ABSTRACT
Survival data with a multi-state structure are frequently observed in follow-up studies. An analytic approach based on a multi-state model (MSM) should be used in longitudinal health studies in which a patient experiences a sequence of clinical progression events. One main objective in the MSM framework is variable selection, where attempts are made to identify the risk factors associated with the transition hazard rates or probabilities of disease progression. The usual variable selection methods, including stepwise and penalized methods, do not provide information about the importance of variables. In this context, we present a two-step algorithm to evaluate the importance of variables formulti-state data. Three differentmachine learning approaches (randomforest, gradient boosting, and neural network) as themost widely usedmethods are considered to estimate the variable importance in order to identify the factors affecting disease progression and rank these factors according to their importance. The performance of our proposed methods is validated by simulation and applied to the COVID-19 data set. The results revealed that the proposed two-stage method has promising performance for estimating variable importance.
ABSTRACT
Survival data with a multi-state structure are frequently observed in follow-up studies. An analytic approach based on a multi-state model (MSM) should be used in longitudinal health studies in which a patient experiences a sequence of clinical progression events. One main objective in the MSM framework is variable selection, where attempts are made to identify the risk factors associated with the transition hazard rates or probabilities of disease progression. The usual variable selection methods, including stepwise and penalized methods, do not provide information about the importance of variables. In this context, we present a two-step algorithm to evaluate the importance of variables for multi-state data. Three different machine learning approaches (random forest, gradient boosting, and neural network) as the most widely used methods are considered to estimate the variable importance in order to identify the factors affecting disease progression and rank these factors according to their importance. The performance of our proposed methods is validated by simulation and applied to the COVID-19 data set. The results revealed that the proposed two-stage method has promising performance for estimating variable importance. © 2023 Tech Science Press. All rights reserved.
ABSTRACT
Survival data with a multi-state structure are frequently observed in follow-up studies. An analytic approach based on a multi-state model (MSM) should be used in longitudinal health studies in which a patient experiences a sequence of clinical progression events. One main objective in the MSM framework is variable selection, where attempts are made to identify the risk factors associated with the transition hazard rates or probabilities of disease progression. The usual variable selection methods, including stepwise and penalized methods, do not provide information about the importance of variables. In this context, we present a two-step algorithm to evaluate the importance of variables for multi-state data. Three different machine learning approaches (random forest, gradient boosting, and neural network) as the most widely used methods are considered to estimate the variable importance in order to identify the factors affecting disease progression and rank these factors according to their importance. The performance of our proposed methods is validated by simulation and applied to the COVID-19 data set. The results revealed that the proposed two-stage method has promising performance for estimating variable importance.
ABSTRACT
Background and Objectives: Hospitals are often as an important goal focus on length of stay (LOS) for controlling the cost of trauma care. This study aimed to investigate the factors affecting hospital LOS in trauma patients before and during the COVID-19 pandemic in a regional trauma center in Iran. Methods: In this retrospective study, all trauma patients referred to Be'sat Hospital, the only regional trauma center in Hamadan, from 19 February to 20 November 2020 (during the COVID-19) were compared to the same date in 2019 (before the COVID-19). Data were collected using the recorded information in the Health Information Management Center of the Be'sat Hospital. Multivariate logistic regression was used to simultaneously evaluate the effect of covariates on LOS (≤3 days;>3 days) in two periods. All statistical analyses were performed using SPSS version 24.0. Results: The mean values for LOS in trauma patients before and during the COVID-19 were 3.9 and 3.2 days, respectively. Before the COVID-19, the age of <18 years old (odds ratio [OR] = 1.59;95% confidence interval CI: 1.33-1.92, P < 0.01), winter season (OR = 1.33;95% CI: 0.99-1.76, P = 0.04), and burn trauma (OR = 1.35;95% CI = 1.02-1.79, P = 0.03) were significant. During the COVID-19, the age of <18 years old (OR = 1.39;95% CI: 1.04-1.90, P = 0.04), males (OR = 1.48;95% CI: 1.10-1.99, P = 0.01), burn trauma (OR = 1.77;95% CI: 1.15-2.73, P = 0.01), and history of hospitalization (OR = 1.77;95% CI: 1.15-2.73, P = 0.01) were significant. Conclusions: According to the results, before the COVID-19, the factors such as age, season, and mechanism of trauma were predictors of LOS. Furthermore, age, sex, mechanism of trauma, and history of hospitalization were factors that affected LOS in trauma patients during the COVID-19. Thus, this study may be helpful in improving the discharge planning in trauma patients.