- Email: [email protected]
Episode-based Payment Variation for Urologic Cancer Surgery
Accepted Manuscript Title: Episode-Based Payment Variation for Urological Cancer Surgery Author: Chad Ellimoottil, Jonathan Li, Zaojun Ye, James M. Dupree, Hye Sung Min, Deborah Kaye, Lindsey A. Herrel, David C. Miller PII: DOI: Reference:
S0090-4295(17)31082-8 https://doi.org/doi:10.1016/j.urology.2017.08.053 URL 20704
To appear in:
Urology
Received date: Accepted date:
7-4-2017 29-8-2017
Please cite this article as: Chad Ellimoottil, Jonathan Li, Zaojun Ye, James M. Dupree, Hye Sung Min, Deborah Kaye, Lindsey A. Herrel, David C. Miller, Episode-Based Payment Variation for Urological Cancer Surgery, Urology (2017), https://doi.org/doi:10.1016/j.urology.2017.08.053. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Episode-based payment variation for urological cancer surgery
Chad Ellimoottil, M.D., M.S.1,2, Jonathan Li, B.S.E. 1,2, Zaojun Ye, M.S. 1,2, James M. Dupree, M.D., M.P.H. 1,2, Hye Sung Min, M.S. 1,2, Deborah Kaye, M.D. 1,2, Lindsey A. Herrel, M.D., M.S. 1,2, David C. Miller, M.D., M.P.H. 1,2
University of Michigan, Ann Arbor 1.
Institute for Healthcare Policy and Innovation
2.
Dow Division of Health Services Research, Department of Urology
Address Correspondence to: Chad Ellimoottil, M.D., M.S. University of Michigan 2800 Plymouth Rd Bldg 16, 1st Floor, Room 152S Ann Arbor, MI 48109 [email protected] Office: 734-232-2247 Fax: 734-232-2400 Keywords: Bundled Payments; Urological Cancer Surgery; 90-day Episode Payments; Cystectomy; Prostatectomy; Nephrectomy Funding: This research was supported by funding from the National Cancer Institute (1R01-CA-174768-01-A1 to Dr. David Miller). Abstract word count: 246 1 Page 1 of 27
Manuscript word count: 2,761 (updated)
Abstract Objective: To investigate payment variation for three common urological cancer surgeries and evaluate the potential for applying bundled payment programs to these procedures. Methods: Using 2008-2011 Surveillance, Epidemiology, and End Results-Medicare linked data, we identified all beneficiaries aged greater than 65 years who underwent cystectomy, prostatectomy, or nephrectomy for cancer. Total episode payments were determined by aggregating hospital, professional, and post-acute care claims from the index surgical hospitalization through 90 days post-discharge. Total episode payments were then compared to examine hospital level-variation within each procedure type and the specific payment components (i.e. index hospitalization, professional, readmission, and post-acute care) driving spending variation. Results: Ninety-day episodes of care were identified for 1,849 cystectomies, 8,770 prostatectomies, and 4,304 nephrectomies. We observed wide variation in mean episode payments for all three conditions (cystectomy mean $35,102: range $24,112 to $57,238, prostatectomy mean $10,803: range $8,816 to $17,877, nephrectomy mean $17,475: range $11,681 to $26,711). Majority of payment variation was attributable to index hospitalization and post-acute care for cystectomy and nephrectomy and professional payments for prostatectomy. The most expensive hospitals by procedure
2 Page 2 of 27
each demonstrated a unique opportunity for spending reduction due to individual differences in component payment patterns between hospitals. Conclusion: Ninety-day episode payments for urological cancer surgery vary widely across hospitals in the United States. The key drivers of this payment variation differ for individual procedures and hospitals. Accordingly, hospitals will need individualized data and clinical re-design strategies to succeed with implementation of episode-based payment models for urological cancer care.
Introduction The National Center for Health Statistics estimates that total expenditures for inpatient surgery in the United States approached $160 billion dollars in 20131. For many common surgical conditions, the total “episode payment” (i.e., payments made for patient’s surgical admission, readmissions, professional fees, and post-discharge care) is known to vary substantially across hospitals2. To mitigate variation in episodic spending, public and private payers are investigating alternative payment models such as episode-based bundled payments to financially incentivize spending reductions achieved through improving quality and coordination of care and reducing repetitive tests and procedures for inpatient hospitalization3,4. Accordingly, supporters of such alternative payment models believe that bundled payments may help reduce episodebased payment variation for many conditions including cancer surgery. While the sources of payment variation for common surgical procedures, including colectomy, coronary artery bypass graft (CABG), and joint replacement surgery, have been investigated extensively5–7, episode-based payment variation for 3 Page 3 of 27
urological cancer surgery is not well defined. For instance, sources of divergent episode costs for urological cancer surgery may mimic those of common orthopedic procedures, in which post-acute care spending accounts for the majority of payment variation5,7. One may expect this to be the case for surgical procedures like radical cystectomy, in which patients are generally older with more comorbid conditions8–11. On the other hand, episodic payment variation for urological cancer surgery may result from additional hospitalization payments due to perioperative complications, which may be the case for patients undergoing prostatectomy or nephrectomy given that utilization of post-acute care and readmissions in this population is likely to be relatively infrequent 12–16. Previous investigators have examined payment variation for select urological cancers17–19. Herein, we build on this previous work by using national data to explore the episode payment variation for three common urological cancers. Specifically, we used Surveillance, Epidemiology, and End Results-linked Medicare data to characterize total episode and component spending for patients undergoing radical cystectomy, nephrectomy and prostatectomy. In addition, we examined patterns at high- and lowpayment hospitals to identify opportunities for spending reduction among the most expensive hospitals. By virtue of this approach, our findings will provide policymakers, hospital administrators, and physicians with insight on sources of episode payment variation for urological cancer surgery and whether alternative payment models such as bundled payments might be appropriate for these conditions.
4 Page 4 of 27
Methods Data sources We utilized data from the 2008-2011 Surveillance, Epidemiology and End Results registry linked to Medicare claims (SEER-Medicare) as our primary dataset. SEER-Medicare is a national dataset that includes clinical, demographic, and outcomes data from 17 major cancer registries across the United States20. Claims from this dataset are divided into four files: 1) MEDPAR, which includes claims from inpatient admissions to acute care hospitals and skilled nursing facility visits, 2) Carrier, which includes claims from healthcare professionals including urologists and anesthesiologists, 3) Outpatient claims, and 4) Home Health Agency claims.
Patient Cohort Our study cohort included all Medicare beneficiaries greater than 65 years of age who underwent radical prostatectomy, cystectomy, or nephrectomy from 2008 through 2011. We identified these patients using International Classification of Diseases, Ninth Edition (ICD-9) diagnosis and procedure codes, Current Procedural Terminology (CPT) codes, and information from the cancer registry to confirm whether a subject had the urological cancer of interest. We excluded patients who were missing any of the following criteria: 1) ICD-9 diagnosis code for the cancer of interest, 2) ICD-9 procedure code for the cancer of interest, 3) CPT code for the surgery of interest, or 4) Cancer diagnosis listed in the SEER file. Moreover, we excluded patients who did not have full enrollment data, complete claims, Medicare coverage, or who were participating in a
5 Page 5 of 27
Medicare HMO from 12 months before diagnosis to the end of the study period (December 31st, 2011) or time of death.
