多物項共同採購管控模式之探討--以醫院藥品為例

Pharmaceuticals Inventory Management in A Hospital-A Case Study of A Private Hospital

陳淑華(Shu-Hua Chen)

指導教授 ： 丁慶榮

元智大學/工程學院/工業工程與管理學系/碩士(2002年)

摘要

全民健康保險的實施，對我國醫療環境造成極大的衝擊，由於醫療服務的同質性高，使得醫療市場競爭更加激烈，過去研究發現醫療資材成本平均約佔整體營運成本之30%~40%，因此，各大公私立醫院、診所，無不加強醫院管理的功能及經營成本的掌控，希望藉由建立成本會計制度、加強物品採購效益、提高醫事人員生產力、強調投資效益分析等方式，使醫院的營運成本降低、獲利增加。
由於醫院中藥品物項數往往有數百種甚至上千種之多，傳統單物項管理之經濟訂購量模式(economic order quantity, EOQ)，無法提供醫院多物項資材有效率的存貨管理。多物項存貨管理(multi-item inventory control)模式之探討，已有文獻證實，多物項採購策略確實可比單物項單獨採購模式節省成本。因此，本研究嘗試以多物項管理模式應用於醫院藥品存貨管理，探討其對醫院藥品採購與管理之影響，第一步驟將藥品分群，共採用三種不同分群法：1. Chakravarty分群、2.藥理屬性分群、3.供應商分群法，進一步再藉由可訂購點策略(( , , )policy)，找出各藥品之相關管理變數，以期能降低藥品採購、存貨持有成本，以提供較佳之服務水準。
由本研究個案醫院實例驗證，可得以下結論：1.由Chakravarty分群、藥理屬性分群、供應商分群三模式輔以可訂購點策略，其採購成本均會比醫院現行方式高，但存貨持有成本卻可大幅降低，且其總成本均較醫院實際總成本為低，分別可降低$487,957(15.20%)、$505,460(15.75%)、$814,377(25.38%)，其中以供應商分群並輔以可訂購點策略，其採購及管理成本降低幅度為最多。2.分配之需求量平均數( )、標準差( )、存貨持有成本比率( )、配送前置時間( )與總成本成正比，主要採購成本( )與總成本則無較明顯的正比關係存在。3.供應商分群法所求解之( )變數值較小，可以有效控制藥品存量，並降低藥品存貨持有成本，使其總成本相對較低，隨著需求量倍增，供應商分群法其總成本增加幅度明顯較Chakravarty分群、藥理屬性分群法小，亦即當需求量持續增加，供應商分群輔以可訂購點策略之存貨管控模式會比Chakravarty分群、藥理屬性分群模式為佳。

關鍵字

多物項 ； 共同採購 ； 分群法 ； 可訂購點策略 ； 醫院藥品

並列摘要

Since the inception of National Health Insurance Program in March 1995 in Taiwan, lots of new hospitals are constructed and operating around the nation. To compete with other hospitals, the hospital administration should find a way to improve productivity or reduce operating cost. Pharmaceuticals (drug) inventories constitute an important hospital investment. If too much drug is kept in stock, lots of opportunity cost will be found in the yearly balance sheet. However, if too less stock is kept, there is a chance of drug shortage. In such a condition, the hospital not only loses money but also its reputation as well. The extra ordering and transportation costs will also increase at the same time. Therefore, a good control policy is necessary for the pharmaceuticals inventory management.
In this research, an “can-order” policy of the (s, c, S) type, where “s” is the reorder point, “c” is the can-order point, and “S” is the order up to level, is employed. Can-order system is specifically geared to the situation where savings in setup costs are of primary concern as opposed to achieving a specified total replenishment size. Since some of the drugs have a common supplier and can share the setup cost associated with a replenishments placed on the same day, the (s, c, S) model could be applied in hospital pharmacy control. However, to implement such a control policy, you need to find out what other drugs should be checked whenever one item drops to or below its reorder point and triggers a replenishment action. This research clusters the drugs based on three different grouping techniques: Chakravarty approach, pharmacological mechanism of action of the drug, and drug provider/source.
Transaction reports carried at a private hospital pharmacy in 2001 were the main source for the data required to evaluate the (s, c, S) control policy. Other important sources of information were price lists and useful conversations and interviews with pharmacists. There are 270 items of drug to be considered in this research. Based on the three different clustering techniques, the three control parameters, s, c, and S, are computed. Compared to the current control strategy, our results show that the (s, c, S)control policy with three different grouping techniques reduce the total purchasing cost by $487,957(15.20%), $505,460(15.75%), and $814,377(25.38%), respectively. The (s, c, S) control policy grouping by drug provider yields the lowest total purchasing cost. From the sensitivity analysis of the parameters, when the inventory holding cost or the suppliers’ distribution lead time increases, the control variables (s, c, S) and the resulting total purchasing costs increase; while the major setup cost does not show the same impact. Based on the simulation of different coefficients of variation on demand, the (s, c, S) control policy grouping by drug provider shows the best results among three grouping policies.

並列關鍵字

multi-item ； can-order policy ； pharmaceuticals inventory management ； total purchasing cost ； simulation