Development and validation of a nomogram for individualizing fall risk in patients with hematologic malignancies

Inpatient falls are common adverse events especially for patients with hematologic malignancies. A fall-risk prediction model for patients with hematologic malignancies are still needed. Here we conducted a multicenter study that prospectively included 516 hospitalized patients with hematologic malignancies, and developed a nomogram for fall risk prediction. Patients were divided into the modeling group (n = 389) and the validation group (n = 127). A questionnaire containing sociodemographic factors, general health factors, disease-related factors, medication factors, and physical activity factors was administered to all patients. Logistic regression analysis revealed that peripheral neuropathy, pain intensity, Morse fall scale score, chemotherapy courses, and myelosuppression days were risk factors for falls in patients with hematologic malignancies. The nomogram model had a sensitivity of 0.790 and specificity of 0.800. The calibration curves demonstrated acceptable agreement between the predicted and observed outcomes. Therefore, the nomogram model has promising accuracy in predicting fall risk in patients with hematologic malignancies.