醫療服務改善方案能否避免糖尿病相關之住院

Can pay-for-performance program reduce diabetes-related avoidable hospitalization?

吳慧俞(Huei Yu Wu)

指導教授 ： 林文德

長榮大學/健康科學學院/醫務管理學研究所/碩士(2008年)

摘要

中央健保局於2001年11月推行糖尿病醫療服務改善方案試辦計畫，過去研究顯示該計畫藉由定期的生化檢查及衛教等活動，有助於降低糖尿病相關住院，惟此類研究並未考慮加入計畫者的選擇偏差問題，因此本研究目的再進一步釐清糖尿病論質計酬計畫之成效。
本研究採類實驗之雙組前後測設計，自全民健保研究資料庫之承保抽樣歸人檔中，擷取就醫日期為2000年11月至2003年10月連續三年至門診就醫且診斷為糖尿病之病人，藉由傾向分數(Propensity Score)配對方法，依加入計畫者的人口、就醫情形及各項生化檢查之有無等特性，配對加入機率相當之已配對控制組樣本(422人)，復利用差異中的差異(Difference-in-Difference)法分析是否因政策介入對糖尿病相關住院有所影響。
未加入計畫者在計畫實施前一年之人口、就醫情形及各項生化檢查等特性與加入者有顯著差異，但經配對後的控制組與加入者之各特性均無顯著差異。計畫實施後，加入計畫者之各項檢查比例均顯著高於已配對控制組，但糖尿病相關住院則與已配對控制組相當。經差異中的差異分析，計畫實施前後的住院率差異在兩組間亦無顯著差異。
以傾向分數配對法可形成糖尿病論質計酬計畫的可比較已配對控制組，其計畫前一年的各項特性與加入計畫者沒有差異。計畫實施後，加入計畫者生化檢查比例雖顯著高於已配對控制組，但糖尿病相關住院沒有顯著較低，計畫成效頗值商榷。

關鍵字

糖尿病 ； 論質計酬 ； 醫療服務改善方案 ； 可避免住院 ； 門診照護敏感疾病

並列摘要

Purpose:
The Bureau of National Health Insurance (NHI) started the NHI Pay-for-Performance Program in November 2001. Literature review has indicated that through regular laboratory examinations and health educational activities, the admission rate of diabetes patients has dropped, yet these types of studies did not consider the selection bias of program participants. As a result, the purpose of this study is to examine the Pay-for-Performance Program to reduct diabetes avoidable hospitalization?

Material and Method:
This study adopts a two-group pre-test and post-test design and collect subjects diagnosed with diabetes from those who had data filed in the National Health Research Institute datasets of the NHI and continued their clinic visits three years in a row from November 2000 to October 2003. With propensity score matching, subjects (422) at certain metched-control group ratio were added in accordance with the properties of the number of experiment group, clinic visits, and laboratory examinations. In addition, the difference-in-difference method was also used to analyze the influence of police intervention on the hospitalization of diabetes patients.

Result:
Compared group one year before the experiment implementation demonstrated a significant difference in properties such as subject number, clinic visits, and laboratory examination than those of the experiment group. After matching, subjects from the metched-control group did not find significant properties different from those of experiment group. After the program’s implementation, the ratio of each examination result for the experiment group is significantly higher than that of the metched-control group. However, these two groups showed similar ramission rates. After the introduction of the difference-in-difference method, we also found no significant different admission rates for these two groups.
Conclusion:
The propensity score matching method helps to form a comparable diabetes matched-control group in the Pay-for-Performance Program. There has been no significance shown in the properties for each item between experiment group and compare study one year before the program implementation. After the implementation, although the experiment group were found to have a higher latoratory examination ratio than that of the metched-control group and there has been low significance levels in the hospitalization of diabetes patients, which shows a need for further discussion on the program’s effectiveness.

並列關鍵字

Diabetes ； Pay-for-Performance ； Avoidable Hospitalization ； Ambulatory Care Sensitive Conditions ； ACSC

參考文獻

王俊雄、紀櫻珍、吳振龍等：高血壓疾病管理之臨床結果及醫療費用評估。北市醫學雜誌 2006; 3:248-259。

王冠懿、鄭守夏：民眾就醫層級與醫療利用分析：SARS疫情前後的觀察。臺灣公共衛生雜誌 2006; 25:75-82。

王馨儀：氣喘門診照護對可避免住院之影響。臺灣大學醫療機構管理研究所碩士論文，2005。

李待弟：全民健康保險糖尿病醫療給付改善方案初步影響評估。臺灣大學衛生政策與管理研究所碩士論文，2006。

林文德、張睿詒、楊志良：不同醫師專科別間糖尿病門診照護品質之差異。醫學教育 2003; 7:271-281。