在心理咨询与治疗的效果研究领域, 研究者普遍认为存在剂量效应(dose-effect), 即来访者必须要坚持参与治疗并达到足够多的次数之后才有可能发生积极的改变(Fernández-Arias et al., 2016; Gearing, Townsend, Elkins, El-Bassel, & Osterberg, 2014; Hansen, Lambert, & Forman, 2002)。这表明足够次数的治疗是治疗成功的必要前提。但在临床实践中, 普遍存在来访者由于种种原因在心理困扰或问题尚未解决之前就终止治疗的现象, 即从治疗中脱落(Swift & Greenberg, 2012)。很多研究表明, 在实际的心理咨询和治疗过程中, 脱落的发生率很高。例如, Wierzbicki和Pekarik (1993)的元分析研究显示, 约46.8%的来访者会从治疗中脱落; 还有研究发现, 多个公立心理治疗临床中心的来访者脱落率超过50% (Garfield, 1994; Lee, 2015), 甚至有研究发现心理咨询与治疗过程中的脱落率高达50%~70% (Olver, Stockdale, & Wormith, 2011)。
脱落通常被看作治疗中的显著负性治疗事件(negative therapy event), 它会给来访者、治疗师、服务机构乃至全社会造成消极影响。例如, 研究发现, 与坚持治疗的来访者相比, 脱落者的疗效更差, 对治疗更不满意, 无望感更强(Berke et al., 2019; Knox et al., 2011); 脱落者的家庭、朋友、同事等也承受着由其症状和损伤所带来的负担(Swift, Greenberg, Whipple, & Kominiak, 2012); 治疗师会因来访者脱落而产生挫败感, 感觉被来访者拒绝或抛弃; 服务机构也因脱落而面临收入减少和时间成本增加的问题(Klein, Stone, Hicks, & Pritchard, 2003; Piselli, Halgin, & MacEwan, 2011)。
由于来访者脱落的普遍性以及脱落对于治疗效果、治疗师、服务机构等各方面的消极影响, 过去60余年来国外研究者对脱落的发生率、预测变量及应对策略等方面进行了大量研究。近年来, 脱落研究的重点也发生了一些新变化, 包括逐渐从静态的影响因素研究转向动态的作用机制研究, 更关注脱落的研究设计问题以及重新看待脱落的消极影响等。而国内对脱落的研究尚处于起步阶段, 鲜有对脱落研究的文献梳理和理论思考。在中国文化情境下开展来访者脱落的相关研究, 有助于我们深入认识这一现象, 理解中国文化情境下脱落对来访者、治疗师各方面的消极影响以及脱落的形成原因, 从而为找到降低脱落率的方法做好理论上的准备, 并为相关的教育和服务机构工作的有效开展提供依据和建议。基于此, 本文拟对来访者脱落的定义、发生率、预测变量以及预防策略进行梳理, 并对未来研究进行展望, 以求为后续研究提供参考和启示。
1 脱落的定义
1.1 脱落的概念内涵
和脱落对应的英文词汇有多个, 其中包括dropout, premature discontinuation, premature termination, unilateral termination, attrition, early withdraw, therapeutic abandonment等。不同的词汇反映了研究者对脱落的不同理解。比如, premature discontinuation和premature termination强调脱落为一种不成熟的治疗中断行为, unilateral termination则突出脱落为未与治疗师达成一致的单方面退出, 而withdraw单纯表达来访者从治疗中主动退出。研究者对于以上哪个词汇最能表达脱落概念这一点上存在争议, 但对于脱落的概念内涵有着共识性的看法, 即认为来访者开始了治疗, 但在导致他前来寻求治疗的问题(包括症状、功能损伤、痛苦等)没有解决之前, 来访者就在没有获得治疗师同意或没有与治疗师商量的情况下中断了治疗(Garfield, 1994; Swift & Greenberg, 2012)。
1.2 脱落的操作性定义
尽管研究者都同意将脱落看作对治疗性干预的不成熟的中断, 但在对脱落下操作性定义的方法上存在着显著差异。下面介绍几种被研究者广泛使用的脱落的操作性定义。脱落的第一种操作性定义采用“特定剂量” (specified dose)法, 认为来访者没有完成特定次数的会谈即为脱落。这种操作性定义基于剂量效应理念, 认为当事人要获得治疗性改变就必须达到最低会谈次数(Baekeland & Lundwall, 1975; Hansen et al., 2002)。“特定剂量”法在早期研究中使用较多, 也往往以真实临床工作中的来访者为研究被试。“特定剂量”标准只要确定好最低会谈次数, 就可以客观判断来访者是否脱落, 易于操作。但不同研究者对于最低会谈次数的设定标准存在争议。Lowry和Ross (1997)调查了治疗师主观上认为治好某类症状需要的最低会谈次数, 结果差异非常大。同时, 来访者的个体差异也非常大, 有的来访者可能通过几次会谈就能解决问题, 而有的来访者可能会谈100次仍不能解决问题(Swift & Greenberg, 2012)。脱落的第二种操作性定义以“来访者爽约” (missed last appointment)为标准。如果来访者在会谈结束时同意继续治疗, 但下一次会谈时既没有出现, 也没有提前取消, 即可视为脱落(Erbes, Curry, & Leskela, 2009)。“来访者爽约”标准具有较高的信度和表面效度(Wierzbicki & Pekarik, 1993), 但可能存在着来访者主动告知治疗师终止治疗可自身问题并没有得到解决的现象, 而这种不成熟的终止治疗行为在“来访者爽约”标准下却不被视为脱落, 这显然和脱落的内涵相悖(Hatchett & Park, 2003; Pekarik, 1985)。第三种操作性定义以来访者“未完成治疗协议” (noncompletion of the treatment protocol)为标准。这种操作性定义在近20年来被大量地运用在效力研究(efficacy studies)中, 研究者往往招募特定的患者为被试, 设定某种治疗方法和临床设置, 并签订治疗协议, 中途退出者即被视为脱落者(Southam-Gerow, Weisz, & Kendall, 2003)。“未完成治疗协议”方法中治疗次数的设定往往基于剂量效应, 因而它的利弊与“特定剂量”标准的利弊非常相似, 主要的不同之处在于, “特定剂量”标准往往应用于真实临床情境中, 而“未完成治疗协议”标准往往应用于实验设计类的心理治疗研究中。脱落的第四种操作性定义为“治疗师评定”法(therapist judgment), 即由治疗师来判断自己的来访者是否脱落(Hatchett & Park, 2003), 来访者未达到事先设定的治疗目标就终止治疗, 即被视为脱落。该方法近年来得到越来越多的应用。基于对来访者的熟知和对治疗过程与结果的了解, 治疗师对于自己的来访者是否属于脱落有着更为恰当的认识和判断(Pekarik, 1985; Swift & Greenberg, 2012), 因而“治疗师评定”这个方法被认为更可靠, 与其他操作方法相比具有更高的一致性(Pekarik, 1985)。但这种依赖于临床的判定方法可能也会存在偏差, 因为来访者和治疗师对于治疗目标和所需会谈次数的期望可能不一致, 治疗师的主观判定也可能会有明显误判(Garb, 2005; Garfield, 1994)。此外, 还有“临床显著变化”、“平均分割程序”、“单方面终止”等界定脱落的方法(Baekeland & Lundwall, 1975; Richmond, 1992), 但这些操作性定义在近20年来研究中应用较少, 在此不作详述。
