Advertisement

Reengineering valve patients' postdischarge management for adapting to bundled payment models

Open ArchivePublished:March 10, 2017DOI:https://doi.org/10.1016/j.jtcvs.2016.10.109

      Abstract

      Background

      Bundled Payments for Care Improvement (BPCI) initiatives were developed by Medicare in an effort to reduce expenditures while preserving quality of care. Payment model 2 reimburses based on a target price for 90-day episode of care postprocedure. The challenge for valve patients is the historically high (>35%) 90-day readmission rate. We analyzed our institutional cardiac surgical service line adaptation to this initiative.

      Methods

      On May 1, 2015, we instituted a readmission reduction initiative (RRI) that included presurgical risk stratification, comprehensive predischarge planning, and standardized postdischarge management led by cardiac nurse practitioners (CNPs) who attempt to guide any postdischarge encounters (PDEs). A prospective database also was developed, accruing data on all cardiac surgery patients discharged after RRI initiation. We analyzed detailed PDEs for all valve patients with complete 30-day follow-up through November 2015.

      Results

      Patients included 219 surgical patients and 126 transcatheter patients. Sixty-four patients had 79 PDEs. Of these 79 PDEs, 46 (58.2%) were guided by CNPs. PDEs were due to fluid overload/effusion (21, 27%), arrhythmia (17, 22%), bleeding/thromboembolic events (13, 16%), and falls/somatic complaints (12, 15%). Thirty-day readmission rate was 10.1% (35/345). Patients with transcatheter aortic valve replacement had a higher rate of readmission than surgical patients (15.0% vs 6.9%), but were older with more comorbidities. The median readmission length of stay was 2.0 days (interquartile range 1.0–5.0 days). Compared with 2014, the 30-day readmission rate for BPCI decreased from 18% (44/248) to 11% (20/175), P = .05.

      Conclusions

      Our reengineering of pre/postdischarge management of BPCI valve patients under tight CNP control has significantly reduced costly 30-day readmissions in this high-risk population.

      Key Words

      Abbreviations and Acronyms:

      ACA (Affordable Care Act), BPCI (Bundled Payments for Care Improvement), CMMI (Center for Medicare and Medicaid Innovation), CMS (Centers for Medicare and Medicaid Services), CNP (cardiac nurse practitioner), CSSL (Cardiac Surgical Service Line), EOC (episode of care), HRRP (Hospital Readmissions Reduction Program), PDE (postdischarge encounter), RRI (readmission reduction initiative), SAVR (surgical aortic valve replacement), SNF (skilled nursing facilities), STS (Society of Thoracic Surgery), TAVR (transcatheter aortic valve replacement), TCV (transcatheter valve), TVT (transcatheter valve therapy)
      Figure thumbnail fx1
      Logo for the CMS Innovation Center created by the Affordable Care Act.
      Surgeons must take ownership of patient management in the postdischarge period to survive in the era of financial risk bearing.
      The Affordable Care Act created an innovation center to explore solutions for health care cost reduction. What has been perceived as a postprocedural failure (high readmission rates after valves) has become a target of this reform. This has placed pressure on the cardiac surgeon to optimize postdischarge management with an institutional approach to avoid unnecessary and costly hospital readmissions.
      See Editorial Commentary page 199.
      The Centers for Medicare and Medicaid Services (CMS) has historically based reimbursement on a fee-for-service mechanism. However, this approach has been criticized as a major driver of health care costs as it contains unintended incentives for increasing the volume of care while not necessarily improving quality.
      • Schroeder S.A.
      • Frist W.
      National Commission on Physician Payment R. Phasing out fee-for-service payment.
      In an effort to respond to these concerns, the Center for Medicare and Medicaid Innovation (CMMI) was created in 2010 by the Affordable Care Act (ACA) and instituted the Bundled Payments for Care Improvement (BPCI) initiative.
      The BPCI program was developed to test payment and service delivery models with the potential to reduce expenditures during clinical episodes while preserving or enhancing the quality of care.

      Centers for Medicare and Medicaid Services. Bundled payments for care improvement (BPCI) initiative. Available at: https://innovation.cms.gov/initiatives/bundled-payments/. Accessed April 1, 2016.

