The number of total knee arthroplasties performed in the United States more than tripled from 1993 to 2009, while the number of total hip arthroplasties approximately doubled during the same time period1. In the current environment of health-care reform and cost containment, increasing surgical volumes have been the subject of much scrutiny2. A multitude of interacting factors, including obesity, population aging, supplier-induced demand, increases in sports-related injuries, direct-to-consumer advertising, changing expectations about quality of active life, and expanding surgical indications, have been proposed as potential explanations for the increasing prevalence of total joint arthroplasties2-4.
Several observational studies have described this growth in arthroplasties, have quantified its economic ramifications, have projected future trends, and have correlated arthroplasty trends with other observed phenomena2,3,5. This study compiles and analyzes historical data on a variety of clinical, demographic, and economic factors related to the growth in joint replacement procedures. We have concentrated on identifying factors contributing to the faster rate of growth of total knee arthroplasties compared with total hip arthroplasties. The growth in overweight and obesity is of particular interest in this context, as the association between overweight and obesity and the likelihood of developing arthritis appears to be more robust for the knee than the hip6-13.
Materials and Methods
Annual total hip arthroplasty and total knee arthroplasty volumes were compiled and were plotted to determine growth trends. Longitudinal data from a number of sources1,14-17 were analyzed to identify factors that would accelerate the utilization of total knee arthroplasty more than total hip arthroplasty, thus potentially explaining the differential growth rate. Supply-side factors included the size of the surgical workforce, per-surgeon output, surgeon and hospital reimbursement, and length of hospital stay. Demand-side factors were in-hospital mortality (used as a proxy for procedural safety) and population demographics (age and overweight or obesity prevalence).
Operative Volumes, Length of Stay, and In-Hospital Mortality
Annual procedure counts, average length of stay, and in-hospital mortality rates for total knee arthroplasties and total hip arthroplasties performed from 1993 to 2009 were obtained from the Nationwide Inpatient Sample (NIS), a database sponsored by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project (HCUP)1. The most recent NIS data utilized in this study (2009) contain details of approximately eight million admissions at more than 1000 hospitals located in forty-four states.
The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure codes for primary total knee arthroplasty (81.54) and primary total hip arthroplasty (81.51) were used to define relevant procedures. Surgical volumes for each procedure were subdivided into two age groups (eighteen to sixty-four years and sixty-five years or older). Age-stratified procedure volumes were available from 1999 to 2009. Here and throughout the study, the number of years of longitudinal data available in each data source dictated the duration of time that could be represented in subsequent analyses, and 95% confidence intervals (95% CIs) were used whenever possible.
Orthopaedic Surgery Manpower
The number of active fellows in the American Academy of Orthopaedic Surgeons (AAOS), the largest association of musculoskeletal specialists18, was used as a proxy for the size of the U.S. orthopaedic surgery workforce. This information was obtained for all years from 1993 to 2009 from annual AAOS membership data.
The percentage of surgeons specializing in knee and hip arthroplasty was gleaned from data contained in the AAOS member census14. In this survey, surgeons who identified “adult hip” as their main practice area were used to calculate the total number of practicing surgeons involved in total hip arthroplasty, and surgeons who identified “adult knee” as their main practice area were used to calculate the total number of practicing surgeons involved in total knee arthroplasty. Because the census is biennial, linear extrapolation was used to produce estimates for intervening years.
Physician and Hospital Reimbursement
Reimbursement rates were approximated with use of the Medicare payment scale; these rates are fairly constant nationwide and generally guide the fees of private insurers as well. All inputs for calculating Medicare reimbursement are published annually in the Federal Register or made publicly available by the Centers for Medicare & Medicaid Services (CMS)15,16.
