糖尿病心脏功能超声数据模型的初步研究

发布时间：2018-05-26 02:04

本文选题：数据库 + 数据挖掘　；参考：《皖南医学院》2014年硕士论文

【摘要】：目的：以糖尿病患者作为研究对象，，采用先进的数据库开发平台，获取糖尿病患者临床信息、基本信息并通过采集超声心动图原始构型及功能数据信息建立数据库，根据数据挖掘（Data Mining，DM）原理，利用数据挖掘工具对大样本的糖尿病患者的临床相关数据及心脏功能的有关超声数据进行数据挖掘，得到影响糖尿病心脏功能（左室舒张功能）变化的关键影响因素及特征性指标，并揭示出各影响因素与心脏功能改变之间的潜在的、有价值的规律，分析各指标之间地变化规律，了解糖尿病心脏功能的动态变化，并通过数据模型的建立对糖尿病心脏功能的变化进行预测并对糖尿病心脏并发症的早期干预、动态监测、临床治疗、疗效观察等提供可行性的决策支持。方法：选择2012年7月至2014年2月来我院进行超声心动图检查检查并已临床确诊为糖尿病的360例患者，采用高档彩色多普勒超声诊断仪按照统一规范化检查、测量进行操作，获取糖尿病患者心脏构型、功能超声数据，并且记录受检者身高、体重、腹围、体表面积等形体指标，通过与医院信息系统（hospital informationsystem，HIS）连接查询获取糖尿病患者临床信息及基本信息（患者有无糖尿病家族史、血糖、血脂、血压、是否高糖、高热量饮食、是否高蛋白饮食、是否抽烟、是否饮酒、以及心电图的改变等相关数据），用以上数据源建立糖尿病心脏功能数据库，以SQL Server2008作为数据库管理工具，利用SQL Server2008数据挖掘工具对糖尿病患者的基本信息、临床信息及反映心脏功能的超声数据进行数据挖掘，挖掘出各影响因素与糖尿病心脏功能（左室舒张功能）改变的潜在的、有价值的规律，分析各指标之间地变化规律，了解糖尿病心脏功能（左室舒张功能）的动态变化，初步设计了糖尿病患者心脏功能（左室舒张功能）改变的简单模型。结果：建立包含糖尿病患者的基本信息、临床信息及心脏构型、功能数据的关系数据库，并对关系数据库进行预处理转换成适合数据挖掘的事务数据库，进行数据挖掘初步设计了糖尿病患者心脏功能（左室舒张功能）改变的初步模型。通过功能数据模型的建立对糖尿病心脏功能的变化进行预测并对糖尿病心脏并发症的早期干预、动态监测、临床治疗、疗效观察等提供可行性的决策支持。提供了一种对糖尿病心脏功能变化的影响因素、相关的功能指标进行分析的方法，通过数据挖掘我们可以得到糖尿病心脏功能改变与糖尿病遗传史、血糖、血压、血脂、BMI及腹围、左房容积指数等密切相关，可以初步建立反映糖尿病患者左室舒张功能变化的数据模型。结论：利用数据挖掘技术可以有效地对大量的医学信息进行挖掘，可以从中提取有价值的规则并获取知识，可以及时准确地对心脏功能改变进行预测，对糖尿病心脏并发症的早期干预、动态监测、临床治疗、疗效观察等提供可行性的决策支持。对糖尿病心血管并发症的临床早期诊断具有一定的现实意义。
[Abstract]:Objective: Taking the diabetic patients as the research object, using the advanced database development platform to obtain the clinical information of the patients with diabetes, the basic information and the establishment of the database by collecting the original configuration and functional data of echocardiography, according to the principle of data mining (Data Mining, DM), and using data mining tools for the diabetes of large samples The clinical data of the patients and the data of echocardiography related to cardiac function were excavated, and the key influencing factors and characteristic indexes of the changes of cardiac function (left ventricular diastolic function) were obtained, and the potential, valuable rules between the factors and the changes of cardiac function were revealed, and the changes between the indexes were analyzed. Regularity, understand the dynamic changes of cardiac function of diabetes, predict the changes of diabetic cardiac function through the establishment of data model and provide the feasible decision support for the early intervention, dynamic monitoring, clinical treatment and therapeutic observation of diabetic cardiac complications.
Methods: from July 2012 to February 2014, 360 patients with diabetes mellitus were diagnosed by echocardiography in our hospital and 360 cases of diabetes have been diagnosed. The high grade color Doppler ultrasonic diagnostic instrument was used in accordance with the unified and standardized examination, and the operation was carried out to obtain the cardiac structure, function ultrasound data and record the height of the subjects. Body weight, abdominal circumference, body surface area and other physical indicators, through the hospital information system (hospital informationsystem, HIS) access to access the clinical information and basic information of diabetes patients (patients have no diabetes family history, blood sugar, blood lipids, blood pressure, high sugar, high calorie diet, high protein diet, smoking, whether or not drinking, and The diabetes heart function database was set up with the above data sources. SQL Server2008 was used as the database management tool. The basic information of diabetic patients, the clinical information and the ultrasonic data reflecting the heart function were excavated with the SQL Server2008 data mining tool, and the influencing factors were excavated. With the potential and valuable rules of the changes in cardiac function (left ventricular diastolic function), the changes in the changes in the ground function (left ventricular diastolic function) of diabetic patients were analyzed, and a simple model for the changes of cardiac function (left ventricular diastolic function) in diabetic patients was preliminarily designed.
Results: the basic information, clinical information and the relationship database of cardiac configuration and functional data were established, and the relational database was preprocessed into a transaction database suitable for data mining, and a preliminary design of the cardiac function (left ventricular diastolic function) of diabetic patients was preliminarily designed by data mining. The establishment of an over functional data model predicts the changes in diabetic cardiac function and provides a feasible decision support for the early intervention of diabetic cardiac complications, dynamic monitoring, clinical treatment, and therapeutic observation. Through data mining, we can obtain diabetes heart function change and genetic history of diabetes, blood sugar, blood pressure, blood lipid, BMI and abdominal circumference, left atrial volume index and so on. We can establish a data model that reflects the changes of left ventricular diastolic function in diabetic patients.
Conclusion: the use of data mining technology can effectively excavate a large number of medical information. It can extract valuable rules and acquire knowledge. It can predict the changes of cardiac function in time and accurately. It can provide the feasibility of early intervention, dynamic monitoring, clinical treatment and observation of curative effect of diabetic cardiac complications. Policy support is of practical significance for the early diagnosis of diabetic cardiovascular complications.
【学位授予单位】：皖南医学院
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R587.1;R445.1

【参考文献】