Beyond Bed Numbers: A Proposed Quality Improvement Framework for Reviewing ICU Acuity, Workload, and Medical Staffing

Background Intensive care unit (ICU) medical staffing is often planned using bed numbers or fixed ratios. These counts do not describe the work required when patients deteriorate, urgent tasks overlap, or clinicians have duties outside the unit. A structured review could help teams discuss whether available cover can meet these demands. Methods We undertook a targeted narrative synthesis of adult ICU staffing guidance, acuity and workload measures, physician workload studies, and capacity-strain research available through 8 September 2026. We examined study settings, findings, and limitations to inform the framework domains. The authors proposed the review prompts and response options. No intervention was implemented, and no new patient or staff data were collected. Results Illness severity scores and nursing workload tools do not measure the medical work required during a shift. The intensivist workload and ICU strain studies reviewed did not establish when additional bedside medical cover is needed. We propose the ICU Medical Staffing Review Framework, comprising 7 domains: patient allocation; organ support and dependency; instability and bedside demand; admissions, discharges and transfers; procedures and complex decisions; responsibilities outside ICU; and available capability and support. Consultant- or attending-led reviews at handover, or when demand or staffing changes, would identify work the team may be unable to cover, document an agreed response, and establish a time for reassessment. Conclusion The proposed framework offers a structure for discussing ICU medical staffing alongside professional standards and local escalation arrangements. Prospective evaluation should assess whether it is practical, produces consistent assessments, and helps identify unmet staffing needs. Its effects on staffing decisions, patient care, and documentation workload also require evaluation before wider use.