Development and validation of a nomogram for predicting cognitive frailty in patients on cancer

To investigate the current status of cognitive frailty in older cancer patients and construct a risk prediction model for cognitive frailty in older cancer patients.

Using convenience sampling, 308 older cancer patients from four wards in the oncology department of a grade-A tertiary hospital in Jiangsu Province from November 2023 to May 2024 were selected as the research subjects, including a training set of 215 cases (70%) and a validation set of 93 cases (30%). Data were collected through a general information questionnaire, Activities of Daily Living Scale, Mini-Nutritional Assessment Scale, Geriatric Depression Rating Scale, Pittsburgh Sleep Quality Index, Fried Frailty Scale, and Mini-Mental State Examination. A prediction model was established using Logistic regression, and a visual nomogram was constructed using R software. The model's discriminative ability and calibration were evaluated using the receiver operating characteristic (ROC) curve and calibration curve, respectively, and the clinical effectiveness was assessed using the clinical decision curve (DCA).

The incidence of cognitive frailty in older cancer patients was 26.7%. Logistic regression analysis revealed that education level, depression, sleep disorders, and malnutrition were influencing factors for cognitive frailty (P < 0.05). The Hosmer-Leme-show test of the nomogram model showed $$ =10.342, P = 0.242. The area under the ROC curve was 0.934, with a sensitivity and specificity of 81.1% and 94.1%, respectively.

Older cancer patients are at risk of cognitive frailty. The risk prediction model constructed in this study can conveniently and accurately predict the risk of cognitive frailty in older cancer patients, providing an important reference for clinical medical staff to conduct early assessment, screening, and intervention.