      American Hospital Association. Medicare's bundled payment initiatives: considerations for providers. Available at: http://www.aha.org/content/16/issbrief-bundledpmt.pdf. Accessed April 1, 2016.

      Four different models are currently being tested that link payments for the multiple services rendered during a clinical care episode. An episode of care (EOC) is designated based on disease process or procedure as defined by disease-related group codes. BPCI model 2 EOCs include acute care and physician fees for the index procedure as well as related post-acute care, including readmissions and all related services through 90 days postdischarge. Model 2 includes a retrospective bundled payment arrangement in which actual expenditures are reconciled against a baseline target price for the EOC. The target price is set by CMS, individualized for an institution based on 3-year historic claims data, EOC category, episode length, and a 2% discount. For a “bundle” patient, the actual 90-day expenditures from the CMS master claims file are tabulated and reconciled. If total expenditures are below the target amount, the difference is paid to the institution; if above, recoupment is accordingly collected by CMS. Readmissions are therefore a costly penalty to the institution. Proponents of this model believe it will incentivize the more judicial use of health care resources and lead to less fragmentation of care.
      • Schroeder S.A.
      • Frist W.
      National Commission on Physician Payment R. Phasing out fee-for-service payment.
      • Hussey P.S.
      • Eibner C.
      • Ridgely M.S.
      • McGlynn E.A.
      Controlling U.S. health care spending—separating promising from unpromising approaches.
      • Mechanic R.E.
      • Altman S.H.
      Payment reform options: episode payment is a good place to start.
      One of the challenges in caring for Medicare valve patients is the historically high (>40%) 90-day readmission rates present in this elderly, high-risk population.
      • Fry D.E.
      • Pine M.
      • Nedza S.M.
      • Locke D.G.
      • Reband A.M.
      • Pine G.
      Inpatient and 90-day postdischarge outcomes in cardiac surgery.
      It has been shown that between 50% and 80% of cardiac surgery readmissions occur in the first 30 days postdischarge.
      • Iribarne A.
      • Chang H.
      • Alexander J.H.
      • Gillinov A.M.
      • Moquete E.
      • Puskas J.D.
      • et al.
      Readmissions after cardiac surgery: experience of the National Institutes of Health/Canadian Institutes of Health research cardiothoracic surgical trials network.
      From a fiscal perspective, Medicare payments for unplanned rehospitalizations in 2004 accounted for approximately $17 billion of the $100 billion that Medicare awarded that year.
      • Jencks S.F.
      • Williams M.V.
      • Coleman E.A.
      Rehospitalizations among patients in the Medicare fee-for-service program.
      Previously, extended postdischarge management and concomitant fiscal burden of late complications was typically beyond the surgeon's purview. Now that this expenditure has become a target for CMS penalty, this perspective has changed. We analyzed our institution's cardiac surgical service line (CSSL) adaptation to this BPCI model.

      Methods

       BPCI Details

      Our institution decided to participate for valve surgery in BPCI model 2 with a 90-day EOC. The phase 1 preparation period started in January 2013, with the financial risk-bearing phase starting in October 2013. Historically, before and during the preparation period, our discharge strategy had high reliance (approximately 70%) on inpatient rehabilitation and skilled nursing facilities (SNFs), which was driven by length-of-stay concerns. Our initial approach to reduce postdischarge inpatient expenditures was to move away from facility-based post-acute care.
      By the end of 2013, use of inpatient rehabilitation and SNFs was down to 21%, resulting in beneficiary savings of more than $7000 per patient.
      • Jubelt L.E.
      • Goldfeld K.S.
      • Chung W.
      • Blecker S.B.
      • Horwitz L.I.
      Changes in discharge location and readmission rates under Medicare bundled payment.
      However, 30-day readmission remained at approximately 20%.