Surgeon reimbursement for each of the relevant procedures was determined by multiplying the appropriate number of relative value units by the conversion factor for each year. Current Procedural Terminology (CPT) codes (27130 for total hip arthroplasty and 27447 for total knee arthroplasty) are used to define procedures for the assignment of relative value units. Hospital reimbursement for each year (1995 to 2009) was estimated by multiplying that year’s diagnosis-related group relative weight by the federal payment rate. Both total knee arthroplasty and total hip arthroplasty fall under the same diagnosis-related group (DRG 209 from 1990 to 2005, DRG 544 from 2006 to 2007, and DRG 470 from 2008 to the present). All dollar figures were converted to 2012 dollars using the Consumer Price Index coefficients provided by the U.S. Department of Labor’s Bureau of Labor Statistics19.
Obesity and Overweight Prevalence
Obesity and overweight trends from 1984 to 2010 were estimated using data from the Behavioral Risk Factor Surveillance System, a nationwide cross-sectional health survey run by the Centers for Disease Control and Prevention, which collects data on more than 350,000 adults annually17. Sample data from U.S. territories were excluded, as these were not included in other data sets used in our analyses. Overweight prevalence was defined as the percentage of the population with a body mass index (BMI) of 25 to 29.9 kg/m2 and obesity prevalence was defined as the percentage of the population with a BMI of ≥30 kg/m2. Because the Behavioral Risk Factor Surveillance System is a probabilistic sample survey, sampling weights were used to adjust for unequal sampling probabilities. Data were stratified into two age groups (eighteen to sixty-four years and sixty-five years and older).
We sought to understand the differential growth in total knee arthroplasties and total hip arthroplasties in the context of labor supply. Growth in procedure volume can reflect an increase in the number of procedures that each surgeon performs, an increase in the number of surgeons, or both. To understand the relative contribution of the extensive margins (change in the number of surgeons) compared with the intensive margins (change in the average per-surgeon volume), we used the following decomposition of the annual difference in the total volume of total knee arthroplasties and total hip arthroplasties:
Here total knee arthroplasty (TKA) and total hip arthroplasty (THA) correspond to the annual volumes of such procedures, SK is the number of orthopaedic surgeons reporting adult knee as a major specialty area, and SH is the number of orthopaedic surgeons reporting adult hip as a major specialty area. The term (1) represents the intensive margin effect of changes in per-surgeon volume, weighted by the number of surgeons reporting adult knee as a major specialty area. The term (2) represents the extensive margin effect of changes in surgical manpower, weighted by the average per-surgeon total hip arthroplasty volume. We are able to calculate annual measures of both (1) and (2) terms for each year in our study period. These results were then plotted to illustrate the trend in the relative importance of individual productivity compared with surgeon manpower over time. Here, and whenever possible throughout this study, 95% CIs were calculated and were plotted with use of standard rules for the propagation of uncertainty20,21.
In addition, we evaluated the makeup of the patient population for total knee arthroplasty and total hip arthroplasty in terms of BMI. The number of overweight and obese individuals represented in the difference between total knee arthroplasty and total hip arthroplasty volume was determined using the following formulae:
Here, the term TKA represents the national volume of total knee arthroplasties performed in a given year (t), and the term THA represents the national volume of total hip arthroplasties performed in a given year (t). The variable B represents the percentage of the adult population in each year (t) that is obese, the variable V represents the percentage of the population that is overweight, and the variable N represents the percentage of the population that is underweight or normal weight (those with a BMI of <25 kg/m2). The variable F, with an input of 4.65 (95% CI, 2.70 to 8.10), represents the risk of requiring total knee arthroplasty for obese individuals compared with the risk of requiring it for underweight or normal-weight individuals. The variable G, with an input of 1.65 (95% CI, 1.35 to 2.40), represents the risk of requiring total knee arthroplasty for overweight individuals compared with the risk of requiring it for underweight or normal-weight individuals. The variable Y, with an input of 1.35 (95% CI, 0.80 to 2.20), represents the risk of requiring total hip arthroplasty for obese individuals compared with the risk of requiring it for underweight or normal-weight individuals. The variable Z, with an input of 1.05 (95% CI, 0.85 to 1.40), represents the risk of requiring total hip arthroplasty for overweight individuals compared with the risk of requiring it for underweight or normal-weight individuals. These inputs were derived from a study by Franklin et al. that quantified the relationship between BMI and joint arthroplasty12. These results were plotted to illustrate the trend in the contribution of overweight and obesity to the relative difference in total knee and total hip arthroplasty volumes. Results were reported as odds ratios rather than relative risks because Franklin et al. employed a case-control design12. Although our formulae are designed to utilize relative risks, this statistic cannot be calculated from case-control data. However, they are closely approximated by odds ratios when the disease or outcome in the population of interest is rare22,23. Robertsson et al. combined data from the Swedish Knee Arthroplasty Register and Swedish census registers (the same population as Franklin’s study at around the same time frame) and showed an overall total hip arthroplasty and total knee arthroplasty incidence of sixty-three per 1,000,000 (0.0063% of the population)24. This is sufficiently rare to justify the use of odds ratios as approximations of relative risks, and our calculations were therefore performed as such.
