43卷1期

• Link 政策論壇 Policy Forum
• 由生命末期照護需求看病人自主權利法Examining Patient Right to Autonomy Act from the perspective of end-of-life care needs
• 施至遠、程劭儀
  Chih-Yuan Shih, Shao-Yi Cheng
• 病人自主權利法、生命末期照護、預立醫療照護諮商
  the Patient Right to Autonomy Act, end-of-life care, advance care planning
• 根據推估，未來25年臺灣的每年死亡人數將從19萬上升到30萬，使得生命末期照護需求日益增加。病人自主權利法跟安寧緩和醫療條例的主要差異反映在於適用對象、意願人表達治療抉擇的項目，以及需經過預立醫療照護諮商的過程，才能完成預立醫療決定的文件，且諮商機構得酌收諮商費用。目前衛福部擬透過給付來減少諮商費用造成的善終權不平等，然而除了給付，照護歷程的各部分都需被重視，例如將advance care planning（ACP）納入個案管理照護流程、調整諮商團隊的人數與職類要求。未來仍需法令持續的調整與政策的支持，讓我國死亡照護品質持續精進。
  It is estimated that the number of deaths in Taiwan annually will rise from 190,000 to 300,000, which will increase the need for end-of-life care. The main difference between the Patient Right to Autonomy Act and the Hospice Palliative Care Act are the people they apply, the items for declarants to express their decisions, and the process of making advance decisions. According to the Patient Right to Autonomy Act, advance care planning (ACP) is necessary before making advance decisions (AD). The Ministry of Health and Welfare plans to include ACP in National Health Insurance (NHI) coverage to reduce the inequity of end-of-life rights caused by consultation fees. In addition to the inclusion of NHI coverage, it is crucial to emphasize every part in the process of care, such as incorporating ACP into the case management or adjusting the quorum of ACP teams and the qualification of members. Laws and supportive policies should be dynamically adjusted to improve the quality of end-of-life care.
• 1-3
• 10.6288/TJPH.202402_43(1).PF01

• Link 公衛論壇 Public Health Forum
• 建構「資訊驅動的病人為中心照護」：一個台灣新模式的倡議Building information-driven patient-centered care: an initiative of a new Taiwanese mode
• 鄭守夏、莊智鈞、林青青
  Shou-Hsia Cheng, Chih-Chun Chuang, Ching-Ching Claire Lin
• 無none
  無none
• 無none
  無none
• 4-8
• 10.6288/TJPH.202402_43(1).112114

• Link 原著 Original Article
• 探索口腔健康與骨密度關係—環口X光片在成人的應用Exploring the relationship between oral health and bone density: utilizing panoramic radiography in adults
• 鍾秉宸、詹大千
  Ping-Chen Chung, Ta-Chien Chan
• 下顎下緣皮質骨厚度、下顎皮質指數、骨密度、環口X光片，咀嚼力檢測
  mandibular cortical thickness, mandibular cortical index, BMD, panoramic radiography, masticatory test
• 目標：骨質疏鬆症患者易有其他併發症，需在骨質缺乏初期即早介入，預防疾病發展。本研究目標使用牙科常規檢查常用的環口X光片，探討其與骨密度之間的關係，可做為簡易的骨密度篩檢工具。方法：本研究為橫斷性研究，於衛生福利部朴子醫院牙科門診，招募35歲至80歲的成人參加，收集受試者病史、股骨頸骨密度、環口X光片、咀嚼力檢測、台灣版老人口腔健康評估（GOHAI）量表、牙菌斑指數、身高、體重、抽血的血鈣資料和24小時飲食回憶紀錄，使用線性、邏輯斯迴歸進行統計分析。結果：226位參與者平均年齡65.6歲，其中女性佔58.4%，男性、咀嚼力好、GOHAI分數高與下顎下緣皮質骨厚度有顯著正相關；年輕、GOHAI分數高與下顎皮質指數C3有顯著負相關；在校正心臟疾病和骨折病史後，年輕、下顎下緣皮質骨厚度較厚、下顎皮質指數C1、身體質量指數較高、血鈣濃度較高者，與股骨頸骨密度有顯著正相關。結論：利用可近性高、省時簡便的方法，於一般牙科檢查時，分析環口X光片，可即時發現骨密度低下的可能，讓民眾能提高警覺，再進一步至醫療院所進行更精密的檢查。
