There are many limitations to this study, most significant of which was the use of procedural rates starting in 2006, a delay of 3 years after the introduction of a unique arthroscopic rotator cuff repair CPT code in 2003, and nearly a decade after arthroscopic techniques were well described in the literature [22, 43, 44]. Recognizing this delay is important as it frames the phase of diffusion that we were able to study. Diffusion science defines innovators or early adopters to be the first 16% of individuals to adopt innovation, while the next 34% are generally termed the early majority [6, 18]. In 2006, 50% of rotator cuff repair procedures were already being completed arthroscopically suggesting that we are unable to meaningfully evaluate the characteristics of regions that might be considered innovators or early adopters. Instead, our 2006 analysis is effective for evaluating the early majority phase of diffusion as the critical mass had been reached and widespread adoption had already begun to occur [6, 18]. In 2014, 72% of rotator cuff repair were being completed arthroscopically and the analysis from this time is more effective for understanding patterns related to the late majority or laggards [6, 18]. Altogether, this limitation makes it impossible for us to determine what regions or characteristics are associated with early adoption, but instead provides insight into how novel technology spreads throughout the orthopaedic community after it is a proven option. A second major limitation was the use of the Medicare population as our study cohort, since this group of patients is not necessarily where innovative procedures are introduced. The trends that we observed within this group of patients may not hold true for younger patients where innovative procedures may be offered or pursued more often. Another major limitation was our assumption that the adoption of arthroscopic rotator cuff repair represented a novel step forward that should be pursued, while studies including a recent high-quality randomized control trial , continue to suggest that long-term outcomes between open and arthroscopic rotator cuff repair are equivalent. It is possible that individual surgeons or institutions realized early on that arthroscopic repair did not provide superior outcomes and chose to stay with their proven open techniques.
An additional limitation was the use of hospital referral regions that varied in population, since they were defined according to referral patterns and could vary substantially across rural or urban areas. In regions with a relatively small population, the preferences of a single surgeon can affect the overall proportion of procedures being completed arthroscopically in a more significant manner than in areas with larger populations and many surgeons. This issue could help to explain why several relatively small hospital referral regions including Provo, UT, USA, or Altoona, PA, USA, ended up at the top of the adoption list. To adjust for these discrepancies, we completed population-weighted regression to better understand trends across regions. Another limitation was the use of administrative CPT code-based data that may include errors made at the time of billing. For example, there was evidence that a small number (< 1%) of rotator cuff repair procedures were coded as both open and arthroscopic within the Medicare database. Because of the methods used for extraction, we could not individually identify these procedures to remove them from the analysis; thus, they remained in the overall counts. However, this should not have affected our overall conclusions because the number of such occurrences was small. A statistical limitation was our inability to perform a Bonferroni correction to the p value cutoff considered to be significant, a technique that could have been applied since we evaluated several independent associations within our regression models. Our sample size was relatively small and limited by the number of regions within the United States, making it impossible to gain the power necessary to utilize a p value cutoff less than 0.05, as would be indicated by the Bonferroni correction. This issue raises the possibility of observing spurious statistically significant results simply by chance, making it necessary to consider the plausibility and reproducibility of the associations. Furthermore, we did not have information to model regional differences in disease severity or surgical conditions, which may influence the decision whether to pursue open or arthroscopic repair. Therefore, we assumed that these factors were similar between regions. We measured innovation by determining the proportion of procedures performed arthroscopically so that differences in underlying use among regions would not influence our overall conclusions. Similarly, we were unable to quantify the density of shoulder specialists by region, another factor that could have influenced the use of novel technology, because there is no single professional organization or designation to identify these surgeons. Another limitation was the use of the membership of the Council of Teaching Hospitals  as our marker of an academic medical center because some institutions that are not on this list likely consider themselves academically focused. However, we felt that this list effectively captured information on the largest and most influential educational institutions in the United States. Importantly, our results only apply to the adoption of arthroscopic rotator cuff repair and would likely be different if evaluating the diffusion of other procedures, such as hip arthroscopy, where different surgeons and/or regions might be found to be relatively earlier adopters.
We found substantial variation in the adoption of arthroscopic rotator cuff repair. With considerable geographic variation in orthopaedic surgical practice at baseline [5, 13, 29, 45], we expected substantial heterogeneity in the adoption of arthroscopic rotator cuff repair. Using data from 2007, Chung et al.  noted 10-fold variation in the use of open reduction and internal fixation for distal radius fractures, a technique considered new at the time. This is larger than the fivefold variation in the adoption of arthroscopic rotator cuff repair we observed and may be owing to differences in the nature of the innovations or timepoints along the innovation curve at which the data were collected. Chung et al.  suggested that the high degree of variation in the adoption of internal fixation was because of ambiguity in previous studies regarding the optimal treatment; this theory also applies to the adoption of arthroscopic rotator cuff repair because other studies continued to show equivalent outcomes between open and arthroscopic repair [3, 11, 27, 31, 37, 40, 49]. Altogether, the considerable geographic variation observed in the use of arthroscopic rotator cuff repair highlights that Medicare beneficiaries do not have equal access to this procedure across the United States.
