Predicting the Individual Treatment Effect of Neurosurgery for Patients with Traumatic Brain Injury in the Low-Resource Setting: A Machine Learning Approach in Uganda.

Traumatic brain injury (TBI) disproportionately affects low- and middle-income countries (LMICs). In these low-resource settings, effective triage of patients with TBI-including the decision of whether or not to perform neurosurgery-is critical in optimizing patient outcomes and healthcare resource utilization. Machine learning may allow for effective predictions of patient outcomes both with and without surgery. Data from patients with TBI was collected prospectively at Mulago National Referral Hospital in Kampala, Uganda, from 2016 to 2019. One linear and six non-linear machine learning models were designed to predict good versus poor outcome near hospital discharge and internally validated using nested five-fold cross-validation. The 13 predictors included clinical variables easily acquired on admission and whether or not the patient received surgery. Using an elastic-net regularized logistic regression model (GLMnet), with predictions calibrated using Platt scaling, the probability of poor outcome was calculated for each patient both with and without surgery (with the difference quantifying the "individual treatment effect," ITE). Relative ITE represents the percent reduction in chance of poor outcome, equaling this ITE divided by the probability of poor outcome with no surgery. Ultimately, 1766 patients were included. Areas under the receiver operating characteristic curve (AUROCs) ranged from 83.1% (single C5.0 ruleset) to 88.5% (random forest), with the GLMnet at 87.5%. The two variables promoting good outcomes in the GLMnet model were high Glasgow Coma Scale score and receiving surgery. For the subgroup not receiving surgery, the median relative ITE was 42.9% (interquartile range [IQR], 32.7% to 53.5%); similarly, in those receiving surgery, it was 43.2% (IQR, 32.9% to 54.3%). We provide the first machine learning-based model to predict TBI outcomes with and without surgery in LMICs, thus enabling more effective surgical decision making in the resource-limited setting. Predicted ITE similarity between surgical and non-surgical groups suggests that, currently, patients are not being chosen optimally for neurosurgical intervention. Our clinical decision aid has the potential to improve outcomes.

Head Computed Tomography Scan in Isolated Traumatic Brain Injury in a Low-Income Country.

BACKGROUND: Head computed tomography scan (HCTS) is the cornerstone of the management of traumatic brain injury (TBI). The impact of performing a HCTS in TBI has been scarcely investigated in low-income countries (LICs). Furthermore, the cost of a HCTS is a burden for family finances. METHODS: A prospective observational study was conducted in Burkina Faso. All consecutive patients with isolated TBI needing a HCTS were included. Result and impact of HCTS were evaluated. RESULTS: There were 183 patients prescribed a HCTS for an isolated TBI. Mild, moderate, and severe TBIs represented 55%, 31%, and 14% of the cases, respectively. In 72 patients, HCTS was not performed because of economic barrier. Among the 110 HCTSs performed, there were intracranial lesions in 81 (74%) patients. Among the 110 performed HCTS, 34 (31% [22.3%-39.5%]) HCTSs altered the management of TBI, with 16 (15%) cases of surgical indications, and 20 (18%) cases of modification of the medical treatment. In patients without neurologic signs, the rate of alteration of management was 28%. The realization of the HCTSs was associated with the presence of neurologic signs and income level. In-hospital mortality was 11% (n = 21). Among the 162 patients discharged alive from the hospital, 27 (20%) were discharged with a severe disability state (Glasgow Outcome Scale score ≤3). The rate of return to work was 77%. CONCLUSIONS: No modification of guidelines can be advocated from this study. However, given the financial burden on family of performing HCTS, research may identify criteria allowing for avoiding HCTS. Guidelines specific to LICs are needed to get closer to the best interest of patients.

Influencia de la recuperación de la marcha sobre el tiempo de ingreso hospitalario en el niño y adolescente con traumatismo craneoencefálico / Influence of the recovery of walking on the hospital impatient time in children and adolescents with traumatica brain injury

Introducción. El traumatismo craneoencefálico (TCE) es la primera causa de discapacidad adquirida entrela población infantil y juvenil, y la recuperación de la marchauno de los principales objetivos de los programas de rehabilitación. La finalidad de este trabajo es estudiar la influencia de la capacidad de la marcha en el momento del alta hospitalaria,y su repercusión sobre el tiempo de estancia hospitalaria en un grupo de niños y adolescentes con TCE ingresadosen el Servicio de Rehabilitación de un hospital infantil. Pacientes y métodos. Evaluamos 40 niños con TCE y conuna edad media de 13,5 años ingresados en la planta de Rehabilitaciónde un hospital infantil desde el año 2001 hasta el 2006. En el momento del alta hospitalaria la marcha se clasificó de forma cuantitativa por su presencia o ausencia, y de forma cualitativa por ser autónoma o asistida. Las variablesanalizadas fueron: tipo de lesión según la neuroimagen, lesiones asociadas en extremidades inferiores y/o pelvis, días de estancia en la Unidad de Cuidados Intensivos (UCI), en la planta de Rehabilitación y días de estancia total, presencia del síndrome de disfunción autonómica (SDA) y pronóstico funcionalal alta según la escala de resultados de Glasgow (GOS). Resultados. En el momento del alta hospitalaria 35 niños(87,5 %) realizaban marcha; 29 de ellos (72,5 %) lo hacían de forma autónoma y 6 (15,0 %) de forma asistida. Los 5 restantes (12,5 %) no deambulaban. Tres pacientes presentaron el síndrome de disfunción autonómica, ninguno de los cualescaminaba. La estancia media hospitalaria fue de 32,15 días enlos ambulantes y de 46,35 en los no ambulantes. Se encontró correlación entre la capacidad de marcha, el tiempo medio de coma y los días de estancia en la UCI. Otras variables correlacionadas,aunque estadísticamente menos significativas, fueron los días de estancia en la planta de Rehabilitación y el tiempo total de ingreso (AU)

Introduction. Traumatic brain injury (TBI) is the most important cause of acquired disability in children and young people. Restoring walking ability is one of the primary purposes of our rehabilitation program. This work has aimed to study the incidence of the walking ability on hospital discharge time and its effects on the average hospital stay in children and young people with TBI admitted to a Children¿s Hospital Rehabilitation Service. Patients and methods. Forty children with TBI and an average age of 13.5 were admitted to the rehabilitation service between 2001 and 2006. Walking was evaluated at the time of discharge quantitatively (presence or absence of walking ability) and qualitatively (independent or device assisted walking). The variables analyzed were: lesion type through neuroimaging, associated lower limb injuries, average stay in ICU (Intensive care Unit) and Rehabilitation Service, presence of Autonomic Dysfunction Syndrome (SDA) and functional outcome on discharge by Glasgow Outcome Scale (GOS). Results. Thirty-five (87.5 %) of the 40 children and adolescents were walkers on discharge, 29(72.5 %) as independent walkers, 6(15.0 %) walked with device assistance and 5 (12.5 %) were non-walkers. Average Hospital stay was 32.15 days for walking children and 46.35 for non-walking ones. A correlation was found between walking ability, length of coma and length of the impatient stay in the ICU. Other correlated variables but with less statistical significance were average Rehabilitation service stay and total length of impatient stay. We did not find any correlation with initial TBI severity or with lower limb or pelvis injuries. Conclusions. Most of the children were walkers at discharge. The average Rehabilitation impatient stay and total average inpatient stay were lower in non-walking patients. We conclude that walking ability has an influence over total average Hospital inpatient stay (AU)