The Study Of The Correlation Between Olfactory Disorder Assessment And Chronic Neurodegenerative Disease By Clinical Decision Support System And Clinical Data Model

Objects:Olfaction is one of the basic feelings of human beings,and working as an essential sensation in human daily life.Olfactory disorders may commonly keep the elderly at risk in surrounding environment,including burnt foods,deteriorated foods,fire,toxic vapor,leaked natural gas.While seriously very few people knows that it can cause lethal loss of appetite and bodyweight,especially for the older.The elderly is an important part of the world's population and the main population effected by the olfactory dysfunction.Globally,the number of the elderly is growing faster than the numbers of people in any other age group.Meanwhile many diseases related to the elder are getting more and more people's attention,such as the olfactory disorders and the neurodegenerative disease.The current treatment of olfactory disorders is far from satisfactory.Doctors suffer from limited targeted treatments of olfactory disorders caused by partial causes,while patients can only accept such unsatisfactory treatment results.Many olfactory disorders involving unobtrusive or uncommon causes are ignored by both doctors and patients.Neurodegenerative diseases are a kind of chronic disease with irreversible course of disease,weak treatment effect,which will eventually lead to severe dysfunctions.Alzheimer's disease and Parkinson's disease are the two highest prevalence rates among the dysfunctions.It has been documented that olfactory impairment is associated with aging,and it is also an early sign of Parkinson's disease.Further studies have shown that at least 10%of immediate family members of patients with Parkinson's disease with primary olfactory impairment will have clinically defined Parkinson's disease within two years.And olfactory impairment is more likely to be an early clinical sign of dementia caused by Alzheimer's disease.The clinical decision support system is a computerized system that can assist healthcare providers,patients and other potential users to intelligently acquire or screen clinical disease data and knowledge,assist judgments on specific problems,and achieve the purpose of improving medical services and improving medical quality.CDSS is an effective and advanced application of electronic health records(EHR).Drawing on the mature experience of previous clinical decision support systems in managing refractory chronic systemic diseases,this project explores how to apply the approaches of medical informatics to effectively manage olfactory disorders,from the perspective of both doctors and patients,including family members,make personalized informatics prescriptions,and improve patients'quality of life.Further,the correlation between the severity of olfactory dysfunction and the occurrence of neurodegenerative diseases was explored through data analysis,and the factors and indicators related to the occurrence of neurodegenerative diseases in the common diagnosis and treatment of ENT were identified,and their contribution indexes were evaluated to provide targeted guidance for patients with different risk factors.Finally,the peripheral blood samples of patients with different risk levels above were detected by TREM2,and the correlation analysis was carried out to explore the supplement of laboratory objective indicators to the clinical decision-making system.Methods:1.Under the guidance of the 5R principle of CDSS theoretical framework,we systematically searched literature related to senile olfactory disorders in Pub Med and Medline over the past 14 years.We carefully scrutinized the morbidity,onset characteristics,inducing and prognostic factors,treatment methods and cure rates of olfactory disorders,and gradually analyzed the entire treatment process and identified the main indicators of concern.2.According to the results of previous studies,the characteristics of the disease,the degree of risk events,the quality of life score and the family history of neurodegenerative diseases are the main indicators for use to pay attention to,but there are still many indicators related to olfactory disorders and neurodegenerative diseases.The questionnaire was designed to collect data in six sections,including general condition,incidence,previous history,family history and disease history of immediate family members,adverse events caused by olfactory disorders,and relevant nasal examinations,with a total of 34items.The first draft of the questionnaire was evaluated and repeatedly revised by three nasal physician experts,and the data collection was graded,in an accurate and standardized manner.The statistics of adverse events,including the Chinese version of the general quality of life score in the United States,applied the subjective feelings of patients to the evaluation of the condition,and its reliability and validity have been confirmed