健保藥價調整在不同藥品市場競爭特質下對於處方型態之影響：以口服降血糖用藥長期分析

The Impact of Drug Price Adjustment on Prescribing Patterns in Different Pharmaceutical Market Competition Environment: A Longitudinal Analysis of Oral Anti-hyperglycemic Drugs

蘇浩然(Hao-Jan Su)

指導教授 ： 鄭守夏

國立臺灣大學/公共衛生學院/健康政策與管理研究所/碩士(2011年)

https://doi.org/10.6342/NTU.2011.00692

摘要

背景：歷年來台灣健保局藉由藥價調整來控制藥費的成長，但藥費仍逐年上升且約占總醫療費用的25%。除了慢性病人增加和醫藥科技發展外，過去已有許多文獻證明醫師處方行為會受到外在誘因影響而改變，但尚無探究在不同藥品市場競爭特質 (PMCCs)的醫院，其醫師開立的處方型態是否有不同變化。本研究選擇口服糖尿病藥品 (OADs)市場進行分析，探討藥價調整時，處方型態在不同PMCCs的醫院下是否有不同變化。

目的：1) 描述2002年至2009年OADs處方型態整體的變化情形，並比較不同就醫類型、PMCCs和醫院特質下的變化趨勢。2) 探討藥價調整前後不同PMCCs的醫院，其處方型態在更換藥品和處方劑量上是否有不同的變化情形。

方法：使用國家衛生研究院提供之2002年至2009年50萬人糖尿病特殊需求檔，描述病人的處方型態變化趨勢，包括平均每日處方DDD總數、平均處方DDD總數和平均處方天數之變化。第一部分研究先選取期間內每年均使用OADs之病人，描述該群病人處方型態的變化趨勢，並依就醫類型分為五組，計算每人每年之DCSI分數，分別描述每組病人之處方型態變化；再挑出就醫類型完全在醫院者，加入醫院權屬和層級別，及HHI、廠商家數和市場大小三種PMCCs後，分別描述各組病人之處方型態變化情形。第二部分研究再篩選出第3次和第5-2次藥價調整前後，各自均在同家醫院看診，且用藥組合和慢箋使用率均相同之病人，計算其處方總藥價調整幅度和處方型態變化情形，利用廣義估計方程式 (GEE)之線性模式，控制病患特質和醫院特質，分析調整前後各PMCC對於處方總藥價調整幅度和處方型態變化之影響。

結果：傳統的SUs和BGs在8年期間內處方占率呈現下降趨勢，同時TZDs, AGIs和MGs則呈現上升趨勢；在處方上以合併2種用藥最多，約占50%，單一用藥從40%減至不到30%，合併三種用藥從不到10%升至約30%；最常使用的用藥組合為BGs+SUs, SUs, BGs, BGs+SUs+TZDs和BGs+SUs+AGIs五種，其歷年總占率皆超過7成。而在8年期間內，糖尿病患DCSI分數漸增，大致上就醫類型越屬於醫院者分數越高，但混和就醫組增加速度最快；PMCCs屬於高競爭組如HHI低組、廠商多組、市場大組其每日處方DDD總數、處方DDD總數和處方天數均越多；在醫院特質上則是醫學中心>地區醫院>區域醫院，法人醫院>私立醫院>公立醫院。而第3次和第5-2次藥價調整後，整體上會換成調降幅度較少之藥品，且每日處方DDD總數、處方DDD總數和處方天數均增加。但PMCC競爭程度低組，較會換成調降幅度較少之藥品；PMCC競爭程度高組，在第3次調整後處方劑量增加幅度較大，第5-2次調整後其結果不明顯甚至反向。

結論：使用HHI、廠商家數和市場大小來推估PMCC在藥價調整前後之影響，其結果均類似，顯示確實會影響醫院醫師選擇藥品的行為和處方劑量。整體上面對藥價調整，醫院醫師均會採用更換調降成分較少的藥品和增加處方劑量，但在藥品市場競爭程度較低之醫院採用的主要方式為更換調降成分較少之藥品，而競爭程度較高之醫院則主要採用增加處方劑量，但劑量增加現象在長期後有趨緩現象，可能為劑量已達上限或藥物產生續發性失效而併用非口服治療，但是否為長期服用較高劑量所導致則無法定論。基於以上現象，健保局應在藥價調整時調查醫師處方行為的改變是否會影響病人照護結果，且監控藥品劑量增加的情形，尤其在藥品市場競爭程度較高之醫院，必須評估劑量的持續上升是否確實有達到臨床治療益處的經濟效益，以確保病患用藥安全及減少不必要的藥品花費支出。

關鍵字

口服降血糖藥物 ； 處方型態 ； 藥價調整 ； 藥品市場競爭 ； 就醫型態

並列摘要

Background
The Bureau of National Health Insurance in Taiwan uses drug price adjustment (DPA) to control the growth of pharmaceutical expenditures for years. However, they still grow up steadily and account for about 25% of total NHI medical expenditures. In addition to the increase of the patients with chronic diseases and the development of pharmaceutical technology, the association between physician's behavior and incentives beyond the clinical field has been well-established. Nevertheless, few studies focused on the impact of different pharmaceutical market competition characteristics (PMCCs) while assessing the impact of DPA on the changes of physician's prescription decisions. This study selected oral anti-hyperglycemic drugs (OADs) market to examine the relationship between PMCCs and the prescribing patterns in the period of DPA.

