Mortality, Cost, and Downstream Disease of Total Hip Arthroplasty Patients in the Medicare Population

The Journal of Arthroplasty 29 (2014) 242–246

Contents lists available at ScienceDirect

The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org

Mortality, Cost, and Downstream Disease of Total Hip Arthroplasty Patients in the Medicare Population Scott T. Lovald, PhD, MBA a, Kevin L. Ong, PhD b, Edmund C. Lau, MS a, Jordana K. Schmier, MA c, Kevin J. Bozic, MD, MBA d, Steve M. Kurtz, PhD b a b c d

Exponent, Inc., Menlo Park, California Exponent, Inc., Philadelphia, Pennsylvania Exponent, Inc., Alexandria, Virginia Department of Orthopaedic Surgery, University of California San Francisco, San Francisco, California

a r t i c l e

i n f o

Article history: Received 5 February 2013 Accepted 17 April 2013 Keywords: downstream disease total hip arthroplasty Medicare mortality

a b s t r a c t The purpose of this study is to compare the differences in downstream cost and health outcomes between Medicare hip OA patients who undergo total hip arthroplasty (THA) and those who do not. All OA patients in the Medicare 5% sample (1998–2009) were separated into non-THA and THA groups. Differences in costs and risk ratios for mortality and new disease diagnoses were adjusted using logistic regression for age, sex, race, socioeconomic status, region, and Charlson score. Mortality, heart failure, depression, and diabetes were all reduced in the THA group, though there was an increased risk for atherosclerosis in the short term. The potential for selection bias was investigated with two separate propensity score analyses. This study demonstrates the potential benefit of THA in reducing mortality and improving aspects of overall health in OA patients. © 2014 Elsevier Inc. All rights reserved.

It is well accepted that total hip arthroplasty (THA) leads to significant improvements in pain, functioning and quality of life [1]. Despite this, not all patients who are appropriate for the procedure are treated with a THA [2]. Furthermore, there are wide variations in the utilization of THA throughout the country [3]. Due to the degenerative nature of OA, patients delaying surgery will likely become less active with increasing pain, likely compounding disability associated with the disease. Inactivity associated with OA can lead to the loss of function, disability, increased risk of cardiovascular disease, and a reduced quality of life [4]. It is hypothesized that pain relief and restored function can have a positive effect on a patient's overall health. Patients experiencing less pain are likely to be more active, thus potentially reducing the risk of cardiovascular and other diseases [4-6]. It has been suggested that improved mobility and function, along with reduced pain, may also lead to a reduced mortality risk in the long term [7]. Several studies

Study design, manuscript preparation, and review occurred at the offices of each author independently. The initial study design and data processing took place at Exponent, Inc., in Philadelphia, PA, and Menlo Park, CA. All statistical analyses took place at Exponent, Inc., in Menlo Park, CA. The Conflict of Interest statement associated with this article can be found at http:// dx.doi.org/10.1016/j.arth.2013.04.031. Reprint requests: Scott T. Lovald, PhD, MBA, Exponent, Inc. 149 Commonwealth Drive, Menlo Park, CA 94025. 0883-5403/2901-0048$36.00/0 – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.arth.2013.04.031

have shown arthroplasty patients to have a reduced mortality risk compared to the general population, although the exact reasons remain unclear [8-13]. While the short-term benefits of surgery are well known and longterm benefits have been proposed, there are few data that quantify the long-term cost, mortality, and burden of downstream disease associated with hip osteoarthritis. The purpose of the present study was to determine the differences in health outcomes, including mortality and new disease diagnoses, between patients with hip OA who underwent THA and patients with hip OA who did not receive a THA. A secondary goal of this study is to determine the difference in health costs over an extended postoperative period between THA and non-THA patients. Patients and Methods The Medicare 5% Limited Data Set (LDS) sample was used to identify patients with OA of the hip during the period of 1998 to 2009. Hip osteoarthritis patients were identified using ICD-9 codes within the 715.x5 family. This included patients with localized, primary or secondary hip OA. The OA code was required to be the principal diagnosis and occur more than once during a 12-month period to be included in the study. Patients were recruited continuously during the study period and followed until the end of the study period (December 31, 2009), until their benefits were terminated, or until death. Exclusions included hemiarthroplasty patients, those who

