- Email: [email protected]
Does Orthopaedic Training Compromise the Outcome in Knee Joint Arthroplasty?
ORIGINAL REPORTS
Does Orthopaedic Training Compromise the Outcome in Knee Joint Arthroplasty? Richard Storey, MD,* Chris Frampton, PhD,† David Kieser, MD, FRACS, PhD,* Ramez Ailabouni, MD, FRACS,* and Gary Hooper, MD, FRACS* *
Department of Orthopaedic Surgery and Musculoskeletal Medicine, University of Otago, Christchurch, New Zealand; and †Department of Medicine, University of Otago, Christchurch, New Zealand
OBJECTIVE: This study investigates knee joint arthroplasty and compares the outcomes between attending (consultant) orthopedic surgeons and resident (trainee) surgeons. DESIGN: Retrospective review and comparison of knee
joint arthroplasty outcomes between 4 surgeon groups (attending, supervised senior and junior residents, and unsupervised senior residents). Measured outcomes were implant survival (revision rate) and patient reported functional outcomes, measured by Oxford knee score (OKS). SETTING: New Zealand arthroplasty service. PARTICIPANTS: Seventeen years of knee joint arthroplasty
not increased in resident surgeon groups. Postoperative function was not reduced by a clinically significant amount in TKA in any of the resident surgeon groups but was reduced in supervised junior resident and unsupervised senior resident surgeon groups for UKA. ( J Surg Ed C ]:]]]-]]]. J 2018 Association of Program Directors in Surgery. Published by Elsevier Inc. All rights reserved.) KEY WORDS: knee joint arthroplasty, learning curve,
functional outcomes, orthopedic training, resident surgeon COMPETENCIES: Patient Care, Practice Based Learning
and Improvement and Systems Based Practice
data from the New Zealand Joint Registry (NZJR) was reviewed. RESULTS: The New Zealand Joint Registry (NZJR) data
showed 79,671 total knee arthroplasties (TKA) and 8854 unicompartmental knee arthroplasties (UKA) performed between 1999 and 2016. Attending surgeons performed 90% and 97% of TKA and UKA, respectively. The number and proportion of resident performed knee joint arthroplasty has decreased. Faster operation times was observed in the attending surgeon group. Attending surgeon revision rate was 0.49 and 1.19/100 component years for TKA and UKA, respectively, this was not significantly increased in resident surgeon groups. Postoperative OKS was 37.7 and 39.7 for attending surgeon performed TKA and UKA, respectively. Mean OKS were less than 2 points worse in resident groups (resident range: 36.3-36.9) compared to attending colleagues for TKA, but for UKA scores were up to 11 points worse (resident range: 28.9-38.8). CONCLUSIONS: New Zealand has a high rate of attend-
ing surgeon performed TKA and UKA. Revision rates were Correspondence: Inquiries to GeorgeRichard Storey, Department of Orthopaedics, Christchurch Hospital, 2 Riccarton Avenue, Christchurch, New Zealand; e-mail: [email protected]
INTRODUCTION Surgical training is a complex challenge in health care systems worldwide.1 Surgeon experience is fundamental for resident surgeons before qualifying as independent practitioners, however, patient care and service provision cannot be compromised. The New Zealand orthopedic surgical training program follows a 5-year course which would commence the earliest at 4 years postgraduation. Residents (trainees) are considered senior (advanced) once they have completed their first set of examinations generally after 2 years. Residents are introduced to new surgical procedures as observers or assistants and as their experience progresses are expected to develop into independent surgeons. Almost all resident operating in New Zealand is performed under the supervision of an attending surgeon. The classification of “supervised” used in the New Zealand Joint Registry (NZJR) can range from the attending surgeon being a scrubbed assistant to being out of the operating room available to assist or complete the case if required. Unsupervised training opportunities exist where an attending surgeon allocates appropriate cases to an operating list that is
Journal of Surgical Education & 2018 Association of Program Directors in Surgery. Published by 1931-7204/$30.00 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jsurg.2018.02.011
1
completed independently by a senior resident with adequate experience. The surgeon’s operating experience and patient outcome has received increased attention recently. Working hour restrictions and reduced resident log book numbers can result in lower independent surgeon competency at completion of training.2,3 It is important to note that poor surgical outcomes due to procedural learning curves are not isolated to resident but also affect attending surgeons.4 Joint arthroplasty has been identified as having a considerable learning curve.5 To date, joint arthroplasty has not been well replicated by surgical simulation.5 Hip joint arthroplasty was analyzed most recently by Reidy et al.6 who performed a 10-year review of 584 cases, 43.6% of which were completed by residents. Revision rates and functional outcomes were not significantly different between the 2 surgeon groups.6 Inglis et al.7 analyzed total hip arthroplasty data from the NZJR and found that orthopedic training does not adversely affect Oxford hip scores or revision rate. Edelstein et al.8 investigated the effect of resident involvement and perioperative complications in orthopedic surgery. Of 30,628 cases, 11,071 (36.1%) involved residents and the most common procedure was total knee arthroplasty (TKA) with 5638 cases (18.4%). The authors found that resident involvement did not increase readmission, 30-day mortality, medical or surgical complications. However, the surgical time was greater in cases involving residents (88 min compared to attending surgeons 71 min).8 United Kingdom trainees (resident) have demonstrated no difference in radiological parameters, length of stay, transfusion rate, tourniquet time, or Oxford knee scores (OKS) in TKA compared to their consultant (attending) colleagues.9,10 Surgical outcomes and compilations related to surgeon experience have been studied across a wide range of surgical specialties.8,11-15 Studies in cardiothoracic surgery, colorectal surgery, and gynecological surgery do not show a significant increase in complications or poor outcomes with both supervised and unsupervised resident surgeons.11-15 Increased surgical time among resident surgeons compared to their attending colleagues is a common finding.14,15 One recent NSQIP database study assessed resident participation in a matched cohort of 83,790 commonly performed emergency procedures such as, appendectomy, exploratory laparotomy, and adhesionolysis. Resident participation was independently associated with longer surgical times, intraoperative, and postoperative surgical and medical events—wound, pulmonary, and venous thromboembolic complications.16 This study aims to compare the results for TKA and unicompartmental knee arthroplasty (UKA) between attending (consultant) surgeon and resident (trainee) surgeon groups using two primary outcomes: knee replacement survivorship, measured by revision rate, and the patient′s reported function. We hypothesized that there would be no difference in these outcomes between groups. 2
MATERIALS AND METHODS A retrospective review of data from the NZJR on all elective TKA and UKA from 1999 to 2016 was performed. Data in the NZJR is collected from both the public and private sectors. A recent audit of the registry from 2013 showed a procedure capture rate of 98%.17 Data were collected across 4 surgeon groups where the principle surgeon was an attending surgeon (consultant), a supervised senior resident (advanced trainee), a supervised junior resident (basic trainee), or an unsupervised senior resident. Baseline data included: patient demographic, American Society of Anesthesiologists (ASA) score, and body mass index (BMI). Our primary outcome measures were arthroplasty survivorship, quantified by revision rate per 100 component years and patient reported functional outcome using the OKS. The OKS is a validated, patient reported outcome questionnaire for knee arthroplasty. The score employs a 12 item, 48 point scoring questionnaire which patients complete to assess their pain and function. Since its development in 1998 it has undergone rigorous validation and is used in joint registries in England, New Zealand, and Sweden.18 The NZJR sends OKS questionnaires to a randomly selected 20% of patients 6 months following their arthroplasty and of these 70% are completed and returned. Baseline demographic data were first assessed to determine the comparability of all 4 surgeon groups. Then the revision rate and OKS for each group were directly compared. Secondary points of interest including number of cases in each group and surgical time were also compared.