Measuring 90-day episode payments First, we identified the index hospitalization (inpatient admission for surgery) for each patient meeting our inclusion criteria. We then linked the index hospitalization with all claims that occurred 90 days after surgery. We selected a 90-day post-discharge window for two major reasons. First, this period is consistent with the global period for surgery. Second, this interval is employed by existing bundled payment-based programs7,21. Finally, our choice of episode length allowed us to assess claims for postacute care and other services that may extend beyond 30 days after the initial surgery. Total episode payments consisted of the components related to index hospitalization, readmission, professional, and post-acute care payments per single episode of care. Claims for index hospitalization and readmission payments within 90 days after surgery were obtained from the MEDPAR file. We utilized the Carrier file to evaluate professional payments. Post-acute care spending was obtained from claims listed in the Outpatient, Home Health, and MEDPAR files (inpatient rehabilitation and skilled nursing facility). Prices were then standardized using previously established methods4,5. The purpose of price standardization is to remove geographic variation in reimbursements due to adjustments that the Centers for Medicare and Medicaid Services (CMS) make for local labor costs, hospital teaching status, and factors specific to individual hospitals. To reduce the influence of outliers, we also performed a 99% Winsorization, whereby
6 Page 6 of 27
we adjusted episode payments above the 99th percentile to the 99th percentile value— and likewise for payments below the first percentile. The mean component payments and total 90-day episode payments were then calculated for cystectomy, prostatectomy, and nephrectomy. To increase precision, we limited our analysis to hospitals that had at least 10 cases of the procedure of interest during the study period.
Examining variation in episode payments We first examined variation in total and component episode payments for urological cancers by comparing mean episode payments across individual hospitals. Next, we assessed payment variation among the most and least expensive hospitals. To do this, we classified hospitals into four quartiles based on mean total episode payments for each individual procedure. We then evaluated the difference between the lowest and highest spending quartiles by calculating the difference in mean total episode payments and differences in index hospitalization, readmission, professional, and post-acute care payments. Additionally, we calculated the degree to which differences in index hospitalization, readmission, professional, and post-acute care payments contributed to the total variation in 90-day episode payments between highand low-spending quartiles. Finally, we performed a Spearman rank correlation to investigate the association of high spending in one payment component with high spending in total episode payments for cystectomy, prostatectomy, and nephrectomy.
Examining opportunities to reduce episode payments at the individual hospital level
7 Page 7 of 27
We then focused our analysis on individual hospitals in order to determine possible ‘pathways’ that each high-spending hospitals might take in order to reduce spending to the national average (the target for a presumed bundle payment). For each procedure type, we selected 10 hospitals with the highest total episode payments and ranked these hospitals according to component spending (index hospitalization, readmission, profession, and post-acute care) relative to the remaining hospitals in the dataset. With this step, we could determine the components that contributed the most to an individual hospital’s high spending status based on whether that particular component was ranked in the top quarter among all hospitals for the procedure of interest. Additionally, the component with the greatest opportunity for spending reduction at an individual hospital was determined as the component with the greatest deviation (via standardized z-score) from the average component payment across all hospitals for the procedure of interest.
Sensitivity Analysis Due to the rising prevalence and importance of nephron-sparing approaches in the treatment of kidney cancer, we divided our nephrectomy cohort into partial nephrectomy and radical nephrectomy groups based on corresponding ICD-9 and CPT codes. We then examined differences in mean total episode payments and sources of payment variation between the partial and radical nephrectomy groups. Second, we stratified episode payments by quartiles of hospital surgical volume to determine the effect of volume on payment variation for urological cancer surgery.
8 Page 8 of 27
We then used the ANOVA test of significance to determine whether total episode and component spending varied among the four volume quartiles in our cohort.
All analyses were performed using SAS (Cary, NC) and STATA 14/SE (College Station, TX), and at the 5% significance level. This study was deemed exempt from review by the Institutional Research Board at our institution.
9 Page 9 of 27
Results We identified 14,923 patients who underwent surgery for urological cancers from 2008 through 2011. This cohort included 1,849 patients who underwent cystectomy, 8,770 patients with prostatectomy, and 4,304 patients with nephrectomy. Mean age for patients undergoing cystectomy was 76.3 years old (range 66-94), 70.5 years old (range 65-87) for prostatectomy, and 75.0 years old (range 66-99) for nephrectomy. Total episode payments varied widely across hospitals for all three surgical procedures. Figure 1 presents the range of values for mean total episode payments among hospitals for cystectomy, prostatectomy, and nephrectomy. We observed wide variation in mean episode payments for all three conditions (cystectomy mean $35,102: range $24,112 to $57,238, prostatectomy mean $10,803: range $8,816 to $17,877, nephrectomy mean $17,475: range $11,681 to $26,711). Considerable differences were noted between hospitals at the highest and lowest episode payment quartiles for cystectomy, prostatectomy, and nephrectomy (Table 1). The greatest difference in total episode payments between high-spending and low-spending hospitals was noted for cystectomy among all procedures studied, with a $13,057 difference in total episode expenditures between the highest and lowest payment quartiles. Using Spearman’s rank correlation to examine the relationship between total and component episode payments for each cancer (Table 2), we observed that high payments for index hospitalization (R=0.79, p<0.001) and post-acute care (R=0.70, p<0.001) had the strongest correlation with high total episode payments for cystectomy. Meanwhile, total episode payments were strongly correlated positively with professional
10 Page 10 of 27
payments (R=0.70, p<0.001) for prostatectomy and index hospitalization (R=0.69, p<0.001) and post-acute care (R=0.71, p<0.001) payments for nephrectomy. Finally, we observed considerable differences in patterns of component payments among individual high-spending hospitals (Figure 2). Among the 10 highest spending hospitals for cystectomy, the component payment that drove differences in total episode payments varied among individual hospitals. Specifically, each of the top hospitals for 90-day episode spending varied by the types of component payments within the top quarter for that particular category (e.g. index hospitalization, readmissions, etc). In addition, the greatest opportunity for spending reduction (i.e. greatest deviation from the mean component payment for the entire cohort) varied for cystectomy among the most expensive hospitals. For instance, among the 10 most expensive hospitals for cystectomy in Figure 2A, the greatest opportunity for spending reduction was readmission payments for some hospitals (Hospitals A, B, and E) and post-acute care payments for other hospitals (Hospitals D, G, H, and J). Similar levels of inter-hospital variability in component payment patterns were observed for the most expensive hospitals for nephrectomy and prostatectomy (Figures 2B-C).
Sensitivity Analysis In our sensitivity analysis, we observed that average 90-day episode payments to hospitals were $18,052 (range: $12,810 to $26,711) for radical nephrectomies and $15,910 (range: $10,494 to $25,091) for partial nephrectomies (Supplementary Table 1). Variation in index hospitalization payments and post-acute care payments represented the major sources of variation between hospitals at the highest and lowest
11 Page 11 of 27
payment quartiles for both partial and radical nephrectomies—similar to trends we noted in the overall nephrectomy cohort (Supplementary Table 2). Stratification of total episode payments by quartiles of hospital urological surgical volume showed no significant differences in total episode payments among the hospital volume quartiles (p=0.09). Significant association was observed only for post-acute care payments with hospital volume (p=0.02). No other associations between component payments and hospital volume were observed otherwise (Supplementary Table 3).