  Objectives: Early detection of bone loss is an effective strategy for preventing osteoporosis. In this study, we used routine dental panoramic radiography to examine the relationship between oral health and bone mineral density (BMD) and ultimately establish panoramic radiography as a useful screening tool for BMD. Methods: A total of 226 patients (35–80 years of age, mean age: 65.6 years) receiving treatment at Puzi Hospital were included in this cross-sectional study. Data were collected on the patients’ medical history, BMD of the femoral neck, panoramic radiography results, masticatory test results, Geriatric Oral Health Assessment Index (GOHAI) scores, plaque index, body mass index (BMI), serum calcium levels, and 24-hour dietary recall. Linear regression, and logistic regression analyses were conducted. Results: Being male and having a high GOHAI score were significantly and positively associated with mandibular cortical thickness but significantly negatively associated with mandibular cortical index C3. After adjusting for heart disease and bone fracture history, young adult, wider mandibular cortical thickness and index C1, high BMI, and high serum calcium were all found to be significantly positively associated with BMD of the femoral neck. Conclusions: Panoramic radiography is highly accessible and convenient and enables the timely detection of potential bone loss, thereby encouraging individuals to become more vigilant and willing to undergo more detailed examinations.
• 9-20
• 10.6288/TJPH.202402_43(1).112102

• Link 原著 Original Article
• 十大死因排序：兩種分類表比較Ranks of the 10 leading causes of death: a comparison of two lists
• 韓伊晴、雷子瑩、戴書郁、簡玉雯、呂宗學
  Yi-Ching Han, Tsu-Ying Lei, Shu-Yu Tai, Yu-Wen Chien, Tsung-Hsueh Lu
• 十大死因、死因統計、可避免死因、死亡數、優先順序決定
  ten leading causes of death, causes of death statistics, avoidable mortality, number of deaths, priority setting
• 目標：比較政府傳統分類表（以下簡稱傳統表）與世界衛生組織2006年修訂分類表（以下簡稱世衛表）呈現台灣三個年代性別年齡別十大死因排序的差異。方法：本研究由死因統計資料集擷取2011，2016與2021年的死因別死亡數依照兩個分類表產出性別年齡別十大死因排序。結果：依照傳統表，除了85歲及以上外，每一年代每一個年齡層的第一大死因都是惡性腫瘤。但是，依照世衛表在2011，2016與2021年第一名死因：在15-44歲分別是交通事故、自殺與自殺；在45-64歲分別是肝癌、肝癌與肺癌，65-84歲分別是腦血管疾病、腦血管疾病與糖尿病，85歲及以上在三個年代都是流感與肺炎。在15-44歲，肝癌的名次由2011的第四名，降到2016年的第七名，到2021年跳到十大死因外。結論：同時使用兩種分類表產出十大死因排序資訊，可以提供更多預防與管制政策擬定優先順序資訊，讓衛生政策決策者做出更佳決策。
  Objectives: To compare the ranks of the 10 leading causes of death (CODs) between a traditional government list (traditional list) and the 2006 World Health Organization (WHO) list. Methods: From open data sets, COD data were collected for the years 2011, 2016, and 2021 to estimate the number of deaths for each ranking category. Subgroup analyses by age and sex were performed. Results: According to the traditional list, the leading COD was malignant neoplasm in individuals of all ages except for those aged ≥85 years. However, according to the WHO list, the leading CODs in 2011, 2016, and 2021 were, respectively, transport injury, suicide, and suicide in individuals aged 15–44 years; liver cancer, liver cancer, and lung cancer in those aged 45–64 years; stroke, stroke, and diabetes in those aged 65–84 years; and influenza/pneumonia, influenza/pneumonia, and influenza/pneumonia in those aged ≥85 years. Among individuals aged 15–44 years, liver cancer was the fourth most prevalent COD in 2011; however, liver cancer obtained was ranked seventh in 2016 and was absent from the list in 2021. Conclusions: Insights into the top 10 CODs from the aforementioned two lists may facilitate mortality prevention and health-related policymaking.
