The limitations of using operating room utilisation to allocate surgeons more or less surgical block time in the USA
Professor of Anesthesia and (by courtesy) of Health Research & Policy Program Director, Anesthesia Residency Department of Anesthesia Stanford University School of Medicine Stanford, California, USA E-mail: email@example.com
The article by Dr Pandit and colleagues  in this issue reviews the trade-off between providing surgeons enough operating room (OR) time capacity such that all operations can be performed without incurring an unreasonable waiting time and the inevitable additional cost of making fully staffed OR time available but not necessarily used.
Here, I review the method that many USA hospitals use to make decisions about planning OR block time for surgeons. Traditionally, this decision is reassessed every 3–6 months, and is based on the measured utilisation of that block of OR time by the surgeon. For instance, if a surgeon consistently has 98% utilisation of the OR time, the hospital will try to open another OR on another day to accommodate the surgeon’s business. At first glance, this general approach seems to make sense. However, there are important limitations and potential problems that need to be understood. Among these is how utilisation is defined.
Most USA hospitals perform all cases scheduled by its surgeons, provided a case can be done safely. This reflects the desire to retain and grow surgeons’ practices, to enhance market share and reputation, and to fulfill community-service missions. Also, on average the hospital payment for nearly every surgical case exceeds its variable cost, so almost all cases contribute positively to a hospital’s finances .
On which day was the OR suite less efficient?
A fundamental question in OR management is described in Fig. 1. Imagine an OR suite with two ORs. On Monday (Fig. 1a), three cases by one surgeon were performed in OR 1, finishing at 14.00 with the nurses and an anaesthetist scheduled to work an 8-h day from 07.00 to 15.00. That same Monday, in OR 2, a separate surgeon had a list of cases that ran late until 16.00, such that staff had to stay 1 h past their regularly scheduled 8-h shift. This hour would be considered an over-utilised hour. The next day (Fig. 1b), OR 1 finishes at 14.30, only 30 min before the end of the shift. Meanwhile, OR 2 runs on longer than expected with the last patient not leaving the OR until 17.00, 2 h after the scheduled finishing time. The fundamental question is: ‘On which day, Monday or Tuesday, has the OR suite been less efficient, taking into account both ORs?’
Measuring OR utilisation
The traditional way many hospitals in the USA answer the question of determining OR efficiency is to measure OR utilisation. This can be computed in two ways: raw and adjusted. Raw utilisation is the total hours of elective surgery performed within OR block time divided by the hours of allocated block time. Raw utilisation penalises surgeons who have many short cases because the turnover times are not included, thus artifically reducing the measured OR utilisation. To compensate, most USA hospitals use adjusted utilisation, which equals the time used (occupancy + set-up + clean-up) divided by the time available.
Using traditional OR utilisation criteria in our example in Fig. 1, we conclude that on Monday in OR 1, utilisation equals 87.5% (7 h of cases ÷ 8 h of staffed OR time). However, computing utilisation in OR 2 is less clear. Some USA hospitals would define OR utilisation as 100% for OR 2 on Monday. Others would define it as 112.5% to account for the overrun. This is crucial because without standardising what to do with late-running ORs, comparing hospitals’ OR utilisations for benchmarking will yield misleading results.
For Tuesday in OR 1, utilisation equals 93.75% (7.5 h of cases ÷ 8 h of staffed OR time). Utilisation in OR 2 could be 100% if late running cases were not factored in, or 125% if they were.
Yet another complexity in computing OR utilisation is whether to include cases added on the day of surgery. For example, if the OR was scheduled to finish at 14.00, and did finish at 14.00, but then a 1-h case was added and finished at 15.00, would OR utilisation equal 100%? Once the hospital manager determines how to address this in his/her OR information computer system the next questions that usually follow are: (i) ‘Does it matter whether the add-on case was performed by the same surgeon who was in the room, or a different surgeon?’ and (ii) ‘What about a surgeon who does cases out of usual OR block time, in the evening for example – should those be added to the OR utilisation determination?’ Different hospitals address these questions differently.
Historically, overall OR suite utilisation above 80% has been considered very good, with 90% a hard-to-achieve maximum given the need for some unused capacity to ‘fit in’ cases from the waiting list (Table 1).