Comment Across a nationally representative sample of hospitals, we identified wide variation in 90-day episode payments for cystectomy, prostatectomy, and nephrectomy. In addition, when we compared spending patterns between the most and least expensive hospitals, we observed that the drivers of payment variation differed by surgical procedure. Finally, we found that even among hospitals considered the “most expensive,” the sources of variation were not uniform. Collectively, these findings suggest that while there is sufficient Medicare payment variation for urological cancer surgery to be a potential candidate for bundled payment programs, hospitals will need to evaluate their own data and design organization and procedure-specific strategies to achieve success under such programs. Our finding that total episode payment variation exists for urological cancer surgery is consistent with current literature for other surgical conditions1,5,16,20,21,22. Specifically, previous studies have demonstrated wide variation in episode-based payments for common surgical procedures such as CABG, colectomy, joint 12 Page 12 of 27
replacement, and other oncological surgical procedures. However, the sources of payment variation identified in our analysis are not uniformly consistent with the existing literature. For example, in one study index hospitalization payments were the major source of variation between high and low spending hospitals for cystectomy and prostatectomy22. In contrast, another study on episode spending for prostatectomy suggested that variation in total episode payments is distributed more evenly among multiple components—a result more consistent with our own findings17. Our observation that sources of payment variation are highly variable among individual hospital may explain some of the discrepancies in the literature regarding general trends among specific hospital-level cohorts. Our study has several limitations. First, by utilizing SEER-Medicare data, we restrict our study to Medicare beneficiaries greater than 65 years of age. While we would expect the average age of surgery for all patients undergoing cystectomy to be greater than 65 years old, the average age may be lower for patients undergoing prostatectomy or nephrectomy. However, we deemed Medicare beneficiaries as the most appropriate population given the aggressive efforts by the CMS to develop and implement alternative payment models. Second, our study was limited to patients who had surgery at hospitals within the SEER regions. While this may limit the generalizability of our results beyond these regions, the SEER regions are considered nationally representative of the general oncologic population in the United States. Thirdly, we only examined hospital-level variation in our analysis. While recent literature has shown that the majority of payment variation occurs at the hospital-level, physicianlevel spending variation still plays an important role in current bundled payment
13 Page 13 of 27
initiatives21,25. In addition, with the increasing prevalence of co-managed care between surgeon and internal medicine specialties, the topic of physician-level variation will be the subject of future investigation26. Finally, we did not examine the impact of case-mix (e.g., comorbidity, socioeconomic status, tumor staging) on episode spending. Nonetheless, at present, Medicare does not include risk-adjustment in their current, bundled payment models23. Additionally, we believe that the use of risk-adjustment for episode payment is a highly complex topic that, while beyond the scope of this study, will be a necessary focus of future work. These limitations notwithstanding, our findings will help many stakeholders anticipate the potential role and impact of bundled payments for urological cancer surgery. Relevant to payers and policymakers, our finding that fairly large episode payment variation exists for the three most common urological cancer surgeries implies that these conditions may be candidates for bundled payment programs. In fact, the American Urological Association is in the process of developing bundled payment programs for these conditions27. For hospital administrators, we show that health systems will need to compile institution-specific data and clinical re-design strategies to succeed with implementation of episode-based payment models for urological cancer surgery. For instance, the optimal strategy for one hospital may be to reduce readmissions for cystectomy patients, while a different hospital may find post-acute care spending to be its major source of high expenditures. Much like the preceding example, the hospitals within our study cohort demonstrate different characteristics in payments that may require further investigation of the administrative and clinical factors driving high spending in a
14 Page 14 of 27
particular component. As a result, we would anticipate different health systems to execute varying methods of quality improvement and clinical redesign tailored to the areas of disproportionate spending within that specific institution.
15 Page 15 of 27
Conclusions Our collective findings suggest that radical cystectomy, prostatectomy and nephrectomy may be suitable candidates for bundled payment programs due to wide episode payment variation. Additionally, hospitals will need to design specific strategies based on their individual spending patterns to succeed in these types of payment programs for urological cancer surgery. Moving forward, research in this area will also need to examine the influence of patient characteristics such as age, comorbidities, cancer stage and socioeconomic status on 90-day episode spending. In the end, the opportunity for policymakers to reduce urological cancer surgery spending will depend on the magnitude of unwarranted episode payment variation across hospitals. Moreover, the opportunity for hospitals to succeed in these programs will depend on the ability to identify and address local sources of excessive spending, while at the same time maintaining high-quality outcomes and patient experience.
16 Page 16 of 27
References 1.
National Center for Health Statistics. Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities. Hyattsville, MD.; 2016: 300-303. https://www.cdc.gov/nchs/data/hus/hus15.pdf#096.
2.
Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45(6 Pt 1):1783-1795. doi:10.1111/j.1475-6773.2010.01150.x.
3.
Iglehart JK. Bundled payment for ESRD--including ESAs in Medicare’s dialysis package. N Engl J Med. 2011;364(7):593-595. doi:10.1056/NEJMp1014187.
4.
Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare Payments for Common Inpatient Procedures: Implications for Episode-Based Payment Bundling. Health Serv Res. 2010;45(6p1):1783-1795. doi:c.
5.
Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. Health Aff (Millwood). 2011;30(11):2107-2115. doi:10.1377/hlthaff.2011.0783.
6.
Abdelsattar ZM, Birkmeyer JD, Wong SL. Variation in Medicare Payments for Colorectal Cancer Surgery. J Oncol Pract. 2015;11(5):391-395. doi:10.1200/JOP.2015.004036.
7.
Iorio R, Clair AJ, Inneh IA, Slover JD, Bosco JA, Zuckerman JD. Early Results of Medicare’s Bundled Payment Initiative for a 90-Day Total Joint Arthroplasty Episode of Care. J Arthroplasty. 2016;31(2):343-350. doi:10.1016/j.arth.2015.09.004.
8.
Jacobs BL, Zhang Y, Tan H-J, Ye Z, Skolarus TA, Hollenbeck BK. Hospitalization Trends After Prostate and Bladder Surgery: Implications of Potential Payment Reforms. J Urol. 2013;189(1):59-65. doi:10.1016/j.juro.2012.08.182.
9.
Skolarus TA, Jacobs BL, Schroeck FR, et al. Understanding Hospital Readmission Intensity after Radical Cystectomy. J Urol. 2015;193(5):1500-1506. doi:10.1016/j.juro.2014.10.107.
10.
Minnillo BJ, Maurice MJ, Schiltz N, et al. Few modifiable factors predict readmission following radical cystectomy. Can Urol Assoc J. 2015;9(7-8):E43946. doi:10.5489/cuaj.2793.
11.
Aghazadeh MA, Barocas DA, Salem S, et al. Determining factors for hospital discharge status after radical cystectomy in a large contemporary cohort. J Urol. 2011;185(1):85-89. doi:10.1016/j.juro.2010.08.016.
12.
Liu J-J, Maxwell BG, Panousis P, Chung BI. Perioperative outcomes for laparoscopic and robotic compared with open prostatectomy using the National Surgical Quality Improvement Program (NSQIP) database. Urology. 17 Page 17 of 27
2013;82(3):579-583. doi:10.1016/j.urology.2013.03.080. 13.
Trinh Q-D, Sammon J, Sun M, et al. Perioperative outcomes of robot-assisted radical prostatectomy compared with open radical prostatectomy: results from the nationwide inpatient sample. Eur Urol. 2012;61(4):679-685. doi:10.1016/j.eururo.2011.12.027.
14.
Moschini M, Gandaglia G, Dell’Oglio P, et al. Incidence and Predictors of 30-Day Readmission in Patients Treated With Radical Cystectomy: A Single Center European Experience. Clin Genitourin Cancer. 2016;14(4):e341-6. doi:10.1016/j.clgc.2015.12.017.
15.
Autorino R, Zargar H, Butler S, Laydner H, Kaouk JH. Incidence and risk factors for 30-day readmission in patients undergoing nephrectomy procedures: a contemporary analysis of 5276 cases from the National Surgical Quality Improvement Program database. Urology. 2015;85(4):843-849. doi:10.1016/j.urology.2014.11.044.
16.
Schmid M, Chiang HA, Sood A, et al. Causes of hospital readmissions after urologic cancer surgery. Urol Oncol. 2016;34(5):236.e1-236.e11. doi:10.1016/j.urolonc.2015.11.019.
17.
Herrel LA, Syrjamaki JD, Linsell SM, Miller DC, Dupree JM. Identifying Drivers of Episode Cost Variation With Radical Prostatectomy. Urology. 2016;97:105-110. doi:10.1016/j.urology.2016.05.071.
18.
Pate SC, Uhlman MA, Rosenthal JA, Cram P, Erickson BA. Variations in the open market costs for prostate cancer surgery: a survey of US hospitals. Urology. 2014;83(3):626-630. doi:10.1016/j.urology.2013.09.066.
19.
Wang S-Y, Wang R, Yu JB, et al. Understanding regional variation in Medicare expenditures for initial episodes of prostate cancer care. Med Care. 2014;52(8):680-687. doi:10.1097/MLR.0000000000000158.
20.
Overview of the SEER Program. National Cancer Institute: Surveillance, Epidemiology, and End Results (SEER) Program. https://seer.cancer.gov/about/overview.html. Accessed February 8, 2017.
21.
