabetes patients with BMI ≥35 kg/m2.
Methods and Procedures
We identified 808 diabetes patients who had bariatric surgery and a matched cohort of 808 diabetes patients who did not have surgery using an administrative claims database of privately insured patients covering 8.5 million lives from 1999 through 2007 at 40 large nationwide companies. This database contains deidentified patient medical and pharmacy claims, demographics, enrollment history, date of service, associated diagnoses, performed procedures, billed charges, and actual insurance payments. Pharmacy claims contain complete information on prescribed medications (date filled, days of supply, quantity, and payment amount).
Bariatric surgery patients with diabetes were identified using the following criteria: (i) at least one bariatric surgery claim, identified using HCPCS codes 43644, 43645, 43770, 43842, 43843, 43845, 43846, 43847, S2082, S2083, and S2085; (ii) at least one medical claim diagnosis of obesity for which BMI ≥35 kg/m2 (ICD-9-CM 278.01) any time before the index/surgery date; (iii) continuous enrollment for at least 6 months before and 1 month after the index date; (iv) 18–65 years old at the index date; and (v) diabetes diagnosis before the index date. Because the claims data do not record clinical outcomes that could be used to identify diabetes at baseline, such as glycosylated hemoglobin levels, we used the criteria outlined by Pladevall et al. (11), and defined patients as having diabetes (type 1 or 2) if, in the 3 months before the index date, the patient had both: (i) one or more medical claims for diabetes (ICD-9-CM 250.xx), dyslipidemia (ICD-9-CM 272.xx), or hypertension (ICD-9-CM 401.xx-405.xx); and (ii) one or more claims for diabetes medication.
Laparoscopic procedures were not recorded separately until 2004. Between 2004 and 2007, 64% of procedures were performed laparoscopically. By 2007, laparoscopic procedures constituted 94% of which nearly half were laparoscopic banding.
One control patient with diabetes was matched with each bariatric surgery patient with diabetes on the following criteria: (i) age; (ii) sex; (iii) state of residence; (iv) comorbidities (asthma, coronary artery disease, gallstones, gastroesophageal reflux, nonalcoholic fatty liver disease, sleep apnea, and urinary incontinence); and (iv) health-care cumulative costs in months −6 to −2 before the index date within 1 s.d. of the surgery patient's cumulative costs in the same period. Where more than one control could be matched with a surgery patient, one control patient was randomly selected.
Data for surgery patients and their matched control patients were obtained for 6 months before and up to 84 months after surgery. We calculated overall RoI based on total direct medical costs and assessed three postindex/surgery outcome measures: (i) diagnostic claims for diabetes; (ii) claims for diabetes medication; and (iii) average total costs of diabetes medications and supplies.
Using the methodology in Cremieux et al. (12), we calculated RoI by comparing monthly costs for surgery and control patients over time. To avoid bias from imbalances across the two samples, the multivariate RoI analysis accounts for eight additional comorbidities (breast cancer, congestive heart failure, lymphedema, major depression, osteoarthritis, polycystic ovary syndrome, pseudotumor cerebri, and venous stasis/leg ulcers) based on existing literature and American Society for Bariatric Surgery guidelines (13). Total health-care costs were deflated to 2007 dollars using the CPI for Medical Care, and incremental savings from surgery were discounted using the mean return on a 3-month Treasury bill, 3.43%.
The cost differential between surgery and control patients was calculated by estimating a cluster Tobit model on the pooled population (surgery and control patients) and using the coefficients on interacting bariatric surgery with the relevant time period. For surgery patients, the investment in bariatric surgery is the sum of all incremental claims incurred in the month before surgery, during surgery and 2 months after surgery relative to the claims of the control patients during the same period. RoI is calculated by offsetting the investment in bariatric surgery against the returns from incremental cost-savings postsurgery.
The numbers of diabetes diagnoses and medication claims were calculated every 3 months postsurgery for 36 months. Claims for surgery and control patients were compared using χ2 tests. Monthly drug costs were calculated postsurgery for 36 months and adjusted by the CPI for Medical Care, and were compared using Wilcoxon rank-sum tests.