• 21-31
• 10.6288/TJPH.202402_43(1).112096

• Link 原著 Original Article
• 以項目反應理論協助發展簡版慢性疾病生活品質量表Developing an abbreviated WHOQOL-BREF for patients with major chronic diseases with the assistance of Item Response Theory (IRT)
• 冷芷涵、姚開屏
  Chih-Han Leng, Grace Yao
• 台灣簡明版世界衛生組織生活品質問卷、試題反應理論、慢性疾病、量表縮減、生活品質
  WHOQOL-BREF, item response theory, chronic disease, shortened scale, quality of life
• 目標：為減輕慢性病患者在門診、急診時，以及後續定期追蹤時的評估時間及負擔，本文針對六種重大慢性病患者，除考量傳統計量方法，並藉由項目反應理論（IRT）的輔助來開發簡版慢性病患WHOQOL-BREF問卷，以協助評估該些族群健康相關生活品質。方法：本文的發展目標為將有四個範疇26個題目的WHOQOL-BREF問卷，縮減為每個範疇有三個題目的結構。研究數據來自國民健康署與國家衛生研究院的國民健康訪問調查（National Health Interview Survey, NHIS）。在NHIS中，13,008名年齡在20歲至65歲之間的參與者完整填答了WHOQOL-BREF問卷。我們使用了其中1,263名患有經過醫生診斷的主要慢性疾病的個案（594名女性）和7,657名健康人（3,644名女性）來發展此簡版慢性病問卷。在問卷發展階段，我們使用了四個標準：校正後題目與總分相關係數（corrected item-total correlation）、預測各整體範疇的反向迴歸模型（backward regression model）、內外適配度均方誤差指標(outfit and infit mean squares）、以及項目訊息量(item information)。在驗證階段，量表信度以受測者信度評估，並估計同時效度和區別效度。結果：研究結果顯示本研究所開發的量表具有良好的受測者信度並與相關聯的量表存在可接受的同時效度。此外，本量表在驗證上能區分健康個體和慢性病患者在每個領域的得分。結論：本文所開發的 WHOQOL-BREF 慢性疾病版本是可靠且有效的。
  Objectives: To lower the burden and shorten the evaluation time of patients with chronic diseases before outpatient clinic appointments, in the emergency room, and during regular followups, a short version of the WHOQOL-BREF assessment was developed for patients with major chronic diseases to assess health-related quality of life with classical test theory methods and item response theory (IRT) analysis. Methods: The proposed assessment has four domains, with three items in each domain. This version was developed based on data from the National Health Interview Survey (NHIS) in Taiwan. In the NHIS, 13,008 Taiwanese participants between the ages of 20 and 65 completed the WHOQOL-BREF assessment. Among the participants, 1,263 individuals (594 females) with major chronic diseases and 7,657 healthy individuals (3,644 females) were included in the development and validation processes. In the development stage, four criteria were used: the corrected item-total correlation, backward regression models of the general domains, the outfit and infit mean squares, and the item information. In the validation stage, the reliability was measured according to the person reliability, and the validity was estimated based on the concurrent and known-groups validities. Results: The findings indicated good person reliability and acceptable concurrent validity with related scales. Additionally, the new assessment distinguished scores between healthy individuals and those with chronic diseases in each domain. Conclusions: The proposed short version of the WHOQOL-BREF was reliable and valid. Therefore, this new form is recommended for use in clinical settings.