Table 1. Overall adjusted operating room (OR) utilisation rates for a hospital surgical suite (traditional thinking).
Not realistic on regular basis; need capacity for add-ons
Exceptional, based on maximum effort
Good scheduling algorithm
Poor management, excessive constraints, uneven caseload (but n.b. free-standing surgery centres have low utilisation so are able to provide OR time tomorrow if requested)
Computation of OR utilisation remains important in many facilties in the USA because of the way in which data are used. If a surgeon has a high OR utilisation then efforts are made to provide more OR time. On the other hand, if a surgeon has low OR utilisation, then some of the OR time may be removed . This approach has important limitations (Table 2).
Table 2. Top 10 limitations to using operating room (OR) utilisation as criteria for changing the amount of OR block time provided to a surgeon .
Optimal OR utilisation differs among surgical specialties: • Not all surgeons can achieve utilisation > 90% . For example, orthopaedic surgeons specialising in joint replacements are much more likely to know in advance which cases they are doing and fill up their OR time months ahead of time, than cardiac surgeons who only know a day or two in advance which patients admitted with unstable angina will need revascularisation.
Utilisation is poorly related to contribution margin: • Contribution margin = hospital revenue generated by a surgical case, less all the variable costs due to the hospitalisation. • Theoretically, any case with a contribution margin > 0 that can be done safely is financially worth performing from the facility’s perspective. • In the USA, fee-for-service hospitals have a positive contribution margin for almost all elective cases, mostly due to a large percentage of OR costs’ being fixed. As much as tenfold variability in contribution margin exists among surgeons in different specialties . • If a surgical procedure is paid by case, then a slow surgeon will have high utilisation, but low profitability for hospital.
Estimates of utilisation are not accurate for individual surgeons: • For example, during the preceding 3 months, a surgeon’s measured adjusted OR utilisation was 65%. The 95% CI for the surgeon’s average adjusted utilisation was 38–83% . • The reason for the wide confidence intervals in OR utilisation is random variations in the numbers of patients each week who request to be scheduled for surgery. To illustrate, when a quality control mean chart of block utilisation is made, decisions should be based on 3 × SD above and below the mean which is a large range.
Efforts to increase utilisation can impair growth of some subspecialties: • For example, local media have publicised the opening of a new robotic prostatectomy service but it has a low utilisation. The hospital benefits strategically, in spite of low utilisation of OR time.
Utilisation percentages can be artificially inflated: • For example, a surgeon purposefully slows down at the end of the day to utilise his block fully, worrying that his block time will be reduced if his utilisation is low. This will lead to a lower contribution margin, assuming that payments for that type of surgery are fixed.
Subspecialties with longer cases will have lower utilisations: • This is because long cases cannot be fitted neatly into block time as easily as several short cases. • For example, a surgeon performs complex abdominal cancer cases that average 5 h. Only a single case can be scheduled each day without going past the regularly scheduled end time of 15.00.
Utilisation correlates poorly with waiting times for quick access to free OR time: • Although many facilities believe that high utilisation can be accomplished without sacrificing access to OR time at short notice, these two aspects are not separable. The greater the ‘OR utilisation’, the less the convenience (ability to book cases when desired) for surgeons and patients. • This mathematical reality is defined by the Pollaczek-Khintchine curve which shows that as utilisation of any system (not necessarily the OR) increases, the waiting time for those wanting to enter the system will increase. Waiting time doubles when utilisation goes from 80% to 90%, and tenfold from 90% to 99% utilisation .
Increasing utilisation to high levels can reduce revenues: • For example, overall adjusted OR utilisation at a hospital is 90%. Administrators see the ORs as 10% empty, and think that revenues could be increased without any change in staffing costs. The hospital signs a contract to become a preferred provider for an additional insurance plan: the hospital will perform surgery on plan members at a discounted rate in the hopes of increasing OR utilisation above 90%. However, overall revenues decrease, to their surprise !• This is because new discounted patients compete and displace more lucrative patients for a fixed amount of OR time.
