White Kamila S, Allen Laura B, Barlow David H, Gorman Jack M, Shear M Katherine, Woods Scott W
Department of Psychology, University of Missouri-Saint Louis, Saint Louis, MO 63121, USA.
J Nerv Ment Dis. 2010 Sep;198(9):665-71. doi: 10.1097/NMD.0b013e3181ef3627.
This study examined attrition in a multisite clinical trial for panic disorder. Of 379 eligible patients, 19 refused treatment (5% Refusal Rate), 104 discontinued treatment prematurely (19% Dropout Rate) or were withdrawn by the investigators (8% Withdrawal Rate), and 256 completed the treatment (68% Completion Rate). Logistic regression was used to examine 5 domains theorized to be related to attrition (e.g., diagnostic severity, treatment factors). Few variables were associated with increased odds of attrition at padj < 0.004. Younger age was the only independent predictor of attrition in the demographic factor model. Diagnostic severity and comorbidity, panic disorder symptom severity, treatment factors, and therapist factors were unrelated to study attrition. Patient dropout was highest after treatment sessions that targeted interoceptive and situational exposure exercises. Findings suggest that attrition may not strongly threaten the validity of results from treatment outcome studies.
本研究考察了惊恐障碍多中心临床试验中的失访情况。在379名符合条件的患者中，19人拒绝治疗（拒绝率5%），104人过早中断治疗（脱落率19%）或被研究者撤组（撤组率8%），256人完成了治疗（完成率68%）。采用逻辑回归分析来检验理论上与失访相关的5个领域（如诊断严重程度、治疗因素）。在p校正<0.004时，几乎没有变量与失访几率增加相关。在人口统计学因素模型中，年龄较小是失访的唯一独立预测因素。诊断严重程度与共病、惊恐障碍症状严重程度、治疗因素及治疗师因素均与研究失访无关。在针对内感受性和情境暴露练习的治疗阶段后，患者脱落率最高。研究结果表明，失访可能不会对治疗结果研究的结果效度构成严重威胁。