STATISTICAL METHODS Revision rates were calculated from number of “events” (revisions) and sum total of all patients’ follow-up years. CIs were estimated using a Poisson approximation. Independent t tests were used to compare OKS scores between surgeon groups. A 2-tailed p value of o0.05 used to indicated statistical significance. Regression modeling was used to adjust results for difference is age, ASA and BMI.
RESULTS Baseline Data Overall there were 79,671 TKA and 8854 UKA performed between 1999 and 2016 in New Zealand. The annual incidence of all knee arthroplasty increased on average 9% per year. Baseline data showed that, with the exception of 1 small group (supervised junior residents UKA), the attending surgeon groups were younger, had lower BMI and lower ASA than resident groups, but the absolute differences were Journal of Surgical Education Volume ]/Number ] ] 2018
TABLE 1. Baseline Data Total Knee Joint Arthroplasty Surgeon Attending surgeon Mean (SD) N Supervised Senior resident Mean (SD) N Junior resident Mean (SD) N Unsupervised Senior resident Mean (SD) N Totals Mean (SD) N Unicompartmental knee arthroplasty Attending surgeon Mean (SD) N Supervised Senior resident Mean (SD) N Junior resident Mean (SD) N Unsupervised Senior resident Mean (SD) N Totals Mean (SD) N
BMI
Age
Surgical Time
31.2 (6.0) 18,966
67.5 (9.6) 71,519
81.8 (25.5) 68,633
31.7 (6.2) 1836
69.4 (9.2) 5073
94.3 (24.6) 4875
31.2 (5.6) 500
71.7 (8.0) 1770
87.2 (21.6) 1692
31.9 (6.7) 543
71.8 (9.0) 1309
104.5 (25.3) 1272
31.2 (6.0) 21845
67.7 (9.6) 79671
83.1 (25.7) 76472
29.8 (5.0) 2761
65.7 (9.5) 8550
75.0 (23.5) 8283
31.3 (4.9) 136
68.3 (9.5) 276
83.4 (22.1) 272
27.9 (9) 9
69.0 (14) 14
108.8 (34.0) 14
31.8 (3.2) 4
66.4 (6.3) 14
112.1 (46.5) 13
29.8 (5.0) 2910
65.8 (9.5) 8854
75.4 (23.7) 8582
SD, standard deviation; N, number.
small. Age ranged between 65.7 years and 71.8 years and BMI between 27.9 kg/m2 and 31.9 kg/m2 (Table 1). Comparing ASA between attending surgeon and resident surgeon groups demonstrated that attending surgeons operated on proportionally less ASA 3 and 4 patients compared to resident surgeon groups. For TKA, 24% of patients in the attending surgeon group had an ASA of 3 or 4 compared to 32% of resident patients, for UKA this was 16% and 21% in attending surgeon and resident groups, respectively. For TKA 52.1% were performed in public and 47.9% in private, compared to UKA with 39.6% and 60.4%, respectively. Attending surgeons operated on the highest percentage of all knee joint arthroplasty completing 90% and 97% of TKA and UKA, respectively. The highest percentage of knee arthroplasty performed by residents was 16% in 2010 and this has steadily decreased to a current rate of 9.8%. Mean surgical time for TKA was shortest for attending surgeons (81.8 min) and longest for unsupervised senior
residents (104.5 min), a total of 23 minutes longer. In UKA, shortest times were among the attending surgeon group with a mean of 75.0 minutes and longest in the unsupervised senior resident group with 112.1 minutes, 37 minutes longer (Fig).
Journal of Surgical Education Volume ]/Number ] ] 2018
3
FIGURE. Mean surgical time (min) for attending surgeon and residents in total and unicompartmental knee arthroplasty.
TABLE 2. Revision Rates Total Knee Joint Arthroplasty Revision Rate Confidence Interval (95%) Total Component Events, (Per 100 Lower Upper Years n Component Years)
N (%)
Surgeon
Attending surgeon 67,539 (90%) Supervised Senior resident 4861 (6.4%) Junior resident 1716 (2.3%) Unsupervised Senior resident 1272 (1.7%) Unicompartmental knee arthroplasty Attending surgeon 8550 (96.7%) Supervised Senior resident 276 (3.1%) Junior resident 14 (0.1%) Unsupervised Senior resident 14 (0.1%)
457,823.8
2226
0.49
0.47
0.51
26,074.8 11,822.0
126 26
0.48 0.22
0.40 0.14
0.57 0.32
6707.7
31
0.46
0.31
0.66
59,699.9
710
1.19
1.10
1.28
1660.7 111.3
20 1
1.20 0.90
0.74 0.02
1.86 5.00
83.7
2
2.39
0.29
8.63
supervised senior residents with rates of 0.49 and 0.48/100 component years, respectively. Although the unsupervised senior resident group compromised a small subset (N ¼ 1272 TKAs), their revision rate of 0.46/100 component years was not significantly different from the other groups. Attending surgeons performed most of the UKA cases in the registry (96.7%). There was no difference in revision rates between attending surgeons and supervised senior residents despite the small numbers (1.19 and 1.20/100 component years). The supervised junior residents and unsupervised senior residents performed very few cases making any statistical comparison meaningless.
