Introduction
People entering psychological treatment vary in their motivations for change (Holtforth & Mikalak, 2012). Concepts of readiness to change (RTC) and treatment motivation have received most research attention within the field of addiction (DiClemente et al., 2004), where it is widely considered that an individual needs to be ‘ready to change’ for treatment to succeed. However, clinical experience suggests that RTC may play a role in treatment outcome across a range of conditions. Although various concepts have been considered relevant to understanding motivation within therapy (Holtforth & Mikalak), the Trans-Theoretical Model of Stages of Change, (TMSC; Prochaska & DiClemente, 1984; Prochaska, DiClemente & Norcross, 1993) is one of the most influential.
The TMSC describes the process and tasks of behaviour change, within or outside therapy contexts. It outlines five successive stages: precontemplation, in which the individual does not regard themselves as having a problem, or has no intention of changing; contemplation, the process of weighing up pros and cons of change; preparation, anticipation of imminent action and initiation of small steps; action, active engagement in effortful behaviours in the service of change; and maintenance, engagement in relapse prevention or maintenance behaviours.
Since the early 1980s, the TMSC has generated a huge volume of research. It has also been subject to critique (for reviews see Littell & Girvin, 2002; Drieschner et al., 2004). Key challenges include recognition that the process of behaviour change is non-linear, with relapse as a normal part of any recovery journey, the fact that people completing measures of RTC often endorse items reflecting multiple stages of change simultaneously and evidence that people rarely make sequential transitions from one adjacent stage to another, as predicted by the model. Despite this, the TMSC is still regarded as a useful heuristic for conceptualising variations in client motivational readiness. Littell and Girvin (2002, pp.255) in their critical review state that “The stage model’s contributions—emphasis on cognitive precursors and correlates of behavioral change, the concept of readiness, and nonlinear movement—can be retained without a stage theory”. Reflecting this, measures derived from the model, including those which explicitly operationalise RTC as a continuum, continue to be used to explore the relationship between RTC and therapy outcomes.
Other attempts to define and measure treatment motivation have also been criticised for conceptual confusion (e.g., Drieschner et al., 2004). For example, Drieschner et al. (2004) highlight the fact that the Neijmegen Motivation List – 2 (Keisjers et al., 1999), developed specifically to assess motivation for treatment in Cognitive Behaviour Therapy (CBT), includes items which conflate treatment motivation with factors that are likely precursors or associates of motivation (e.g., treatment expectancy), or arise as its consequence (e.g., session attendance). This results in a degree of circularity when the measure is used to explore the relationship between treatment motivation and CBT outcome and limits use in other treatment contexts.
RTC and CBT
CBT is time-limited and requires active collaboration from the start. As such it might be expected that client pre-treatment RTC would be predictive of outcome. Research has demonstrated that client active engagement during therapy influences therapeutic response. For example, individuals with high levels of attendance, adherence to homework and compliance with key behavioural techniques in CBT for anxiety disorders, achieve better outcomes than those with lower levels of treatment engagement (e.g., Glenn et al., 2013). There is also evidence to suggest that RTC at entry to therapy is associated with better outcome. For example, a systematic review and meta-analysis of the relationship between stage of change and clinical outcome across 76 studies of different forms of face-to-face psychotherapy (Krebs et al., 2018) identified a small to moderate effect (d = 0.41) of RTC on treatment outcome, with those clients who had greater RTC showing better outcomes. However, although around half the studies in this review involved CBT-based treatments, a majority addressed substance misuse and 95% were published before 2016.
It has been argued that the concepts of RTC apply more to problems which have a clear behavioural manifestation (e.g., substance misuse, binging) than to those where it is unclear what behavioural changes might be required to bring about improvement (e.g., depression, chronic pain, e.g., Drieschner et al., 2004), and the bulk of research into RTC and therapy outcomes reflects this. However, clients may approach therapy with varying degrees of readiness to ‘do what it takes’ to make progress with their problems even if they are not sure what this might entail. Conversely, even where required behavioural changes may appear clear (e.g., to stop drinking alcohol) the process of achieving these changes may require unanticipated effortful commitments. It is therefore plausible that RTC may be relevant across a broader range of common mental health problems, where it might manifest as a willingness to come into contact with distressing emotions, be open to considering different ways of responding to difficult situations, or to confront distressing past experiences in the service of change.