Centers for Medicare & Medicaid Services (CMS) H. Medicare Program; Advancing Care Coordination Through Episode Payment Models (EPMs); Cardiac Rehabilitation Incentive Payment Model; and Changes to the Comprehensive Care for Joint Replacement Model (CJR). Final rule. Fed Regist. 2017;82(1):180651. http://www.ncbi.nlm.nih.gov/pubmed/28071874.
22.
Nathan H, Atoria CL, Bach PB, Elkin EB. Hospital volume, complications, and cost of cancer surgery in the elderly. J Clin Oncol. 2015;33(1):107-114. doi:10.1200/JCO.2014.57.7155.
23.
Ellimoottil C, Ryan AM, Hou H, Dupree J, Hallstrom B, Miller DC. Medicares New Bundled Payment For Joint Replacement May Penalize Hospitals That Treat Medically Complex Patients. Health Aff. 2016;35(9):1651-1657. 18 Page 18 of 27
doi:10.1377/hlthaff.2016.0263. 24.
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.
25.
Ejaz A, Gani F, Kim Y, Pawlik TM. Variation in inpatient hospital and physician payments among patients undergoing general versus orthopedic operations. Surgery. 2016;160(6):1657-1665. doi:10.1016/j.surg.2016.07.013.
26.
Sharma G, Kuo Y-F, Freeman J, Zhang DD, Goodwin JS. Comanagement of hospitalized surgical patients by medicine physicians in the United States. Arch Intern Med. 2010;170(4):363-368. doi:10.1001/archinternmed.2009.553.
27.
Dupree JM. Alternative Payment Models in Urology. AUA News. February 2017:23.
Figure 1. Mean Total Episode Payments by Hospital for (A) Cystectomy, (B) Prostatectomy and (C) Nephrectomy, 2008-2011 SEER-Medicare. Individual marker denotes single hospital with caseload ≥ 10 for procedure of interest.
Figure 2. Episode-based Payment Patterns among the Highest Spending Hospitals for (A) Cystectomy, (B) Prostatectomy and (C) Nephrectomy, 2008-2011 SEER-Medicare. Shaded cells indicate that the hospital is in the top quartile for the particular spending component (e.g. professional, readmissions). Star indicates greatest opportunity for spending reduction for the hospital of interest. Greatest opportunity for spending reduction for an individual hospital was determined as the component with the greatest deviation from the average component payment across all hospitals for the procedure of interest.
19 Page 19 of 27
Table 1. Comparison of Component Payments between Hospitals at the Highest and Lowest Episode Payment Quartiles, 2008-2011 SEER Medicare
Procedure Cystectomy
Prostatectomy
Nephrectomy
Payment Componenta Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment
Low Payment Quartile 16,252 2,355 5,328 4,185 28,121 6,818 134 2,523 206 9,680 9,342 589 2,906 1,222 14,059
High Payment Quartile 22,441 3,447 7,079 8,211 41,177 7,207 672 3,512 797 12,187 11,485 1,907 3,980 3,772 21,144
Difference High and Low Quartile 6,189 1,092 1,750 4,025 13,057 388 538 989 591 2,507 2,143 1,318 1,074 2,550 7,085
% Total Difference Attributed to Component 47.4 8.4 13.4 30.8 100.0 15.5 21.5 39.5 23.6 100.0 30.2 18.6 15.2 36.0 100.0
a
All amounts in US Dollars
20 Page 20 of 27
Table 2. Spearman Rank Correlation between Total Episode Payment and Individual Component Payments, 20082011 SEER Medicare Spearman Correlation Coefficient of Total Episode vs Component, R (p-value) Procedure
Index
Professional
Readmission
Post-acute Care
Cystectomy
0.79 (<0.001)
0.66 (<0.001)
0.35 (0.029)
0.70 (<0.001)
Prostatectomy
0.51 (<0.001)
0.70 (<0.001)
0.49 (<0.001)
0.61 (<0.001)
Nephrectomy
0.69 (<0.001)
0.57 (<0.001)
0.56 (<0.001)
0.71 (<0.001)
21 Page 21 of 27
Figure 1.
22 Page 22 of 27
Figure 2. (A) Payment Patterns for Highest Spending Hospitals for Cystectomy Hospital
Index Hospitalization
Professional
B
D
E F
Post-acute Care
A C
Readmissions
G H
I
J
(B) Payment Patterns for Highest Spending Hospitals for Prostatectomy Hospital
Index Hospitalization
Professional
Readmissions
A
B
C D
E F
G H I J
Post-acute Care
23 Page 23 of 27
(C) Payment Patterns for Highest Spending Hospitals for Nephrectomy Hospital
Index Hospitalization
Professional
C D
E F G
H I J
Post-acute Care
A B
Readmissions
24 Page 24 of 27
Editorial Comment
Christopher P. Filson, MD, MS Department of Urology, Emory University School of Medicine
Contact Information: Department of Urology Emory University School of Medicine 1365 Clifton Rd NE, Suite B1400 Atlanta, GA, 30322 404-778-4583 [email protected]
Amidst ever-rising health care expenditures in the United States, stakeholders and policymakers have turned to alternative payment models as mechanisms to control costs. These tools also aim to maintain—and possibly improve—the quality of health care delivered in this country. Bundled payment programs are an alternative payment model that have shown promise in curtailing costs and maintaining quality among patients undergoing lower extremity joint (e.g., hip and knee) replacement.1 As Dr. Ellimoottil and co-authors demonstrate in the above article, the wide two-fold variation in episode costs for urological cancer operations
25 Page 25 of 27
makes consideration of bundled payment programs in these instances quite appealing at first glance.
However, the most compelling finding from this work was the observation that the costliest hospitals varied considerably in which components contributed most to episode-based spending, suggesting that a one-size-fits-all approach for hospital-based cost containment would likely fail for urologic cancers. Instead, this work highlights the importance for medical centers to actively measure and understand their practice patterns and delivery of care throughout the entire episode surrounding urological cancer operations. From there, hospitals can decide on interventions that suit them best. For example, surgeon intraoperative cost scorecards2 and pathways to shorten hospitalization3 have both been shown to decrease episode payments. Whether these would be applicable for a particular hospital would depend on whether intraoperative equipment use or prolonged hospitalizations were cost drivers, respectively.
The authors also highlight an important limitation of their work, in that this analysis did not consider cost variation attributed to physician-level practice patterns. As just one example, the delivery of low-value peri- and postoperative care based on dogma—in the absence of evidence of benefit—may quickly rack up costs. A recent analysis of the Virginia All-Payer Claims database highlighted that the majority of low-value care was not related to expensive “big-ticket” services, but instead high-volume, low-cost services.4 In fact, the costliest low-value service overall (at $227.8 million in 2014) was receipt of preoperative labs for low-risk patients undergoing low-risk surgery. The findings from that study are worth reflecting on. When ordering an X-ray to check stent position following a radical cystectomy, or checking fluid creatinine
26 Page 26 of 27
levels on a low-output drain following prostatectomy, are we delivering a high-value service, or reassuring ourselves (as providers) that everything is stable? In the big picture, are the costs associated with those studies worth the low likelihood of being abnormal? I don’t have the answer to these questions, and medico-legal concerns also exist, but bundled payment programs may provide the impetus for hospitals to delve into these issues and “trim the fat”, so to speak.
References 1. Dummit LA, Kahvecioglu D, Marrufo G, et al. Association between hospital participation in a Medicare bundled payment intitiative and payments and quality outcomes for lower extremity joint replacement episodes. JAMA 2016; 316(12):1267–1278. 2. Zygourakis CC, Valencia V, Moriates C, et al. Association between surgeon scorecard use and operating room costs. JAMA Surgery 2017; 152(3): 284 – 291. 3. Regenbogen SE, Cain-Nielsen AH, Norton EC, et al. Costs and consequences of early hospital discharge after major inpatient surgery in older adults. JAMA Surgery 2017; 152(5) [epub ahead of print] 4. Mafi JN, Russell K, Bortz BA, et al. Low-cost, high-volume health services contribute the most to unnecessary health spending. Health Aff 2017; 36(10): 1701 – 1704.