Outcome Analysis The great majority of knee replacements were performed due to primary osteoarthritis (94.6% of TKA and 97.8% of UKA). The number of cases due to other indications were too small for any meaningful comparative analysis and therefore only osteoarthritis cases were included in the survival and functional analysis. Revision rates for TKA and UKA are shown for each surgeon group in Table 2. For TKA, supervised junior residents demonstrated the lowest revision rate (0.22/100 component years). This was statistically significant when compared to attending surgeons and TABLE 3. Oxford Knee Scores
Total Knee Joint Arthroplasty p Value As Compared to
Surgeon
N
Attending surgeon 20,404 Supervised Senior resident 1029 Junior resident 476 Unsupervised Senior resident 246
Senior Junior Senior Mean Score (SD) Attending Surgeon Resident (S) Resident (S) Resident (US) 37.7 (8.0)
-
o0.001
0.016
0.126
36.3 (8.4) 36.8 (7.9)
o0.001 0.016
0.294
0.294 -
0.310 0.860
36.9 (7.5)
0.126
0.310
0.860
-
-
0.102
o0.001
0.010
0.102 o0.001
o0.001
o0.001 -
0.029 0.237
0.010
0.029
0.237
-
Unicompartmental knee arthroplasty Attending surgeon 5800 39.7 Supervised Senior resident 181 38.8 Junior resident 8 28.9 Unsupervised Senior resident 8 33.1
(7.2) (6.6) (10.3) (9.0)
SD, standard deviation; S, supervised; US, unsupervised. 4
Journal of Surgical Education Volume ]/Number ] ] 2018
Surgical training is a considerable challenge in all health care systems. It is essential that residents progress so they can eventually operate unsupervised. Knowledge of surgical outcomes from resident surgeons at each level is invaluable for planning supervised and unsupervised training while maintaining optimal provision of health care. Total knee joint arthroplasty is an increasingly common elective procedure performed in New Zealand, most often for osteoarthritis. Proficiency in this procedure is expected of an orthopedic surgeon at the completion of their training. In this study patients were investigated in 4 surgeon groups (attending surgeon [consultant], supervised senior and junior residents [advanced and basic trainees], and unsupervised senior residents). Baseline data demonstrated that attending surgeons operated on younger, lower BMI, and lower ASA population (with the exception of the small junior resident UKA group having the lowest BMI). These differences were adjusted for in the outcome assessment of implant survivorship and patient reported functional outcome. Attending surgeons performed 90% of TKA and 97% of UKA in New Zealand between 1999 and 2016. United Kingdom residents performed 33.6% of total hip arthroplasty (5% unsupervised) in 2003 to 2004 and United States residents participated in 64% of orthopedic procedures in 2011.6,8 The number of resident performed procedures in New Zealand has been decreasing by 1% a year for the last 5 years. This trend is of concern for future service provision as qualifying surgeons may be less experienced that previous. Currently the New Zealand training program does not have a minimum requirement for the number of surgical procedures that must be performed prior to completing training. This is different to other New Zealand medical and surgical specialties where case load requirements are in place. Comparable industries such as aviation require minimum experience before final qualifications are reached.
Attending surgeons were significantly faster operators than residents in both TKA (82 vs 105 min) and UKA (75 vs 112 min). This finding was consistent with Edelstein et al.8 who showed an, 71 vs 88 minutes for attending surgeon only and resident involved cases respectively. Increased surgical time in resident cases is not isolated to orthopedics and has been noted across a range of surgical specialties.14-16 This increased surgical time is significant for service provision planning. Implant survival measured by revision rates were not significantly different in any of the surgical groups for TKA and UKA. Attending surgeon revision rate for TKA was 0.49/100 component years. Remarkably supervised junior residents showed the lowest TKA revision rate at 0.22/100 component years, however, this was not statistically significant. This is consistent with NZJR data reviewed by Inglis et al.7 on THJR who also did not show a significant difference in revision rate of THJR between attending surgeon and resident surgeon groups. A possible explanation is that junior residents are very closely supervised during their operations and naturally less technically demanding cases would have been selected as appropriate for training initiation. Unfortunately, the NZJR does not record preoperative disease severity hence a comparison between the groups could not be made. However, it is reasonable to expect that attending surgeons would be more likely to operate on the more difficult patients who may be prone to poorer survival and functional outcomes. Furthermore, although the registry contained a large number of cases, the numbers in the resident groups were low and therefore revision events were low. The finding may therefore be a type 2 statistical error, where a true difference was not detected due to the smaller numbers. Functional outcome measured by the validated OKS tool was highest for attending surgeons in both TKA and UKA. Mean scores of 37.7 and 39.7 for TKA and UKA, respectively were achieved in the attending surgeon group. For TKA, supervised senior residents and supervised junior residents had statistically significant lower mean OKS, 1.4 and 0.9 points below that of their attending surgeon colleagues, respectively. The minimum clinically important difference for the OKS is estimated to be between 1 and 5 points. As such, the statistically significant mean differences listed, at best represents a very small clinically relevant difference. This is concordant with a recent study by Beattie et al.10 who found no difference between resident and attending surgeon outcomes in a retrospective study of 636 TKA with both groups achieving median OKS of 34 at 5 years. Ideally, the absolute change in OKS between preoperative and postoperative scores would have been compared in this study but this data is not available from the NZJR. Given that training arthroplasty cases are performed almost exclusively in the public health care sector in New Zealand, the difference may actually represent an underlying different patient demographic with worse
Journal of Surgical Education Volume ]/Number ] ] 2018
5
Six month postoperative OKS were available for a total of 22,155 knee procedures. OKS and comparison p values for TKA and UKA are shown in Table 3. In TKA the scores were highest in the attending surgeon group with a mean score of 37.7. Scores amongst the resident groups ranged from 36.3 for supervised senior residents to 36.9 for unsupervised senior residents. For UKA, supervised junior residents had the lowest OKS at 28.9 and attending surgeons had the highest at 39.7. Poor scores in the supervised junior resident (28.9) and unsupervised senior resident (33.1) groups were statistically significant (p o 0.05) when compared to attending surgeons and supervised senior residents.
