Predictive accuracy of comorbidity index models in assessing mortality risk among hemodialysis patients: A comprehensive single-center observational cohort study.

Abstract

Objectives: Comorbidity prediction models have been demonstrated to offer more comprehensive and accurate predictions of death risk compared to single indices. However, their application in China has been limited, particularly among maintenance hemodialysis (MHD) patients. Therefore, the objective of this study was to evaluate the utility of comorbidity index models in predicting mortality risk among Chinese MHD patients.

Methodology: The MHD patients in the First Affiliated Hospital of Guangzhou University of Traditional Chinese Medicine were taken as the subjects. Claims-based disease-specific refinements matching translation to ICD-10 and flexibility (CDMF-CCI) model and Liu model were selected as the candidate models for this verification research. Univariate and multivariate Cox regression calculations were used to analyze the independent predictive effect of the models on survival rate.

Results: Annually, nearly 500 patients undergo hemodialysis treatment. From January 2019 to June 2022, a total of 199 patients succumbed, with a mean age of 65.2 years. During these 4 years, the mortality rates were 13.04%, 9.68%, 11.69%, and 6.39%, respectively. The leading causes of death were sudden demise (82 patients, 41.2%), cardiovascular disease (48 patients, 24.1%), pulmonary infection (33 patients, 16.5%), and stroke (19 patients, 9.5%). When compared to individual indices, the CDMF-CCI model displayed more accurate and predictive results, with an HR of 1.190 (P = 0.037). Conversely, the Liu model failed to identify high-risk individuals.

Conclusion: The MHD patients face a significant risk of mortality. When compared to univariate parameters and the Liu model, the CDMF-CCI model exhibits superior predictive accuracy for mortality in MHD patients.