.png){kind=icon}
Return
A data-driven approach for surgical case assignment problem
.png){kind=icon}
PRE
AI
DOI：10.1016/j.cor.2026.107675.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Inaccurate estimates of surgery duration reduce operating room efficiency, leading to under- or overutilized time blocks, costly overtime, and prolonged patient waiting times. Current scheduling practices in the National Health Service (NHS) in England rely on fixed duration benchmarks or clinician judgment, which fail to capture patient- and procedure-specific variability. We develop a data-driven framework that integrates machine learning predictions with deterministic and robust optimization models for surgical case assignment. Using three years of data from an NHS Foundation Trust, we compare alternative predictive models and show that our approach reduces the average error of duration estimates from 38 min to 11 min relative to NHS benchmarks. These predictions are embedded within a robust optimization formulation solved via a column-and-constraint generation algorithm to construct schedules that maximize utilization while hedging against uncertainty. Incorporating predictive inputs narrows the gap between expected and realized utilization from 25.8% to 1.5%, increases patient throughput by up to 45%, and reduces waiting list length by more than 70%. The proposed framework demonstrates that combining predictive analytics with robust optimization yields significant improvements in efficiency and service delivery. The approach is directly applicable to elective surgery scheduling and generalizable to other healthcare systems seeking data-driven, uncertainty-aware decision tools.
Journal
C
IF:
4.3
Papers:
254
Citations:
0
Organization
Cited Papers
No cited papers available