分段常数强度Markov模型在阿尔茨海默病进程研究中的应用

发布时间：2018-07-14 12:15

【摘要】：目的尝试将分段常数强度Markov模型引入到轻度认知损害(Mild Cognitive Impairment, MCI)向阿尔茨海默病(Alzheimer's disease, AD)转归过程中,深入研究影响转归过程的影响因素,为制定不同发展阶段的预防措施提供理论依据。从应用角度出发,对分段常数强度Markov模型在慢性病中的研究进行方法学的探试,为其他多状态慢性病不同发展阶段影响因素的探讨提供方法学借鉴。方法应用太原市600名社区老年人的4次随访资料,以MCI为状态1,中重度认知损害为状态2,AD为状态3,拟合分段常数强度Markov模型,分析MCI向AD转归不同发展阶段的影响因素。结果经假设检验，，,数据满足Markov性(P=0.89),不满足时齐性(P0.001),可应用分段常数强度Markov模型拟合。采用分段常数强度Markov模型,对于第个时间区间[0,12]月,经多因素筛选,发现性别、年龄、文化程度、职业和高血压对状态1到状态2的转移有统计学意义,年龄、文化程度、职业、高血压和糖尿病对状态1到状态3的转移有统计学意义,性别、年龄、文化程度、高血压和糖尿病对状态2到状态3的转移有统计学意义。对于第二个时间区间[12,18]月,经多因素筛选,发现性别、年龄、文化程度、职业、吸烟和高血压对状态1到状态2的转移有统计学意义,文化程度、职业、吸烟、高血压和糖尿病对状态1到状态3的转移有统计学意义,年龄、文化程度、职业、吸烟、高血压和糖尿病对状态2到状态3的转移有统计学意义。因此,女性、年龄、吸烟、高血压和糖尿病是MCI向AD转归的危险因素,高文化程度和从事脑力劳动是MCI向AD转归的保护因素。根据模型作出生存曲线并计算转移强度矩阵和3年转移概率矩阵。结论为延缓尔茨海默病的疾病进程，应根据影响疾病不同阶段转移的影响因素,开展疾病分阶段重点防治。与齐次Markov模型相比,分段常数强度Markov模型对疾病转归过程的相关影响因素及其变化规律的分析更为全面。应根据收集数据的实际情况,选择合适的模型进行分析。
[Abstract]:Objective to introduce the piecewise constant intensity Markov model into the process of mild cognitive impairment (MCI) to (Alzheimer's disease, AD, and to study the factors influencing the outcome. To provide theoretical basis for formulating preventive measures at different stages of development. From the point of view of application, the research of piecewise constant intensity Markov model in chronic diseases is explored, which provides methodological reference for the discussion of the influencing factors of other multi-state chronic diseases in different stages of development. Methods six hundred elderly people in Taiyuan were followed up for 4 times. MCI was taken as state 1, moderate and severe cognitive impairment as state 2 and AD as state 3, and the Markov model of piecewise constant strength was fitted to analyze the influencing factors of MCI to AD in different stages of development. Results after hypothesis test, the data satisfied Markov property (P0. 89) and unsatisfied time homogeneity (P0. 001), which can be fitted by piecewise constant strength Markov model. Using a piecewise constant intensity Markov model, for the first time interval of [0 ~ 12] months, we found that sex, age, education, occupation and hypertension had statistical significance in the transfer of state 1 to state 2. Occupation, hypertension and diabetes had statistical significance in the metastasis of state 1 to state 3, sex, age, education, and hypertension and diabetes had statistical significance in the metastasis of state 2 to state 3. For the second time interval [12 ~ 18] months, sex, age, education, occupation, smoking and hypertension were found to be statistically significant in the transfer of state 1 to state 2, education level, occupation, smoking. Hypertension and diabetes were statistically significant for the metastasis from state 1 to state 3, and age, education, occupation, smoking, hypertension and diabetes were statistically significant for the metastasis from state 2 to state 3. Therefore, women, age, smoking, hypertension and diabetes are risk factors for MCI to AD, high education and mental labor are the protective factors of MCI to AD. The survival curve is made according to the model and the transfer intensity matrix and the 3-year transition probability matrix are calculated. Conclusion in order to delay the progression of Alzheimer's disease, the prevention and treatment of the disease should be carried out in different stages according to the influencing factors of different stages of disease metastasis. Compared with the homogeneous Markov model, the piecewise constant intensity Markov model is more comprehensive in the analysis of the related factors and the changing rules of the disease outcome process. According to the actual situation of data collection, the appropriate model should be selected for analysis.
【学位授予单位】：山西医科大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R749.16

【参考文献】