Construction And Application Of Comprehensive Management Program For Elderly Patients With Fragility Fractures Based On Accurate Identification Of Fall Risk

ObjectiveFragility fractures and refractures caused by falls have become serious problems in the elderly,significantly affecting their quality of life and health.This thesis aims to construct an accurate prediction model of fall risk in the elderly using machine learning algorithms to identify people at high risk of falls and determine the key factors affecting falls.On this basis,and combined with the findings of the modeling analyses,literature results,and theoretical review,a comprehensive intervention management program was constructed for elderly patients with a high risk of falls.Ultimately,comprehensive management for the elderly after the fragility fracture can be promoted,reducing the risk of falls and preventing the occurrence of refractures.MethodsThis thesis can be divided into three parts:1.Prediction model of fall risk in the elderly based on machine learningWe extracted and labeled fall risk factors and samples of the elderly collected in the previous period to form different data sets.The risk prediction model was constructed using five different machine learning algorithms based on Python,including K-Nearest Neighbor,Support Vector Machine,Random Forest,Extreme Gradient Boosting,and Multilayer Perceptron.The accuracy,F1 value,and AUC value of the model were compared to determine the optimal model and the critical intervention factors of falling in the elderly extracted from the corresponding model.2.Construction of a comprehensive intervention management program for the elderly with a high risk of fragility fractures caused by fallsBased on the discovery of the fall risk prediction model,through a systematic review of the literature and relevant theories,a draft of the comprehensive intervention management program for the elderly with a high risk of fragility fractures caused by falls was formed.On this basis,experts in related fields are invited to revise the scheme further.3.Application of the comprehensive intervention management for the elderly patients with a high risk of fragility fractures caused by fallsThe constructed comprehensive intervention management program the elderly patients with a high risk of fragility fractures caused by falls will be applied in clinical practice.The program’s effectiveness was evaluated by comparing the following indicators before and after the intervention,including the risk of falling,balance ability,fear of falling,fracture risk,osteoporosis knowledge level,self-efficacy,self-management behavior,medication compliance,the incidence of adverse events,quality of life,etc.Results1.After data pre-treatment,a total of 3940 effective fall-related data samples and 81 influential characteristic factors were included,forming four characteristic data sets.The model construction results showed that the prediction model for the risk of falls and fall risk classification based on the random forest algorithm had the highest accuracy,reaching 97.33%and 100%,respectively.The corresponding AUC values were 0.984 and 0.999,respectively.By extracting and combining the features of the two optimal models,it was found that the balance ability was the most critical intervention factor affecting falls in the elderly.2.Based on the discovery of the risk prediction model,the intervention target,intervention object inclusion,and evaluation index content elements were determined.The draft plan was prepared on the basis of the fracture liaison service model element framework,the optimal practice framework standards,and the guidance of communication theory.Moreover,twelve experts in related fields with an authority coefficient of 0.871 were selected.The ultimate scheme includes intervention objectives,intervention objects,intervention location,intervention timing,intervention means,core intervention content,intervention personnel,implementation process,main evaluation indicators,and so on.3.80 elderly patients with a high risk of falls who met the inclusion and exclusion criteria of the fall risk prediction model were selected.The number of elderly patients who completed the entire intervention study in the intervention and control groups was 36 and 35,respectively.During the intervention period,there was no statistical difference in the occurrence of falls and refractures between the two groups(P > 0.05).After the intervention,the risk of falls,balance ability,and fear of falling for patients in the intervention group showed essentially improvement compared to the control group(P < 0.05).Meanwhile,the osteoporosis-related knowledge,selfefficacy and self-management behavior,medication compliance,and quality of life also improved significantly(P < 0.05).However,there is no significant difference between the two groups of patients with fracture risk(P > 0.05).Conclusion1.The prediction model of the risk of falls and fall risk classification for the elderly based on the random forest algorithm shows better accurately,which can reasonably identify the population with a high risk of falls and provide support for targeted fall prevention.2.The comprehensive intervention management program guided by the fracture liaison service model element framework and the communication theory for patients with a high risk of fragility fracture caused by falls in the elderly has scientific solid and feasibility.It can improve patients’ balance ability,self-management behavior,treatment compliance,quality of life,and reduce the risk of falls.However,the effect on reducing the risk of refractures needs to be further explored.