27 Page 27 of 27
S0090-4295(17)31082-8 https://doi.org/doi:10.1016/j.urology.2017.08.053 URL 20704
To appear in:
Urology
Received date: Accepted date:
7-4-2017 29-8-2017
Please cite this article as: Chad Ellimoottil, Jonathan Li, Zaojun Ye, James M. Dupree, Hye Sung Min, Deborah Kaye, Lindsey A. Herrel, David C. Miller, Episode-Based Payment Variation for Urological Cancer Surgery, Urology (2017), https://doi.org/doi:10.1016/j.urology.2017.08.053. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Episode-based payment variation for urological cancer surgery
Chad Ellimoottil, M.D., M.S.1,2, Jonathan Li, B.S.E. 1,2, Zaojun Ye, M.S. 1,2, James M. Dupree, M.D., M.P.H. 1,2, Hye Sung Min, M.S. 1,2, Deborah Kaye, M.D. 1,2, Lindsey A. Herrel, M.D., M.S. 1,2, David C. Miller, M.D., M.P.H. 1,2
University of Michigan, Ann Arbor 1.
Institute for Healthcare Policy and Innovation
2.
Dow Division of Health Services Research, Department of Urology
Address Correspondence to: Chad Ellimoottil, M.D., M.S. University of Michigan 2800 Plymouth Rd Bldg 16, 1st Floor, Room 152S Ann Arbor, MI 48109 [email protected] Office: 734-232-2247 Fax: 734-232-2400 Keywords: Bundled Payments; Urological Cancer Surgery; 90-day Episode Payments; Cystectomy; Prostatectomy; Nephrectomy Funding: This research was supported by funding from the National Cancer Institute (1R01-CA-174768-01-A1 to Dr. David Miller). Abstract word count: 246 1 Page 1 of 27
Manuscript word count: 2,761 (updated)
Abstract Objective: To investigate payment variation for three common urological cancer surgeries and evaluate the potential for applying bundled payment programs to these procedures. Methods: Using 2008-2011 Surveillance, Epidemiology, and End Results-Medicare linked data, we identified all beneficiaries aged greater than 65 years who underwent cystectomy, prostatectomy, or nephrectomy for cancer. Total episode payments were determined by aggregating hospital, professional, and post-acute care claims from the index surgical hospitalization through 90 days post-discharge. Total episode payments were then compared to examine hospital level-variation within each procedure type and the specific payment components (i.e. index hospitalization, professional, readmission, and post-acute care) driving spending variation. Results: Ninety-day episodes of care were identified for 1,849 cystectomies, 8,770 prostatectomies, and 4,304 nephrectomies. We observed wide variation in mean episode payments for all three conditions (cystectomy mean $35,102: range $24,112 to $57,238, prostatectomy mean $10,803: range $8,816 to $17,877, nephrectomy mean $17,475: range $11,681 to $26,711). Majority of payment variation was attributable to index hospitalization and post-acute care for cystectomy and nephrectomy and professional payments for prostatectomy. The most expensive hospitals by procedure
2 Page 2 of 27
each demonstrated a unique opportunity for spending reduction due to individual differences in component payment patterns between hospitals. Conclusion: Ninety-day episode payments for urological cancer surgery vary widely across hospitals in the United States. The key drivers of this payment variation differ for individual procedures and hospitals. Accordingly, hospitals will need individualized data and clinical re-design strategies to succeed with implementation of episode-based payment models for urological cancer care.
Introduction The National Center for Health Statistics estimates that total expenditures for inpatient surgery in the United States approached $160 billion dollars in 20131. For many common surgical conditions, the total “episode payment” (i.e., payments made for patient’s surgical admission, readmissions, professional fees, and post-discharge care) is known to vary substantially across hospitals2. To mitigate variation in episodic spending, public and private payers are investigating alternative payment models such as episode-based bundled payments to financially incentivize spending reductions achieved through improving quality and coordination of care and reducing repetitive tests and procedures for inpatient hospitalization3,4. Accordingly, supporters of such alternative payment models believe that bundled payments may help reduce episodebased payment variation for many conditions including cancer surgery. While the sources of payment variation for common surgical procedures, including colectomy, coronary artery bypass graft (CABG), and joint replacement surgery, have been investigated extensively5–7, episode-based payment variation for 3 Page 3 of 27
urological cancer surgery is not well defined. For instance, sources of divergent episode costs for urological cancer surgery may mimic those of common orthopedic procedures, in which post-acute care spending accounts for the majority of payment variation5,7. One may expect this to be the case for surgical procedures like radical cystectomy, in which patients are generally older with more comorbid conditions8–11. On the other hand, episodic payment variation for urological cancer surgery may result from additional hospitalization payments due to perioperative complications, which may be the case for patients undergoing prostatectomy or nephrectomy given that utilization of post-acute care and readmissions in this population is likely to be relatively infrequent 12–16. Previous investigators have examined payment variation for select urological cancers17–19. Herein, we build on this previous work by using national data to explore the episode payment variation for three common urological cancers. Specifically, we used Surveillance, Epidemiology, and End Results-linked Medicare data to characterize total episode and component spending for patients undergoing radical cystectomy, nephrectomy and prostatectomy. In addition, we examined patterns at high- and lowpayment hospitals to identify opportunities for spending reduction among the most expensive hospitals. By virtue of this approach, our findings will provide policymakers, hospital administrators, and physicians with insight on sources of episode payment variation for urological cancer surgery and whether alternative payment models such as bundled payments might be appropriate for these conditions.
4 Page 4 of 27
Methods Data sources We utilized data from the 2008-2011 Surveillance, Epidemiology and End Results registry linked to Medicare claims (SEER-Medicare) as our primary dataset. SEER-Medicare is a national dataset that includes clinical, demographic, and outcomes data from 17 major cancer registries across the United States20. Claims from this dataset are divided into four files: 1) MEDPAR, which includes claims from inpatient admissions to acute care hospitals and skilled nursing facility visits, 2) Carrier, which includes claims from healthcare professionals including urologists and anesthesiologists, 3) Outpatient claims, and 4) Home Health Agency claims.
Patient Cohort Our study cohort included all Medicare beneficiaries greater than 65 years of age who underwent radical prostatectomy, cystectomy, or nephrectomy from 2008 through 2011. We identified these patients using International Classification of Diseases, Ninth Edition (ICD-9) diagnosis and procedure codes, Current Procedural Terminology (CPT) codes, and information from the cancer registry to confirm whether a subject had the urological cancer of interest. We excluded patients who were missing any of the following criteria: 1) ICD-9 diagnosis code for the cancer of interest, 2) ICD-9 procedure code for the cancer of interest, 3) CPT code for the surgery of interest, or 4) Cancer diagnosis listed in the SEER file. Moreover, we excluded patients who did not have full enrollment data, complete claims, Medicare coverage, or who were participating in a
5 Page 5 of 27
Medicare HMO from 12 months before diagnosis to the end of the study period (December 31st, 2011) or time of death.
Measuring 90-day episode payments First, we identified the index hospitalization (inpatient admission for surgery) for each patient meeting our inclusion criteria. We then linked the index hospitalization with all claims that occurred 90 days after surgery. We selected a 90-day post-discharge window for two major reasons. First, this period is consistent with the global period for surgery. Second, this interval is employed by existing bundled payment-based programs7,21. Finally, our choice of episode length allowed us to assess claims for postacute care and other services that may extend beyond 30 days after the initial surgery. Total episode payments consisted of the components related to index hospitalization, readmission, professional, and post-acute care payments per single episode of care. Claims for index hospitalization and readmission payments within 90 days after surgery were obtained from the MEDPAR file. We utilized the Carrier file to evaluate professional payments. Post-acute care spending was obtained from claims listed in the Outpatient, Home Health, and MEDPAR files (inpatient rehabilitation and skilled nursing facility). Prices were then standardized using previously established methods4,5. The purpose of price standardization is to remove geographic variation in reimbursements due to adjustments that the Centers for Medicare and Medicaid Services (CMS) make for local labor costs, hospital teaching status, and factors specific to individual hospitals. To reduce the influence of outliers, we also performed a 99% Winsorization, whereby
6 Page 6 of 27
we adjusted episode payments above the 99th percentile to the 99th percentile value— and likewise for payments below the first percentile. The mean component payments and total 90-day episode payments were then calculated for cystectomy, prostatectomy, and nephrectomy. To increase precision, we limited our analysis to hospitals that had at least 10 cases of the procedure of interest during the study period.