脱落的以上几种操作性定义都在某一个面向上与脱落的内涵相符, 但并没有哪一种操作性定义得到研究者的一致认可。研究者往往根据自己的研究目的或研究设计而采用不同的操作性定义。在研究的实际实施过程中, 脱落的以上操作性定义也有重合之处。例如, 进行脱落元分析的研究者往往将“特定剂量”标准和“未完成治疗协议”标准合并为“预设会谈次数”, 因为它们都事先设定会谈次数, 有客观的次数指标(de Haan, Boon, de Jong, Hoeve, & Vermeiren, 2013; Warnick, Gonzalez, Weersing, Scahill, & Woolston, 2012); 也有研究者将“单方面终止”、“显著临床改变”等标准纳入“治疗师判定”标准中, 因为是否发生了单方面终止或是否有显著临床改变, 往往由治疗师来判断。在脱落研究中, 脱落的操作性定义决定着脱落的具体测量方式和测量程序。对同一被试群体的心理治疗脱落率进行测量时, 就可能因为所采用操作性定义的不同而导致结果的差异。因此, 研究者在脱落的操作性定义使用上没有达成共识, 这使得操作性定义本身影响着脱落的发生率和脱落前因变量的预测力, 也导致研究结果在跨研究间难以进行有意义的比较(Dunn, 2008)。这将在下文详细阐述。
2 脱落的发生率
已有的脱落研究主要探究了脱落的发生率和预测变量。尽管很多研究报告了较高的脱落率, 但不同研究得出的脱落率差异很大, 甚至出现相互矛盾的结果。de Haan等人(2013)综合分析了48项儿童青少年心理治疗脱落的研究后发现, 脱落率最低为16%, 最高为75%, 波动幅度近60%。有研究者考察了患焦虑症的儿童青少年的心理治疗脱落现象, 发现脱落率在14.4%~51.3%之间波动(Gonzalez, Weersing, Warnick, Scahill, & Woolston, 2011; Wergeland et al., 2015)。还有研究发现焦虑障碍患者的脱落率要高于抑郁症患者(Lamers et al., 2012); 但另一项研究显示焦虑症患者的脱落率又低于包括抑郁症在内的其他心理障碍患者(Bados, laguer, & Saldana, 2007)。考虑到单个研究之间脱落率的巨大差异, 有研究者综合性别、治疗形式、疾病类型等因素采用元分析方法得出了平均脱落率。但不同元分析得出的结果也存在明显差异。例如, 以成人的心理治疗脱落为例, Wierzbicki和Pekarik (1993)的元分析研究涉及到125篇心理治疗脱落文献, 发现成人的平均脱落率为47%, 而Swift和Greenberg (2012)开展的涵盖了669项心理治疗脱落研究的元分析显示, 成人的平均脱落率为19.7%。此外, 在脱落的元分析研究中, 研究者通常采用研究样本中脱落率的中位数(median)或加权平均值(weighted average)来表示平均脱落率, 但由于方差较大, 所得出的结果只是综合了不同研究结果后的平均值, 抹去了单个研究的特征, 难以推论到具体的单个研究中。脱落率之所以在不同研究间存在巨大差异, 是因为脱落率极易受患者群体特征、治疗师特征、治疗设置等各种因素的影响。实际上, 以往研究在以上各个因素上不尽相同, 而每个因素的差异都可能引起脱落率的波动, 这导致了研究者在探讨某个群体或某种治疗方法的脱落率时难以得出具有跨研究一致性的结论。
在影响脱落率的众多因素中, 操作性定义和研究设计类型是比较特殊的研究设计方面的变量。关于脱落的每个实证研究均会采用某种研究设计类型, 也都需要选用一种测量脱落率的操作性定义, 它们属于研究程序的一部分, 和来访者脱落没有必然的因果关系, 但会直接影响到脱落率的测量结果。研究者认为这两个变量可能是造成脱落率在不同研究间存在巨大差异的最重要的原因(de Haan et al., 2013)。
2.1 脱落的操作性定义对脱落率的影响
Warnick等人(2012)曾使用3种操作性定义对一家儿童青少年心理门诊4年间的1098个家庭案例进行研究, 结果发现, 采用“来访者爽约”标准得出的脱落率为56.6%, 采用“治疗师判定”标准得出的脱落率为63.1%, 而采用“特定剂量”标准得出的脱落率为88.1%。还有研究者采用元分析方法探讨了脱落的操作性定义对脱落率的影响。综合不同时期的元分析研究结果来看, 以“来访者爽约”为标准得出的脱落率通常比较低, 以“治疗师判定”为标准得出的脱落率通常很高, 而采用“特定剂量”和“未完成治疗协议”这两个标准得出的脱落率很不稳定, 有时高于、有时又低于其他标准下得出的脱落率(Swift & Greenberg, 2012; Warnick et al., 2012; Wierzbicki & Pekarik, 1993)。究其原因, 可能正如前文所述, “来访者爽约”主要测量的是回避与治疗师讨论结案而单方面终止的情况, 排除了其他许多临床状况(Hatchett & Park, 2003; Pekarik, 1985), 因而测量出的脱落率偏低。而“治疗师判定”法通常被更符合真实临床情况的效果研究(effectiveness studies)所采用, 涵盖了更多不成熟结案的临床情况, 因而测量出的脱落率较高。而“特定剂量”或“未完成治疗协议”这两种标准均涉及到事先主观规定的会谈次数, 其脱落率容易受规定次数的多少、被试群体特征等因素的影响而出现波动。
笔者认为, 在符合脱落的内涵前提下, 操作性定义所涵盖的脱落的临床情况越多, 采用该操作性定义的研究所测量到的脱落率就会越高。从这个角度来讲, “来访者爽约”标准直接以一种脱落的具体临床情况来定义脱落, 所以该操作性定义下的脱落率就会相对较低; 而“治疗师判定”标准涵盖了更多可能的脱落的临床情况, 所得脱落率相对会更高。“特定剂量”标准和“未完成治疗协议”标准可能看上去容易操作, 但存在的问题是不能准确地反映脱落内涵(Hatchett & Park, 2003)。比如, 有的来访者于预先约定的次数之前解决了心理困扰而决定终止治疗, 但这种情况会被判为脱落; 也有的来访者虽然完成了预定次数的会谈, 但并没有解决最初来就诊时的心理困扰, 只是因为会谈次数不够多而没有机会脱落。因而有研究者认为, 采用预先设定次数来定义的这两个操作性定义测量的脱落并非真实的脱落, 并使得脱落的研究结果复杂化(Pekarik, 1985)。
2.2 脱落研究设计类型对脱落率的影响
脱落领域的研究设计通常分为效力研究和效果研究。效力研究通常采用实验设计法。为了得出在控制了各种条件后的脱落率或脱落影响因素的理想结果, 研究者事先设定治疗方案和协议, 随后招募某一特定疾病的患者, 采用随机对照实验设计; 研究过程中遵循统一的治疗方法和治疗时长, 由特定治疗师开展治疗(La Greca, Silverman, & Lochman, 2009; Lambert & Ogles, 2004; Southam- Gerow et al., 2003), 其内部效度较高。而效果研究发生在真实的临床设置中, 它以现实中的门诊或住院病人为被试, 对治疗方法和治疗过程不采取或很少采取干预措施, 数据的收集可能需要长达数年, 因而效果研究更接近真实的临床情境, 它的外部效度更高(Lambert & Ogles, 2004)。不少研究发现, 效力研究中的脱落率要低于效果研究中的脱落率。在de Haan等人(2013)的元分析研究中, 效力研究(17篇)中的平均脱落率为28.4%, 效果研究(30篇)中的平均脱落率为50%。在Swift和Greenberg (2012)的元分析研究中, 效力研究(398篇)的平均脱落率为17.0%, 效果研究(235篇)的平均脱落率为26.0%。Swift和Greenberg认为, 近10年来的元分析研究得出的平均脱落率远低于2000年以前的元分析研究中的平均脱落率, 可能是因为近20年来出现了越来越多的效力研究, 导致整体平均脱落率的降低。
