Related Experiment Video
Updated: Jul 8, 2025
Evaluation of a Point-of-Care Testing Analyzer for Measuring Peripheral Blood Leukocytes
Published on: March 22, 2022
Adjustment for collider bias in the hospitalized Covid-19 setting
Moslem Taheri Soodejani1, Seyyed Mohammad Tabatabaei2, Mohammad Hassan Lotfi1
1Center for Healthcare Data Modeling, Departments of Biostatistics and Epidemiology, School of Public Health, Shahid Sadoughi University of Medical Sciences, Yazd, Iran.
Insights
Causal directed acyclic graphs (cDAGs) help adjust for collider bias in hospitalized COVID-19 patients. Accounting for age and comorbidities reveals vaccination’s protective effect against COVID-19 death.
Area of Science:
- Epidemiology
- Biostatistics
- Causal Inference
Background:
- Causal directed acyclic graphs (cDAGs) are essential for identifying confounding and collider bias.
- This study focuses on adjusting for collider bias within the context of hospitalized COVID-19 patients.
Purpose of the Study:
- To demonstrate the application of cDAGs for mitigating collider bias in COVID-19 research.
- To investigate the association between vaccination and mortality in hospitalized COVID-19 patients while addressing selection bias.
Main Methods:
- Three statistical models were employed, progressively incorporating covariates.
- Model 1: Vaccination as the sole independent variable.
- Models 2 and 3: Included age and comorbidities, respectively, to adjust for collider bias introduced by conditioning on hospitalization.
Main Results:
- Initial analysis (Model 1) showed no significant effect of vaccination on COVID-19 mortality.
- Adjusting for age (Model 2) revealed a protective effect of vaccination.
- Further adjustment for comorbidities (Model 3) strengthened this observed protective effect.
Conclusions:
- Hospitalized patient cohorts are susceptible to collider-stratification bias.
- This bias can be effectively managed by including risk factors associated with both the outcome and the selection criteria (hospitalization) in regression models.
Background:
Causal directed acyclic graphs (cDAGs) are frequently used to identify confounding and collider bias. We demonstrate how to use causal directed acyclic graphs to adjust for collider bias in the hospitalized Covid-19 setting.
Materials And Methods:
According to the cDAGs, three types of modeling have been performed. In model 1, only vaccination is entered as an independent variable. In model 2, in addition to vaccination, age is entered the model to adjust for collider bias due to the conditioning of hospitalization. In model 3, comorbidities are also included for adjustment of collider bias due to the conditioning of hospitalization in different biasing paths intercepting age and comorbidities.
Results:
There was no evidence of the effect of vaccination on preventing death due to Covid-19 in model 1. In the second model, where age was included as a covariate, a protective role for vaccination became evident. In model 3, after including chronic diseases as other covariates, the protective effect was slightly strengthened.
Conclusion:
Studying hospitalized patients is subject to collider-stratification bias. Like confounding, this type of selection bias can be adjusted for by inclusion of the risk factors of the outcome which also affect hospitalization in the regression model.
More Related Videos
Related Concept Videos
01:25Transmission-based Precautions II: Airborne and Protective Environment
Airborne precautions:
Use airborne precautions when treating patients known or suspected to have diseases that spread through the air—for example, tuberculosis or measles. These organisms are present in smaller droplets expelled by an infected person and...
01:23Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast,...
01:12Hazard Ratio
For example, in a clinical trial...
01:13Controls in Experiments
01:09Censoring Survival Data
01:29Contingency Table