以Tw-DRGs分類探討全民健保住院日,疾病嚴重度及醫療費用之相關性

As Tw-DRGs classifications inquire into the national health insuranceprotect be length of stay,disease seriousness and medical expense relativity

饒玲瑜(Ling-Yu Jao)

指導教授 ： 楊志良

亞洲大學/健康學院/健康管理研究所/碩士(2006年)

摘要

前言：
台灣實施全民健保，最重要的意義是造福無數弱勢族群達到社會互助之目的，各界無不期盼全民健保能夠永續經營。而改革支付制度，是必要之過程。中央健康保險局擬藉由導入DRG支付制度解決財務問題，已於93年公告Tw-DRGs規劃草案。
目的：
擬藉由健保資料庫之資料，以統計分析方法，對於健保局目前所規劃之第三版DRGs（Tw-DRGs V3.0）支付草案，檢視其是否符合疾病診斷組合之支付邏輯？並利用該分類方法進行住院日、疾病嚴重度與醫療費用相關性之探討。
方法：
資料來源為93年度健保資料庫之住院申報資料檔，經進行Tw-DRGs編輯及統計分析，探討下列問題：一、Tw-DRGs編碼後，可解釋多少住院費用的變異（variation）？又住院日數與住院費用有否相關？二、疾病嚴重度（DRG的相對權重Relative weight）與住院日數有否相關？三、醫院特性是否影響住院日數、疾病嚴重度及住院費用？四、病人特性及醫院特性是否影響住院日數及住院費用？
結果：
醫療費用及住院日數：病人特性之「年齡」及「性別」、醫院特性之「層級別」不同，其平均醫療費用及住院日數有顯著性差異，醫學中心之住院日數及醫療費用均較其他層級醫院為高。「權屬別」之公立醫院與財團法人醫院並無差異；「分局別」之台北分局與東區分局其醫療費用及住院日數均相當並無顯著之差異。
疾病嚴重度（CMI）：醫院權屬別間並無明顯差異；醫院層級間之醫學中心所處置的患者顯然較其他層級為高，與國內外過去之研究結果有相似之處。惟區域醫院與地區醫院間並無顯著之差異，此結果是否因為二者間之規模、形態及照護之患者相近有關？則有待進一步探討之；以健保分局別區分，東區分局可能礙於地形幅員遼闊且醫療資源較缺乏，病況嚴重之病患轉送不易，故其所照顧之患者相對疾病嚴重度會較北區、區中、高屏等分局所屬之西部醫院為高，惟分析之結果並未達統計上之顯著性差異。
Tw-DRGs對於醫療費用之解釋力，與美國學者之研究結果相近；但對住院日數之解釋力，則低於美國之研究結果。
結論：
一、Tw-DRGs之醫療費用解釋力未去極值為49.84%、去極值為58.05%與國外相近，顯示其應為可做為支付之方法。
二、Tw-DRGs之醫療費用變異性大、各DRG組內一致性差。與國外研究相同，因此以Tw-DRGs分類給付可能會面臨與國外相同之DRG取巧行為（DRG Creep）。
三、同一DRG，於不同層級醫院間存在有醫療費用之顯著性差異，顯示Tw-DRGs之相對權重(RW)要用於不同層級醫院應有所調整。

關鍵字

診斷關聯群、住院日、疾病嚴重度、醫療費用

並列摘要

Foreword:
In Taiwan, about the implementation of National Health Insurance, the most important meaning is aimed to benefit countless members of the disadvantaged minority to reach the goal of helping each other in the society. People from various fields sincerely hope for the eternal operation of National Health Insurance. Thus, the reform of payment systems is naturally the inevitably required process. The Bureau of National Health Insurance（BNHI）is now planning to introduce the DRG payment systems with an effort to sooth financial problems. In 2004, the Tw-DRG plan draft had been publicized.

Purpose:
Within this research, by using the data derived from national health insurance database, we carry out statistical analysis to inspect whether the currently projected Tw-DRGs V3.0 Payment Draft can comply with the payment logics of disease seriousness? Also, by using the said classification methods, we explore the correlation among Length of Stay, disease seriousness and medical expenditures.

Method and material:
The data source is retrieved from the inpatient claims of National Health Insurance in 2004. After compiled and analyzed statistically by Tw-DRGs, the major questions to be asked:
1. By Tw-DRGs, how much variation of medical expenditures can be explained? Is there any correlation between Length of Stay and medical expenditures?
2. Is there any correlation between disease seriousness (DRG relative weight) and Length of Stay ?
3. Is there any influence for hospital characteristics to affect Length of Stay , disease seriousness and medical expenditures?
4. Is there any influence for patient characteristics and hospital characteristics to affect Length of Stay and medical expenditures?

Result:
Medical expenditures and Length of Stay : Patient characteristics such as genders and ages, together with hospital characteristics such as levels of hospital, come with the significant different in average medical expenditures and Length of Stay . The medical centers come with higher Length of Stay and medical expenditures higher than those of other levels of hospitals. The public hospitals come without variance between the hospitals of fund juridical persons. The branch-leveled Taipei Branch and East District Branch come without significant variance in medical expenditures mutually.

Disease Seriousness (CMI): It exists in no significant variance among various ownership of hospitals. The patients treated by the hospital accreditation levels medical centers come with higher seriousness than that of patients from other levels of hospitals. The result is similar to past researches made by other overseas organizations.

Conclusion:
1. The explanation strength for Tw-DRGs medical expenditures of no extreme value eliminated is rated at 49.84%. The value after extreme values eliminated is 58.05% and it is similar to those results made by overseas organizations. It reveals that the said Tw-DRGs are availably used as the payment method.
2. It comes with large variance for Tw-DRGs medical expenditures with inferior internal congruence within every DRG group. Results are similar to those made by overseas organizations. Thus, the Tw-DRGs classified payment could possibly confront the opportunistic events the same as the overseas environment. (DRG Creep)
3. Within the same DRG, hospitals from various levels come with the significant variance in medical expenditures. It reveals that it is basically required for further adjustment if the relative weight (RW) of Tw-DRGs is applied to hospitals from various levels.

並列關鍵字

Diagnosis Related Groups, Length of Stay, Case Mix Index, Medical Expenditures

參考文獻

21、韓揆：診斷組合制在台推行問題（上）-DRGs之支付精神，支付邏輯及侷限，醫務管理，6（1）：1-19。

22、韓揆：診斷組合制在台推行問題（下）-論病例計酬及實施準備，醫務管理，6（1）：20-36。

15、張秀薇（2002）：病歷組合複雜性與出院狀況及資源耗用之相關性探討，臺北醫學大學醫務管理學系，碩士論文。

18、劉靜玲（2001）：合併症和併發症影響醫療資源利用之分析：以內科系三種疾病為例，高雄醫學大學公共衛生研究所，碩士論文。

23、謝惠閔（1991）：影響醫療資源耗用因素及DRGs有效校正因素探討-以內科心導管處置為例。高雄醫學大學公共衛生研究所，碩士論文。