- 學位論文
醫療業歷史需求之非限化估計研究應用於動態預約系統
A Case Study on Dynamic Reservation Systems Using Unconstraining Methods for Censored Data in Healthcare
摘要
本研究主要探討醫療業歷史需求受限下非限化估計方法的套用以及台大醫學中心之健康管理部門多類型醫療資源與多重顧客間關係與配置，以減少資源浪費、最大化其利潤為目標。此部門顧客由普通健檢及精密健檢兩類顧客組成。精密健檢顧客主要使用之關鍵資源為正子掃瞄（PET）及核磁共振掃瞄（MRI）; 而普通健檢顧客之關鍵資源為MRI。由於MRI為兩類顧客共用關鍵資源。每當普通健檢顧客量大增或提前到來，且現行預約系統在MRI服務量可滿足顧客需求量條件下，即允許顧客預約要求，在未考慮預留MRI給未來精密健檢顧客的條件下，PET關鍵資源也就會因此閒置，即為浪費，利潤將無法最佳化。針對此問題的特性，我們先利用不同非限化估計方法改善此健檢中心需求設限問題，再利用動態規劃求解，本研究結論可提供部門管理者即時且精準之決策參考，對於安排顧客預約規劃有顯著的效益。本案例雖有其特殊性，但我們的模型在許多方面可套用在不同的醫療資源規劃，如一般病患會需要利用到斷層掃描儀器，但癌症病患在做放射性治療前，一定需要先經過斷層掃描儀器檢查後，才可接著做放射性治療，此類資源規劃問題皆可利用本研究成果加以應用。未來研究方面，由於多項文獻指出顧客的預約行為會受到等候時間長短而影響顧客的預約取消率與未到診之人數變化，因此，未來我們將針對不同取消率之顧客分群，進而決定超額預約之決策，以達到長期利益與服務水準最大化。
並列摘要
In this research, we endeavor to streamline the demand management by developing a better scheduling system. This modified decision-making system would present decisions on how to use resource protection levels to increase profits in the long term. Specifically, we observed that the National Taiwan University Hospital’s Health Management Center (NTUH-HMC) does not reserve any MRI installations for precise comprehensive health examination (PCHE) customers and that it ends up being booked at capacity. Considering different types of customers’ demand, cancellations, and no-shows, setting up the dynamic booking limit could help improve the center’s income by satisfying the potential future PCHE customer, who wants to be examined more extensively. This kind of policy adjustment is called dynamic booking limit policy. By implementing these policies it will also improve the management of the center’s resource usage and increase its revenue. However, the historical data we obtained were partially censored. Lacking in unconstrained demand data makes it difficult to make better customer sets and capacity allocation decisions. Therefore, we also demonstrate several unconstrained methods and adapt a few models in order to be more precise to predict the true demand distribution. Our method of setting protection levels to incorporate a dynamic booking limit policy could also apply to different reservation systems, especially for perishable production and service industries, such as restaurants, hotels, car rental businesses, and travel agencies.