如何理解效力研究中脱落率更低?研究者认为, 效力研究中脱落率更低可能既和该研究设计中所采用的脱落操作性定义有关, 也和效力研究的设计特征有关。效力研究有一套严格而清楚的治疗框架和时长, 相比于效果研究中的不确定感, 效力研究中的来访者有更清楚的期待和确定感, 他们可以预期发生什么, 因而更不容易脱落; 再者, 效力研究中的患者被试通过了严格筛选后才进入治疗, 因此也有更强的动力来完成整个治疗过程(de Haan et al., 2013)。
实际上, 脱落的操作性定义和研究设计对脱落率的影响可能比上述分析更加复杂, 在得出概括性结论的过程中须慎之又慎。研究者认为不同研究中得出的脱落率的差异可能说明脱落率受多种潜在调节变量和协变量的影响, 同时也说明平均脱落率可能并非评估所有脱落研究最合适的指标(Fernández-Arias et al., 2016; Swift & Greenberg, 2012; Wierzbicki & Pekarik, 1993)。
3 脱落的预测变量
由于脱落对治疗效果的重要影响, 脱落的预测变量或影响因素研究一直以来都是心理治疗领域研究的热点。影响脱落的变量大体上可以分为来访者变量、治疗师变量、临床机构变量、治疗设置变量、治疗过程变量五类(唐小娟, 2015; Swift & Greenberg, 2012), 下面分别总结。
3.1 不同类别的变量对脱落的预测力
来访者变量在脱落预测变量研究中的研究成果最为丰富。影响脱落的来访者变量主要包括性别、年龄、受教育水平等人口统计学变量, 人格、社会支持等心理特征变量以及问题类型、严重程度等临床特征变量。有关来访者人口统计学变量对脱落预测力的研究结果很不一致。以年龄为例, 一些研究发现年轻的来访者更容易脱落(Lamers et al., 2012; Schneibel et al., 2017), 而另一些研究发现年长的来访者比年轻的来访者脱落率更高(Hamberger, Lohr, & Gottlieb, 2000), 但更多研究发现, 包括年龄在内的各种人口学变量(比如性别、种族、受教育水平、收入等)对脱落均没有预测作用(de Haan et al., 2013; Fernández-Arias et al., 2016; Wergeland et al., 2015)。Swift和Greenberg (2012)对669项研究中83834名来访者的脱落进行元分析研究发现, 在众多预测来访者脱落的相关变量中, 只有来访者疾病诊断和年龄两个变量能一致性地预测脱落。具体而言, 没有被划分为具体的疾病类型的来访者以及患有人格障碍或进食障碍的来访者相比于患其他疾病的来访者更容易从治疗中脱落, 年轻的来访者更容易从治疗中脱落。来访者种族、性别、婚姻状态、是否就业等人口统计学变量均不能一致性地预测脱落。来访者相关的很多变量对脱落的预测力在跨研究间的一致性很低, 这导致某种情境中脱落的预测变量被用于理解和解释其他情境中的脱落现象时受到质疑(de Haan et al., 2013; Warnick et al., 2012; Wierzbicki & Pekarik, 1993)。
相比于来访者变量, 与治疗师相关的预测变量没有得到充分研究和重视(Zimmermann, Rubel, Page, & Lutz, 2017)。Swift和Greenberg (2012)元分析发现, 治疗师的年龄、性别、种族这3个人口学变量对脱落没有显著预测力。但研究发现, 与治疗师能力相关的变量, 如治疗师的卷入程度、经验水平、对反移情的处理等都对脱落有负向预测力, 并且这些变量的研究结果更具有跨研究间的一致性(杨雪岭 等, 2017; Chernus, 2015; Rodolfa, Rapaport, & Lee, 1983; Zimmermann et al., 2017)。影响脱落的临床机构变量在早期研究中得到比较多的关注, 包括预约排队的时间、临床机构性质、接待人员态度、治疗前告知等。其中排队预约时间对脱落的影响作用的研究结论并不一致(Rodolfa et al., 1983; Warnick et al., 2012), 而治疗前告知和治疗师提醒(初次咨询前提醒来访者)可以有效降低来访者的预约未到率(Greenfield, 1983; Mennicake, Lent, & Burgoyne, 1998)。但临床机构变量主要对来访者的早期脱落产生影响, 随着治疗深入, 治疗师变量等其他变量对来访者脱落的影响作用开始变大(Rodolfa et al., 1983)。治疗设置变量包括治疗的理论取向、治疗形式、治疗时长、治疗场所等。心理治疗需要在一定设置内进行, 比如在某种类型的治疗场所采用某种治疗取向进行某种形式的心理治疗。但目前绝大多数实证研究并未考察自身的治疗设置变量对脱落的影响, 这就需要对不同研究的治疗设置进行比较分析以考察其对脱落的影响, 因而治疗设置变量对脱落的预测力探索通常见于元分析研究中。Swift和Greenberg (2012)的元分析发现, 相比于没有确切的治疗时长、无结构的治疗方案, 明确的治疗时长、操作化的治疗方案能显著地减少来访者脱落; 相较于医院门诊、私人执业、其他临床研究中心而言, 大学附设的心理诊所(包括心理学院临床实习诊所和大学心理咨询中心)脱落率更高; 治疗取向(认知行为取向、心理动力学取向、焦点解决取向、整合取向等)与治疗形式(个体治疗、团体治疗、伴侣或家庭治疗)对来访者脱落并不具有显著预测力。
以上提到的来访者与治疗师的人口统计学变量, 临床机构变量, 治疗形式、时长、场所等治疗设置变量等绝大多数为静态的变量, 它们在治疗开始前已经存在, 并在治疗过程中基本保持稳定。识别出这些能够预测脱落的静态的变量具有临床意义, 它们能提示治疗师在何种情境下来访者更易脱落, 便于治疗师做出预判和采取相应的应对措施。
3.2 脱落的静态预测变量的局限与治疗过程变量研究的兴起
尽管识别出脱落的静态预测变量具有重要意义, 但这些变量本身并不能解释为什么来访者选择脱落。脱落的静态预测变量存在两个局限。其一, 如前文所述, 很多静态的预测变量对脱落的影响作用不能在不同研究间都得到验证, 甚至还会出现相互矛盾的结果(Krishnamurthy, Khare, Klenck, & Norton, 2015)。其二, 尽管不少来访者人口统计学变量、临床变量等静态预测变量与脱落存在相关, 但这些变量并不能为我们了解来访者不能完成治疗的真正原因提供深入的信息(Armbruster & Kazdin, 1994; de Haan et al., 2013)。例如, 虽然研究发现来访者的种族身份有助于预测谁更可能从治疗中脱落, 但它并非导致来访者脱落的直接原因, 而是可能间接提供了脱落原因的动态解释——由于治疗师没有在来访者的文化背景中概念化地理解治疗而让来访者对治疗产生不满, 从而导致来访者脱落(Garfield, 1994; Krishnamurthy et al., 2015)。由此可见, 如果只识别出与来访者脱落相关的静态变量而没有研究其潜在的过程机制, 并不能有效增加我们对脱落的深入理解。
