- Email: [email protected]
Age and Frailty Influence Hip and Knee Arthroplasty Reimbursement in a Bundled Payment Care Improvement Initiative
The Journal of Arthroplasty 34 (2019) S80eS83
Contents lists available at ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Health Policy & Economics
Age and Frailty Influence Hip and Knee Arthroplasty Reimbursement in a Bundled Payment Care Improvement Initiative Andrew M. Pepper, MD a, b, *, David Novikov, BS a, Zlatan Cizmic, MD a, John T. Barrett, BS a, Michael Collins, BS a, Richard Iorio, MD a, Ran Schwarzkopf, MD, MSc a, William J. Long, MD, FRCSC a, b a b
Langone Orthopedic Hospital, Department of Orthopaedic Surgery, New York University, New York, NY Insall Scott Kelly Institute, New York, NY
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 November 2018 Received in revised form 6 January 2019 Accepted 22 January 2019 Available online 30 January 2019
Background: The Bundled Payment Care Improvement (BPCI) initiative aims to improve quality of patient care while mitigating cost. How patient age and frailty affect reimbursement after hip and knee total joint arthroplasty (TJA) is not known. This study evaluates if patient age and frailty affect cost of care. Methods: A retrospective review of prospectively collected data of 1821 patients undergoing TJA at our institution under the BPCI initiative was performed from 2013 to 2016. We recorded demographics for patients and calculated their modified frailty index (mFI). Cost of care was obtained for each patient. Statistical analyses included t-test and analysis of variance to evaluate age and frailty as independent categorical variables. Beta coefficients were utilized to evaluate age as a continuous variable. Multivariate linear regression models evaluated age and frailty’s combined contribution to cost. Results: Age was evaluated as a categorical variable, with the median age of our sample population the categorical cutoff. Age 72 years and increasing mFI score were associated with statistically significant increased cost. Increasing age demonstrated a statistically significant increase in cost of 0.68% per incremental age increase. Multivariate evaluation of increasing age and mFI revealed a statistically significant increase in cost for mFI score 2. Conclusion: Increasing age and frailty increase cost associated with TJA. The BPCI initiative oversimplifies the cost associated with TJA. Concerningly, this information could deincentivize care to older, higher risk patients. Objective patient-specific and risk-adjusted stratification of BPCI pricing is necessary to be considered as a valid financial model. © 2019 Elsevier Inc. All rights reserved.
Keywords: total knee arthroplasty total hip arthroplasty bundled payment cost age frailty
It is well documented that the cost of healthcare in the United States is among the highest in the world and is projected to increase to approximately 20% of gross domestic product by the year 2026 [1]. For this reason, there has been increased focus on alternative
One or more of the authors of this paper have disclosed potential or pertinent conflicts of interest, which may include receipt of payment, either direct or indirect, institutional support, or association with an entity in the biomedical field which may be perceived to have potential conflict of interest with this work. For full disclosure statements refer to https://doi.org/10.1016/j.arth.2019.01.050. Source of Funding: No funding was utilized or required to perform this study at our institution. * Reprint requests: Andrew M. Pepper, MD, Andrews Institute for Orthopaedics and Sports Medicine, Baptist Medical Park - Airport, 5100 North 12th Avenue, Suite 102, Pensacola, FL 32504. https://doi.org/10.1016/j.arth.2019.01.050 0883-5403/© 2019 Elsevier Inc. All rights reserved.
payment models (APMs) to attempt to mitigate the costs of providing care in the United States. One such model is the Bundled Payment Care Improvement (BPCI) initiative, which aims to improve the quality of patient care while simultaneously mitigating the associated cost [2]. The BPCI initiative was created out of the Affordable Care Act with the distinct goal to “test innovative payment and service delivery models” while reducing the expenditures of Medicare/Medicaid and ultimately improve the quality of patient care [2e4]. Intense focus on the cost of total joint arthroplasty (TJA) is due to the large proportion of the Centers for Medicare and Medicaid Services’ expenditure contribution attributed to total knee and hip arthroplasty (TKA/THA), with projections that the population’s need for TJA will increase exponentially in the future [5]. The cost for care in the BPCI initiative is determined by a variety of factors [3]. In early models of the BPCI initiative, cost of care
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
considered are: historical cost data, regional data supplements to unavailable historical cost data, episode costs were trended compared to national episode-specific growth rates, with discounts applied depending on model participation [3]. Although many multiple considerations contribute to cost determination in this payment model, there is no episode cost stratification based on individual patient demographics, including age or patient-specific risk factors [4]. Frailty is defined as a decreased physiological reserve and it typically increases with increasing age, but is considered to be a measure of overall health status independent of physiologic age [6e8]. Frailty has been demonstrated as an effective tool in predicting perioperative risk for multiple surgical subspecialties [9e12]. The modified frailty index (mFI) is a simple tool that utilizes a subset of standardized patient parameters that are commonly collected during standard patient perioperative encounters from the original Canadian Study of Health and Aging Frailty Index [13]. mFI consists of 11 patient comorbid variables: (1) nonindependent functional status (partially or totally dependent activities of daily living); (2) history of no diabetes or diabetes controlled by diet alone, diabetes treated with oral antihyperglycemic therapy, or diabetes treated