DISCUSSION
preoperative disease severity. There is a large amount of unmet need in publically funded services for elective arthroplasty and it is widely acknowledged that the surgical intervention threshold is higher in the public sector.19 If anything, one would expect this difference in public patients to become more apparent with even lower OKS results for patients undergoing operations by residents. The UKA group was significantly different. Functional scores were very different between unsupervised senior residents, supervised junior residents, and attending surgeons. Unsupervised senior residents and supervised junior residents showed respective mean OKS 6.6 and 10.9 points less than the attending surgeon standard, which is clinically significant. Notably the groups sizes were very small (8 cases in each surgeon group). This may represent a greater learning curve in UKA and increased supervision requirement or a markedly different patient group. Previous studies have confirmed a larger learning curve in UKA compared to TKA.20 The limitations of the present study include the retrospective nature of data assessment. Definition of “supervision” is variable in registry data and can range from the supervising surgeon being present and assisting the primary surgeon to not being present in the operating room. Given that preoperative OKS were not available for comparison, absolute improvements in scores could not be analyzed. Finally the registry’s postoperative function questionnaire data is limited by response rate and response quality.
REFERENCES 1. Holt G, Nunn T, Gregori A. Ethical dilemmas in
orthopaedic surgical training. J Bone Joint Surg Am. 2008;90(12):2798-2808. 2. Sher JL, Reed MR, Calvert P, et al. Influencing the
national training agenda, the UK and Ireland orthopaedic elogbook. J Bone Joint Surg Br. 2005;87(9): 1182-1186. 3. Lonergan PE, Mulsow J, Tanner WA, et al. Analysing
the operative experience of basic surgical trainees in Ireland using a web-based logbook. BMC Med Educ. 2011;11:70. 4. Roberts TV, Lawless M, Bali SJ, et al. Surgical
outcomes and safety of femtosecond laser cataract surgery—a prospective study of 1500 consecutive cases. Opthalmology. 2013;120(2):227-233. 5. Nzeako O, Back D. Learning curves in arthroplasty in
orthopaedic trainees. J Surg Educ. 2016;73(4): 689-693. 6. Reidy MJ, Faulkner A, Shitole B, et al. Do trainee
surgeons have an adverse effect on the outcome after total hip arthroplasty? Bone Joint J. 2016;98-B(3): 301-306. 7. Inglis T, Dalzell K, Hooper G, et al. Does orthopaedic
training compromise the outcome in total hip arthroplasty? J Surg Educ. 2013;70(1):76-80. 8. Edelstein AI, Lovecchio FC, Sujata S, et al. Impact of
CONCLUSIONS New Zealand has a high and increasing rate of attending (consultant) surgeon performed TKA and UKA with attending surgeons having faster surgical times for both procedures. Implant survival was not compromised in resident (trainee) performed TKA or UKA. Functional outcomes were comparable between resident and attending surgeons for TKA. In UKA 2 of 3 resident surgeon groups showed significantly worse OKS compared to the attending surgeon standard. This potentially reflects a greater learning curve and supervision requirement for UKA compared to TKA.
resident involvement on orthopaedic surgery outcomes: an analysis of 30,628 patients from the American College of Surgeons National Surgical Quality Improvement Program Database. J Bone Joint Surg Am. 2014;96(15):e131 1-11. 9. Mahaluxmivala J, Bankes MJ, Nicolai P, et al. The effect
of surgeon experience on component positioning in 673 pres fit condylar posterior cruciate sacrificing total knee arthroplasties. J Arthroplasty. 2001;16(5):635-640. 10. Beattie N, Maempel S, Roberts G, Brown P, Walmsley
P. Supervised registrar-performed surgery does not adversely affect medium-term function outcomes following total knee replacement. Bone Joint J. 2016;98-B (Supp 12):3.
ACKNOWLEDGMENTS
11. Singh
No external funding or author payment was received for this study. The NZJR is funded by contributions from surgeons, Accident Compensation Corporation (ACC), the New Zealand government, and Southern Cross Hospital, however, these funds are used to run the registry only.
12. Goodwin AT, Birdi I, Ramesh TPJ, et al. Effect of
6
Journal of Surgical Education Volume ]/Number ] ] 2018
KK, Aitken RJ. Outcome in patients with colorectal cancer managed by surgical trainees. Br J Surg. 1999;86(10):1332-1336. surgical training on outcome and hospital costs in coronary surgery. Heart. 2001;85(4):454-457.
13. Borowski DW, Ratcliffe AA, Bharathan B, et al.
Involvement of surgical trainees in surgery for colorectal cancer and their effect on outcome. Colorectal Dis. 2008;10(8):837-845. 14. Igwe E, Hernandez E, Rose S, et al. Resident
participation in laparoscopic hysterectomy: impact of trainee involvement on operative times and surgical outcome. Am J Obstet Gynecol. 2014;211(5):484 e1-7. 15. Virk S, Bowman S, Bannon PG. Equivalent outcomes
after coronary artery bypass graft surgery performed by consultant and trainee surgeons: a systematic review and meta-analysis. Cardiothoracic Surg Educ Train. 2015 (3):1-8. 16. Kasostakis G, Aliya L, Beda S, et al. Trainee partic-
ipation is associated with adverse outcomes in
Journal of Surgical Education Volume ]/Number ] ] 2018
emergency general surgery: an analysis of the National Surgical Quality Improvement Program Database. Ann Surg. 2014;260(3):483-493. 17. Hooper G. The ageing population and increased
demand for joint replacement. NZ Med J. 2013; 126:1377. 18. Murray DW, Fitzpatrick R, Rogers K, et al. The use of
the Oxford hip and knee scores. J Bone Joint Surg Br. 2007;89-B(8):1010-1014. 19. Gwynne-Jones D. Quantifying the demand for hip
and knee replacement in Otago, New Zealand. NZMJ. 2013;126(1377):7-17. 20. Borus T, Thornhill T. Unicompartmental knee arthro-
plasty. J Am Acad Orthop Surg. 2008;16(1):9-18.