基于上述有关脱落的静态的预测变量的两个局限, 越来越多的研究者呼吁将关注点放在对脱落起作用的因素及其作用机制上, 比如治疗关系、来访者治疗动机等(Armbruster & Kazdin, 1994; Sly, 2009; Zaitsoff, Pullmer, Cyr, & Aime, 2015)。在心理咨询与治疗的“过程研究”领域, 研究者已经发现求助动机、问题严重程度等来访者变量, 治疗关系、移情与反移情等治疗共同因素以及各个流派中特定的理念和干预技术等都会对治疗效果起作用(夏勉, 江光荣, 2005; Kazdin, 2007)。而且这类变量通常是动态的, 在治疗过程中可能发生变化, 能够进行监测和干预以提升心理治疗效果。这些和治疗过程直接相关的动态变量同样会直接作用于来访者, 影响脱落。在脱落的研究领域里, 有学者将这类变量称之为与治疗相关的共同因素(Common Factors) (Sly, 2009), 也有学者称之为与过程相关的变量(Process Related Variables) (de Haan et al., 2013; Jordan et al., 2017; Warnick et al., 2012)。综前人所述, 我们将这类变量称为治疗过程变量, 并将其界定为“与治疗过程直接相关的、对脱落起作用的动态变量”。
众多研究证实, 治疗过程变量对来访者脱落具有更强的预测力和跨研究间结论的一致性(de Haan et al., 2013; Fernández-Arias et al., 2016; Jordan et al., 2017; ter Huurne, Postel, de Haan, van der Palen, & DeJong, 2017)。常见的治疗过程变量有治疗同盟(therapeutic alliance)、来访者的治疗信念(treatment belief)、治疗可信度(treatment credibility)和治疗动机(treatment motivation)等。治疗同盟是研究者最为关注的过程变量, 它对于脱落具有跨治疗形式(个体治疗、家庭治疗)和跨疾病类型(抑郁患者、酒精依赖患者等)的稳定一致的负向预测力(DeRubeis, Gelfand, German, Fournier, & Forand, 2014; Meier, Donmall, Mcelduff, Barrowclough, & Heller, 2006; Yoo, Bartle-Haring, & Gangamma, 2016)。良好的治疗同盟让来访者感到被理解和关心, 增加了来访者改变的信心和动力, 从而促进来访者持续参与治疗, 降低脱落风险(Sly, Morgan, Mountford, & Lacey, 2013)。来访者治疗信念是指来访者有关治疗时长、改变过程等方面的预期和观念(Dunn, 2008)。研究发现, 来访者往往并不是依据“治疗前设定的次数或治愈所需要达到的次数”, 而是依据自己预期的治疗次数来决定是否要结束治疗(Ciarlo, 1979; Jordan et al., 2017)。治疗可信度是指来访者在了解了治疗的理论依据和方案内容后, 对治疗的可靠性与自己从中受益的可能性的感知(Kazdin, Holland, & Crowley, 1997)。当来访者觉得治疗不可靠时, 他们更少投入和坚持治疗。有回归分析发现, 治疗可信度这一因素能够解释脱落原因总体变异(total dropout odds)的22% (Wergeland et al., 2015)。治疗动机是指来访者承认自己的问题或困扰并愿意通过心理治疗来改变, 多项研究发现治疗动机能显著地负向预测脱落(Keijsers, Schaap, Hoogduin, Hoogsteyns, & de Kemp, 1999; Wergeland et al., 2015)。当然, 治疗过程变量远不止上述变量。Kazdin等人(1997)在研究儿童青少年心理治疗脱落时提出了“治疗坚持障碍”模型(Barriers to Treatment Participation Model), 尝试整合治疗过程中对脱落起作用的各类变量, 系统地考察其对脱落的影响。治疗坚持障碍是指儿童及其家庭在参与治疗的过程中所经历的各种可能会增加脱落风险的阻碍, 包括治疗过程中的压力源、来访者的治疗动机和治疗可信度、咨访关系、重要临床事件等。近年来研究者运用《治疗坚持障碍量表》来考察治疗坚持障碍对脱落的影响, 结果发现相比于众多静态预测变量, 治疗坚持障碍更能一致地、显著地预测脱落(de Haan et al., 2013; Warnick et al., 2012; Wergeland et al., 2015)。
这些研究给了临床实践宝贵的启示, 相比于在治疗过程中无法改变的静态预测变量, 识别出这些在治疗过程中能够改变的、预测作用更一致和明确的治疗过程变量并在治疗中加以监控和干预, 对于提高治疗的持续性和最终效果具有更重要的意义(de Haan et al., 2013; Wergeland et al., 2015)。
4 减少脱落的策略
识别出能够预测脱落的变量之后, 如何根据这些有效的预测变量发展出减少脱落的干预策略, 是研究者和临床工作者需要思考的重要问题。Ogrodniczuk, Joyce和Piper (2005)回顾了39篇有关降低脱落的策略的文献后, 总结提出了6条减少脱落的建议：(1)筛选来访者, 为特定类型的治疗选择合适的来访者; (2)就如何当来访者进行角色教育(role induction)访谈; (3)与来访者就治疗时长制定契约; (4)每次治疗前给来访者发送提醒; (5)整个过程中密切关注治疗同盟; (6)创建安全的氛围以促进来访者对负性情绪和正性情绪的探索。但以上一些干预策略的有效性存在异议。比如心理动力学取向治疗师并不认可“治疗前给来访者发送提醒”的做法, 因为他们认为出席会谈是来访者的责任, 同时缺席或忘记会谈本身也是有价值的分析材料(Reis & Brown, 1999)。还有研究者认为, “将来访者分配到合适的治疗”策略在现实中实行起来并不容易, 因为进行“准确的来访者筛选”本身就是一项颇为复杂的高难度技能(Malan, 1979)。Swift等人(2012)认为, 坚持参加心理治疗对很多来访者而言是一项困难的任务, 治疗费用、时间、治疗过程中的自我暴露和痛苦处理等都可能让来访者难以忍受, 因此来访者必须相信坚持治疗的获益要大于成本, 才能提高对治疗的投入。基于这个观点, 他们也总结了6条降低脱落的策略：(1)为来访者提供治疗持续时间和变化规律的教育; (2)就如何当来访者进行角色教育; (3)考虑来访者的偏好; (4)在治疗早期加强来访者获益的希望; (5)促进治疗同盟; (6)和来访者一起评估、讨论治疗进展。
以上两篇研究综述的建议存在一些相似之处, 并且多项策略与治疗过程变量(如治疗同盟、来访者动机、症状严重程度等)密切相关。尽管不同研究者提出了不同的策略, 但大多数研究者都很认同为来访者提供治疗前准备(pretherapy preparation)以及对治疗中的一些重要变量进行监测。治疗前准备是指为来访者提供关于心理治疗的本质和理论依据、来访者的角色与义务以及治疗中可能出现的困难等方面的信息, 从而澄清来访者对治疗的误解, 减少来访者与治疗师对治疗期待的不一致, 以提高来访者的治疗准备(Ogrodniczuk et al., 2005)。不少研究发现, 来访者对于治疗时长、治疗师的作用、治疗效果等抱有的不切实际的预期会显著增加脱落, 因此治疗师应了解来访者的期待并进行心理教育(赵宏祥, 江光荣, 朱旭, 2014; de Haan et al., 2013; Thylstrup & Hesse, 2016; Zimmermann et al., 2017)。在治疗过程中, 对来访者投入程度、治疗同盟、治疗进展等进行监控和干预能有效降低脱落率(周忠英, 夏勉, 江光荣, 孙启武, 2016; Clark, 2011; Duncan et al., 2003; Lambert, 2013)。例如, 在每次会谈结束后使用治疗同盟问卷, 或者将对治疗同盟的感知写进个案记录中, 以帮助治疗师对来访者的脱落风险保持觉察, 以便及时对潜在脱落者进行干预(Dyer, Hooke, & Page, 2014)。另外, 最好的预防脱落的策略是什么?这其实并没有一个明确的答案, 因为心理治疗是一项高度个性化的工作, 治疗策略要因人、因时而异。研究者建议在实践中要根据来访者的需要、个性、文化背景等选用最适宜的干预策略(Swift et al., 2012)。例如, 对于有很多现实生活问题的来访者, 可以使用个案管理策略帮助来访者处理日常生活问题, 从而帮助来访者坚持心理治疗(Ogrodniczuk et al., 2005)。