with insulin therapy; (3) history of chronic obstructive pulmonary disease exacerbation or pneumonia within 30 days; (4) history of congestive heart failure exacerbation within 30 days; (5) history of myocardial infarction within 6 months; (6) history of angina within 30 days or any percutaneous coronary intervention or coronary artery bypass grafting; (7) hypertension requiring medication; (8) history of peripheral vascular disease; (9) acutely impaired sensorium; (10) transient ischemic attack or cerebrovascular accident without deficits; and (11) cerebrovascular accident with deficits. Each variable was scored dichotomously and the final mFI was the sum of present variables, for simple interpretation. The mFI tool has proven to be accurate in predicting adverse perioperative outcomes in a variety of surgical subspecialties and in orthopedics specifically, it is a reliable tool for risk stratifying patients undergoing TJA [7,12]. To our knowledge, there is no information in the literature evaluating how mFI affects cost of care in TJA. The aim of this study is to evaluate the effect that both patient age and mFI score have on the cost of care for TKA and THA and this cost’s effect in a BPCI initiative model. Our hypothesis was that increasing age and mFI score would result in increased episode cost. Materials and Methods After Institutional Review Board approval, a retrospective review of prospectively collected data of Medicare patients undergoing TJA (billing code Diagnosis-Related Groups 469 and 470) at our institution under the BPCI initiative from 2013 to 2016 was analyzed. Demographic data were collected for each patient, including age and medical history. mFI was calculated for each patient based on review of his or her medical history. Cost of care was obtained for each patient, in reference to BPCI initiative data available at our institution. Patients who underwent THA for a hip fracture, had incomplete data that made calculating mFI impossible, or had incomplete cost data were excluded from analysis. One thousand eight hundred twenty-one patients were included in the final analysis. When age was treated as a categorical variable, the median age of our sample population (72 years) was used as the cutoff. All cost data are presented as a percentage and dollar cost compared to the institution-specific reference standard for BPCI initiative at our institution. Positive values indicate profit to the healthcare system in the BPCI initiative; negative values indicate deficit to the healthcare system in the BPCI initiative.
S81
Table 1 Comparison of Cost by Age Groups and mFI Score. Variable
n (%)
% Difference From Targeta/US Dollars (SD)
Overall Age group <72 72 mFI score 0 1 2 3 4 5
1821 (100%)
e
P-Value
<.01 872 (47.9) 949 (52.1)
þ19.2%/þ$6120 (32.3%) þ10.6%/þ$3391 (39.5%) <.01
558 767 368 92 26 6
(30.7) (42.2) (20.3) (5.1) (1.4) (0.3)
þ21.2%/þ$6752 þ16.8%/þ$5369 þ10.4%/þ$3334 10.5%/$3341 14.1%/$4491 47.9%/$15,268
(31.5%) (30.0%) (33.9%) (66.1%) (74.2%) (107.8%)
mFI, modified frailty index. P < .05 represents statistically significant findings, bolded. a Positive values represent percentage of decreased cost of care/US dollar profit compared to the institutional reference standard; negative values represent percentage of increased cost of care/US dollar increase cost compared to the institutional reference standard.
All statistical analyses were performed using SPSS software (IBM SPSS). Statistical analyses included t-test and analysis of variance to evaluate age and frailty as independent categorical variables. Beta coefficients were utilized to evaluate age as a continuous variable. Multivariate linear regression models evaluated age and frailty’s combined contribution to cost. Results When evaluating age as an independent categorical variable, our data revealed that age 72 years was associated with increased cost difference of þ8.6% compared to patients <72 years (P < .01) (Table 1 and Fig. 1). Increasing mFI score as an independent variable was also associated with increased cost (P < .01) per care episode (Table 1). Positive values represent percentage of decreased cost of care/US dollar profit, whereas negative values represent percentage of increased cost of care/US dollar increase cost. When age was evaluated as an independent continuous variable, our data revealed an increase cost per care episode of 0.68%, or $217, for each incremental year increase in age (95% confidence interval 0.47-0.89). In a multivariate linear regression model evaluating both age and mFI score, the data revealed statistically significant (P < .01) incremental increasing cost of care for both continuously increasing age and mFI score, except for an mFI score of 1 (P ¼ 0.06) (Table 2 and Fig. 2). Discussion Our data from the BPCI initiative at our institution, from inception of BPCI in 2013 through the most recent complete set of data for 2016, reveal that both increasing age and mFI score have a significant effect on the cost of care per episode for TKA and THA. This effect is evident for each variable independently and in multivariate analyses. The effect of increasing age and mFI score on the cost of care per episode correlates to decreasing reimbursement in a BPCI initiative model. Our findings are the first to directly evaluate the effects of simple patient factors, such as age and mFI, on episode cost in a BPCI initiative. Currently, the BPCI initiative does not take into consideration patient age and modifiable and nonmodifiable risk factors into determining payment for TJA. Although the goal of the BPCI initiative is to decrease cost and improve the quality of care, the current pricing scheme over-simplifies the cost associated with TJA. Further research in this area and refinement of pricing metrics is advisable to better estimate the cost associated with TJA and
S82
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
Fig. 1. Comparison of cost by age groups and mFI score. Positive values represent US dollar profit compared to the institutional reference standard; negative values represent US dollar increase cost compared to the institutional reference standard. P < .05. mFI, modified frailty index.
