居家藥事服務之智慧化排程初探

Abstract

隨著台灣人口逐漸老化，長照2.0中的居家藥事服務已經成為許多全民健保特約藥局經營者關注的焦點，並且傾全力發展此項業務。而隨著業務逐漸開展，接受服務的慢箋領藥個案人數逐漸增加，每間藥局開始面臨新的問題――「如何找出高效率的送藥排程組合？」，意即在次工作日居家藥事服務需送確定後，如何安排多位執勤藥師的送藥順序，以達到各家特約藥局訂定的獲利與成本營業目標，如掉單率最低，或是送藥路程總距離最短……等。特別值得一提的是，本問題中每日待安排的個案人數眾多，可行的居家藥事服務排程組合數量也會呈指數成長，但特約藥局每日可進行排程的時間有限。在這個狀況下，過往許多已發展的演算法，如「多行程車輛路徑排程演算法」或「天際線路徑查詢演算法」都無法順利應用在這個問題上。本論文探討居家藥事服務之智慧化排程所需要素與執行步驟，並分別調查這些執行步驟的可行演算法，最終測試這些演算法串連後是否適用於這個問題。最後，為了驗證本論文所提出演算法的效率與效能，本論文使用台灣南部六縣市的真實路網結合虛擬的個案資料點位來進行模擬。

With Taiwan’s recent demographics aging gradually, the management and development of homecare pharmaceutical services in the Long-Term Care Services 2.0 becomes the center focus of many institutional pharmacies. The increasing demand of homecare pharmaceutical services brings a critical question for every service provider: how to schedule a more effective delivery routine? Specifically, the logistics and management, of delivery of prescriptions to patient homes, for on-duty pharmacist on a daily basis in the most cost-effective and efficient manner that also meets the business operation goals such as high retention rates. In the past, pharmacies’ delivery scheduling are based on calculations such as the Multi-Trip Vehicle Routing and Scheduling Problem and the Skyline Path Query Algorithm; however, these models cannot sufficiently address the needs of pharmacies delivery system because due to time constraints and the rapid increasing demand, on-duty pharmacists simply cannot fit more scheduling in a day’s work with the current workload. To address the eminent gap seen in the current homecare pharmaceutical services, this research aims to explore the critical elements and procedures required for smart scheduling of homecare pharmaceutical services and examines feasible algorithms for said procedures, and tests whether these algorithms are applicable in solving the core issue. Finally, to validate the efficiency and effectiveness of the algorithm proposed in this research, simulations were ran using the actual roadmap networks across six southern cities with virtual patients as test subjects.