In medical research, missing data is common. In acute diseases, such as traumatic brain injury (TBI), even well-conducted prospective studies may suffer from missing data in baseline... Show moreIn medical research, missing data is common. In acute diseases, such as traumatic brain injury (TBI), even well-conducted prospective studies may suffer from missing data in baseline characteristics and outcomes. Statistical models may simply drop patients with any missing values, potentially leaving a selected subset of the original cohort. Imputation is widely accepted by methodologists as an appropriate way to deal with missing data. We aim to provide practical guidance on handling missing data for prediction modeling. We hereto propose a five-step approach, centered around single and multiple imputation: 1) explore the missing data patterns; 2) choose a method of imputation; 3) perform imputation; 4) assess diagnostics of the imputation; and 5) analyze the imputed data sets. We illustrate these five steps with the estimation and validation of the IMPACT (International Mission on Prognosis and Analysis of Clinical Trials in Traumatic Brain Injury) prognostic model in 1375 patients from the CENTER-TBI database, included in 53 centers across 17 countries, with moderate or severe TBI in the prospective European CENTER-TBI study. Future prediction modeling studies in acute diseases may benefit from following the suggested five steps for optimal statistical analysis and interpretation, after maximal effort has been made to minimize missing data. Show less
Background Prehospital care for traumatic brain injury (TBI) is important to prevent secondary brain injury. We aim to compare prehospital care systems within Europe and investigate the association... Show moreBackground Prehospital care for traumatic brain injury (TBI) is important to prevent secondary brain injury. We aim to compare prehospital care systems within Europe and investigate the association of system characteristics with the stability of patients at hospital arrival. Methods We studied TBI patients who were transported to CENTER-TBI centers, a pan-European, prospective TBI cohort study, by emergency medical services between 2014 and 2017. The association of demographic factors, injury severity, situational factors, and interventions associated with on-scene time was assessed using linear regression. We used mixed effects models to investigate the case mix adjusted variation between countries in prehospital times and interventions. The case mix adjusted impact of on-scene time and interventions on hypoxia (oxygen saturation <90%) and hypotension (systolic blood pressure <100mmHg) at hospital arrival was analyzed with logistic regression. Results Among 3878 patients, the greatest driver of longer on-scene time was intubation (+8.3 min, 95% CI: 5.6-11.1). Secondary referral was associated with shorter on-scene time (-5.0 min 95% CI: -6.2- -3.8). Between countries, there was a large variation in response (range: 12-25 min), on-scene (range: 16-36 min) and travel time (range: 15-32 min) and in prehospital interventions. These variations were not explained by patient factors such as conscious level or severity of injury (expected OR between countries: 1.8 for intubation, 1.8 for IV fluids, 2.0 for helicopter). On-scene time was not associated with the regional EMS policy (p= 0.58). Hypotension and/or hypoxia were seen in 180 (6%) and 97 (3%) patients in the overall cohort and in 13% and 7% of patients with severe TBI (GCS <8). The largest association with secondary insults at hospital arrival was with major extracranial injury: the OR was 3.6 (95% CI: 2.6-5.0) for hypotension and 4.4 (95% CI: 2.9-6.7) for hypoxia. Discussion Hypoxia and hypotension continue to occur in patients who suffer a TBI, and remain relatively common in severe TBI. Substantial variation in prehospital care exists for patients after TBI in Europe, which is only partially explained by patient factors. Show less
Objective: We aimed to explore the added value of common machine learning (ML) algorithms for prediction of outcome for moderate and severe traumatic brain injury.Study Design and Setting: We... Show moreObjective: We aimed to explore the added value of common machine learning (ML) algorithms for prediction of outcome for moderate and severe traumatic brain injury.Study Design and Setting: We performed logistic regression (LR), lasso regression, and ridge regression with key baseline predictors in the IMPACT-II database (15 studies, n = 11,022). ML algorithms included support vector machines, random forests, gradient boosting machines, and artificial neural networks and were trained using the same predictors. To assess generalizability of predictions, we performed internal, internal-external, and external validation on the recent CENTER-TBI study (patients with Glasgow Coma Scale <13, n = 1,554). Both calibration (calibration slope/intercept) and discrimination (area under the curve) was quantified.Results: In the IMPACT-II database, 3,332/11,022 (30%) died and 5,233(48%) had unfavorable outcome (Glasgow Outcome Scale less than 4). In the CENTER-TBI study, 348/1,554(29%) died and 651(54%) had unfavorable outcome. Discrimination and calibration varied widely between the studies and less so between the studied algorithms. The mean area under the curve was 0.82 for mortality and 0.77 for unfavorable outcomes in the CENTER-TBI study.Conclusion: ML algorithms may not outperform traditional regression approaches in a low-dimensional setting for outcome prediction after moderate or severe traumatic brain injury. Similar to regression-based prediction models, ML algorithms should be rigorously validated to ensure applicability to new populations. (C) 2020 The Authors. Published by Elsevier Inc. Show less