continue to drive value and quality of care in the US healthcare system. Our data should be evaluated cautiously. We recognize that it is a limited data set and was analyzed in a retrospective manner, and therefore should not be taken as a definitive evaluation of cost implications related to increasing age and mFI. These data should especially be carefully utilized, as it could easily be used as validation for deferring care to older and/or more frail patients when care is provided in a BPCI initiative or Comprehensive Care for Joint Replacement model. Advanced age and increasing frailty are not a contraindication to TKA and/or THA, as the literature supports TJA as an exceptional treatment for decreasing pain, improving function, and increasing the quality of life of patients with advanced hip and knee osteoarthritis [14]. It is advisable that we avoid institutional cost discrimination for elderly/frail patients in building and implementing APMs, as this can inadvertently lead to care discrimination for patients who are still well indicated for TJA. These findings also raise questions of regional and geographical care distribution and care access equity for patients seeking TJA. As the current BPCI/Comprehensive Care for Joint Replacement models are constructed, the potential for smaller, rural, low-volume arthroplasty centers to absorb the additional costs demonstrated in our data may inadvertently deincentivize equitable care for higher risk patients at these locations, as these patients’ care may be especially cost prohibitive. Alternatively, if these patients are diverted to larger, tertiary care centers, this also can potentially concentrate cost prohibitive inequitable care to larger urban academic centers, causing further volume and cost strain on tertiary care centers. Additionally, this could limit geographic access to care
requiring patients at higher risk to travel long distances for care and limit their access to postoperative care. The greater question raised by this discussion is the structure of bundled payment models, which assume cost of care as a fixed value based on over-simplified estimates of cost of care for TJA. This ignores the fact that patients are, by nature, not equivalent and that appropriate care requires an individualized approach to achieve consistent quality of care [15e17]. This includes careful thorough perioperative evaluation, risk factor identification and optimization, surgical planning, varying degrees of surgical technique/equipment for increasing complexity, and differing implant selection [18]. Without addressing the lack of patient-specific cost consideration in current APMs, we are destined to develop institutionalized discrimination in care quality and access, specifically for patients who require the expertise of well-trained arthroplasty surgeons. This situation amounts to national institutionalized gain-sharing which has the potential to perpetuate care discrimination [19]. This study is not without limitations. The study was designed as a retrospective review of cost data from our institution’s BPCI initiative database. Our data are not appropriately powered to detect subtle differences in cost attributable solely to age and frailty. Additionally, there was no control group to compare directly to our study cohort. The range of ages was limited to those patients in our cohort, and therefore a complete evaluation of age ranges is not provided. Similarly, mFI scores ranged from 0 to 5 and we therefore are unable to comment on higher mFI scores. Also, our higher mFI scores had lower numbers of patients and are underpowered to make definitive conclusions regarding the effect on cost. As an early participant in the BPCI initiative, our institution concurrently focused on perioperative optimization of patients known to be a higher risk of perioperative complications, and this is a potential source of confounding. Studies comparing the effect of optimization on improving perioperative outcomes are diffuse in the literature and have been established as the standard of care for patients undergoing TJA [15,20,21]. Finally, we did not collect data regarding the cause of cost differences related to age and frailty (ie, prolonged length of stay, postdischarge disposition, etc.) and cannot therefore comment on the specific contributors to increased or decreased episode cost. The strengths of our study are the discreet reliable cost information available from our institutions BPCI initiative database. The data we present are also generalizable, as our patients undergoing standardized perioperative optimization, which have been established as the standard of care for patients undergoing TJA [15,20,21]. This method is consistent with the protocols that are generally adopted by hospitals across the nation. Our data are also
Table 2 Multivariable Linear Regression of Cost of Care Predicted by Age (Continuous) and mFI. Variable Age (continuous) mFI score 0 (reference) 1 2 3 4 5
Beta Coefficienta/US Dollars (95% CI) 0.59%/$188 (0.38-0.80) e 3.3%/$1047 9.7%/$3096 29.9%/$9526 33.8%/$10,774 67.4%/$21,483
(0.55 to 7.1) (5.1-14.3) (22.1-37.6) (20.0-47.6) (39.2-95.6)
P-Value <.01 e .09 <.01 <.01 <.01 <.01
CI, confidence interval; mFI, modified frailty index. P <.05 represents statistically significant findings, bolded. a Beta coefficient represents the increase in US dollar cost (as percentage of the institutional reference standard) for every incremental increase in age (in year increments) associated with mFI score.
Fig. 2. Multivariable linear regression of cost of care predicted by age (continuous) and mFI. Beta coefficient represents the increase in US dollar cost (compared to the difference from the institutional reference standard) for every incremental increase in age (in year increments) associated with mFI score. P < .05 represents statistically significant findings. mFI, modified frailty index.