       Implementation of Readmission Reduction Initiative

      In 2015, our cardiac service line team completed development of a comprehensive discharge planning and management approach. It begins with presurgical risk stratification based on a variety of heart failure, frailty, and cognitive measures (eg, Society of Thoracic Surgery [STS] risk score, Kansas City Cardiomyopathy Questionnaire, walk tests). Predischarge planning is then begun early in the postoperative course with patient and family education (emphasizing the importance of medication compliance), as well as coordination with home-care services or acute and subacute facilities. Standardized postdischarge management led by cardiac nurse practitioners (CNPs) begins with a phone call within 48 hours of discharge. CNPs then maintain weekly contact by phone and use telemedicine (Cardiocom, Chanhassen, Minn) to monitor the patient's clinical status and attempt to guide any postdischarge encounters (PDEs) throughout the first 30 days after discharge. Postdischarge office visits also have been regimented, with patients being seen on postdischarge days 7 and 21 and scheduled to see their cardiologist on day 14. We also implemented a number of postdischarge triage and management protocols for patients presenting to our emergency department, which facilitates immediate evaluation by our team and standardizes the approach to common presentations, such as postoperative pain, atrial fibrillation, and volume overload. Whenever possible, less than 48-hour hospital observation was used instead of readmission. Additionally, a prospective database was created in May 2015 to track all adult cardiac surgery patients for 30 days postdischarge.

       Outcomes

      A retrospective analysis of our prospectively collected adult cardiac surgery discharge database was conducted. Additional data were collected from CMS claims as well as STS and transcatheter valve therapy (TVT) data. Local institutional review board approval for retrospective de-identified data analysis was obtained, and the requirement for written informed consent was waived.
      PDE included any unplanned hospital encounter other than office visit, regardless of hospital setting; planned staged procedures were excluded. A patient was considered an inpatient readmission if the patient stayed more than 48 hours in-house, as per CMS definition. For less than 48 hours, the patient was classified as observational or outpatient status.

       Statistics

      Statistical analyses were performed using SPSS 22 (IBM Corp, Armonk, NY). Encounter rate and readmission rate were calculated per index discharge. Continuous variables are reported as mean ± standard deviation and skewed data as median and interquartile range (IQR), unless otherwise stated. Categorical variables are reported as a rate of occurrence. Groups were compared by χ2 methods, Fisher exact tests, and analysis of variance. A P value less than .05 was considered statistically significant. Where appropriate, post hoc analysis was performed for contingency tables, and P values between groups were calculated by using the partitioning method and a Bonferroni correction was applied.

