Differences in mean risk-adjusted Medicare payments (US $) for each procedure by service type per episode of bariatric surgery according to quartiles of total episode payments for all procedures (A), laparoscopic gastric banding (B), laparoscopic gastric bypass (C), and open gastric bypass (D). A, The “all procedures” category includes adjustment for procedure type. B-D, Quartiles of hospital payments are generated within each procedure.
Grenda TR, Pradarelli JC, Thumma JR, Dimick JB. Variation in Hospital Episode Costs With Bariatric Surgery. JAMA Surg. 2015;150(12):1109-1115. doi:10.1001/jamasurg.2015.2394
Under bundled payment programs, hospitals receive a single payment for all services provided surrounding an episode of care. Because health care providers, such as hospitals and physicians, accept more financial risk under these programs, they will need a better understanding of episode costs to identify areas to target improvements in quality and cost-efficiency.
To examine hospital variation in episode costs for a common high-risk procedure that is a prime candidate for bundled payment programs (ie, bariatric surgery).
Design, Setting, and Participants
In this retrospective cohort study, we used national Medicare claims data and identified patients undergoing bariatric procedures in 2011-2012 (N = 24 647 patients; 463 hospitals). We calculated risk-adjusted Medicare payments from the date of admission for the index procedure to 30 days following discharge. We then divided hospitals into equally sized quartiles and examined variation in payments for services around episodes of care. Medicare payments were examined by service payment type (ie, payments to hospitals, payments to physicians, and payments for postacute care services) across hospital quartiles.
Main Outcomes and Measures
Hospital variation in episode costs for services around an episode of bariatric surgery.
Mean total payments for bariatric procedures varied from $11 086 to $13 073 per episode of care, resulting in a mean difference of $1987 (16.5% difference) per episode of care between the lowest and highest hospital quartiles. The index hospitalization was responsible for the largest portion of total payments (75%), followed by physician services (21%) and postacute care services (2.8%). Payments for index hospitalization accounted for the greatest variation in payments.
Conclusions and Relevance
There are variations in hospital episode payments among Medicare patients undergoing bariatric surgery procedures. As hospitals enter bundled payment programs, they will need to target areas with the largest variation in costs for quality- and efficiency-improvement activities.
Quiz Ref IDThere is growing enthusiasm for reducing hospital costs through payment reform. One payment innovation that has gained traction is bundled payments, wherein health care providers receive a single fixed payment for all services (eg, hospital, physician, and postacute care) delivered surrounding an episode of care. Specifically, the Centers for Medicare and Medicaid Services (CMS) Bundled Payments for Care Improvement Initiative has expanded this program to 48 episodes of care that health care providers, such as hospitals and physicians, may elect to have bundled.1 Most existing programs for common high-risk surgical procedures, such as cardiac and orthopedic surgery, have already been implemented.1,2 Bariatric surgery, another common procedure, has been proposed as a potential candidate for inclusion in the next iteration of these programs.
As hospitals and physicians accept more financial liability under these programs, it will be important for them to better understand their own episode costs for surgery. To be successful after enrolling in these programs, hospitals will need to understand where further quality-improvement efforts will have the highest yield (eg, readmissions and complications). Variation in episode payments has been previously explored for a small subset of procedures, demonstrating that for each operation, the root cause of high costs is different (ie, hospital, physician, and postacute care services).3 Presently, it is unclear which aspects of bariatric surgical care exhibit the largest variations in costs and should therefore be targeted for improvement.
In this context, we used national Medicare data to examine variation in costs around episodes of bariatric surgery. In addition, we evaluated the composition of payments across different services provided around an episode of bariatric surgery to better understand the overall variation in episode costs.
We used Medicare claims data to identify all patients undergoing bariatric procedures in 2011-2012. We included patients undergoing open Roux-en-Y gastric bypass, laparoscopic Roux-en-Y gastric bypass, laparoscopic adjustable gastric banding surgery, and other procedures (remaining procedures other than those previously mentioned). Patients were identified from the database using a previously validated coding algorithm that included a combination of procedure codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and diagnosis-related group (DRG) codes.4- 7 We used Current Procedural Terminology (CPT) codes in the cohort of Medicare patients that was identified to further improve the clinical specificity of the algorithm.4,8,9
Patients included in this study were required to have inpatient procedure codes for bariatric surgery (ICD-9-CM codes 43.89, 44.3, 44.31, 44.38, 44.39, 44.68, 44.95, 44.96, 44.97, 44.99, 44.5, 45.51, and 45.9), a primary or secondary diagnosis code for morbid obesity (ICD-9-CM codes 278.0, 278.00, 278.01, and V77.8), and a DRG code for weight loss surgery (MS-DRG code 619-621). Procedures were categorized according to laparoscopic gastric bypass, open gastric bypass, laparoscopic gastric banding, or other procedure based on the procedure codes.
