Center Variation in Medicare Spending for Durable Left Ventricular Assist Device Implant Hospitalizations | Cardiology | JAMA Cardiology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Rank-Ordered Center-Level Mean Observed Price-Standardized Payments for Durable Left Ventricular Assist Device (LVAD) Implant Hospitalizations and Expected Payments Accounting for Patient Case Mix
Rank-Ordered Center-Level Mean Observed Price-Standardized Payments for Durable Left Ventricular Assist Device (LVAD) Implant Hospitalizations and Expected Payments Accounting for Patient Case Mix

LOESS indicates locally estimated scatterplot smoothing.

Figure 2.  Crude and Adjusted Mean Postimplant Length of Stay (and 95% CI) by Price-Standardized Payment Quartile
Crude and Adjusted Mean Postimplant Length of Stay (and 95% CI) by Price-Standardized Payment Quartile
Figure 3.  Risk-Adjusted Adverse Events for Durable Left Ventricular Assist Device (LVAD) Implant Hospitalizations by Price-Standardized Payment Quartile (Compared With Lowest Spending Quartile)
Risk-Adjusted Adverse Events for Durable Left Ventricular Assist Device (LVAD) Implant Hospitalizations by Price-Standardized Payment Quartile (Compared With Lowest Spending Quartile)
Table 1.  Patient Characteristics for the Overall Sample (N = 4442) and by Price-Standardized Payment Quartile for Durable Left Ventricular Assist Device Hospitalizations
Patient Characteristics for the Overall Sample (N = 4442) and by Price-Standardized Payment Quartile for Durable Left Ventricular Assist Device Hospitalizations
Table 2.  Variation in Observed, Price-Standardized, and Price-Standardized and Risk-Standardized Payments for Durable Left Ventricular Assist Device Hospitalizations
Variation in Observed, Price-Standardized, and Price-Standardized and Risk-Standardized Payments for Durable Left Ventricular Assist Device Hospitalizations
1.
Kirklin  JK, Pagani  FD, Kormos  RL,  et al.  Eighth annual INTERMACS report: special focus on framing the impact of adverse events.  J Heart Lung Transplant. 2017;36(10):1080-1086. doi:10.1016/j.healun.2017.07.005PubMedGoogle ScholarCrossref
2.
Hernandez  AF, Shea  AM, Milano  CA,  et al.  Long-term outcomes and costs of ventricular assist devices among Medicare beneficiaries.  JAMA. 2008;300(20):2398-2406. doi:10.1001/jama.2008.716PubMedGoogle ScholarCrossref
3.
Khazanie  P, Hammill  BG, Patel  CB,  et al.  Trends in the use and outcomes of ventricular assist devices among medicare beneficiaries, 2006 through 2011.  J Am Coll Cardiol. 2014;63(14):1395-1404. doi:10.1016/j.jacc.2013.12.020PubMedGoogle ScholarCrossref
4.
Centers for Medicare and Medicaid Services. National Coverage Determination (NCD) for Artificial Hearts and Related Devices (20.9). https://www.cms.gov/medicare-coverage-database/(S(fifmbk55zkfkik552iwqmivv))/details/ncd-details.aspx?NCDId=246&ncdver=5&NCAId=243&ver=5&ExpandComments=n&NcaName=Ventricular+Assist+Devices+as+Destination+Therapy+(2nd+Recon)&bc=BEAAAAAAEAgA&. Published October 30, 2013. Accessed September 8, 2018.
5.
Krim  SR, Vivo  RP, Campbell  P,  et al.  Regional differences in use and outcomes of left ventricular assist devices: Insights from the Interagency Registry for Mechanically Assisted Circulatory Support Registry.  J Heart Lung Transplant. 2015;34(7):912-920. doi:10.1016/j.healun.2015.01.007PubMedGoogle ScholarCrossref
6.
Cowger  JA, Stulak  JM, Shah  P,  et al.  Impact of center left ventricular assist device volume on outcomes after implantation: an INTERMACS analysis.  JACC Heart Fail. 2017;5(10):691-699. doi:10.1016/j.jchf.2017.05.011PubMedGoogle ScholarCrossref
7.
Baras Shreibati  J, Goldhaber-Fiebert  JD, Banerjee  D, Owens  DK, Hlatky  MA.  Cost-effectiveness of left ventricular assist devices in ambulatory patients with advanced heart failure.  JACC Heart Fail. 2017;5(2):110-119. doi:10.1016/j.jchf.2016.09.008PubMedGoogle ScholarCrossref
8.
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.xPubMedGoogle ScholarCrossref
9.
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.2011.0783PubMedGoogle ScholarCrossref
10.
Liang  Q, Ward  S, Pagani  FD,  et al.  Linkage of medicare records to the interagency registry of mechanically assisted circulatory support.  Ann Thorac Surg. 2018;105(5):1397-1402. doi:10.1016/j.athoracsur.2017.11.044PubMedGoogle ScholarCrossref
11.
Weichle  T, Hynes  DM, Durazo-Arvizu  R, Tarlov  E, Zhang  Q.  Impact of alternative approaches to assess outlying and influential observations on health care costs.  Springerplus. 2013;2:614. doi:10.1186/2193-1801-2-614PubMedGoogle ScholarCrossref
12.
Gottlieb  DJ, Zhou  W, Song  Y, Andrews  KG, Skinner  JS, Sutherland  JM.  Prices don’t drive regional Medicare spending variations.  Health Aff (Millwood). 2010;29(3):537-543. doi:10.1377/hlthaff.2009.0609PubMedGoogle ScholarCrossref
13.
Kirklin  JK, Naftel  DC, Pagani  FD,  et al.  Seventh INTERMACS annual report: 15,000 patients and counting.  J Heart Lung Transplant. 2015;34(12):1495-1504. doi:10.1016/j.healun.2015.10.003PubMedGoogle ScholarCrossref
14.
Krumholz  HM, Wang  Y, Mattera  JA,  et al.  An administrative claims model suitable for profiling hospital performance based on 30-day mortality rates among patients with heart failure.  Circulation. 2006;113(13):1693-1701. doi:10.1161/CIRCULATIONAHA.105.611194PubMedGoogle ScholarCrossref
15.
Keenan  PS, Normand  S-LT, Lin  Z,  et al.  An administrative claims measure suitable for profiling hospital performance on the basis of 30-day all-cause readmission rates among patients with heart failure.  Circ Cardiovasc Qual Outcomes. 2008;1(1):29-37. doi:10.1161/CIRCOUTCOMES.108.802686PubMedGoogle ScholarCrossref