• 32-45
• 10.6288/TJPH.202402_43(1).112130

• Link 原著 Original Article
• 醫院慢性病患能否成功下轉至基層診所之相關因素—以糖尿病為例Factors influencing downward referrals for patients with diabetes
• 蘇靖惠、郭年真
  Jing-Hui Su, Raymond N. Kuo
• 分級醫療、轉診、慢性病、醫療品質、基層醫療
  hierarchical healthcare, referral, chronic disease, healthcare quality, primary care
• 目標：過去有關轉診之研究多針對個案醫院或特定區域之轉診進行探討，缺少關於轉診的全國性研究。本研究使用健保資料庫分析分級醫療政策中，雙向轉診的「下轉」案件，探討醫療院所哪些特質與病患下轉後就醫與否及是否在下轉診所持續就醫（成功下轉）有關。方法：使用2018-2020年之全民健康保險資料，篩選出由醫院向下轉診至基層診所且主診斷為糖尿病之門診病患，根據後續就醫情形再區分為成功轉診組與未成功轉診組。本研究計算診所之醫療品質指標，分析醫療院所基本特性與醫療品質，與病患是否轉診就醫有關。因健保資料庫所能提供資訊有限，當病患被轉診後未至原定之轉診院所就醫，則無法得知原定的院所，因此本研究將分兩階段進行分析與探討。首先探討病患被下轉後，是否前往該下轉診所就醫的相關因素；第二階段則針對下轉至基層診所、且有就醫的病患進行分析。結果：病患過去曾至下轉診所就醫（OR=3.451）與成功下轉達統計上顯著差異，病患轉診前醫院層級為區域醫院及地區醫院，與成功下轉呈現負相關（區域醫院：OR=0.692、地區醫院：OR=0.622）。病患下轉至「糖化血紅素」（OR: 1.523）、「血清肌酸酐」（OR: 1.359）與「血清麩胺酸丙酮酸轉胺基脢」（OR: 1.301）檢查比例高的診所，與成功下轉呈現正相關。結論：病患過去就醫經驗、轉診前醫院層級以及醫療品質與病患是否成功下轉有關，建議衛生主管機關針對提升下轉診所之醫療品質擬定轉診策略，以促進病患成功下轉的機會。
  Objectives: Most studies regarding patient referrals in Taiwan have focused on individual hospitals or specific regions and thus have reported findings with limited generalizability. Consequently, the overall referral process remains unclear. Accordingly, this study analyzed downward referral patterns in the bidirectional referral system under Taiwan’s hierarchical health-care policy. Relevant data were obtained from the National Health Insurance Research Database. We investigated whether patients sought medical care after downward referrals and whether they continued to receive medical care at referred facilities (i.e., successful downward referral). Methods: This study included patients with diabetes who were referred from hospitals to primary care clinics for treatment - a process known as downward referral. The patients were divided into two groups, namely successful and unsuccessful referral groups. Indicators of each clinic’s health-care quality were analyzed to identify the correlation between health-care quality and downward referral. Because of data limitations, we could not identify referred clinics for patients who did not seek medical care despite referral. Thus, a two-part approach was adopted for analysis. First, we explored factors influencing patients’ receipt of medical care at referred clinics. Second, we explored factors influencing continuation of care (at fewest two outpatient visits) at referred clinics. Results: Successful downward referral was significantly likely for patients who had previously visited a referred clinic (odds ratio: 3.451). However, the likelihood of successful downward referral was low for patients referred from regional or community hospitals (regional vs. community hospitals, odds ratios: 0.692 vs. 0.622, respectively). By contrast, this likelihood was high for patients who were referred to clinics with high proportions of hemoglobin A1C, creatinine, and alanine aminotransferase tests. Conclusions: Successful downward referral is associated with prior health-care experiences of patients, levels of clinics visited before referral, and quality of health care. Accordingly, health authorities are recommended to implement continual monitoring of referrals, develop referral strategies for health-care institutions of all levels, and enhance the quality of health care to improve the likelihood of successful downward referral.
