Arizona Trauma System: Past, Present, and Future

Abstract

Background: In Arizona, the number of adult level 1 trauma centers (L1TCs) increased from seven in 2012 to thirteen in 2020, almost doubling in number. Similarly, from 2008 to 2012, 21 adult lower level trauma centers (LLTCs), consisiting of level 3 and level 4 trauma centers, were added to the Arizona trauma system. This rapid proliferation of trauma centers (TCs) in Arizona is an extreme example of the national TC proliferation trend. This thesis outlines the major findings of our studies that evaluated the impact of L1TC and LLTC proliferation in Arizona on clinical outcomes. We hypothesized that L1TC and LLTC proliferation may worsen patient outcomes, but may increase trauma system access. Finally, we propose a pragmatic method of trauma system design using geospatial analysis. Methods: L1TC and LLTC AnalysesAdult patients age 15 or greater in the Arizona state trauma registry from 2007 to 2020 were queried for demographic, injury, and outcome variables. These variables were compared across two time periods. In the L1TC analysis, 2007-2012 was used as pre-proliferation (PRE) and 2013-2020 as post-proliferation (POST). In the LLTC analysis, 2007 was used as PRE, and 2008–2012 as POST cohorts. Multivariate logistic regression was performed to assess independent predictors of mortality. Subgroup analyses were done for Injury Severity Score (ISS) ≥15, age ≥65, and trauma mechanisms. Geospatial Analysis: Publicly available data from the Arizona Department of Health Services and the Arizona Open GIS database were used to identify trauma center and hospital locations in Arizona. Then performed location suitability analysis via ESRI ArcGIS Pro 3.1.2 (Redlands, CA) to find possible future TC locations using a set of criteria. Finally, the network analysis tool from ESRI was used to create service areas where a trauma patient can be transported to a TC within 60 minutes by ground ambulance. These service areas were built for the following scenarios: 1) service area currently covered by existing TCs. 2) service area covered by adding two of the four most suitable TCs that are farthest away from other TCs. 3) service area covered by adding all four of the most suitable locations. Results: L1TC analysis: A total of 482,896 trauma patients were included in the L1TC analysis. 40% were female, 29% were geriatric patients, and 8.6% sustained penetrating trauma. The median injury severity score (ISS) was 4. Inpatient mortality overall was 2.7%. POST consisted of more female, geriatric, and blunt trauma patients (p < 0.001). Both periods had similar median ISS. POST had more interfacility transfers (14.5% vs 10.3%, p < 0.001). Inpatient, unadjusted mortality decreased by 0.5% in POST (p < 0.001). After adjusting for age, gender, ISS, and trauma mechanism, being in POST was predictive of death (OR: 1.4, CI:1.3-1.5, p < 0.001). This was consistent across all subgroups except for the geriatric subgroup. LLTC analysis: A total of 143,919 adults were included in this analysis. POST consisted of significantly more female, geriatric, and blunt traumas (all p <0.001). ISS was similar across the two cohorts. Interfacility transfers increased by 10.2 %. Overall mortality decreased by 0.6 % (p <0.001). After adjusting for differences in demographic and injury variables, multivariate logistic regression analysis showed that being in POST was not associated with survival (OR: 1.07, CI: 0.96–1.18, p = 0.227). Subgroup analyses showed small reductions in mortality, except for geriatric patients. After adjusting for covariates, POST was not associated with survival in any subgroup and trended toward being a predictor for death in penetrating traumas (OR: 1.23; CI: 1.00–1.53, p = 0.059). Geospatial-Guided Expansion of Trauma SystemFuture suitable TC locations were determined using the following criteria: 1. Located within 1 mile from highway ramps 2. Located at least 5 miles or greater from existing TCs 3. Located at preexisting hospitals 4. Located in urban areas as defined by US Census Bureau A total of 14 locations satisfied the criteria above. After weighted, ranked analysis giving preference to shorter distances to highway ramps and longer distances to existing TCs, four TCs had the highest suitability score. The service area analysis showed that approximately 28% of the state of Arizona is covered by existing TCs. This coverage area increased by 2688 mi2 or 30.6% if two of the four most suitable TCs that are farthest away from other TCs were added to the state trauma system. If all four of the most suitable locations became TCs, the coverage would increase by 0.2% to 30.8%. Discussion: Our studies showed that the unplanned proliferation of TCs in Arizona was associated with increased interfacility transfers without a significant increase in the number of trauma patients treated. Furthermore, it may worsen regional patient mortality despite significant resource use and financial cost. We have used a set of logical criteria and showed a pragmatic method of trauma system design using geospatial analysis that would increase trauma coverage in Arizona. We urge federal or state lead agencies to develop a set of evidence-based criteria to guide further future trauma system development.