We found strong evidence for a geographic pattern of adoption on a larger regional level as demonstrated by mapping and regression modeling. In 2006, there was higher adoption in the Northeast and Mountain West than in other regions, with widespread adoption nationally by 2014. Our study also demonstrated that specific regions—the Northwest and Midwest—had lower rates of adoption than did other regions. A previous study on surgeries performed by newly trained surgeons observed that there were more arthroscopic shoulder procedures in the Northeast, and more open procedures in the Northwest and Midwest , consistent with our results. Regional trends may be further reinforced if surgeons who train in a given area choose to stay there and ultimately practice in ways similar to their mentors. These trends in the spread of new technology do not necessarily cross surgical disciplines; a recent study in urology demonstrated that fewer robotic procedures, considered a recent innovation, were performed in the South than in other regions , an area noted in our study to adopt arthroscopic rotator cuff repair relatively early. The reason for this discrepancy is unclear, but it highlights that technology diffusion patterns may vary based on the cost of technology; a surgical robot is a much more expensive innovation than arthroscopic equipment.
We found that several independent factors were associated with earlier adoption of arthroscopic repair within a given hospital referral region including a more educated population. The characteristics of a community are an important component affecting technology diffusion [6, 18], and these results demonstrate a potential connection between this factor and the motivation for surgeons to innovate. Studies on arthroplasty have demonstrated that more educated patients are more likely to pursue joint arthroplasty than less educated patients [24, 25, 38], and our results suggest that education may also be associated with the desire for the most innovative surgical techniques. When considering the cost and benefits of innovation [6, 18], surgeons may find novel technology particularly advantageous in more educated markets. Increased Medicare expenditures were also associated with adoption, and higher regional spending may occur in areas that are more innovative, as technological innovation is a major contributor to health care costs . It is important to consider that novel procedures are not always a step forward, as has been demonstrated by synthetic ACL grafts , thermal capsulorrhaphy , and metal-on-metal hip replacements , and the presence of checkpoints along the pathway of adoption is important. The observation that regions home to academic medical centers had lower adoption rates in 2006 is consistent with diffusion science principles, which suggest that thought-leaders wait to endorse novel technology until its effectiveness has been proven by early adopters [6, 18], and may suggest that academic institutions play a gatekeeper role in technology dissemination. In contrast to basic science innovations, which often originate in large academic centers and then spread outwards , orthopaedic surgical innovation may be more likely to be developed by private industry, as was the case with arthroscopic rotator cuff repair , and then spread back into academia. Alternatively, physicians in academic institutions may not have been influenced by the financial  and competitive advantages associated with arthroscopic repair, or may have appreciated the equivalence between open and arthroscopic repair , either of which could have led to lower adoption. Diffusion science has long recognized the impact of context on the diffusion of innovation, and these results help to frame regional factors that were associated with adoption of arthroscopic rotator cuff repair.
Although the use of arthroscopic rotator cuff repair grew rapidly over the last two decades, adoption of this procedure in the Medicare population occurred heterogeneously in the United States demonstrating that patients do not have equal access to innovative techniques across the country. While the Mountain West was noted to adopt relatively earlier than other areas, the Northwest and Midwest were associated with lower adoption of arthroscpic procedures, highlighting how larger regional trends in innovation are present. A more educated population was associated with innovation and highlights the connection between the context of a community and the motivation for surgeons to adopt novel techniques. Higher healthcare spending was also associated with adoption and illustrates the tension between controlling costs and encouraging innovation within a region. Future prospective studies that account for patient and surgeon characteristics, in addition to surgical indications, are necessary for understanding whether these trends in innovation are true across other orthopaedic procedures.
We thank JoAnna Leyenaar MD, PhD for her assistance with statistical analysis and manuscript editing.
2. Ahmad CS, Levine WN, Bigliani LU. Arthroscopic rotator cuff repair. Orthopedics. 2004;27:570-574.
3. Barnes LA, Kim HM, Caldwell JM, Buza J, Ahmad CS, Bigliani LU, Levine WN. Satisfaction, function and repair integrity after arthroscopic versus mini-open rotator cuff repair. Bone Joint J. 2017;99:245-249.
4. Batty LM, Norsworthy CJ, Lash NJ, Wasiak J, Richmond AK, Feller JA. Synthetic devices for reconstructive surgery of the cruciate ligaments: a systematic review. Arthroscopy. 2015;31:957-968.