7
Does Orthopaedic Training Compromise the Outcome in Knee Joint Arthroplasty? Richard Storey, MD,* Chris Frampton, PhD,† David Kieser, MD, FRACS, PhD,* Ramez Ailabouni, MD, FRACS,* and Gary Hooper, MD, FRACS* *
Department of Orthopaedic Surgery and Musculoskeletal Medicine, University of Otago, Christchurch, New Zealand; and †Department of Medicine, University of Otago, Christchurch, New Zealand
OBJECTIVE: This study investigates knee joint arthroplasty and compares the outcomes between attending (consultant) orthopedic surgeons and resident (trainee) surgeons. DESIGN: Retrospective review and comparison of knee
joint arthroplasty outcomes between 4 surgeon groups (attending, supervised senior and junior residents, and unsupervised senior residents). Measured outcomes were implant survival (revision rate) and patient reported functional outcomes, measured by Oxford knee score (OKS). SETTING: New Zealand arthroplasty service. PARTICIPANTS: Seventeen years of knee joint arthroplasty
not increased in resident surgeon groups. Postoperative function was not reduced by a clinically significant amount in TKA in any of the resident surgeon groups but was reduced in supervised junior resident and unsupervised senior resident surgeon groups for UKA. ( J Surg Ed C ]:]]]-]]]. J 2018 Association of Program Directors in Surgery. Published by Elsevier Inc. All rights reserved.) KEY WORDS: knee joint arthroplasty, learning curve,
functional outcomes, orthopedic training, resident surgeon COMPETENCIES: Patient Care, Practice Based Learning
and Improvement and Systems Based Practice
data from the New Zealand Joint Registry (NZJR) was reviewed. RESULTS: The New Zealand Joint Registry (NZJR) data
showed 79,671 total knee arthroplasties (TKA) and 8854 unicompartmental knee arthroplasties (UKA) performed between 1999 and 2016. Attending surgeons performed 90% and 97% of TKA and UKA, respectively. The number and proportion of resident performed knee joint arthroplasty has decreased. Faster operation times was observed in the attending surgeon group. Attending surgeon revision rate was 0.49 and 1.19/100 component years for TKA and UKA, respectively, this was not significantly increased in resident surgeon groups. Postoperative OKS was 37.7 and 39.7 for attending surgeon performed TKA and UKA, respectively. Mean OKS were less than 2 points worse in resident groups (resident range: 36.3-36.9) compared to attending colleagues for TKA, but for UKA scores were up to 11 points worse (resident range: 28.9-38.8). CONCLUSIONS: New Zealand has a high rate of attend-
ing surgeon performed TKA and UKA. Revision rates were Correspondence: Inquiries to GeorgeRichard Storey, Department of Orthopaedics, Christchurch Hospital, 2 Riccarton Avenue, Christchurch, New Zealand; e-mail: [email protected]
INTRODUCTION Surgical training is a complex challenge in health care systems worldwide.1 Surgeon experience is fundamental for resident surgeons before qualifying as independent practitioners, however, patient care and service provision cannot be compromised. The New Zealand orthopedic surgical training program follows a 5-year course which would commence the earliest at 4 years postgraduation. Residents (trainees) are considered senior (advanced) once they have completed their first set of examinations generally after 2 years. Residents are introduced to new surgical procedures as observers or assistants and as their experience progresses are expected to develop into independent surgeons. Almost all resident operating in New Zealand is performed under the supervision of an attending surgeon. The classification of “supervised” used in the New Zealand Joint Registry (NZJR) can range from the attending surgeon being a scrubbed assistant to being out of the operating room available to assist or complete the case if required. Unsupervised training opportunities exist where an attending surgeon allocates appropriate cases to an operating list that is
Journal of Surgical Education & 2018 Association of Program Directors in Surgery. Published by 1931-7204/$30.00 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jsurg.2018.02.011
1
completed independently by a senior resident with adequate experience. The surgeon’s operating experience and patient outcome has received increased attention recently. Working hour restrictions and reduced resident log book numbers can result in lower independent surgeon competency at completion of training.2,3 It is important to note that poor surgical outcomes due to procedural learning curves are not isolated to resident but also affect attending surgeons.4 Joint arthroplasty has been identified as having a considerable learning curve.5 To date, joint arthroplasty has not been well replicated by surgical simulation.5 Hip joint arthroplasty was analyzed most recently by Reidy et al.6 who performed a 10-year review of 584 cases, 43.6% of which were completed by residents. Revision rates and functional outcomes were not significantly different between the 2 surgeon groups.6 Inglis et al.7 analyzed total hip arthroplasty data from the NZJR and found that orthopedic training does not adversely affect Oxford hip scores or revision rate. Edelstein et al.8 investigated the effect of resident involvement and perioperative complications in orthopedic surgery. Of 30,628 cases, 11,071 (36.1%) involved residents and the most common procedure was total knee arthroplasty (TKA) with 5638 cases (18.4%). The authors found that resident involvement did not increase readmission, 30-day mortality, medical or surgical complications. However, the surgical time was greater in cases involving residents (88 min compared to attending surgeons 71 min).8 United Kingdom trainees (resident) have demonstrated no difference in radiological parameters, length of stay, transfusion rate, tourniquet time, or Oxford knee scores (OKS) in TKA compared to their consultant (attending) colleagues.9,10 Surgical outcomes and compilations related to surgeon experience have been studied across a wide range of surgical specialties.8,11-15 Studies in cardiothoracic surgery, colorectal surgery, and gynecological surgery do not show a significant increase in complications or poor outcomes with both supervised and unsupervised resident surgeons.11-15 Increased surgical time among resident surgeons compared to their attending colleagues is a common finding.14,15 One recent NSQIP database study assessed resident participation in a matched cohort of 83,790 commonly performed emergency procedures such as, appendectomy, exploratory laparotomy, and adhesionolysis. Resident participation was independently associated with longer surgical times, intraoperative, and postoperative surgical and medical events—wound, pulmonary, and venous thromboembolic complications.16 This study aims to compare the results for TKA and unicompartmental knee arthroplasty (UKA) between attending (consultant) surgeon and resident (trainee) surgeon groups using two primary outcomes: knee replacement survivorship, measured by revision rate, and the patient′s reported function. We hypothesized that there would be no difference in these outcomes between groups. 2
MATERIALS AND METHODS A retrospective review of data from the NZJR on all elective TKA and UKA from 1999 to 2016 was performed. Data in the NZJR is collected from both the public and private sectors. A recent audit of the registry from 2013 showed a procedure capture rate of 98%.17 Data were collected across 4 surgeon groups where the principle surgeon was an attending surgeon (consultant), a supervised senior resident (advanced trainee), a supervised junior resident (basic trainee), or an unsupervised senior resident. Baseline data included: patient demographic, American Society of Anesthesiologists (ASA) score, and body mass index (BMI). Our primary outcome measures were arthroplasty survivorship, quantified by revision rate per 100 component years and patient reported functional outcome using the OKS. The OKS is a validated, patient reported outcome questionnaire for knee arthroplasty. The score employs a 12 item, 48 point scoring questionnaire which patients complete to assess their pain and function. Since its development in 1998 it has undergone rigorous validation and is used in joint registries in England, New Zealand, and Sweden.18 The NZJR sends OKS questionnaires to a randomly selected 20% of patients 6 months following their arthroplasty and of these 70% are completed and returned. Baseline demographic data were first assessed to determine the comparability of all 4 surgeon groups. Then the revision rate and OKS for each group were directly compared. Secondary points of interest including number of cases in each group and surgical time were also compared.