      Results

       Patients

      From May through November 2015, 345 valve patients were discharged; 219 open surgical patients (Open) and 126 transcatheter valve patients (TCV). The TCV patients were predominantly transcatheter aortic valve replacements (TAVRs) but also included 3 transcatheter mitral valve replacements. Of the Open valves, there were 89 isolated aortic valve replacements, 72 mitral valve repair or replacements, 25 multivalve cases, 17 combined valve/coronary bypasses, and 16 others.
      The average age was 71 ± 14 years and 63.8% were men (220/345). STS risk score ranged from 0.2% to 28.0% with a median of 4.2% (IQR 1.0%-6.3%) and 51.9% of patients were in New York Heart Association heart failure class III or IV. Demographic data are summarized in Table 1. TCV patients were older (82.0 ± 8.6 years vs 65.0 ± 12.9 years, P < .001) and had a higher STS risk score of 6.6% (IQR 5.2-8.8), P < .001. Additionally, TCVs had worse heart failure, twice the incidence of several major comorbidities, and were frailer (Table 1). The median length of stay postprocedure was 2 days (IQR 1-2 days) for TCV patients and 5 days (IQR 4-7 days) for Open patients (P < .001). Discharge home with health services occurred in 80.3% (377/345). There was a significant difference in discharge patterns with more Open patients (9.6%) requiring inpatient rehabilitation than TCV patients (2.4%) (Table 2).
      Table 1Demographics
      CharacteristicAll valve patients, n = 345, n (%)Transcatheter, n = 126, n (%)Open surgical, n = 219, n (%)P
      Age, y71.3 ± 14.282.3 ± 8.665.0 ± 12.9<.001
      Male gender220 (63.8)81 (64.3)139 (63.5).913
      White race298 (86.4)117 (92.9)181 (82.6).008
      BMI, kg/m227.5 ± 6.227.2 ± 5.527.8 ± 6.6.365
      STS risk score, %<.001
       Median (IQR)3.0 (1.0-6.25)6.6 (5.2-8.8)1.4 (0.7-2.8)
      Ejection fraction, %57.5 ± 12.655.8 ± 14.358.5 ± 11.4.080
      NYHA III-IV179 (51.9)108 (85.7)71 (32.4)<.001
      Katz index5.7 ± 0.95.4 ± 1.25.9 ± 0.7<.001
      Lives alone95 (27.5)39 (31.0)56 (25.6).256
      Two falls <6 mo10 (2.9)8 (6.3)2 (0.9).006
      Prior sternotomy66 (19.1)41 (32.5)25 (11.4)<.001
      Atrial fibrillation90 (26.1)36 (28.6)54 (24.7).446
      Diabetes86 (24.9)40 (31.7)46 (21.0).026
      COPD61 (17.7)30 (23.8)31 (14.2).024
      PVD38 (11.0)26 (20.6)12 (5.5)<.001
      CKD77 (22.3)44 (34.9)33 (15.1)<.001
      CVA29 (8.4)13 (10.3)16 (7.3).332
      Coronary artery disease157 (45.5)84 (66.7)73 (33.3)<.001
      BPCI participant175 (50.7)84 (66.7)91 (41.6)<.001
      Index postoperative LOS, d, median (IQR)4 (2-6)2 (1-2)5 (4-7)<.001
      30-d mortality3 (0.9)1 (0.8)2 (0.9)1.000
      BMI, Body-mass index; STS, Society of Thoracic Surgeons; IQR, interquartile range; NYHA, New York Heart Association; COPD, chronic obstructive pulmonary disease; PVD, peripheral vascular disease; CKD, chronic kidney disease; CVA, cerebral vascular accident; BPCI, Bundled Payments for Care Improvement; LOS, length of stay.
      Table 2Discharge disposition
      Discharge dispositionAll valve patients, n = 345, n (%)Transcatheter, n = 126, n (%)Open surgical, n = 219, n (%)P
      P value for contingency table listed first. Bonferroni correction applied with significant P < .013.
      .042
      Home32 (9.3)9 (7.1)23 (1.5).300
      Home with services277 (8.3)110 (87.3)167 (76.3).013
      Inpatient rehabilitation24 (7.0)3 (2.4)21 (9.6).011
      Skilled nursing facility12 (3.5)4 (3.2)8 (3.7).815
      P value for contingency table listed first. Bonferroni correction applied with significant P < .013.