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
some of the earliest cost data available, as our institution was an early participant in the BPCI initiative at its outset in 2013. Conclusions We demonstrate in this study that simple patient factors such as age and frailty significantly affect the cost of care and reimbursement in a BPCI initiative for TKA and THA. At present, there is no patient-specific cost consideration to account for these differences in current bundled payment models. This information demonstrates that further study is required to better elucidate cost considerations for arthroplasty patients. The authors also highlight the potential implications of fixed-fee payment models and advocate for equitable access to arthroplasty care without penalizing surgeons and/or institutions who are dedicated to providing quality care to all patients. References [1] National healthcare expenditure projections, 2017-2026. https://www cmsgov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/ NationalHealthExpendData/NationalHealthAccountsProjected.html; 2011 [accessed 9.7.18]. [2] Bundled payments for care improvement (BPCI) initiative: general information. https://innovationcmsgov/initiatives/bundled-payments/ [accessed 09.07.18]. [3] Bundled payments for care improvement initiative (BPCI). https://www cmsgov/Newsroom/MediaReleaseDatabase/Fact-sheets/2016-Fact-sheetsitems/2016-04-18html [accessed 09.07.18]. [4] Press MJ, Rajkumar R, Conway PH. Medicare’s new bundled payments: design, strategy, and evolution. JAMA 2016;315:131e2. [5] Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89:780e5. [6] Farhat JS, Velanovich V, Falvo AJ, Horst HM, Swartz A, Patton JH, et al. Are the frail destined to fail? Frailty index as predictor of surgical morbidity and mortality in the elderly. J Trauma Acute Care Surg 2012;72:1526e30. [7] Shin JI, Keswani A, Lovy AJ, Moucha CS. Simplified frailty index as a predictor of adverse outcomes in total hip and knee arthroplasty. J Arthroplasty 2016;31:2389e94.
S83
[8] Kulminski AM, Ukraintseva SV, Kulminskaya IV, Arbeev KG, Land K, Yashin AI. Cumulative deficits better characterize susceptibility to death in elderly people than phenotypic frailty: lessons from the cardiovascular health study. J Am Geriatr Soc 2008;56:898e903. [9] Bellamy JL, Runner RP, Vu CCL, Schenker ML, Bradbury TL, Roberson JR. Modified frailty index is an effective risk assessment tool in primary total hip arthroplasty. J Arthroplasty 2017;32:2963e8. [10] Runner RP, Bellamy JL, Vu CCL, Erens GA, Schenker ML, Guild GN. Modified frailty index is an effective risk assessment tool in primary total knee arthroplasty. J Arthroplasty 2017;32:S177e82. [11] Patel KV, Brennan KL, Brennan ML, Jupiter DC, Shar A, Davis ML. Association of a modified frailty index with mortality after femoral neck fracture in patients aged 60 years and older. Clin Orthop Relat Res 2014;472:1010e7. [12] Wahl TS, Graham LA, Hawn MT, Richman J, Hollis RH, Jones CE, et al. Association of the modified frailty index with 30-day surgical readmission. JAMA Surg 2017;152:749e57. [13] Rockwood K, Song X, Macknight C, Bergman H, Hogan DB, Mcdowell I, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ 2005;173:489e95. [14] Mont MA, Tanzer M. Orthopaedic knowledge update: hip and knee reconstruction 5. Am Acad Orthop Surg 2017;173:489e95. Chapter 18, Chapter 30. [15] Chimento GF, Thomas LC. The perioperative surgical home: improving the value and quality of care in total joint replacement. Curr Rev Musculoskelet Med 2017;10:365e9. [16] Clement RC, Soo AE, Kheir MM, Derman PB, Flynn DN, Levin LS, et al. What incentives are created by Medicare payments for total hip arthroplasty? J Arthroplasty 2016;31(9 Suppl):69e72. [17] Magone K, Kemker 3rd BP, Pilipenko N, O’Connor E, Walter N, Atkinson T. The new surgical technique for improving total knee and hip arthroplasty outcomes: patient selection. J Arthroplasty 2017;32:2070e6. [18] Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR, Manley MT. Which clinical and patient factors influence the national economic burden of hospital readmissions after total joint arthroplasty? Clin Orthop Relat Res 2017;475: 2926e37. [19] Ramkumar PN. How bundling care hurts the patients who need it the most. Forbes Magazine; 2018. https://www.forbes.com/sites/premramkumar/2018/ 07/11/how-bundling-care-hurts-the-patients-who-need-it-the-most/#42e3efeb 3250 [accessed 13.07.18]. [20] Boraiah S, Joo L, Inneh IA, Rathod P, Meftah M, Band P, et al. Management of modifiable risk factors prior to primary hip and knee arthroplasty: a readmission risk assessment tool. J Bone Joint Surg Am 2015;97:1921e8. [21] Paxton EW, Inacio MC, Singh JA, Love R, Bini SA, Namba RS. Are there modifiable risk factors for hospital readmission after total hip arthroplasty in a US healthcare system? Clin Orthop Relat Res 2015;473:3446e55.