Examining variation in episode payments We first examined variation in total and component episode payments for urological cancers by comparing mean episode payments across individual hospitals. Next, we assessed payment variation among the most and least expensive hospitals. To do this, we classified hospitals into four quartiles based on mean total episode payments for each individual procedure. We then evaluated the difference between the lowest and highest spending quartiles by calculating the difference in mean total episode payments and differences in index hospitalization, readmission, professional, and post-acute care payments. Additionally, we calculated the degree to which differences in index hospitalization, readmission, professional, and post-acute care payments contributed to the total variation in 90-day episode payments between highand low-spending quartiles. Finally, we performed a Spearman rank correlation to investigate the association of high spending in one payment component with high spending in total episode payments for cystectomy, prostatectomy, and nephrectomy.
Examining opportunities to reduce episode payments at the individual hospital level
7 Page 7 of 27
We then focused our analysis on individual hospitals in order to determine possible ‘pathways’ that each high-spending hospitals might take in order to reduce spending to the national average (the target for a presumed bundle payment). For each procedure type, we selected 10 hospitals with the highest total episode payments and ranked these hospitals according to component spending (index hospitalization, readmission, profession, and post-acute care) relative to the remaining hospitals in the dataset. With this step, we could determine the components that contributed the most to an individual hospital’s high spending status based on whether that particular component was ranked in the top quarter among all hospitals for the procedure of interest. Additionally, the component with the greatest opportunity for spending reduction at an individual hospital was determined as the component with the greatest deviation (via standardized z-score) from the average component payment across all hospitals for the procedure of interest.
Sensitivity Analysis Due to the rising prevalence and importance of nephron-sparing approaches in the treatment of kidney cancer, we divided our nephrectomy cohort into partial nephrectomy and radical nephrectomy groups based on corresponding ICD-9 and CPT codes. We then examined differences in mean total episode payments and sources of payment variation between the partial and radical nephrectomy groups. Second, we stratified episode payments by quartiles of hospital surgical volume to determine the effect of volume on payment variation for urological cancer surgery.
8 Page 8 of 27
We then used the ANOVA test of significance to determine whether total episode and component spending varied among the four volume quartiles in our cohort.
All analyses were performed using SAS (Cary, NC) and STATA 14/SE (College Station, TX), and at the 5% significance level. This study was deemed exempt from review by the Institutional Research Board at our institution.
9 Page 9 of 27
Results We identified 14,923 patients who underwent surgery for urological cancers from 2008 through 2011. This cohort included 1,849 patients who underwent cystectomy, 8,770 patients with prostatectomy, and 4,304 patients with nephrectomy. Mean age for patients undergoing cystectomy was 76.3 years old (range 66-94), 70.5 years old (range 65-87) for prostatectomy, and 75.0 years old (range 66-99) for nephrectomy. Total episode payments varied widely across hospitals for all three surgical procedures. Figure 1 presents the range of values for mean total episode payments among hospitals for cystectomy, prostatectomy, and nephrectomy. We observed wide variation in mean episode payments for all three conditions (cystectomy mean $35,102: range $24,112 to $57,238, prostatectomy mean $10,803: range $8,816 to $17,877, nephrectomy mean $17,475: range $11,681 to $26,711). Considerable differences were noted between hospitals at the highest and lowest episode payment quartiles for cystectomy, prostatectomy, and nephrectomy (Table 1). The greatest difference in total episode payments between high-spending and low-spending hospitals was noted for cystectomy among all procedures studied, with a $13,057 difference in total episode expenditures between the highest and lowest payment quartiles. Using Spearman’s rank correlation to examine the relationship between total and component episode payments for each cancer (Table 2), we observed that high payments for index hospitalization (R=0.79, p<0.001) and post-acute care (R=0.70, p<0.001) had the strongest correlation with high total episode payments for cystectomy. Meanwhile, total episode payments were strongly correlated positively with professional
10 Page 10 of 27
payments (R=0.70, p<0.001) for prostatectomy and index hospitalization (R=0.69, p<0.001) and post-acute care (R=0.71, p<0.001) payments for nephrectomy. Finally, we observed considerable differences in patterns of component payments among individual high-spending hospitals (Figure 2). Among the 10 highest spending hospitals for cystectomy, the component payment that drove differences in total episode payments varied among individual hospitals. Specifically, each of the top hospitals for 90-day episode spending varied by the types of component payments within the top quarter for that particular category (e.g. index hospitalization, readmissions, etc). In addition, the greatest opportunity for spending reduction (i.e. greatest deviation from the mean component payment for the entire cohort) varied for cystectomy among the most expensive hospitals. For instance, among the 10 most expensive hospitals for cystectomy in Figure 2A, the greatest opportunity for spending reduction was readmission payments for some hospitals (Hospitals A, B, and E) and post-acute care payments for other hospitals (Hospitals D, G, H, and J). Similar levels of inter-hospital variability in component payment patterns were observed for the most expensive hospitals for nephrectomy and prostatectomy (Figures 2B-C).
Sensitivity Analysis In our sensitivity analysis, we observed that average 90-day episode payments to hospitals were $18,052 (range: $12,810 to $26,711) for radical nephrectomies and $15,910 (range: $10,494 to $25,091) for partial nephrectomies (Supplementary Table 1). Variation in index hospitalization payments and post-acute care payments represented the major sources of variation between hospitals at the highest and lowest
11 Page 11 of 27
payment quartiles for both partial and radical nephrectomies—similar to trends we noted in the overall nephrectomy cohort (Supplementary Table 2). Stratification of total episode payments by quartiles of hospital urological surgical volume showed no significant differences in total episode payments among the hospital volume quartiles (p=0.09). Significant association was observed only for post-acute care payments with hospital volume (p=0.02). No other associations between component payments and hospital volume were observed otherwise (Supplementary Table 3).
