We appreciate the interest in our work from the officers and committee chairs of the American Society of Anesthesiologists. To lessen the possibility that we have misrepresented or misinterpreted the comments of these authors, in our reply, we quote from their respective letters and identify the specific authors. We provide explanations as to why we think the results and conclusions of our original article are reliable and valid.
Our research was motivated by the previous report from Paoletti and Marty of France, who performed a simulation study to calculate the percentage of days in which there would be waiting for an anesthesiologist in at least one operating room (OR).1Their results were published in the British Journal of Anaesthesia in 2007. Cohen et al. state that the percentage should be high; Paoletti and Marty's simulation study found it was high,1and so did our data analysis.2Thus, the scientifically useful results of our research were principally the time of the day when the percentage risk of waiting was the largest (our second hypothesis) and the parameters most highly affecting those percentage waits.2
1. Cohen et al. write, “Every day in this country, anesthesiologists prioritize which cases to start first, when they may safely leave, and what aspects of care require their presence. Although in some systems staggered starts may not be structurally embedded in the formal operating room ‘schedule,’ they are a reality in practice as anesthesiologists focus on patient safety.”
Paoletti and Marty's mathematical model1was, in retrospect, excellent, based on their results being nearly identical to those we obtained by summing events from anesthesia information management systems data.2Given the concordance, we can rely on the1mathematical model's finding that the principal factors affecting the incidence of waiting in six ORs supervised by two anesthesiologists are: use of staggered starts, mean durations of the critical periods of inductions and emergence, and the shortest durations of cases. These findings and the calculated percentages assume conservatively that any anesthesiologist not occupied in a nonpreemptive task can cross-cover, thus creating a massive team.1,2
2. Cohen et al. comment that “this study was published … based on a methodologically-suspect mathematical model.”
The following is the first sentence of the Abstract from Paoletti and Marty's British Journal of Anaesthesia paper, with the italic emphasis being our addition1:
3. Elsewhere, Abouleish and Stead state that “in the United States, ‘medical supervision’ of anesthesia care by an anesthesiologist differs from ‘medical direction’ of anesthesia care.” Similarly, Cohen et al. comment that “this study … included terminology that was confusing and acted to obfuscate a conclusion relevant to the study hypotheses. … The authors seem to erroneously interchange the terms ‘supervision’ and ‘medical direction.’”
“Numerous hospitals implement a ratio of one anesthetist supervising nonmedically qualified anesthetist practitioners in two or more operating theaters.”
The same language is used by the Brazilian authors, de Oliveira Filho et al. , in the title of their 2008 article in Anesthesia and Analgesia , reporting on how to evaluate the quality of anesthesiologists' supervisory skills3:
“An instrument designed for faculty supervision evaluation by anesthesia residents and its psychometric properties.”
Aware of the possible confusion resulting from the use of differing terminologies by leading anesthesia journals and international authors, by design we did not use the phrases “medical supervision” or “medical direction” in our paper.2Furthermore, the word “direction” was not used in any form anywhere in our paper.2We conducted (and repeated before authoring our response) a PubMed search of the phrases “medical supervision” and “medical direction” in the abstracts of articles in the following journals: Anesthesiology, Anesthesia & Analgesia, Acta Anaesthesiologica Scandinavica, Anaesthesia and Intensive Care, Anaesthesia, British Journal of Anaesthesia, Canadian Journal of Anaesthesia, European Journal of Anaesthesiology, the Japanese Journal of Anesthesiology, the Middle East Journal of Anesthesiology , and the Journal of Clinical Anesthesia. The phrase “medical supervision” was never used in these articles' abstracts. “Medical direction” was used in one Anesthesiology article abstract, one Anesthesia & Analgesia article abstract, and not once in any of the other journals' abstracts. Since Anesthesiology is read globally, and most articles are from countries other than the United States, our use of the generic noun “supervision” seemed more appropriate to us than use of a U.S. billing term.
Yes, both sets of authors are correct. Our results show that2:
4. The two letters commenting on our original research article again hit on similar themes when Abouleish and Stead state that “the word ‘lapses’ is misleading since really what the authors found were ‘overlaps’ based on their self-defined critical portions. They did not demonstrate any lapses in care by the anesthesiologist or the team. They did not study what actually happened; rather they used their broad definitions to determine if potential overlaps would occur.” To this, Cohen et al. add, “an alleged ‘supervision lapse’ could occur when the induction of an anesthetic is delayed for a few minutes while waiting for the medically-directing anesthesiologist …. At most, this delay would result in a possible inconvenience to the surgeon and a decrement in efficiency of perioperative resources.”
“Administrators who want to reduce their anesthesia group's costs by encouraging them to decrease their anesthesiologist supervision ratios need to consider the effect of our findings on the timeliness of first-case starts, which is often a major institutional focus.4At a ratio of one anesthesiologist to three anesthesia providers, it will not be possible to start all ORs simultaneously and have sufficient anesthesiologists to supervise all critical portions of cases on most days. Either the administrators will need to accept the fact that the additional OR often will be delayed from its scheduled start time, or agree to rearrange the OR schedule so that first cases supervised simultaneously by each anesthesiologist will have staggered start times.”