Patients with a diagnosis code for abdominal cancer (ICD-9-CM codes 150.0-159.9 or 230.1-230.9) were excluded to ensure that the study only included patients undergoing bariatric procedures. Patients who were missing an accompanying CPT code (43846-43848, 43644, 43645, 43844, 43659, S2085, 43770, 43771, 43773, S2082, 43842, 43843, and 43845) for physician payment were excluded to ensure payment information was complete. Finally, only patients with continuous enrollment into Medicare Part A and B (Medicare Fee-for-Service) with no health maintenance organization for 6 months prior to the admission and during the study were included.
Once patients were appropriately identified, we linked patients’ records with other CMS files containing additional claims associated with the surgical episode, including skilled nursing facilities, outpatient facilities, and home health care.
This study was judged to be exempt from human participant review by the institutional review board of the University of Michigan, Ann Arbor.
Quiz Ref IDThe actual Medicare payments, rather than the submitted charges, for each patient were assessed in this analysis. To assess payments, we obtained payment information for all services from the time of hospital admission for the index procedure to 30 days following discharge, in accordance with the Medicare Physician Advisory Commission recommendations. All payments were price-standardized according to methods previously described by Gottlieb et al.10
In our analysis of payments to hospitals, we assessed those related to the index hospitalization (DRG payments plus outlier payments, if applicable) and readmissions occurring within 30 days of discharge, as previously described by Birkmeyer et al.3 Diagnosis-related group payments were determined from the Medicare Price Amount, which includes the actual hospital payment plus any patient liability amounts that are applicable. Payments for capital expenses, used by CMS to cover bad debt, medical education, and disproportionate share of low-income patients, were excluded given that these were small for most of the hospitals.
In assessment of payments to physicians, each was categorized according to the service (ie, surgery, anesthesia, imaging, and laboratory) provided. This information was determined from claims using CPT codes identifying the index procedure and related anesthetic services. Those physician payments related to imaging or laboratory services were identified based on CPT codes associated with those services. All other physician payments, both inpatient and outpatient, were included under a broader category specified as other medical.
We then assessed payments related to postacute care services using Medicare claims for outpatient and home health care. For payments related to rehabilitation hospital stays, we prorated the episode payment to include only payments within a 30-day window from discharge. Similarly, payments for skilled nursing facilities and nursing homes were determined using per-diem payments within this 30-day window.
We first calculated average payments to hospitals for episodes of bariatric surgery by DRG payment, outlier payments, 30-day readmissions, and total hospital payment. We then determined average payments to physicians, for postacute care services, and total episode payments including all components (hospital, physician, and postacute care services). All payments were adjusted for patient age, sex, race/ethnicity, procedure type (ie, laparoscopic gastric bypass, open gastric bypass, laparoscopic gastric banding, or other procedure), and 29 comorbidities as defined by Elixhauser and colleagues11 and Southern and colleagues12 using a linear mixed regression model. To analyze variation in payments across hospitals, we ranked hospitals from lowest to highest in overall price-standardized risk-adjusted Medicare payments. We then divided hospitals into equally sized quartiles by payment rank for overall payments and then within each procedure. Next, we compared mean payments between hospital payment quartiles for overall episode payments and also for each service provided. To improve reliability and minimize chance variation, we restricted this analysis to hospitals with at least 10 bariatric procedures.
Furthermore, we examined the influence of hospital case volume on episode costs using data from 12 state inpatient databases (n = 32 595 total patients with 8641 Medicare patients; 182 hospitals) because Medicare data would not reflect true hospital volume across all payers. We determined hospital volume based on all patients undergoing bariatric procedures (eg, all adult patients across all payers) in 2011. Hospitals were then placed into equally sized quartiles based on case volume and then mean Medicare total episode payments were determined for each quartile.
All statistical analyses were conducted using SAS version 9.4 (SAS Institute).
