「國民健康訪問調查」之調查設計、內容、執行方式與樣本人口特性

Abstract

目標：「國民健康訪問調查」是跨機構的合作計畫，目的在建立一機制，定期調查國民健康狀況及健康相關之需求，提供決策者制訂政策之依據。方法：本調查由國家衛生研究院醫療保健政策研究組統籌規劃，國民健康局衛生教育中心負責問卷設計，人口與健康調查研究中心負責訪視調查之執行，國家衛生研究院生物統計與生物資訊研究組負責統籌抽樣設計和資料庫之建立。本計畫採用多段分層系統抽樣設計(multistage stratified systematic sampling design)，將臺灣地區359個鄉鎮市區依地理位置和都市生活圈分為七大層，每層內採用抽取率與單位大小成比例方式(Probability Proportional to Size, PPS)抽出鄉鎮市區，被抽到的鄉鎮市區內再抽出鄰，最後每鄰抽出四戶。訪問內容以影響健康的因素為基礎，如個人、社會和物質環境及醫療政策等。結果：臺灣地區共抽出6,592戶(26,685人)。山地離島、地區因人口少，被抽到的機率小，另行加重抽樣，山地地區抽出608戶(2,797人)，離島地區抽出432戶(1,954人)。問卷的設計在考慮健康指標的需求與面訪的特性後，設計成五種問卷包括有家戶問卷、12歲以上個人問卷、12歲以下個人問卷、12∼19歲青少年自填問卷、20∼65歲自填生活品質問卷。結論：問卷經過專家效度檢定；抽出之樣本經過整理後，臺灣地區應訪戶有6,364戶，其中5,798戶完訪，完訪率為91.1％，完訪戶中之實住人口有23,473人，其中22,121人(94.2％)完訪。收到的所有問卷資料均經過嚴格的品質管制，並建立於資料庫中，可供研究人員使用，並可提供政府衛生相關機構制訂政策之重要參考。

Objectives: NHIS is a survey to understand the general health of the resident civilian non-institutionalized population in Taiwan. Once the regular survey system is well established, it will provide information in making timely health related policies. Method: This survey incorporated a multi-stage stratified systematic sampling scheme. It first divided 359 townships/districts of Taiwan into 7 strata according to their geographical location and degree of urbanization. Townships or districts in each stratum were selected with selection probability proportional to their sizes (PPS). In each selected township/district, lins (the smallest administrative unit) were selected with PPS. Four households were selected randomly from each selected lin. Each member in the selected household was to be interviewed. Results: Altogether, 6,592 households (26,658 persons) were sampled from the whole Taiwan area. Due to the small population sizes of mountainous areas and off-shore islands, these areas were oversampled. In mountainous areas, another 608 households (2,797persons) were sampled. In off-shore islands, 432 households (1,954 persons) were sampled. This resulted in equal probability samples. Development of questionnaires was based on the DOH needs of health indicators and the nature and nurture factors affecting health. The final version of questionnaire was composed of 5 parts: questionnaire for household, for individuals older than 12 years, for individuals younger than 12 years, self-administered questionnaire for teenagers between 12 and 19 years, WHO quality of life for adults aged between 20 and 65. Conclusion: This survey was well designed. The demographic characteristics of the samples were consistent with the population. The response rate was 91.1% for households and 94.2% for individuals. The information gathered by this survey was very valuable in designing health related programs and policies.