Contents lists available at ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Health Policy & Economics
Age and Frailty Influence Hip and Knee Arthroplasty Reimbursement in a Bundled Payment Care Improvement Initiative Andrew M. Pepper, MD a, b, *, David Novikov, BS a, Zlatan Cizmic, MD a, John T. Barrett, BS a, Michael Collins, BS a, Richard Iorio, MD a, Ran Schwarzkopf, MD, MSc a, William J. Long, MD, FRCSC a, b a b
Langone Orthopedic Hospital, Department of Orthopaedic Surgery, New York University, New York, NY Insall Scott Kelly Institute, New York, NY
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 November 2018 Received in revised form 6 January 2019 Accepted 22 January 2019 Available online 30 January 2019
Background: The Bundled Payment Care Improvement (BPCI) initiative aims to improve quality of patient care while mitigating cost. How patient age and frailty affect reimbursement after hip and knee total joint arthroplasty (TJA) is not known. This study evaluates if patient age and frailty affect cost of care. Methods: A retrospective review of prospectively collected data of 1821 patients undergoing TJA at our institution under the BPCI initiative was performed from 2013 to 2016. We recorded demographics for patients and calculated their modified frailty index (mFI). Cost of care was obtained for each patient. Statistical analyses included t-test and analysis of variance to evaluate age and frailty as independent categorical variables. Beta coefficients were utilized to evaluate age as a continuous variable. Multivariate linear regression models evaluated age and frailty’s combined contribution to cost. Results: Age was evaluated as a categorical variable, with the median age of our sample population the categorical cutoff. Age 72 years and increasing mFI score were associated with statistically significant increased cost. Increasing age demonstrated a statistically significant increase in cost of 0.68% per incremental age increase. Multivariate evaluation of increasing age and mFI revealed a statistically significant increase in cost for mFI score 2. Conclusion: Increasing age and frailty increase cost associated with TJA. The BPCI initiative oversimplifies the cost associated with TJA. Concerningly, this information could deincentivize care to older, higher risk patients. Objective patient-specific and risk-adjusted stratification of BPCI pricing is necessary to be considered as a valid financial model. © 2019 Elsevier Inc. All rights reserved.
Keywords: total knee arthroplasty total hip arthroplasty bundled payment cost age frailty
It is well documented that the cost of healthcare in the United States is among the highest in the world and is projected to increase to approximately 20% of gross domestic product by the year 2026 [1]. For this reason, there has been increased focus on alternative
One or more of the authors of this paper have disclosed potential or pertinent conflicts of interest, which may include receipt of payment, either direct or indirect, institutional support, or association with an entity in the biomedical field which may be perceived to have potential conflict of interest with this work. For full disclosure statements refer to https://doi.org/10.1016/j.arth.2019.01.050. Source of Funding: No funding was utilized or required to perform this study at our institution. * Reprint requests: Andrew M. Pepper, MD, Andrews Institute for Orthopaedics and Sports Medicine, Baptist Medical Park - Airport, 5100 North 12th Avenue, Suite 102, Pensacola, FL 32504. https://doi.org/10.1016/j.arth.2019.01.050 0883-5403/© 2019 Elsevier Inc. All rights reserved.
payment models (APMs) to attempt to mitigate the costs of providing care in the United States. One such model is the Bundled Payment Care Improvement (BPCI) initiative, which aims to improve the quality of patient care while simultaneously mitigating the associated cost [2]. The BPCI initiative was created out of the Affordable Care Act with the distinct goal to “test innovative payment and service delivery models” while reducing the expenditures of Medicare/Medicaid and ultimately improve the quality of patient care [2e4]. Intense focus on the cost of total joint arthroplasty (TJA) is due to the large proportion of the Centers for Medicare and Medicaid Services’ expenditure contribution attributed to total knee and hip arthroplasty (TKA/THA), with projections that the population’s need for TJA will increase exponentially in the future [5]. The cost for care in the BPCI initiative is determined by a variety of factors [3]. In early models of the BPCI initiative, cost of care
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
considered are: historical cost data, regional data supplements to unavailable historical cost data, episode costs were trended compared to national episode-specific growth rates, with discounts applied depending on model participation [3]. Although many multiple considerations contribute to cost determination in this payment model, there is no episode cost stratification based on individual patient demographics, including age or patient-specific risk factors [4]. Frailty is defined as a decreased physiological reserve and it typically increases with increasing age, but is considered to be a measure of overall health status independent of physiologic age [6e8]. Frailty has been demonstrated as an effective tool in predicting perioperative risk for multiple surgical subspecialties [9e12]. The modified frailty index (mFI) is a simple tool that utilizes a subset of standardized patient parameters that are commonly collected during standard patient perioperative encounters from the original Canadian Study of Health and Aging Frailty Index [13]. mFI consists of 11 patient comorbid variables: (1) nonindependent functional status (partially or totally dependent activities of daily living); (2) history of no diabetes or diabetes controlled by diet alone, diabetes treated with oral antihyperglycemic therapy, or diabetes treated with insulin therapy; (3) history of chronic obstructive pulmonary disease exacerbation or