國家發展委員會於人口重要指標大事紀預估臺灣地區人口紅利將於2028年結束，伴隨高齡化之醫療需求增加，醫療建築逐漸朝大型化與複合化發展，空間、構件與機電系統均需在不中斷醫療服務下維持穩定。相較於一般建築，醫療建築之維護與修繕更直接關係病患安全、感染控制、醫護效率與營運韌性。然而，在健保財政與院方預算排擠之有限條件下，若仍以損壞後修復或經驗式應變作為主要管理方式，將難以掌握老朽化趨勢，也不利既有醫院建築之長期使用。因此，如何在不中斷醫療服務之前提下，辨識關鍵維護項目，並將日常請修紀錄轉化為長壽命化管理決策，為本研究探討課題。
本研究以國立成功大學醫學院附設醫院之表單流程控管系統請修申請作業資料為對象，蒐集2016年至2024年九年間共計233,910筆數據庫，經篩選後以門診大樓與住院大樓之124,677筆有效工單作為分析基礎。研究方法包括文獻回顧、量化統計及管理者深度訪談，並參照日本國土交通省長期修繕計畫標準樣式及厚生勞動省醫療機構基礎設施長壽命化計畫(個別設施計畫)制定指南之概念，檢視醫療建築維護管理由被動修繕轉向預防性維護可行性。本研究將醫療建築長壽命化界定為「故障類別與空間連動」生命週期管理，聚焦營繕、水電、弱電、空調與消防五項類別，分析其修繕頻率、部位分布與空間使用之關聯。
研究結果顯示，修繕負荷並非平均分布，而集中於高接觸、高使用頻率及直接影響醫療流程之部位。其中營繕類以門(39%)與室內裝修(31%)為主，水電類以照明(39%)、衛生設備(24%)及給排水設備(29%)為主，弱電類集中於監視設備(25%)與病房呼叫(24%)，空調類則以空調設備(76%)為大宗，消防類為避難方向指示(17%)、緊急照明(16%)及滅火器(14%)，攸關生命安全須維持維護優先性。維護成本每筆11,772元/筆、單位面積年度維修費每平方公尺981元。建築生命週期階段與空間機能會影響修繕型態，住院大樓以病房、急診、行政單位及影像醫學部門呈現高維護負荷，門診大樓則以診間、行政單位及影像醫學部門較為明顯。院內之請修申請作業平台類別錯置、空間資訊不足與資產履歷未完整連結等問題，會降低數據庫對異常辨識、修繕排序及中長期預算配置之支援能力。此外，深度訪談指出院方雖具備保養、年限汰換與跨單位協調等維護基礎，惟實務管理模式偏向事後修繕，且判斷多仰賴經驗導向之即時應變，現行數據庫尚未充分轉化為劣化評估並支援決策功能。
本研究建議院方應以既有請修申請作業平台為基礎，落實空間編號、資產編號、維修金額與修復說明等欄位，並由工務登記桌採取人機協作以強化類別把關功能，使請修資料由行政派工紀錄轉化為設施維護管理資料庫。其次，應導入「劣化度×重要度」風險矩陣，將關診關床營運影響納入修繕排序，作為預防性維護與年度預算編列依據。最後，對於仍須持續營運之既有醫院建築，宜以五十年為長壽命化計畫週期，每年更新現況資料，每五年滾動修正實施計畫，並預留可支援分期施工、病床調度與感染控制之備援空間。透過上述策略，醫療建築維護管理得以由零散修繕轉向具資料基礎、風險意識與營運韌性之建築生命週期管理模式。

According to the National Development Council’s chronology of key demographic indicators, Taiwan’s demographic dividend is projected to come to an end in 2028. As population aging drives an increase in healthcare demand, hospital buildings have gradually developed toward larger scales and greater functional complexity, and their spaces, building components, and building service systems must remain stable without interrupting medical services. Compared with ordinary buildings, maintenance and repair in healthcare facilities are more directly associated with patient safety, infection control, staff efficiency, and operational resilience. However, under the dual constraints of National Health Insurance financing and increasingly limited hospital budgets, continued reliance on failure-based repair or experience-driven response as the primary management approach makes it difficult to identify long-term deterioration trends and is unfavorable to the sustained use of existing hospital buildings. Accordingly, this study investigates how critical maintenance items can be identified and how routine repair-request records can be transformed into service-life extension management decisions strategies without disrupting medical services.
