Editor’s Spotlight/Take 5: Resident Participation is Not Associated With Worse Outcomes After TKA

Not every surgeon reading this teaches residents or registrars, but every practicing surgeon reading this once was a trainee. And it’s likely that everyone reading this will someday be in the care of someone who has come through surgical training. For those reasons, the findings from Dr. David Jevsevar’s study from the Geisel School of Medicine at Dartmouth College, “Resident Participation is Not Associated With Worse Outcomes After TKA” [1] should be of compelling interest.

This is a difficult question to study. Large numbers are needed if one wishes to detect differences in uncommon complications, and so use of a large, national registry to answer it might be tempting. Unfortunately, those registries lack the necessary detail on the endpoint that matters most to patients undergoing orthopaedic surgery: Whether the operation resulted in less pain and greater function. To get that sort of granularity, a single-institution drill-down is better, but studies of that sort often are underpowered to identify small but important differences that might result from having residents involved during surgery.

Dr. Jevsevar’s group found the sweet spot between those two methodological approaches by plumbing the depths of a large registry at an institution that has made extensive commitments to tracking its results. They found that having residents involved in surgery was not associated with lower validated outcomes scores or a lower likelihood of achieving a clinically important improvement on the Patient-Reported Outcomes Measurement Information System (PROMIS) outcomes tool used. While resident participation often was associated with longer surgical times, the differences (4 to 12 minutes of added time when residents of varying levels were involved) were generally smaller than the between-surgeon differences, where 28 minutes separated the fastest attending from the slowest. And more importantly, resident involvement was not associated with increased length of stay nor an increased likelihood that the patient would be discharged to an extended-care facility, which might be a surrogate for serious, in-hospital complications. I think that we and our patients can find all of this reassuring.

Although they would have liked to evaluate intraoperative complications, the number of these (four of 1626 TKAs) was far too small, and probably is too small for serious statistical analysis even in large, national registries. We should find this reassuring, as well.

These results extend and clarify those of a recent study drawn from a large, national database of patients having hip-fracture surgery [9]. That study concluded that resident involvement was not associated with a greater risk of death or serious morbidity, but that patients cared for in collaboration with residents were more likely to have surgical times in excess of 90 minutes and lengths of stay longer than 2 weeks. These seemingly concerning findings prompted a commentary that I believe drew immodest inferences from a dataset that had considerable shortcomings, and conflated association with causation: “There is a price to be paid for resident education … If one were to extrapolate those added time-related costs across all intertrochanteric fracture surgeries performed in the US each year, the total added annual costs could be astronomical” [7]. That study [9] did not compare the association of resident involvement with those of other between-surgeon differences as did Dr. Jevsevar’s group [1], it dichotomized important continuous variables in seemingly arbitrary ways (why treat surgical time or length of stay as an over/under?). And most importantly, it did not control for the many important differences among the centers that contributed patients to the National Surgical Quality Improvement Program® (NSQIP) database. It seems possible, if not likely, that differences among hospitals dwarfed any impact that resident participation may have had. Case complexity at tertiary-care referral centers (which are largely populated by residents) is vastly different from that of community hospitals (which are not). It is important to note that both of those kinds of hospitals—and others as well—contribute data to NSQIP. Factors like these must be considered before one uses that sort of dataset to suggest that “resident involvement in a case be stopped after a certain amount of operative time” [7].

While the findings of Dr. Jevsevar’s team are reassuring in many ways, some important questions deserve further discussion and future study. Join me in the Take 5 interview with Dr. Jevsevar as we get his insights on these important topics.

Take Five Interview with David S. Jevsevar, MD MBA, senior author of “Resident Participation is Not Associated With Worse Outcomes After TKA”

Seth S. Leopold MD: Congratulations on this thoughtfully done study. I think the most emotional part of this topic for patients is the perception that that they could be harmed as the result of a resident’s participation. With fewer than a handful of intraoperative complications in more than 1600 procedures, you could not study this. Given your experience and the other research that has tried to address this topic, how do you discuss it with patients?