Source of Funding
There was no external source of funding.
The number of total knee arthroplasties performed in the United States more than tripled during our study period, and the number of total hip arthroplasties approximately doubled (Fig. 1). The ratio of total knee arthroplasty to total hip arthroplasty grew from 1.47 in 1993 to 2.18 in 2009. Moreover, because the total knee arthroplasty volume already exceeded the total hip arthroplasty volume in 1993, the gap in procedure volumes widened over time.
Surgical Workforce and Per-Surgeon Output
The ratio of the number of knee surgeons to hip surgeons increased slightly (1.27 in 1993 to 1.36 in 2009), but this shift in physician specialization only accounts for a relatively small share of the divergence in total knee arthroplasty and total hip arthroplasty volumes (Fig. 2-A). However, growth in per-surgeon output was responsible for the majority of the difference. In 1993, surgeons performed 1.16 total knee arthroplasties for every total hip arthroplasty, but this ratio grew to 1.60 by 2009. According to the decomposition equation described above, 70% of the difference between total knee arthroplasty and total hip arthroplasty volumes in 2009 can be explained by a shift toward individual knee surgeons increasing their volume relatively more than hip surgeons rather than by disproportionate growth of knee specialists (Fig. 2-B). In the case of the typical arthroplasty surgeon, this results from increasing total knee arthroplasty volume faster than total hip arthroplasty volume.
Surgeon and Hospital Reimbursement
Surgeon per-case reimbursement for total knee arthroplasty fell from approximately $3000 in 1995 to $1560 in 2009, and surgeon fees for total hip arthroplasty fell from $2840 to $1460 (Fig. 2-C). Because these fees fell equivalently (48% drop in both), surgeons do not appear to be performing relatively more knee replacements than hip replacements in response to disproportionate declines in reimbursement between the procedures.
Hip and knee replacements fall into the same diagnosis-related group, making hospital reimbursement identical for both procedures. Although hospital payment declined 16% (from roughly $13,300 to $11,100) from 1995 to 2009, this fee contraction affected total knee arthroplasty and total hip arthroplasty equivalently (Fig. 2-C). This suggests that economic incentives for surgeons and hospitals to induce disproportionate demand for total knee arthroplasty did not exist on the revenue side and are therefore unlikely to explain the observed phenomena.
Length of Stay
Length of hospital stay is relevant for two reasons. First, inpatient hospital days are a major component of a hospital’s cost for arthroplasty; hence, a decrease in the length of stay of total knee arthroplasty relative to total hip arthroplasty would increase the relative attractiveness of total knee arthroplasty on a cost basis and might induce hospitals to favor these procedures. Second, the availability of hospital beds might be the constraining factor toward increasing output in some settings. If the average length of stay for total knee arthroplasty fell relatively more than total hip arthroplasty, this would promote a shift toward total knee arthroplasty as suppliers strive to enhance their productivity.