This study uses repair-request data extracted from the Form Workflow Control System of National Cheng Kung University Hospital. A total of 233,910 repair-request records collected between 2016 and 2024 were analyzes. After screening, 124,677 valid work orders from the Outpatient Building and the Inpatient Building were retained as the basis for analysis. The research methods included a literature review, quantitative statistical analysis, and in-depth interviews with managers. The study also drew on the Standard Format for Long-Term Repair Planning issued by Japan’s Ministry of Land, Infrastructure, Transport and Tourism, as well as the Guidelines for Formulating the Infrastructure Longevity Plan (Individual Facility Plan) for Medical Facilities issued by Japan’s Ministry of Health, Labour and Welfare, in order to examine the feasibility of shifting healthcare building maintenance management from reactive repair to preventive and predictive maintenance. In this study, the service-life extension of healthcare buildings is defined as a lifecycle management approach based on the linkage between failure category and spatial function. The analysis focuses on five categories, including architectural works, plumbing and electrical systems, low-voltage systems, HVAC systems, and fire protection systems, and examines the relationships among repair frequency, component distribution, and intensity of space use.
The results indicate that repair demand is not evenly distributed; rather, it is concentrated in components characterized by frequent contact, intensive use, and direct influence on medical workflows. Within architectural works, doors (39%) and interior finishes (31%) accounted for the highest proportions. Within plumbing and electrical services, lighting (39%), sanitary fixtures (24%), and water supply and drainage equipment (29%) were the major items. Low-voltage repairs were concentrated in surveillance systems (25%) and nurse call systems (24%), whereas HVAC equipment accounted for the majority of HVAC-related repairs (76%). Fire protection system repairs were primarily concentrated in exit signs (17%), emergency lighting (16%), and fire extinguishers (14%), reflecting the maintenance priorities of life-safety facilities required to remain operational under emergency conditions. The average maintenance expenditure was NT$11,772 per work order, while the annual maintenance cost averaged NT$981 per square meter of gross floor area. Repair patterns were also influenced by the stage of the building lifecycle and by spatial function. In the Inpatient Building, high maintenance loads were concentrated in wards, the emergency department, administrative units, and imaging departments. In the Outpatient Building, high maintenance loads were most evident in consultation rooms, administrative units, and imaging departments. At the same time, problems such as category misclassification within the internal repair-request platform, insufficient spatial information, and incomplete linkage of asset histories reduce the database’s capacity to support anomaly identification, repair prioritization, and medium- to long-term budget allocation. In addition, the in-depth interviews revealed that although the hospital has established a basic maintenance foundation-including routine servicing, service-life-based replacement, and cross-departmental coordination-its practical management model remains predominantly reactive. Decision-making still relies heavily on experience-based, real-time responses, and the existing database has not yet been fully converted into a tool capable of supporting deterioration assessment and maintenance decision-making.
This study therefore recommends that the hospital build upon its existing repair-request platform by standardizing key fields such as space identification numbers, asset identification numbers, repair costs, and restoration descriptions. The facilities registration desk should adopt a human-AI collaborative approach to strengthen category verification, thereby transforming repair request records from administrative dispatch logs into a facility maintenance management database. Second, a risk matrix based on “degree of deterioration × criticality” should be introduced so that the operational impacts of clinic closures and bed closures can be incorporated into repair prioritization and used as a basis for preventive maintenance planning and annual budget preparation. Finally, for existing hospital buildings that must remain in service, a 50-year service-life extension planning horizon is recommended, with current-condition data updated annually and implementation plans revised on a rolling five-year basis. Backup spaces capable of supporting phased construction, bed reallocation, and infection control should also be reserved in advance. Through these strategies, healthcare facility maintenance management can evole from fragmented repair practices into a data-driven lifecycle management framework that emphasized risk awareness and operational resilience.

