Development of Clinical Decision Support System for Severe Patients With Polytrauma

This study aimed to develop a Clinical Decision Support System for the management of severe polytrauma patients at the early inpatient care stage.

Scientific Hypothesis: It is assumed that the use of a clinical decision Support System at the stage of early inpatient care may reduce the mortality rate and adverse outcomes in severe patients with polytrauma.

Study setting The study is planned as a multicenter, non-interventional study, with participation from clinics providing inpatient trauma care for patients with polytrauma. The participating clinics will be selected based on their ability to provide access to a large and diverse population of patients with polytrauma. The research subgroup will be determined by clinics in compliance with the necessary ethical standards. The study will be conducted both retrospectively and prospectively, with data collection spanning a period of at least two years. Retrospective data will be collected from patients' medical records, including information on patient demographics, injury characteristics, diagnostic tests, and treatment modalities. Prospective data will be obtained through standardized patient interviews and clinical assessments, including patient-reported outcomes, clinical measures of physical and psychological functioning, and healthcare utilization. All participating clinics will be required to comply with the rules of medical ethics and deontology as well as local regulations governing access to personal data. In addition, researchers will be required to obtain informed consent from patients or their legal representatives before collecting any data for the study.

Obtained data:

Gender, Age, AIS code for every registered injury, NISS, Vitals on Admission (pulse rate, systolic and diastolic blood pressure, respiratory rate), Complete Blood Count on Admission (Red Blood Cells, Hematocrit, Hemoglobin, Platelets, White Blood Cells), biochemical analysis of blood on admission (total protein, total bilirubin, glucose, urea, creatinine), Blood Coagulation test (Activated Partial Thromboplastin Time, International Normalized Ratio, Fibrinogen), amount of total transfused blood and plasma packs, presence of complications (systemic inflammatory response syndrome, shock, severe hemorrhage [>25% of blood loss], acute traumatic coagulopathy, respiratory distress, pneumonia), length of stay in hospital, length of stay in the intensive care unit, and outcome at hospital discharge.

Additional Data Generation After the initial data were collected, additional fields will be created and calculated. These fields include the AIS field, which will be split into nine fields for each body region, with each field recording the square root of the sum of severity scores for that region; values for MCV, MCH, and MCHC, which will be calculated using the corresponding formulas; the "initial iron deficiency anemia" field, which will be calculated based on the initial MCV value; the "post-hemorrhagic anemia" field, which is calculated based on the initial hemoglobin level and relationship with initial iron deficiency anemia.

Informed Consent will be obtained by a research member who is an emergency/traumatology/ICU doctor, who was trained and explained to all the details of the current study. Consent will be obtained on patient admission to the hospital. Written Consent will be obtained from the patient or his/her legal representative if the patient is temporarily or completely disabled.

Recruitment will be started for every participating clinic separately during their recruitment periods, as defined by personal agreements. The main research clinic (Hospital of Emergency Aid, Semey, the Republic of Kazakhstan) has been participating in the recruitment process since December 1, 2020.

The reason for the recruitment end is the achievement of the minimum sample size by the research group or cancellation of participation in the trial.

The research group will collect a minimum sample size of data that will be used in the process of developing ML models. All the data were divided into training and test samples. Any record with less than 10% missing variables will have missing values imputed, whereas records with more than 10% missing variables will be excluded. Furthermore, additional data will be generated as outlined in the Eligibility Criteria section.

The selected features will be used to predict the following endpoints: the presence of complications (such as systemic inflammatory response syndrome, severe hemorrhage [>25% of blood loss], acute traumatic coagulopathy, respiratory distress, and pneumonia), and the outcome at hospital discharge.

All the developed models will be used for further CDSS development.

Confidentiality No identifiable patient information, such as names, IDs, phone numbers, emails, or addresses, will be collected. All information collected in accordance with the eligibility criteria will be transferred to trial participants via authorized access. After the study ends, any access not related to further work, such as statistical analysis or publications, will be blocked by the main researcher.