健康檢查中心的磁振造影排程優化：榮科醫學影像中心個案研究

Abstract

背景介紹：自費影像健檢是醫療業的一個利潤豐厚的領域，涉及一系列的檢查。為了提高顧客滿意度和維護穩定的顧客基礎，有效的工作流程整合至關重要。手動排程基於經驗數據，相對較不靈活，但容易監控工作流程。但是，每次排程的最佳化會因多種變量而異，例如檢查類型、顧客合作和員工協調，這在臨床實踐中很難管理。此研究主旨在利用導入整數規劃(Integral Programming)改善優化顧客等待時間和工作流程的可行性。 材料和方法：我們基於整體規劃算法設計了一種排程方法。而人工手動排程方法則作為對造組。磁振造影(MR)和醫師(MD)分別被視為主要的機器資源和人力資源；時間限制包括工作時間，前置時間等限制。基於工作經驗，我們提供5種場景包含不同顧客數和不同健檢組套來測試兩種方法的性能。客人可以選取六種不同組套之一，其中總時長超過100分的為大程序包，反之為小程序包。目標式為最小化顧客閒置時間和資源超時時間。MR閒置和MD閒置時間則列為資源使用率作為參考指標。 結果：相較於手動方法，整數規劃方法在各種情境中通常導致整體客戶閒置時間略有增加。可能的原因是我們限制了整數規劃程式尋找最佳解決方案的運行時間為20分鐘。在客戶較少的情況下，整數規劃方法在減少客戶超時時間方面比手動方法更有效率，在20分鐘就達到接近100%的最佳化結果，比手動方式約2小時明顯較省時。相反，在客戶數量爆量的場景中，手動方法在減少客戶超時方面反而更為有效。經內部分析，主要衝突是由於MD解釋時間的重疊所引起的，凸顯了MD可用性的重要性。在MR閒置時間方面，整數規劃方法在5個場景中的3個場景表現比手動方法優越。另一方面，對於減少MD閒置時間，手動方法在5個案例中有3個比整數規劃方法更為有效。不過，值得注意的是，這些差異通常都是微小的。整體而言，隨機案例場景的目的是測試整數規劃方法對於不可預測狀況的適應能力，和手動方法比較顯示出相對穩定的結果。 結論：我們首次嘗試在有限時間內使用整數規劃方法安排影像健康檢查排程，雖然比傳統人工方法稍微遜色一些，但是不失為一個減少對人工勞力的依賴，並量化資源利用等有效方法，這樣的數位化有助於我們能夠適應動態的不可預見因素，如緊急情況或設備故障，為更精密和以顧客為中心的解決方案展開解決之道。通過縮短等待時間，我們最終能提升當代健康檢查服務中的顧客體驗。

Introduction: Out-of-pocket Imaging health checkups are a lucrative aspect of the medical industry that involves a range of examinations. Effective workflow integration is crucial for enhancing customer satisfaction and maintaining a consistent customer base. Manual scheduling is based on empirical data and is relatively inflexible, allowing for easy monitoring of workflow. However, the duration of each examination varies depending on multiple variables, such as the type of examination, customer cooperation, and staff coordination, making it difficult to manage in clinical practice. This study aims to utilize integral programming (IP) to design a workflow that optimizes customer waiting time and workflow simultaneously. Materials and Methods: We designed a IP scheduling method based on the overall planning algorithm, with the manual scheduling method as a control method. Magnetic Resonance (MR) and Medical Doctors (MD) are considered as the main machine resources and human resources respectively; time constraints include working hours, lead time, etc. Based on work experience, we provide 5 scenarios that contain different numbers of customers and different health check-up packages to test the performance of the two methods. Customers can select one of six different packages, where those with a total duration of over 100 minutes are considered large packages, and those less are small packages. The objective is to minimize customer idle time and resource overtime. MR idle and MD idle times are listed as resource utilization rates for reference. Results: Results: Compared to the manual method, the IP method typically leads to a slight increase in overall customer idle time across various scenarios. A possible reason is that we limited the run time of the IP process to find the best solution to 20 minutes. In cases with fewer customers, the IP method is more effective in reducing customer overtime and reaches close to 100% optimization in just 20 minutes, which is significantly more time-efficient than the manual method that takes about 2 hours. Conversely, in scenarios with a large number of customers, the manual method is actually more effective in reducing customer overtime. Internal analysis revealed that the main conflict was caused by the overlap of MD interpretation times, highlighting the importance of MD availability. In terms of MR idle time, the IP method performed better than the manual method in 3 out of 5 scenarios. On the other hand, for reducing MD idle time, the manual method was more effective in 3 out of 5 cases. However, it's worth noting that these differences are usually minimal. Overall, the purpose of the random case scenarios is to test the adaptability of the IP method to unpredictable situations, and when compared with the manual method, it demonstrates relatively stable results. Conclusion: This is our first attempt to use the IP method to arrange imaging health check-up scheduling within a limited time, and although it is slightly inferior to the traditional manual method, it is still an effective way to reduce reliance on manual labor and quantify resource utilization. Such digitization helps us adapt to dynamic unforeseen factors such as emergencies or equipment failures, paving the way for more precise and customer-centered solutions. By shortening waiting times, we ultimately improve the customer experience in contemporary health check services.