學術論文(博碩士論文)
[1]黃業展。建築物長期修繕健檢快篩研究。逢甲大學碩士論文，2023年。
[2]顏瑞庭。預測性維修與預防性維修的比較研究。國立臺灣科技大學碩士論文，2017年。
[3]江立偉。建築物外牆公共安全目視診斷評估方法之研究。國立高雄大學碩士論文，2009年。
期刊文章／報導／書籍
[1]江支川。21世紀醫院規劃設計(入門篇)，詹氏書局，臺北市，2022年3月。
[2]杜功仁、林坤緯。台北地區醫療機構設施管理部門之管理模式與營運績效，台灣物業管理學會，2008研究成果論文發表會，2008年。
[3]黃世孟、杜功仁、張智元、卜遠程、王順治、楊詩弘、顏世禮。物業管理名詞彙編，麗文文化，高雄市，2009年。
[4]張智元。建築醫學與物業管理，台灣物業管理學會，第1屆台灣物業管理學會論文發表會論文集，2007年3月。
[5]羅紫萍、郭斯傑。大型醫院建築維護成本分析-以國立台灣大學醫學院附設醫院為例，物業管理學報，2011秋季號。
[6]國立文化資產保存中心籌備處。台南市市定古蹟原日軍台南衛戍病院調查與修復計畫期末報告，2005年。
[7]蔡侑樺、徐明福。衛戍榮光：原日軍臺灣步兵第二聯隊營舍級臺南衛戍病院建設歷程，成大出版社，2022年。
[8]陳瑞鈴、林憲德。台灣建築物生命週期使用年限調查之研究，內政部建築研究所研究報告，2001年。
[9]Lavy S., Shohet I. M., Computer-Aided Healthcare Facility Management. Journal of Computing in Civil Engineering 2007, 21:5 363-372. DOI: 10.1061/(ASCE)0887-3801(2007)21:5(363).
[10]Scholz et al., Rapid assessment of infrastructure of primary health care facilities - a relevant instrument for health care systems management; BMC Health Services Research (2015) 15:183. DOI 10.1186/s12913-015-0838-8
[11]成大醫院，表單流程控管系統使用手冊簡報。
法令規範
[1]內政部國土管理署。建築物公共安全檢查簽證及申報辦法。2022年11月28日。檢自https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=D0070149
[2]財政部。固定資產耐用年數表。2017年2月3日台財稅字第10604512060號令。檢自 https://law-out.mof.gov.tw/LawQuery.aspx
[3]交通部高速公路局。高速公路橋梁目視檢測手冊。2024年10月。檢自https://www.freeway.gov.tw/Mp/MPPeriodical.aspx?id=61&cnid=1644
[4]內政部國土管理署。建築技術規則建築設計施工編。2026年2月23日。檢自https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=D0070115
[5]內政部國土管理署。防火避難綜合檢討執行要點。2019年12月10日。檢自https://glrs.moi.gov.tw/LawContent.aspx?id=FL028016
網路資源
[1]國家發展委員會人口推估查詢系統。檢自https://pop-proj.ndc.gov.tw/News.aspx?n=3&sms=10347
[2]全國法規資料庫。檢自https://law.moj.gov.tw/Index.aspx
[3]內政部主管法規查詢系統。檢自https://glrs.moi.gov.tw/index.aspx
[4]國立成功大學醫學院附設醫院。檢自https://nckuh.hosp.ncku.edu.tw/p/412-1000-29534.php、https://esg.hosp.ncku.edu.tw/p/406-1110-269845,r2799.php?Lang=zh-tw、https://eng.hosp.ncku.edu.tw/app/index.php
[5]日本國土交通省。長期修繕計画標準様式。檢自https://www.mlit.go.jp/jutakukentiku/house/jutakukentiku_house_tk5_000052.html
[6]日本厚生勞動省。インフラ長寿命化計画(行動計画)。個別施設計画策定のためのマニュアル・ガイドライン等。檢自https://www.mhlw.go.jp/topics/2015/04/dl/tp0416-01-01.pdf
[7]日本國土交通省。マンションの建替えか修繕かを判断するためのマニュアル。檢自https://www.mlit.go.jp/jutakukentiku/house/content/001374067.pdf
[8]財團法人台灣綠色生產力基金會。服務業節能服務網。節能典範案例。檢自https://ea01.moeaea.gov.tw/e0406/01/Knowledge/knowledge_more?id=8844d27aebd94274adfe8e2737e00d47
[9]Lounge Talk 採訪計畫。築醫者-許常吉建築師。檢自https://loungetalk.weebly.com/part-1-322273621565292332742877137291302742431431689303403380433469/part-1
[10]衛生福利部疾病管制署。醫療機構感染管制措施指引，檢自https://www.cdc.gov.tw/Category/ListPage/NO6oWHDwvVfwb2sbWzvHWQ