A model predicting the 6-year all cause mortality of patients with advanced schistosomiasis after discharge: Derived from a large population-based cohort study

Lanyue Pan, Chunmei Wu, Ping Li, Jiaquan Huang, Yizhi Wu and Guo Li

PLOS Neglected Tropical Diseases, 2025, vol. 19, issue 5, 1-16

Abstract: Background: Advanced schistosomiasis imposed a heavy economic burden on society and had a high rate of mortality and disability. However, methods for assessing its long-term prognosis were currently insufficient, and there was a lack of predictive tools to aid clinical decision-making and personalized follow-up plans for patients. We sought to determine risk factors associated with six-year all-cause mortality in advanced schistosomiasis, deriving and validating a six-year all-cause mortality prediction model through a retrospective cohort study based on a large population-based cohort. Methodology: We collected information from 4,136 patients with advanced schistosomiasis who were discharged between December 2014 and January 2015. After excluding 17 patients with the less common subtypes of colonic tumoroid proliferation and dwarfism, as well as 92 patients who were lost to follow-up or had incomplete information, data from 4,027 patients were included in the study. These patients were randomly assigned to the derivation cohort and the external validation cohort in a 7:3 ratio, with 1,400 patients randomly selected from the derivation cohort for internal validation. Sixteen candidate variables were collected: age, gender, nutritional status, splenectomy history, presence of other conditions (such as cardiovascular and digestive diseases), clinical classification, disease duration, ascites occurrence frequency, levels of serum total bilirubin (TBil), direct bilirubin (DBil), aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin (ALB), alkaline phosphatase (ALP), Hepatitis B surface antigen (HBsAg), and alpha-fetoprotein (AFP). High-risk factors associated with the 6-year mortality outcome were identified through univariate and multivariate Cox proportional hazards regression analyses. The predictive value of different models was evaluated and compared using the receiver operating characteristic (ROC) curves, Akaike information criterion (AIC), net reclassification improvement (NRI), C statistic, and integrated discrimination improvement (IDI). Findings: The derivation cohort comprised 2819 patients and we randomly selected 1400 cases from this cohort for internal validation. The external cohort consisted of 1208 patients. The mortality rate for three groups was around 27%-28%. We identified ten variables associated with increased risk of death, including age, course of disease, frequence of ascites, hepatitis B co-infection, and levels of DBil, ALT, AST, ALP, ALB, and AFP at baseline. Using these variables, we developed a ten-variable model and three simpler models. In the derivation cohort, the ten-variable model showed the highest C statistic (0.759; 95% CI, 0.739-0.778) and the lowest AIC (2834.2). ROC curves indicated an AUC of 0.759 for the ten-variable model, outperforming the simpler models. External validation also demonstrated superior performance of the ten-variable model with a higher C statistic (0.774; 95% CI, 0.749-0.797). This model consistently showed better results in ROC curves, IDI, continuous NRI, and categorical NRI analyses compared to the reduced models in external validation cohort. Conclusions: This study developed a multivariate model to predict the 6-year all-cause mortality rate in patients with advanced schistosomiasis, which demonstrated good performance. This convenient tool may potentially assist clinicians in formulating patient follow-up plans. Author summary: Existing methods for evaluating the long-term prognosis of advanced schistosomiasis were inadequate and needed improvement. This study aimed to identify risk factors associated with six-year all-cause mortality and to develop a predictive model using data from a large population-based cohort. We identified ten factors linked to an increased risk of death in patients with advanced schistosomiasis. Using these variables, we constructed a predictive model and validated its performance through various statistical measures, including ROC, AIC, NRI, and IDI, in both internal and external validation cohorts. The model that included all ten variables outperformed the reduced models. These findings highlight key indicators associated with six-year mortality in patients with advanced schistosomiasis, and the predictive model demonstrated strong performance through validation. This model may assist clinicians in making more informed decisions and developing personalized follow-up plans, potentially improving survival outcomes for patients with advanced schistosomiasis.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0013134 (text/html)
https://journals.plos.org/plosntds/article/file?id ... 13134&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pntd00:0013134

DOI: 10.1371/journal.pntd.0013134

Access Statistics for this article

More articles in PLOS Neglected Tropical Diseases from Public Library of Science
Bibliographic data for series maintained by plosntds ().