尽管发展出预防脱落的策略很重要, 但目前这方面的研究成果很有限, 远不如脱落的发生率和预测变量的研究成果丰富。很多帮助来访者坚持治疗的策略还处于理论构想或临床描述的阶段, 鲜有研究者在临床实践中专门使用这些策略并进行深入的实证研究。这导致了我们对于这些策略的适用情况、使用条件和可能的风险知之甚少, 也缺乏成熟的评估量表和操作化方案, 这使得脱落的预防策略在临床工作中的应用受到限制。为了更好地给临床工作者提供与特定情境相适应的预防脱落的有效建议(situation-recommendations), 使其在特定时机对特定来访者采取有针对性的策略以促进其坚持治疗(Ogrodniczuk et al., 2005), 我们迫切需要加强这个领域的研究。
5 未来研究展望
5.1 脱落的操作性定义需要改进
近年来不少研究者提议应重新审视“脱落者”的定义和对“脱落者”的看法。研究者认为脱落者可能已通过早期治疗而明显获益, 甚至一些脱落者恰恰是因为早期治疗中的症状快速改善而选择了终止治疗, 可能他们并不需要完全实现治疗目标, 因而研究者主张将其看成早期治疗反应者(early treatment responders) (Erbes et al., 2009; Szafranski, Smith, Gros, & Resick, 2017)。还有研究者认为, 有些来访者不适合进行心理治疗, 故而对治疗没有反应甚至会出现心理问题恶化, 因此这类来访者也不应被视为脱落者(Lambert & Ogles, 2004)。遗憾的是这些假设目前仅停留在理论层面。
我们认为, “早期治疗反应者”的脱落可能本身也与心理功能受损有关。根据脱落的核心内涵, 只要来访者在最初求助的问题没有解决前(而不是根据是否完成全部的治疗协议)就终止治疗, 就可以将其看作脱落者。研究者需要进一步完善脱落的操作性定义, 脱落的操作性定义既要尽可能准确地反映出脱落的真正内涵, 也要尽可能多地涵盖脱落的各种临床情况。我国研究者唐小娟(2015)在对大学生来访者脱落的研究中采用了“治疗师评定”法来定义脱落, 并给治疗师提供了脱落的统一的操作性定义：“至少参加过一次咨询会谈, 并且没有经过咨询师同意而自己停止继续咨询。比如, 没有参加约定好的会谈; 没有达到预定的咨询目标就不再咨询; 促使当事人寻求心理咨询的问题没有解决; 当事人单方面终止或者不与咨询师讨论结案就自己结束治疗等。”这个操作性定义结合了治疗师判定、来访者爽约、单方面终止、临床显著变化等多个标准, 其明显的优势在于为治疗师提供了一个统一定义, 也涵盖了脱落的各种临床情况。这种操作性定义方法与不少研究者所建议的操作性定义界定方式相契合(Swift & Greenberg, 2012; Warnick et al., 2012)。总之, 发展出一个新的、更有说服力的脱落操作性定义是这个领域亟待解决的问题。在一个具有共识性的操作性定义基础上开展研究, 有助于我们更深入地理解心理治疗脱落现象和切实降低脱落率。
5.2 未来研究需更多地在自然治疗设置下进行
当前的脱落研究仍以实验情境下进行的效力研究为主。但实际上, 效力研究中严格的主试标准、被试标准及实施过程标准等设置在真实的临床情境中很难达到(Lambert & Ogles, 2004)。临床中的来访者通常为有共病、而非单一疾病的患者, 而治疗师也不会挑选患者, 并且治疗师会根据每个患者的情况做出适合于他的治疗方案而非使用统一方案(Maniaci et al., 2017; Southam-Gerow et al., 2003)。这些因素限制了效力研究中的研究结果在临床实践中的应用。更有研究者提出质疑, 认为效力研究中的治疗师多为大学培训中心受训中的治疗师, 他们比拥有执照的全职治疗师临床经验更少, 这可能会影响治疗同盟而导致来访者脱落的增加(Fernández-Arias et al., 2016; Raytek, McCrady, Epstein, & Hirsch, 1999)。但是, 相关研究结果却发现, 效力研究中的脱落率显著低于效果研究中的脱落率, 这其中或许还有其他重要的暂未阐明的因素在起作用。
笔者认为, 国内外研究者很少关注的两种研究设计中治疗费用差异这一因素可能是导致两种研究设计中脱落率差异的一个重要因素。在真实临床治疗中, 来访者需要付费, 而效力研究中招募而来的患者被试通常是免费参加治疗。“是否收费”可能对来访者的参与动机、治疗同盟和治疗效果产生复杂影响。例如, 对一些来访者而言, 参与治疗相当于完成一个实验, 当发现自己并不能获益(即症状改善)且退出实验没有损失时, 他们可能更容易从治疗中脱落; 但另一些来访者可能在没有明显获益的情况下仍会继续接受治疗, 因为治疗免费, 这给予了他们更多尝试和坚持的理由(Krishnamurthy et al., 2015)。当然这些只是假设, 至今还没有研究探讨过效力研究中“是否免费”这一和治疗设置有关的变量对脱落的影响。但治疗费用对来访者脱落的影响不能忽视。虽然效力研究中对各类研究变量的操纵和控制更严格, 但其与真实治疗情景差异太大, 我们甚至有理由质疑招募而来的患者免费参与的治疗项目是否能称得上真正的心理治疗。
5.3 脱落的预测变量的研究需要进一步细化
尽管许多重要的治疗过程预测变量已被识别出来, 但实际临床工作的复杂性使得深入研究这些动态过程变量成为必要。例如, 即使大量证据支持“治疗同盟”对脱落具有显著预测力, 但在不同的评定者、评定次数上得到的治疗同盟质量对脱落的影响也不一致(Yoo et al., 2016; Zaitsoff et al., 2015)。此外, 静态的预测变量也并非不重要, 而是需要进一步澄清其对脱落的影响机制。尽管很多静态变量不能一致地预测脱落, 但也有少数静态变量对脱落的预测力在跨研究间的一致性较高, 可能是因为在这些静态变量与脱落率之间还存在稳定的中介变量。例如, 来访者的婚姻状况可能可以预测其责任感和承诺水平的强弱(Rooney & Hanson, 2001), 从而进一步预测其是否会选择脱落。因此我们有必要检视一些人口统计学变量如何通过一些中介变量影响脱落, 了解其内在作用机制(Warnick et al., 2012), 这样才能增强我们对脱落现象的深入理解, 帮助治疗师发展出更有效的干预脱落的措施。
5.4 关注治疗外重要事件及文化背景对脱落的影响
治疗师要加强对来访者特殊生活事件的觉察和干预。以往研究关注来访者治疗设置内变量, 但不够重视来访者在治疗设置外的生活中发生的重要事件, 例如搬迁、失业、家庭成员死亡等。研究发现, 和坚持治疗的来访者相比, 特殊的生活事件在脱落者身上出现的比率更高, 但没有被当作阻碍治疗的重要因素来解释脱落现象, 治疗师通常也不一定意识到这些重要因素的存在(Blackshaw, Evans, & Cooper, 2018; O’Keeffe, Martin, Targe, & Midgley, 2019)。另外, 文化差异对来访者脱落的影响也值得关注。现有研究大部分以西方国家来访者为被试, 以亚洲来访者为被试的研究寥寥无几。东方文化在很多方面不同于西方文化。笔者认为, 东方文化对于治疗师这一权威角色的态度可能更加复杂; 同时, 东方被试因为接受心理治疗而产生的羞耻感可能也会对脱落产生影响。遗憾的是目前还没有研究对这些可能影响脱落的因素进行探索。