16.
Austin  PC.  A comparison of Bayesian methods for profiling hospital performance.  Med Decis Making. 2002;22(2):163-172. doi:10.1177/0272989X0202200213PubMedGoogle ScholarCrossref
17.
Normand  S-LT, Glickman  ME, Gatsonis  CA.  Statistical Methods for Profiling Providers of Medical Care: Issues and Applications.  J Am Stat Assoc. 1997;92:803-814. doi:10.1080/01621459.1997.10474036Google ScholarCrossref
18.
Slaughter  MS, Bostic  R, Tong  K, Russo  M, Rogers  JG.  Temporal changes in hospital costs for left ventricular assist device implantation.  J Card Surg. 2011;26(5):535-541. doi:10.1111/j.1540-8191.2011.01292.xPubMedGoogle ScholarCrossref
19.
Slaughter  MS, Pagani  FD, Rogers  JG,  et al; HeartMate II Clinical Investigators.  Clinical management of continuous-flow left ventricular assist devices in advanced heart failure.  J Heart Lung Transplant. 2010;29(4)(suppl):S1-S39. doi:10.1016/j.healun.2010.01.011PubMedGoogle ScholarCrossref
20.
Mehra  MR, Salerno  C, Cleveland  JC,  et al.  Health care resource use and cost implications in the MOMENTUM 3 long-term outcome study: a randomized controlled trial of a magnetically levitated cardiac pump in advanced heart failure.  Circulation. 2018;(May). doi:10.1161/CIRCULATIONAHA.118.035722Google Scholar
21.
Kilic  A, Phillips  G, Chimanji  N,  et al.  Cost Comparison between Heart Transplantation and Left Ventricular Assist Device Implantation.  J Card Fail. 2014;20(8):S84. doi:10.1016/j.cardfail.2014.06.238Google ScholarCrossref
22.
Digiorgi  PL, Reel  MS, Thornton  B, Burton  E, Naka  Y, Oz  MC.  Heart transplant and left ventricular assist device costs.  J Heart Lung Transplant. 2005;24(2):200-204. doi:10.1016/j.healun.2003.11.397PubMedGoogle ScholarCrossref
23.
Patel  N, Kalra  R, Doshi  R, Bajaj  NS, Arora  G, Arora  P.  Trends and cost of heart transplantation and left ventricular assist devices: impact of proposed federal cuts.  JACC Heart Fail. 2018;6(5):424-432. doi:10.1016/j.jchf.2018.03.005PubMedGoogle ScholarCrossref
24.
Boyle  AJ, Ascheim  DD, Russo  MJ,  et al.  Clinical outcomes for continuous-flow left ventricular assist device patients stratified by pre-operative INTERMACS classification.  J Heart Lung Transplant. 2011;30(4):402-407. doi:10.1016/j.healun.2010.10.016PubMedGoogle ScholarCrossref
25.
Cotts  WG, McGee  EC  Jr, Myers  SL,  et al.  Predictors of hospital length of stay after implantation of a left ventricular assist device: an analysis of the INTERMACS registry.  J Heart Lung Transplant. 2014;33(7):682-688. doi:10.1016/j.healun.2014.02.022PubMedGoogle ScholarCrossref
26.
Shah  P, Pagani  FD, Desai  SS,  et al; Mechanical Circulatory Support Research Network.  Outcomes of patients receiving temporary circulatory support before durable ventricular assist device.  Ann Thorac Surg. 2017;103(1):106-112. doi:10.1016/j.athoracsur.2016.06.002PubMedGoogle ScholarCrossref
27.
Zayat  R, Menon  AK, Goetzenich  A,  et al.  Benefits of ultra-fast-track anesthesia in left ventricular assist device implantation: a retrospective, propensity score matched cohort study of a four-year single center experience.  J Cardiothorac Surg. 2017;12(1):10. doi:10.1186/s13019-017-0573-9PubMedGoogle ScholarCrossref
28.
Cuckler  GA, Sisko  AM, Poisal  JA,  et al.  National health expenditure projections, 2017-26: despite uncertainty, fundamentals primarily drive spending growth.  Health Aff (Millwood). 2018;37(3):482-492. doi:10.1377/hlthaff.2017.1655PubMedGoogle ScholarCrossref
29.
Shah  N, Agarwal  V, Patel  N,  et al.  National trends in utilization, mortality, complications, and cost of care after left ventricular assist device implantation from 2005 to 2011.  Ann Thorac Surg. 2016;101(4):1477-1484. doi:10.1016/j.athoracsur.2015.09.013PubMedGoogle ScholarCrossref
30.
Bachmann  JM, Duncan  MS, Shah  AS,  et al.  Association of cardiac rehabilitation with decreased hospitalizations and mortality after ventricular assist device implantation.  JACC Heart Fail. 2018;6(2):130-139. doi:10.1016/j.jchf.2017.11.002PubMedGoogle ScholarCrossref
31.
Kerrigan  DJ, Williams  CT, Ehrman  JK,  et al.  Cardiac rehabilitation improves functional capacity and patient-reported health status in patients with continuous-flow left ventricular assist devices: the Rehab-VAD randomized controlled trial.  JACC Heart Fail. 2014;2(6):653-659. doi:10.1016/j.jchf.2014.06.011PubMedGoogle ScholarCrossref
32.
Tsiouris  A, Paone  G, Nemeh  HW, Brewer  RJ, Morgan  JA.  Factors determining post-operative readmissions after left ventricular assist device implantation.  J Heart Lung Transplant. 2014;33(10):1041-1047. doi:10.1016/j.healun.2014.05.009PubMedGoogle ScholarCrossref
33.
Akhter  SA, Badami  A, Murray  M,  et al.  Hospital readmissions after continuous-flow left ventricular assist device implantation: incidence, causes, and cost analysis.  Ann Thorac Surg. 2015;100(3):884-889. doi:10.1016/j.athoracsur.2015.03.010PubMedGoogle ScholarCrossref
34.
Patel  S, Poojary  P, Pawar  S,  et al.  National landscape of unplanned 30-day readmissions in patients with left ventricular assist device implantation.  Am J Cardiol. 2018;122(2):261-267. doi:10.1016/j.amjcard.2018.03.363PubMedGoogle ScholarCrossref
35.
Pope  GC, Kautter  J, Ellis  RP,  et al.  Risk adjustment of Medicare capitation payments using the CMS-HCC model.  Health Care Financ Rev. 2004;25(4):119-141.PubMedGoogle Scholar
Original Investigation
January 30, 2019

