藥物流行病學研究次級與三級資料的應用:健保用藥資料之共病測量與多項藥品安全性之比較統合分析

Use of Secondary and Tertiary Data in Pharmacoepidemiologic Studies:Pharmacy-Based Comorbidity Measure and Mixed Treatment Comparison Meta-Analysis

東雅惠(Yaa-Hui Dong)

指導教授 ： 賴美淑

共同指導教授 ： 陳建煒

國立臺灣大學/公共衛生學院/流行病學與預防醫學研究所/博士(2012年)

https://doi.org/10.6342/NTU.2012.10098

摘要

背景.
藥物流行病學研究次級與三級資料的應用越來越廣泛。在以次級資料中的健康保險申報資料庫進行的研究中，共病測量 (Comorbidity measure) 是用以描繪研究族群健康狀態相當重要的工具。在以三級資料中的臨床試驗數據進行的統合分析中，多重比較統合分析 (Mixed treatment comparison meta-analysis) 是用以整合所有試驗資料、且同時比較多種治療療效/副作用相當有效的技術。

第一部分研究之目標.
本研究欲發展一運用健保用藥資料之共病測量工具- Pharmacy-Based Disease Indicator (PBDI)，並將之與運用診斷資料之Deyo- Charlson Index作比較，以評估其在成年族群預測未來一年是否住院的能力。
材料與方法.
資料來源為自台灣健康保險申報資料庫中抽樣而得的『2005年承保抽樣歸人檔』。我們選取兩成年族群作為研究族群：在2005年1月1日大於等於18歲者 (Dataset 2005) 及在2006年1月1日大於18歲者(Dataset 2006)。我們根據慢性疾病分數(Chronic Disease Score)的架構與解剖學、治療學及化學分類系統 (Anatomical Therapeutic Chemical classification system)，發展出PBDI，其為涵括台灣37大類主要疾病所使用藥品之共病測量工具。我們以邏輯斯迴歸模式評估每個人PBDI score與未來一年是否住院的關連性。模式中的自變項包括年齡分組、性別、PBDI score 或Deyo score。
結果.
我們選取出兩成年族群，分別為697,823人 (Dataset 2005) 與714,072人 (Dataset 2006)，並建構出每個人PBDI score及Deyo score。以未來一年是否住院為研究終點，以PBDI score 或Deyo score作為預測因子，在兩成年族群中PBDI與Deyo-Charlson Index的c-statistics分別為0.715與0.691 (Dataset 2005) 及0.716 versus 0.694 (Dataset 2006)。而經由其他統計量包括Akaike Information Criterion、Bayesian Information Criterion、 Model calibration 及 Reclassification measures也驗證了PBDI的預測能力。
結論.
PBDI對於未來一年是否住院具有良好的預測能力，並可應用在含有族群資料的健康保險申報資料庫中，作為一般共病測量工具。

第二部分研究之目標.
在慢性阻塞性肺病病人，吸入性藥物使用安全性的比較資料較為闕如。因此，本研究欲針對慢性阻塞性肺病病人，比較使用不同吸入性藥物的總死亡及心因性死亡情形。
材料與方法.
本研究透過系統性資料庫搜尋，選取治療期間至少6個月的隨機對照試驗，試驗研究的治療藥物包括： Titropium Soft Mist™ Inhaler、Tiotropium HandiHaler®、Long-acting beta-2 agonists (LABA), Inhaled corticosteroids (ICS), and LABA-ICS combination。我們進行直接比較與多重比較統合分析，以估計每種藥物死亡的相對危險性，以Pooled odds ratios (ORs)表示之。
結果.
本研究總共收納42個試驗，計有52,516位病人進行分析。在多重比較統合分析Fixed effect model下，Tiotropium Soft Mist™ Inhaler無論與任何藥物比較，均有較高的總死亡相對危險性。Tiotropium Soft Mist™ Inhaler與各藥物比較的的相對危險性如下：Placebo (OR 1.51; 95% CrI: 1.06 to 2.19)、Tiotropium HandiHaler® (OR 1.65; 95% CrI 1.13 to 2.43)、LABA (OR 1.63; 95% CrI 1.10 to 2.44) 及LABA-ICS (OR 1.90; 95% CrI 1.28 to 2.86)。在心因性死亡、嚴重慢性阻塞性肺病病人及高劑量使用下，Tiotropium Soft Mist™ Inhaler呈現更高的危險性。相較之下，LABA-ICS是死亡風險最低的治療藥物。Tiotropium HandiHaler®或LABA則沒有呈現額外的風險。多重比較與直接比較統合分析所得到的結果相近，在Random effects model下其估計結果較不精確。
限制.
本研究中包含Titropium Soft Mist™ Inhaler的試驗數目較少，因此無法針對劑量-反應關係作進一步探討。
結論.
本研究提供不同種類的吸入性藥物及不同劑型的Tiotropium安全性之資料。Tiotropium Soft Mist™ Inhaler有較高的死亡風險，臨床上使用必須格外小心。

關鍵字

健康保險申報資料庫 ； 藥物流行病學 ； 健康照護研究 ； 共病測量 ； 慢性疾病分數 ； 查爾森指標 ； 多重比較統合分析 ； 慢性阻塞性肺部疾病 ； 適喘樂 ； 長效乙二型擬交感神經作用劑 ； 吸入性類固醇 ； 死亡

並列摘要

Background.
There has been enormous growth in the use of secondary and tertiary data in pharmacoepidemiologic studies and health services research. In claims-based studies using secondary data from health insurance claims databases, comorbidity measures are important tools to characterize health status of study populations. In meta-analyses using tertiary data from experimental trials, mixed treatment comparison (MTC) approach are useful techniques to integrate all available evidence in one analysis and to compare effects of various treatments efficiently.