来访者从治疗中脱落一直是治疗师临床工作中普遍存在的棘手问题, 受到多种因素的综合影响。如果能够通过实证研究发现预测脱落的更多变量, 并采用整合的、动态的视角理解来访者脱落的心理过程及其影响因素, 我们就能在此基础上更好地提出有效的干预策略, 切实降低临床心理治疗中的脱落率。然而, 目前国内心理治疗领域有关来访者脱落的研究还处于起步阶段, 亟需更多的治疗师和研究者加入进来, 共同推动我国心理治疗脱落领域的理论研究和临床应用的发展。
参考文献
Attrition in child therapy.In T. H. Ollendick, & T. J. Prinz (Eds.)
Outcome of cognitive-behavioural therapy in training practice with anxiety disorder patients
Dropping out of treatment: A critical review
Predictors of attendance and dropout in three randomized controlled trials of PTSD treatment for active duty service members
When life gets in the way: Systematic review of life events, socioeconomic deprivation, and their impact on counselling and psychotherapy with children and adolescents
A self psychologist approaches retirement: ‘‘Forced’’ termination with highly vulnerable clients
Annual evaluation report for 1975 of the Northwest Denver Mental Health Center. In C. Windle (Ed.)
Implementing NICE guidelines for the psychological treatment of depression and anxiety disorders: The IAPT experience
A meta-analytic review on treatment dropout in child and adolescent outpatient mental health care
URL PMID:23742782 [本文引用: 13]
Understanding processes of change: How some patients reveal more than others—and some groups of therapists less—about what matters in psychotherapy
The Session Rating Scale: Preliminary psychometric properties of a “working” alliance measure
Reducing the premature termination of children from psychotherapy through research based program evaluation (Unpublished doctorial dissertation)
Effects of providing domain specific progress monitoring and feedback to therapists and patients on outcome
Treatment presentation and adherence of Iraq/Afghanistan era veterans in outpatient care for posttraumatic stress disorder
DOI:10.1037/a0016662 URL [本文引用: 2]
Premature termination of psychological treatment for anxiety disorders in a clinical setting
Clinical judgment and decision making
Research on client variables in psychotherapy. In A. E. Bergin & S. L. Garfield (Eds.)
Strategies to predict, measure, and improve psychosocial treatment adherence
DOI:10.1097/HRP.10.1097/HRP.0000000000000005
URL
PMID:24394220
[本文引用: 1]
Nonadherence to psychosocial and behavioral treatment is a significant public health problem that presents a barrier to recovery and effective treatment. An estimated 20% to 70% of individuals who initiate psychosocial mental health services discontinue treatment prior to clinicians' recommendations. Empirically supported, evidence-based, stand-alone or adjunctive psychosocial interventions treat an increasingly wide range of mental health conditions. A core assumption of most, if not all, interventions is that clients will fully and actively engage in the treatment protocol. Adherence to psychosocial treatment has received much less scientific attention, however, than adherence to medical treatment. Drawing extensively from existing research, this comprehensive review conceptualizes several types of psychosocial and behavioral treatment adherence, examines predictors of adherence to psychosocial treatment, summarizes measures of adherence, and describes existing interventions to enhance psychosocial treatment adherence.