S.T. Lovald et al. / The Journal of Arthroplasty 29 (2014) 242–246

received benefits for a reason other than age (i.e., end-stage renal disease or disability) or Medicare beneficiaries who received their care through the Medicare Advantage (Part C) program. Patients' enrollment status, age, and date of death were tracked using the matching 1998–2009 Medicare denominator files. OA patients were separated into arthroplasty and non-arthroplasty groups using the presence of ICD-9 code 81.51 for the hip arthroplasty group. The arthroplasty group was allowed to have received their THA at any point after their OA diagnosis. The arthroplasty procedure was confirmed by the appropriate CPT4 code (27130) in the Part B (Physician Services) portion of the database. To normalize the comparison, only newly OA diagnosed patients were considered in the analysis. To ensure this, patients must have had an entire year without a hip OA diagnosis prior to the index OA diagnosis. Outcomes of interest included average annual payments (payer cost, adjusted to January 2011 US dollars), mortality, and diagnoses of ischemic heart disease and atherosclerosis (410.xx–414.xx, 440.xx), cardiovascular disease unspecified (429.2), heart failure (428.x, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93), heart disease unspecified (428.8, 429.9), diabetes (.250.xx), and depression (ICD-9 309.xx and 311.xx). Average medical costs were quantified over each period for patients who were fully observable for the entire period (i.e. they had the full period of follow-up and did not die or withdraw from Medicare). Differences in risk ratios between arthroplasty and non-arthroplasty groups for each outcome were adjusted using logistic regression, controlling for age, sex, region of residence, comorbidities (Charlson comorbidity index), socioeconomic status (using the Medicare buy-in status as surrogate), and previous disease diagnoses. The Charlson index quantifies the presence of comorbid conditions into a single score, and has been determined to be a valid method for estimating the risk of death from comorbid disease [14]. The Medicare buy-in status identifies patients whose Medicare premiums and deductibles were subsidized by the state. Although considered a valuable data point, OA severity data are not included in the database and thus were not accounted for in the model. A generalized linear model comparing costs for each group adjusted for the same factors. Year of diagnosis was treated as a covariate in the analyses, testing whether the treatment effects and costs were changing over time. Costs included all direct health care payments covered under Medicare Parts A and B, including those payments classified under Inpatient, Skilled Nursing, Hospice, Outpatient, Physician Services, Home Health Agency Services, and Durable Medical Equipment. The following cost metrics were considered: all costs (for any diagnosis) and those that included only arthritis-attributable (with OA as the principal diagnosis). Medicare costs, mortality, and disease diagnoses were obtained at 1, 3, 5, and 7 years after OA diagnosis and compared for both groups. In an attempt to balance the study cohorts and minimize selection bias, two separate propensity score analyses were conducted to investigate the sensitivity of the outcomes to a patient's likelihood for surgery. A propensity score was calculated using a logistic regression model with patient demographic characteristics and morbidity conditions as independent variables. The propensity score represents the likelihood (i.e., “propensity”) of receiving a THA (or no THA) treatment given a particular combination of patient characteristics and morbidity conditions. Patients having similar scores are considered “comparable.” The propensity score approach was conducted in two ways: (a) using the score itself as an adjustment factor (covariate) in subsequent Cox regression analyses and (b) using the score as a metric for matching THA and non-THA patients to form sub-samples of matched cases and controls to be used in stratified Cox regressions. In either form, the Cox regression remains the outcome or “response” model to access differences in survivorship.

243

Results The study included 43,369 patients with at least 1 year of followup. Of these 19,706 received a THA, while the remaining 23,663 patients did not have a THA. The number of patients with 7-year follow-up was 13,443 in the non-THA and 10,523 in the THA group. Patients in the THA group were younger and in better general health (Fig. 1). Survival analysis curves (indicating the time between diagnosis of OA and surgery) show that 40% of OA patients receive a THA within the first 24 months after OA diagnosis (Fig. 2). The curves represent the percentage of patients with OA who received a THA at a particular time point. The rate at which patients receive a THA tapered slightly with 50% of patients having been treated with a THA at 77 months after their OA diagnosis. Based on the examination of survival curves, it appears that patients received a THA more quickly after the OA diagnosis in the latter years of the study. The mortality hazard ratio (adjusted) of the THA group ranged from 0.26 to 0.52 through 7 years (all P b 0.001) (Table 1). The risk of heart failure in the THA group was 23% at 3 years (adjusted HR = 0.85, P b 0.001) and 43% at 7 years (adjusted HR = 0.92, P = 0.005). There was a lower risk of depression for the THA group at 3 years (rate =18.2%, adjusted HR = 0.86, P b 0.0001) to 7 years (rate = 31.8%, adjusted HR = 0.90, P = 0.0003), and a lower risk for diabetes up to 3 years (adjusted HR = 0.91–0.93, P b 0.02) There was a higher risk for ischemic heart disease/ atherosclerosis (rate = 23.4%, adjusted HR = 1.14, P b 0.0001) at 1 year, though at no time after. Cardiovascular disease (unspecified) was higher in the THA group at all follow-up points (rate = 4.9% to 15.7%, adjusted HR = 1.10–1.28, P b 0.02). Results for the standard analysis, the propensity score-adjusted analysis, and the propensity score-matched analysis in Table 2 show that the results were insensitive to our efforts to balance the study and reduce bias using two types of propensity score analyses. At 1 year, the average cumulative payment was $19,274 for the non-THA group and $29,840 for the THA group, an incremental cost of $10,566 (Fig. 3). The 7-year cumulative average Medicare payments for all treatments for all medical care were $82,788 for the non-THA group and $89,154 for the THA group, an incremental 7-year cost of $6366.