       Postdischarge Encounters

      Sixty-four patients (18.6%) had 1 or more PDE. A total of 79 PDEs occurred for a rate of PDE per index discharge of 22.9%. Fourteen patients (4.1%) had more than 1 PDE, with 13 (3.8%) having 2 PDEs and 1 patient having 3 PDEs. The PDEs included 40 observational stays, 4 (nonoffice) outpatient visits, and 35 (44.2%) inpatient readmissions. The total readmission rate was 10.1% (35/345); 2 patients had 2 readmissions. TCV patients had a higher readmission rate of 15.9% (20/126) versus 6.8% (15/219) among Open patients, P = .035. Median length of stay for readmissions was 4 (IQR 3-7) days for transcatheter patients and 6 (IQR 3-6) days for open surgical patients. One readmitted patient died from a stroke after readmission to an outside hospital on postdischarge day 21.
      The median time to PDE was 11 days (IQR 6-19) and 35.4% (28/79) occurred during the first week after discharge. Of these first-week encounters, 35.7% resulted in readmission (10/28). Subsequently, PDEs and readmissions were distributed equally across postdischarge weeks 2 to 4 (Tables 3 and 4). The time to reencounter and readmission were not different between groups.
      Table 3Time to postdischarge encounter
      Time to postdischarge encounterAll valve patients, n = 79, n (%)Transcatheter, n = 38, n (%)Open surgical, n = 41, n (%)P
      Week 128 (35.4)13 (34.2)15 (36.6).247
      Week 218 (22.8)10 (26.3)8 (19.5)
      Week 317 (21.5)5 (13.2)12 (29.3)
      Week 416 (20.3)10 (26.3)6 (14.6)
      Median (interquartile range), d11 (6-19)10.5 (6-23)13 (6-17).697
      Table 4Time to readmission
      Time to readmissionAll valve patients, n = 35, n (%)Transcatheter, n = 20, n (%)Open surgical, n = 15, n (%)P
      Week 110 (35.7)6 (30.0)4 (26.7).389
      Week 26 (17.1)4 (20.0)2 (13.3)
      Week 39 (25.7)3 (15.0)6 (40.0)
      Week 410 (28.6)7 (35.0)3 (20.0)
      Median (interquartile range), d15 (7-23)14.5 (6.5-25.5)16 (7-20).826
      Indication for PDE is summarized in Figure 1. The most common reason for PDE was fluid overload or effusion (21/79, 26.6%) followed by arrhythmia (17/79, 21.5%), somatic pain (9/79, 11.4%), and thromboembolic events (8/79, 10.1%). Other indications for reencounter included gastrointestinal disease (6/79, 7.6%), infection (5/79, 6.3%), bleeding (5/79, 6.3%), falls (3/79, 3.8%), and other indications (5/79, 6.3%). Fluid overload encounters resulted in readmission 57.1% of the time (12/21), whereas PDE for arrhythmia resulted in readmission 29.4% of the time (5/17). PDEs were more likely to be associated with arrhythmia in Open patients, whereas PDEs were more associated with volume overload in TCV patients (Figure 2).
      Figure thumbnail gr1
      Figure 1Postdischarge encounter and readmission rates. PDE, Postdischarge encounter; GI, gastrointestinal.
      Figure thumbnail gr2
      Figure 2Distribution of PDE indication by procedure. PDE, Postdischarge encounter; GI, gastrointestinal.
      CNPs guided 58.2% (46/79) of PDEs. If the encounter was guided, the patient was more likely to present to our institution rather than an outside hospital, 74% versus 33% of nonguided encounters (P < .001). Compared with 2014, the 30-day readmission rate for BPCI decreased from 18% (44/248) to 11% (20/175), P = .05.

      Discussion

      The Hospital Readmissions Reduction Program (HRRP) was created with CMMI as part of the ACA in 2010.

      Centers for Medicare and Medicaid Services. Readmissions reduction program. Available at: http://www.cms.gov/Medicare/medicare-fee-for-service-payment/acuteinpatientPPS/readmissions-reduction-program.html. Accessed April 1, 2016.