• 46-56
• 10.6288/TJPH.202402_43(1).112063

• Link 原著 Original Article
• 新冠肺炎大流行期間「不同時期」對國中生身體活動及睡眠之影響Variations in physical activity and sleep patterns among junior high students across COVID-19 pandemic phases
• 劉蕙怡、王振興、楊雅婷
  Huey-Yi Liou, Jeen-Shing Wang, Ya-Ting Carolyn Yang
• 新冠肺炎、青少年、身體活動、睡眠、智慧手環
  COVID-19, adolescents, physical activity, sleep, smart fitness trackers
• 目標：本研究旨在了解新冠肺炎大流行前後對青少年身體活動及睡眠的影響。方法：本研究採「準實驗研究」之單一組前後測來進行，並以便利抽樣方式，以台灣台南地區某國中八年級56位學生，於2021年4月~9月，以「智慧手環」及「克里夫蘭青少年嗜睡量表（CASQ）」來蒐集身體活動之每日步數、輕度、中度活動、睡眠時間，及青少年嗜睡情形之數據，並以描述性統計、相依樣本單因子變異數分析國二下學期實體上課、線上上課初期、期末，國二升國三暑假及國三上學期實體上課，共五時期之身體活動及睡眠時間之情形，及「克里夫蘭青少年嗜睡量表（CASQ）」在國二下實體上課、線上上課後，及國三上實體上課，共三時期之變化情形。結果：身體活動不論是每日步數和輕度與中度活動時間都是實體上課較線上上課（初期與期末）和暑假較多。反觀睡眠時間則是於國二下學期實體上課期間為354分鐘最少，線上上課初期為395分鐘，線上上課期末為409分鐘，一直延續到暑假已增加為428分鐘，且到國三上學期實體上課仍持續增加為432分鐘。結論：依研究結果及文獻綜論提出之具體建議如下：一、身體活動方面：實行短時間多回合「低強度身體活動」，如以站立、步行等活動來取代長時間的靜態行為。2.提供線上跨領域雲端課程（家政、童軍、體育、音樂）等學習任務，如完成規定之跑走步數、親子運動活動來養成運動習慣；二、睡眠方面：可於彈性課程中實施睡眠教育，並實施睡眠問卷來了解學生睡眠之實際情形。未來相關單位可利用本研究所提出具體的因應措施及建議，在寒、暑假、或下一波疾病大流行期間，來避免身體活動量不足及改善睡眠問題的情形。
  Objectives: This study was conducted to identify the physical activity and sleep patterns of adolescents in southern Taiwan during different phases of the COVID-19 pandemic. Methods: For this study, we adopted a quasi-experimental design with a single-group pretest–posttest approach. Through convenience sampling, we enrolled 56 junior high school (eighth grade) students from Tainan, Taiwan. Data (April to September 2021) on daily steps, low-intensity and moderate-intensity exercise levels, and sleep duration were collected using smart fitness trackers and the Cleveland Adolescent Sleepiness Questionnaire. The students’ physical activity levels and sleep durations were measured at five time points: during the second semester of the eighth grade with in-person classes, online classes, and summer vacation and the first semester of the ninth grade with in-person classes. Changes in their Cleveland Adolescent Sleepiness Questionnaire scores were assessed at three time points: during the second semester of the eighth grade with in-person classes, after transitioning to online classes, and during the first semester of the ninth grade with in-person classes. Data were analyzed using descriptive statistics and the repeated measures analysis of variance test. Results: Daily steps, low-intensity exercise levels, and moderate-intensity exercise levels were higher during in-person classes than during online classes and summer vacation. Sleep duration was the shortest during the second semester of the eighth grade (354 min); however, it increased to 395 min at the beginning of online classes, 409 min at the end of online classes, 428 min during summer vacation, and 431 min during the first semester of the ninth grade (in-person classes). Conclusions: On the basis of the findings of this study and a review of the literature, we propose the following recommendations for improving adolescents’ health. Physical activity: (1) introduce short, multiple rounds of low-intensity exercises (e.g., standing and walking) to avoid prolonged sedentary behavior and (2) offer online interdisciplinary courses (e.g., home economics, scouting, physical education, and music) as learning tasks, integrating step goals and parent–child activities, to promote exercise habits. Sleep: (1) integrate sleep education into the curriculum and (2) conduct sleep surveys to understand students’ sleep patterns. These measures may help address adolescents’ inadequate physical activity levels and improve their sleep during school breaks (e.g., holidays and summer vacation) or pandemics.
