| 摘要: | 自1998年健保財務出現入不敷出的困境，健保單位試圖抑制健保支出，諸如；增加部分負擔、高門診次使用者加重負擔、門診合理量、醫院自主管理等抑制費用成長的措施，已嚴重威脅到醫院之永續經營。醫院經營主管為使醫院在此環境中維持應有之競爭能力，對各部門之人力及設備需有完善之規劃，以達到具效能之營運。 門診收入乃為現行醫院營收之主要部分，而門診檢驗是醫院營運收入重要的部分，佔醫院收入約10%，因此醫院對門診檢驗利用之趨勢有所瞭解，以對未來醫院檢驗人力及檢驗設備能做有效之規劃，為極為重要之課題。 本研究以健保制度下某醫學中心門診檢驗利用率為例，以利用門診病人數、門診檢驗病人數、門診診數、門診醫師數、健保制度介入因子等相關資料進行分析，選用Box and Jenkins之時間數列分析來建構以下三式的預測模式：1.單變量ARIMA模型2.轉換函數模型3.介入模型。模式之實證研究則獲得到下列三式的預測模式之應用。一、ARIMA Model AR（1）；經研究顯示門診檢驗利用率數列的諸個觀察值間僅前後一個時期有相關性。二、Transfer function Model；若單僅就門診檢驗利用率對門診診數而言，則統計顯示，若增加門診診數每增加 1000診數，則檢驗利用率將平均增加 0.8 單位的檢驗利用率。三、Intervention Model；門診檢驗利用率，若對放入以上六項介入因子而言，則僅以門診三級制之部分負擔對門診檢驗利用率有顯著影響，增加門診檢驗平均利用率 2.2099 ﹪。 由研究結果可知未來三期預測值對此三模型與實際值之預測模型評估結果的MAPE、RMSE、MAD皆很小，在預測能力指標皆為不錯的預測模型。皆能作為臨床檢驗利用需求預測的管理策略之參考。; In Taiwan, since the National Health Insurance (NHI) program was first implemented in 1995, we have seen an increase in expenditure from under NT $ 137 billion to over NT $ 230 billion in 2000. As a result, several cost containment mechanisms including cost-sharing, financial burden on those frequent visitors in the outpatient clinic, NHI has implemented rational outpatient volume and a hospital self-managerial strategy to solve the deficit problem. Those mechanisms have serious impacts on the continuous operating of hospitals. In order to maintain competence with other healthcare providers, hospital’s manager should have a long-term planning for manpower and facilities within the hospital. Utilization of clinical laboratory services has been an important sector, around 10%, of the overall hospital revenue. It is necessary that hospital’s manager understand the trend of laboratory service utilization and establish an efficient strategic planning for manpower and facilities in the department of clinical laboratory services. The purpose of this study was to build-up a forecasting model for laboratory service utilization. The independent variables are total outpatient volume, % outpatients utilizing laboratory services, the number of physicians served in the outpatient clinic, and a few other intervening factors. The method of Box and Jenkins’s time series was used to construct three prediction models for comparisons. The dependent variable is the rate of laboratory utilization. 1.ARIMA Model AR(1): Our statistical data indicated that there was an association among the observed values in terms of utilization rates, but a close correlation was only present in any two consecutive periods. 2.Transfer Function Model: If only utilization rates over total outpatient volume were considered, it was estimated that there is an increase of 0.8 unit of laboratory service utilization in every 1000 cases. 3.Intervention Model: If all other intervening factors were considered, only cost-sharing factor under a three-level outpatient service system (i.e., $150 for medical centers, $100 for regional hospitals and $50 for all other Preliminary clinics as the outpatient service co-payment under current NHI system) has an impact on utilization rates. This has caused an increase of up to 2.2%. Base on the rules of MAPE, RMSE, and MAD, we concluded that although all of the three models are good enough to serve as predicated model, AR(1) is chosen as the best one. |
