Key result
Anesthesiologists with the lowest case volume were associated with a higher 30-day hospital readmission rate compared to those with the highest volume.
Why the study?
Does low anesthesiologist case volume increase hospital readmission rates in patients undergoing radical cystectomies and colorectal surgeries?
Does low anesthesiologist case volume increase hospital readmission rates in patients undergoing radical cystectomies and colorectal surgeries?
Answers are grounded in this paper and cite their sources.
This editorial highlights that while low anesthesiologist case volume is associated with higher readmission rates, experience (caseload) may not be a perfect surrogate for true clinical expertise.
Robert Collier, a famous American self-help author, once said, “success is the sum of small efforts, repeated day-in and day-out.”1 Thus, based on a growing number of medical studies, the mantra of “Do More, Be Greater” has been increasingly promoted and marketed by hospital administrators and health system leaders. The data are compelling correlating low-volume hospitals and low-volume surgeons with poor outcomes.2,3 However, there are few studies investigating anesthesiologists’ case volume and outcomes. In this issue of Anesthesia & Analgesia, Jaeger et al4 adopted these methodologies to retrospectively evaluate a large cancer registry database in Ontario, Canada, of radical cystectomy case volumes for anesthesiologists and associated outcomes. Using multivariable analysis while accounting for surgeon and hospital volume, the anesthesiologists who had the lowest volume of these patients (<6.5 cases/year) were associated with a significantly higher hospital readmission rate at 30 days (odds ratio [95% confidence interval] 1.36 [1.1–1.7]) and 90 days (odds ratio 1.36 [1.1–1.7]) when compared with the outcomes of anesthesiologists with the highest volumes (>11.8 cases/year).4 Notably, postoperative mortality and long-term survival were not associated with anesthesiologist volume.4 So, do repetition and experience inevitably lead to success? We believe repetition and experience will not lead to expertise without feedback. K. Anders Ericsson5 described this in his account of deliberate practice and expertise in medicine. Through repetition, one can reach a state of automaticity and competence; however, expert performers move beyond automaticity to a higher level of cognitive and associative performance. In a recent trend, researchers in medicine have tried to tie levels of hospital or provider experience (not levels of expertise) to patient outcome. The results have been mixed because experience may not be a good surrogate for expertise! The present findings by Jaeger et al4 are somewhat surprising and different from the very few previous studies investigating any association between low anesthesiologist volume and poor outcome. Papachristofi et al6 examined nearly 20,000 cardiac surgical patients in a single center over a 10-year period for the effect of anesthesiologist and surgeon volume on mortality. The results suggested that outcome was determined primarily by patient risk profile and that anesthesiologists did not affect patient outcome. Similarly, Gani et al7 studied surgeon and anesthesia provider volume and associated outcomes in 969 patients undergoing pancreatic surgery at a tertiary care center. Variations in patient outcomes were not associated with anesthesia provider volume. Finally, Papachristofi et al8 re-examined anesthesiologist case volume and its potential association with mortality among 10 centers in the United Kingdom and over 110,000 cardiac surgical patients. The authors concluded that there was no association between anesthesiologist case volume and mortality. So, despite the common assumption that experience (in this example, caseload) counts and should lead to improved outcomes, the data are far from conclusive. How can we place the findings of Jaeger et al4 into appropriate perspective and explain them in light of previous findings on this topic involving anesthesiologists? First, the patient population studied is different than previous studies. The anesthesia care for patients undergoing radical cystectomies and colorectal cases is different from patients undergoing cardiac and pancreatic surgery in duration of anesthesia and anesthesia techniques, fluid and blood product requirements, amount of muscle relaxants and narcotics used, and surgical approach. These factors could partially explain the observed differences. Second, the present study used a Canadian cancer database registry to investigate the outcomes. Different national standards and practices (ie, the development of Early Recovery After Surgery [ERAS] initiatives) could play a role in the different outcomes. Third, these studies analyzed different time periods that may not account for the potential improvements in technology, medicines, and surgical approaches that could lead to better or worse outcomes. Despite differences noted between Jaeger’s study and others on this topic, it is important to consider the plausibility of the article’s conclusions; lower volume anesthesiologists could have higher readmission rates when caring for patients undergoing radical cystectomies/colorectal surgeries. This conclusion invokes a presumed mechanism that lower volume anesthesiologists are “less expert.” Could lower volume anesthesiologists lack the experience (ie, expertise) to care for these patients in the intraoperative period and therefore provide them with flawed care? Could the same anesthesiologists provide unnecessary amounts of fluids, blood products, muscle relaxants, and/or narcotics that could lead to an increase in readmission rates? Could the processes of care among lower volume anesthesiologists be less standardized (the lack of ERAS effect) and lead to readmissions? These questions might be answered if we could accurately define “best practice” for such patients. We could then outline deviations in care that likely account for poor outcome. However, these questions were not answered in this article.4 Provider expertise was not explicitly defined, so readers cannot deduce whether an anesthesia provider’s experience (caseload) is indeed the predominant factor in readmission