Research in THA and TKA From the United States Has Declined Over the Past Decade Relative to Other Countries

Materials and Methods

All original research published in The Journal of Bone and Joint Surgery, American Volume (JBJS-Am), Journal of Arthroplasty (JOA), and Clinical Orthopaedics and Related Research (CORR) from January 1999 to December 2008 was queried. Every article in every issue during that time period was examined, and all original research involving THA and TKA was identified and examined in detail. Research manuscript inclusion criteria included original clinical or basic science research involving primary or revision THA or TKA. Research involving epidemiology, resource use, or cost analysis of THA and TKA was included. Additional topics selected included unicompartmental knee arthroplasty and hip resurfacing. If the manuscript involved treatment of the proximal femur, it was only included if THA was used in at least part of the study. Manuscripts isolated to hemiarthroplasty or internal fixation of proximal femoral fractures were excluded. In addition, supplements of meeting proceedings and surgical techniques were excluded due to the altered peer-review process and selection bias that is typically involved with those particular publications.

The proceedings of the Hip and Knee Society and the annual American Association of Hip and Knee Surgeons were reported as well, but addressed separately because the manuscripts are invited and may introduce bias by virtue of the disproportionate number of members from the United States in those respective societies. Review articles and opinion publications were excluded.

Each manuscript that met the inclusion criteria was examined in detail through review of the abstract and all pertinent accompanying manuscript information. The research country of origin was identified and documented. In addition, each manuscript was examined for the level of evidence to provide a more objective classification of the research quality. The manuscript on clinical research was assigned 1 of the following levels of evidence: Level I (randomized, prospective), Level II (lesser quality randomized, prospective), Level III (case control study), and Level IV (case series). If a manuscript included a level of evidence classification assigned to it by that particular journal and peer-review process, that level of evidence classification was accepted and recorded. All manuscripts that did not have a documented level of evidence were carefully reviewed and a level of evidence was assigned based on the established classification criteria of JBJS-Am. ⁴ Manuscripts dealing with basic science, such as cadaver or biomechanical studies, were classified as “basic science” and were not classified with the levels of evidence for clinical studies.

All information was recorded into a database for statistical analysis. The absolute number of manuscripts published annually was determined and examined with respect to country of origin and level of evidence. The number of publications from the United States and from non-US countries as a percentage of total annual research publications was determined in addition to the level of evidence and compared.

Results

From January 1999 to December 2008, 2272 original peer-reviewed research studies involving THA and TKA were identified. After exclusion of the annual Hip Society and Knee Society meetings and the American Association of Hip and Knee Surgeons annual meeting proceedings’ studies to minimize bias, 1724 original research publications remained and were analyzed. There were 399 articles published in JBJS-Am, 329 articles in CORR, and 996 articles in JOA. The topic distribution included 767 primary THA articles, 727 primary TKA articles, 89 revision THA articles, and 33 revision TKA articles, with the remaining 108 articles including both THA and TKA. Examined on an annual basis, the number of overall THA and TKA publications in all 3 journals steadily increased from 146 articles in 1999 to 194 articles in 2008.

Over the 10-year period, US institutions published 908 articles involving THA and/or TKA, while 135 articles came from the United Kingdom, 107 from Japan, 87 from Canada, and 73 from Korea. Thirty-three other countries published <50 articles during this time period (). The percentage of publications from US institutions steadily declined from 65.8% in 1999 to 46.9% in 2008 (Figure ). In contrast, the annual percentage of publications from non-US institutions increased from 34.2% in 1999 to 53.1% in 2008. The annual absolute number of US publications has remained relatively constant during the past decade, with an average of 91 per year, while the non-US countries effectively doubled their annual number of publications from 50 in 1999 to 103 in 2008 (Figure ).

