Study Of Fall Risks And Fall Prediction Model In Patients With Hematological Malignancies

Objectives To explore the risk factors for falls in patients with hematological malignancies(HM),develop a fall risk prediction model,and provide evidence to support fall prevention and management in clinical practice.Methods Extraction of risk factors: A research team was formed to conduct a comprehensive literature review of databases and extract relevant risk factors for falls in HM patients.Following the literature review,group discussions and two rounds of expert consultation,a questionnaire on risk factors for falls in HM patients was formulated.Development and evaluation of the risk prediction model: A cross-sectional study was conducted with 347 HM patients from five tertiary hospitals in Jinan City,using convenience sampling.A self-designed general information questionnaire,falls incidence questionnaire,and risk factors questionnaire were used to collect patient data.SPSS 26.0software and R language were used to analyze the data.Univariate and multivariate analyses were used to screen the predictive factors for the model.The predictive factors were used as independent variables,and the occurrence of falls in the past year was used as the outcome variable in the model.Receiver operating characteristic(ROC)curves and the Hosmer-Lemeshow(H-L)goodness-of-fit test were used to evaluate the discriminative and calibration performance of the model,respectively.Results Results of risk factor extraction: Based on literature review and expert consultation,a questionnaire on risk factors for falls in HM patients was developed,including six aspects: general information,function and activity,disease and treatment-related factors,medication factors,psychological and cognitive factors,and caregiver factors.Results of risk prediction model development: A total of 347 HM patients were included in the study,and 67 patients experienced falls in the past year,with a fall incidence rate of 19.3%.Univariate analysis showed that there were statistically significant differences(P < 0.05)between the fall and non-fall groups in terms of dynamic balance or gait abnormalities,assistive devices,exercise status,dizziness history in the past year,peripheral neuropathy,comorbidities,hypertension,diabetes,steroid use,antihypertensive medication,hypoglycemic medication,polypharmacy,cognitive status,fear of falling,patient’s perception of fall risk,caregiver’s perception of fall risk,chronic pain,Barthel index,and PHQ-9 score.Logistic regression analysis showed that the independent risk factors for falls in HM patients were dizziness history in the past year(OR=2.795,95%CI:1.556-5.020),steroid use(OR=2.017,95%CI:1.121-3.628),Barthel index(OR=0.981,95%CI:0.967-0.995),and PHQ-9 score(OR=1.065,95%CI:1.008-1.125).According to the logistic regression-based fall risk prediction model,the area under the curve(ROC)was 0.748(95% CI:0.681-0.816),the c-statistic was 0.417,the sensitivity was0.567,and the specificity was 0.850.The Homers-Lemeshow(H-L)test yielded a P-value of 0.477,and the calibration plot showed good agreement between predicted and observed fall events.Conclusions Our study developed a fall risk prediction model for HM patients which comprises of four predictors: history of dizziness in the past year,use of steroids,Barthel index,and PHQ-9 score.The model can aid in identifying individuals at high risk of falls and inform the development of targeted prevention strategies in this population.