Registry of Patients With Severe Trauma Admitted to the ICU in Tarragona

• Study design: observational, retrospective, one centre (Hospital Universitari Joan XXIII) in Tarragona, Catalonia, Spain.
• Study period: Seven-years period: Since 01 January 2015 - 31 December 2022 (The registry will be open and active recruiting after this time point).
• Sample size calculation: Not applicable. The sample size will be equal to the total number of patients consecutively admitted to the Intensive Care Unit (ICU) due to severe trauma during the study period.
• Data extraction:

Data will be extracted by twofold:

1. Hand-driven: extracted data from the medical reports by reviewing the clinical history system performed by the study investigators, who are part of the staff of the ICU involved in the care of trauma patients. This information was collected into a large-scale database and, once introduced, the data was checked and verified by two investigators (and principal investigator).

These manually extracted data included pre-hospital variables (such as the time of the injury, the mechanism of injury, first vital signs, pre-hospital procedures), in-hospital information (demographic's, primary survey of trauma in emergency department-ED, use of Extended Focused Assesment with Sonography in Trauma, complete laboratory results in ED), description of detailed anatomic injuries (Traumatic Brain Injury-TBI, thoracic trauma, abdominal trauma, Pelvic trauma, orthopedic injuries), severity scores (Acute Physiology And Chronic Health Evaluation-APACHE II, Sequential Organ Failure Assessment-SOFA score, Injury Severity Score-ISS, Abbreviated Injury Score-AIS, etc), resources during ICU admission (surgeries, transfusions, organ support, etc), complications (organ failures and nosocomial infections such as ventilator-associated pneumonia, catheter-related infection, etc), and clinical outcomes (ICU mortality, cause of death, in-hospital mortality, one-year survival, mechanical ventilation days, ICU and hospital length of stay, disability at ICU discharge and destination at discharge. 2. Automatic data: automated data collected with the Electronic Health System (Centricity Critical Care) provide solid and valid extraction of information. Automatic variables were checked and validated by the expert and specialized members of the ICU in data quality assessment and the management of data to ensure that recorded information was reliable. The accuracy, completeness and representativeness of automatic data was double checked when entered into the large-scale database for data quality by the principal investigator and a biotechnologist, who were responsible for the quality assurance plan that addressed data validation, monitoring and auditing (identifying inconsistencies, outliers or incorrect ranges for variables entered into the registry). The automated variables collected (not recorded manually) included demographic data, comorbidities and previous risk factors, dates (hospital and ICU admission, hospital and ICU discharge), laboratory tests during the first week of ICU admission, resources consumption during ICU stay, complications during ICU admission and outcomes.

• Statistical plan: descriptive and analytical statistics using binary logistic regression (for association), survival analysis with Kaplan Meier analysis and Cox regression (for clinical outcomes), nonlinear association test such as CHAID decision analysis and restricted cubic splines, and matching learning analysis (for cluster or subgroups identification). These analyzes will be carried out using SPSS v24 and R software. Missing data values of automated variables will be handled using multiple imputation. Hand-driven data had negligible missing values.
• Variable definitions (all definitions met the ATLS-Advanced Trauma Life Support criteria):