Objectives
1) To describe the trends of OADs prescribing pattern from 2002 to 2009 among different types of medical care institution visited, PMCCs, and hospital attributes. 2) To examine the effects of DPA on physician's prescription decision in replacing drugs or adjusting dosage while taking different PMCCs into account.

Methods
Using the NHI claims data, we identified diabetic patients taking OADs every year from 2002 to 2009, and described the trends of their prescribing patterns, including average prescribed daily DDD amount per prescription, average DDD amount per prescription, and average lengths of one prescription. In the first part of the study, the patients were divided into 5 groups (A-E) by the types of medical care institution visited and we described their trends of prescribing patterns by different DCSI scores. Furthermore, patients in group A, that patients only visited their doctors in hospitals, were selected to examine the trend of prescription patterns while considering the three PMCCs (HHI, the number of firms, and the market size) levels and hospital attribute. In the second part of the study, we selected those who visited the physicians in the same hospitals with the same combination of OADs class and utilization rate of refillable prescriptions for chronic diseases during the period of drug price adjustment in group A. Finally we examined the relationship between PMCCs and the prescribing patterns in the period of two DPAs by GEE liner model.

Results
Within the study period, SUs and BGs were prescribed decreasingly. In the meanwhile, the amount of TZDs, AGIs and MGs increased slightly. The top 5 combinations of OADs class that account for more than 70% of all prescriptions were "BGs+SUs", "SUs", "BGs", "BGs+SUs+TZDs", and "BGs+SUs+AGIs". Combination therapy was more widely prescribed than monotherapy which was decreased from 40% to less than 30%. The regimen for two OAD classes account for over 50% in all prescriptions every year and the regiment for three OAD classes ranged from less than 10% to about 30%. DCSI scores of diabetic patients were increasing in the time period and those who didn’t have the regular medical care institution type increased most quickly.

Generally speaking, if patient's medical care institution type was hospitals, the DCSI scores were higher. Moreover, we found if PMCCs were high competition, (including lower HHI, more firms, and bigger market size,) or if the hospital attributes were "medical center" or "proprietary hospital", prescribed daily DDD amount per prescription, DDD amount per prescription, and lengths of one prescription tended to be higher or longer. We also found that after DPA, the current drugs were likely to be replaced by other drugs that had lower adjustment magnitude. In addition, the dosage and lengths of one prescription tended to be higher or longer. Besides, in the low competition groups of PMCCs, the physicians replaced the drugs more frequently. However, in the high competition groups of PMCCs, the patient's prescription dosage was increased more at the 3rd DPA, but not at the 5th DPA.

Conclusions
This study revealed that pharmaceutical competition level was associated with physician's prescribing behavior. In general, when facing the drug price adjustment, the responding strategies of the hospitals were replacing the current drugs by other drugs that had lower adjustment magnitude and increasing the dosage. In addition, the changes of physician's prescribing behavior differ when the PMCCs were considered. Based on the findings, the Bureau of National Health Insurance in Taiwan should investigate if the change of prescription patterns might result in poor treatment outcome after DPA. On top of that, they should monitor the increase of the dosage, especially in the hospitals with high pharmaceutical competition level, and evaluate the cost-effectiveness of higher dosage to ensure drug safety and prevent the unnecessary pharmaceutical expenditures.

並列關鍵字

Oral anti-hyperglycemic drugs ； Prescribing patterns ； Drug price adjustment ； Pharmaceutical market competition ； Type of medical care institution visited

參考文獻

江瑞坤，陳政男，陳欣欣(2008)。新糖尿病用藥－incretin類似物、DPP-4抑制劑及其他。基層醫學，23(5)，137-140。

林于婷(2010)。健保藥價調整與醫師處方行為之相關研究：以糖尿病用藥為例。國立台灣大學公共衛生學院衛生政策與管理研究所碩士論文。

林世崇(2010)。糖尿病治療的新策略。臺灣醫界，53(9)，458-465。

洪乙禎(2007)。健保體系下藥品費用分攤制度的分析。人文及社會科學集刊，19(4)，473-504。

高淑真，李玉春，黃文鴻，李龍騰(2006)。全民健保糖尿病門診問題處方之分析一以北台灣為中心之研究。台灣衛誌，25(1)，58-64。