Center Variation in Medicare Spending for Durable Left Ventricular Assist Device Implant Hospitalizations

Author Affiliations
  • 1Department of Cardiac Surgery, University of Michigan Medical School, Ann Arbor
  • 2School of Public Health, Department of Biostatistics, University of Michigan, Ann Arbor
  • 3University of Michigan Medical School, Ann Arbor
  • 4Department of Health Management and Policy, School of Public Health, University of Michigan
  • 5Division of Cardiovascular Medicine, Department of Internal Medicine, University of Michigan Medical School, Ann Arbor
  • 6Department of Cardiothoracic Surgery, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
JAMA Cardiol. 2019;4(2):153-160. doi:10.1001/jamacardio.2018.4717
Key Points

Question  Does Medicare spending on hospitalizations for durable left ventricular assist device implants vary across centers, and is spending variation associated with clinical outcomes?

Findings  In this cohort study of 4442 durable left ventricular assist device implant hospitalizations across 106 centers, price-standardized and risk-standardized Medicare spending varied by 35% between the lowest and highest spending quartile centers, which was primarily driven by differences in outlier payments between hospitals. Patients treated in higher-spending hospitals had longer postimplant length of stay but similar clinical outcomes.

Meaning  As the supply and demand for durable left ventricular assist device therapy continues to rise, identifying opportunities to reduce variation in spending from both explained and unexplained sources will ensure high-value use.

Abstract

Importance  Hospitalizations for durable left ventricular assist device (LVAD) implants are expensive and increasingly common. Insights into center-level variation in Medicare spending for these hospitalizations are needed to inform value improvement efforts.

Objective  To examine center-level variation in Medicare spending for durable LVAD implant hospitalizations and its association with clinical outcomes.

Design, Setting, and Participants  Retrospective cohort study of linked Medicare administrative claims and Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) clinical data comprising 106 centers in the United States providing durable LVAD implant. Centers were grouped into quartiles based on the mean price-standardized Medicare spending of their patients. The study included Medicare beneficiaries receiving primary durable LVAD implant between January 2008 and December 2014. Data were analyzed between November 2017 and October 2018.

Main Outcomes and Measures  Price-standardized Medicare payments and clinical outcomes. Overall and component (facility diagnosis-related group payments, outlier payments, physician services) payments and clinical outcomes (postimplant length of stay and adverse events) were compared across payment quartiles.

Results  The study sample included 4442 hospitalized patients, with mean (SD) age of 63.0 (10.8) years, 18.7% female, 27.2% nonwhite, and 6.1% Hispanic ethnicity. Among 4442 hospitalizations, the mean (SD) price-standardized Medicare payment was $176 825 ($60 286) and ranged from $122 953 to $271 472 across 106 centers. The difference in price-standardized payments between lowest and highest spending quartiles was $55 446 ($152 714 vs $208 160; 36%; P < .001), with outlier payments making up most of the difference ($42 742; 77%), followed by DRG ($6929; 13%) and physician services ($5774; 10%). After risk standardization, there was a modest decline in the difference in payments between quartiles ($53 221; 35%), with outlier payments accounting for a larger proportion of the difference (84%). After adjusting for patient characteristics, higher price-standardized payment quartiles were associated with longer postimplant length of stay but were not associated with any adverse events.

Conclusions and Relevance  Medicare payments for durable LVAD implant hospitalizations vary widely across centers; this was not well explained by prices or case mix. While associated with longer postimplant length of stay, increased spending was not associated with adverse events. As the supply and demand for durable LVAD therapy continues to rise, identifying opportunities to reduce variation in spending from both explained and unexplained sources will ensure high-value use.

Introduction

The use of left ventricular assist device (LVAD) therapy has surpassed heart transplant as the predominant surgical therapy for patients with medically refractory end-stage heart failure.1 Most of these patients receive durable LVADs as destination therapy, which are devices intended as long-term support outside the hospital. Hospitalizations for implant of a durable LVAD are among the most expensive inpatient stays billed to Medicare, ranging from $175 000 to more than $200 000 per hospitalization.2,3 In response to the rapid growth of durable LVAD therapy, Medicare issued a revised National Coverage Decision in 2013, which expanded its reimbursement policies to include a broader range of eligible beneficiaries and centers qualified to implant durable devices.4 Consequently, durable LVADs are likely to become a greater financial burden for Medicare and other payers.

The extent to which Medicare spending for durable LVAD implant hospitalizations varies across centers is unknown. Analyses that have explored center-level variation in clinical outcomes and health care use associated with durable LVAD therapy have primarily focused on clinical registry data, which lack information on payments associated with care.3,5,6 Conversely, analyses of administrative claims on durable LVAD implant spending lack important clinical factors needed to account for case-mix differences between centers, are unable to identify primary durable LVAD implants, and lack valid data on clinically relevant outcomes.2,7 Moreover, to our knowledge, no studies have examined spending specifically during the index hospitalization, which is often the period of highest spending for surgical procedures.8,9 Informing these gaps in knowledge would provide relevant stakeholders a better understanding of health care spending and outcomes (ie, value) in the use of durable LVAD implants.

In this context, the objective of this study was to describe center-level variation in Medicare spending for durable LVAD implant hospitalizations. We hypothesized that significant variation in spending exists between centers implanting durable LVADs, which may not be explained by traditional risk factors. We also sought to explore the association between center-level spending and clinical outcomes, including postimplant length of stay and adverse events. To accomplish this objective, we leveraged a novel linkage between Medicare Parts A and B administrative billing data and rich clinical registry data from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS).