Comment Across a nationally representative sample of hospitals, we identified wide variation in 90-day episode payments for cystectomy, prostatectomy, and nephrectomy. In addition, when we compared spending patterns between the most and least expensive hospitals, we observed that the drivers of payment variation differed by surgical procedure. Finally, we found that even among hospitals considered the “most expensive,” the sources of variation were not uniform. Collectively, these findings suggest that while there is sufficient Medicare payment variation for urological cancer surgery to be a potential candidate for bundled payment programs, hospitals will need to evaluate their own data and design organization and procedure-specific strategies to achieve success under such programs. Our finding that total episode payment variation exists for urological cancer surgery is consistent with current literature for other surgical conditions1,5,16,20,21,22. Specifically, previous studies have demonstrated wide variation in episode-based payments for common surgical procedures such as CABG, colectomy, joint 12 Page 12 of 27
replacement, and other oncological surgical procedures. However, the sources of payment variation identified in our analysis are not uniformly consistent with the existing literature. For example, in one study index hospitalization payments were the major source of variation between high and low spending hospitals for cystectomy and prostatectomy22. In contrast, another study on episode spending for prostatectomy suggested that variation in total episode payments is distributed more evenly among multiple components—a result more consistent with our own findings17. Our observation that sources of payment variation are highly variable among individual hospital may explain some of the discrepancies in the literature regarding general trends among specific hospital-level cohorts. Our study has several limitations. First, by utilizing SEER-Medicare data, we restrict our study to Medicare beneficiaries greater than 65 years of age. While we would expect the average age of surgery for all patients undergoing cystectomy to be greater than 65 years old, the average age may be lower for patients undergoing prostatectomy or nephrectomy. However, we deemed Medicare beneficiaries as the most appropriate population given the aggressive efforts by the CMS to develop and implement alternative payment models. Second, our study was limited to patients who had surgery at hospitals within the SEER regions. While this may limit the generalizability of our results beyond these regions, the SEER regions are considered nationally representative of the general oncologic population in the United States. Thirdly, we only examined hospital-level variation in our analysis. While recent literature has shown that the majority of payment variation occurs at the hospital-level, physicianlevel spending variation still plays an important role in current bundled payment
13 Page 13 of 27
initiatives21,25. In addition, with the increasing prevalence of co-managed care between surgeon and internal medicine specialties, the topic of physician-level variation will be the subject of future investigation26. Finally, we did not examine the impact of case-mix (e.g., comorbidity, socioeconomic status, tumor staging) on episode spending. Nonetheless, at present, Medicare does not include risk-adjustment in their current, bundled payment models23. Additionally, we believe that the use of risk-adjustment for episode payment is a highly complex topic that, while beyond the scope of this study, will be a necessary focus of future work. These limitations notwithstanding, our findings will help many stakeholders anticipate the potential role and impact of bundled payments for urological cancer surgery. Relevant to payers and policymakers, our finding that fairly large episode payment variation exists for the three most common urological cancer surgeries implies that these conditions may be candidates for bundled payment programs. In fact, the American Urological Association is in the process of developing bundled payment programs for these conditions27. For hospital administrators, we show that health systems will need to compile institution-specific data and clinical re-design strategies to succeed with implementation of episode-based payment models for urological cancer surgery. For instance, the optimal strategy for one hospital may be to reduce readmissions for cystectomy patients, while a different hospital may find post-acute care spending to be its major source of high expenditures. Much like the preceding example, the hospitals within our study cohort demonstrate different characteristics in payments that may require further investigation of the administrative and clinical factors driving high spending in a
14 Page 14 of 27
particular component. As a result, we would anticipate different health systems to execute varying methods of quality improvement and clinical redesign tailored to the areas of disproportionate spending within that specific institution.
15 Page 15 of 27
Conclusions Our collective findings suggest that radical cystectomy, prostatectomy and nephrectomy may be suitable candidates for bundled payment programs due to wide episode payment variation. Additionally, hospitals will need to design specific strategies based on their individual spending patterns to succeed in these types of payment programs for urological cancer surgery. Moving forward, research in this area will also need to examine the influence of patient characteristics such as age, comorbidities, cancer stage and socioeconomic status on 90-day episode spending. In the end, the opportunity for policymakers to reduce urological cancer surgery spending will depend on the magnitude of unwarranted episode payment variation across hospitals. Moreover, the opportunity for hospitals to succeed in these programs will depend on the ability to identify and address local sources of excessive spending, while at the same time maintaining high-quality outcomes and patient experience.
16 Page 16 of 27
References 1.
National Center for Health Statistics. Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities. Hyattsville, MD.; 2016: 300-303. https://www.cdc.gov/nchs/data/hus/hus15.pdf#096.
2.
Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare payments for common inpatient procedures: implications for episode-based payment bundling. Health Serv Res. 2010;45(6 Pt 1):1783-1795. doi:10.1111/j.1475-6773.2010.01150.x.
3.
Iglehart JK. Bundled payment for ESRD--including ESAs in Medicare’s dialysis package. N Engl J Med. 2011;364(7):593-595. doi:10.1056/NEJMp1014187.
4.
Birkmeyer JD, Gust C, Baser O, Dimick JB, Sutherland JM, Skinner JS. Medicare Payments for Common Inpatient Procedures: Implications for Episode-Based Payment Bundling. Health Serv Res. 2010;45(6p1):1783-1795. doi:c.
5.
Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. Health Aff (Millwood). 2011;30(11):2107-2115. doi:10.1377/hlthaff.2011.0783.
6.
Abdelsattar ZM, Birkmeyer JD, Wong SL. Variation in Medicare Payments for Colorectal Cancer Surgery. J Oncol Pract. 2015;11(5):391-395. doi:10.1200/JOP.2015.004036.
7.
Iorio R, Clair AJ, Inneh IA, Slover JD, Bosco JA, Zuckerman JD. Early Results of Medicare’s Bundled Payment Initiative for a 90-Day Total Joint Arthroplasty Episode of Care. J Arthroplasty. 2016;31(2):343-350. doi:10.1016/j.arth.2015.09.004.
8.
Jacobs BL, Zhang Y, Tan H-J, Ye Z, Skolarus TA, Hollenbeck BK. Hospitalization Trends After Prostate and Bladder Surgery: Implications of Potential Payment Reforms. J Urol. 2013;189(1):59-65. doi:10.1016/j.juro.2012.08.182.
9.
Skolarus TA, Jacobs BL, Schroeck FR, et al. Understanding Hospital Readmission Intensity after Radical Cystectomy. J Urol. 2015;193(5):1500-1506. doi:10.1016/j.juro.2014.10.107.
10.
Minnillo BJ, Maurice MJ, Schiltz N, et al. Few modifiable factors predict readmission following radical cystectomy. Can Urol Assoc J. 2015;9(7-8):E43946. doi:10.5489/cuaj.2793.
11.
Aghazadeh MA, Barocas DA, Salem S, et al. Determining factors for hospital discharge status after radical cystectomy in a large contemporary cohort. J Urol. 2011;185(1):85-89. doi:10.1016/j.juro.2010.08.016.
12.
Liu J-J, Maxwell BG, Panousis P, Chung BI. Perioperative outcomes for laparoscopic and robotic compared with open prostatectomy using the National Surgical Quality Improvement Program (NSQIP) database. Urology. 17 Page 17 of 27
2013;82(3):579-583. doi:10.1016/j.urology.2013.03.080. 13.
Trinh Q-D, Sammon J, Sun M, et al. Perioperative outcomes of robot-assisted radical prostatectomy compared with open radical prostatectomy: results from the nationwide inpatient sample. Eur Urol. 2012;61(4):679-685. doi:10.1016/j.eururo.2011.12.027.
14.
Moschini M, Gandaglia G, Dell’Oglio P, et al. Incidence and Predictors of 30-Day Readmission in Patients Treated With Radical Cystectomy: A Single Center European Experience. Clin Genitourin Cancer. 2016;14(4):e341-6. doi:10.1016/j.clgc.2015.12.017.
15.
Autorino R, Zargar H, Butler S, Laydner H, Kaouk JH. Incidence and risk factors for 30-day readmission in patients undergoing nephrectomy procedures: a contemporary analysis of 5276 cases from the National Surgical Quality Improvement Program database. Urology. 2015;85(4):843-849. doi:10.1016/j.urology.2014.11.044.
16.
Schmid M, Chiang HA, Sood A, et al. Causes of hospital readmissions after urologic cancer surgery. Urol Oncol. 2016;34(5):236.e1-236.e11. doi:10.1016/j.urolonc.2015.11.019.
17.
Herrel LA, Syrjamaki JD, Linsell SM, Miller DC, Dupree JM. Identifying Drivers of Episode Cost Variation With Radical Prostatectomy. Urology. 2016;97:105-110. doi:10.1016/j.urology.2016.05.071.
18.
Pate SC, Uhlman MA, Rosenthal JA, Cram P, Erickson BA. Variations in the open market costs for prostate cancer surgery: a survey of US hospitals. Urology. 2014;83(3):626-630. doi:10.1016/j.urology.2013.09.066.
19.
Wang S-Y, Wang R, Yu JB, et al. Understanding regional variation in Medicare expenditures for initial episodes of prostate cancer care. Med Care. 2014;52(8):680-687. doi:10.1097/MLR.0000000000000158.
20.
Overview of the SEER Program. National Cancer Institute: Surveillance, Epidemiology, and End Results (SEER) Program. https://seer.cancer.gov/about/overview.html. Accessed February 8, 2017.
21.
Centers for Medicare & Medicaid Services (CMS) H. Medicare Program; Advancing Care Coordination Through Episode Payment Models (EPMs); Cardiac Rehabilitation Incentive Payment Model; and Changes to the Comprehensive Care for Joint Replacement Model (CJR). Final rule. Fed Regist. 2017;82(1):180651. http://www.ncbi.nlm.nih.gov/pubmed/28071874.
