As a fellow in an era when surgeons and engineer teams had successfully introduced metal backing to the tibial component and thought that the same innovation would remedy patellar complications, I was instructed in the vicissitudes of new technology. In an attempt to eliminate fractures, the length of the central lug on a new, cemented, metal-backed patellar implant was shortened to conserve bone. The implant was under compression, wasn’t it? Before patellar clamps were introduced, my contribution to the reconstruction was often to squeeze the implant to the bone while methacrylate polymerized. Perhaps my thumb grip was not up to the task, or, as it turns out, there really is considerable shear at that interface; however, most of the early prototypes promptly came loose and were revised. The surgeon, my mentor, passed on the message ruefully: “Be careful, even with small changes to something that works.”
I have had a long involvement with knee arthroplasty design and must disclose that implants I know well are featured in this study. I have been among the first to use new devices and have supported a general introduction on a few occasions. The study by Peltola et al. has identified a major opportunity for all of us to improve patient safety and the results of surgery. It has done so by bringing forward data on a topic that surgeons may have faced with ignorance if not denial: the dangers inherent in introducing new technology, specifically new knee replacements.
Modern life, and not just medicine, is everywhere fraught with the need to adapt. Our culture is predicated on innovation with an economy whose very survival depends on change. Incessant, ubiquitous mini-learning curves are a way of life in hospitals: the intern on a new rotation, the scrub technician covering trauma cases, or the attentive worker in the central supply room using the autoclave independently for the first time. Every trainee has a first knee with a valgus deformity to correct. Although supervised, the surgery will probably take longer. Somebody, somewhere in the long chain of collaboration, is ascending a personal learning curve every time an arthroplasty is performed.
For more than a century, the good physician has been expected to “keep up” and embrace progress. “Newness” has been synonymous with “goodness.” Tragically, the metal-on-metal hip arthroplasty debacle is headline news around the world and has sensitized the public to innovation. These are two distinct issues, however: failures that originate with inherent flaws in the implants versus the challenge of introducing something better, which happens to be different. The two can overlap, when introducing new devices with inherent, although unrecognized, flaws. The database for this study, although excellent, is not sufficiently detailed to discriminate among the many elements that constitute a learning curve. The learning curve is probably less about the products the registry tracks than the processes associated with their introduction.
Peltola and colleagues monitored the first fifteen implantations of ten designs for reoperations in the three years after surgery. These results were compared with the experience with established implantations (starting at the 100th) of (1) the same prosthesis and (2) the most commonly used prosthesis of the ten. The database at the heart of this study, the Finnish Arthroplasty Register, is of very high quality, but, like all registries, it tracks implants and basic patient demographics. Surgeon identity is not available even under coding. It is conceivable that all fifteen cases could have been performed by one surgeon for one implant or by fifteen different surgeons for another. Neither the cause nor the nature of the reoperation is available, so deeper analysis is unavailable. An inopportune infection in the first fifteen would be fundamentally different from failure to lock in the modular polyethylene accurately or an intraoperative fracture.
Prostheses identified as “new” in this study may have been recognizable modifications of a predecessor or a radical departure from anything previously used in a particular hospital. The prosthesis might be implanted with similar instrumentation or something completely different functionally and conceptually. Poor teamwork in surgical teams has been implicated in adverse events to patients1. Educational efforts (or the lack thereof) directed at surgeons, manufacturers’ representatives, or hospital staff are beyond the scope of the study. The need for preparation, experience, and familiarity with a device has implications beyond the individual surgeon. Four of the ten implants studied demonstrated increased hazard ratios for failure in the first fifteen cases, but six did not. Two of the four implants with early increased hazard ratios for failure later exhibited superior performance in the established period. Adverse results during the learning curve are not inevitable.
Several points in this excellent paper merit discussion. One lies in the hypothesis: “We hypothesized that, in total knee arthroplasty, the learning effect is dependent on the implant model, with some implants easier and safer to begin using than others.” This implies some inherently troublesome physical quality of the four manufactured components that are implanted in the patient. It would be more accurate to state that “it appears some implants were more difficult and hazardous to begin to use than others.” “Were” of course implies that it could have been otherwise with the right planning and support. The greater unarticulated message of the study is that implants are primarily indicators of a “process,” and that process is primarily education.
“The manufacturers should consider the learning effect when designing implants and instrumentation.” For generations, surgeons involved in the development of implants have recognized that if the best conceivable implant required extraordinary skills to ensure success, it would be an unmitigated failure. I can assure the investigators that ease of use and clarity of concept are the driving forces in the world of prosthesis development.
“The surgeons should thoroughly familiarize themselves with the new knee implants before use.” This point is clear, basic, and irrefutable. Failure to prepare is indefensible. But the question then becomes, to what extent? How much time can be devoted and with what resources? Few practitioners have independent access to cadavers. Few parties other than industry, with its incentive to ensure strong clinical results, have been supportive in the past. Budgets and regulatory bodies now regulate industry activity rigidly. It is the essence of professionalism that surgeons educate themselves, but increasingly, as expenses mount and reimbursements decline, time and money for substantive education are in short supply.
The ill-fated patellar button from the early days of knee arthroplasty development was never used more than a half dozen times and only by one surgeon. The short peg was an intended improvement based on a flawed model of knee mechanics. It would have been easy to introduce, but the results were abysmal. It was not a “learning curve” phenomenon, as no amount of time would have improved the result.
More importantly for the current study, the admonition directed narrowly to surgeons in the conclusion of this study misses a key facet of surgical success: coordination of the extended surgical team2-4. Surgeon education must be dynamic and shared5. Preparation for crisis6 and change creates a stronger, collectively smarter team that is better equipped for daily routine. A new implant is the perfect opportunity to convene the extended team outside the pressure of regular hours to communicate and coordinate. An egalitarian approach where every team member contributes will reveal important and pertinent issues far beyond the new implant that have a bearing on how the team functions. The best surgeons are those who inspire and educate, who organize, confide in, and listen to their team7. You can’t climb that learning curve alone.