Prediction Models for Psychological Distress in Patients With Malignant Tumors: A Scoping Review

Area of Science:

• Oncology
• Psychological Medicine
• Biostatistics

Background:

• Psychological distress is prevalent in cancer patients, impacting treatment adherence and quality of life.
• Existing prediction models for identifying at-risk patients have seen limited clinical implementation.

Purpose of the Study:

• To synthesize the development methods, performance, validation, and limitations of psychological distress prediction models in cancer patients.
• To inform future research directions and facilitate clinical translation of these models.

Main Methods:

• A scoping review following Joanna Briggs Institute (JBI) methodology.
• Searched eight databases from inception to June 10, 2026.
• Independent screening, data extraction, and quality assessment by two reviewers.

Main Results:

• Included 13 studies with 26 prediction models; common methods included Logistic Regression (LR), Random Forests (RF), eXtreme Gradient Boosting (XGBoost), and Artificial Neural Network (ANN).
• Reported performance metrics varied widely (Sensitivities: 0.518-0.968, Specificities: 0.651-1.000, AUCs: 0.673-1.000).
• All studies reported high risk of bias, with only internal validation and no external validation; key predictors included tumor stage, sleep quality, pain, age, financial issues, and coping style.

Conclusions:

• Models using XGBoost, RF, and ANN reported high performance, but these results may be overestimated due to small sample sizes, low Events Per Variable (EPV), and lack of external validation.
• Methodological rigor, increased sample sizes, and external validation are crucial for future model development.
• Strengthening these aspects will support the clinical adoption of psychological distress prediction models in oncology care.