Our endpoint was appropriate because although staggered starts are easy to implement, neither formally adjusting start times nor recommending to anesthesiologists the night before which ORs to start first can be done accurately without statistical calculations of historical workload.4,–,6Anesthesiologists who make the decisions without those calculations make decisions that are worse than random chance.4,6,7
Yes, that is why we separately analyzed these scenarios. We calculated that2:
5. Cohen et al. also state, “in a first scenario, an alleged ‘supervision lapse’ could occur when the induction of an anesthetic is delayed for a few minutes while waiting for the medically-directing anesthesiologist. … In a second scenario … during a potentially deleterious physiologic event. This would be a patient safety concern. Disentanglement of these two distinct scenarios is essential.”
“Fewer than 20% of the minutes of critical portions … were accounted for by minutes with … physiologic events [“considered as critical portions of cases”] (P < 0.0001, mean 14.7%, SE 0.5%). Excluding physiologic events occurring during critical portions reduced the percentage to 13.8% (SE 0.4%).”
We observed that2:
6. According to Abouleish and Stead, “the authors chose to define [induction of general anesthesia]… as when the patient enters the OR to intubation (or the equivalent) + 3 min. Therefore, they include within their definition of the induction the following events: transportation into the OR, movement of the patient from the stretcher to the bed, placement of the IV (if not done in holding), placement of standard monitors, and waiting for the surgeon to arrive. This overly broad definition creates artificial ‘conflicts,’ where none in fact occur.”
“The average peak activity (total providers needed) during cases occurred at the start of the workday for most days (P < 0.0001).”
Thus, what matters are the behaviors, described by Abouleish and Stead, among the first cases of the day. ‘Conflicts’ were underestimated, because few such cases were so brief that emergence would start while induction was ongoing in other first cases1,2:
“The fact that we studied a tertiary hospital with many long cases rather than an outpatient surgery center with short cases is not a limitation because, from the simulation study,1our results would be even stronger for short cases.”
At the studied hospital, the placement of the IV is done in the holding area for nonpediatric cases. The surgeon or a surgical resident is available because otherwise the patient is not brought to the OR. The patient is logged into the anesthesia information management system as entering the OR after the stretcher is brought into the OR and positioned next to the OR table. Thus, the definitions we used to analyze the data from the study hospital were appropriate for that hospital. Because conditions may differ among facilities, we evaluated whether our results were typical1:
“From … the French simulation study1with 24 ORs, a staffing ratio of 1:2, and one additional floater anesthesiologist (i.e. , effective supervision ratio of 1:1.8), the expected incidence of supervision lapses is 12%. We observed a 12% incidence with a supervision ratio of 1:1.7.”
We also evaluated whether our time periods were comparable with those of another U.S. hospital with detailed observational data:8
“The mean number of minutes of critical portions of first-case starts was 22.2 min (95% CI 21.8 to 22.8 min). This observation matched [the] observational finding [of 22 min] reported previously from Yale-New Haven Hospital8(P = 0.29). Thus, the third hypothesis that the mean number of critical minutes for first-case starts would match the anesthesia release time measured by observers8was confirmed.”
Please refer to our previous response to see that we did not make assumption of 22 min. The value of 22 min was used to check that Thomas Jefferson University Hospital data were comparable with that of the hospital (Yale) that had studied induction times by observers.8We did so because indeed the “tasks considered as critical portions of the anesthetic” were the principal limitation of our study. We used the actual time of intubation to determine when the anesthesiologist could likely have left. The 3 min we allowed after intubation accounted for supervisory tasks such as confirming proper placement of the endotracheal tube, ensuring adequate taping of the endotracheal tube, and verifying that the postinduction/intubation blood pressure was in a satisfactory range.
7. Cohen et al. also comment that “of paramount concern to us … a requirement that the anesthesiologist cannot leave the first patient for which he or she induces general anesthesia under medical direction until the patient is ‘turned over to the surgical team’ (mean anesthesia release time was 22.2 min in the study population) … the methodology … create[s] false and overstated ‘supervision lapses.’ Of note, the authors acknowledge this concern as ‘the principal limitations of our study…’ in the Discussion.”
We could not assume a priori that the first cases of the day were the most important from the perspective of supervision. That needed to be tested:
8. Abouleish and Stead comment that, “instead of looking at the first-case starts, the authors chose to also look at other portions of anesthesia care as well.”
“As predicted [by the French simulation study]… the average peak activity (total providers needed) during cases occurred at the start of the workday for most days (P < 0.0001) …. The start of the OR day is the period of time when the anesthesiologist supervision requirement is greatest. Even with lunch breaks included, this result is so robust that changes in the anesthesiologist supervision ratio can be described to administrators simply in terms of the effect on first-case starts. This finding is useful because the psychology of first-case starts is already understood (e.g. , how they are interpreted economically).4As the economics of first-case starts are also fully developed, the decision to stagger first-case starts appropriately versus having more anesthesiologists can be modeled for each facility.”4,–,6,9
Anesthesiologists have led the development of the science of OR management. Hopefully they will also play a large role in its application at their facilities. We stand by the appropriateness of the methodology and the conclusions of our paper.2