and applied.3.Support vector machine(SVM)algorithm was deemed the best approach to study the data in this study.Since support vector machines are dimensionally transformed,there is no rigid requirement for the sample size of training set.Due to the limitation of the quantity of data,support vector machine is the best choice.Application of support vector machine(SVM)regression model(SMO),use Polynomial kernel function,the use of computing tools for Weka3.8.3,take 5–folded validation.The weight of each feature to the classifier,namely contribution index,is analyzed,and the evaluation model is obtained.4.Using the ROC curve,the performance of the model equation running on the new test dataset is investigated.5.Patients with different risk levels of Nd were grouped according to the quartile method,the expression levels of TREM2 in peripheral blood of high-risk group and low-risk group were compared,and their correlation was measured.Result:1.The current treatment status of olfactory disorders is not satisfactory.2.CDSS is suitable for the diagnosis and treatment of patients with olfactory disorders.The study benefited doctors,patients and guardians.3.Five key points for intervention were identified throughout the workflow:age,course of disease,occurrence of adverse events in life,quality of life score,and family history of neurodegenerative disease.4.Indicators with significant positive contribution to olfactory disorders include burnt food,disease course of 10-20 years,food deterioration undetected,disease course of 20-30 years,family history of neurodegenerative diseases,weight change,VAS evaluation as severe,QOD evaluation as severe,age of 80 years or above,disease course of 30-40 years;The indicators of reverse contribution were age 70-79 years old,course of disease of 1 year or less,course of disease of 1-10 years,age 60-69 years old,age 50-59 years old,excessive salt and sugar,degree of olfactory disturbance,normal or slight VAS,normal or slight QOD.5.A model for evaluating the incidence of neurodegenerative diseases using indicators related to olfactory disorders is as follows:P=0.3962×S_CR+0.3614×S_SF+0.3520×S_C10S+0.0130×S_C20S+0.0112×S_NDF+0.0081×S_VASS+0.0074×S_C30S+0.0072×S_QODS+0.0016×S_BWC-0.2431×S_C1-0.2095×S_CS-0.1397×S_A50S-0.1374×S_A60S-0.1361×S_A70S-0.0102×S_ODE-0.0045×S_VASM-0.0002×S_OSS-0.0001×S_QODM6.The ROC curve of model equation on the new dataset is significantly better than that of stochastic cases.7.The expression level of TREM2 m RNA in peripheral blood monocytes and the level of TREM2 protein in plasma were positively correlated with the risk coefficient of Nd in the high-risk group of OD patients,and the results were statistically significant.Conclusion:1.The application of CDSS clinical workflow can standardize the diagnostic process and remind doctors to pay attention to the problems that are easy to be ignored,especially the junior otolaryngologists,community doctors,even general practitioners,neurologists,doctors in the physical examination center,etc.,a series of medical workers who may contact patients with olfactory disorders.2.CDSS interventional diagnosis and treatment system can supplement diagnostic information and remind doctors to pay attention to the existence of olfactory disorders under the cover of many diseases.3.CDSS is widely used in the treatment of chronic refractory diseases,and the experience is worth learning.Informatics prescription is based on clinical diagnosis and pharmaceutical prescription.It is a guidance for patients'follow-up treatment and suggestion for life by applying medical informatics method.Health education is targeted at high-risk groups of different levels,namely patients themselves,their family members and relevant guardians,to avoid risks and improve the quality of life.4.Burn food and suffering from neurodegenerative disease are the most significant contributing factors.In terms of measuring the"course of disease"and"early manifestation of neurodegenerative diseases",the condition of patients with olfactory disorder tends to be stable when the"course of disease"is 10-20 years,and one year is far from enough to evaluate the prognosis of the disease.Although the severity of olfactory dysfunction subjective assessment(QOD and VAS)can reflect the impact on the quality of life of patients to some extent,they cannot be used to assess the prognosis of patients,especially the prediction of neurodegenerative diseases.5.The primary model is intended to be strengthened through larger sample size and wider application in both ENT and general medical practices.6.The expression of TREM2 in peripheral blood of OD patients with Nd in high-risk group was higher than that in high-risk group.7.In the high-risk group of OD patients with Nd,the expression level of TREM2 in peripheral blood was positively correlated with the risk coefficient of Nd.In the high-risk group of OD patients with Nd,the correlation was not statistically significant.