tetano
Editor, Senior Moderator
BMC Pulm Med
. 2026 Jul 24;26(1):423.
doi: 10.1186/s12890-026-04409-6.
Zhang Haojing 1 , Kan Lin 1 , Pan Dianzhu 2
Affiliations Expand
Objective: To explore the risk factors of Long COVID and to construct a nomogram to predict the occurrence of Long COVID.
Method: This study was an observational study. Clinical data were collected from patients diagnosed with COVID-19 and hospitalized at the First Affiliated Hospital of Jinzhou Medical University Hospital from December 7, 2022, to February 1, 2023. The prediction model was constructed using a nomogram. The clinical data of patients in Panjin Central Hospital Hospital from December 7, 2022 to December 7, 2023 were used for external validation.
Results: In the development cohort and the validation cohort of this study, 60.3% and 62.3% of the patients developed Long COVID, respectively. After Least absolute shrinkage and selection operator regression, the final variables included in the prediction model were percentage of lymphocyte, the Charlson Comorbidity Index, Computed Tomography score, and oxygen requirement. The Area Under the Receiver Operating Characteristic for external validation of the model is 0.786, and the p value of Spiegelhalter test was 0.126. The p value of the Hosmer and Lemeshow chi-square statistic was 0.098. The decision curve analysis indicates that the model performs well.
Conclusion: The prediction model developed in this study is useful for assessing the likelihood of developing Long COVID in hospitalized patients.
Keywords: COVID-19; Forecast; Long COVID; Nomogram; Risk factor.
. 2026 Jul 24;26(1):423.
doi: 10.1186/s12890-026-04409-6.
Building and externally validating a prediction model for long COVID in severe and critical COVID-19 patients: a multi-center cohort study
Zhang Haojing 1 , Kan Lin 1 , Pan Dianzhu 2
Affiliations Expand
- PMID: 42778908
- PMCID: PMC13602745
- DOI: 10.1186/s12890-026-04409-6
Abstract
Objective: To explore the risk factors of Long COVID and to construct a nomogram to predict the occurrence of Long COVID.
Method: This study was an observational study. Clinical data were collected from patients diagnosed with COVID-19 and hospitalized at the First Affiliated Hospital of Jinzhou Medical University Hospital from December 7, 2022, to February 1, 2023. The prediction model was constructed using a nomogram. The clinical data of patients in Panjin Central Hospital Hospital from December 7, 2022 to December 7, 2023 were used for external validation.
Results: In the development cohort and the validation cohort of this study, 60.3% and 62.3% of the patients developed Long COVID, respectively. After Least absolute shrinkage and selection operator regression, the final variables included in the prediction model were percentage of lymphocyte, the Charlson Comorbidity Index, Computed Tomography score, and oxygen requirement. The Area Under the Receiver Operating Characteristic for external validation of the model is 0.786, and the p value of Spiegelhalter test was 0.126. The p value of the Hosmer and Lemeshow chi-square statistic was 0.098. The decision curve analysis indicates that the model performs well.
Conclusion: The prediction model developed in this study is useful for assessing the likelihood of developing Long COVID in hospitalized patients.
Keywords: COVID-19; Forecast; Long COVID; Nomogram; Risk factor.