緩起訴甲基安非他命戒癮治療參與者遵從門診治療之相關因素

Abstract

目的：本研究針對參加北部某區域醫院緩起訴戒癮治療的甲基安非他命使用疾患者，藉由比較退出治療組與完成一年期治療組來找出與治療遵從性相關的因子；本研究也試圖運用與治療遵從性相關的因子，發展出一套簡易的計分系統，協助臨床工作者用以預測容易提早退出治療的危險族群。 方法：收案對象為2011年8月至2012年7月期間因為使用甲基安非他命而由司法機關附命至北部某區域醫院參加緩起訴戒癮治療的成年甲基安非他命濫用或依賴個案，本研究採取病歷回顧方法，回朔性擷取個案資料。研究中主要的應變項為提早退出治療或是完成一年期治療，自變項包括治療院區、社會人口統計學變項、甲基安非他命使用相關變項、初診時精神共病變項與初診實驗室檢驗變項，之後運用邏輯斯回歸分析找出與治療遵從性相關的因子。再進一步利用Cox比例風險回歸模型換算相關因子的風險分數，藉以發展出一套遵從性計分系統，並以受試者工作特性曲線（Receiver Operator Characteristic curves）來檢定其預測能力。 結果：收案的221位個案中，139位完成一年期的緩起訴戒癮治療（62.9%），平均治療存留月數是9.6月（標準差：3.7月）。婚姻狀態、來院距離、甲基安非他命依賴、甲基安非他命初用年齡、甲基安非他命終身使用年數、初診前一個月使用天數與初診驗尿報告等變相與完成治療的遵從性有相關。Cox比例風險回歸模型分析後，六個相關變項被選入遵從性計分系統，建立的計分系統敏感度82.61%、特異度69.54%，受試者工作特性曲線的曲線下面積為0.79。計分系統區分出的高危險群與低危險群以Kaplan-Meier存活曲線分析可以達到顯著差異（p < 0.0001）。 結論：本研究中發現若干與甲基安非他命緩起訴戒癮治療遵從性相關的因子，並運用相關因子建立一簡易計算而能有效預測治療遵從性的計分系統，藉以提供臨床工作，早期針對容易退出治療的高風險族群進行介入，以提高治療成功性。

Objectives: The objectives of this study were to compare between cases that dropped out from or completed a judicially supervised outpatient treatment programs for methamphetamine use disorder and to identify associated factors of treatment compliance. This investigation also utilized the associated factors to develop a risk scoring model that can help clinicians predict treatment compliance. Methods: Subjects in this study were adults with the diagnosis of methamphetamine use disorder participated in a judicially supervised outpatient drug abuse treatment in New Taipei City Hospital during August 2011 to July 2013. Using a retrospective design, the investigation collected data entirely from medical charts. The major dependent variable was one-year treatment retention. The independent variables selected as potential predictors including sociodemographic data, factors related to methamphetamine use and factors related to psychiatric comorbidity and laboratory results at first visit. Logistic regression analysis was used to define the associated factors with treatment dropout. Then multivariate Cox’s proportional hazards analysis was performed to calculate the risk scores and develop a predictive scoring method. Finally the scoring method was evaluated based on Receiver Operator Characteristic (ROC) curves with leave-one-out cross-validation. Results: There were 139 individuals (62.9%) who completed the whole treatment course among the 221 participants. The mean value for treatment retention time was 9.6 months (S.D. = 3.71 months). Marital status, distance from residence to hospital and onset age of MA use, MA dependence, lifetime MA use years, use days during the last month preceding first visit and urinalysis result at first visit had significant difference between dropout and completion groups. By multivariate Cox’s proportional hazards analysis, six variables were selected into developing the predictive scoring method. This scoring method has fair predictability with sensitivity 82.61% and specificity 69.54%. The area under the curves (AUC) of ROC was 0.79. Using the scoring model we divide subjects into lower risk group (0 to 7 points) and higher risk group (8 to 10 points) for treatment dropout. Kaplan-Meier survival analysis showed significant difference between these two groups (p < 0.0001). Conclusion: Several factors were associated with the compliance in the judicially supervised outpatient methamphetamine abuse treatment program. A simple risk scoring method was developed to predict compliance of outpatient MA abuse treatment that can remind clinicians to take strategies enhancing compliance for the higher risk group of treatment dropout as earlier as possible.