Discussion Our study is the first attempt we are aware of to evaluate cost and disease risk for Medicare patients with osteoarthritis (OA) of the hip. The study was a broad population-based observational analysis across a patient cohort that constitutes a significant burden of osteoarthritis on the U.S. healthcare system. A goal of the study was to observe whether OA patients treated with hip arthroplasty will realize a greater overall long-term health benefit. Twenty-one million Americans have arthritis-attributable activity limitation [15]. As OA is degenerative, it is believed that there is a point when it is time to move from conservative care to surgery [16]. If this point is never reached, patients with increasing pain become less active, limiting their daily exercise. Physical activity is an important part of the prevention of disability associated with OA. Inactivity associated with OA can lead to the loss of function, disability, increased risk of cardiovascular disease, and a reduced quality of life [4]. A recent study associated fatigue with pain from OA and pain medications in a small sample of patients, and further noted that this had an effect on the mental health, physical functioning, and daily activities of OA sufferers [5]. The current study examined the downstream effect that THA may have on improved hip function and cardiovascular fitness by examining mortality and new diagnoses of cardiovascular conditions. The consistently reduced risk of mortality, heart failure, and diabetes, suggests that increased activity after THA may

244

S.T. Lovald et al. / The Journal of Arthroplasty 29 (2014) 242–246

Fig. 1. The distribution of patient age, gender, race and Charlson Index is shown for the THA and non-THA groups.

have a beneficial effect in reducing long-term disease and death among OA patients. Limitations of the current study are in line with other retrospective database reviews. A major drawback of using the Medicare database is that claims data lack specific details, including indices for patient pain and functioning. While a generalized comparison of THA and non-THA groups is a suitable analysis for this database, it is difficult to discern why some patients elected to forego arthroplasty procedures. Also confounding the data is that comorbid illnesses are suspected to be under-coded. Due to this, the methodology may miss identifying all comorbidities in patients who were not selected for surgery. Further, there is the potential that patients with mild OA were not escalated to an OA diagnosis (715.x5) and were instead designated an alternate code, for example hip pain (719.15). Although this potentially excludes a segment of the early OA population which would likely

Fig. 2. Survival analysis curves are shown for time between OA diagnosis and surgery. The curves represent the percentage of patients with OA who receive a THA at a particular time point.

be treated conservatively, the intent of the study was to directly compare patients in each group who were on the same “level” of disease according to their diagnosis. Further, patients with potentially non-OA diagnoses who eventually received an OA diagnosis during the later years of the study were included, limiting the excluded group to those whose condition never graduated into the diagnosed “OA” category through the entire study. Despite the drawbacks, the Medicare database primary advantage is that it is a very large patient sample, generalizable to the US elderly population. Most studies on outcomes for arthroplasty patients tend to be limited to one or a few surgeons and hospitals, and therefore have limited value for generalization [17]. Medicare patients also represent a very significant proportion of arthroplasty patients. It has been reported that 66.4% of total arthritis-related surgeries are performed in patients 65 years or older, while over half of people over the age of 65 years are estimated to have arthritis [18]. Our study determined that patients with hip OA treated with a THA have a consistently lower mortality risk of 26% to 52% of the nonTHA group over the full study period. Heart failure was similar between groups in the first year, but there was a consistent reduced risk (0.85–0.92, P b 0.005) at 3 to 7 years after the surgery. There was a reduced risk of diabetes at 1 and 3 years and a reduction of depression diagnoses starting at 3 years. On the other hand, there is an increased risk of Ischemic Heart Disease and Atherosclerosis at 1 year, as well as an increased risk of “Cardiovascular Disease Unspecified” over all time points that were considered. Early deaths from THA have previously been reported from cardiopulmonary arrest, acute coronary syndrome, stroke, pulmonary embolism, arrhythmias, respiratory complications, malignancy, and sepsis [19]. Data from the Rochester epidemiology project and the Mayo Clinic total joint registry demonstrated a 6.9% rate of all cardiac events at 90 days after THA, including congestive heart failure (2.7%), myocardial infarction (1.2%), and arrhythmia (4.8%). The rate for congestive heart failure, and other diseases was higher in the present study, though initial follow-up was over a much longer period and the

