視覺圖評估醫院病人的用藥:抗血壓重複率高於抗血糖和抗血脂

透過健康保險資料開放服務平台，只見數據卻缺乏視覺圖的分析與比較。本研究評估台灣醫院三高高血壓、高血糖、和高血脂病人的用藥重複天數率：驗證假設三特定類別間是否一致：（1）三高間；（2）三醫院層級醫學中心、區域醫院、和地區醫院間；及（3）三高在其用藥重複率類別與醫院層級間的家數分佈。利用視覺圖分析台灣422家醫院的三高用藥重複率，包括19家醫學中心、87家區域醫院、和316家地區醫院，分析最近12季（2022Q4至2023Q3）的處方重複（天數）率，包括羅序人題變數圖、資料趨勢熱點圖、和4象限散佈圖。結果發現：（1）抗血壓的用藥重複天數率高於抗血糖和抗血脂；（2）醫院層級間的用藥重複（天數）率呈顯著差異，地區醫院的比率最低；及（3）用藥重複率類別與醫院層級間的家數分佈，只有抗血壓和抗血脂呈顯著差異。視覺圖幫助分析醫院開立三高病人的用藥重複趨勢和異常，提出追蹤改善的特定醫院。此外，也探討三高用藥重複（天數）率對病人治療結果的潛在影響，包括不良反應和增加的醫療成本。建議強化處方箋的開立，提高病人安全，並有效管理慢性病的資源。本研究示範視覺圖的分析與比較，對提高健康保險資料開放服務平台的利用，提供參考的依據。

Through the Health Insurance Data Open Service Platform (HIDOSP), data is available for public access, yet it lacks comprehensive visual information. This study assesses therapeutic duplication (TD) in Taiwan hospitals among patients with hypertension, and compares it to duplication in treatments for , diabetes, and hyperlipidemia. Our hypotheses tested whether TD rates were consistent across three specific categories: (1) the targeted treatment areas of hypertension, diabetes, and lipid disorders (the ''3 heights''); (2) three levels of hospital facilities (medical centers, regional hospitals, and local hospitals); and (3) observed patterns of frequency across these 3 heights and levels. We utilized advanced visual health mapping tools to analyze TD rates from 422 hospitals across Taiwan, which included 19 medical centers, 87 regional hospitals, and 316 local hospitals, focusing on prescription data from the last 12 seasons (2022Q4 to 2023Q3). Three types of visualizations were used to test the hypotheses: Rasch Wright Maps, data-trend heatmaps, and 4-quadrant scatter plots. The findings indicated: (1) higher TD rates in the treatment of hypertension compared to diabetes and lipid disorders; (2) significant differences in TD rates among hospital levels, with the lowest rates at local hospitals; and (3) distinct patterns of frequency across the 3 heights and levels. The visual tools helped identify specific trends and anomalies in prescription practices across hospitals, pinpointing critical areas for intervention. Additionally, the study addressed the potential impact of TD on patient outcomes, including adverse effects and increased healthcare costs. Recommendations are provided to enhance prescribing practices, improve patient safety, and optimize resource use in managing chronic conditions. This research highlights the critical role of technological aids, such as health mapping tools, in enhancing the quality of visual information provided by HIDOSP for future improvements.