應用資料倉儲系統於台灣防疫作為之支援——以支援 H1N1新型流感防治為例

Application of data warehouse system for communicable disease control in Taiwan – A case of novel influenza A (H1N1)

過去十年以來，台灣疾病管制局不斷的建置各種通報系統，使得疾管局對於傳染病感測的觸角，能延伸到更多角落，以儘可能涵蓋所有可能發生傳染病的地方。然而這些作業導向、以線上交易功能為主的通報系統，在台灣 2003 年 SARS 風暴中，面臨通報資料難以整合與共享，致使指揮官需要資訊時，需要花費許多時間經過層層轉送才能得到，不具時效性。另外，因為疫情通報資料有時間差，當不同時間向承辦人索取資料時，可能會得到不盡相同之數據，進而發生許多紛擾。 為達成資料整合共享，本倉儲系統使用了 Informatica、Oracle、Business Objects 等 ETL、DB 與 OLAP技術，建立資料倉儲，藉由整合局內各個分散系統，並搭配疫情啟動時之指揮中心架構，建立傳染病倉儲網頁查詢系統、線上分析系統、流行疫情決策支援系統，以即時支援指揮中心進駐人員不同類型的資訊需求。 本研究即以疾管局因應 2009 年 H1N1 新型流感防制為例，探討資料倉儲導入後，對於台灣中央流行疫情指揮中心進駐人員在防疫作為上之資訊需求支援，並評估倉儲導入後，此新的資訊架構，能否改善通報資料難以整合與共享的問題，達成防疫決策所需之ㄧ致性與即時性資訊。

Over the past decade, the Taiwan Centers for Disease Control (TCDC) has built a variety of disease monitoring systems, making TCDC capable for expanding the infectious surveillance network toward all possible corners. However, these monitoring systems, established mainly with the mission of online transaction or disease reporting, were confronted with difficulties in information integration and sharing amongst different systems during the 2003 Taiwan SARS outbreak. In this regard, the commanders acquired the epidemic information through several levels of information transfer and not in a timely manner. In addition, the information received from different epidemic analysts may be different due to different end time of data obtained, which was confusing. To achieve the goal of information integration and sharing, TCDC has built an epidemic data warehousing system by using ETL, DB and OLAP tools from companies like Informatica, Oracle, and Business Objects. With the advantage of the data warehousing system, data from individual surveillance systems are allowed to be shared and utilized by integrated information systems, which support staffs working in the epidemic command center in real-time, such as Infectious disease web-based querying and reporting system. This study was to demonstrate the role of TCDC’s epidemic data warehousing system in providing various supports to staffs working in the Central Epidemic Command Center for the 2009 Novel influenza A/H1N1 outbreak. This new information infrastructure was also evaluated for the proposed purpose of improving surveillance data integration and sharing as to provide concordant and timely information for epidemic decision support.