S.T. Lovald et al. / The Journal of Arthroplasty 29 (2014) 242–246

245

Table 1 Adjusted Hazard Ratios for Mortality and New Disease Diagnoses After OA. Years After Surgery 1y Rate Rate (%) (%) (No Hazard (THA) THA) Ratio Mortality Heart failure Ischemic heart disease and atherosclerosis Cardiovascular disease unspecified Heart disease unspecified Diabetes Major or minor depression

3y

P

Rate Rate (%) (%) (No (THA) THA)

5y

Hazard Ratio

P

Rate Rate (%) (%) (No (THA) THA)

7y

Hazard Ratio

P

Rate Rate (%) (%) (No Hazard (THA) THA) Ratio

P

0.9 12.3 23.4

5.4 19.4 27.2

0.257 0.996 1.139

b0.0001 0.902 b0.0001

5.7 23.1 40.1

19.7 35.0 47.8

0.386 0.851 0.971

b0.0001 b0.0001 0.221

14.3 32.9 53.0

36.9 46.4 60.8

0.446 0.854 0.974

b0.0001 b0.0001 0.316

28.8 42.9 63.8

56.3 55.5 70.1

0.523 0.919 0.998

b0.0001 0.005 0.950

4.9

5.0

1.278

b0.0001

9.0

9.9

1.099

0.020

12.6

13.0

1.134

0.001

15.7

15.5

1.146

0.0007

2.6

2.9

1.088

0.190

5.4

6.5

0.947

0.262

8.0

9.2

0.970

0.501

10.4

11.1

1.019

0.688

20.3 10.4

26.8 13.5

0.914 1.022

0.010 0.527

29.3 18.2

35.7 25.0

0.930 0.864

0.017 b0.0001

35.6 25.0

41.0 32.8

0.974 0.857

0.385 b0.0001

40.9 31.8

44.6 38.7

1.005 0.895

0.882 0.0003

The rate for each diagnosis is reported for the THA and non-THA groups.

and the use of anti-inflammatory drugs [11]. Concerning patient selection, arthroplasty surgery may exclude patients with concomitant comorbidities [16], including cardiac and other medical conditions [12]. Access to care for minority patients and those of lower social status may also play a role [25]. Our study sought to minimize this bias by controlling for disease using the Charlson Comorbidity Index, allowing an appropriate comparison with patients of similar general health [11]. Previous studies have included a control population loosely controlled for some demographic factors but have not included patients with the same disease. To further account for potential selection bias between study cohorts, the study methodology included performing two types of propensity score analyses (adjustment and matching), both of which showed minimal differences in comparison to the standard analysis. The current study used non-treated OA Medicare patients as the reference group. By including control patients with OA, and not the general population, it is ensured that patients have free access and the means to utilize healthcare through Medicare, and a demonstration that they are concerned enough with their condition to be diagnosed with OA. Also, it is plausible that both groups in the current study may have utilized medications such as anti-inflammatory drugs, which may influence survivorship of the patients and the development of comorbidities, but this could not be confirmed in the available data set. Direct medical costs for primary THR have been previously reported to be between $10,732 and $13,339 [26,27]. The total direct 1-year incremental cost of $10,566, identified for the THA group in this study, agrees well with previous findings. This incremental cost, however, does not take into account the expense of prescription drugs, which have been reported to be much higher in non-arthroplasty