Methods
Data Sources and Sample

Our study linked 100% Medicare Part A (hospital) and B (physician) administrative claims to INTERMACS registry data, which contains detailed clinical data on patients receiving durable LVADs (2008-2014) using previously published methods.10 This algorithm linked up to 77% of Medicare beneficiaries receiving durable LVAD implants. We included Part A and B payments that occurred within the admission and discharge dates as documented in INTERMACS. We excluded hospitalizations that were unmatched in the INTERMACS registry and Medicare claims, were not enrolled in Medicare Part A and B during the duration of the hospital stay, were LVAD exchanges and not primary implants, were right VAD/total artificial heart implants, or had missing covariate data. The top and bottom 1% of price-standardized payments were excluded to limit the effect of extreme outliers on our estimates.11 To limit the effect of centers with small samples, we excluded patients from centers with 10 or fewer hospitalizations. Our final sample included 4442 patients receiving primary durable LVAD implants from 106 centers (eFigure 1 in the Supplement).

Medicare Payments

Our primary outcome in this study was patient-level price-standardized Medicare spending occurring during the durable LVAD implant hospitalization including discharge diagnosis related group (DRG) payments, outlier payments, and payments for physician services. Outlier payments are made to hospitals for cases with estimated costs greater than a fixed-loss cost threshold or when costs for a procedure (submitted charges multiplied by the hospital’s cost-to-charge ratio) exceed the DRG rate by a specified amount. Thus, the total observed Medicare payment for the hospitalization is the sum of DRG, outlier, and physician payments amounts. Price standardization is achieved by applying the Medicare fee schedule prices to services rendered, which eliminates various price adjustments such as geographic reimbursement rates, disproportionate share status, and indirect medical education. The purpose of price standardization is to ensure that differences in payments between centers reflect differences in health care use and not differences in geography or hospital characteristics.8,12 Payments were also adjusted for inflation (in 2014 US dollars) to ensure equal weighting of payments across years.

We rank ordered centers according to their mean observed Medicare payment and grouped them into quartiles. We then repeated the center rank ordering and quartile assignment process using price-standardized payments instead of observed payments. The quartile of price-standardized payments served as the primary independent variable in our analysis. This study was approved by the University of Michigan institutional review board. An identified INTERMACS dataset was provided by the INTERMACS Data Coordinating Center (University of Alabama) to the University of Michigan through permission from the National Heart, Lung, and Blood Institute for the purposes of data set linkage. Written informed consent for registrant participation in INTERMACS was required until Protocol v4.0 (February 27, 2014).

Clinical Outcomes

The secondary outcomes in this study were patient-level postimplant length of stay and adverse events from the INTERMACS registry. For this study, we focused on common and important clinical outcomes in durable LVAD recipients that are actively collected and monitored using the INTERMACS registry including major infection, major bleeding event, stroke and/or neurological dysfunction, device malfunction, right heart failure, respiratory failure, renal failure, and in-hospital mortality.1,13 We also created a binary indicator for whether the patient experienced at least 1 of these adverse events.

Patient Characteristics

Using INTERMACS, we identified patient factors commonly reported to describe patient case mix and used to risk adjust hospital-level outcomes. Demographic information included age, sex, race/ethnicty (white or nonwhite), Hispanic ethnicity, body mass index category (calculated as weight in kilograms divided by height in meters squared; <25, 25-30, or >30), and blood type O. We also used several variables to capture clinical status such as preimplant length of stay (in days); INTERMACS profile category (1, 2, 3, or 4-7); listed or likely bridge to transplant vs destination therapy; placement of biventricular assist device; the use of dialysis, ventilator, or intra-aortic balloon pump within 48 hours of surgery; history of coronary artery bypass grafting, valve surgery, or mechanical circulatory support device; and concomitant cardiac surgery.

Using International Classification of Diseases, Ninth Revision diagnosis codes available in Medicare Part A claims, we created binary indicators for 31 common comorbidities used to create hierarchical condition categories (HCCs) and then summed the total number of HCCs present in each patient. These HCCs are often used to adjust for center-level differences in heart failure outcomes publicly reported by Medicare.14,15

Statistical Analysis

We described the distribution of patient-level observed and price-standardized payments for durable LVAD implant hospitalizations using box and whisker plots. We also described patient characteristics of the sample by quartile of price-standardized payments and tested for differences in characteristics across quartiles using χ2 tests and analysis of variance for categorical and continuous variables, respectively.

Next, we used a generalized linear mixed model to estimate price-standardized and risk-standardized payments, accounting for the patient characteristics and clustering of patients within centers (ie, the center random effect). Using the empirical Bayes approach, we estimated price-standardized and risk-standardized payments as the ratio of predicted price-standardized payment, including the center random effect from the model to the predicted price-standardized payment excluding the center random effect, multiplied by the sample mean price-standardized payment. This approach minimizes chance variation in payment estimates by shrinking estimates toward the mean, which will ultimately yield more conservative estimates of variation in payments across centers.16,17 We then estimated the coefficient of determination (r2) between center-level price-standardized payments and estimated payments (minus the center random effect) to assess the extent to which patient factors explain variation in payments across centers.

We rank ordered centers according to their price-standardized and risk-standardized payments and grouped them into quartiles. We compared the absolute and relative difference in mean payments between the first (lowest) and fourth (highest) payment quartiles using observed payments, price-standardized payments, and price-standardized and risk-standardized payments. We used generalized linear model t tests to assess statistical differences in absolute and relative differences in overall and component payments (DRG, outlier, and physician payments) between the first and fourth quartile. We also estimated the percentage that each payment component contributed to the overall variation in payments between the first and fourth quartile.

We compared crude and adjusted postimplant length of stay by price-standardized payment quartile using generalized linear mixed models. Similarly, we compared the relative odds of any adverse event and individual adverse events using hierarchical logistic regression models across payment quartiles. Both models adjusted for the same patient factors (eg, demographics, clinical status, and comorbidities) used to risk standardize Medicare payments. All analyses were performed using SAS, version 9.4 (SAS Institute Inc), and statistical tests were deemed significant at α = .05 (2-sided).