All estimations were performed using statistical software Intercooled STATA 10 (Stata Corp, College Station, TX) and SAS 9.2 (SAS Institute, Cary, NC).
Table 1 shows the baseline characteristics of study participants. The sample is predominantly female (72.8%), with a mean age of 51.5 for surgery and 52.3 for control patients. Breast cancer and major depression were the only two comorbidities with a statistically different prevalence across the two groups, and were among the eight comorbidities included as control variables in the multivariate RoI analysis. The two groups had similar baseline health-care costs in months −6 to −2 before index date. Surgery patients had 28.7% higher baseline medical service costs, whereas control patients had 36.2% higher prescription drug costs.
Table 1. Comorbidities and medical care utilization before surgery index date
Of the 808 surgery and control patients, 713 surgery patients and 725 control patients remained at 6 months; 283 surgery patients and 307 controls remained at 36 months. The attrition rate was not statistically different between surgery and control patients.
RoI results are presented in Table 2 and Figure 1. For surgery patients, the initial investment averaged ∼$25,000 for all surgeries 1999–2007, $31,000 for open surgeries 1999–2003, $29,000 for open surgeries 2004–2007, and $19,000 for laparoscopic surgeries 2004–2007. Cost savings associated with surgery started accruing at month 3. Total surgery costs were fully recovered on average after 30 months in 1999–2007 for all types of surgeries; after 29 months for open surgeries in 2004–2007; and after 26 months for laparoscopic surgeries in 2004–2007.
Table 2. RoI to bariatric surgery for diabetes patients, multivariate analysis
Clinical benefits are the underlying driver of the RoI results. For diagnostic claims of diabetes, by the first 3-month period after surgery, 40.7% of surgery patients had a diabetes-related claim compared to 72.1% of control patients (P < 0.001). The postindex decline in diabetes diagnostic claims for the control group likely reflects underdiagnosis as opposed to a true decline. By month 6, only 28.2% of surgery patients reported a claim of diabetes vs. 73.5% of control patients (P < 0.001; Figure 2).
The drug utilization analysis is presented in Figure 3. By the first 3-month period postindex, 45.6% of surgery patients had filled a prescription for diabetes medication in the previous 3 months, compared to 90.8% of control patients. At month 6, the percentages were 33.5% and 89.7%, respectively (P < 0.001). Among patients who had insulin claims before index date, insulin claims dropped to 42.8% for surgery patients and remained at 83.8% for control patients at month 3 after index (P < 0.001). Among surgery patients who had claims for noninsulin diabetes medications before surgery, 37.3% had claims for noninsulin medications at month 3, compared with 86.3% of control patients (P < 0.001); 84.5% of surgery patients who had claims for noninsulin medication at index had no claims for any diabetes medications by month 36 (Figures 3 and 4a,b).
Total diabetes medication costs decreased significantly among surgery patients relative to control patients. By the first 3-month period after index, the average total cost of diabetes medications and supplies for surgery patients was $33, compared to $123 for control patients (P < 0.001). This diabetes drug cost-savings trend is sustained for the duration of the study period (Figure 5).
Bariatric surgery is the most effective weight-loss therapy for obese (BMI ≥35 kg/m2) patients. Current financial concerns surrounding health care has led to an increased interest in evaluating the economic effects of medical and surgical therapies. The results of this study demonstrate that the clinical benefits of bariatric surgery in diabetes patients with BMI ≥35 kg/m2 translate into considerable economic benefits. On average, cost savings began to accrue to third-party payers at 3 months and surgery costs were recovered at 30 months (for open and laparoscopic surgeries, 1999–2007). These data suggest that surgical therapy is clinically more effective and ultimately less expensive than standard therapy for diabetes patients with BMI ≥35 kg/m2.
