| Summary: | 碩士 === 國立高雄應用科技大學 === 資訊工程系 === 106 === Many medical businesses start with "finding the right patients from medical records". How to efficiently and quickly find patients meeting certain criteria from a large number of medical records is a frequently encountered problem by medical personnel. Electronic medical records (EHR) are an important source of data for medical big data. EHR have sufficient integrity and representativeness for the record of patients' medical history.
However, the use of EHR by medical institutions is often limited to the most original purpose of medical services and fee declarations. It may not be possible to use EHR for secondary use in an efficient manner because of the lack of introduction of relevant information technology, and the value of medical big data cannot be realized.
There are restrictions on the methods available to existing medical institutions to "find the right patients from medical records.” Searching for patients based on a single type of information (such as only based on administrative data, only based on radiology reports) may not accurately identify patients who meet the conditions. The performance and scalability of multi-condition query, which queries multiple type of medical record at the same time, is a challenge in the context of medical big data
Big-data technology, which has received rapid attention in recent years, is being gradually adopted in the medical and biomedical fields and has achieved certain results, due to characteristics such as easier horizontal expansion which is suitable for processing medical big data.
This study proposes two different versions of system to solve the problem of multi-condition query: one system applies big-data technology such as NoSQL and Apache Spark, and another one applies traditional relational database management system. The study then compare the performance between the two to determine whether big-data technology is more suitable for processing multi-condition queries and explore its future feasibility in other medical businesses.
|
|---|