STATISTICAL METHODS Revision rates were calculated from number of “events” (revisions) and sum total of all patients’ follow-up years. CIs were estimated using a Poisson approximation. Independent t tests were used to compare OKS scores between surgeon groups. A 2-tailed p value of o0.05 used to indicated statistical significance. Regression modeling was used to adjust results for difference is age, ASA and BMI.
RESULTS Baseline Data Overall there were 79,671 TKA and 8854 UKA performed between 1999 and 2016 in New Zealand. The annual incidence of all knee arthroplasty increased on average 9% per year. Baseline data showed that, with the exception of 1 small group (supervised junior residents UKA), the attending surgeon groups were younger, had lower BMI and lower ASA than resident groups, but the absolute differences were Journal of Surgical Education Volume ]/Number ] ] 2018
TABLE 1. Baseline Data Total Knee Joint Arthroplasty Surgeon Attending surgeon Mean (SD) N Supervised Senior resident Mean (SD) N Junior resident Mean (SD) N Unsupervised Senior resident Mean (SD) N Totals Mean (SD) N Unicompartmental knee arthroplasty Attending surgeon Mean (SD) N Supervised Senior resident Mean (SD) N Junior resident Mean (SD) N Unsupervised Senior resident Mean (SD) N Totals Mean (SD) N
BMI
Age
Surgical Time
31.2 (6.0) 18,966
67.5 (9.6) 71,519
81.8 (25.5) 68,633
31.7 (6.2) 1836
69.4 (9.2) 5073
94.3 (24.6) 4875
31.2 (5.6) 500
71.7 (8.0) 1770
87.2 (21.6) 1692
31.9 (6.7) 543
71.8 (9.0) 1309
104.5 (25.3) 1272
31.2 (6.0) 21845
67.7 (9.6) 79671
83.1 (25.7) 76472
29.8 (5.0) 2761
65.7 (9.5) 8550
75.0 (23.5) 8283
31.3 (4.9) 136
68.3 (9.5) 276
83.4 (22.1) 272
27.9 (9) 9
69.0 (14) 14
108.8 (34.0) 14
31.8 (3.2) 4
66.4 (6.3) 14
112.1 (46.5) 13
29.8 (5.0) 2910
65.8 (9.5) 8854
75.4 (23.7) 8582
SD, standard deviation; N, number.
small. Age ranged between 65.7 years and 71.8 years and BMI between 27.9 kg/m2 and 31.9 kg/m2 (Table 1). Comparing ASA between attending surgeon and resident surgeon groups demonstrated that attending surgeons operated on proportionally less ASA 3 and 4 patients compared to resident surgeon groups. For TKA, 24% of patients in the attending surgeon group had an ASA of 3 or 4 compared to 32% of resident patients, for UKA this was 16% and 21% in attending surgeon and resident groups, respectively. For TKA 52.1% were performed in public and 47.9% in private, compared to UKA with 39.6% and 60.4%, respectively. Attending surgeons operated on the highest percentage of all knee joint arthroplasty completing 90% and 97% of TKA and UKA, respectively. The highest percentage of knee arthroplasty performed by residents was 16% in 2010 and this has steadily decreased to a current rate of 9.8%. Mean surgical time for TKA was shortest for attending surgeons (81.8 min) and longest for unsupervised senior
residents (104.5 min), a total of 23 minutes longer. In UKA, shortest times were among the attending surgeon group with a mean of 75.0 minutes and longest in the unsupervised senior resident group with 112.1 minutes, 37 minutes longer (Fig).
Journal of Surgical Education Volume ]/Number ] ] 2018
3
FIGURE. Mean surgical time (min) for attending surgeon and residents in total and unicompartmental knee arthroplasty.
TABLE 2. Revision Rates Total Knee Joint Arthroplasty Revision Rate Confidence Interval (95%) Total Component Events, (Per 100 Lower Upper Years n Component Years)
N (%)
Surgeon
Attending surgeon 67,539 (90%) Supervised Senior resident 4861 (6.4%) Junior resident 1716 (2.3%) Unsupervised Senior resident 1272 (1.7%) Unicompartmental knee arthroplasty Attending surgeon 8550 (96.7%) Supervised Senior resident 276 (3.1%) Junior resident 14 (0.1%) Unsupervised Senior resident 14 (0.1%)
457,823.8
2226
0.49
0.47
0.51
26,074.8 11,822.0
126 26
0.48 0.22
0.40 0.14
0.57 0.32
6707.7
31
0.46
0.31
0.66
59,699.9
710
1.19
1.10
1.28
1660.7 111.3
20 1
1.20 0.90
0.74 0.02
1.86 5.00
83.7
2
2.39
0.29
8.63
supervised senior residents with rates of 0.49 and 0.48/100 component years, respectively. Although the unsupervised senior resident group compromised a small subset (N ¼ 1272 TKAs), their revision rate of 0.46/100 component years was not significantly different from the other groups. Attending surgeons performed most of the UKA cases in the registry (96.7%). There was no difference in revision rates between attending surgeons and supervised senior residents despite the small numbers (1.19 and 1.20/100 component years). The supervised junior residents and unsupervised senior residents performed very few cases making any statistical comparison meaningless.
