Table 4 First of the 3 scenarios for which the described practice of the operating room manager of Hospital A did not match decision-making based on maximizing efficiency of use of operating room time

4a. Adapted scenario printed on single page in landscape orientation using Arial 16 point font

  Among the 13 ORs typically started on Thursdays, there are 12 ORs allocated to specific

services. The services "General Surgery," "Orthopedics," and "Pain Medicine" have each been

allocated one OR. General Surgery has scheduled 9 hr of cases in its OR. Orthopedics has 6 hr

of cases scheduled into its OR. Surgeons in Pain Medicine have scheduled a 3 hr case. The

remaining ORs are fully booked. General Surgery wants to schedule another case.

  OR management, staffing,^† and case scheduling decisions are made based on four ordered

priorities: Safety, Access, OR efficiency, and Reducing patient waiting on the day of surgery.

  General Surgery can book the case, because it has access to OR time on what ever work day

the service chooses. If based on surgeon availability, staff knowledge, equipment availability, etc.,

the case can be performed safely in the OR time allocated for Pain Medicine, then Pain Medicine

would have its OR time released for the General Surgery case. This is because Pain Medicine has

the largest difference between the allocated and scheduled hours of cases (i.e., would be expected

to have the most under-utilized OR time on the day of surgery).

4b. Scenario as stored in computerized library (line breaks added for clarity)

  Among the [S6] ORs typically started on [S5], there are [S1] ORs allocated to specific

services. The services "[S2]," "[S3]," and "[S4]" have each been

allocated one OR. [S2] has scheduled 9 hr of cases in its OR. [S3] has 6 hr

of cases scheduled into its OR. Surgeons in [S4] have scheduled a 3 hr case. The

remaining ORs are fully booked. [S2] wants to schedule another case.

  OR management, staffing,^† and case scheduling decisions are made based on four ordered

priorities: Safety, Access, OR efficiency, and Reducing patient waiting on the day of surgery.

  [S2] can book the case, because it has access to OR time on what ever work day

the service chooses. If based on surgeon availability, staff knowledge, equipment availability, etc.,

the case can be performed safely in the OR time allocated for [S4], then [S4]

would have its OR time released for the [S2] case. This is because [S4] has

the largest difference between the allocated and scheduled hours of cases (i.e., would be expected

to have the most under-utilized OR time on the day of surgery).

4c. Unsolicited comments based on non-adapted scenario, shown with the corresponding stored parameters. None of the comments relate to the process of decision-making.

S5, S6 We don't run 13 ORs on Thursdays

S1, S6 All of our ORs are planned for specific services every day

S2, S3, S4 We allocate "block" time by surgeon, not specialty

S2, S5 General Surgery doesn't do many cases on Thursdays

S6 We don't do pain medicine in our ORs, but at a clinic

4 d. Steps performed by software to adapt parameters S1 through S5 automatically

1.

Using the most recent 9 four-week periods of data [1, 12], calculated turnover times between each pair of sequential cases in the same OR on the same day. Turnover time was defined as time from when a patient exited the OR until the next patient entered the same OR on the same day [2]. Turnovers longer than 90 min were set equal to 90 min. The use of 9 periods is based on previous study using training-testing datasets, which showed that each increase in the number of four-week periods resulted in a statistically significant reduction in mean labor costs [12].

2.

Calculated the OR workload for each service on each day [1]. The workload was defined as total hours of OR time and turnover times, excluding urgent cases [1]. Although the services were specialties for the specific hospital shown in the example, no such assumption was made [1].

3.

Assigned each combination of service and day of the week with a mean OR workload less than 5.60 hr to the pseudo-specialty representing open, unblocked, first-scheduled, first-served "OTHER" time for low workload specialties (i.e., the service is not assigned its own OR) [1]. The value of 5.60 hr is the optimal break even point [13, 14] based on over-utilized OR time costing 1.5 times as much as under-utilized OR time and OR workload by service and day of the week having a standard deviation of 0. Derivation can be shown by simple algebraic manipulation of equation (11) in Reference [21]. By Monte-Carlo simulation, results are insensitive to typical ranges of standard deviations [22].

4.

For each combination of service and day of the week, the total inefficiency of use of OR time was calculated for choices of 0 ORs, 1 OR, 2 ORs, etc., and the choice with the smallest inefficiency was used as the allocation.

5.

Tested the statistical assumption of randomness (e.g., no trend over time) as described in our review article [1], and if assumption violated, stopped and used non-adapted scenario for facility. Note that this situation has not occurred, as expected from the sample size chosen [1, 12].

6.

Set S1 to be the number of ORs allocated to individual specialties, excluding OTHER.

7.

Determined which combinations of service and day of the week were allocated 1 OR.

8.

Chose the earliest day of the week with at least 3 specialties allocated 1 OR. If situation did not exist, then scenario was not included in the collection of scenarios given to the facility. Otherwise, set S2, S3, and S4 to be the first 3 services allocated 1 OR for that day of the week, alphabetically. Set S5 to be the day of the week.

4e. Automatic selection of S6

9.

From among all cases performed during the most recent 9 four-week periods, selected those starting between 6:45 AM and 9:30 AM.

10.

From among those cases, selected the first case of the day in each OR on each workday.

11.

Calculated the median number of first case of the day starts separately for each day of the week.

12.

Set S6 to be the number of ORs from step 11 on the S5 day of the week.