Objective of the first study.
We aimed to develop the Pharmacy-Based Disease Indicator (PBDI) and evaluated its performance in prediction of subsequent-year hospitalization in comparison with the diagnosis-based Deyo version of Charlson Index among adult populations.
Materials and Methods.
Data source was the Longitudinal Health Insurance Database 2005 derived from the national health insurance system in Taiwan. Two adult populations were identified for study populations: individuals who were at least 18 years of age on January 1 of 2005 (Dataset 2005) and who were at least 18 years of age on January 1 of 2006 (Dataset 2006). Based on the Chronic Disease Score framework and Anatomical Therapeutic Chemical classification system, we developed the PBDI, a comorbidity measure that is a function of 37 drug categories that correspond to major diseases in Taiwan. Relationship between individuals’ PBDI score and subsequent-year hospitalization was evaluated by logistic regression models. Covariates in the models included age group, sex, the PBDI score or the Deyo score.
Results.
Using two overlapping adult populations (n=697,823 in 2005 [Dataset 2005] and n=714,072 in 2006 [Dataset 2006]), we constructed both the PBDI score and the Deyo score for each individual in each year. Using subsequent-year hospitalization as the outcome and each comorbidity measure as the predictor, we demonstrated that c-statistics of PBDI versus Deyo version of Charlson Index were 0.715 versus 0.691 and 0.716 versus 0.694 for the 2005 and 2006 populations, respectively. The Akaike Information Criterion, Bayesian Information Criterion, model calibration and reclassification measures also confirmed the utility of PBDI.
Conclusion.
The PBDI had acceptable predictive performance for subsequent-year hospitalization and can be used as a general comorbidity measure for data derived from population-based health insurance claims databases.

Objective of the second study.
The active-treatment comparative safety information for all inhaled medications in patients with chronic obstructive pulmonary disease (COPD) is limited. We aimed to compare the risk of overall and cardiovascular death for inhaled medications in patients with COPD.
Materials and Methods.
Through systematic database searching, we identified randomized controlled trials of tiotropium Soft Mist™ Inhaler, tiotropium HandiHaler®, long-acting beta-2 agonists (LABA), inhaled corticosteroids (ICS), and LABA-ICS combination with at least a 6-month treatment duration. Direct comparison and MTC meta-analyses were conducted to estimate the pooled odds ratios (ORs) of death for each comparison.
Results.
A total of 42 trials with 52,516 subjects were included in the analysis. The MTC meta-analysis with the fixed effect model indicated tiotropium Soft Mist™ Inhaler was associated with an universally increased risk of overall death compared with placebo (OR 1.51; 95% CrI: 1.06 to 2.19), tiotropium HandiHaler® (OR 1.65; 95% CrI 1.13 to 2.43), LABA (OR 1.63; 95% CrI 1.10 to 2.44), and LABA-ICS (OR 1.90; 95% CrI 1.28 to 2.86). The risk was more evident for cardiovascular death, in severe COPD patients, and at a higher daily dose. Instead, LABA-ICS was associated with the lowest risk of death among all treatments. No excess risk was noted for tiotropium HandiHaler® or LABA. Results were similar between MTC and direct comparison meta-analyses, with less precision in the random effects model.
Limitations.
The number of tiotropium Soft Mist™ Inhaler trials was limited and this precluded more exploration of the dose-response effect.
Conclusion.
Our study provided a comparative safety spectrum for each category of inhaled medications. Tiotropium Soft Mist™ Inhaler had a higher risk of mortality and should be used with caution.

並列關鍵字

health insurance claims database ； pharmacoepidemiology ； health services research ； comorbidity measures ； Chronic disease score ； Charlson Index ； mixed treatment comparison meta-analysis ； chronic obstructive pulmonary disease ； tiotropium ； long-acting beta-2 agonists ； inhaled corticosteroids ； mortality

參考文獻

44. National Health Research Institute. National Health Insurance Research Database. Available from http://w3.nhri.org.tw/nhird/en/Background.html. Accessed April 18, 2011.

1. Klungel OH, Martens EP, Psaty BM, et al. Methods to assess intended effects of drug treatment in observational studies are reviewed. J Clin Epidemiol 2004;57:1223-31.

2. Schneeweiss S, Avorn J. A review of uses of health care utilization databases for epidemiologic research on therapeutics. J Clin Epidemiol 2005;58:323-37.

3. Schneeweiss S, Gagne JJ, Glynn RJ, et al. Assessing the comparative effectiveness of newly marketed medications: methodological challenges and implications for drug development. Clin Pharmacol Ther 2011;90:777-90.

4. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. Updated March 2011. The Cochrane Collaboration, 2011. Available from http://www.cochrane-handbook.org Accessed 28 July 2011.