      A primary motivator for this reform was the article by Jencks and colleagues
      • Jencks S.F.
      • Williams M.V.
      • Coleman E.A.
      Rehospitalizations among patients in the Medicare fee-for-service program.
      in the New England Journal of Medicine the year prior that estimated $17.4 billion was spent by Medicare on 30-day readmissions; this made 30-day readmission reduction an opportune reform target. To that end, HRRP has begun to implement financial penalties for institutions having a higher-than-average 30-day readmission rate for selected clinical conditions. Although the purview of these financial penalties has not yet been extended to include readmission after cardiac surgery, this is widely expected to be forthcoming. Due to this concern, many institutions, including our own, have implemented readmission reduction strategies in patients after cardiac surgery.
      Historically, cardiac surgeons have left much of the postdischarge management in the hands of cardiologists. As at many tertiary centers relying on a large referral base, many private physicians following our patients are not part of our network. This typically leaves little control over patient management choices and limits the ability to accurately track outcomes. Our early reduction of postdischarge SNF utilization avoided the scenario in which these facilities would just transfer back to a local emergency room when any complication occurred, thus preventing the opportunity for our outpatient management. Therefore, our readmission reduction initiative (RRI) program was created with the focus of maintaining ownership of our patients after discharge. By creating a regimented postdischarge process by using frequent phone calls and multiple office visits starting within the first week of discharge, we demonstrated success at early identification of many postoperative issues and treating them efficiently without readmission. With close follow-up and a 24/7 on-call hotline, we also have had success in guiding patients in need of care escalation to the hospital, specifically to NYU. Directing patients to our institution is not only safer in terms of continuity of care, but is also likely to be much more cost-effective by obviating the need for redundant diagnostic tests and workup.
      Overall, we found a 30-day PDE rate of 22.9% with a readmission rate of 10.1%. The median time to PDE and readmission was 11 and 15 days, respectively, with approximately one-third of all PDEs occurring in the first week. One concern was that PDEs could be affected by shorter length-of-index admission,
      • Zuckerman R.B.
      • Sheingold S.H.
      • Orav E.J.
      • Ruhter J.
      • Epstein A.M.
      Readmissions, observation, and the Hospital Readmissions Reduction Program.
      but this is not obviously true from our experience. Readmission rate for the TCV patients was higher than for Open cases, 15.9% and 6.8%, respectively; however, the transcatheter patients represent an older and sicker population. Although the transcatheter patients were crudely differentiated in this initial analysis, in-depth study of this cohort is pending and may afford insight as to which patient-specific features are associated with readmission. We identified 2 major sources for PDE and readmission, fluid overload/heart failure and arrhythmic issues, that together account for almost half of all PDEs and readmissions, with arrhythmia more frequent among Open patients. Hopefully, as we accrue more data, we will gain insight on how to predict these complications.
      Similar to some of our findings, a recent study by Hannan and colleagues
      • Hannan E.L.
      • Samadashvili Z.
      • Jordan D.
      • Sundt T.M.
      • Stamato N.J.
      • Lahey S.J.
      • et al.
      Thirty-day readmissions after transcatheter aortic valve implantation versus surgical aortic valve replacement in patients with severe aortic stenosis in New York State.
      analyzed 30-day readmission for TAVR/SAVR by using New York's administrative acute-care database. With multiple patient exclusions, emergency room visit exclusion, and propensity matching, 30-day readmissions between SVAR and TAVR groups were similar at 18.8% and 19.3%; the most common reasons for readmission were heart failure and arrhythmia. Iribarne and colleagues
      • Iribarne A.
      • Chang H.
      • Alexander J.H.
      • Gillinov A.M.
      • Moquete E.
      • Puskas J.D.
      • et al.
      Readmissions after cardiac surgery: experience of the National Institutes of Health/Canadian Institutes of Health research cardiothoracic surgical trials network.
      recently published readmission data from the Cardiothoracic Surgery Trials Network of more than 5000 open surgical patients followed for 65 days postdischarge. Readmission rate was 18.7% overall, with 80% occurring in the first 30 days. Rates for valve surgery and valve/coronary artery bypass grafting were 18.3% and 24.0%. For this study, they did not distinguish between observation and inpatient status, excluded emergency room visits, and included patients with ventricular assist devices, who had the highest readmission rate (>35%).
      One of the major challenges of bundled payment models (particularly for such high-risk EOC as cardiac surgery) is the random variation in episode spending that can occur, although one method of correcting this is the use of risk stratification. However, Yount and colleagues
      • Yount K.W.
      • Isbell J.M.
      • Lichtendahl C.
      • Dietch Z.
      • Ailawadi G.
      • Kron I.L.
      • et al.
      Bundled payments in cardiac surgery: is risk adjustment sufficient to make it feasible?.
      recently showed that using the STS mortality risk score, the most validated surgical risk score available, did not adequately predict variation in costs. If bundled payment models are to succeed, particularly for cardiac valve surgery, more sophisticated tools must be developed to adjust risk and predict resource consumption.

       Limitations

      There were several limitations to this study. This is a single-center analysis of our initial process results with a relatively small sample size. Before the initiation of our RRI and associated database, the availability of readmission data is scant and we relied on CMS claims data for a control group. Although we had the benefit of our electronic medical record for patients seen at our institution, the accuracy of outside hospital claims data may be unreliable.