patient population is older and more susceptible to disease than that reported in previous studies. A number of previous studies have examined the mortality risk of hip arthroplasty in the short term (up to 90 days after surgery) or compared to the general population in the long term. Short-term outcomes and mortality after THA have been reported in localized patient samples, typically from specialist centers [20]. Studies of this type have reported 30-day mortality from 0.24% to 0.29% [19,21] and 90-day mortality from 0.3% to 0.7% [1,19,22]. In line with these previous studies, the current study reports a 1-year mortality of 0.85% in the THA population. Medium- to long-term mortality rates previously reported also compare well to those found in the current work. Studies from Norway, Sweden, and Finland reported mortality rates of 19% to 27% in THA patients with OA from 6.2 to 8 years after surgery [11,13,23]. When compared to age- and gender-matched general population, hazard ratios were reported from 0.68 to 0.73. In a previous study of US Medicare patients, those undergoing THA were reported to have higher 5-year survival when compared to matched (age, sex, race, and income) controls [24]. Despite this evidence, no previous study has studied mortality and new disease diagnoses in the THA population compared to OA patients without the treatment. The long-term health benefit after arthroplasty has been partially attributed to the possibility that patients receiving surgery are generally healthier than those who decide against surgery or deemed not to be candidates for surgery [7]. It should be noted that it has been reported that the mortality benefit of joint arthroplasty does not extend to all patient populations, including those with rheumatoid arthritis, younger patients, and those with a femoral neck fracture [11,12,23]. Specific factors speculated to affect mortality include the preoperative patient selection of healthier patients, patient activity,

Table 2 Hazard Ratios Are Compared for the Standard, Propensity Score Adjustment, and Propensity Score Matching Analyses. 7 y After Surgery

Mortality Heart failure Ischemic heart disease and atherosclerosis Cardiovascular disease unspecified Heart disease unspecified Diabetes Major or minor depression

Standard Analysis

Propensity Score Adjusted

Propensity Score Matched

Rate (%)

Adjusted Hazard Ratio

P

Adjusted Hazard Ratio

P

Adjusted Hazard Ratio

P

28.8 42.9 63.8 15.7 10.4 40.9 31.8

0.52 0.92 1.00 1.15 1.02 1.01 0.90

b0.0001 0.0050 0.9502 0.0007 0.6876 0.8823 0.0003

0.52 0.92 1.00 1.15 1.02 1.00 0.90

b0.0001 0.0059 0.9978 0.0006 0.7349 0.8938 0.0005

0.52 0.92 1.00 1.15 1.02 1.00 0.90

b0.0001 0.0065 0.9915 0.0007 0.6899 0.9444 0.0004

The rate for each diagnosis is reported for the THA group.

246

S.T. Lovald et al. / The Journal of Arthroplasty 29 (2014) 242–246

Fig. 3. Cumulative medical costs for both THA and non-THA groups. Costs are categorized for “all treatments,” which include all Medicare payments, and for “OA treatment only,” which only include costs for which osteoarthritis is the principal diagnosis.

patients. One study found the annual cost of prescription drugs to be $1994 in large joint OA patients who avoid arthroplasty, vs. $1261 in arthroplasty patients [28]. Over a 1-year study period, the cumulative difference between these groups would be $5131. Further, indirect costs of inactivity also need to be considered. It has been estimated that the proportion of direct medical costs that are incurred due to inactivity is 12.4% in patients with OA [4]. Similarly, the proportion of medical costs for cardiovascular disease that were directly associated to inactivity was 13.1% [29]. Considering the differences in indirect costs and prescription drug costs between THA and non-THA patients, the 7-year incremental cost of THA treatment is likely much less than the $6366 determined here. For example, factoring in prescription drug costs using the calculations above, the adjusted 7-year cost of a THA would be $1235. Joint arthroplasty has been proposed as possibly cost-saving in the management of disability related to arthritis [2]. While these results stop short of confirming this, they support the conclusion that THA is a “bargain to society” [30]. Despite the aforementioned limitations and the higher reported incidence of early mortality and affective disorders among patients undergoing THA, the study has demonstrated that THA confers a potential long-term benefit in terms of prolonged lifespan and reduced burden of disease in Medicare patients with OA of the hip. References 1. Singh JA, Jensen MR, Harmsen WS, et al. Cardiac and thromboembolic complications and mortality in patients undergoing total hip and total knee arthroplasty. Ann Rheum Dis 2011;70(12):2082.