David S. Jevsevar MD, MBA: For me, this is a complicated but important question. I practiced in a community, nonacademic setting before returning to academic practice. It was sobering to think about giving up some control within the operating room when working with residents. In the end, I have found my own practice transitions (from community practice back to academics) helpful in answering the question.

Like most of us practicing in academic settings, I am confronted with patients who either are concerned about the level of resident involvement or frankly, ask that the resident not be involved with the procedure at all. My approach in responding to these patients is to educate them on the facts as we know them. Specifically, I address in a general fashion the available evidence [2, 6, 8, 9] and specifically that I have found no compelling research that suggests that involving residents results in poorer clinical or patient-reported outcomes with resident participation. As you note, we are also fortunate in that intraoperative complications are rare. I also try to explain that my role is to “teach” the resident the technical as well as cognitive aspects of hip and knee arthroplasty. The level of resident-specific participation is determined by level of experience, procedural proficiency, and preparation for contributing to the care of the specific patient. I also add, and thoroughly believe, that operating with an inquisitive and attentive resident improves my care because the resident might appropriately challenge me on my decision-making or technique; and is another set of eyes to look for problems.

The opportunity to lead, mentor, and work with young orthopaedic surgeons in training is an exhilarating experience, and one that I am truly grateful and honored to be able to enjoy. Our specialty moves forward because of the drive and inquisitive nature of our trainees. Patient care is undoubtedly improved because of their diligence, thoughtfulness, dedication, and commitment to our patients and our specialty.

Dr. Leopold: Do you think that question has been adequately answered by the studies that have specifically addressed the associations between resident participation and morbidity and surgical complications [2, 6, 8, 9]? If not, how might future studies get the answers you think we need on that association?

Dr. Jevsevar: The challenging aspect of answering this question is understanding the baseline for complications and mortality associated with and without resident participation. While we currently spend a great deal of time risk-adjusting for comorbidities as well as educational and socioeconomic factors, our current institutional and national registries do not offer sufficiently fine detail to assess the technical differences of the procedures performed. It seems obvious to me, for example, that the surgical complexity of even primary hip and knee arthroplasties that I perform at a safety-net academic medical center typically are more difficult and complicated than those I performed in a community practice setting. These relevant but difficult-to-measure differences confound comparisons about the influence residents may have on surgical results, since the settings and patients themselves are different in so many important ways.

With complications and mortality being so uncommon, prospective clinical trials are not ideal because of the massive numbers needed. Multicenter national or international registry data often lack the clinical granularity and detail necessary to draw scientifically valid conclusions for these rare clinical occurrences. Unfortunately, our current medicolegal climate also creates challenges in accurate reporting of intraoperative complications. As in our work, we had to rely on formal chart review to identify the presence or absence of an intraoperative complication. Future study designs would need to create a “safe” reporting environment where surgeons felt no threat for accurately and comprehensively reporting safety and complication issues encountered during surgery.

Dr. Leopold: Most of the large studies I’ve read on the questions you explored have used the NSQIP database; you decided not to. Though I see the temptation to use NSQIP, I think its participating institutions differ too widely to try to attribute differences to a single element of the academic medical center care experience (resident participation). What alternatives did you consider in terms of study setting for your work, and how did you eventually settle on your chosen single-center approach?

Dr. Jevsevar: As you know, one of the biggest issues is that most national databases, including NSQIP, do not include patient-reported outcomes. We believe that reporting on both clinical and patient-reported outcomes is critically important when measuring overall outcome on an intervention. For this kind of analysis, our institutional database reduces the number of confounding factors in play. First, we do not have fellows participating in the surgical procedures. In addition, most procedures in our dataset involved only a single resident, rather than more than one. Finally, our center uses a single-vendor implant strategy. This decreased variation in implant choice and improved proficiency because of the benefits of repeated use.

With fewer confounding variables to address, we could investigate more closely the impact of residents at specific levels of training. We found the variability of the level of resident participation between surgeons to be interesting, and likely worthy of greater investigation. While we asked the attending surgeons for their estimation of resident participation, we are now investigating the difference between levels of participation between faculty and resident point of views. Although the factors I noted earlier supported our decision to use a single-center approach, I do believe there is a benefit to expanding this work to other centers to evaluate between-institution differences if they exist.