The aim of the current review is to examine the association between pre-treatment RTC and treatment outcome in studies examining the efficacy/effectiveness of CBT for mental health problems, excluding addictive behaviours. The review focuses on studies which have used measures conceptually linked to the TMSC, as one of the most prominent theories of motivational processes in psychological and behavioural change, whilst holding in mind some of the recognised limitations of this theory. Importantly the TMSC is concerned with the person’s RTC (hereafter termed ‘readiness’ for continuous measures and ‘stage of change’ for categorical measures) in relation to a problem, rather than their desire to engage with a particular change method (Carey et al., 1999) and thus measures linked to this model are applicable when considering the potential role of RTC across different treatment (or no-treatment) conditions. This is important because many studies appear to operate under the assumption that where an association between RTC and treatment outcome is observed, it results from some influence of RTC on the way a person engages with treatment. However, a first step in establishing whether RTC exerts an influence on treatment response rather than, for example, simply being associated with positive change over time, would be to demonstrate a differential impact of RTC on symptom change in those receiving psychological treatment as compared to those in those in waitlist / non-treatment control conditions. This question has important theoretical implications and requires a measure which is relevant across contexts. Second, various authors have hypothesised that an individual’s RTC may be more relevant to some treatment interventions than others depending on their proposed mechanism of change (e.g., Wolk & Devlin, 2000; Lewis et al., 2009, Gorrell et al., 2019). This review will therefore consider the following questions:
1.
Is there a relationship between an individual’s readiness/stage of change at the start of CBT and (a) treatment attendance and/or (b) symptom change, over the treatment period?
2.
For controlled studies only: Is the association between readiness/stage of change and symptom change greater in those receiving CBT than those in waitlist or no-treatment control groups? (e.g., is any effect observed more substantial than might be expected from a relationship between readiness/stage of change and spontaneous symptom change? )
3.
For controlled studies only: Is the association between readiness/stage of change and symptom change greater for those receiving CBT than those receiving other therapeutic interventions?
Method
Protocol and Registration
The protocol for the review was registered with Prospero REF: CRD42020209173.
Search Strategy
Four databases: PsycINFO; MEDLINE; Embase and CINAHL were searched to identify relevant studies using the search terms shown in Table 1. Terms were searched in the following domains: title, abstract, heading word, table of contents, key concepts, original title, tests and measures, Medical Subject Headings (MESH; OVID). Terms addressing each criterion were combined with OR and returns across criteria were combined with AND. Returns were limited to English language publications from 1st January 1980 to current. Initial searches were conducted on November 16th and 17th 2020. A search update was conducted on 7th May 2024 (see later). Following identification of eligible full texts, reference screening and forward citation searching were conducted to determine the final review sample. The full search run in PsycInfo, is provided in the online supplementary materials (Appendix A).
Table 1
Search terms used to identify potentially eligible articles
Criterion for Initial Literature Search | Search Terms |
|---|---|
Study includes a measure of motivational readiness conceptually linked to the transtheoretical model of stages of change. | Readiness to Change OR Stages of Change OR Transtheoretical Model OR Motivation OR readiness ruler OR URICA OR SOCRATES OR Contemplation ladder |
The study includes the delivery of a cognitive behavioural psychological intervention | Cognitive behavio$ OR Cognitive Therap$ OR CBT |
The study reports engagement or outcomes of the intervention | outcom$ OR effect$ OR prognos$ OR impact$ OR treatment$ OR benefi$ OR engag$ OR attend$ or dropout$. |
Study Selection
A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA: Moher et al., 2009) flow chart is shown in Fig. 1. Searches were conducted in each database and returns were combined and deduplicated. The exclusion criteria applied at the title/abstract screening phase and the full text screening phase are shown in Table 2 (see also Appendix B). Separate criteria were applied at the title/abstract and full text stage because certain qualifying features of studies could rarely be determined by inspection of the abstract alone.
After removal of duplicates n = 4830 unique returns were identified in the initial search. Following title/abstract screening n = 304 were moved to full text screening and n = 19 were identified as eligible for inclusion in the review. No further eligible papers were identified through backward and forward citation searching.
