Construction of a predictive model for cognitive impairment risk in patients with advanced cancer

AimsThe purpose of this study was to identify risk factors for cognitive impairment in advanced cancer patients and to develop predictive models based on these risk factors. BackgroundCancer-related cognitive impairment seriously affects the quality of life of advanced cancer patients. However, neural network models of cognitive impairment in patients with advanced cancer have not yet been identified. DesignA cross-sectional design was used. MethodsThis study collected 494 questionnaires between January and June 2022. Statistically significant clinical indicators were selected by univariate analysis, and the artificial neural network model and logistic regression model were used for multivariate analysis. The predicted value of the model was estimated using the area under the subject's working characteristic curve. ResultThe artificial neural network and the logistic regression models suggested that cancer course, anxiety and age were the major risk factors for cognitive impairment in advanced cancer patients. All the indexes of artificial neural network model constructed in this study are better than those of the logistic model. ConclusionThe artificial neural network model can better predict the risk factors of cognitive impairment in patients with advanced cancer. Better prediction will enable nurses and other healthcare professionals to provide better targeted and timely support.