Case duration accuracy is essential for running an efficient operating room. It’s the building block of your plan. Everything else you do is built upon this simple attribute. You use the case duration for determining how many cases you can do in each operating room and in what order you will do them. You also rely on its accuracy to decide the number of staff you will need at different hours. It impacts your planning for your pre-op area, your PACU bays, and even has ramifications for when beds need to be available for the patient to go back to their ward.
What is surgical case duration accuracy?
Surgical case duration accuracy is how closely the time you schedule for a case matches the time the case actually takes. A case booked for 120 minutes that runs 125 is accurate. One that runs 170 throws off every case scheduled after it in that room, along with the staff, the PACU bays, and the beds planned around them.
Why are most case duration estimates inaccurate?
You would think an attribute with such profound effect on the operations of not only the operating room but also the hospital at large would be scrutinized very closely. Yet, it seems most organizations have bought into the idea that the best they can do is rely on simple historical averages and surgical team estimates. It has been shown over and over in academic publications that such simplistic approaches are inaccurate and do have costly consequences on operating suite utilization and patient, physician, and staff satisfaction.
If your hospital is using an EMR such as Epic or Cerner, you probably feel pretty good when their scheduling module gives you the option to use their “prediction” for case duration. If you dig deeper, this prediction is some variation of taking the last say 10 procedures of the given type, maybe throwing away the highest and lowest of the bunch and doing an average! Basically using 5th grade math on sample size of 10 to decide the single most important attribute of your planning for a multimillion dollar resource!
What factors affect surgical case duration?
Think about all the possible factors that impact a procedure:
Patient
- Age and sex
- Height and weight
- Allergies
- History and physical examination (H&P)
- ASA status (American Society of Anesthesiologists physical status)
Providers
- Surgeon(s)
- Anesthesiologist(s)
- Scrub nurse(s) and tech(s)
- Circulating nurse(s) and tech(s)
Facility and room
- Bed census
- Equipment
- Staff
- Day of week
- Time of day
Procedure
- Procedure type
- Surgeon comments
- Procedure modifiers
- Anesthesia type
- Implants and tissues
Prior events
- Timing of prior perioperative milestones
- Case delays
- Cancellations
- Turnover
Any single of one of these factors can have drastic impact on how the case will progress and the duration of the surgery. Now let’s think back to what our scheduling system used! We throw all of this info (which we incidentally capture & document anyways) away and just picked the procedure type. Is it any wonder why such predictions can be grossly inaccurate?
Case duration explained
Let’s think of an analogy! Imagine you wanted to drive from point A to point B! You rely on an application such as Google Maps or Waze to tell you how long it will take you to make the trip. The app can take a number of factors into consideration: distance, possible routes, traffic conditions, current construction projects, weather conditions, …. Very similar to our surgical case, any number of these factors can impact the duration of the trip. Would you rely on an app that throw all this info out except for the distance?
How do you measure case duration accuracy?
To measure case duration accuracy, compare each case's scheduled duration with its actual duration, using the same start and end milestones your schedule is built on (for example, patient in room to patient out of room). Then look at it from a few angles:
- Average error in minutes. How far off is the typical case, in either direction? This is the simplest number to track over time.
- Share of cases within tolerance. What percentage of cases finish within a set window of the scheduled time, such as ±15 minutes or ±10%? This shows how often the schedule can actually be trusted.
- Over vs. under. Are you consistently running long or short? Cases that run long push back every case after them and drive overtime. Cases that finish early leave rooms idle and blocks underused.
- Where the misses cluster. Break the error down by surgeon, procedure, service line, and time of day. Most organizations find the inaccuracy isn't spread evenly; it concentrates in a handful of places.
If you've never run these numbers, start there. You'll quickly see how much your current estimates are costing you. Getting the measurement right also depends on clean milestone data; see cleaning up your OR data.
How machine learning improves case duration prediction
The good news is that we don't have to settle for just one dimension (procedure type), a handful of data points, and simple math! With today's technology, we can take all the factors listed above into consideration. We can run multiple sophisticated algorithms over all of our data and pick the ones with the best predictive ability! We can continuously train these algorithms as new information is captured. Every day that your team performs more cases, it's creating more data to build smarter predictions from. It's a shame that for many, this trove of insight is sitting unused in an EMR database instead of being leveraged to make your team more successful.
At Leap Rail, we put over 1,500 distinct dimensions from your existing data in your EMR to use for predicting surgical case duration. A study by Harvard Medical School, published in the Journal of Medical Systems, shows the Leap Rail method resulted in a 70% reduction in overall scheduling inaccuracy. This improved level of accuracy has given our customers much better visibility into their operations. They can plan for rooms, resources, equipment, and staff much more confidently. They can give physicians reliable start times, or tell them when an add-on case might be accommodated, with higher confidence.
Better case duration predictions feed directly into surgical scheduling and block utilization. When the durations are right, the schedule built on them holds.
By using Leap Rail's state-of-the-art technology and benefiting from their existing data through artificial intelligence and machine learning, our customers are not only delivering a higher quality of care, they are a lot more efficient with their resources, and their physicians and surgical teams are a lot more satisfied.
Is your organization ready?
Are you curious if your organization is ready to take advantage of artificial intelligence and machine learning for your perioperative environment? Let the Leap Rail team do an assessment of your organization to find out. Our team will tell you how big of an opportunity your organization is sitting on without realizing it! What are you waiting for?