Overall, we found that most cases were laparoscopic gastric bypass (69.9%) and laparoscopic gastric banding (21.8%). Open gastric bypass (3.7%) and other procedures (4.6%) accounted for a much smaller proportion of all bariatric procedures. Sleeve gastrectomy composed most of the other procedures category (3.5%).
There were 24 647 Medicare patients in our cohort who underwent bariatric surgery. Mean total payments for bariatric surgery varied from $11 086 per episode of care in the lowest quartile to $13 073 per episode of care in the highest quartile, resulting in an average difference of $1987 (16.5% difference) per episode of care across hospital quartiles (Figure). Quiz Ref IDPayments to hospitals, in particular payments for the index hospitalization, accounted for the largest portion of payments and also the largest source of variation across hospital quartiles for all procedures, accounting for 74% of total payments in the lowest quartile and 71% of total payments in the highest quartile. The mean payments to physicians varied from 21% to 22% across quartiles. Payments for postacute care services varied from 2.4% to 3.8% across quartiles. Quiz Ref IDWhen examining the influence of hospital volume on costs, there was no association between hospital volume and total episode payments, with only a small difference in mean payments between the lowest-volume ($11 922) and highest-volume ($11 981) quartiles.
Diagnosis-related group payments for the index hospitalization contributed the single largest portion of both payments to hospitals and overall payments for episodes of care associated with bariatric surgery (Table 1). When assessing readmissions, 8.3% of patients were readmitted within 30 days of discharge from the index admission, varying from 3.9% to 14.9% across procedures. Readmissions accounted for 2.8% of total episode payments (with a mean of $4138 per readmission and a mean of $326 per patient), ranging from 1.3% to 4.8% across procedures. Only 1.5% of patients in our cohort had outlier payments for episodes surrounding bariatric surgery, accounting for 0.7% of total episode payments.
Overall, payments to physicians accounted for approximately 21.7% of total episode payments. A mean of 16.3% ($1971) of total episode payments was paid to the operating surgeon (Table 2), while anesthesiologists received 2% of total payments. Physician laboratory and imaging payments each accounted for less than 1% of total payments. Payments to physicians for other medical care provided during episodes of bariatric surgery accounted for approximately 2.6% of total payments.
Quiz Ref IDPayments for postacute care services accounted for 2.8% of total payments. These payments contributed to only a small portion of variation, with postacute care accounting for 2.4% of mean episode payments in the lowest quartile to 3.8% in the highest quartile. Outpatient care was the single largest contributor to payments for postacute care services, with 47% of all patients requiring outpatient care services (Table 3). This contributed to 1.6% of total payments. Home health care, skilled nursing facilities, rehabilitation, and nursing homes each accounted for less than 1% of the overall total payments.
New payment policies, such as the CMS Bundled Payments for Care Improvement Initiative, aim to improve quality and cost-efficiency by bundling together payments for hospitals, physicians, and postacute care for selected surgical specialty areas (eg, cardiac and orthopedic surgery). As enthusiasm for these programs continues to grow, these initiatives will extend to other common operations, such as bariatric surgery procedures. Because these policies hold hospitals, physicians, and postacute care services accountable for the total costs of care, they essentially shift the financial responsibility for poor quality (eg, complications) and inefficiency (eg, overuse of resources) to hospitals. In this policy environment, a detailed understanding of variation in the costs for bariatric surgery will be essential for hospitals to identify areas of risk and opportunities for improvement. Our study revealed variation in total episode costs across hospitals and procedures, suggesting that hospitals may have varying opportunities for success under bundled payment programs.
Previous work by Birkmeyer and colleagues3 demonstrated that Medicare payments for several different procedures varied significantly across hospitals. Our study builds on this work by specifically evaluating variation in hospital episode costs for another common procedure, bariatric surgery. This type of surgery appears to have a distinct pattern of hospital cost variation. In particular, we demonstrated that payments for the index hospitalization were the largest source of variation among hospitals, likely owing to inpatient complications that drive DRG upcoding (ie, DRGs with complications result in higher Medicare payments). Compared with high-risk procedures evaluated in the Birkmeyer et al3 study, this study demonstrated that the pattern of variation is inherently different in bariatric surgery. For example, postacute care services following hip fracture repair accounted for a large portion of variation in total payments, while much less variation was observed in payments for the index hospitalization. This difference in the pattern of variation emphasizes the importance of understanding cost variation specific to each procedure.