5. Bell JE, Leung BC, Spratt KF, Koval KJ, Weinstein JD, Goodman DC, Tosteson AN. Trends and variation in incidence, surgical treatment, and repeat surgery of proximal humeral fractures in the elderly. J Bone Joint Surg Am. 2011;93:121-131.
6. Berwick DM. Disseminating innovations in health care. JAMA. 2003;289:1969-1975.
7. Birkmeyer JD, Sharp SM, Finlayson SR, Fisher ES, Wennberg JE. Variation profiles of common surgical procedures. Surgery. 1998;124:917-923.
8. Bodenheimer T. High and rising health care costs. part 2: technologic innovation. Ann Intern Med. 2005;142:932-937.
9. Burkhart SS. The burden of craft in arthroscopic rotator cuff repair: where have we been and where we are going. Am J Orthop (Belle Mead, NJ). 2015;44:353-358.
10. Burkhart SS, Danaceau SM, Pearce CE Jr. Arthroscopic rotator cuff repair: analysis of results by tear size and by repair technique-margin convergence versus direct tendon-to-bone repair. Arthroscopy. 2001;17:905-912.
11. Carr A, Cooper C, Campbell MK, Rees J, Moser J, Beard DJ, Fitzpatrick R, Gray A, Dawson J, Murphy J, Bruhn H, Cooper D, Ramsay C. Effectiveness of open and arthroscopic rotator cuff repair (UKUFF): a randomised controlled trial. Bone Joint J. 2017;99:107-115.
13. Chung KC, Shauver MJ, Yin H, Kim HM, Baser O, Birkmeyer JD. Variations in the use of internal fixation for distal radial fracture in the United States medicare population. The J Bone Jont Surg Am. 2011;93:2154-2162.
14. Codman EA. Complete rupture of the supraspinatus tendon. Operative treatment with report of two successful cases. 1911. J Shoulder Elbow Surg. 2011;20:347-349.
15. Colvin AC, Egorova N, Harrison AK, Moskowitz A, Flatow EL. National trends in rotator cuff repair. J Bone Joint Surg Am. 2012;94:227-233.
18. Dearing JW, Cox JG. Diffusion of innovations theory, principles, and practice. Health Aff (Millwood). 2018;37:183-190.
19. Ensor KL, Kwon YW, Dibeneditto MR, Zuckerman JD, Rokito AS. The rising incidence of rotator cuff repairs. J Shoulder Elbow Surgery. 2013;22:1628-1632.
20. Gachter A, Seelig W. Arthroscopy of the shoulder joint. Arthroscopy. 1992;8:89-97.
21. Garrett WE Jr, Swiontkowski MF, Weinstein JN, Callaghan J, Rosier RN, Berry DJ, Harrast J, Derosa GP. American Board of Orthopaedic Surgery Practice of the Orthopaedic Surgeon: Part-II, certification examination case mix. J Bone Joint Surg Am. 2006;88:660-667.
22. Gartsman GM, Khan M, Hammerman SM. Arthroscopic repair of full-thickness tears of the rotator cuff. J Bone Joint Surg Am. 1998;80:832-840.
23. Good CR, Shindle MK, Kelly BT, Wanich T, Warren RF. Glenohumeral chondrolysis after shoulder arthroscopy with thermal capsulorrhaphy. Arthroscopy. 2007;23:797.e1-5.
24. Hanchate AD, Zhang Y, Felson DT, Ash AS. Exploring the determinants of racial and ethnic disparities in total knee arthroplasty: health insurance, income, and assets. Med Care. 2008;46:481-488.
25. Hawker GA, Guan J, Croxford R, Coyte PC, Glazier RH, Harvey BJ, Wright JG, Williams JI, Badley EM. A prospective population-based study of the predictors of undergoing total joint arthroplasty. Arthritis Rheum. 2006;54:3212-3220.
26. Jensen AR, Cha PS, Devana SK, Ishmael C, Di Pauli von Treuheim T, D’Oro A, Wang JC, McAllister DR, Petrigliano FA. Evaluation of the trends, concomitant procedures, and complications with open and arthroscopic rotator cuff repairs in the Medicare population. Ortho J Sports Med. 2017;5:2325967117731310.
27. Ji X, Bi C, Wang F, Wang Q. Arthroscopic versus mini-open rotator cuff repair: an up-to-date meta-analysis of randomized controlled trials. Arthroscopy. 2015;31:118-124.
28. Katz JN, Keller RB, Simmons BP, Rogers WD, Bessette L, Fossel AH, Mooney NA. Maine Carpal Tunnel Study: outcomes of operative and nonoperative therapy for carpal tunnel syndrome in a community-based cohort. J Hand Surg Am. 1998;23:697-710.
29. Keller RB, Largay AM, Soule DN, Katz JN. Maine Carpal Tunnel Study: small area variations. J Hand Surg Am. 1998;23:692-696.