Table. Number of Publications Per Country of Origin From January 1999 to December 2008

Figure 1:. The percentage of annual publications from the United States and non-US institutions (excluding American Association of Hip and Knee Surgeons, Hip Society, and Knee Society meeting proceedings) demonstrating the decreasing trend in percentage of publications from the United States and the conversely increasing percentage of publications from non-US institutions (A). The absolute number of annual publications from US and non-US institutions (excluding American Association of Hip and Knee Surgeons, Hip Society, and Knee Society meeting proceedings) demonstrating the relatively constant number of publications from the United States and the conversely increasing number of publications from non-US institutions (B). The percentage of annual publications from US and non-US institutions, including the American Association of Hip and Knee Surgeons, Hip Society, and Knee Society meeting proceeding publications. The decreasing trend in percentage of publications from the United States is demonstrated (C).

When including the annual meeting proceeding publications from the American Association of Hip and Knee Surgeons, Hip Society, and Knee Society, the same trends are evident with regard to the steady decline in the percentage of US publications. Because these meeting proceedings are predominantly US institution-based, including these publications in the analysis increases the overall annual number of US publications; however, the trend of a decreasing percentage remains relative to other countries. The percentage of publications from US institutions steadily declined from 72.8% in 1999 to 52.8% in 2008 (Figure ). The annual absolute number of US publications has again remained relatively constant during the past decade, with 142 publications in 1999, 133 publications in 2008, and a mean 140.9 publications over that 10-year period. This relatively constant rate of annual publications from the United States remains evident even when the American Association of Hip and Knee Surgeons, Hip Society, and Knee Society meeting proceedings’ publications are included in the analysis.

Of the 1724 published articles, the largest number were classified as Level IV evidence, with 568 studies (33%), whereas 131 were classified as Level I (8%), 184 Level II (10%), and 344 Level III (20%). There were 494 basic science publications (29%). The percentage of high-quality Level I and II evidence studies from the United States increased from 9.7% in 1999 to 23.3% in 2008, which parallels the increase seen from all countries from 7.8% to 24.8% during the same decade (Figure ).

Figure 2:. The percentage of level of evidence classification for publications from all countries, including the United States, annually in all 3 journals. The linear regression line is shown for each level of evidence classification to more clearly see the trends (A). The percentage of level of evidence classification for the publications from US institutions annually in all 3 journals. The linear regression line is shown for each level of evidence classification to more clearly see the trends (B).

Discussion

The results of this study reveal a relative decrease in joint arthroplasty publications from US institutions relative to other countries in the world, with the percentage of publications from US institutions steadily declining from 65.8% in 1999 to 46.9% in 2008. In addition, despite the total annual number of overall THA and TKA publications in the 3 journals queried increasing from 146 articles in 1999 to 194 articles in 2008, the annual number of US publications has remained relatively constant during the past decade, with an average of 91 per year. Alternatively, non-US countries have effectively doubled their annual number of publications, from 50 in 1999 to 103 in 2008 (Figure ).

A potential explanation for the relative decline in the THA and TKA publication rate in these journals by US institutions may be a difference in the scientific quality of the research performed. In JBJS-AM, there has been a documented improvement in the level of evidence of the research studies published over the past 30 years. ⁵ However, the results of this study demonstrate the percentage of high-quality Level I and II evidence studies from the United States increased from 9.7% in 1999 to 23.3% in 2008, which parallels the increase seen in all countries, from 7.8% to 24.8%, during the same decade (Figure ). Therefore, the level of evidence as an indicator of a study’s scientific quality does not appear to be a factor in the difference between US and non-US institution publication rates over the past decade in THA and TKA.

Another possibility for the increase in publications from other countries is that their interest and ability to perform and publish high-quality research is increasing, while the United States remains relatively stagnant or is at its peak capacity within the current health care environment. The changing economic climate in the practice of THA and TKA in the United States, combined with the increased costs of doing research because of institutional review board requirements, may be making it more difficult for surgeons in the United States to conduct high-quality clinical research.

