住院診療費用審查決策支援系統之開發與評估

Development and Evaluation of a Decision Support System for Hospitalization and Treatment Payment Auditing

本研究是以健保局中區分局醫療院所住院資料為基礎，建立住院診療費用審查決策支援系統，希望運用決策支援系統來輔助健保審查，以節省審查人力和成本，加強審查效果與效率，並且追求審查的一致性與公平性，進而提高醫療品質與改善醫療行為。我們利用Visual Basic實際開發出一套決策支援系統，以中區健保局的住院資料來實際評估，相同的資料分別以人工審查與決策支援系統審查來加以比較，評估系統的效率與效果，並且調查使用者滿意度。研究所得的主要結論如下：(1)在效果面，審查可歸納的醫令群組時，住院診療費用審查決策支援系統比人工審查效果好。以中區健保局91年11月住院資料，分別經過住院診療費用審查決策支援系統與人工審查兩種方式進行異常資料篩選，不論是篩選出異常件數或金額，或同一醫令篩選出的異常件數與金額，決策支援系統審查都比人工審查效果好。(2)在效率面，不論在審查時間、人力減少、與成本節省上，住院診療費用審查決策支援系統，都比人工審查的效率好。(3)在滿意度面，滿意度調查介於5.40～6.60，屬於滿意與非常滿意之間，平均滿意度為6.03屬於滿意。醫療費用審查在我國是以人工審查為主，以抽樣方式來決定審查個案，有時難失公平性與一致性，並且須花費許多人力與行政成本。本研究結果證實不論是審查效果、效率、使用者滿意度都有顯著成效，在審查公平性、一致性、醫療品質的提昇亦有正面效果。然而，因為醫療行為有其專業性及存有不可歸納項目，所以決策支援系統應定位於支援的角色，在現行的審查與流程模式下支援審查人員的審查工作。

The study takes the inpatient databases of hospitals / clinics of the Central Taiwan Branch of the National Health Insurance Bureau as the grounds to develop a decision support system for hospitalization and treatment payment auditing. It is intended to find ways to economize labor costs of review and strengthen the review performance and efficiency. In turn, it will strive for unanimity and equality of review to upgrade medical service quality and improve medication behaviors. The study employs Visual Basic to develop a decision support system for substantial enforcement with the inpatient databases of the Central Taiwan Branch of the National Health Insurance Bureau. It assesses the systematic efficiency and effects of the survey for user satisfaction. The primary findings yielded through the study are concluded as follows: 1. In terms of the results analyzed through effects in the generalized medical instruction cluster review, the decision support system for hospitalization and treatment payment auditing outperforms the manual review. The inpatient databases of the Central Taiwan Branch of the National Health Insurance Bureau as of November 2002 were screened for abnormalities through both decision support system for hospitalization and treatment payment auditing outperforms manual review. In all cases of the number of abnormal cases and amounts so sorted out or number of abnormal cases and amounts sorted out in a same medical order, the inpatient review decision support system outperforms manual review. 2. In terms of efficiency, in time, reduced labor, and economized costs for the review, the decision support system for hospitalization and treatment payment auditing yielded better effects than manual review. 3. The survey results indicate that the mean score of user satisfaction is 6.03 (The items were measured using a 1~7 seven-points Likert-type scale.). That is, users are highly satisfied with the decision support system. In Taiwan, medical charges are primarily reviewed manually. The determination of the cases of review by means of sample check is inevitably unfair and inconsistent. Moreover, it also calls for numerous labor and administrative costs. The results of this study prove that the system brings significant effect in review results, efficiency, and user satisfaction, and yields positive results in fair review, unanimity and upgrade of medical services. Nevertheless, the decision support system should be positioned as a help to reviewers in manual work under the existing review process.