Clinical outcomes and opportunities for improvement across modified trauma quality improvement cohorts: a registry-based cohort study

Abstract

<title>Abstract</title> Background The American College of Surgeons Trauma Quality Improvement Program (TQIP) relies primarily on risk-adjusted mortality benchmarking, yet mortality alone may overlook non-fatal morbidity and care-process failures, particularly in cohorts where outcomes are driven by injury severity. This study aimed to analyse and compare 30-day mortality, cause of death, functional outcomes as measured by the Glasgow Outcome Scale (GOS), and opportunities for improvement (OFIs) identified through peer review across modified TQIP cohorts. Methods Registry-based cohort study of 8,298 trauma patients at Karolinska University Hospital (2013–2023), classified into four modified TQIP cohorts: isolated severe TBI, blunt multisystem with TBI, blunt multisystem without TBI, and penetrating truncal injury. Remaining patients, predominantly lower-severity, formed a non-TQIP reference cohort. Cohort associations with mortality, unfavourable GOS, and OFI were assessed using cumulative sequential logistic regression (reference: non-TQIP cohort), progressively adjusting for patient characteristics, physiological status, and care processes, pooled across 20 imputed datasets. Cause of death, GOS levels, and OFI categories were compared using multinomial logistic regression. Results All TQIP cohorts had elevated odds of mortality and unfavourable GOS. The blunt multisystem cohorts had the highest OFI rates (10.6% with TBI, 16.9% without) and were the only cohorts with elevations across all OFI categories, including potentially preventable deaths and clinical judgement errors. Care-process adjustment (e.g. emergency interventions, time to radiology, care level) reduced their mortality ORs, suggesting potentially modifiable processes. Isolated severe TBI had the highest mortality (52.8%) but no OFI excess, suggesting outcomes driven by primary injury rather than care-process failures. Penetrating truncal injury showed high, potentially preventable, haemorrhage-related mortality but favourable functional recovery in survivors after care-process adjustment. Conclusions Outcome profiles differed across TQIP cohorts and were not captured by mortality alone: isolated severe TBI had the highest mortality but no OFI excess, while the blunt multisystem cohorts had the greatest OFI burden and potentially preventable deaths despite lower mortality. Penetrating truncal injury showed high haemorrhage-related mortality but excellent functional recovery in survivors. Integrating functional outcomes and structured peer review alongside mortality identifies cohorts amenable to quality improvement and reveals cohort-specific processes to target.