rates. Furthermore, it is unclear as to whether cases were taken over by other anesthesia providers (less or more experienced) that could confound the results. The authors did not investigate factors such as the administration of fluids, blood products, narcotics, muscle relaxants, and the very important postoperative care that could play a role in these readmission rates. Finally, Jaeger et al4 did not investigate the potential effect that standardization and the development of improved processes of care (ERAS, laparoscopy) had on readmission rates. Further prospective studies will need to investigate these factors to better delineate best care and outline the effect of experience on patient outcomes. Additional limitations should be recognized when reflecting on these findings. First, this is a retrospective analysis of a database susceptible to not including all potentially confounding variables, which by definition are variables associated with both the exposure (anesthesiologist volume) and outcome (readmission, mortality, survival). Authors adjust for only 7 patient, provider, and hospital characteristics, but there are undoubtedly many more characteristics in each of these categories that could be contributing to small or large amounts of bias in these results by their absence. Like with most database studies, there can be substantial bias from misclassification of the exposure, the outcomes, and particularly the confounding variables themselves. For any observational study, after adjusting for as many observed confounding variables as possible, an excellent statistical approach is to conduct sensitivity analyses assessing the magnitude of unobserved confounding (eg, size of odds ratios from 1 or more variables) that would have to be present to change the conclusions of the study.9,10 Had Jaeger et al4 conducted such a sensitivity analysis, it is quite possible that only small amounts of unobserved confounding would be needed to change conclusions, which would highlight the fragility of their findings. This follows because their results were close to the significance criterion and only significant if not adjusting for multiple testing. Moreover, the observational study design lends itself to further potential bias in that the data were not specifically collected for the questions posed by the study. In addition, the authors chose to combine colorectal cases and radical cystectomies to increase the sample size of caseloads. Although these cases are similar in duration and surgical approach, they are nonetheless unique. Studies have demonstrated that radical cystectomies may be associated with larger blood loss that could contribute to varying outcomes.11,12 Furthermore, Jaeger et al4 did not focus on a single primary outcome but rather assessed 6 main outcome variables. After controlling for the chance of false-positive findings (maintaining either type I error or, less conservatively, the false discovery rate,13 at 5%), none was significant. Surprisingly, Jaeger et al4 only adjusted for multiple testing in their univariable results, not the primary multivariable results. Finally, Jaeger et al4 formulate their primary exposure of anesthesiologist volume as categorical variables, placing each provider into 1 of 4 quartiles of volume. Although categorizing a continuous exposure might be clinically intuitive because one can discuss the lowest quartile versus other “groups,” substantial information is lost, and the rationale was not obvious. Jaeger et al4 could have first assessed the linearity of the relationship between continuous anesthesiologist volume and outcomes, and they used that assessment to help decide whether categorizing was necessary or appropriate. However, to their credit, they do present results for continuous volume as a sensitivity analysis, and those relationships are, not surprisingly, stronger than those for categorical volume. One could have made an assessment of whether there was a dose–response relationship between anesthesiologist volume and outcomes, and whether the relationship was linear. In summary, we applaud Jaeger et al4 for investigating a fundamental anesthesiology practice and staffing question: Does anesthesiologist case volume contribute to patient outcome? This is the first study evaluating this question in patients undergoing radical cystectomies/colorectal cases. Future studies need to look at this hypothesis in a prospective fashion and meticulously examine the anesthesiologist’s experience but, more importantly, expertise. There may be a yet unidentified bimodal curve in an anesthesiologist’s career where too little or too much experience (ie, “too old”) may adversely affect patient outcomes.14 These questions require rigorous investigation. For now, Jaeger provides some tantalizing insights regarding the effect of experience on outcome. Future studies will seek to define best practice as well as a link between experience and expertise before trying to identify a link between experience and outcome. After all, doing more of the wrong thing…won’t make you better! DISCLOSURES Name: Steven Greenberg, MD, FCCP, FCCM. Contribution: This author involved in manuscript preparation and revisions. Conflicts of Interest: Dr Steven Greenberg has served as a consultant for both CASMED and MERCK. Neither of these conflicts has anything to do with the present manuscript. Name: Torin Shear, MD. Contribution: This author involved in manuscript preparation and revisions. Conflicts of Interest: None. Name: Edward Mascha, PhD. Contribution: This author involved in manuscript preparation and revisions. Conflicts of Interest: None. This manuscript was handled by: Richard C. Prielipp, MD.
No takes yet. Share an insight, caveat, or question.
Greenberg et al. (2017) conducted an editorial in Radical cystectomy and colorectal surgery. Lowest anesthesiologist case volume (<6.5 cases/year) vs. Highest anesthesiologist case volume (>11.8 cases/year) was evaluated on Hospital readmission rate at 30 days (OR 1.36, 95% CI 1.1-1.7). Anesthesiologists with the lowest case volume were associated with a higher 30-day hospital readmission rate compared to those with the highest volume.
Synapse has enriched 10 related papers. Consider them for comparative context: