台灣小區域民眾就醫經驗之分析

Patient experiences in small area of Taiwan

目標：本研究於台灣就醫權益調查之下的民眾就醫經驗進行小區域估計（Small Area Estimation; SAE）；並比較95小區域分類與醫療網50次區域之差異，強化在資料既有框架下改善其細微度且不需額外耗費其他人力及資源之推估方法。方法：研究係以健保民眾就醫權益調查之跨總額聯合資料為基礎，各分項調查樣本數均達1,030份以上（抽樣誤差為±3.05%），總有效回收樣本達5,152份。以等比機率抽樣（Probability Predictionate to Size Sampling, PPS）為基礎下向下切分至鄉鎮市區單位之區位內人口數、輔以性別、年齡與教育程度等三因子進行數值加權。爾後進一步透過多層次變異數成份估計組內相關係數，衡量組間變異數與組內變異數的相對程度。結果：95小區域分類與醫療網50次區域有四項民眾就醫經驗存在有顯著差異性，依序為『就診時醫師是否與您共同討論照護或治療方式？』、『就診時醫師是否有提供您問問題或表達顧慮的機會？』、『請問過去一年來，那家醫療院所的醫護人員有沒有為您進行衛生教育指導（例如：解說病情、營養及飲食諮詢、預防保健方法等）？』、『過去一年內，請問那家醫療院所的醫護人員，是否常用您簡單易懂的方式來向您解說病情與照護方法？』。若以醫療網50次區域為估計基礎，可能會因此低估了各區塊中不能忽視的小區域差異。結論：本研究於健保就醫權益調查的框架下，提出95個小區域的分類，可做為未來概觀式大型調查於小區域推估與城鄉差距比較之應用方法參考，進而掌握小區域間之差異性與重要趨勢。

Objectives: This study conducted small area estimation on health care experiences in Taiwan, with a focus on health care rights. The study compared 95 small area classifications and 50 medical areas in terms of the health care experiences of their residents. This study’s leveraged an existing data framework, and the proposed method does not require additional resources. Methods: The study used a cross-sectional survey conducted nationwide in 2019 to evaluate public opinions on the National Health Insurance program in Taiwan. Each subgroup survey had a sample size of at least 1,030 respondents (with a sampling error of±3.05%), resulting in a total of 5,152 valid responses. Proportional probability sampling was used to divide the population into district-level units within townships and municipalities. This division was further refined by factors including gender, age, and education level, and numerical weighting was applied. Multilevel variance component estimates were then used to measure the relative levels of between-group and within-group variances. Results: Health-care experiences significantly differed between the 95 small area classifications and the 50 medical areas. These differences pertained to whether the doctor discussed care or treatment options with the patient during the visit, whether the doctor provided an opportunity for the patient to ask questions or express concerns, whether health care personnel at the medical institution provided hygiene education, and whether health care personnel at the medical institution used accessible language to explain the patient’s medical condition and the available methods of care. Significant local differences within each block may be underestimated if only the 50 medical areas are used as a basis for estimation. Conclusions: This study proposed a classification system of 95 small areas that can serve as reference points for future large-scale surveys for fine-grained estimations and comparisons of regional disparities.