论文部分内容阅读
目的:黄疸待查是一个常见而复杂的临床问题,涉及到内、外、妇、儿等多个学科。目前我国医学专家存在数量相对不足,分布不均匀等情况,导致了区域性和部门性医疗服务水平不足。本研究旨在建立一个客观的黄疸待查智能诊断系统,以提高医学诊断的正确性,提升基层医院及急诊的诊断水平,同时减少病人的花费。创新点:本研究采用了国际先进的动态不确定性因果图(DUCG)模型,建立了黄疸待查相关疾病的知识库,通过203例临床病例的测试,其准确率达99.01%。文章以图形化的方式给出了疾病的诊断过程,方便医师理解和学习。方法:本研究采用了DUCG模型进行疾病诊断,首先根据DUCG模型的定义和黄疸诊断思路建立了包含27种黄疸相关疾病(表4)的知识库(图2),其中包括了疾病的危险因素、临床症状和体征、客观检查检验结果等。然后与根据DUCG算法(公式1–4)编写的推理软件相结合形成诊断系统,对203例临床黄疸患者进行智能诊断,准确率达99.01%。最后对一例丙型病毒性肝炎患者的具体诊断过程进行了拆解阐述,体现了DUCG模型适用于复杂逻辑关系、计算效率高、不依赖推理概率和结果易于理解等优点。结论:DUCG模型成功实现了对黄疸待查相关疾病的智能诊断,准确率高,实用性好。该方法具有在其他医学领域推广应用的价值。
Objective: Jaundice pending investigation is a common and complicated clinical problem involving many subjects, including internal and external, women and children. At present, the relative shortage of medical experts in our country and their uneven distribution have led to the lack of regional and departmental medical services. The purpose of this study is to establish an objective diagnostic system for jaundice to be diagnosed to improve the accuracy of medical diagnosis and to improve the diagnosis of primary hospitals and emergency departments and to reduce the patient’s costs. Innovative point: In this study, the international advanced dynamic uncertainty causality diagram (DUCG) model was used to establish a knowledge base of jaundice related diseases. The accuracy rate of 99 cases was tested by 203 clinical cases. The article gives a graphical way of the diagnosis of the disease process, to facilitate understanding and learning of the doctor. Methods: This study used DUCG model for disease diagnosis. Firstly, according to the definition of DUCG model and the diagnosis of jaundice, a knowledge base of 27 jaundice related diseases (Table 4) was established (Figure 2), including the risk factors of the disease, Clinical symptoms and signs, objective test results and so on. Then a diagnostic system based on inference software based on DUCG algorithm (Formula 1-4) was formed to diagnose 203 cases of clinical jaundice with an accuracy rate of 99.01%. Finally, a case of hepatitis C patients with a specific diagnosis of dismantling described, embodies the DUCG model for complex logic, high computational efficiency, independent of the probability of reasoning and the results easy to understand and so on. Conclusion: The DUCG model has successfully realized the intelligent diagnosis of jaundice related diseases, with high accuracy and good practicability. This method has the value of popularization and application in other medical fields.