30. Koval KJ, Lurie J, Zhou W, Sparks MB, Cantu RV, Sporer SM, Weinstein J. Ankle fractures in the elderly: what you get depends on where you live and who you see. J Orthop Trauma. 2005;19:635-639.
31. Liu J, Fan L, Zhu Y, Yu H, Xu T, Li G. Comparison of clinical outcomes in all-arthroscopic versus mini-open repair of rotator cuff tears: a randomized clinical trial. Medicine. 2017;96:e6322.
32. Maurer-Ertl W, Friesenbichler J, Holzer LA, Leitner L, Ogris K, Maier M, et al. Recall of the ASR XL Head and Hip Resurfacing Systems. Orthopedics. 2017;40:e340-e7.
33. Mauro CS, Jordan SS, Irrgang JJ, Harner CD. Practice patterns for subacromial decompression and rotator cuff repair: an analysis of the American Board of Orthopaedic Surgery database. J Bone Joint Surg Am. 2012;94:1492-1499.
34. Millstein ES, Snyder SJ. Arthroscopic evaluation and management of rotator cuff tears. Orthop Clin North Am. 2003;34:507-520.
35. Monn MF, Cary KC, Kaimakliotis HZ, Flack CK, Koch MO. National trends in the utilization of robotic-assisted radical cystectomy: an analysis using the Nationwide Inpatient Sample. Urol Oncol. 2014;32:785-790.
36. Morris ZS, Wooding S, Grant J. The answer is 17 years, what is the question: understanding time lags in translational research. J R Soc Med. 2011;104:510-520.
37. Morse K, Davis AD, Afra R, Kaye EK, Schepsis A, Voloshin I. Arthroscopic versus mini-open rotator cuff repair: a comprehensive review and meta-analysis. Am J Sports Med. 2008;36:1824-1828.
38. Mota RE, Tarricone R, Ciani O, Bridges JF, Drummond M. Determinants of demand for total hip and knee arthroplasty: a systematic literature review. BMC Health Serv Res. 2012;12:225-225.
39. Norberg FB, Field LD, Savoie FH 3rd. Repair of the rotator cuff. Mini-open and arthroscopic repairs. Clinics Sports Med. 2000;19:77-99.
40. Pham TT, Bayle Iniguez X, Mansat P, Maubisson L, Bonnevialle N. Postoperative pain after arthroscopic versus open rotator cuff repair. A prospective study. Orthop Traumatol Surg Res. 2016;102:13-17.
41. Rogers EM. Diffusion of Innovations. New York, New York. The Free Press; 1983.
42. Skinner J, Weinstein JN, Sporer SM, Wennberg JE. Racial, ethnic, and geographic disparities in rates of knee arthroplasty among Medicare patients. N Engl J Med 2003;349:1350-1359.
43. Tauro JC. Arthroscopic rotator cuff repair: analysis of technique and results at 2- and 3-year follow-up. Arthroscopy. 1998;14:45-51.
44. Thal R. A technique for arthroscopic mattress suture placement. Arthroscopy. 1993;9:605-607.
45. Vitale MG, Krant JJ, Gelijns AC, Heitjan DF, Arons RR, Bigliani LU, Flatow EL. Geographic variations in the rates of operative procedures involving the shoulder, including total shoulder replacement, humeral head replacement, and rotator cuff repair. J Bone Joint Surg Am. 1999;81:763-772.
46. Weinstein JN, Bronner KK, Morgan TS, Wennberg JE. Trends and geographic variations in major surgery for degenerative diseases of the hip, knee, and spine. Health Aff (Millwood). 2004;Suppl Variation:Var81-89.
47. Weinstein JN, Lurie JD, Olson PR, Bronner KK, Fisher ES. United States' trends and regional variations in lumbar spine surgery: 1992-2003. Spine (Phila Pa 1976). 2006;31:2707-2714.
48. Wennberg JE, Cooper MM, Birkmeyer JD. The Quality of Medical Care in the United States: A Report on the Medicare Program, The Dartmouth Atlas of Health Care.
Chicago, IL: American Hospital Association Press; 1999.
49. Williams G Jr, Kraeutler MJ, Zmistowski B, Fenlin JM Jr. No difference in postoperative pain after arthroscopic versus open rotator cuff repair. Clin Orthop Relat Res. 2014;472:2759-2765.
50. Yamaguchi K, Levine WN, Marra G, Galatz LM, Klepps S, Flatow EL. Transitioning to arthroscopic rotator cuff repair: the pros and cons. Inst Course Lect. 2003;52:81-92.
51. Zhang AL, Montgomery SR, Ngo SS, Hame SL, Wang JC, Gamradt SC. Analysis of rotator cuff repair trends in a large private insurance population. Arthroscopy. 2013;29:623-629.