A steady decline in physician reimbursement for THA and TKA has been observed over the past 10 to 15 years. ⁶ There has been a 20% to 23% decrease in primary and revision THA and TKA Medicare reimbursement from 1998 to 2007. ⁶ When adjusted for inflation with the consumer price index, there has been a dramatic decrease of >40% in Medicare reimbursement for primary and revision THA and TKA since 1992. ⁶ This decrease in reimbursement has made it difficult for surgeons, as both practice and employee expenses continue to escalate. Furthermore, in the same time interval of decreasing reimbursement, the surgical volume of THA and TKA in the United States has increased steadily. The annual number of TKAs performed in the United States has steadily increased from 266,000 in 1998 to 542,000 in 2006. ⁷ Likewise, 160,000 THAs were performed in 1998 and increased to 231,000 in 2006. ⁸

Decreasing physician reimbursement, rising fixed business costs, and rising malpractice premiums have created an environment where an arthroplasty surgeon must increase the volume of THAs and TKAs performed to maintain their given income level or even allow practice survival. ¹ To accommodate this volume, much emphasis has been placed on optimizing the efficiency of THA and TKA surgical care; however, regardless of the degree of practice efficiency, one could surmise that the increased surgical volume with the accompanying patient care and complication volumes expected may result in less time for the busy arthroplasty surgeon to devote to clinical and/or basic science research. Furthermore, the burden of revision procedures, which are more time consuming and demanding, is typically borne by fellowship-trained arthroplasty surgeons at institutions that might be more apt to perform research. The rates of revision THA and TKA in the United States have risen over the past decade and are expected to continue rising over the next 2 decades. ²

It has been reported that current and future orthopedic surgery residents are becoming less likely to choose adult hip and knee reconstruction as a subspecialty. ^2,3 Ten adult reconstructive fellowship programs have been discontinued over the past 5 years due to an inability to attract applicants, and only 77% of the 119 fellowship positions were filled in the academic year 2006–2007. ² This percentage decreased to only 62% of fellowship positions being filled in the following 2007–2008 academic year. ² This decrease in the number of fellows reduces the number of individuals available to perform the procedures in the present and future. Finally, over the past decade, meeting institutional review board requirements to perform clinical research has become more stringent and more expensive. Therefore, the combination of increasing clinical workload, decreased revenue, and increased costs associated with performing research may be contributing to this decline in arthroplasty productivity research in the United States. Further studies will be necessary to confirm this hypothesis.

This study has limitations, which primarily involve the selection of the orthopedic journals reviewed. We limited our research publication query to 3 journals that were deemed to be of the highest quality and yield for THA and TKA research by the 2 senior authors (R.M.M., J.R.L.). This may represent a bias, as research from United States and non-US institutions may preferentially be submitted to other journals of high quality. However, the authors believe that other journals that frequently publish arthroplasty articles, such as Acta Orthopaedica Scandanavica and the Journal of Bone and Joint Surgery, British Volume, likely publish a lower percentage of articles from US centers than the 3 journals evaluated in this study. In addition, these 3 journals demonstrate an emphasis on clinical research rather than basic science research, as is emphasized by journals such as the Journal of Orthopaedic Research. Furthermore, we intentionally analyzed US-based specialty meeting proceedings separately, as we felt artificial bias may be introduced into the conclusions based on the altered peer-review process that may occur with invited manuscripts, as opposed to those that are submitted without specific invitation.

Whether this relative decline in US-based THA and TKA research is specifically related to decreased physician reimbursement, decreased time available for research due to increased practice volume, or a decline in interest from orthopedic residents in this subspecialty remains to be determined and requires further study. However, with health care reform a top national priority, the role of the United States as a leader in THA and TKA research, development, and innovation must be concomitantly addressed. Further study is warranted to elicit the etiology of the trends revealed in this study and to implement changes and reform that enable and promote opportunities and incentives for the busy arthroplasty surgeon and clinician-scientist to perform and publish high-quality research in THA and TKA.