加入論質計酬對照護過程與照護結果之影響─以氣喘為例

The Effect of Introducing Pay-for-Performance Program on Care Process and Care Outcomes─An Example of Asthma

簡怡祥(Yi-Hsiang Chien)

指導教授 ： 鍾國彪

國立臺灣大學/公共衛生學院/健康政策與管理研究所/碩士(2017年)

https://doi.org/10.6342/NTU201702202

摘要

研究背景與目的：氣喘是常見的慢性疾病，病患需要定期接受醫師診療與服用藥物，並且有自我管理之能力，以避免急性發作的危險。故台灣實施氣喘論質計酬方案鼓勵醫療院所加入，又稱「氣喘醫療給付改善方案」，以加強氣喘患者之追蹤管理與衛教。然而台灣的氣喘論質計酬制度並非論成效計酬，而是包裹給付，過去關於氣喘論質計酬是否能提升照護過程和結果的研究鮮少，雖認為該制度有助於提升門診次數、降低急性醫療耗用和改善病情，但多為部分區域與單組前後測類實驗設計（single group before-and-after design），也尚缺乏以全國為範圍之氣喘論質計酬研究。此外，過去研究發現糖尿病論質計酬方案有風險選擇問題，但尚未探討氣喘論質計酬方案是否也有此情形。究竟氣喘論質計酬對照護過程和結果的影響為何，和是否有風險選擇的健康不平等現象，是本研究欲釐清之問題。然而欲進行氣喘的大樣本資料庫研究，需先找出能將病人的嚴重度適當分級之方法。嚴重度是重要的疾病特性，不同嚴重度其預後與治療方式皆不同，國外已有數個較常使用的嚴重度分級準則，但目前尚無定論最適用者為何，國內氣喘嚴重度的研究也付之闕如。因此，本研究欲建立以處方用藥為病人氣喘嚴重度分級的評估模式，並探討氣喘病患的病人特性是否會影響有無加入論質計酬方案，形成風險選擇問題。且有無加入氣喘論質計酬方案是否會造成氣喘追蹤照護的不同，進而影響照護成果。

材料與方法：本研究為回溯型世代研究，採用全民健康保險研究資料庫，並篩選2007-2008年4,157位成年氣喘新發個案進行資料追蹤。先發展一氣喘嚴重度分級的評估模式，再以存活分析探討用藥分級下，病人的急性發作是否有差異，以驗證所建立之分級標準是否合宜；之後以羅吉斯與卜瓦松迴歸模型探討有無加入氣喘論質計酬方案是否受病人特性與嚴重度的影響；最後在控制病人特性與嚴重度後，檢驗病人有無加入氣喘論質計酬方案與其照護成果的關係，是否會受到照護過程的中介效果影響。
研究結果：以用藥為氣喘嚴重度分級的方式，在合併部分組別後組間的後續急診與住院情形達到顯著差異，評估模式可以將病患區分為三級別或二級別；病人的嚴重度、城鄉別、就醫醫師年齡、就醫院所層級皆會影響到是否加入氣喘論質計酬方案，嚴重度為中重度者、高度都市化區域、基層診所、醫師較年輕者加入的勝算較高。而檢驗病人有無加入氣喘論質計酬方案與照護過程、照護結果的關係，控制其他變項後，結果為自變項「有無加入論質計酬」並不會顯著影響照護過程之中介變項「是否一年內回門診追蹤三次以上」，中介變項「是否一年內回門診追蹤三次以上」對部分的照護結果達統計顯著；但自變項「有無加入論質計酬」並不會顯著影響照護結果，故三者中介關係不成立。

結論與建議：氣喘論質計酬沒有發生糖尿病論質計酬之採櫻桃謬誤(Cherry-picking)，也未有對氣喘的追蹤與照護結果具顯著影響。再加上參與人數不多，可能需要制度面調整和增加誘因。對未來研究建議包含對分級模式的進一步改善，可配合臨床初級資料來比較確認；發展適合未成年患者的分級模式，以利將研究對象擴大進行相關探究。

