Nomogram model for predicting frailty of patients with hematologic malignancies – A cross-sectional survey

This study aimed to develop and validate an assessment tool for predicting and mitigating the risk of frailty in patients diagnosed with hematologic malignancies.A total of 342 patients with hematologic malignancies participated in this study, providing data on various demographics, disease-related information, daily activities, nutritional status, psychological well-being, frailty assessments, and laboratory indicators. The participants were randomly divided into training and validation groups at a 7:3 ratio. We employed Lasso regression analysis and cross-validation techniques to identify predictive factors. Subsequently, a nomogram prediction model was developed using multivariable logistic regression analysis. Discrimination ability, accuracy, and clinical utility were assessed through receiver operating characteristic (ROC) curves, C-index, calibration curves, and decision curve analysis (DCA).Seven predictors, namely disease duration of 6-12 months, disease duration exceeding 12 months, Charlson Comorbidity Index (CCI), prealbumin levels, hemoglobin levels, Generalized Anxiety Disorder-7 (GAD-7) scores, and Patient Health Questionnaire-9 (PHQ-9) scores, were identified as influential factors for frailty through Lasso regression analysis. The area under the ROC curve was 0.893 for the training set and 0.891 for the validation set. The Hosmer-Lemeshow goodness-of-fit test confirmed a good model fit. The C-index values for the training and validation sets were 0.889 and 0.811, respectively. The DCA curve illustrated a higher net benefit when using the nomogram prediction model within patients threshold probabilities ranging from 10% to 98%.This study has successfully developed and validated an effective nomogram model for predicting frailty in patients diagnosed with hematologic malignancies. The model incorporates disease duration (6-12 months and>12 months), CCI, prealbumin and hemoglobin levels, GAD-7, and PHQ-9 scores as predictive variables.