The authors developed SQDSS, combining two-stage stochastic programming with multi-class priority queueing to triage surgical critical-care patients under uncertain arrivals, changing acuity, and deterioration while waiting. Across 48 synthetic datasets, comparisons with first-come-first-served and a fixed-severity rule showed about 22.8% shorter high-acuity waits, 16% greater bed utilization, roughly 28% less queue overflow, and roughly 18% fewer simulated complications. Benefits were greatest under surge and trauma-dominant conditions. No real-patient data or implementation trial were used, so these results demonstrate modeled feasibility rather than clinical effectiveness.
Key findings
- In simulation, SQDSS reduced mean high-acuity waiting time by about 22.8% relative to benchmark policies; no real-patient queue was evaluated.
- Bed utilization increased by about 16% and queue overflow fell by about 28%, suggesting modeled capacity to coordinate priority and constrained resources under the stated assumptions.
- The simulated postoperative-complication rate fell by roughly 18%, with the largest gains under surge and trauma-dominant conditions. This is not an observed clinical rate.
- Comparisons with FCFS and a fixed-severity threshold demonstrate the value of dynamic prioritization within the model, but not superiority over experienced clinical teams or advanced ICU systems in practice.
Why this matters globally
The model offers a useful framework for hospitals allocating critical-care beds under volatile demand and makes trade-offs among urgency, efficiency, and deterioration risk explicit. If prospectively validated, it could support recommendations or policy testing. It should not yet automate admission decisions, because misspecified assumptions could directly affect safety and equity.
Thai researcher contribution
The entire team is from Khon Kaen University: Jarun Bootdachi from the Faculty of Interdisciplinary Studies and Parichat Wonggom and Wiphawadee Potisopha from the Faculty of Nursing. The work integrates business administration and adult nursing and was supported through Khon Kaen University's Fundamental Fund. This cross-disciplinary framing joins operational and clinical risk, but the article does not report deployment in a Thai hospital.
Limitations to consider
All outcomes, including complications, are simulation-generated and sensitive to assumptions about arrivals, service times, acuity transitions, deterioration, and objective weights. Only FCFS and a fixed rule were comparators, potentially simpler than real practice. There is no external validation with records, prospective testing, clinician-decision assessment, usability work, IT integration, override analysis, fairness audit, or adverse-event evaluation. The model therefore does not prove prevention of real complications and cannot replace clinical judgment.
Verify the original sources
Applied System InnovationApplied System Innovation↗DOI: 10.3390/asi9100210