Journal of critical care · 2018
Healthcare-associated ventriculitis and meningitis in a neuro-ICU: Incidence and risk factors selected by machine learning approach.
СтатьяНаблюдательное исследованиеНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Savin Ivan, Ershova Ksenia, Kurdyumova Nataliya, Ershova Olga, Khomenko Oleg, Danilov Gleb, Shifrin Michael, Zelman Vladimir
Аннотация
PURPOSE: To define the incidence of healthcare-associated ventriculitis and meningitis (HAVM) in the neuro-ICU and to identify HAVM risk factors using tree-based machine learning (ML) algorithms.
METHODS: An observational cohort study was conducted in Russia from 2010 to 2017, and included high-risk neuro-ICU patients. We utilized relative risk analysis, regressions, and ML to identify factors associated with HAVM development.
RESULTS: 2286 patients of all ages were included, 216 of them had HAVM. The cumulative incidence of HAVM was 9.45% [95% CI 8.25-10.65]. The incidence of EVD-associated HAVM was 17.2 per 1000 EVD-days or 4.3% [95% CI 3.47-5.13] per 100 patients. Combining all three methods, we selected four important factors contributing to HAVM development: EVD, craniotomy, superficial surgical site infections after neurosurgery, and CSF leakage. The ML models performed better than regressions.
CONCLUSION: We first reported HAVM incidence in a neuro-ICU in Russia. We showed that tree-based ML is an effective approach to study risk factors because it enables the identification of nonlinear interaction across factors. We suggest that the number of found risk factors and the duration of their presence in patients should be reduced to prevent HAVM.
METHODS: An observational cohort study was conducted in Russia from 2010 to 2017, and included high-risk neuro-ICU patients. We utilized relative risk analysis, regressions, and ML to identify factors associated with HAVM development.
RESULTS: 2286 patients of all ages were included, 216 of them had HAVM. The cumulative incidence of HAVM was 9.45% [95% CI 8.25-10.65]. The incidence of EVD-associated HAVM was 17.2 per 1000 EVD-days or 4.3% [95% CI 3.47-5.13] per 100 patients. Combining all three methods, we selected four important factors contributing to HAVM development: EVD, craniotomy, superficial surgical site infections after neurosurgery, and CSF leakage. The ML models performed better than regressions.
CONCLUSION: We first reported HAVM incidence in a neuro-ICU in Russia. We showed that tree-based ML is an effective approach to study risk factors because it enables the identification of nonlinear interaction across factors. We suggest that the number of found risk factors and the duration of their presence in patients should be reduced to prevent HAVM.