Fig. 1
PRISMA flow-chart illustrating the study selection process
* One study was excluded from narrative synthesis after discussion with the corresponding author due to an inconsistency in reporting the results which rendered them uninterpretable
×
![]()
Table 2
Title and abstract and full text exclusion criteria
Title/Abstract Exclusion Criteria | Additional Full Text Exclusion Criteria |
|---|---|
The study did not include a CBT arm | Study did not include a measure of motivational readiness linked to transtheoretical model of stages of change |
The population under investigation was not defined by the presence of a mental health problem (studies where the population of interest was defined only by the presence of an addictive behaviour were excluded) | The measure of motivational readiness was not taken prior to intervention or in session 1 |
The clinical target of the intervention was not a mental health problem (studies where the clinical target was related to an addictive behaviour, e.g., craving, abstinence, were excluded) | No measure of clinical symptoms post-intervention or no measure of engagement with intervention |
The report did not include original data (e.g., a commentary, opinion piece or protocol); | Study did not include a CBT intervention (note: ERP included for OCD spectrum disorders) |
The study was a qualitative study, a case study or case series, was a cross-sectional study or a review. | Study did not report analysis which directly examined link between motivational readiness and outcome/engagement in those receiving CBT |
Titles and abstracts identified in the initial search were screened by the first author (CC), with the first 20% (n = 968) also screened by an independent reviewer (LS). This yielded an initial concordance of 91% and an inter-rater reliability of k = 0.45 (SE = 0.05; t = 14.00, p < .001, agreement moderate). Review of discrepancies indicated an equal proportion in which CC made an inclusion decision, and the independent reviewer, LS, made an exclusion decision and vice versa. In both cases most disagreements reflected missed exclusions: that is, when abstracts were reinspected, they had a clearly stated reason for exclusion which had been overlooked. Recoding these missed exclusions as agreements in discussion with a third reviewer (RKB, based on re-review of content in relation to exclusion criteria) yielded a revised concordance rate of 97%, k = 0.75 (SE = 0.04, t = 23.21, p < .001, agreement substantial). In remaining, ambiguous, cases the third reviewer’s (RKB) judgement was used to determine whether a paper moved to full text screening. Given that most disagreements between CC and LS were conservative (e.g., comprised inclusion of ineligible abstracts), the first reviewer’s decisions were retained for the remainder of the titles/abstracts screened.
At the full text screening stage, a sample of n = 60 (20%) texts were screened for eligibility by CC and LS. There was 100% concordance between raters and CC’s decisions were accepted for the remainder of the full texts. Reasons for exclusion of full texts are shown in Fig. 1.
Data Extraction
A data extraction tool was developed incorporating: authors, study design, study location, publication year, journal, hypotheses/research question, population, participant eligibility criteria, number of participants, age (range, mean), participant gender, participant ethnicity, CBT intervention description, control intervention description, measure of motivational readiness (including reliability/validity of measure), symptom outcome measures, attendance and engagement outcome measures, statistical analysis used to examine relationship between RTC and treatment outcome, findings, conclusions, limitations. A second reviewer, LS, conducted independent data extraction for 20% of full text articles (n = 4, 68 data items extracted), yielding a very high degree of consistency (96%). Disagreements were resolved through discussion.
Quality Assessment
Quality assessment was conducted using the Downs and Black (1998) Quality Index, which is designed to assess quality across both randomised and non-randomised studies of healthcare interventions. It yields an overall quality rating, alongside subscales addressing quality of reporting, external validity, internal validity (bias), internal validity (confounding, selection bias), and power. Following Morton, Barton, Rice and Morrisey (2014) the item addressing power was simplified to record only whether a power calculation was reported. Coding of items that related to non-relevant aspects of study design (e.g., items assessing the equivalence of recruitment procedures between study arms in single group designs) was standardised to ensure consistency across studies. The independent reviewer, LS, assessed 20% of included studies (n = 4) yielding an agreement rate of 89%, k = 0.77 (SE = 0.06, t = 8.07, p < .001, agreement substantial) across the 108 items rated. Discrepancies were resolved through discussion. Information was also gathered from each study concerning any details of: the training/qualifications of those delivering CBT interventions; intervention fidelity, and reliability and validity of measures of readiness.