• 57-68
• 10.6288/TJPH.202402_43(1).112113

• Link 原著 Original Article
• 誰是受害人？台灣數位親密關係暴力的群體特性與暴力特性Who is the victim? Demographic characteristics and violence patterns associated with digital intimate partner violence in Taiwan
• 陳怡青
  Yi-Ching Chen
• 數位、網路、親密關係暴力、科技暴力
  digital, internet, intimate partner violence, technology-facilitated abuse
• 目標：隨著數位工具的進步，親密關係暴力的樣態出現明顯的變化。本研究欲從年齡、性別、性傾向等人口特性來探討台灣數位親密關係暴力的現象，以及數位親密關係暴力的主要類型。方法：本研究透過文獻查證、受害者訪談、專家焦點團體及預試，建立數位親密關係暴力調查問卷。接著針對台灣民眾，在政府與民間的各網頁或社群發放問卷，以網路調查的方式蒐集資料，將所得的問卷，依全國人口年齡、性別比例隨機抽取，最終以2,388位樣本進行統計分析。結果：整體而言，過往曾經歷過數位親密關係暴力者佔14.7%。18至30歲族群發生數位親密關係暴力的比例為24.1%，隨著年齡遞減；男性的受害風險略高於女性，尤其非異性戀男性的風險達顯著差異，為25.3%。最常見的數位親密關係暴力型態為跟蹤肉搜（11.1%），其次是控制（7.2%）。結論：年輕、非異性戀男性群體的數位親密關係暴力風險尤其需要受到關注，在防治工作上，建議從世代、群體網路生活特性來思考，突破實體親密關係暴力思維框架的限制。
  Objectives: With the evolution of digital tools, considerable changes have been noted in the patterns of intimate partner violence. In this study, we explored technology-facilitated intimate partner violence (TFIPV) in Taiwan, considering demographic characteristics such as age, gender, and sexual orientation. In addition, we identified the primary forms of TFIPV. Methods: A TFIPV questionnaire was developed through literature review, victim interviews, expert focus groups, and a pretest. Using this questionnaire, online surveys were conducted among Taiwanese individuals through various governmental/private websites or social media platforms. Respondents were randomly selected on the basis of age and gender distributions in Taiwan. Ultimately, 2,388 respondents were included in the final analysis. Results: Of the respondents, 14.7% reported experiencing TFIPV previously. Among individuals aged 18–30 years, 24.1% experienced TFIPV; this proportion gradually decreased with age. The risk of victimization was marginally higher for men than for women, with a marked disparity noted among nonheterosexual men (25.3%). The predominant form of TFIPV was cyberstalking (11.1%), followed by control (7.2%). Conclusions: Our findings suggest that younger individuals and nonheterosexual men are the primary populations at an elevated risk of TFIPV. This study highlights the need for designing tailored prevention policies considering the Internet usage patterns of these demographic groups. The implementation of such policies may overcome the limitations of traditional face-to-face intimate partner violence frameworks.