The average length of stay for total knee arthroplasty (4.48 days) from 1993 to 2009 was 0.28 day shorter than that of total hip arthroplasty (4.76 days). However, length of stay fell equivalently for both procedures (Fig. 2-D). This negates the hypothesis that changes in relative length of stay may explain the differential rate of growth for knee and hip replacement surgeries. Figure 2 suggests a minor role, if any, for supply-side drivers in explaining the different trends for total knee arthroplasties and total hip arthroplasties.
In-hospital mortality was used as a proxy for the safety of each procedure. Between 1993 and 2009, in-hospital mortality fell 68% for total knee arthroplasty (from 225 to seventy-two deaths per 100,000 operations) and 73% for total hip arthroplasty (from 590 to 158 deaths per 100,000 operations) (Fig. 3-A), suggesting that the safety of total knee arthroplasty did not improve more rapidly than that of total hip arthroplasty. This stands as evidence against the hypothesis that changes in the relative safety of the procedures were responsible for the increase in the volumes of total knee arthroplasty over total hip arthroplasty.
Although the proportion of arthroplasties performed on patients sixty-five years of age and older has fallen, the decline in the average age of patients has been more marked for total knee arthroplasty than total hip arthroplasty. From 1997 to 2009, the share of patients undergoing arthroplasty who were eighteen to sixty-four years of age grew 56% for those undergoing total knee arthroplasty compared with only 35% for those undergoing total hip arthroplasty. As a result, younger patients made up an increasingly large portion of the difference between total knee arthroplasty and total hip arthroplasty volumes (Fig. 3-B).
Overweight and Obesity
Overweight and obesity have both been increasing in prevalence for more than twenty-five years (Fig. 4-A). Because higher BMIs have been associated more strongly with increased odds of developing knee rather than hip arthritis6-10, this nationwide growth in BMI may largely explain the growth in total knee arthroplasty over total hip arthroplasty. Analysis using the BMI-related equations above reveals that growth in total knee arthroplasty volume has far outpaced that of total hip arthroplasty among those with a BMI of ≥25 kg/m2, and total knee arthroplasty and total hip arthroplasty volumes have grown proportionately among individuals with non-elevated BMIs (Fig. 4-B). This more rapid increase in utilization of total knee arthroplasty over total hip arthroplasty in this cohort was responsible for 95% of the growth in total knee arthroplasty over total hip arthroplasty during the study period. The magnitude of this effect will continue to expand if obesity rates continue to increase.
Age was used as a second source of differentiation. The magnitude of the increase in overweight and obese status was greater in the age group of eighteen to sixty-four years than in the age group of sixty-five years and older (Fig. 4-C), which is consistent with the disproportionate growth of total knee arthroplasty to total hip arthroplasty in the age group of eighteen to sixty-four years compared with the sixty-five years and older cohort (Fig. 4-D). The fact that both BMI and the total knee arthroplasty/total hip arthroplasty ratio experienced faster growth in the younger age group lends additional credence to the finding that rising rates of overweight and obesity have preferentially boosted total knee arthroplasty volumes.
These results suggest that, of the factors analyzed, rising BMIs played the most substantial role in increasing demand for total knee arthroplasty above that of total hip arthroplasty, that younger individuals were more affected, and that the surgical community responded to increased demand for procedures primarily by increasing individual surgeon volume rather than by increasing labor supply.
The role of obesity in the development of knee and hip osteoarthritis has been thoroughly investigated. Although strong positive correlations between BMI and the development of knee osteoarthritis have been shown, the effect appears to be of lesser magnitude in the hip6-11. This may be because hinge joints with malalignment or prior trauma might be biomechanically more susceptible to damage from elevated forces associated with obesity than malaligned or previously injured ball-and-socket joints8,25. An extensive literature search identified a single study12 analyzing the effect of BMI on both knees and hips in a mixed population of men and women that showed arthroplasty as a primary outcome. In that study, Franklin et al. reported that such increases in BMI translate to higher rates of arthroplasty, with knees being more sensitive than hips to the effect of excess weight12. Because these findings were consistent with those of related studies6-11,13, the data from Franklin et al. were selected as coefficients in the analysis presented here.