Search Update
A search update was conducted on 7th May 2024. A further 1030 titles/abstracts were identified and screened by CC across the four databases, yielding three additional papers eligible for inclusion.
Data Synthesis
Findings were discussed with reference to the outcome examined; clinical problem addressed; measure of readiness; study design – whether the study allowed for a comparison between the relationship between readiness and outcome in CBT and another psychotherapeutic intervention or non-intervention control condition. The heterogeneity of study designs and analytic approaches precluded extraction and comparison of effect sizes. However, strengths and limitations of analytic approaches, and their comparability, are discussed.
Results
Study Characteristics
Overview
Table 3 provides an overview of the studies included in the review. Each study was assigned an ID number, used for identification in the remainder of the results section.
Table 3
Studies included in the review
ID | Author, Year and Country of Origin | Population/ Setting and Study Design | Sample characteristics (total N, mean age (SD), % female, ethnicity | Treatment Intervention | Measure of Readiness | Outcome(s) | Relevant Findings |
|---|---|---|---|---|---|---|---|
1 | Berry (2010) USA | Adult women admitted to a psychiatric inpatient unit Naturalistic pre-post study. | N = 160 Mean Age = 34.3 (11.0) Gender: 100% female | Group CBT on inpatient unit (5 x groups per day) with adjunctive medical/psychosocial interventions | URICA | Length of inpatient stay (days to discharge) Number of CBT group sessions attended. | Higher URICA score at admission associated with a longer length of stay after controlling for baseline symptom measures which were positively associated with baseline URICA score. URICA score at admission not associated with number of CBT group sessions attended. |
2 | Boswell et al. (2012) USA | Adults seeking treatment for an anxiety disorder at University Anxiety Disorder Clinic Quasi-experimental pre-post study (data collapsed across immediate and waitlist arms of RCT) | N = 29 Mean Age = 29.76 (9.54) Gender: 56% Female | Individual Transdiagnostic CBT for Anxiety Disorder (unified protocol) 8–18 × 60 min sessions covering 8 modules. | URICA | HAM-A HAM-D WSAS None defined as primary outcome, but treatment protocol is for anxiety disorders. Outcomes assessed at end of treatment. | URICA score was significantly positively associated with HAM-A change; HAM-D change and WSAS change: i.e., greater readiness significantly associated with greater symptom reductions across treatment. URICA score moderated the effect of initial symptom severity on symptom change for both HAM-A and HAM-D - High baseline severity was associated with greater change when URICA score was high. |
3 | Castro-Fornieles et al. (2011) Spain | Consecutive referrals to child and adolescent Eating Disorders unit. Naturalistic pre-post study | N = 31 Mean Age = 16.2 (1.10) Gender: 98% female | Individual and Group CBT following Fairburn and Wilson (2003). Sessions 2 x weekly for outpatients and daily for inpatients. All patients receive adjunctive medical management | BNSCQ Total Score | EDI − 2–11 subscales derived). Number of binges per week Number of vomits per week [assessed at end of treatment] | Higher BNSCQ scores at baseline significantly associated with greater reductions in binges per week at end of therapy BNSCQ significantly positively associated with changes on EDI-2 bulimia subscale and EDI-2 interoceptive awareness subscale No significant association between BNSCQ and any of the other EDI-2 subscale scores examined. BNSCQ score was the only variable significantly postively associated with clinical change in binges and decrease in EDI bulimia subscale after controlling for significant univariate correlates and known potential confounders. |
4 | Clarke et al. (2013) USA | Women with (PwSA) and without (PwoSA) a substance use problem admitted to acute psychiatric inpatient unit. Naturalistic pre-post study | N = 117 (n = 50 PwSA, n = 67 PwoSA) Mean Age - PwSA = 32.5 (11.3) PwoSA = 35.6 (10) Gender: 100% female | 4 x CBT groups per day plus individual sessions with adjunctive substance use group and medical management (variable duration of admission) | URICA | OQ [Assessed 24 h prior to discharge] | URICA score at treatment entry was associated with better psychological functioning at discharge, after controlling for potential confounders (marital status, length of stay, baseline symptoms), and accounted for 24% variance in discharge symptoms. |