Dr. Leopold: As you know, statistical differences are not always clinically important; increasing operative time by 4 to 12 minutes as you found doesn’t seem like much to me, though perhaps others might disagree. How should we interpret your findings about surgical time as we think about training residents in real-world settings? Specifically, how should surgeons balance their educational commitments to residents and the risks associated with prolonged surgical time?

Dr. Jevsevar: As our data showed, the difference in operative time was greater among attending surgeons (28 min) than it was between those procedures that did or did not involve a resident. If we take the argument that increased surgical time leads to increased complications to its logical conclusion, we then would need to suggest that all patients be sent to the fastest surgeons. In the management of clinical practices, trade-offs are made based on many factors. A technically sound and fast surgeon is preferable to a less technically proficient and slow surgeon. I would argue that a technically sound but slow surgeon is preferable to a fast but technically challenged surgeon. We need to continue to define our outcomes of interest more precisely so that we can better determine our goals.

Anecdotally, we found that closure was a primary driver of resident participation surgical time. Those surgeons more actively involved in fascial and skin closure had shorter operative times, regardless of the resident’s year in training. If the attending surgeon was not actively involved with closure, resident times varied depending on year in training and experience. I try to impart to our residents that there are portions of procedures that should normally move expeditiously so that we can devote more time to those elements of the procedures that are especially complicated or demanding without unduly extending the overall duration of the intervention. My advice to my residents is the same as one of my mentors gave me: “You don’t have to be fast to be good, but you don’t have to be slow to be careful.”

Dr. Leopold: If the pressures to perform surgery with maximum efficiency continue to mount, resident involvement risks becoming less substantive. And yet we know that if we do not adequately educate residents, the costs will be a great deal higher as complications accrue after our residents graduate. Let’s imagine for a moment that resident involvement incurs added costs, as seems likely at least with respect to operative time. How should the stakeholders involved—whether insurers or the Centers for Medicare & Medicaid Services (CMS), and perhaps hospitals and medical centers—help align the incentives so meaningful graduate medical education in surgery can continue?

Dr. Jevsevar: This is a difficult and challenging question. The decision to have our federal government pay for resident training was made during Lyndon Johnson’s presidency in 1965. The legislation stated: “Educational activities enhance the quality of care in an institution, and it is intended, until the community undertakes to bear such education costs in some other way, that a part of the net cost of such activities (including stipends of trainees, as well as compensation of teachers and other costs) should be borne to an appropriate extent by the hospital insurance program” [3]. To date, no other members of the community have come forward to directly help to defray the costs of medical education. The US Congress capped the number of resident trainees to 100,000 in 1997, and this results in an aggregate cost of training of more than USD 5 billion [4, 5].

These numbers are staggering and are accentuated by the productivity demands of most orthopaedic surgeons practicing in academic medical centers. While a core ethical tenet of healthcare is to practice efficiently to create a higher value and sustainable health system, our data show that residents do slow the process to some degree. In theory, this is to be accounted for by the higher Diagnosis Related Groups reimbursements that academic medical centers receive from CMS. We don’t see these differential dollars being redistributed at the department level to offset the potential diminished productivity of surgeons engaging in resident training.

Some programs have addressed this issue internally by creating service and nonservice operating room days. On service days, residents participate as usual in the care of patients and productivity expectations are lessened. Nonservice days are typically characterized by no resident participation with support by dedicated mid-level providers or fellows and higher productivity expectations. The obvious downside of this approach is that two distinct levels of patient care are generated within the same facility.

With academic medical centers largely fulfilling the role of safety-net hospitals, I believe that it is imperative that our government continue to support our training endeavors. I also firmly believe that all of the stakeholders in healthcare, government, payers, employers, physicians, and patients are responsible for helping to sustain the system. Without continued and even greater support of those dedicated faculty willing to educate our future surgeons, the quality of our surgeons will diminish. Inherent to this process, sustainable and equitable compensation formulas for orthopaedic surgeons who participate in resident education will be required.

David S. Jevsevar MD, MBA