台灣擁有著名的全民健保計畫，並且幾乎涵蓋了所有的台灣公民。不同於其他保險公司，醫院的全民健保病歷資料儲存了所有不論是富有、貧窮、年輕、老年病人的ICD9 疾病代碼。本研究針對地區醫院的病歷資料進行分析，並且整合當地的天氣資訊。地區醫院具有特定看診人口、天氣型態的區域特性，因此本研究結果更能體現地區醫院所在之地域的獨特疾病型態。透過分析各天氣條件下看診人數的變化趨勢，我們可以瞭解疾病好發於何種氣候狀況，並且探索新的、潛在的季節性疾病。許多疾病彼此之間可能是相互關聯的，而這樣的關聯性可以透過分析臨床資料來取得，並且透過結合所有可能關聯的疾病來衍生出疾病網路。然而可能有許多因子可能涉及並影響疾病間的關聯性，藥物就是其中一個重要的因素。本研究分析開立藥物與疾病間的影響性，並結合至疾病網路，建立藥物-疾病網路。我們的研究協助醫生在進行臨床診斷時根據環境與藥物等因子來決定治療方針，並進行後續風險評估。;Taiwan has a famous National Health Insurance (NHI) program which covers almost all citizens. Unlike other data from insurance company, the NHI diagnosis record of hospitals store International Statistical Classification of Diseases and Related Health Problems version 9 (ICD9) codes of every patients no matter rich, poor, young, or old, which form a unbiased sampling of disease occurrences. This study analyzed NHI medical record of local hospital and integrated with local weather condition information. Local hospital had a characteristic with specific population of admission and weather conditions, our result represent unique patterns of diseases occurrence on local hospital region. The seasonal predilection of diseases could be discovered through analyzing the variation of number of admission on different weather conditions, also investigated potential new seasonal diseases. The diseases network could be derived from analyzing the relevance between diseases by clinical data. However, several factors may involve the relevance of diseases, and drug is one of important factor. This study further analyzed prescribing drug effect relevance of diseases, and combined into disease network to construct a drugs-disorders network. Our research assisted doctors to make decision on treatment strategy and evaluate risk when clinical diagnosis according to environment and prescribing drug factors.