Outcome Analysis The great majority of knee replacements were performed due to primary osteoarthritis (94.6% of TKA and 97.8% of UKA). The number of cases due to other indications were too small for any meaningful comparative analysis and therefore only osteoarthritis cases were included in the survival and functional analysis. Revision rates for TKA and UKA are shown for each surgeon group in Table 2. For TKA, supervised junior residents demonstrated the lowest revision rate (0.22/100 component years). This was statistically significant when compared to attending surgeons and TABLE 3. Oxford Knee Scores
Total Knee Joint Arthroplasty p Value As Compared to
Surgeon
N
Attending surgeon 20,404 Supervised Senior resident 1029 Junior resident 476 Unsupervised Senior resident 246
Senior Junior Senior Mean Score (SD) Attending Surgeon Resident (S) Resident (S) Resident (US) 37.7 (8.0)
-
o0.001
0.016
0.126
36.3 (8.4) 36.8 (7.9)
o0.001 0.016
0.294
0.294 -
0.310 0.860
36.9 (7.5)
0.126
0.310
0.860
-
-
0.102
o0.001
0.010
0.102 o0.001
o0.001
o0.001 -
0.029 0.237
0.010
0.029
0.237
-
Unicompartmental knee arthroplasty Attending surgeon 5800 39.7 Supervised Senior resident 181 38.8 Junior resident 8 28.9 Unsupervised Senior resident 8 33.1
(7.2) (6.6) (10.3) (9.0)
SD, standard deviation; S, supervised; US, unsupervised. 4
Journal of Surgical Education Volume ]/Number ] ] 2018
Surgical training is a considerable challenge in all health care systems. It is essential that residents progress so they can eventually operate unsupervised. Knowledge of surgical outcomes from resident surgeons at each level is invaluable for planning supervised and unsupervised training while maintaining optimal provision of health care. Total knee joint arthroplasty is an increasingly common elective procedure performed in New Zealand, most often for osteoarthritis. Proficiency in this procedure is expected of an orthopedic surgeon at the completion of their training. In this study patients were investigated in 4 surgeon groups (attending surgeon [consultant], supervised senior and junior residents [advanced and basic trainees], and unsupervised senior residents). Baseline data demonstrated that attending surgeons operated on younger, lower BMI, and lower ASA population (with the exception of the small junior resident UKA group having the lowest BMI). These differences were adjusted for in the outcome assessment of implant survivorship and patient reported functional outcome. Attending surgeons performed 90% of TKA and 97% of UKA in New Zealand between 1999 and 2016. United Kingdom residents performed 33.6% of total hip arthroplasty (5% unsupervised) in 2003 to 2004 and United States residents participated in 64% of orthopedic procedures in 2011.6,8 The number of resident performed procedures in New Zealand has been decreasing by 1% a year for the last 5 years. This trend is of concern for future service provision as qualifying surgeons may be less experienced that previous. Currently the New Zealand training program does not have a minimum requirement for the number of surgical procedures that must be performed prior to completing training. This is different to other New Zealand medical and surgical specialties where case load requirements are in place. Comparable industries such as aviation require minimum experience before final qualifications are reached.
Attending surgeons were significantly faster operators than residents in both TKA (82 vs 105 min) and UKA (75 vs 112 min). This finding was consistent with Edelstein et al.8 who showed an, 71 vs 88 minutes for attending surgeon only and resident involved cases respectively. Increased surgical time in resident cases is not isolated to orthopedics and has been noted across a range of surgical specialties.14-16 This increased surgical time is significant for service provision planning. Implant survival measured by revision rates were not significantly different in any of the surgical groups for TKA and UKA. Attending surgeon revision rate for TKA was 0.49/100 component years. Remarkably supervised junior residents showed the lowest TKA revision rate at 0.22/100 component years, however, this was not statistically significant. This is consistent with NZJR data reviewed by Inglis et al.7 on THJR who also did not show a significant difference in revision rate of THJR between attending surgeon and resident surgeon groups. A possible explanation is that junior residents are very closely supervised during their operations and naturally less technically demanding cases would have been selected as appropriate for training initiation. Unfortunately, the NZJR does not record preoperative disease severity hence a comparison between the groups could not be made. However, it is reasonable to expect that attending surgeons would be more likely to operate on the more difficult patients who may be prone to poorer survival and functional outcomes. Furthermore, although the registry contained a large number of cases, the numbers in the resident groups were low and therefore revision events were low. The finding may therefore be a type 2 statistical error, where a true difference was not detected due to the smaller numbers. Functional outcome measured by the validated OKS tool was highest for attending surgeons in both TKA and UKA. Mean scores of 37.7 and 39.7 for TKA and UKA, respectively were achieved in the attending surgeon group. For TKA, supervised senior residents and supervised junior residents had statistically significant lower mean OKS, 1.4 and 0.9 points below that of their attending surgeon colleagues, respectively. The minimum clinically important difference for the OKS is estimated to be between 1 and 5 points. As such, the statistically significant mean differences listed, at best represents a very small clinically relevant difference. This is concordant with a recent study by Beattie et al.10 who found no difference between resident and attending surgeon outcomes in a retrospective study of 636 TKA with both groups achieving median OKS of 34 at 5 years. Ideally, the absolute change in OKS between preoperative and postoperative scores would have been compared in this study but this data is not available from the NZJR. Given that training arthroplasty cases are performed almost exclusively in the public health care sector in New Zealand, the difference may actually represent an underlying different patient demographic with worse
Journal of Surgical Education Volume ]/Number ] ] 2018
5
Six month postoperative OKS were available for a total of 22,155 knee procedures. OKS and comparison p values for TKA and UKA are shown in Table 3. In TKA the scores were highest in the attending surgeon group with a mean score of 37.7. Scores amongst the resident groups ranged from 36.3 for supervised senior residents to 36.9 for unsupervised senior residents. For UKA, supervised junior residents had the lowest OKS at 28.9 and attending surgeons had the highest at 39.7. Poor scores in the supervised junior resident (28.9) and unsupervised senior resident (33.1) groups were statistically significant (p o 0.05) when compared to attending surgeons and supervised senior residents.