This study had several limitations. First, this study used administrative data, which have recognized limitations in capturing patient comorbidities and risk factors. Thus, some of the variation in payments may be a result of differences in case mix. Nonetheless, we used standard methods to adjust for case severity using administrative data. Second, we captured variation at the hospital level rather than with individual surgeons. Some of the variation may be driven by individual surgeon practices (eg, complications and radiology studies) that we did not capture with this analysis. However, the findings from this study could be used to identify variation in areas of care that may be driven by individual surgeon practices. Third, our study used episode payments as a surrogate for hospital costs. While these are not truly costs for the hospitals, they are costs from the payer perspective and thus provide a reflection of relative resource use surrounding episodes of bariatric surgery. Finally, our analysis included only Medicare patients, which may limit generalizability to the private sector. Nonetheless, Medicare is one key area where these programs are being piloted, so it is valuable for future policy decision making to understand variation in episode costs within this population. Furthermore, because private payers tend to follow Medicare with reimbursement reform, the findings from this study may have much broader implications as other non-Medicare hospitals and physicians implement bundled payment programs.
The results from our study have important implications for both payers and hospitals. Both parties have an interest in understanding the reasons underlying cost variation in bariatric surgery because they share the financial consequences from poor-quality care.13 Hospitals will assume more financial risk under a bundled-payments approach for the quality of the care they provide, increasing the stakes for providing these services. As a result, hospitals and physicians will have increased need for understanding areas of financial risk around episodes of care to succeed under these newer payment policies. Identifying the root cause of episode cost variation will be essential to target areas of highest financial risk, such as outlier payments, readmissions, and complications, that are associated with higher costs due to poor-quality care.14,15 Hospitals can use this information on a more granular level to help target specific modifiable factors that may drive costs (eg, routine radiology studies following bariatric surgery) in a particular area of care. While improving quality will remain paramount, other high-yield areas for improvement in bariatric surgery may rely on an increased emphasis on preventing overuse of discretionary resources to effectively reduce cost variation.
Payers can also use these findings to provide insight into designing bundled payment programs for specific procedures. For example, services such as postacute care are a small source of variation and may potentially be excluded from bundled payments for bariatric surgery. However, these could not be eliminated from episodes related to procedures such as total joint arthroplasty, where postacute care accounts for a substantial portion of episode costs.3,16 Furthermore, this information will also be important to payers because it demonstrates potential opportunities for cost-savings.17 Finally, while we have demonstrated variation in costs for services across all hospitals, there may be different patterns of variation within individual hospitals. As a result, it will also be imperative to understand the pattern of cost variation at the individual hospital level to respond to the risks and demands of episode payments.
In summary, as new bundled payment programs are implemented to promote cost-efficiency and high-quality care, new challenges in understanding hospital and physician financial risk surrounding bariatric surgery will come to the forefront. While these policies intend to provide incentives to improve quality and cost-efficient care, they will also create financial burden for hospitals and physicians that do not deliver high-quality care or manage resources effectively. To confront these challenges and succeed under newer payment systems, bariatric surgery hospitals and physicians will need to recognize where variations in the cost of care exist among the many services provided surrounding bariatric surgery. This will allow hospitals to understand their areas of greatest liability and opportunities for improvement.
Corresponding Author: Tyler R. Grenda, MD, MS, Center for Healthcare Outcomes and Policy, University of Michigan, 2800 Plymouth Rd, Bldg 16, Room 016-100N-12, Ann Arbor, MI 48109-2800 (firstname.lastname@example.org).
Accepted for Publication: May 29, 2015.
Published Online: September 16, 2015. doi:10.1001/jamasurg.2015.2394.
Author Contributions: Drs Grenda and Dimick had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Grenda, Dimick
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Grenda, Thumma, Dimick.
Obtained funding: Dimick.
Administrative, technical, or material support: Dimick.
Study supervision: Dimick.
Conflict of Interest Disclosures: Dr Dimick is a consultant and equity owner of ArborMetrix Inc, an Ann Arbor–based health care analytics and information technology firm, which was not involved, in whole or in part, in the collection or analysis of any data presented herein. No other disclosures were reported.
Funding/Support: Dr Grenda is supported by Agency for Healthcare Research and Quality grant 2T32HS000053.
Role of the Funder/Sponsor: The Agency for Healthcare Research and Quality had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: Dr Dimick is the Surgical Innovation Editor of JAMA Surgery. He was not involved in the editorial evaluation or decision to accept this article for publication.