Predictors of treatment attrition among an outpatient clinic sample of youths with clinically significant anxiety
The role of client satisfaction in evaluating university counseling services
DOI:10.1016/0149-7189(83)90011-3
URL
PMID:10267259
[本文引用: 1]
Though ubiquitously used for evaluating university counseling services, client satisfaction assessment has been hampered by inadequate instrumentation. Systematic use of a short form of the Client Satisfaction Questionnaire (CSQ) in one such center over the past 5 years is described, together with strategies to ensure maximal accuracy and utilization of results. Several method factors were investigated. Optional respondent identification was not found to reduce response rate or increase reported satisfaction compared to anonymity, while a substudy obtaining very high response suggested bias from nonresponse to the routine survey (response rate = 40%) was not great. The CSQ was found to have excellent psychometric properties and many advantages for use in student service settings. Relationships between satisfaction and a number of other variables such as demographics, precounseling expectancies, problem type and severity, counselor differences, and duration of counseling are reported. Ways such findings were incorporated in service planning are briefly discussed.
Predictors of treatment dropout from a spouse abuse abatement program
The psychotherapy dose-response effect and its implications for treatment delivery services
Comparison of four operational definitions of premature termination
Predictors of premature termination from psychotherapy for anorexia nervosa: Low treatment credibility, early therapy alliance, and self-transcendence
Mediators and mechanisms of change in psychotherapy research
Family experience of barriers to treatment and premature termination from child therapy
Preliminary results of a new instrument to assess patient motivation for treatment in cognitive-behaviour therapy
Understanding dropouts
Clients’ perspectives on therapy termination
Survival modeling of discontinuation from psychotherapy: A consumer decision-making perspective
Moving beyond efficacy and effectiveness in child and adolescent intervention research
Outcome in psychotherapy: The past and important advances
DOI:10.1037/a0030682
URL
PMID:23505980
[本文引用: 1]
Fifty years after the 1963 debate between Strupp and Eysenck, as recorded in their articles in Psychotherapy, it is clear that Eysenck overstated the case against psychoanalysis and dynamic psychotherapy (Bergin, 1971), while inflating the magnitude of improvement in untreated individuals (Lambert, 1976). Eysenck was probably correct about the beneficial effects of behavior therapies, but did not foresee that behavior therapy would be supplanted by cognitive behavior therapies (CBT) and eclectic mixtures of CBT that incorporate elements of eastern religion, humanistic interventions, and psychodynamic constructs. Fortunately, most of the treatments that have been tested in rigorous investigations have been found to be effective, but few have distinguished themselves as uniquely superior. Many of the problems of how to measure the effects of treatment have been solved and suggest that about two thirds of treated individuals improve or recover. This leaves a sizable portion of nonresponding individuals, but emerging methods involving in tracking treatment response are being used to decrease deterioration and enhance positive outcomes.
The efficacy and effectiveness of psychotherapy. In M. J. Lambert (Ed.)
Sociodemographic and psychiatric determinants of attrition in the Netherlands Study of Depression and Anxiety (NESDA)
Treatment dropout at the Berkeley CBT Clinic: An outcome study (Unpublished doctorial dissertation)
Expectations of psychotherapy duration: How long should psychotherapy last?
Individual psychotherapy and the science of psychodynamics
Predictors of early dropout in treatment for gambling disorder: The role of personality disorders and clinical syndromes
The role of the early therapeutic alliance in predicting drug treatment dropout
Premature termination from university counseling centers: A review
Strategies for reducing patient-initiated premature termination of psychotherapy
‘I just stopped going’: A mixed methods investigation into types of therapy dropout in adolescents with depression
A meta-analysis of predictors of offender treatment attrition and its relationship to recidivism
The effects of employing different termination classification criteria in dropout research
What went wrong? Therapists’ reflections on their role in premature termination
Therapeutic alliance and the retention of couples in conjoint alcoholism treatment
Reducing psychotherapy dropouts: Maximizing perspective convergence in the psychotherapy dyad
DOI:10.1037/h0087822 URL [本文引用: 1]
Discriminating variables among psychotherapy dropouts from a psychological training clinic
Variables related to premature termination in a university counseling service
Predicting attrition from treatment programs for abusive men
Adverse events of group psychotherapy in the in-patient setting - Results of a naturalistic trial
DOI:10.1111/acps.12747
URL
PMID:28561929
[本文引用: 1]
OBJECTIVE: Adverse events of psychotherapy have often been neglected in research. In this study, potential adverse events of group psychotherapies in a psychiatric hospital were systematically assessed, explored for predictors and linked to treatment outcome. METHOD: A naturalistic trial was conducted in 180 in-patients attending different group psychotherapies. Adverse events were assessed using three different measures: (i) weekly reporting of unwanted treatment reactions, (ii) mood changes in response to every single group session and (iii) premature group termination. RESULTS: Different measures of adverse events were weakly associated. Deterioration of mood state and/or unwanted treatment reactions were experienced by 60-65% of all patients. Reports of unwanted treatment reactions decreased over time and were negatively associated with symptom improvement. However, mood state deterioration was constant and unrelated to treatment outcome. The rate of premature group termination was 34%. Significant predictors of adverse events included patient characteristics as well as disadvantageous group conditions. CONCLUSIONS: For the majority of patients, group psychotherapy in the in-patient setting is associated with adverse events. Changes over time and a strong correlation with general symptom severity must be considered in the assessment and interpretation of adverse events. Predictors should be considered as potential risk factors in future research.
What’s in a name? Classifying ‘the dropout’ from treatment for anorexia nervosa
DOI:10.1002/erv.964
URL
PMID:19851991
[本文引用: 2]
Treatment drop-out rates for anorexia nervosa (AN) are notoriously high, and have remained so over the course of time, and across different treatment settings and modalities. The outcome for such patients is frequently poor. Despite the prevalence of this problem, there is little quality evidence as to why this is the case, and what the predictors of drop-out are. Methodological uniformity has been called for in order to develop a cohesive literature in this field, and this paper argues that we must also reflect upon how our labelling of 'the dropout' has shaped what we have examined to date, and how that has in turn led to a weak evidence base. A new, less pejorative term for those who do not complete treatment is needed to reflect the direction research in this area needs to take.
Predicting premature termination of hospitalised treatment for anorexia nervosa: The roles of therapeutic alliance, motivation, and behavior change
DOI:10.1016/j.eatbeh.2013.01.007 URL [本文引用: 1]
Youth with anxiety disorders in research and service clinics: Examining client differences and similarities
DOI:10.1207/S15374424JCCP3203_06
URL
PMID:12881026
[本文引用: 3]
Compared 2 groups of children with anxiety disorders: those treated in a university-based research clinic (RC) and those treated in community-based service clinics (SCs). A widely endorsed goal in intervention research is to disseminate evidence-based treatments from RCs to SCs. Attaining this goal requires an understanding of the similarities and differences between clients in these 2 settings. Youth from SCs showed more comorbid externalizing diagnoses and externalizing problems and were more likely to come from low-income and single-parent families. On measures of internalizing symptomatology and diagnoses, youth from RCs were very similar to SC youth. To facilitate development of treatments with real-world applicability, we describe a model involving the testing of treatments in real-world settings. We also discuss limitations to this project.
