Rationale for Anesthesia Groups to Run Additional Flexible Operating Rooms for Multiple Surgeons Who Have Scheduled More than 8 Hours of Cases

摘要

In this month's issue, Sessler et al.1 show absence of an association between the time of day at which scheduled (elective) general surgery and orthopedic procedures start and both 30-day risk-adjusted mortality and incidence of in-hospital complications. For start times between 7:00 AM and 5:00 PM, the confidence intervals for the odds ratios were narrow (i.e., unlikely that absence of an association was attributable to a type II error). These results are especially important because they differ from recent findings in gastrointestinal endoscopy.2–5 Sessler et al.1 also show no difference in patient outcome between July and August, when there are new trainees, and other months. In this Editorial, we review research in operating room (OR) management to help readers apply the results obtained by Sessler et al. to managerial decision-making. We apply the authors' findings to the various decisions that may be made, from decisions made on the day of surgery to decisions made several months before surgery. On the day of surgery, urgent (add-on) cases are scheduled. The findings by Sessler et al.1 do not apply in this setting, because they studied scheduled cases. There may be provider fatigue and/or resource availability late at night affecting patient outcome, as detected for some transplantation procedures.6,7 Weeks to days before the day of surgery, cases are scheduled into allocated OR time.8–10 The decision is made rationally by not scheduling a case into overutilized OR time unless there is insufficient remaining allocated time for the case.8–10 Issues regarding patient outcome related to the time of day do not arise in the decision.8 A few months before the day of surgery (e.g., before staff scheduling is done), allocated OR time is calculated based on forecasted workload.10–12 This decision is made based on the total workload, not case start times.10–12 The results of Sessler et al.1 do not apply to this decision either. Several months before the day of surgery, surgeons' block times can be readjusted (e.g., days of the week that they operate).13 Because such decisions do not change mean start times or numbers of first case starts, these decisions too are unaffected by the Sessler et al. results. Several months before the day of surgery, additional block time may be planned for one or more surgeons in the hope of future growth in the surgeon's OR workload.14–17 That decision frequently does change the duration of the workday (e.g., cases scheduled up to 6:00 PM instead of up to 3:30 PM) and could be influenced by the findings of Sessler et al. However, in practice, the decision should not be affected. Sessler et al.1 pooled procedures (see the authors' Fig. 1). Their confidence intervals for odds ratios were narrow when pooled, unlike what the result would be for each specified procedure and/or surgeon. Although quality can be built into the contribution margin per OR hour analysis, either as increased utility (revenue) or as reduced cost, the resulting confidence intervals by surgeon for contribution margin per OR hour would be impractically wide.14,18 The potential incremental increase in margin from increasing block time to one specified surgeon would thus be offset by a large increase in risk that the realized contribution margin per OR hour would be significantly different from the expected value.19 The analysis would result in the surgeon not being allocated additional block time, in lieu of the alternative (salvage) decision to allocate the additional block time as first-come, first-scheduled unblocked open OTHER overflow time.16,17 Several months before the day of surgery, the decision may be made to open another such OTHER flexible OR. This is the decision to which the results of Sessler et al.1 do apply. Opening such an OR at a facility with many cases starting beyond an 8-hour workday will result in fewer cases starting that late. The value of the findings of Sessler and colleagues is in showing that patient outc