Constructing a Predictive Model for Psychological Distress of Young- and Middle-Aged Gynaecological Cancer Patients

Abstract

Background

Cancer patients experience substantial psychological distress which causes the reduction of the quality of life. However, the risk of psychological distress has not been well predicted especially in young- and middle-aged gynaecological cancer patients. This study aimed to develop a prediction model for psychological distress in young- and middle-aged gynaecological cancer patients using the artificial neural network (ANN).

Methods

A cross-sectional study of young- and middle-aged gynaecological cancer patients (n = 368) was conducted between March and December 2022. We used the univariate analysis to determine the factors affecting psychological distress. ANN was used for psychological distress prediction in young- and middle-aged gynaecological cancer patients. Also, a traditional logistic regression (LR) model was constructed for comparison. The area under the receiver's operating characteristic curve (AUC) was used to evaluate the model's predictive performance.

Results

ANN and LR showed that self-efficacy, economic income and sleep duration were the top risk variables for psychological distress in young- and middle-aged gynaecological cancer patients. The AUC of the ANN was 0.977, the sensitivity was 94.83% and the specificity was 86.44%, whereas logistic regression's were 0.920, 85.57% and 82.76%, respectively.

Conclusion

Compared with the LR model, the ANN model shows obvious superiority across all assessment index outcomes, and it may be used as a decision-support tool for early identification of young- and middle-aged gynaecological cancer patients suffering from psychological distress.