Operating Room Utilization: When 85% Beats 99%

Every page I read on this topic treats utilization as the headline number for an operating room, and in the classic definition that number counts turnover time as time used. The consequence is not subtle. In the simulation that produced the 85 to 90 percent figure everyone quotes, one modelled room running the same eight cases scores 76 percent, 87 percent or 99 percent, and the only thing that changes between those three numbers is how long the room takes between patients.

This guide covers the three different numbers people call OR utilization, where the benchmark actually comes from and what it assumes, what the pages recommending it leave out, what a minute of operating room time really costs, and what to report beside the percentage so it stops being able to flatter you.

Three different numbers share one name

Ask three people for your OR utilization and you can get three answers that are all correct, because they are answers to different questions.

• Overall or raw utilization. The classic definition, the one Tyler's paper uses and credits to the OR-management literature before it, is the sum of the time it takes to perform each surgical procedure, including patient preparation in the room, induction and emergence, plus total turnover time, divided by the time available.
• Prime time utilization. The same ratio restricted to the hours the department is actually staffed and resourced, typically something like 7:30 to 15:30 on weekdays. Minutes worked past that window are usually reported separately as over-utilized time instead of being folded into the same percentage, which is also why a turnover-inclusive figure can pass 100 percent: Tyler's table has rows at 110, 112 and 125 percent.
• Block utilization. Minutes a named surgeon or service used inside the block allocated to them, divided by the minutes allocated. This one is about the allocation decision, not the room.

Two decisions inside that formula matter more than the division. The first is whether turnover counts as used, which decides the direction the number moves when your room gets slower. The second is the denominator: staffed hours inside your operating window, not 24 hours, and not rooms nobody was rostered to run. A suite that counts unstaffed evenings in the denominator will look permanently mediocre, and one that counts only the hours it happened to be busy will look permanently excellent.

The same eight cases, three different numbers

The clearest demonstration of what the classic formula rewards sits in a table inside the study everyone cites for the benchmark, and the authors point at it themselves. Tyler, Pasquariello and Chen ran a simulated operating room 50,000 times per scenario and varied the turnover time while holding everything else still. The staffed day is 615 minutes, 7:15 to 17:30. Here is what happened to a day of eight cases.

One simulated room, 615 staffed minutes, log normal case times of 48 minutes give or take 15, no emergencies.

Eight patients in all three rows. Same surgeons, same procedures, same staffed day. The percentage climbs 23 points because more of the day went on turnover, which is cleaning, setup and instrument prep, not idleness, and at 30 minute turnovers the staff also go home about 21 minutes late. Tyler says the mechanism outright: "Because turnover time is included in the calculation of utilization, there was no difference in the utilization that could be achieved within the guidelines when turnover time was lengthened or shortened." What did change was the work. Nine cases fit the day at 10 minute turnovers, seven at 30.

That is the comparison worth carrying: nine cases score 85 percent, seven cases score 86 percent. Two rooms within a point of each other, two cases apart, and the higher number belongs to the room that did less.

Before anyone reaches for a replacement metric, the same table disposes of the obvious candidate. Tyler's own efficiency measure adds unused minutes inside the staffed day to 1.5 times the minutes worked past the end of it. Across those three days it reads 146, 80 and 58, so it prefers the 99 percent day: a room at 76 percent is paying staff for 146 idle minutes. On the nine-versus-seven pair it reads 90 and 85 and prefers the seven-case room too. It is a staffing-match measure, not a throughput measure, and neither it nor utilization can see that eight cases got done. That is the actual lesson here, and it is not that one number should be swapped for another.

All of this holds where turnover sits inside the numerator, which is why the definitional question is not pedantry. Under a case minutes only definition turnover drops out of the numerator entirely: eight 48 minute cases are 62 percent of a 615 minute day whether the turnovers run 10 minutes or 30, and the number falls only when a slower room loses a case. Before you compare your figure with anyone else's, or with your own from two years ago, find out which of the two you have.