      Conclusions

      Our approach of reengineering postdischarge management of BPCI valve patients under tight cardiac surgical service line and CNP control has significantly reduced costly 30-day readmissions in this high-risk population. See Video 1
      Figure thumbnail fx2
      Video 1The senior author (A.C.G.) discussing the significance of the Bundle Initiative and impact on cardiothoracic practice. Video available at: http://www.jtcvsonline.org/article/S0022-5223(17)30403-8/addons.
      for commentary. We must take ownership of the postdischarge period to survive in the era of financial risk sharing and we cannot rely on administrative databases, but must create our own database and processes.
      The full fiscal impact of this strategy on BPCI valve patients will be understood next year when the 90-day CMS master claims data are reconciled. Additionally, further data and analysis may allow focusing postdischarge management resources on those patients identified at high risk for 30/90-day readmissions.

       Conflict of Interest Statement

      Authors have nothing to disclose with regard to commercial support.

      Supplementary Data

      References

        • Schroeder S.A.
        • Frist W.
        National Commission on Physician Payment R. Phasing out fee-for-service payment.
        N Engl J Med. 2013; 368: 2029-2032
      1. Centers for Medicare and Medicaid Services. Bundled payments for care improvement (BPCI) initiative. Available at: https://innovation.cms.gov/initiatives/bundled-payments/. Accessed April 1, 2016.

      2. American Hospital Association. Medicare's bundled payment initiatives: considerations for providers. Available at: http://www.aha.org/content/16/issbrief-bundledpmt.pdf. Accessed April 1, 2016.

        • Hussey P.S.
        • Eibner C.
        • Ridgely M.S.
        • McGlynn E.A.
        Controlling U.S. health care spending—separating promising from unpromising approaches.
        N Engl J Med. 2009; 361: 2109-2111
        • Mechanic R.E.
        • Altman S.H.
        Payment reform options: episode payment is a good place to start.
        Health Aff (Millwood). 2009; 28: w262-w271
        • Fry D.E.
        • Pine M.
        • Nedza S.M.
        • Locke D.G.
        • Reband A.M.
        • Pine G.
        Inpatient and 90-day postdischarge outcomes in cardiac surgery.
        Am J Account Care. 2016; 3: 16-21
        • Iribarne A.
        • Chang H.
        • Alexander J.H.
        • Gillinov A.M.
        • Moquete E.
        • Puskas J.D.
        • et al.
        Readmissions after cardiac surgery: experience of the National Institutes of Health/Canadian Institutes of Health research cardiothoracic surgical trials network.
        Ann Thorac Surg. 2014; 98: 1274-1280
        • Jencks S.F.
        • Williams M.V.
        • Coleman E.A.
        Rehospitalizations among patients in the Medicare fee-for-service program.
        N Engl J Med. 2009; 360: 1418-1428
        • Jubelt L.E.
        • Goldfeld K.S.
        • Chung W.
        • Blecker S.B.
        • Horwitz L.I.
        Changes in discharge location and readmission rates under Medicare bundled payment.
        JAMA Intern Med. 2016; 176: 115-117
      3. Centers for Medicare and Medicaid Services. Readmissions reduction program. Available at: http://www.cms.gov/Medicare/medicare-fee-for-service-payment/acuteinpatientPPS/readmissions-reduction-program.html. Accessed April 1, 2016.

        • Zuckerman R.B.
        • Sheingold S.H.
        • Orav E.J.
        • Ruhter J.
        • Epstein A.M.
        Readmissions, observation, and the Hospital Readmissions Reduction Program.
        N Engl J Med. 2016; 374: 1543-1551
        • Hannan E.L.
        • Samadashvili Z.
        • Jordan D.
        • Sundt T.M.
        • Stamato N.J.
        • Lahey S.J.
        • et al.
        Thirty-day readmissions after transcatheter aortic valve implantation versus surgical aortic valve replacement in patients with severe aortic stenosis in New York State.
        Circ Cardiovasc Interv. 2015; 8: e002744
        • Yount K.W.
        • Isbell J.M.
        • Lichtendahl C.
        • Dietch Z.
        • Ailawadi G.
        • Kron I.L.
        • et al.
        Bundled payments in cardiac surgery: is risk adjustment sufficient to make it feasible?.
        Ann Thorac Surg. 2015; 100: 1646-1652