Utilisation does not indicate the potential for future expansion: • Some procedures are becoming extinct, while others are growing. Thus, there is always a need to anticipate growth areas and provide more OR time for those .
Utilisation is poorly related to other constraints such as ICU capacity: • If a surgeon with high utilisation performs cases needing ICU postoperatively, and the ICUs are full, then giving more OR time to this surgeon ultimately won’t make a difference because of the ICU bottleneck.
Measuring overall OR efficiency
How well the OR suite runs depends on whom you ask. The majority of hospital costs for surgical inpatients accrue in the short time the patient is in the OR . As a result, the hospital administrator may want the most ‘throughput’ of cases with the least total cost. On the other hand, the surgeon in the USA typically desires an early morning start, rapid turnover, low cancellation rates, and cases that start on time . Nurse managers may focus on flexibility to move cases around, disposable supply costs per case, the percentage of cases complying with sterilisation policy, and adequate reserve capacity for add-on/emergency cases. Risk managers, on the other hand, will want to know the percentage of patients without injury (e.g. wrong-sided surgery). A scoring system to asses how well an OR suite is functioning from the hospital’s perspective is available  (Table 3).
Table 3. An operating room (OR) efficiency scoring system, with representative figures from the best performing USA hospitals .
Excess staffing costs
Start-time tardiness (mean total delay of start times for elective cases per OR per day)
< 45 min
Case cancellation rate
Admission delays to recovery unit (workdays with at least one delay > 10 min because of lack of space)
Prediction bias (bias in case duration estimates per 8 h of OR time) [16, 17]
< 5 min
Prolonged turnovers (proportion > 60 min)
Nothing is more important in OR management than to allocate the right amount of OR time to each service on each day of the week. Two thirds of all USA hospital expenses are for wages, salaries and benefits of caregivers and other staff, and this is also the single most important factor driving up hospital spending (http://www.aha.org/aha/content/2010/pdf/10costofcaring.pdf (accessed 18/03/2010). As a result, correctly matching staffing with the surgical caseload is the priority.
Optimal allocation of OR time needs to be based on historical use by a particular service (i.e. unit of OR allocation such as surgeon, group, department or specialty), and then using computer software to minimise the amount of underutilised time and the more expensive overutilised time [18, 19].
Getting back to our example (Fig. 1), on Monday in OR 1, there would be 1 h of underutilised time. The excess staffing cost would be 12.5% (1 ÷ 8 h). In contrast, in OR 2 on the same day, 9 h of OR cases are performed with staff scheduled to work 8 h. This means that the excess staffing cost is 25% (1 ÷ 8 h = 12.5%, which is then multiplied by a factor of 2 to include the additional overtime costs which include monetary and morale costs of staff’s staying late). On Tuesday in OR 1, there would be 0.5 h of underutilised time. The excess staffing cost would be 6.25% (0.5 ÷ 8 h). In OR 2 that day, 10 h of OR cases are performed with staff scheduled to work 8 h. This means that the excess staffing cost is 50% (2 ÷ 8 h = 25%, multiplied by same ‘fudge’ factor of 2). Thus, we can say that Monday has been more efficient than Tuesday!
It can also be stated that the inefficiency of use of OR time for Monday equals 3 h (1 h of underutilised time in OR 1, plus 1 h of overutilised time in OR 2 multiplied by 2 for the overtime costs). And, for Tuesday, OR inefficiency totals 4.5 h (0.5 h underutilised time for the room that finished 30 min early, plus 2 h overutilised time for the room that finished late multiplied by 2).
Thus, optimising OR staffing costs requires finding a balance between overtime and finishing early. As Pandit et al. so correctly point out: as you increase utilisation, waiting times go up, and it is difficult and probably counter-productive to try to improve utilisation above a certain point. Utilisation of the OR depends greatly on the average length of time that patients wait for surgery, and increases as this queue time increases.
In summary, many USA facilities allocate OR time based on historical utilisation. This assumes that there is a fixed amount of regularly scheduled OR time, usually called block time. However, this fixed hours system does not apply when facilities make OR time available for all its surgeons’ patients, even if the day’s list of cases is expected to finish after the end of block time. In this setting, OR time should be allocated to maximise OR efficiency, not historical utilisation!