• 69-81
• 10.6288/TJPH.202402_43(1).112119

• Link 實務 Public Health Practice
• AI聊天機器人Bing Chat能否準確回答PICO臨床問題？—以術前口服碳水化合物為例Ability of Bing Chat to accurately answer PICO clinical questions: a case study regarding preoperative oral carbohydrate intake
• 汪秀玲
  Hsiu-Ling Wang
• 聊天機器人、人工智慧、ChatGPT、大型語言模型、實證醫學
  chatbot, artificial intelligence (AI), GPT (Generative Pre-trained Transformer), large language model (LLM), evidence-based medicine
• 目標：新一波人工智慧熱潮－GPT（Generative Pre-trained Transformer）是基於大型語言模型（Large Language Model, LLM）的科技，能夠生成不同主題和語境的文本，然而，它在回答醫學問題的表現尚未被充分評估。本研究旨在評估聊天機器人在回答複雜臨床問題時的表現。方法：於2023年6月12日使用微軟開發的Bing Chat作為GPT的代表，它能在Edge瀏覽器和Bing搜尋引擎中使用。我們從考科藍台灣研究中心的網站上隨機抽取了一個臨床問題，即術前口服碳水化合物是否能減少術後不適感？我們要求Bing Chat使用PICO框架（Population、Intervention、Comparison、Outcome）來回答，並提供相關的參考文獻和證據摘要。我們將英文提示輸入到Bing Chat的會話框中，並記錄它的回答。然後，經由與PubMed中的真實文獻進行比對，以審閱證據的完整性、準確性。結果：Bing Chat能夠快速地識別PICO框架的各要素，並根據一些文獻來總結出簡明的答案。然而，Bing Chat提供有限的參考列表（7篇），其中有些文獻的刊名和作者不正確（所謂GPT的幻覺（hallucination）），而且它提供的證據摘要不夠詳盡，沒有涵蓋研究方法和結果數據。Bing Chat在經過多次提示後，能提供更具體的重點數據。結論：Bing Chat在回答醫學問題時有一定的能力，但仍需改進其資料收集和處理的準確度、完整度，醫療提供者在使用它時也應該審慎檢查其提供的信息，並了解其局限性。
  Objectives: The recent surge in enthusiasm for artificial intelligence has revolved around GPT (Generative Pretrained Transformer) technology, a technology that relies on large language models to generate diverse texts across topics and contexts. However, a GPT’s performance in answering clinical questions remains to be comprehensively evaluated. Thus, this study was conducted to evaluate the performance of chatbots in answering complex clinical questions. Methods: In this study, Bing Chat (developed by Microsoft) served as a representative of GPT technology. This chatbot can be used in the Edge browser or Bing search engine. We randomly extracted a clinical question from the website of the Cochrane Taiwan Research Center (regarding whether preoperative oral carbohydrate could reduce postoperative discomfort) and asked Bing Chat to answer it using the PICO (Population, Intervention, Comparison, Outcome) framework and to provide relevant references and evidence summaries. A relevant English prompt was entered into the conversation box of Bing Chat, and its answer was recorded. The completeness and accuracy of the chatbot-provided evidence were analyzed through a comparison with the literature (PubMed). Results: Bing Chat rapidly identified the elements of the PICO framework and provided a concise answer on the basis of the literature. However, it offered a limited number of references (seven articles), some of which had incorrect names of journals and authors (a phenomenon known as GPT hallucination). Furthermore, the chatbot provided insufficient evidence summaries that did not cover research methods or results. Nonetheless, after receiving multiple prompts, Bing Chat provided relatively specific information. Conclusions: Bing Chat can answer medical questions to some extent. However, the accuracy and completeness of its data collection and processing methods require further improvement. Therefore, when using this chatbot, health-care providers are recommended to carefully check the information that it provides and to remain aware of its limitations.