Where 85 to 90 percent actually comes from

The number is real, it is peer reviewed, and it does not mean what it is usually used to mean. It comes from a 2003 paper in Anesthesia & Analgesia whose abstract says: "We created a simulation of an OR to define optimum utilization." Not a survey of hospitals, not a registry. A model: case times fitted to a distribution with Arena, the day itself simulated in an Excel spreadsheet, run 50,000 times per scenario.

What went into the model is the part worth knowing:

• Case times came from every adenoidectomy with or without tonsillectomy performed at the Children's Hospital of Philadelphia between 1 January and 30 June 1999. The best fit was a log normal distribution with a mean of 48 minutes and a standard deviation of 15, a coefficient of variation of 0.31.
• The modelled room ran from 7:15 to 17:30, with no emergencies, and patients could not start before their scheduled time because they had not arrived.
• Two service goals were held fixed: an average patient wait of 15 minutes, and finishing the schedule within 15 minutes of the scheduled end of the day.

Inside those conditions, the authors write that "a utilization of 85% to 90% is the highest that can be achieved without delay or running late." The very next sentence of the abstract is the one that almost never travels with the number: "Increasing the variability of case duration decreases the utilization that can be achieved within these targets." Their conclusion repeats it, putting optimum utilization at 85 to 90 percent "depending to a large extent on the variability of case duration."

So the ceiling is a function of how predictable your cases are, and it was derived on one of the most predictable case mixes in surgery: short, high volume, paediatric ENT procedures with a coefficient of variation of 0.31. A general theatre running revisions, cardiac work and unscheduled trauma has a far messier distribution, and by the paper's own logic a lower achievable ceiling. Holding that room to 85 percent is holding it to someone else's variance.

The paper does not settle on one figure for itself either. For the simplest scenario it reports optimum utilization of approximately 90 percent; against the service goals, 85 to 90 percent; as a working figure, "a target utilization of 85%"; approximately 85 percent once longer cases are mixed with shorter ones; and on its own cost measure, a most efficient band of 85 to 95 percent. Five defensible answers in one study, which is a fair signal of how much precision the quoted number can carry.

I should correct my own page here. Until today our guide to provider and room utilization presented this as the one setting where a real benchmark exists. It is a solid finding about a modelled room, which is not the same thing, and I have reworded that page to say so.

What the seven pages on page one leave out

I read what else searchers on this query are handed. I took the pages returned on the first page of results for operating room utilization and how to improve operating room utilization, and asked each one the same five questions: does it name a target, does it cite a source for that target, does its formula include turnover, does it define the denominator, and does it say anywhere that a slower room can score higher. Eight pages, one of which refuses automated requests and returned a 403, leaving seven I could read.

Read on 1 August 2026 through a search API instead of a logged out browser. It samples the advice in circulation; it is not a ranked snapshot of the SERP. These pages also compete with this one in search, and I scored them myself against questions of my own choosing, so read the table as an argument rather than an audit. The five questions are the instrument; anyone can rerun them.

Three things fall out of that table. Three of the seven name a target and a fourth calls anything above 80 percent high, and only one links a source for it, LiveData, which points at an OR Manager article and carries Tyler in its reference list; none of the four derives its target from Tyler's conditions or mentions that the ceiling moves with case-duration variability. The arithmetic underneath is not the same page to page, so a figure like 75 to 85 percent is being applied to numerators that do not contain the same minutes. And of the two pages that publish a turnover-inclusive formula, neither says which way turnover pushes the result, which is the first thing a manager needs before acting on it.

This is the same pattern I keep finding in clinic benchmarks generally. A figure gets published once with its conditions attached, the conditions fall off in the retelling, and what survives is a round number with the authority of research and none of its qualifications.

What a minute of OR time actually costs

The other number attached to this topic is money, usually as 62 dollars a minute. It is worth tracing, because it is quoted as a cost and it is not one.

The trail runs through Alex Macario's 2010 editorial in the Journal of Clinical Anesthesia, titled with the question directly: what does one minute of operating room time cost. His answer opens "It depends," and he states plainly that "There are no published formal data on true OR costs." The 62 dollars appears there as a report that "A 2005 study of 100 U.S. hospitals found that OR charges averaged $62/min (range: $22 to $133/min)." Charges, not costs. His reference for it points to a PDF hosted on a surgical products vendor's own website, under the filename Time-saving.pdf.