DISCUSSION
preoperative disease severity. There is a large amount of unmet need in publically funded services for elective arthroplasty and it is widely acknowledged that the surgical intervention threshold is higher in the public sector.19 If anything, one would expect this difference in public patients to become more apparent with even lower OKS results for patients undergoing operations by residents. The UKA group was significantly different. Functional scores were very different between unsupervised senior residents, supervised junior residents, and attending surgeons. Unsupervised senior residents and supervised junior residents showed respective mean OKS 6.6 and 10.9 points less than the attending surgeon standard, which is clinically significant. Notably the groups sizes were very small (8 cases in each surgeon group). This may represent a greater learning curve in UKA and increased supervision requirement or a markedly different patient group. Previous studies have confirmed a larger learning curve in UKA compared to TKA.20 The limitations of the present study include the retrospective nature of data assessment. Definition of “supervision” is variable in registry data and can range from the supervising surgeon being present and assisting the primary surgeon to not being present in the operating room. Given that preoperative OKS were not available for comparison, absolute improvements in scores could not be analyzed. Finally the registry’s postoperative function questionnaire data is limited by response rate and response quality.
REFERENCES 1. Holt G, Nunn T, Gregori A. Ethical dilemmas in
orthopaedic surgical training. J Bone Joint Surg Am. 2008;90(12):2798-2808. 2. Sher JL, Reed MR, Calvert P, et al. Influencing the
national training agenda, the UK and Ireland orthopaedic elogbook. J Bone Joint Surg Br. 2005;87(9): 1182-1186. 3. Lonergan PE, Mulsow J, Tanner WA, et al. Analysing
the operative experience of basic surgical trainees in Ireland using a web-based logbook. BMC Med Educ. 2011;11:70. 4. Roberts TV, Lawless M, Bali SJ, et al. Surgical
outcomes and safety of femtosecond laser cataract surgery—a prospective study of 1500 consecutive cases. Opthalmology. 2013;120(2):227-233. 5. Nzeako O, Back D. Learning curves in arthroplasty in
orthopaedic trainees. J Surg Educ. 2016;73(4): 689-693. 6. Reidy MJ, Faulkner A, Shitole B, et al. Do trainee
surgeons have an adverse effect on the outcome after total hip arthroplasty? Bone Joint J. 2016;98-B(3): 301-306. 7. Inglis T, Dalzell K, Hooper G, et al. Does orthopaedic
training compromise the outcome in total hip arthroplasty? J Surg Educ. 2013;70(1):76-80. 8. Edelstein AI, Lovecchio FC, Sujata S, et al. Impact of
CONCLUSIONS New Zealand has a high and increasing rate of attending (consultant) surgeon performed TKA and UKA with attending surgeons having faster surgical times for both procedures. Implant survival was not compromised in resident (trainee) performed TKA or UKA. Functional outcomes were comparable between resident and attending surgeons for TKA. In UKA 2 of 3 resident surgeon groups showed significantly worse OKS compared to the attending surgeon standard. This potentially reflects a greater learning curve and supervision requirement for UKA compared to TKA.
resident involvement on orthopaedic surgery outcomes: an analysis of 30,628 patients from the American College of Surgeons National Surgical Quality Improvement Program Database. J Bone Joint Surg Am. 2014;96(15):e131 1-11. 9. Mahaluxmivala J, Bankes MJ, Nicolai P, et al. The effect
of surgeon experience on component positioning in 673 pres fit condylar posterior cruciate sacrificing total knee arthroplasties. J Arthroplasty. 2001;16(5):635-640. 10. Beattie N, Maempel S, Roberts G, Brown P, Walmsley
P. Supervised registrar-performed surgery does not adversely affect medium-term function outcomes following total knee replacement. Bone Joint J. 2016;98-B (Supp 12):3.
ACKNOWLEDGMENTS
11. Singh
No external funding or author payment was received for this study. The NZJR is funded by contributions from surgeons, Accident Compensation Corporation (ACC), the New Zealand government, and Southern Cross Hospital, however, these funds are used to run the registry only.
12. Goodwin AT, Birdi I, Ramesh TPJ, et al. Effect of
6
Journal of Surgical Education Volume ]/Number ] ] 2018
KK, Aitken RJ. Outcome in patients with colorectal cancer managed by surgical trainees. Br J Surg. 1999;86(10):1332-1336. surgical training on outcome and hospital costs in coronary surgery. Heart. 2001;85(4):454-457.
13. Borowski DW, Ratcliffe AA, Bharathan B, et al.
Involvement of surgical trainees in surgery for colorectal cancer and their effect on outcome. Colorectal Dis. 2008;10(8):837-845. 14. Igwe E, Hernandez E, Rose S, et al. Resident
participation in laparoscopic hysterectomy: impact of trainee involvement on operative times and surgical outcome. Am J Obstet Gynecol. 2014;211(5):484 e1-7. 15. Virk S, Bowman S, Bannon PG. Equivalent outcomes
after coronary artery bypass graft surgery performed by consultant and trainee surgeons: a systematic review and meta-analysis. Cardiothoracic Surg Educ Train. 2015 (3):1-8. 16. Kasostakis G, Aliya L, Beda S, et al. Trainee partic-
ipation is associated with adverse outcomes in
Journal of Surgical Education Volume ]/Number ] ] 2018
emergency general surgery: an analysis of the National Surgical Quality Improvement Program Database. Ann Surg. 2014;260(3):483-493. 17. Hooper G. The ageing population and increased
demand for joint replacement. NZ Med J. 2013; 126:1377. 18. Murray DW, Fitzpatrick R, Rogers K, et al. The use of
the Oxford hip and knee scores. J Bone Joint Surg Br. 2007;89-B(8):1010-1014. 19. Gwynne-Jones D. Quantifying the demand for hip
and knee replacement in Otago, New Zealand. NZMJ. 2013;126(1377):7-17. 20. Borus T, Thornhill T. Unicompartmental knee arthro-
plasty. J Am Acad Orthop Surg. 2008;16(1):9-18.
7