Results

The mean (SD) Medicare payment for durable LVAD implant hospitalization in this sample was $202 094 ($130 371) for observed payments and $176 825 ($60 286) for price-standardized payments (eFigure 2 in the Supplement). Descriptive information on the overall study sample and by price-standardized payment quartile can be found in Table 1. Briefly, we found significant differences in the distribution of several patient factors across price-standardized payment quartile including race; Hispanic ethnicity; blood type O; number of HCCs; INTERMACS profile category; listed or likely bridge to transplant vs destination therapy; biventricular assist device placement; the use of dialysis; intra-aortic balloon pump or ventilator in the 48 hours prior to implant; history of coronary artery bypass graft surgery; and concomitant cardiac surgery. The distribution of individual HCCs across payment quartile can be found in eTable 1 in the Supplement.

Mean center-level, price-standardized payments for durable LVAD implant hospitalizations ranged from $122 953 to $271 472. Figure 1 shows the center-level mean durable LVAD implant hospitalization price-standardized payments in rank order, with their corresponding expected mean payment adjusted for patient characteristics. We found that patient factors accounted for 10.8% of the variation in payment across centers (r2 = 0.108; P < .001). The unadjusted and adjusted model coefficients for patient characteristics on price-standardized payments can be found in eTable 2 in the Supplement, and model diagnostics can be found in eFigure 3 in the Supplement.

On average, observed payments for durable LVAD implant hospitalizations were 70% higher ($106 820; 95% CI, $96 053-$117 586; P < .001) for patients in the highest (fourth, $260 086) compared with lowest (fourth, $153 266) payment quartile (Table 2). The relative increases in spending for DRG payments (69%), outlier (80%), and physician services (66%) between lowest and highest spending quartile were similar, although DRG payments composed 83% of the difference between the 2 quartiles. After price adjustment, the difference in payments between lowest and highest quartiles nearly halved to $55 446 (36%), with outlier payments now making up most of the difference in payments (77%). After additional risk standardization, there was a modest decline in the difference in payments between quartiles ($53 221; 35%), with outlier payments accounting for a larger proportion of the difference in payments (84%). As a sensitivity analysis, we also compared payments between the first (lowest) and fourth (highest) payment quartiles without including the random effect for the center (ie, not using the shrinkage estimator) and found less variation in spending across quartiles, but similar attribution of payment variation by component category (DRG, outlier, and physician) (eTable 3 in the Supplement).

Across all patients, the mean (SD) postimplant length of stay was 24.8 (20.0) days. Compared with the lowest price-standardized payment quartile (20.9 days), mean postimplant length of stay was significantly higher in the second (24.0 days), third (24.8 days), and fourth quartiles (29.3 days) (Figure 2). Similarly, the median postimplant length of stay increases as payment quartile increases: first: 17 days (interquartile range [IQR], 13-25 days), second: 19 days (IQR, 14-28 days), third: 20 days (IQR, 14-30 days), and fourth: 23 days (IQR, 15-36 days). After adjustment, postimplant length of stay was significantly longer among patients in second (23.8 days; 95% CI, 22.8-22.4 days), third (24.4 days; 95% CI, 23.4-25.4 days), and fourth (28.5 days, 95% CI, 27.4-29.6 days) payment quartiles, compared with the first quartile (21.1 days, 95% CI, 19.9-22.3 days) (all P < .001).

The rate of any adverse event was 46.1% (2046 of 4442), and the rates of individual adverse events in our sample were 23.8% for major bleeding (n = 1057), 20.2% for major infection (n = 897), 16.3% for respiratory failure (n = 724), 11.6% for right heart failure (n = 515), 9.3% for renal failure (n = 413), 2.9% for stroke and/or neurologic dysfunction (n = 129), 1.8% for device malfunction (n = 80), and 10.5% for in-hospital mortality (n = 466). We were unable to find evidence of an association between price-standardized payment quartile and adverse events, adjusted for patient factors (Figure 3).

Discussion

In this study, we highlight wide variation in Medicare spending for durable LVAD implant hospitalizations across centers, using a unique Medicare-INTERMACS data set. Most of the difference in observed payments across centers was driven by price variation for Medicare services, while differences in patient case mix accounted for little of the variation in payments. However, even after accounting for differences in price and case mix, payments for durable LVAD hospitalizations were 35% higher in higher-spending quartile centers compared with low-spending centers. After accounting for differences in prices and case mix between centers, outlier payments for particularly high-cost hospitalizations accounted for more than 80% of this variation in payments. This finding was also highlighted by the significantly longer postimplant length of stay for patients in higher-payment quartile centers. Finally, our study found no evidence of an association between center-level spending quartile and adverse events.

To our knowledge, our study is the first to highlight wide-scale variability in inpatient spending for durable LVAD implant hospitalizations. Specifically, our findings suggest that reducing excessive outlier payments, particularly in the postimplant period, will be critical for improving value. Slaughter et al18 also found that hospital length of stay was associated with hospital-wide costs for LVAD patients. Guidelines on the clinical treatment of new durable LVAD recipients in the hospital have been already been developed.19 But whether these guidelines have been universally adopted or whether adoption of these guidelines would reduce variation in postimplant length of stay and hospitalization spending has not been established. Furthermore, advances in LVAD technology may also lead to reduced spending. The HeartMate 3 has already been shown to have significantly lower Medicare spending from shorter inpatient stays.20

We also highlight opportunities to improve value in durable LVAD hospitalizations, with the factors that explained center variation in spending: prices and patient case mix. Our findings suggest that differences in prices of Medicare payments for durable LVAD hospitalizations accounted for much of the variation in actual spending between centers. Specifically, DRG payments varied widely between centers, although the reasons for this variation remain unclear. This finding, which contrasts with prior studies within other medical settings, suggests that prices are not a key determinant of geographic variation in spending.12 Moreover, the frequency and magnitude of outlier payments for these hospitalizations suggest that current DRG payment rates are not sufficient to cover the cost of durable LVAD implants in hospitals, which may be because the DRG codes used for durable LVAD implants are the same used for potentially less costly procedures including heart transplant.21-23 Creating a separate DRG payment rate for durable LVAD therapy may result in less spending variation.