pneumonia within 30 days; (4) history of congestive heart failure exacerbation within 30 days; (5) history of myocardial infarction within 6 months; (6) history of angina within 30 days or any percutaneous coronary intervention or coronary artery bypass grafting; (7) hypertension requiring medication; (8) history of peripheral vascular disease; (9) acutely impaired sensorium; (10) transient ischemic attack or cerebrovascular accident without deficits; and (11) cerebrovascular accident with deficits. Each variable was scored dichotomously and the final mFI was the sum of present variables, for simple interpretation. The mFI tool has proven to be accurate in predicting adverse perioperative outcomes in a variety of surgical subspecialties and in orthopedics specifically, it is a reliable tool for risk stratifying patients undergoing TJA [7,12]. To our knowledge, there is no information in the literature evaluating how mFI affects cost of care in TJA. The aim of this study is to evaluate the effect that both patient age and mFI score have on the cost of care for TKA and THA and this cost’s effect in a BPCI initiative model. Our hypothesis was that increasing age and mFI score would result in increased episode cost. Materials and Methods After Institutional Review Board approval, a retrospective review of prospectively collected data of Medicare patients undergoing TJA (billing code Diagnosis-Related Groups 469 and 470) at our institution under the BPCI initiative from 2013 to 2016 was analyzed. Demographic data were collected for each patient, including age and medical history. mFI was calculated for each patient based on review of his or her medical history. Cost of care was obtained for each patient, in reference to BPCI initiative data available at our institution. Patients who underwent THA for a hip fracture, had incomplete data that made calculating mFI impossible, or had incomplete cost data were excluded from analysis. One thousand eight hundred twenty-one patients were included in the final analysis. When age was treated as a categorical variable, the median age of our sample population (72 years) was used as the cutoff. All cost data are presented as a percentage and dollar cost compared to the institution-specific reference standard for BPCI initiative at our institution. Positive values indicate profit to the healthcare system in the BPCI initiative; negative values indicate deficit to the healthcare system in the BPCI initiative.
S81
Table 1 Comparison of Cost by Age Groups and mFI Score. Variable
n (%)
% Difference From Targeta/US Dollars (SD)
Overall Age group <72 72 mFI score 0 1 2 3 4 5
1821 (100%)
e
P-Value
<.01 872 (47.9) 949 (52.1)
þ19.2%/þ$6120 (32.3%) þ10.6%/þ$3391 (39.5%) <.01
558 767 368 92 26 6
(30.7) (42.2) (20.3) (5.1) (1.4) (0.3)
þ21.2%/þ$6752 þ16.8%/þ$5369 þ10.4%/þ$3334 10.5%/$3341 14.1%/$4491 47.9%/$15,268
(31.5%) (30.0%) (33.9%) (66.1%) (74.2%) (107.8%)
mFI, modified frailty index. P < .05 represents statistically significant findings, bolded. a Positive values represent percentage of decreased cost of care/US dollar profit compared to the institutional reference standard; negative values represent percentage of increased cost of care/US dollar increase cost compared to the institutional reference standard.
All statistical analyses were performed using SPSS software (IBM SPSS). Statistical analyses included t-test and analysis of variance to evaluate age and frailty as independent categorical variables. Beta coefficients were utilized to evaluate age as a continuous variable. Multivariate linear regression models evaluated age and frailty’s combined contribution to cost. Results When evaluating age as an independent categorical variable, our data revealed that age 72 years was associated with increased cost difference of þ8.6% compared to patients <72 years (P < .01) (Table 1 and Fig. 1). Increasing mFI score as an independent variable was also associated with increased cost (P < .01) per care episode (Table 1). Positive values represent percentage of decreased cost of care/US dollar profit, whereas negative values represent percentage of increased cost of care/US dollar increase cost. When age was evaluated as an independent continuous variable, our data revealed an increase cost per care episode of 0.68%, or $217, for each incremental year increase in age (95% confidence interval 0.47-0.89). In a multivariate linear regression model evaluating both age and mFI score, the data revealed statistically significant (P < .01) incremental increasing cost of care for both continuously increasing age and mFI score, except for an mFI score of 1 (P ¼ 0.06) (Table 2 and Fig. 2). Discussion Our data from the BPCI initiative at our institution, from inception of BPCI in 2013 through the most recent complete set of data for 2016, reveal that both increasing age and mFI score have a significant effect on the cost of care per episode for TKA and THA. This effect is evident for each variable independently and in multivariate analyses. The effect of increasing age and mFI score on the cost of care per episode correlates to decreasing reimbursement in a BPCI initiative model. Our findings are the first to directly evaluate the effects of simple patient factors, such as age and mFI, on episode cost in a BPCI initiative. Currently, the BPCI initiative does not take into consideration patient age and modifiable and nonmodifiable risk factors into determining payment for TJA. Although the goal of the BPCI initiative is to decrease cost and improve the quality of care, the current pricing scheme over-simplifies the cost associated with TJA. Further research in this area and refinement of pricing metrics is advisable to better estimate the cost associated with TJA and
S82
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
Fig. 1. Comparison of cost by age groups and mFI score. Positive values represent US dollar profit compared to the institutional reference standard; negative values represent US dollar increase cost compared to the institutional reference standard. P < .05. mFI, modified frailty index.
