Association between surgeon training grade and the risk of revision following total knee replacement: An analysis of National Joint Registry data.

BACKGROUND

Total knee replacements (TKRs) are performed by surgeons at different stages in training with varying levels of supervision, but we do not know if this is a safe practice or whether trainees achieve equivalent outcomes to consultant-performed TKR. This study aimed to investigate the association between surgeon grade, the supervision of trainees, and the risk of revision following TKR. Revision is defined by the National Joint Registry (NJR) for England and Wales as any procedure to add, remove, or modify one or more components of an implant construct for any reason.

METHODS AND FINDINGS

We conducted an observational study using prospectively collected data from the NJR. We included 953,081 cases in 788,288 adult patients who underwent primary TKR for osteoarthritis (OA), recorded in the NJR between 2003 and 2019. Exposures were surgeon grade (consultant or trainee) and the level of scrubbed consultant supervision of trainees. The primary outcome was all-cause revision, and the secondary outcome was the number of procedures revised for the following indications: aseptic loosening/lysis, infection, progression of OA, unexplained pain, and instability. Flexible parametric survival models (FPM) were incrementally adjusted in the following manner. Model 1 was unadjusted. Model 2 was adjusted for patient-level factors (age, sex, American Society of Anaesthesiologists [ASA] grade, and index of multiple deprivation [IMD] decile). Model 3 was further adjusted for operation-level factors (anaesthetic, approach, fixation, constraint and whether or not the patella was resurfaced). Model 4 was further adjusted for healthcare setting factors (funding source, and year of operation). Trainees performed 96,544 (10.1%) TKRs and were directly supervised by a scrubbed consultant in 63.2% of trainee-performed cases. Trainees achieved comparable outcomes to consultants in terms of the unadjusted cumulative probability of all-cause revision (e.g., 15 years of follow-up: consultant % Failure 4.79 (95% CI [4.67, 4.92]) versus trainee (overall) % Failure 4.75 (95% CI [4.43, 5.10]). Adjusted FPM analysis indicated evidence of an association between trainee-performed TKR and a small increased risk of early all-cause revision up to, but not exceeding, 4 years follow-up (1 year: HR 1.12 (95% CI [1.05, 1.19]), 4 years: HR 1.00 (95% CI [0.95, 1.06]), 16 years: HR 0.89 (95% CI [0.81, 0.98])). This association was not explained by the level of supervision. Further analysis suggested that this association may be attributable to revisions for aseptic loosening/lysis, infection, and progression of OA (i.e., subsequent patellar resurfacing). Limitations of this study relate to its observational design and include: the potential for non-random allocation of cases by consultants to trainees; residual confounding; and the use of the binary variable 'surgeon grade', which does not capture variations in the level of experience between trainees.

CONCLUSIONS

Trainees in England and Wales achieve safe and acceptable all-cause TKR implant survival, with comparable outcomes to consultants. However, adjusted analyses suggest an association between trainee-performed TKR and a small increase in the risk of early all-cause revision. This association may be attributable to factors including aseptic loosening, infection, and progression of OA. Current training practices for TKR in England and Wales are safe in terms of equivalence of all-cause implant survival to consultant-performed TKR, but we have identified areas for potential improvement in trainee outcomes.