The extent to which patient risk can be targeted to reduce spending variation is unclear. We identified several preimplant risk factors that are associated with higher spending such as the INTERMACS profile.24 Other studies25,26 have also found that certain patient and clinical characteristics were associated with increased hospital length of stay. However, these factors did not fully explain variation between centers, which may be owing to the fact that the treatment durable LVAD recipients receive during the hospitalization is more uniform across patient risk, and, thus, the spending is more uniform across patient risk. Risk-stratified care pathways could optimize inpatient stays, including shorter intensive care unit and hospital stays and fewer physician services.27 It may also be that longer lengths of stay at some hospitals were the product of additional care needed to restore health for patients presenting with advanced heart failure. As such, this additional care may help explain how these hospitals achieved similar outcomes as those hospitals having shorter lengths of stay. Our findings do not support efforts to shorten length of stay absent careful consideration, given that this strategy may result in undesirable outcomes.

Overall, our findings offer new insights into the value of durable LVAD therapy. Value can be broadly defined as the quality of care delivered for a given amount of money spent. While length of stay was longer in higher-spending quartile hospitals, there was no evidence that clinical outcomes were associated with center-level spending. In other words, some centers provide the same quality of care at comparatively higher expense, thus providing lower-value care. As health care spending in the United States trends toward 20% of the national gross domestic product within the next decade, it will be important to capitalize on opportunities to improve value in high-cost services.28 As the supply of and demand for expensive durable LVAD therapy continues to rise, identifying opportunities to reduce variation in spending from both explained and unexplained sources will ensure high-value use.

Limitations

Our study has limitations that should be considered. First, this study excludes durable LVAD implants not found in INTERMACS and non-Medicare durable LVAD recipients. While these findings may not be generalizable to all patients with durable LVAD, Medicare patients account for more than half of the population since 2011.29 Second, our study excludes common sources of postdischarge spending in durable LVAD recipients such as readmissions and inpatient rehabilitation.30-34 However, as with many surgical episodes of care, the index hospitalization often represents the largest share of overall spending and variation in spending.8,9 Third, unmeasured confounding may remain and account for some of the unexplained variation. We do note that addition of INTERMACS data improved the performance of our risk-adjustment models and perform better than standard Medicare payment models.35 Finally, our study did not include clinician or center-level factors, which could explain some of the variation in spending. We are exploring analyses to understand how structural characteristics and practice patterns account for variation in Medicare spending across clinicans and beyond the inpatient stay.

Conclusions

Hospitalizations for durable LVAD implants are expensive, vary widely across centers, and are not well explained by patient case mix. Outlier payments accounted for most of the variation in spending between centers, while payments for facilities and physician services accounted for little of the variation in comparison. Moreover, increased spending was associated with longer postimplant length of stay but not with adverse events. It is critical to identify opportunities to reduce variation in spending from both explained and unexplained sources to ensure high-value use.

Back to top
Article Information

Corresponding Author: Michael P. Thompson, PhD, Department of Cardiac Surgery, University of Michigan Medical School, 1500 E Medical Center Dr, 5331K Frankel Cardiovascular Center, SPC 5864, Ann Arbor, MI 48109 (mthomps@med.umich.edu).

Accepted for Publication: November 30, 2018.

Published Online: January 30, 2019. doi:10.1001/jamacardio.2018.4717

Author Contributions: Dr Thompson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Franko, Kormos, Likosky.

Study concept and design: Thompson.

Acquisition, analysis, or interpretation of data: Thompson, Pagani, Liang, Franko, Zhang, McCullough, Strobel, Kormos, Likosky.

Drafting of the manuscript: Thompson, Franko, Zhang, Strobel, Kormos.

Critical revision of the manuscript for important intellectual content: Thompson, Pagani, Liang, Franko, McCullough, Strobel, Kormos, Likosky.

Statistical analysis: Thompson, Liang, Franko, Zhang, McCullough, Strobel.

Obtained funding: Pagani.

Administrative, technical, or material support: Thompson, Pagani, Franko, Likosky.

Study supervision: Thompson, Pagani, Kormos, Likosky.

Conflict of Interest Disclosures: Dr Thompson reported support from Blue Cross Blue Shield of Michigan outside the submitted work. Dr Aaronson reported grants from Medtronic and Abbott, personal fees from Medtronic and NuPulseCV, and support from Procyrion outside the submitted work. Dr Likosky reported grants from Agency for Healthcare Research and Quality during the conduct of the study; grants from the National Institutes of Health, support from Blue Cross Blue Shield of Michigan, and personal fees from AmSECT outside the submitted work. No other disclosures were reported.

Funding/Support: Drs Thompson and Likosky receive partial salary support from Blue Cross Blue Shield of Michigan. Dr Likosky received extramural support from the Agency for Healthcare Research and Quality and the National Institutes of Health. Dr Kormos serves on the Medtronic Physician Advisory Board. Data for this study were provided in part by Interagency Registry for Mechanically Assisted Circulatory Support and funded in part by the National Heart, Lung, and Blood Institute (grant HHSN268201100025C). Additional funding was provided in part by the National Institutes of Health (grant T32-HL-007853).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The Michigan Congestive Heart Failure Investigators are Donald S. Likosky, PhD, University of Michigan Medical School, Department of Cardiac Surgery; Francis D. Pagani, MD, PhD, University of Michigan Medical School, Department of Cardiac Surgery; Michael P. Thompson, PhD, University of Michigan Medical School, Department of Cardiac Surgery; Keith D. Aaronson, MD, MS, University of Michigan Medical School, Department of Internal Medicine; Min Zhang, PhD, University of Michigan School of Public Health, Department of Biostatistics; Qixing Liang, MS, University of Michigan School of Public Health, Department of Biostatistics; Jeffrey S McCullough, PhD, University of Michigan School of Public Health, Department of Health Management and Policy; Supriya Shore, MD, MS, University of Michigan Medical School, Department of Internal Medicine; Preeti Malani, MD, MS, MSJ, University of Michigan Medical School, Department of Internal Medicine; John M. Hollingsworth, MD, MS, University of Michigan Medical School, Department of Urology; Russell J Funk, PhD, University of Minnesota Carlson School of Management, Strategic Management and Entrepreneurship; Emily Mower Provost, PhD, University of Michigan, Department of Computer Science and Engineering; Robert Kormos, MD, University of Pittsburgh, Department of Medicine, Pittsburgh, Pennsylvania; Alexander A. Brescia, MD, University of Michigan Medical School, Department of Cardiac Surgery; Tessa Watt, MD, University of Michigan Medical School, Department of Cardiac Surgery; Raymond J Strobel, MS, University of Michigan Medical School; Daniel Liesman, MD, MS, University of Michigan Medical School; Lynze R. Franko, BS, University of Michigan Medical School; and Joshua Bourque, BS, University of Michigan Medical School.