continue to drive value and quality of care in the US healthcare system. Our data should be evaluated cautiously. We recognize that it is a limited data set and was analyzed in a retrospective manner, and therefore should not be taken as a definitive evaluation of cost implications related to increasing age and mFI. These data should especially be carefully utilized, as it could easily be used as validation for deferring care to older and/or more frail patients when care is provided in a BPCI initiative or Comprehensive Care for Joint Replacement model. Advanced age and increasing frailty are not a contraindication to TKA and/or THA, as the literature supports TJA as an exceptional treatment for decreasing pain, improving function, and increasing the quality of life of patients with advanced hip and knee osteoarthritis [14]. It is advisable that we avoid institutional cost discrimination for elderly/frail patients in building and implementing APMs, as this can inadvertently lead to care discrimination for patients who are still well indicated for TJA. These findings also raise questions of regional and geographical care distribution and care access equity for patients seeking TJA. As the current BPCI/Comprehensive Care for Joint Replacement models are constructed, the potential for smaller, rural, low-volume arthroplasty centers to absorb the additional costs demonstrated in our data may inadvertently deincentivize equitable care for higher risk patients at these locations, as these patients’ care may be especially cost prohibitive. Alternatively, if these patients are diverted to larger, tertiary care centers, this also can potentially concentrate cost prohibitive inequitable care to larger urban academic centers, causing further volume and cost strain on tertiary care centers. Additionally, this could limit geographic access to care
requiring patients at higher risk to travel long distances for care and limit their access to postoperative care. The greater question raised by this discussion is the structure of bundled payment models, which assume cost of care as a fixed value based on over-simplified estimates of cost of care for TJA. This ignores the fact that patients are, by nature, not equivalent and that appropriate care requires an individualized approach to achieve consistent quality of care [15e17]. This includes careful thorough perioperative evaluation, risk factor identification and optimization, surgical planning, varying degrees of surgical technique/equipment for increasing complexity, and differing implant selection [18]. Without addressing the lack of patient-specific cost consideration in current APMs, we are destined to develop institutionalized discrimination in care quality and access, specifically for patients who require the expertise of well-trained arthroplasty surgeons. This situation amounts to national institutionalized gain-sharing which has the potential to perpetuate care discrimination [19]. This study is not without limitations. The study was designed as a retrospective review of cost data from our institution’s BPCI initiative database. Our data are not appropriately powered to detect subtle differences in cost attributable solely to age and frailty. Additionally, there was no control group to compare directly to our study cohort. The range of ages was limited to those patients in our cohort, and therefore a complete evaluation of age ranges is not provided. Similarly, mFI scores ranged from 0 to 5 and we therefore are unable to comment on higher mFI scores. Also, our higher mFI scores had lower numbers of patients and are underpowered to make definitive conclusions regarding the effect on cost. As an early participant in the BPCI initiative, our institution concurrently focused on perioperative optimization of patients known to be a higher risk of perioperative complications, and this is a potential source of confounding. Studies comparing the effect of optimization on improving perioperative outcomes are diffuse in the literature and have been established as the standard of care for patients undergoing TJA [15,20,21]. Finally, we did not collect data regarding the cause of cost differences related to age and frailty (ie, prolonged length of stay, postdischarge disposition, etc.) and cannot therefore comment on the specific contributors to increased or decreased episode cost. The strengths of our study are the discreet reliable cost information available from our institutions BPCI initiative database. The data we present are also generalizable, as our patients undergoing standardized perioperative optimization, which have been established as the standard of care for patients undergoing TJA [15,20,21]. This method is consistent with the protocols that are generally adopted by hospitals across the nation. Our data are also
Table 2 Multivariable Linear Regression of Cost of Care Predicted by Age (Continuous) and mFI. Variable Age (continuous) mFI score 0 (reference) 1 2 3 4 5
Beta Coefficienta/US Dollars (95% CI) 0.59%/$188 (0.38-0.80) e 3.3%/$1047 9.7%/$3096 29.9%/$9526 33.8%/$10,774 67.4%/$21,483
(0.55 to 7.1) (5.1-14.3) (22.1-37.6) (20.0-47.6) (39.2-95.6)
P-Value <.01 e .09 <.01 <.01 <.01 <.01
CI, confidence interval; mFI, modified frailty index. P <.05 represents statistically significant findings, bolded. a Beta coefficient represents the increase in US dollar cost (as percentage of the institutional reference standard) for every incremental increase in age (in year increments) associated with mFI score.
Fig. 2. Multivariable linear regression of cost of care predicted by age (continuous) and mFI. Beta coefficient represents the increase in US dollar cost (compared to the difference from the institutional reference standard) for every incremental increase in age (in year increments) associated with mFI score. P < .05 represents statistically significant findings. mFI, modified frailty index.
