健保Tw-DRGs費率調整對醫院營收影響之評估研究

A Study of Assessment on Revenue Impacts of Hospitals while Changing DRG Relative Weights

中央健康保險局計劃以DRGs作為醫院醫療費用給付上的依據，其中最受關注的是將原先第一版（舊版）499項DRGs擴增至第三版（新版）的969項，並在未來每年調整相對權重對醫院的衝擊與影響。本研究利用Rasch分析的Winsteps電腦程式進行等距影響力分數之參數估計，以瞭解：（1）是否新版在各月份上及各主要診斷分類（MDC）上優於舊版本；及（2）是否在醫院層級的群體間或是個別醫院內存在有兩版本影響上之差異。DRGs歸判程式分析2004年471家醫院的3,025,964筆出院申報資料，歸判出兩組新舊版本的DRG相對權重，分別有969及499個DRGs。利用Rasch（1960）模式分析轉換後各醫院在各月及各MDC上的Rasch等距分數，並比較醫院間及醫院內的異同。研究結果發現：（1）醫學中心及區域醫院變差，地區醫院則變好；（2）Rasch分析的結果與原始資料取自然對數後的結果相同；（3）推薦採用相減而非相除法來比較分析相對權重在兩版本間利害關係的比較；及（4）MDC被分析出呈二個潛在測量，第一構面呈顯著性變好，第二構面呈顯著性變差（p<.05）。Rasch分析能夠比較醫院在月份及MDC上的CMI變化情形，亦即是可用來分析未來該兩個版本DRG相對權重對醫院利幣得失上的異同比較。

Taiwan’s government plans to pay hospitals for their inpatient discharged services at a predetermined rate (DRGs price) under the prospective payment system (PPS) and to assess the impacts of changes in DRG relative weights between those newly revised and original version. Data analyses in this study were done with Rasch-specific Winsteps computer program to estimate examinee abilities in logits and to understand those of 1) whether the newly revised version of DRG relative weights outperforms the original one in months and/or MDCs; and 2) whether impacts on between-groups or within-individual hospitals exhibited in two versions. Of 3,025,964 inpatient discharged records in 2004 collected from the 471 Taiwan’s hospitals, we classified two sets containing 969 and 499 DRGs, respectively. Rasch modeling (1960) was applied to generating transformed Rasch interval scores for hospitals across months and MDCs, and comparing changes between and within hospitals. The outcomes of this study were those: 1) medical centers and regional hospitals would be worse in contrast to those district hospitals better than before; 2) results were consistent with each other by those transformed Rasch interval scores and the logarithmically raw scores; 3) the subtraction method rather than that of division one in comparing benefits and impacts between the two versions is recommended for future use; and 4) those of MDC constructs were examined to be two underlying measurements, being better and worse consistently to all the hospital levels. Rasch analyses are worthy of being regarded as a comparison method in CMI changes across months and MDCs, which is to analyze benefits and impacts between the two versions on DRG relative weights in the future.