Premature discontinuation in adult psychotherapy: A meta-analysis
DOI:10.1037/a0028226
URL
[本文引用: 6]
Objective: Premature discontinuation from therapy is a widespread problem that impedes the delivery of otherwise effective psychological interventions. The most recent comprehensive review found an average dropout rate of 47% across 125 studies (Wierzbicki & Pekarik, 1993); however, given a number of changes in the field over the past 2 decades, an updated meta-analysis is needed to examine the current phenomenon of therapy dropout. Method: A series of meta-analyses and meta-regressions were conducted in order to identify the rate at which treatment dropout occurs and predictors of its occurrence. This review included 669 studies representing 83,834 clients. Results: Averaging across studies using a random effects model, the weighted dropout rate was 19.7%, 95% CI [18.7%, 20.7%]. Further analyses, also using random effects models, indicated that the overall dropout rate was moderated by client diagnosis and age, provider experience level, setting for the intervention, definition of dropout, type of study (efficacy vs. effectiveness), and other design variables. Dropout was not moderated by orientation of therapy, whether treatment was provided in an individual or group format, and a number of client demographic variables. Conclusions: Although premature discontinuation is occurring at a lower rate than what was estimated 20 years ago (Wierzbicki & Pekarik, 1993), it is still a significant problem, with about 1 in every 5 clients dropping out of therapy. Special efforts should be made to decrease premature discontinuation, particularly with clients who are younger, have a personality or eating disorder diagnosis, and are seen by trainee clinicians.
Practice recommendations for reducing premature termination in therapy
DOI:10.1037/a0028291 URL [本文引用: 10]
High rates of PTSD treatment dropout: A possible red herring?
DOI:10.1016/j.janxdis.2017.01.002
URL
PMID:28117192
[本文引用: 1]
Few studies have examined symptom change among dropouts from posttraumatic stress disorder (PTSD) treatment. However, dropout is widely considered a negative event needing to be addressed. The present study investigated PTSD and depression symptom change in patients with PTSD who discontinued psychotherapy. Female civilians (n=321) diagnosed with PTSD participated in two randomized clinical trials examining PTSD treatment outcomes. Of those, 53 were identified as dropouts and included in this study. Symptom change was assessed by clinically significant change (CSC) criteria and symptom end-state criteria. Results demonstrated that considerable proportions of participants (35.85-55.56%) displayed significant improvement and/or met good end-state criteria for PTSD and depression. Results also revealed that participants who displayed symptom improvement were younger, attended more treatment sessions, were married or partnered, and had higher annual household income. Although preliminary, these findings contradict belief that treatment dropouts do not display symptom improvement.
Treatment dropout in web-based cognitive behavioral therapy for patients with eating disorders
DOI:10.1016/j.psychres.2016.11.034 URL [本文引用: 1]
Impulsive lifestyle counseling to prevent dropout from treatment for substance use disorders in people with antisocial personality disorder: A randomized study
DOI:10.1016/j.addbeh.2016.02.001
URL
PMID:26882500
[本文引用: 1]
Patients with antisocial personality disorder in outpatient treatment for substance use disorders are at high risk of drop-out. Using a randomized design, this study tested the impact of adding a brief psycho-educational program, the Impulsive Lifestyle Counseling program, to outpatient substance abuse treatment in order to prevent treatment dropout. Patients (N=175) were recruited from 13 municipal treatment centers in Denmark, and assigned to treatment as usual or to the experimental condition. In all, 172 patients could be included in the analyses. In the intent-to-treat analysis, the risk of treatment dropout was reduced among patients randomized to the experimental program (hazard ratio=0.63, p=.031), after controlling for age, gender, and substitution treatment status. The study supported the efficacy of the Impulsive Lifestyle Counseling program as a method for preventing treatment dropout for patients with comorbid antisocial personality disorder in substance abuse treatment. Trial registration #ISRCTN67266318.
Defining dropout from youth psychotherapy: How definitions shape the prevalence and predictors of attrition
DOI:10.1111/j.1475-3588.2011.00606.x
URL
[本文引用: 9]
Background and Aims: Attrition is a long-standing problem in mental health centres serving youth. However, attempts to understand attrition have not consistently identified the same risk factors. The way in which attrition was defined across studies may have had a significant impact on findings. This study examines three definitions of attrition across a large sample of children and adolescents receiving outpatient mental health services, and considers the different relationships observed between the identified predictors and each definition. Method: This study examined data collected concurrently from 1098 families who received services at an urban outpatient mental health clinic (OMHC). Logistic regression was used to examine the association between identified predictor variables and attrition, using three distinct definitions of attrition based on clinician judgment, missed last appointment, and specified dose. The results of each regression analysis were qualitatively compared to assess the impact on findings observed when applying different definitions of attrition. Results: As anticipated, observed predictors of attrition varied by definition. Ethnicity predicted attrition across all definitions. Residing in a single-caregiver household predicted attrition across two of the three definitions, while living with a non-biological family, receiving state-funded, low-income insurance support, having low parent-reported youth functioning, routine intakes (as compared to urgent intakes), and longer wait predicted attrition within only one definition. Conclusions: Rates and factors associated with attrition may vary substantially depending on how treatment attrition is defined. In the evaluation of attrition in youth mental health settings, the definition used should be clearly stated and should reflect the research question posed.
Predictors of dropout from community clinic child CBT for anxiety disorders
A meta-analysis of psychotherapy dropout
DOI:10.1037/0735-7028.24.2.190 URL [本文引用: 6]
Predicting premature termination with alliance at sessions 1 and 3: An exploratory study
DOI:10.1111/1467-6427.12031 URL [本文引用: 2]
The role of the therapeutic alliance in eating disorder treatment outcomes: A systematic review
Therapist effects on and predictors of non-consensual dropout in psychotherapy
DOI:10.1002/cpp.2022
URL
PMID:27160543
[本文引用: 3]
BACKGROUND: Whereas therapist effects on outcome have been a research topic for several years, the influence of therapists on premature treatment termination (dropout) has hardly been investigated. Since dropout is common during psychological treatment, and its occurrence has important implications for both the individual patient and the healthcare system, it is important to identify the factors associated with it. METHOD: Participants included 707 patients in outpatient psychotherapy treated by 66 therapists. Multilevel logistic regression models for dichotomous data were used to estimate the impact of therapists on patient dropout. Additionally, sociodemographic variables, symptoms, personality style and treatment expectations were investigated as potential predictors. RESULTS: It was found that 5.7% of variance in dropout could be attributed to therapists. The therapist's effect remained significant after controlling for patient's initial impairment. Furthermore, initial impairment was a predictor of premature termination. Other significant predictors of dropout on a patient level were male sex, lower education status, more histrionic and less compulsive personality style and negative treatment expectations. CONCLUSIONS: The findings indicate that differences between therapists influence the likelihood of dropout in outpatient psychotherapy. Further research should focus on variables, which have the potential to explain these inter-individual differences between therapists (e.g., therapist's experience or self-efficacy). Copyright (c) 2016 John Wiley & Sons, Ltd. KEY PRACTITIONER MESSAGES: There are substantial differences between therapists concerning their average dropout rates. At the patient level, higher initial impairment, male sex, lower education, less compulsive personality style, more histrionic personality style and low treatment expectations seem to be risk factors of non-consensual treatment termination. Psychometric feedback during the course of treatment should be used to identify patients who are at risk for dropout.