5 | Goodwin (2019) Canada / USA | Community volunteers seeking treatment for anxiety Secondary analysis of RCT | N = 85 (MI + CBT = 42; CBT = 43) Mean Age: MI + CBT = 32.45 (10.54) CBT = 34.19 (11.92) Gender: MI-CBT 81% Female; CBT 95% Female | 15 sessions CBT (following Borkovec & Costello, 1993; Borkovec & Matthews, 1988; Borkovec et al., 2002) 15 sessions MI + CBT – 4 initial sessions of Motivational Interviewing (following Westra, 2012’s guidelines) and 11 sessions of CBT. | MISC 1.1 Adapted for CBT. | PSWQ [Assessed weekly during treatment] | Increased Session 1 Change Talk associated with significantly increased hazard of clinical response by end of treatment (controlling for treatment site, baseline motivation – Change Questionnaire - and medication). Session 1 Counter Change Talk associated with significantly reduced hazard of clinical response by end of treatment (controlling for treatment site, baseline motivation – Change Questionnaire - and medication). No significant Change Talk x treatment (CBT or MI-CBT) or Counter Change Talk x treatment (CBT or MI-CBT) interaction. |
6 | Gorrell et al. (2019) USA | Adolescent outpatients with Bulimia Nervosa. Secondary analysis of RCT | N = 110 (CBT n = 58; FBT N = 52 Mean Age = Not reported, Range 12–18 years Gender: 93.6% Female. | CBT for Adolescents adapted from adult CBT for Bulimia (Lock et al., 2005) Family-Based Treatment for Adolescent Bulimia Nervosa adapted from a family-based treatment for Anorexia Nervosa (Lock et al., 2002) 18 sessions over 6 months | YBC-EDS-MC | EDE Derived: EDE Global Score (attitudes: restraint, eating concern, weight concern, shape concern) Abstinence from ED behaviours: 0 = not abstinent/ 1 = abstinent (binging, vomiting, laxatives, diuretics, other medications, excessive exercise, fasting) [Assessed at end of treatment] | Higher motivation to change associated with lower EDE Global score at end of treatment (controlling for age and BMI) No treatment x motivation interaction. No significant association between motivation to change and abstinence at end of treatment (controlling for age and BMI) No motivation to change x treatment interaction for abstinence at end of treatment |
7 | Greenberg et al. (2019) USA | Individuals with a diagnosis of Body Dysmorphic Disorder (BDD) or its delusional variant. Quasi-experimental pre-post study- data combined from uncontrolled pilot trial and treatment arm of waitlist- controlled RCT. | N = 44 Mean Age: 34.8 (10.1) Gender: 61% Female | Individual modular CBT treatment designed to treat core features of BDD. Sessions for 18–22 weeks, with last two sessions spaced at 2-weekly intervals (Wilhelm, Philips & Steketee, 2013) | URICA | BDD-YBOCS Responders defined as those showing a reduction of > = 30% at post-treatment. [Assessed at end of treatment] | Among the 14 independent variables examined, three, including URICA score, were significantly associated with treatment response. For every unit increase in the URICA score the odds of being a post-treatment responder increased by 145% |
8 | Heider et al. (2018) Germany | Patients with a diagnosis of somatization disorder, an undifferentiated somatoform disorder, a pain disorder or a conversion disorder Pre-post study - per-protocol analysis of data from two samples. Participants adhering to the treatment protocol (attending 5/8 sessions) included. | Sample 1: N = 109 Sample 2: N = 35 Mean Age: Sample 1: M = 48 (12.38) Sample 2: M = 46 (12.29) Gender: Sample 1: 76% Female Sample 2: 71% Female. | Sample 1: 8 x weekly sessions of manualised group CBT (see Schroder et al., 2013). Sample 2: 8 x weekly sessions of CBT with some individual and some group sessions Samples combined for secondary analysis | FF-STABS | SOMS-7 [Assessed at end of treatment] | Main effects: No association between precontemplation, preparation, action or maintenance and symptom reduction. Interaction Effects: Significant interactions between symptom levels and both precontemplation and action scores (see conditional effects below): Conditional effects Pre-Contemplation: People with high levels of precontemplation showed no significant symptom reduction. People with moderate or low levels of pre-contemplation showed significant symptom reduction. Preparation: People with high preparation scores showed no significant symptom reduction. People with