A.M. Pepper et al. / The Journal of Arthroplasty 34 (2019) S80eS83
some of the earliest cost data available, as our institution was an early participant in the BPCI initiative at its outset in 2013. Conclusions We demonstrate in this study that simple patient factors such as age and frailty significantly affect the cost of care and reimbursement in a BPCI initiative for TKA and THA. At present, there is no patient-specific cost consideration to account for these differences in current bundled payment models. This information demonstrates that further study is required to better elucidate cost considerations for arthroplasty patients. The authors also highlight the potential implications of fixed-fee payment models and advocate for equitable access to arthroplasty care without penalizing surgeons and/or institutions who are dedicated to providing quality care to all patients. References [1] National healthcare expenditure projections, 2017-2026. https://www cmsgov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/ NationalHealthExpendData/NationalHealthAccountsProjected.html; 2011 [accessed 9.7.18]. [2] Bundled payments for care improvement (BPCI) initiative: general information. https://innovationcmsgov/initiatives/bundled-payments/ [accessed 09.07.18]. [3] Bundled payments for care improvement initiative (BPCI). https://www cmsgov/Newsroom/MediaReleaseDatabase/Fact-sheets/2016-Fact-sheetsitems/2016-04-18html [accessed 09.07.18]. [4] Press MJ, Rajkumar R, Conway PH. Medicare’s new bundled payments: design, strategy, and evolution. JAMA 2016;315:131e2. [5] Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89:780e5. [6] Farhat JS, Velanovich V, Falvo AJ, Horst HM, Swartz A, Patton JH, et al. Are the frail destined to fail? Frailty index as predictor of surgical morbidity and mortality in the elderly. J Trauma Acute Care Surg 2012;72:1526e30. [7] Shin JI, Keswani A, Lovy AJ, Moucha CS. Simplified frailty index as a predictor of adverse outcomes in total hip and knee arthroplasty. J Arthroplasty 2016;31:2389e94.
S83
[8] Kulminski AM, Ukraintseva SV, Kulminskaya IV, Arbeev KG, Land K, Yashin AI. Cumulative deficits better characterize susceptibility to death in elderly people than phenotypic frailty: lessons from the cardiovascular health study. J Am Geriatr Soc 2008;56:898e903. [9] Bellamy JL, Runner RP, Vu CCL, Schenker ML, Bradbury TL, Roberson JR. Modified frailty index is an effective risk assessment tool in primary total hip arthroplasty. J Arthroplasty 2017;32:2963e8. [10] Runner RP, Bellamy JL, Vu CCL, Erens GA, Schenker ML, Guild GN. Modified frailty index is an effective risk assessment tool in primary total knee arthroplasty. J Arthroplasty 2017;32:S177e82. [11] Patel KV, Brennan KL, Brennan ML, Jupiter DC, Shar A, Davis ML. Association of a modified frailty index with mortality after femoral neck fracture in patients aged 60 years and older. Clin Orthop Relat Res 2014;472:1010e7. [12] Wahl TS, Graham LA, Hawn MT, Richman J, Hollis RH, Jones CE, et al. Association of the modified frailty index with 30-day surgical readmission. JAMA Surg 2017;152:749e57. [13] Rockwood K, Song X, Macknight C, Bergman H, Hogan DB, Mcdowell I, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ 2005;173:489e95. [14] Mont MA, Tanzer M. Orthopaedic knowledge update: hip and knee reconstruction 5. Am Acad Orthop Surg 2017;173:489e95. Chapter 18, Chapter 30. [15] Chimento GF, Thomas LC. The perioperative surgical home: improving the value and quality of care in total joint replacement. Curr Rev Musculoskelet Med 2017;10:365e9. [16] Clement RC, Soo AE, Kheir MM, Derman PB, Flynn DN, Levin LS, et al. What incentives are created by Medicare payments for total hip arthroplasty? J Arthroplasty 2016;31(9 Suppl):69e72. [17] Magone K, Kemker 3rd BP, Pilipenko N, O’Connor E, Walter N, Atkinson T. The new surgical technique for improving total knee and hip arthroplasty outcomes: patient selection. J Arthroplasty 2017;32:2070e6. [18] Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR, Manley MT. Which clinical and patient factors influence the national economic burden of hospital readmissions after total joint arthroplasty? Clin Orthop Relat Res 2017;475: 2926e37. [19] Ramkumar PN. How bundling care hurts the patients who need it the most. Forbes Magazine; 2018. https://www.forbes.com/sites/premramkumar/2018/ 07/11/how-bundling-care-hurts-the-patients-who-need-it-the-most/#42e3efeb 3250 [accessed 13.07.18]. [20] Boraiah S, Joo L, Inneh IA, Rathod P, Meftah M, Band P, et al. Management of modifiable risk factors prior to primary hip and knee arthroplasty: a readmission risk assessment tool. J Bone Joint Surg Am 2015;97:1921e8. [21] Paxton EW, Inacio MC, Singh JA, Love R, Bini SA, Namba RS. Are there modifiable risk factors for hospital readmission after total hip arthroplasty in a US healthcare system? Clin Orthop Relat Res 2015;473:3446e55.