Disclaimer: The opinions expressed in this manuscript do not represent those of Interagency Registry for Mechanically Assisted Circulatory Support, National Heart, Lung and Blood Institute, Centers for Medicare and Medicaid Services, or US Food and Drug Administration.

Additional Contributions: We thank Daniel Gottlieb, MS (The Dartmouth Institute for Health Policy and Clinical Practice, Hanover, New Hampshire), for expert contributions in merging the Medicare records and Susan Meyers, BBA, QMIS (Department of Surgery, University of Alabama at Birmingham), at the Interagency Registry for Mechanically Assisted Circulatory SupportData Coordinating Center for expert assistance with data queries. No compensation was received from a funding sponsor.

References
1.
Kirklin  JK, Pagani  FD, Kormos  RL,  et al.  Eighth annual INTERMACS report: special focus on framing the impact of adverse events.  J Heart Lung Transplant. 2017;36(10):1080-1086. doi:10.1016/j.healun.2017.07.005PubMedGoogle ScholarCrossref
2.
Hernandez  AF, Shea  AM, Milano  CA,  et al.  Long-term outcomes and costs of ventricular assist devices among Medicare beneficiaries.  JAMA. 2008;300(20):2398-2406. doi:10.1001/jama.2008.716PubMedGoogle ScholarCrossref
3.
Khazanie  P, Hammill  BG, Patel  CB,  et al.  Trends in the use and outcomes of ventricular assist devices among medicare beneficiaries, 2006 through 2011.  J Am Coll Cardiol. 2014;63(14):1395-1404. doi:10.1016/j.jacc.2013.12.020PubMedGoogle ScholarCrossref
4.
Centers for Medicare and Medicaid Services. National Coverage Determination (NCD) for Artificial Hearts and Related Devices (20.9). https://www.cms.gov/medicare-coverage-database/(S(fifmbk55zkfkik552iwqmivv))/details/ncd-details.aspx?NCDId=246&ncdver=5&NCAId=243&ver=5&ExpandComments=n&NcaName=Ventricular+Assist+Devices+as+Destination+Therapy+(2nd+Recon)&bc=BEAAAAAAEAgA&. Published October 30, 2013. Accessed September 8, 2018.
5.
Krim  SR, Vivo  RP, Campbell  P,  et al.  Regional differences in use and outcomes of left ventricular assist devices: Insights from the Interagency Registry for Mechanically Assisted Circulatory Support Registry.  J Heart Lung Transplant. 2015;34(7):912-920. doi:10.1016/j.healun.2015.01.007PubMedGoogle ScholarCrossref
6.
Cowger  JA, Stulak  JM, Shah  P,  et al.  Impact of center left ventricular assist device volume on outcomes after implantation: an INTERMACS analysis.  JACC Heart Fail. 2017;5(10):691-699. doi:10.1016/j.jchf.2017.05.011PubMedGoogle ScholarCrossref
7.
Baras Shreibati  J, Goldhaber-Fiebert  JD, Banerjee  D, Owens  DK, Hlatky  MA.  Cost-effectiveness of left ventricular assist devices in ambulatory patients with advanced heart failure.  JACC Heart Fail. 2017;5(2):110-119. doi:10.1016/j.jchf.2016.09.008PubMedGoogle ScholarCrossref
8.
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.xPubMedGoogle ScholarCrossref
9.
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.2011.0783PubMedGoogle ScholarCrossref
10.
Liang  Q, Ward  S, Pagani  FD,  et al.  Linkage of medicare records to the interagency registry of mechanically assisted circulatory support.  Ann Thorac Surg. 2018;105(5):1397-1402. doi:10.1016/j.athoracsur.2017.11.044PubMedGoogle ScholarCrossref
11.
Weichle  T, Hynes  DM, Durazo-Arvizu  R, Tarlov  E, Zhang  Q.  Impact of alternative approaches to assess outlying and influential observations on health care costs.  Springerplus. 2013;2:614. doi:10.1186/2193-1801-2-614PubMedGoogle ScholarCrossref
12.
Gottlieb  DJ, Zhou  W, Song  Y, Andrews  KG, Skinner  JS, Sutherland  JM.  Prices don’t drive regional Medicare spending variations.  Health Aff (Millwood). 2010;29(3):537-543. doi:10.1377/hlthaff.2009.0609PubMedGoogle ScholarCrossref
13.
Kirklin  JK, Naftel  DC, Pagani  FD,  et al.  Seventh INTERMACS annual report: 15,000 patients and counting.  J Heart Lung Transplant. 2015;34(12):1495-1504. doi:10.1016/j.healun.2015.10.003PubMedGoogle ScholarCrossref
14.
Krumholz  HM, Wang  Y, Mattera  JA,  et al.  An administrative claims model suitable for profiling hospital performance based on 30-day mortality rates among patients with heart failure.  Circulation. 2006;113(13):1693-1701. doi:10.1161/CIRCULATIONAHA.105.611194PubMedGoogle ScholarCrossref
15.
Keenan  PS, Normand  S-LT, Lin  Z,  et al.  An administrative claims measure suitable for profiling hospital performance on the basis of 30-day all-cause readmission rates among patients with heart failure.  Circ Cardiovasc Qual Outcomes. 2008;1(1):29-37. doi:10.1161/CIRCOUTCOMES.108.802686PubMedGoogle ScholarCrossref
16.
Austin  PC.  A comparison of Bayesian methods for profiling hospital performance.  Med Decis Making. 2002;22(2):163-172. doi:10.1177/0272989X0202200213PubMedGoogle ScholarCrossref
17.
Normand  S-LT, Glickman  ME, Gatsonis  CA.  Statistical Methods for Profiling Providers of Medical Care: Issues and Applications.  J Am Stat Assoc. 1997;92:803-814. doi:10.1080/01621459.1997.10474036Google ScholarCrossref
18.