關鍵字

氣喘 ； 氣喘嚴重度 ； 論質計酬制度 ； 風險選擇 ； 照護過程 ； 照護結果

並列摘要

Background and purpose: Asthma is a common chronic disease. Patients need regular medical treatment and medication, and have self-management ability in order to avoid the risk of acute exacerbations. Therefore, Taiwan has implemented the pay-for-performance (P4P) program for asthma, and encouraged medical institutions to join the program to strengthen the tracking management and health education for asthma patients. However, the P4P program for asthma in Taiwan is not effect-based payment, but bundled payment. There are few past studies regarding whether asthma P4P program can improve the care process and outcome. Although those studies suggest that the program does help increasing the outpatient visits, reducing acute medical resource utilization and improve patients’ condition, most of them are regional and single group before-and-after experimental design. Nation-wide asthma P4P studies are also lacking. In addition, past studies has found there is risk selection issue in diabetes P4P program, but has not explored if this is also the case for asthma. This study aims to explore the effect of asthma P4P program on care process and outcome, and whether there exist the phenomenon of health disparity in risk selection. To carry on large sample database study for asthma, measures for appropriate asthma severity classification are necessary. Severity is an important disease characteristic. Patients with different disease severity grades have different prognoses and are suitable for different methods of treatment. Several severity classification criteria are commonly used abroad, but which of them is the most appropriate is still open to debate. Moreover, Domestic research on asthma severity is also absent. Hence, this study attempts to establish an assessment model of asthma severity classification based on the prescribed medication for each patient, and explore if the patient characteristics of asthma patients affect whether they join the P4P program, leading to risk selection issue, and if joining the program makes difference in asthma tracking and care, thus influencing care outcome.

Materials and methods: This study is a retrospective cohort study. We used National Health Insurance Research Database to filter out 4,157 newly-diagnosed asthma cases between 2007-2008, and conducted data tracking. We first developed an assessment model of asthma severity classification, and then explored whether there is difference in patients’ acute exacerbations under medication-based classification through survival analysis, in order to examine the appropriateness of the established classification criterion. Subsequently, we explored if patient characteristics and severity affects whether patients join the asthma P4P program. Lastly, with patient characteristics and severity controlled, we examined if care process is the mediator for the relationship between whether patients joining the program and their care outcomes.

Results: Under medication-based asthma severity classification, after some of the groups being merged, the difference between the follow-up emergency and hospitalization of each group reached significant level. The assessment model can divide the patients into three or two stages. Each Patient’s severity, degree of urbanization, and his doctor’s age and the level of the medical institution he goes to, all affect whether the patient joins the asthma P4P program. There is higher probability for patients with moderate to severe severity, living in highly urbanized area, receiving treatment in primary clinics and having younger doctors to join the program. Examining the relationship between whether patients joining the asthma P4P program and their care processes and outcomes, after controlling other variables, we found that the independent variable “joining the P4P program” does not significantly influence the care process mediator “outpatient visits three times or above in a year”. Also, the mediator “outpatient visits three times or above in a year” has statistically significant influence on some of the care outcomes, but the independent variable “joining the P4P program” does not significantly influence the care outcomes. Therefore, there is no mediation between the three.

Conclusion and suggestions: The cherry-picking in the diabetes P4P program does not happen in the asthma P4P program, and the asthma P4P program has no significant influence on asthma tracking and care ou tcome. Considering not many patients participating in the program, institutional adjustment and incentive increasing may be necessary. Suggestions for future studies include further improvement to the classification method, with the introduction of clinical primary data for compare and confirmation, and development of classification method suitable for underage patients to extend the subjects of study.

並列關鍵字

Asthma ； Asthma severity ； Pay-for-performance program ； Risk selection ； Care process ； Care outcome

參考文獻

方俞尹(2009)。全民健康保險氣喘醫療給付改善方案長期影響評估。臺北市：國立臺灣大學公共衛生學院衛生政策與管理研究所碩士論文。

吳家興, 林瑞雄, 謝貴雄, 邱文達, 陳麗美, 邱淑媞, . . . 方淑慧. (1998). 台灣北部國中學生氣喘盛行率調查. [Adolescent Asthma in Northern Taiwan]. 中華公共衛生雜誌, 17(3), 214-225. doi:10.6288/cjph1998-17-03-10

宋鴻生, & 蘇喜. (2006). 精神醫療機構對健保論質計酬的態度與預期醫療服務的變化. [The Attitude and Service Changes in Institutes of Mental Health Care in Response Toquality-Based Payment System]. 醫務管理期刊, 7(4), 383-402. doi:10.6174/jhm2006.7(4).383

林文德, 謝其政, 邱尚志, 吳慧俞, & 黃一展. (2010). 以傾向分數配對法評估糖尿病論質計酬方案之成效. [An Effectiveness Evaluation of a pay-for-performance Program for Diabetes Based on the Propensity Score Matching Method]. 臺灣公共衛生雜誌, 29(1), 54-63.

徐慧君, 翁慧卿, 林育慈, 陳淑銘, 李佩儒, 李集美, . . . 李洮俊. (2004). 糖尿病患介入疾病管理在經濟面、臨床面及滿意度成效評估之初探－以南部某區域醫院糖尿病病患為例. [A Study of Economic, Clinical and Satisfaction Evaluation of the Effectiveness of Diabetes Disease Management Program-A Pilot Study of a Community Hospital]. 醫務管理期刊, 5(2), 222-242. doi:10.6174/jhm2004.5(2).222