2. Santaguida PL, Hawker GA, Hudak PL, et al. Patient characteristics affecting the prognosis of total hip and knee joint arthroplasty: a systematic review. Can J Surg 2008;51(6):428. 3. http://www.dartmouthatlas.org/. 4. Wang G, Helmick CG, Macera C, et al. Inactivity-associated medical costs among US adults with arthritis. Arthritis Rheum 2001;45(5):439. 5. Power JD, Badley EM, French MR, et al. Fatigue in osteoarthritis: a qualitative study. BMC Musculoskelet Disord 2008;9:63. 6. Ries MD, Philbin EF, Groff GD, et al. Improvement in cardiovascular fitness after total knee arthroplasty. J Bone Joint Surg Am 1996;78(11):1696. 7. Parry MC, Smith AJ, Blom AW. Early death following primary total knee arthroplasty. J Bone Joint Surg Am 2011;93(10):948. 8. Visuri T, Pulkkinen P, Turula KB, et al. Life expectancy after hip arthroplasty. Case– control study of 1018 cases of primary arthrosis. Acta Orthop Scand 1994;65(1):9. 9. Schroder HM, Kristensen PW, Petersen MB, et al. Patient survival after total knee arthroplasty. 5-Year data in 926 patients. Acta Orthop Scand 1998;69(1):35. 10. Ritter MA, Albohm MJ, Keating EM, et al. Life expectancy after total hip arthroplasty. J Arthroplasty 1998;13(8):874. 11. Paavolainen P, Pukkala E, Pulkkinen P, et al. Causes of death after total hip arthroplasty: a nationwide cohort study with 24,638 patients. J Arthroplasty 2002; 17(3):274. 12. Ohzawa S, Takahara Y, Furumatsu T, et al. Patient survival after total knee arthroplasty. Acta Med Okayama 2001;55(5):295. 13. Garellick G, Malchau H, Herberts P, et al. Life expectancy and cost utility after total hip replacement. Clin Orthop Relat Res 1998;346:141. 14. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40(5):373. 15. Prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation–United States, 2007–2009. MMWR Morb Mortal Wkly Rep 2010;59(39):1261. 16. Gruber WH, Hunter DJ. Transforming osteoarthritis care in an era of health care reform. Clin Geriatr Med 2010;26(3):433. 17. Ethgen O, Bruyere O, Richy F, et al. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86-A(5):963. 18. Kim S, Koebel S, Duffy R. Cost burden of hip and knee replacements in Ohio: estimates from the National Hospital Discharge Survey, 2000. Manag Care Interface 2004;17(8):22. 19. Aynardi M, Pulido L, Parvizi J, et al. Early mortality after modern total hip arthroplasty. Clin Orthop Relat Res 2009;467(1):213. 20. Williams O, Fitzpatrick R, Hajat S, et al. Mortality, morbidity, and 1-year outcomes of primary elective total hip arthroplasty. J Arthroplasty 2002;17(2):165. 21. Parvizi J, Sullivan TA, Trousdale RT, et al. Thirty-day mortality after total knee arthroplasty. J Bone Joint Surg Am 2001;83-A(8):1157. 22. Dearborn JT, Harris WH. Postoperative mortality after total hip arthroplasty. An analysis of deaths after two thousand seven hundred and thirty-six procedures. J Bone Joint Surg Am 1998;80(9):1291. 23. Lie SA, Engesaeter LB, Havelin LI, et al. Mortality after total hip replacement: 0–10year follow-up of 39,543 patients in the Norwegian Arthroplasty Register. Acta Orthop Scand 2000;71(1):19. 24. Barrett J, Losina E, Baron JA, et al. Survival following total hip replacement. J Bone Joint Surg Am 2005;87(9):1965. 25. Nwachukwu BU, Kenny AD, Losina E, et al. Complications for racial and ethnic minority groups after total hip and knee replacement: a review of the literature. J Bone Joint Surg Am 2010;92(2):338. 26. Lavernia CJ, Alcerro JC. Quality of life and cost-effectiveness 1 year after total hip arthroplasty. J Arthroplasty 2011;26(5):705. 27. Antoniou J, Martineau PA, Filion KB, et al. In-hospital cost of total hip arthroplasty in Canada and the United States. J Bone Joint Surg Am 2004;86-A(11):2435. 28. Hawker GA, Badley EM, Croxford R, et al. A population-based nested case–control study of the costs of hip and knee replacement surgery. Med Care 2009;47(7):732. 29. Wang G, Pratt M, Macera CA, et al. Physical activity, cardiovascular disease, and medical expenditures in U.S. adults. Ann Behav Med 2004;28(2):88. 30. Huo MH, Dumont GD, Knight JR, et al. What's new in total hip arthroplasty. J Bone Joint Surg Am 2011;93(20):1944.