Slaughter  MS, Bostic  R, Tong  K, Russo  M, Rogers  JG.  Temporal changes in hospital costs for left ventricular assist device implantation.  J Card Surg. 2011;26(5):535-541. doi:10.1111/j.1540-8191.2011.01292.xPubMedGoogle ScholarCrossref
19.
Slaughter  MS, Pagani  FD, Rogers  JG,  et al; HeartMate II Clinical Investigators.  Clinical management of continuous-flow left ventricular assist devices in advanced heart failure.  J Heart Lung Transplant. 2010;29(4)(suppl):S1-S39. doi:10.1016/j.healun.2010.01.011PubMedGoogle ScholarCrossref
20.
Mehra  MR, Salerno  C, Cleveland  JC,  et al.  Health care resource use and cost implications in the MOMENTUM 3 long-term outcome study: a randomized controlled trial of a magnetically levitated cardiac pump in advanced heart failure.  Circulation. 2018;(May). doi:10.1161/CIRCULATIONAHA.118.035722Google Scholar
21.
Kilic  A, Phillips  G, Chimanji  N,  et al.  Cost Comparison between Heart Transplantation and Left Ventricular Assist Device Implantation.  J Card Fail. 2014;20(8):S84. doi:10.1016/j.cardfail.2014.06.238Google ScholarCrossref
22.
Digiorgi  PL, Reel  MS, Thornton  B, Burton  E, Naka  Y, Oz  MC.  Heart transplant and left ventricular assist device costs.  J Heart Lung Transplant. 2005;24(2):200-204. doi:10.1016/j.healun.2003.11.397PubMedGoogle ScholarCrossref
23.
Patel  N, Kalra  R, Doshi  R, Bajaj  NS, Arora  G, Arora  P.  Trends and cost of heart transplantation and left ventricular assist devices: impact of proposed federal cuts.  JACC Heart Fail. 2018;6(5):424-432. doi:10.1016/j.jchf.2018.03.005PubMedGoogle ScholarCrossref
24.
Boyle  AJ, Ascheim  DD, Russo  MJ,  et al.  Clinical outcomes for continuous-flow left ventricular assist device patients stratified by pre-operative INTERMACS classification.  J Heart Lung Transplant. 2011;30(4):402-407. doi:10.1016/j.healun.2010.10.016PubMedGoogle ScholarCrossref
25.
Cotts  WG, McGee  EC  Jr, Myers  SL,  et al.  Predictors of hospital length of stay after implantation of a left ventricular assist device: an analysis of the INTERMACS registry.  J Heart Lung Transplant. 2014;33(7):682-688. doi:10.1016/j.healun.2014.02.022PubMedGoogle ScholarCrossref
26.
Shah  P, Pagani  FD, Desai  SS,  et al; Mechanical Circulatory Support Research Network.  Outcomes of patients receiving temporary circulatory support before durable ventricular assist device.  Ann Thorac Surg. 2017;103(1):106-112. doi:10.1016/j.athoracsur.2016.06.002PubMedGoogle ScholarCrossref
27.
Zayat  R, Menon  AK, Goetzenich  A,  et al.  Benefits of ultra-fast-track anesthesia in left ventricular assist device implantation: a retrospective, propensity score matched cohort study of a four-year single center experience.  J Cardiothorac Surg. 2017;12(1):10. doi:10.1186/s13019-017-0573-9PubMedGoogle ScholarCrossref
28.
Cuckler  GA, Sisko  AM, Poisal  JA,  et al.  National health expenditure projections, 2017-26: despite uncertainty, fundamentals primarily drive spending growth.  Health Aff (Millwood). 2018;37(3):482-492. doi:10.1377/hlthaff.2017.1655PubMedGoogle ScholarCrossref
29.
Shah  N, Agarwal  V, Patel  N,  et al.  National trends in utilization, mortality, complications, and cost of care after left ventricular assist device implantation from 2005 to 2011.  Ann Thorac Surg. 2016;101(4):1477-1484. doi:10.1016/j.athoracsur.2015.09.013PubMedGoogle ScholarCrossref
30.
Bachmann  JM, Duncan  MS, Shah  AS,  et al.  Association of cardiac rehabilitation with decreased hospitalizations and mortality after ventricular assist device implantation.  JACC Heart Fail. 2018;6(2):130-139. doi:10.1016/j.jchf.2017.11.002PubMedGoogle ScholarCrossref
31.
Kerrigan  DJ, Williams  CT, Ehrman  JK,  et al.  Cardiac rehabilitation improves functional capacity and patient-reported health status in patients with continuous-flow left ventricular assist devices: the Rehab-VAD randomized controlled trial.  JACC Heart Fail. 2014;2(6):653-659. doi:10.1016/j.jchf.2014.06.011PubMedGoogle ScholarCrossref
32.
Tsiouris  A, Paone  G, Nemeh  HW, Brewer  RJ, Morgan  JA.  Factors determining post-operative readmissions after left ventricular assist device implantation.  J Heart Lung Transplant. 2014;33(10):1041-1047. doi:10.1016/j.healun.2014.05.009PubMedGoogle ScholarCrossref
33.
Akhter  SA, Badami  A, Murray  M,  et al.  Hospital readmissions after continuous-flow left ventricular assist device implantation: incidence, causes, and cost analysis.  Ann Thorac Surg. 2015;100(3):884-889. doi:10.1016/j.athoracsur.2015.03.010PubMedGoogle ScholarCrossref
34.
Patel  S, Poojary  P, Pawar  S,  et al.  National landscape of unplanned 30-day readmissions in patients with left ventricular assist device implantation.  Am J Cardiol. 2018;122(2):261-267. doi:10.1016/j.amjcard.2018.03.363PubMedGoogle ScholarCrossref
35.
Pope  GC, Kautter  J, Ellis  RP,  et al.  Risk adjustment of Medicare capitation payments using the CMS-HCC model.  Health Care Financ Rev. 2004;25(4):119-141.PubMedGoogle Scholar
×