以申報標準代碼探討申報疾病組合之有效性：案例式推論及資料探勘之應用

Investigate the efficiency of using DRGs in disease alignments applications: Case-Based Reasoning and Data Mining methods

診斷關係群（DRGs）制度是一種將服務量計酬制度改為同病同酬的支付制度。若是申報時採取高報的方式，將造成全民健保財務虧損、病患權利受損，相反的，若是低報則會造成醫院虧損。因此如何正確的申報DRGs成為醫院重要課題。本研究建立一個預測系統，範圍制定在Tw-DRGs的案件資料以及醫令代碼範圍介於47001C到94205B來進行系統的開發，採用資料探勘及案例式推論，分析案例之有效醫令代碼對應，推論有效以及無效之疾病申報代碼與疾病代碼組合，以達到減少申報錯誤機率的目標。使用者在輸入診斷碼或是醫令代碼後，系統會利用診斷碼或醫令代碼進行搜尋，找出相關的病徵組合及相似案例後，呈現給使用者，讓使用者瞭解以往申報成功之案例與對應之醫令代碼組合。如果沒有找到案例，則使用者須比對健保規則，將可申報成功之案例之診斷碼與醫令代碼組合新增至新案例資料庫中，待有一定筆數的資料後，即可進行整理並更新至規則資料庫與原有之案例資料庫中。為了能讓本研究所提之系統更加穩定以及評估地更精確，因此本系統後續實施數個月的系統效率測試評估報告並著重於分析申報疾病代碼之有效性以及驗證申報健保給付疾病代碼組合與有效疾病代碼組合之差異與正確性來做評估。評估報告將於後續蒐集完相關評估資料或是問卷分析後，待達到整體資訊飽和，再利用報告分析之數據進行系統之改善及修正來更加精進系統的穩定性以及精確性。本研究因研究的限制，未來期望能將ICD-10-CM/PCS也應用在本系統中，以再度提升醫院申報方面的品質。

Diagnosis Related Groups (DRGs) is a payment system that change “fee for service” system to fee schedule and peer review mechanism. If to use the highest rate when apply, then it'll cost financial loss for the National Health Insurance Administration and as well as the patients will lose their rights on what they deserved. Contrary, if to apply at the lowest rate, it'll cost financial loss on the hospital side. And so, it's very important for the hospitals to learn about accurate application for the DRGs. This study established a forecasting system, developed in Tw-DRGs range of information on the case and the NHI Order Codes range from 47001C to 94205B system to be developed, the use of Data Mining and Case-Based Reasoning, effective NHI Order Codes study case of correspondence, valid and invalid inference target of disease reporting codes and code combinations diseases, in order to achieve reduction in the probability of reporting errors. After the user has entered in the diagnostic code or NHI Order Codes, the system will use the diagnostic code or NHI Order Codes to search to find relevant combination of symptoms, similar to the case presented to the user, the user's understanding of the case in the past and to declare success corresponding to the NHI Order Codes. If the case is not found, users should compare health insurance rules, will be declared the diagnostic code and NHI Order Codes Success Stories combination of new cases added to the database, there are certain items of information to organize and update to rule the case with the original database repository. In order to allow this study to mention system more stable, more accurate assessments, so the system efficiency of the system and the need to implement the follow-up test a few months to assess the validity of the report focuses on the analysis and reporting of the disease code, verify difference declaration of reimbursement disease codes and code combinations effective combination of diseases and the correctness of the assessment done. Assessment report will be completed to collect data related to the assessment or follow-up questionnaire analysis, the overall information to be reached saturation, the use of the report analyzes data for system improvements and modifications to improve system stability and accuracy. In this study, due to limitation of the study, expected future be able to ICD-10-CM / PCS also used in this system to improve the quality of hospital re-declared area.