Bayesian model averaging for predicting factors associated with length of COVID-19 hospitalization

BMC Medical Research Methodology

Research

Mohammad Chehrazi¹  Karimollah Hajian-Tilaki²  Masomeh Bayani³  Abazar Amoozadeh⁴  Shabnam Bahrami⁵  Zahra Mohamadi-Pirouz⁵ 

[1]Department of Biostatistics and Epidemiology, School of Public Health, Babol University of Medical Sciences, Babol, Iran

[2]Neonatal Research Unit, Imperial College London, Exhibition Rd, South Kensington, SW7 2BX, London, UK

[3]Department of Biostatistics and Epidemiology, School of Public Health, Babol University of Medical Sciences, Babol, Iran

[4]Social Determinants of Health Research Center, Research Institute, Babol University of Medical Sciences, Babol, Iran

[5]Department of Infectious Diseases, Ayatollah Rohani Hospital, Babol University of Medical Sciences, Babol, Iran

[6]Social Determinants of Health Research Center, Research Institute, Babol University of Medical Sciences, Babol, Iran

[7]Student Research Center, Research Institute, Babol University of Medical Sciences, Babol, Iran

关键词: AIC;    GBDT;    Bayesian model averaging;    BIC;    COVID-19;    Length of hospital stay;    Markov Chain Monte Carlo (MCMC);    Occam's Window;    Stepwise;   

DOI  :  10.1186/s12874-023-01981-x

 received in 2023-01-30, accepted in 2023-06-18,  发布年份 2023

来源: Springer

【 摘 要 】

IntroductionThe length of hospital stay (LOHS) caused by COVID-19 has imposed a financial burden, and cost on the healthcare service system and a high psychological burden on patients and health workers. The purpose of this study is to adopt the Bayesian model averaging (BMA) based on linear regression models and to determine the predictors of the LOHS of COVID-19.MethodsIn this historical cohort study, from 5100 COVID-19 patients who had registered in the hospital database, 4996 patients were eligible to enter the study. The data included demographic, clinical, biomarkers, and LOHS. Factors affecting the LOHS were fitted in six models, including the stepwise method, AIC, BIC in classical linear regression models, two BMA using Occam's Window and Markov Chain Monte Carlo (MCMC) methods, and GBDT algorithm, a new method of machine learning.ResultsThe average length of hospitalization was 6.7 ± 5.7 days. In fitting classical linear models, both stepwise and AIC methods (R2 = 0.168 and adjusted R2 = 0.165) performed better than BIC (R2 = 0.160 and adjusted = 0.158). In fitting the BMA, Occam's Window model has performed better than MCMC with R2 = 0.174. The GBDT method with the value of R2 = 0.64, has performed worse than the BMA in the testing dataset but not in the training dataset. Based on the six fitted models, hospitalized in ICU, respiratory distress, age, diabetes, CRP, PO2, WBC, AST, BUN, and NLR were associated significantly with predicting LOHS of COVID-19.ConclusionThe BMA with Occam's Window method has a better fit and better performance in predicting affecting factors on the LOHS in the testing dataset than other models.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

41116_2023_38_Article_IEq117.gif

41116_2023_38_Article_IEq116.gif

41116_2023_38_Article_IEq115.gif

41116_2023_38_Article_IEq114.gif

Fig. 2

41116_2023_38_Article_IEq112.gif

41116_2023_38_Article_IEq111.gif

41116_2023_38_Article_IEq110.gif

41116_2023_38_Article_IEq109.gif

41116_2023_38_Article_IEq108.gif

41116_2023_38_Article_IEq107.gif

41116_2023_38_Article_IEq106.gif

41116_2023_38_Article_IEq105.gif

41116_2023_38_Article_IEq104.gif

41116_2023_38_Article_IEq103.gif

41116_2023_38_Article_IEq101.gif

41116_2023_38_Article_IEq100.gif

41116_2023_38_Article_IEq99.gif

41116_2023_38_Article_IEq98.gif

41116_2023_38_Article_IEq97.gif

41116_2023_38_Article_IEq96.gif

41116_2023_38_Article_IEq95.gif

41116_2023_38_Article_IEq94.gif

Fig. 1

41116_2023_38_Article_IEq92.gif

41116_2023_38_Article_IEq91.gif

41116_2023_38_Article_IEq90.gif

41116_2023_38_Article_IEq89.gif

41116_2023_38_Article_IEq88.gif

41116_2023_38_Article_IEq87.gif

41116_2023_38_Article_IEq86.gif

41116_2023_38_Article_IEq85.gif

41116_2023_38_Article_IEq84.gif

41116_2023_38_Article_IEq83.gif

41116_2023_38_Article_IEq82.gif

41116_2023_38_Article_IEq81.gif

41116_2023_38_Article_IEq80.gif

41116_2023_38_Article_IEq79.gif

41116_2023_38_Article_IEq78.gif

41116_2023_38_Article_IEq77.gif

41116_2023_38_Article_IEq76.gif

41116_2023_38_Article_IEq75.gif

41116_2023_38_Article_IEq74.gif

41116_2023_38_Article_IEq73.gif

41116_2023_38_Article_IEq72.gif

41116_2023_38_Article_IEq71.gif

41116_2023_38_Article_IEq70.gif

41116_2023_38_Article_IEq69.gif

41116_2023_38_Article_IEq68.gif

41116_2023_38_Article_IEq67.gif

Fig. 4

41116_2023_38_Article_IEq66.gif

41116_2023_38_Article_IEq65.gif

41116_2023_38_Article_IEq64.gif

41116_2023_38_Article_IEq63.gif

13065_2023_981_Figaa_HTML.gif

41116_2023_38_Article_IEq62.gif

41116_2023_38_Article_IEq61.gif

41116_2023_38_Article_IEq60.gif

41116_2023_38_Article_IEq59.gif

41116_2023_38_Article_IEq57.gif

41116_2023_38_Article_IEq56.gif

41116_2023_38_Article_IEq55.gif

Fig. 3

41116_2023_38_Article_IEq54.gif

41116_2023_38_Article_IEq53.gif

41116_2023_38_Article_IEq52.gif

41116_2023_38_Article_IEq51.gif

41116_2023_38_Article_IEq50.gif

41116_2023_38_Article_IEq49.gif

41116_2023_38_Article_IEq48.gif

41116_2023_38_Article_IEq47.gif

41116_2023_38_Article_IEq46.gif

41116_2023_38_Article_IEq45.gif

41116_2023_38_Article_IEq44.gif

41116_2023_38_Article_IEq43.gif

41116_2023_38_Article_IEq42.gif

41116_2023_38_Article_IEq41.gif

41116_2023_38_Article_IEq40.gif

41116_2023_38_Article_IEq39.gif

41116_2023_38_Article_IEq38.gif

41116_2023_38_Article_IEq37.gif

41116_2023_38_Article_IEq36.gif

41116_2023_38_Article_IEq35.gif

41116_2023_38_Article_IEq31.gif

12888_2023_4982_Article_IEq6.gif

12888_2023_4982_Article_IEq5.gif

12888_2023_4982_Article_IEq2.gif