• 82-92
• 10.6288/TJPH.202402_43(1).112087

• Link 實務 Public Health Practice
• 調整後台灣新生兒死亡率：國際與國內縣市比較Adjusted neonatal mortality rates in Taiwan: international and domestic comparisons
• 梁富文、呂宗學、江東亮
  Fu-Wen Liang, Tsung-Hsueh Lu, Tung-liang Chiang
• 新生兒死亡率、早產/死亡率、活產/流行病學、調整死亡率、出生證明、台灣
  neonatal mortality, premature birth/mortality, live birth/epidemiology, adjusted mortality rates, birth certificate, Taiwan
• 目標：由於醫師判定活產行為的差異會影響新生兒死亡率的可比較性，本研究目的是依照世界衛生組織建議，比較去除幾乎不可能存活新生兒前後，台灣與經濟合作發展組織（OECD）國家以及台灣各縣市之間的新生兒死亡率差異。方法：我們連結2018至2021年出生登記檔、出生通報檔與死因統計檔進行死亡率調整。方法一：去除出生體重<500公克或懷孕週數<22週死亡個案；方法二：去除懷孕週數<24週死亡個案。結果：台灣調整前新生兒死亡率（每千活產死亡數）2.52，方法一調整後死亡率降為1.83。與OECD國家比較的百分等級值調整前是60級，屬於中間偏後；調整後是47級，屬於中間偏前，名次改善。方法二調整後新生兒死亡率降為1.49。調整前有四個縣市新生兒死亡率統計檢定顯著高於台灣，方法一調整後，新北市不再顯著高於台灣。方法二調整後，花蓮縣、屏東縣與高雄市都不再顯著高於台灣。結論：幾乎不可能存活新生兒的活產判斷明顯影響新生兒死亡率高低，未來要進行跨國與跨縣市新生兒死亡率比較時，一定要進行調整才能正確解讀數據。
  Objectives: To improve the comparability of neonatal mortality rates (NMRs) across regions, the World Health Organization recommends adjusting for potential artifacts stemming from physicians' judgments of live births for previable newborns. This study compared the NMRs of Taiwanese counties and cities before versus after previable newborns were excluded from the total. Methods: We linked birth registration, birth reporting, and mortality data to obtain birth weight (BW) and gestational age (GA) information for adjustment. Subsequently, we conducted a first round of exclusions of neonatal deaths with BW < 500 g or GA < 22 weeks. After conducting the first round of adjustment, we conducted a second round of exclusions of neonatal deaths with GA < 24 weeks. Results: The NMR (deaths per 1,000 live births) for all of Taiwan before adjustment was 2.52. Following the first round of exclusions, the NMR decreased to 1.83. Taiwan’s percentile ranks relative to OECD countries before and after the first round of adjustment was 66 and 47, respectively, improving from the middle low to middle high tier. Taiwan’s NMR fell to 1.49 after the second round of exclusions. Before adjustment, four cities or counties had NMRs that were significantly higher than Taiwan’s national average; they were New Taipei City, Pintung County, Hualien County, and Kaohsiung City. However, after the first round of adjustments, the NMR of New Taipei City was not significantly higher than Taiwan. After the second round of adjustment, the NMRs of Hualien County, Pintung County, and Kaohsiung City were also not significantly higher than Taiwan. Conclusions: Variations in the classification of previable newborns as live births can substantially alter calculations of NMRs, making adjustment for these births critical to accurate NMR calculations between and within countries.
• 93-102
• 10.6288/TJPH.202402_43(1).112123

• Link 評論 Commentary
• 評論：降低台灣新生兒死亡率—不只是調整活產的認定及收錄標準即可！Commentary: adjusting the criteria of live birth registration is not enough to decrease neonatal mortality in Taiwan
• 林其和
  Chyi-Her Lin
• 無none
  無none
• 無none
  無none
• 103-104
• 10.6288/TJPH.202402_43(1).11212301

• Link 評論 Commentary
• 評論：調整後台灣新生兒死亡率：國際與國內縣市比較Commentary: adjusted neonatal mortality rates in Taiwan: international and domestic comparisons
• 張瑞幸
  Jui-Hsing Chang
• 無none
  無none
• 無none
  無none
• 105
• 10.6288/TJPH.202402_43(1).11212302

• Link 作者回覆 Authors' response to commentary
• 作者回覆：降低台灣新生兒死亡率—標準化死亡率後，還要有系統蒐集相關資料與分析！Authorsʼ response to commentary: reducing neonatal mortality rates in Taiwan: after the adjustment of neonatal mortality rates, we need to collect and analyze relevant data systemically
• 梁富文、呂宗學、江東亮
  Fu-Wen Liang, Tsung-Hsueh Lu, Tung-liang Chiang
• 無none
  無none
• 無none
  無none
• 106-107
• 10.6288/TJPH.202402_43(1).11212303