利用儀表板評估受新冠病毒影響之醫院營收和比較

Dashboards for evaluating and comparing the hospital incomes affected by COVID-19

新型冠狀病毒遍及全球許多國家和地區，卻也影響著許多醫院的營運收入。我們假設：（1）門診收入影響大於住診；（2）大型醫院受疫情影響大於小型醫院。本研究繪製前後，兩年三月和四月的醫院門診和住診方面案件數和收入增長趨勢的散佈圖，然後，驗證上述的兩項假設。自健保險署網站下載468家醫院的案例數和申報的健保點值。用微軟Excel中的三角函數（＝DEGREE（Atan（成長率））定位散佈圖上的座標（即兩軸分別為兩月的成長座標角度而非傳統的成長率）。四項圖例說明：門診和住診的（1）案件數、和（2）健保點值增減情形。由兩個月間醫院成長率的相關係數，驗證新冠病毒影響醫院營收，相關係數越高，表示趨勢的一致性愈強。另再對散佈圖4個象限的數量進行卡方關聯性檢定，檢定不同醫院類型間的影響趨向。本研究於谷歌地圖上設計散佈圖儀表板，呈現研究的成果。我們發現（1）由於兩個月間，在門診案件數和收入成長之相關係數，分別為0.79和0.83，大於住診的0.40和0.18，說明門診收入影響大於住診。醫院類型（例如，醫療中心，區域醫院和地方醫院）間的門診和住診受影響情形一致（案件數和收入之p＝0.14和p＝0.46），而住診則較不一致（p＜0.001），這說明較大醫院的受影響較小醫院來得大。在谷歌地圖上呈現醫院受新冠病毒影響營運件數及申報點值情形，利用本研究的方法，提供未來進行類似研究之參考。

The outbreak of COVID-19 spread worldwide in many countries/areas. What extent to which the hospital revenues eroded by COVID-19 is required for verifications. Hypotheses were made by (1) outpatient departments (OPD) in hospitals were affected more than that as of inpatient departments (IPD), and (2) the bigger hospitals were influenced more than those smaller ones. We drew scatter plots of growth trends of case numbers and revenues in March and April (2020) for hospitals on aspects of OPD and IPD and then verified the hypotheses mentioned above. Case numbers and revenues of 468 hospitals were downloaded from the website of the Taiwan Government-run National Health Insurance Administration (TGNHIA). Scatter plots were drawn for all studied hospitals located on coordinates by using the trigonometric function ( =DEGREES( Atan (growth rate)) in MS Excel. Four domains were inspected, including both (1) case numbers and (2) revenues (denoted by medical fees) claimed to TGNHI in March and April of 2019 and 2020 referring to OPD and IPD, respectively. Hospital incomes affected by COVID-19 were examined using (1) correlation coefficients (CCs) derived from the growth rates of hospitals between two months, and (2) Chi-square tests on the numbers in quadrants of scatter plots. The higher CCs mean the consistency and strength in trend. The Chi-square was used for examining the impact tendency across hospital types. Dashboards of the study scatter plots were designed and online shown on Google Maps. We found that (1) OPDs with CCs=0.79 and 0.83 were hit harder than that in IPDs with CCs=0.40 and 0.18 due to CCs in case numbers and revenues; and (2) all those hospital types (i.e., medical centers, regional hospitals, and local hospitals) present consistent in OPDs (p=0.14 and p=0.46 in case numbers and revenues), but inconstant in IPDs (p<0.001), indicating that bigger hospitals were struck more than the smaller ones, particularly in IPDs. The demonstration of a dashboard using dashboards on a map can inspire other relevant research to replicate the approaches used in this study for other countries struck by COVD-19 in the future.