A new peer-reviewed study published on Cureus finds that data-intensive machine learning is suitable for predicting the case durations of common surgical procedures at ASCs.
The researchers – Franklin Dexter, MD, PhD, professor of anesthesiology at the University of Iowa; Brenda G. Fahy, MD, an anesthesiologist at the University of Florida; and Richard H. Epstein, MD, an anesthesiologist at the University of Miami Miller School of Medicine – posit that Bayesian methods such as applying the surgeon’s or scheduler’s time estimates often perform well for uncommon procedures. But they state that uncommon procedures dramatically affect case duration predictions that are necessary for scheduling cases days to weeks in advance of surgeries.
“The current epidemiology of uncommon procedures is based on datasets from 25 years ago,” they write. “We calculated contemporaneous incidence proportions for rare procedure combinations, those performed at facilities only once or twice per quarter.”
The retrospective cohort study used de-identified, publicly available data from Florida ambulatory surgery databases from 2010 to 2024, comprising distinct combinations of 20,014,189 cases distributed across 5,106,524 combinations of quarter, facility, and CPT codes. It also collected 11,643,813 cases in 2009 to 2024 inpatient databases across 4,772,566 combinations of quarter, facility, and distinct combinations of International Classification of Diseases (ICD) procedure codes.
The researchers found that incidence proportions of procedures performed once or twice at each facility during the quarter performed, referred to as “doubletons,” became progressively less common. “The change from ICD-9 to the more granular ICD-10-PCS in 2015 made singletons and doubletons more common for inpatient surgery. They write, “In 2024, approximately 66%, 78%, and 87% of procedures were observed just once or twice each quarter at the ASC, HOPD and inpatient surgical suite where observed. These doubleton procedures accounted for approximately 18% of cases at ASCs, 36% at HOPDs, and 55% at inpatient surgical suites. Pooling hospital estimates, approximately 84% of procedures and approximately 44% of cases were among procedures performed just once or twice during the quarter.
The team concluded, “Although surgical procedure(s) are the most important predictors of operating room time, many procedures are performed rarely, resulting in little historical case duration data for case duration estimation. Freestanding ASCs have, for more than 10 years, remained different from hospitals in performing more common procedures. Hospitals should plan, when scheduling cases, that approximately 84% of distinct combinations of procedures and 44% of cases will have little to no procedure-specific historical data, and even less so by the combination of procedure and surgeon. Machine learning models for predicting case durations should therefore account for these uncommon procedures, for which Bayesian methods are well suited.”