Losina et al. recently investigated the effect of BMI and population growth on total knee arthroplasty volumes in a Swedish population and found that rising obesity could not fully explain ballooning total knee arthroplasty rates3. The analysis presented here focuses on a related but discrete issue, the relative growth of total knee arthroplasty volumes over those of total hip arthroplasty, and suggests that rising BMIs are closely correlated with the increase in total knee arthroplasty over total hip arthroplasty. Thus, it appears likely that rising BMIs have contributed too much of the acceleration in growth of total knee arthroplasty over total hip arthroplasty but may be a less influential factor when considering the dramatic growth in the absolute number of knee replacements.
There is evidence to suggest that total knee arthroplasty has increased in clinical appeal over the years. Proximal tibial osteotomy, a procedure that was commonly used as a substitute for total knee arthroplasty in younger patients, has been decreasing in popularity despite physician reimbursements that have fallen less steeply than those of total knee arthroplasty and hospital payments that have actually increased (see Appendix). Surgeons are substituting away from proximal tibial osteotomy and toward total knee arthroplasty despite financial incentives to do otherwise; this is likely due to improvements in the total knee arthroplasty procedure. Relative improvements in total knee arthroplasty compared with total hip arthroplasty were not able to be assessed in this study but would be informative and are a potential topic of future investigation.
Other factors that would be interesting to examine as data become available include the effects of changes in operative time; sports-related injuries and meniscectomies; direct-to-consumer marketing; and advances in implant design, clinical management, and surgical techniques and tools. Perhaps some of the disproportionate growth in total knee arthroplasty over total hip arthroplasty is attributable to the fact that total hip arthroplasty was developed and was refined earlier than total knee arthroplasty; hip arthroplasty may have achieved mature levels of adoption while knee arthroplasty is still expanding in prevalence toward its eventual steady state of market penetration4.
The primary challenge in developing this study was identifying high-quality, longitudinal data sources that reflected supply-side and demand-side factors potentially influencing arthroplasty volumes. These data were thus compiled from a number of different sources, which may have introduced some error into comparisons and analyses. Additionally, some data sources were proxies (for example, in-hospital mortality for safety) made necessary by the dearth of more comprehensive nationwide data sets. The use of surrogate data sources would not be expected to directionally affect study results as long as the inherent error introduced by such approximation does not systematically differ between total knee arthroplasty and total hip arthroplasty. Systematic bias is unlikely given the fact that surgeons have been exposed to a similar set of influences nationwide; changes in reimbursement are fairly standard, and BMI growth is widespread.
In conclusion, the results of this study suggest that rising BMIs have played a substantial role in increasing the demand for total knee arthroplasty over total hip arthroplasty, that younger individuals have been affected to a greater degree, and that the surgical community has responded to this added need primarily by increasing per-physician output. These findings help explain the differential growth rates between total knee arthroplasty and total hip arthroplasty and may prove instructive in understanding and planning for future arthroplasty trends, especially if overweight and obesity continue to increase in prevalence. Although we have synthesized a preponderance of epidemiologically based evidence, the experimental design employed herein is suited to identify correlation and not necessarily causation.
An explanation of the data and sources used in conducting an analysis of trends in proximal tibial osteotomy and a figure showing trends in the annual volume of proximal tibial osteotomy and a comparison of trends of reimbursement for proximal tibial osteotomy, total knee arthroplasty, and total hip arthroplasty are available with the online version of this article as a data supplement at jbjs.org.
Investigation performed at the Department of Orthopaedics, Hospital for Special Surgery, New York, NY, and The Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
2. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997-2004. Arthritis Rheum. 2008 ;59(4):481–8.