The cost figure in the same editorial is not his own measurement either. He reports that, "Excluding physician costs, OR administrators may use a ballpark number such as $15 to $20 per OR minute for a basic surgical procedure, with at least half of that figure being fixed overhead costs", and footnotes it to Park and Dickerson, not to data of his own.

The best actual measurement I could find is Childers and Maggard-Gibbons in JAMA Surgery, who went at it through mandatory financial disclosures instead of surveys: every comparable California hospital, ten fiscal years, 302 of them reporting inpatient operating rooms in fiscal 2014. Their result:

One state, one fiscal year, anaesthesia counted as a separate revenue centre, and implants and other billable supplies excluded from the minute entirely.

Roughly two thirds of the direct cost is wages and benefits, and supplies account for under a tenth of the total. That is why saved minutes so rarely turn into saved money: the staff are there either way, and the overhead runs whether the room is occupied or empty. Macario puts the practical consequence plainly: "Superhuman effort, for example, to rush around on the day of surgery trying to reduce turnover times, may be dangerous, stressful, and have little financial justification."

A minute you take out of a turnover becomes money only if it changes something downstream: a staff member goes home earlier on a shift you pay by the hour, or the time freed adds a case that gets billed. Otherwise you have improved a percentage.

What to report beside the percentage

None of this is news inside a theatre suite that already reports turnover time, first-case on-time starts and day-of-surgery cancellations as their own line items, which is exactly why those line items exist: so the percentage cannot absorb them. The failure mode is the executive slide that carries the percentage alone. None of it makes utilization useless either. It makes it a number that cannot travel by itself, which is a different claim. Five things travel well next to it.

Utilization needs company on the dashboard

Utilization earns its place on a dashboard as long as it is not alone on it. Beside the case count, the minutes over and under, and the written down denominator, it tells you something. On its own, at the top of a slide, it is a number nobody in the room can check. The same discipline applies further down the building: provider and room utilization has the same denominator problem and no credible benchmark at all, patient wait times turn out to be three different metrics wearing one name, and both sit in the 12 KPIs worth tracking alongside the five numbers worth a weekly look.

Frequently asked questions

Written by

Olha · clinic data analyst

I build the reporting our managers open every morning at a multi-branch medical clinic, and package it so other practices don't have to start from scratch.

Published 1 August 2026 and corrected the same day after four independent checks, which found nine hard errors, all mine. The worst was self-inflicted: I wrote that the 99 percent day was the one you would wrongly celebrate, then two sections later recommended Tyler's own efficiency measure, which rates that day best of the three (146, 80 and 58 minutes of inefficiency). Both passages are rewritten, and the piece no longer offers a replacement number, because the table refutes every candidate it contains. I claimed none of the seven reviewed pages sourced its target; LiveData does, with an inline link and a reference list that includes Tyler, so the sentence was simply false. Umbrex calls above 80 percent high rather than setting a target, and Surgical Directions names 85 percent as well as 75, so "four name a target in the 75 to 85 range" was wrong twice over. The claim that none of the pages mentions turnover in the numerator contradicted my own table, where two of them publish exactly that. The Macario ballpark was quoted with a word dropped, presented as his own figure when he reports it from Park and Dickerson, and stripped of "excluding physician costs". Childers' ambulatory figure comes from 52 hospitals, not the 302 in the inpatient column; the supplies line is 2.55 dollars of a 37.37 dollar minute and the word is nonbillable. The Charlesworth and Pandit list of flawed metrics has five items, not four, and the one I omitted was the number of operations, which this article recommends: that tension is now stated instead of hidden. Two limits remain. Tyler's rows describe one modelled room, and I use them to show what the formula does, not to predict your theatre; I quote Charlesworth and Pandit from the abstract. The seven page review is my own reading on one day through a search API, the eighth page returned a 403, and those pages compete with this one in search. I have no commercial relationship with any of them. Lucid Vitals is not affiliated with Microsoft.