Trauma Care Capacity and Outcomes in Texas: Regional Comparison by Use of Multilevel Model (MLM)

Objective: We will review one year of Texas hospital discharge data and trauma registry data for the 22 trauma services regions around Texas to determine what regional variations exist in the capacity, process and clinical outcomes for trauma patients and whether there is an association between capacity, process, and outcomes. Methods: ICD9-CM codes (800-959) and ICD/ISS Injury Severity Score(ICD/ISS>15) will be used to select all injured patients and major trauma via ICDMAP-90 software. The indicators for capacity are: staffed beds per capita of the designated trauma centers including Level I to IV. The indicators of process are: ambulance transport hours; over/undertriage; and interhospital transfer hours. The indicators of outcomes are: trauma mortality; hospital/ICU LOS; and hospital costs. Analysis: Data analyses are described by descriptive statistics to examine regional variations in capacity, process, and outcomes. Hierarchical structure are composed of the 3 different levels: Level 1(trauma patients); Level 2 (trauma centers); and Level 3(regions). One-way ANOVA random effects models will be used to test the regional variability at the 5% significance level. Multivariate regression analyses will be used to test the relationships between capacity, process, and outcomes measures. A likelihood ratio test will be used to determine the significance of variances which represent the regional variations for continuous response variables (hospital/ICU LOS and hospital care cost). For discrete response variable (trauma mortality), a Wald test will be used. The p-value will be used to examine whether the null hypothesis of each model will be rejected at the 5% level. Other covariates such as age, gender, ISS and insurance status will be considered in the models to adjust for the confounding effects. Results: We expect that there will be statistically significant differences between the regions in terms of capacity, process, and outcomes. The regions with the highest number of ratios or averages in the indicators such as staffed beds per capita, transport/transfer hours, over/undertriage, trauma mortality, hospital/ICU LOS and hospital costs will be in the metropolitan areas. We expect that the regions with more staffed beds per capita will have lower trauma mortality, shorter hospital/ICU LOS, and lower hospital costs. Also, the regions with less transport hours, transfer hours, and lower under/over-triage will have lower trauma mortality, shorter hospital/ ICU LOS, and lower hospital cost. Conclusions: The regional variations in capacity measures (staffed beds per capita) will be a critical contributor to make regional differences in process and outcomes for trauma patients.