3. Losina E, Thornhill TS, Rome BN, Wright J, Katz JN. The dramatic increase in total knee replacement utilization rates in the United States cannot be fully explained by growth in population size and the obesity epidemic. J Bone Joint Surg Am. 2012 ;94(3):201–7.
4. Crowninshield RD, Rosenberg AG, Sporer SM. Changing demographics of patients with total joint replacement. Clin Orthop Relat Res. 2006 ;443:266–72.
5. Rana AJ, Iorio R, Healy WL. Hospital economics of primary THA decreasing reimbursement and increasing cost, 1990 to 2008. Clin Orthop Relat Res. 2011 ;469(2):355–61.
6. Gelber AC, Hochberg MC, Mead LA, Wang NY, Wigley FM, Klag MJ. Body mass index in young men and the risk of subsequent knee and hip osteoarthritis. Am J Med. 1999 ;107(6):542–8.
7. Grotle M, Hagen KB, Natvig B, Dahl FA, Kvien TK. Obesity and osteoarthritis in knee, hip and/or hand: an epidemiological study in the general population with 10 years follow-up. BMC Musculoskelet Disord. 2008 ;9:132.
8. Reijman M, Pols HAP, Bergink AP, Hazes JM, Belo JN, Lievense AM, Bierma-Zeinstra SM. Body mass index associated with onset and progression of osteoarthritis of the knee but not of the hip: the Rotterdam Study. Ann Rheum Dis. 2007 ;66(2):158–62. Epub 2006 Jul 12.
9. Mork PJ, Holtermann A, Nilsen TIL. Effect of body mass index and physical exercise on risk of knee and hip osteoarthritis: longitudinal data from the Norwegian HUNT Study. J Epidemiol Community Health. 2012 ;66(8):678–83. Epub 2012 Apr 17.
10. Lohmander LS, Gerhardsson de Verdier M, Rollof J, Nilsson PM, Engström G. Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study. Ann Rheum Dis. 2009 ;68(4):490–6. Epub 2008 May 8.
11. Holliday KL, McWilliams DF, Maciewicz RA, Muir KR, Zhang W, Doherty M. Lifetime body mass index, other anthropometric measures of obesity and risk of knee or hip osteoarthritis in the GOAL case-control study. Osteoarthritis Cartilage. 2011 ;19(1):37–43. Epub 2010 Oct 31.
12. Franklin J, Ingvarsson T, Englund M, Lohmander LS. Sex differences in the association between body mass index and total hip or knee joint replacement resulting from osteoarthritis. Ann Rheum Dis. 2009 ;68(4):536–40. Epub 2008 May 26.
13. Liu B, Balkwill A, Banks E, Cooper C, Green J, Beral V. Relationship of height, weight and body mass index to the risk of hip and knee replacements in middle-aged women. Rheumatology (Oxford). 2007 ;46(5):861–7. Epub 2007 Feb 4.
20. Bevington P, Robinson D. Error Analysis. In: Data Reduction and Error Analysis for the Physical Sciences, 3rd ed. New York, NY: McGraw-Hill; 2002. p 36–51.
22. Grimes DA, Schulz KF. Cohort studies: marching towards outcomes. Lancet. 2002 ;359(9303):341–5.
23. Cummings P. The relative merits of risk ratios and odds ratios. Arch Pediatr Adolesc Med. 2009 ;163(5):438–45.
24. Robertsson O, Dunbar MJ, Knutson K, Lidgren L. Past incidence and future demand for knee arthroplasty in Sweden: a report from the Swedish Knee Arthroplasty Register regarding the effect of past and future population changes on the number of arthroplasties performed. Acta Orthop Scand. 2000 ;71(4):376–80.
25. Sharma L, Chang A. Overweight: advancing our understanding of its impact on the knee and the hip. Ann Rheum Dis. 2007 ;66(2):141–2.