Our results are consistent with those of Cremieux et al. (12), who employed the same methodology to examine the impact of bariatric procedures on all patients with BMI ≥35 kg/m2, not just those with diabetes. They found costs of laparoscopic procedures were recovered at 25 months. In contrast, a study by Finkelstein et al. (14) found that full cost recovery after surgery took 5–10 years. These differences might be due to the reliance on survey data (2000–2001 Medical Expenditure Panel Survey) and simulation methods (Finkelstein et al.) rather than actual claims records (12).
Our findings support the existing data on the economic impact of bariatric surgery on drug costs. Snow et al. (15) reported that average annualized prescription drug costs after laparoscopic gastric bypass surgery were lower, relative to presurgery costs, by 68% after 6 months and by 72% after 2 years. Gould et al. (16) showed monthly cost savings of $120 for medications associated with diabetes, hypertension, hypercholesterolemia, GERD, and depression 6 months postsurgery. Monk et al. (17) reported an average monthly medication expenditure reduction of 57% at 6 months after gastric bypass procedures. A 72% monthly medication cost savings was also reported by Nguyen et al. (18), with 69% savings for diabetes medications. Data from the Swedish Obese Subjects Study showed a decrease in diabetes mellitus medication costs after bariatric surgery relative to a matched obese group who did not have surgery (19). Our study shows that average monthly prescription drug costs for surgery patients were 64% lower than for control patients at 6 months after index, 69% lower at 1 year, and 72% lower at 2 years. A similar trend was observed for monthly diabetes medication costs.
Our results reinforce and extend previous reports of the clinical benefits of bariatric surgery in diabetes treatment. Prior studies have shown bariatric surgery to be an effective therapy for diabetes in obese patients (9,10,20,21,22). The results from a meta-analysis (23) found that resolution of diabetes after bariatric surgery was achieved in 73–88% of patients. This is consistent with our analysis showing that by month 6, 72% of surgery patients no longer had diabetes diagnosis claims. In addition, bariatric surgery reduces the future risk of developing diabetes. The Swedish Obese Subjects Study (23) found that the 10-year incidence of diabetes in bariatric surgery patients was one-third that of patients treated with conventional diet therapy. Several recent studies have reported reductions in medication use soon after bariatric surgery (20,24,25), with a rapid decrease in medication use for diabetes, hypertension, and hyperlipidemia at 3 months and a decrease in the mean number of prescription medications per patient from 2.4 before surgery to 0.2 at 12 months after surgery.
This study has some limitations. First, the claims database did not have data on glycosylated hemoglobin levels, blood pressure measurements, or lipid profiles. We used diagnostic codes and prescription fills to identify diabetes patients, which may not reliably estimate true prevalence. However, sensitivity analyses using alternative definitions of diabetes based on the claims information yielded similar results and are available upon request.
Second, BMI, a measure of surgery eligibility and clinical outcome, was not available in the database. However, control patients were matched to surgery patients along multiple demographic factors and 16 comorbidities likely correlated with BMI. More importantly, both surgery and control patients were diagnosed with a BMI of at least 35 kg/m2.
Third, the reliability of the cost-savings estimates depends in part on the accuracy of the matching process. Although an exact match results in a more balanced sample of patients and controls than a propensity score approach, unobserved characteristics unrelated to index costs, age, gender, and selected comorbidities may influence the decision for surgery and cannot be controlled for with the current methodology.
Fourth, there could be differences in lifestyle behaviors, self-management skills, and other diabetes medical treatments between surgery and control patients. Also, some medication use may be misrepresented in the claims database because patients obtain medications from alternative sources (e.g., dual insurance coverage or samples) or may temporarily stop a medication because of a physician recommendation or side effects. However, given the size of the study and its design, it is unlikely that these factors bias the results significantly.
This study demonstrates that bariatric surgery is cost-effective for diabetes patients with BMI ≥35 kg/m2. Additional studies are needed to determine the optimal eligibility criteria for bariatric surgery in diabetes patients based on both clinical and economic cost-benefit relationships.
Some of the authors (P.Y.C., A.G., S.E., and T.J.M.) received funding from Ethicon Endo-Surgery, Inc. The funding source had no role in study design, collection, analysis, and interpretation of data, writing the paper, or in the decision to submit the paper for publication.