22.
Nathan H, Atoria CL, Bach PB, Elkin EB. Hospital volume, complications, and cost of cancer surgery in the elderly. J Clin Oncol. 2015;33(1):107-114. doi:10.1200/JCO.2014.57.7155.
23.
Ellimoottil C, Ryan AM, Hou H, Dupree J, Hallstrom B, Miller DC. Medicares New Bundled Payment For Joint Replacement May Penalize Hospitals That Treat Medically Complex Patients. Health Aff. 2016;35(9):1651-1657. 18 Page 18 of 27
doi:10.1377/hlthaff.2016.0263. 24.
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.
25.
Ejaz A, Gani F, Kim Y, Pawlik TM. Variation in inpatient hospital and physician payments among patients undergoing general versus orthopedic operations. Surgery. 2016;160(6):1657-1665. doi:10.1016/j.surg.2016.07.013.
26.
Sharma G, Kuo Y-F, Freeman J, Zhang DD, Goodwin JS. Comanagement of hospitalized surgical patients by medicine physicians in the United States. Arch Intern Med. 2010;170(4):363-368. doi:10.1001/archinternmed.2009.553.
27.
Dupree JM. Alternative Payment Models in Urology. AUA News. February 2017:23.
Figure 1. Mean Total Episode Payments by Hospital for (A) Cystectomy, (B) Prostatectomy and (C) Nephrectomy, 2008-2011 SEER-Medicare. Individual marker denotes single hospital with caseload ≥ 10 for procedure of interest.
Figure 2. Episode-based Payment Patterns among the Highest Spending Hospitals for (A) Cystectomy, (B) Prostatectomy and (C) Nephrectomy, 2008-2011 SEER-Medicare. Shaded cells indicate that the hospital is in the top quartile for the particular spending component (e.g. professional, readmissions). Star indicates greatest opportunity for spending reduction for the hospital of interest. Greatest opportunity for spending reduction for an individual hospital was determined as the component with the greatest deviation from the average component payment across all hospitals for the procedure of interest.
19 Page 19 of 27
Table 1. Comparison of Component Payments between Hospitals at the Highest and Lowest Episode Payment Quartiles, 2008-2011 SEER Medicare
Procedure Cystectomy
Prostatectomy
Nephrectomy
Payment Componenta Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment Index Hospitalization Readmission Professional Post-acute Care Total Episode Payment
Low Payment Quartile 16,252 2,355 5,328 4,185 28,121 6,818 134 2,523 206 9,680 9,342 589 2,906 1,222 14,059
High Payment Quartile 22,441 3,447 7,079 8,211 41,177 7,207 672 3,512 797 12,187 11,485 1,907 3,980 3,772 21,144
Difference High and Low Quartile 6,189 1,092 1,750 4,025 13,057 388 538 989 591 2,507 2,143 1,318 1,074 2,550 7,085
% Total Difference Attributed to Component 47.4 8.4 13.4 30.8 100.0 15.5 21.5 39.5 23.6 100.0 30.2 18.6 15.2 36.0 100.0
a
All amounts in US Dollars
20 Page 20 of 27
Table 2. Spearman Rank Correlation between Total Episode Payment and Individual Component Payments, 20082011 SEER Medicare Spearman Correlation Coefficient of Total Episode vs Component, R (p-value) Procedure
Index
Professional
Readmission
Post-acute Care
Cystectomy
0.79 (<0.001)
0.66 (<0.001)
0.35 (0.029)
0.70 (<0.001)
Prostatectomy
0.51 (<0.001)
0.70 (<0.001)
0.49 (<0.001)
0.61 (<0.001)
Nephrectomy
0.69 (<0.001)
0.57 (<0.001)
0.56 (<0.001)
0.71 (<0.001)
21 Page 21 of 27
Figure 1.
22 Page 22 of 27
Figure 2. (A) Payment Patterns for Highest Spending Hospitals for Cystectomy Hospital
Index Hospitalization
Professional
B
D
E F
Post-acute Care
A C
Readmissions
G H
I
J
(B) Payment Patterns for Highest Spending Hospitals for Prostatectomy Hospital
Index Hospitalization
Professional
Readmissions
A
B
C D
E F
G H I J
Post-acute Care
23 Page 23 of 27
(C) Payment Patterns for Highest Spending Hospitals for Nephrectomy Hospital
Index Hospitalization
Professional
C D
E F G
H I J
Post-acute Care
A B
Readmissions
24 Page 24 of 27
Editorial Comment
Christopher P. Filson, MD, MS Department of Urology, Emory University School of Medicine
Contact Information: Department of Urology Emory University School of Medicine 1365 Clifton Rd NE, Suite B1400 Atlanta, GA, 30322 404-778-4583 [email protected]
Amidst ever-rising health care expenditures in the United States, stakeholders and policymakers have turned to alternative payment models as mechanisms to control costs. These tools also aim to maintain—and possibly improve—the quality of health care delivered in this country. Bundled payment programs are an alternative payment model that have shown promise in curtailing costs and maintaining quality among patients undergoing lower extremity joint (e.g., hip and knee) replacement.1 As Dr. Ellimoottil and co-authors demonstrate in the above article, the wide two-fold variation in episode costs for urological cancer operations
25 Page 25 of 27
makes consideration of bundled payment programs in these instances quite appealing at first glance.
However, the most compelling finding from this work was the observation that the costliest hospitals varied considerably in which components contributed most to episode-based spending, suggesting that a one-size-fits-all approach for hospital-based cost containment would likely fail for urologic cancers. Instead, this work highlights the importance for medical centers to actively measure and understand their practice patterns and delivery of care throughout the entire episode surrounding urological cancer operations. From there, hospitals can decide on interventions that suit them best. For example, surgeon intraoperative cost scorecards2 and pathways to shorten hospitalization3 have both been shown to decrease episode payments. Whether these would be applicable for a particular hospital would depend on whether intraoperative equipment use or prolonged hospitalizations were cost drivers, respectively.
The authors also highlight an important limitation of their work, in that this analysis did not consider cost variation attributed to physician-level practice patterns. As just one example, the delivery of low-value peri- and postoperative care based on dogma—in the absence of evidence of benefit—may quickly rack up costs. A recent analysis of the Virginia All-Payer Claims database highlighted that the majority of low-value care was not related to expensive “big-ticket” services, but instead high-volume, low-cost services.4 In fact, the costliest low-value service overall (at $227.8 million in 2014) was receipt of preoperative labs for low-risk patients undergoing low-risk surgery. The findings from that study are worth reflecting on. When ordering an X-ray to check stent position following a radical cystectomy, or checking fluid creatinine
26 Page 26 of 27
levels on a low-output drain following prostatectomy, are we delivering a high-value service, or reassuring ourselves (as providers) that everything is stable? In the big picture, are the costs associated with those studies worth the low likelihood of being abnormal? I don’t have the answer to these questions, and medico-legal concerns also exist, but bundled payment programs may provide the impetus for hospitals to delve into these issues and “trim the fat”, so to speak.
References 1. Dummit LA, Kahvecioglu D, Marrufo G, et al. Association between hospital participation in a Medicare bundled payment intitiative and payments and quality outcomes for lower extremity joint replacement episodes. JAMA 2016; 316(12):1267–1278. 2. Zygourakis CC, Valencia V, Moriates C, et al. Association between surgeon scorecard use and operating room costs. JAMA Surgery 2017; 152(3): 284 – 291. 3. Regenbogen SE, Cain-Nielsen AH, Norton EC, et al. Costs and consequences of early hospital discharge after major inpatient surgery in older adults. JAMA Surgery 2017; 152(5) [epub ahead of print] 4. Mafi JN, Russell K, Bortz BA, et al. Low-cost, high-volume health services contribute the most to unnecessary health spending. Health Aff 2017; 36(10): 1701 – 1704.
27 Page 27 of 27