moderate or low preparation scores showed significant symptom reduction. Action. People with high action scores showed no significant symptom reduction. People with moderate or low action scores showed significant symptom reduction. Maintenance: Significant conditional effects at high, moderate and low values of maintenance (in absence of main effects or interaction effects). At all levels of maintenance, participants with higher baseline symptoms showed greater symptom reduction. Robust Regression was consistent with OLS for precontemplation, contemplation and action. Robust Regression for maintenance showed a significant main effect of maintenance on symptoms reduction but no interaction. |
9 | Hudson & Neil (2020) Canada | Female adults treated at an ED outpatient unit over a 2.5 year period. Naturalistic pre-post study | N = 108 Mean Age: 27.24 (10.59) Gender: 100% Female | 16-week CBT group programme for eating disorders (see Mac Neil et al., 2016). Some patients had adjunctive ACT or cognitive remediation therapy to address problems other than ED psychopathology (25% of total). | Readiness Ruler | Number of CBT sessions attended. Engagement – person did or did not attend first session of CBT Dropout – person did or did not attend any of the last 6 sessions of CBT | Readiness Ruler rating not significantly associated with number of CBT sessions attended, initial engagement in CBT or dropout from CBT. |
10 | Lewis et al. (2009) USA | Adolescents aged 12–17 with major depression. Secondary analysis of RCT | N = 332 Mean Age: 14.6 (1.5) Gender: 59.3% Female | CBT – 15 sessions over 12 weeks, 2 parent sessions and 2–3 joint adolescent & parent sessions. Fluoxetine – dose monitored in 6 medication review over 12 weeks. Placebo -. monitored in 6 medication visits over 12 weeks. CBT + Fluoxetine – CBT delivery independent of medication protocol but increases in medication informed by response to CBT in cases of partial response. Fluoxetine and Placebo administered double-blind. | SCQ − 18-item version with wording modified for an adolescent sample. | CDRS-R [Assessed at 12 weeks] | Baseline action score significantly associated with treatment outcome: adolescents with higher action scores at baseline showed a greater symptom reduction across treatment. Interaction between treatment arm and action score did not reach significance but analysis underpowered to examine the interaction effect. |
11 | Lombardi et al. (2014) Canada | Community volunteers with an anxiety disorder. Quasi-experimental pre-post study | N = 37 Mean age: 41.05 (11.85) Gender: 73% Female | CBT following Borkovek & Costello, 1993; Borkovek & Matthews, 1988; Borkovek, Newman, Pincus & Lytle, 2002) 6 × 2-hour sessions (weekly) + 2 × 1 h sessions. | CQ MISC 1.1 (78% session 1 22% session 2) | PSWQ Recovery status: Responder - participant shows reliable change and a post-treatment score below the clinical cut-off on the measure versus non-responder. [Assessed at post-treatment] | Higher Change Questionnaire scores significantly associated with lower post-treatment worry, controlling for baseline worry. Counter Change Talk (session 1 or 2) additionally associated with higher post-treatment worry. Change Talk not associated with post-treatment worry. Participants showing reliable symptom change at post-treatment had significantly lower levels of session 1 or 2 Counter Change Talk than those who did not show reliable symptom change. Change Talk was not associated with reliable symptom change. |
12 | Monaghan et al. (2015) USA | Patients admitted to a residential intensive inpatient programme for OCD. Naturalistic pre-post study | N = 424 Mean Age = 33.89 (14.16) Gender: 50% female | 2–3 individual CBT sessions / week. 2 h/ day therapist-guided ERP + 2 h/day self-guided ERP. Adjunctive groups and milieu experiences to reinforce ERP. Medical management. | URICA | YBOCS [Assessed at discharge] | URICA subscale scores were not significantly associated with change in YBOCS score. |
13 | Pellizzer et al. (2017) UK | Consecutive referrals to Flinders University Eating Disorders Outpatient Clinic Naturalistic pre-post study. | N = 78 Mean Age: 27.19 (9.60) Gender: 92.3% Female | 10 weekly session CBT-T intervention for eating disorders (see Waller et al., 2017) | Single item assessment: How ready are you to change? 0 to 100 | EDE - Global Score (restraint, eating concern, weight concern, shape concern) [Assessed mid treatment, post-treatment, follow-up 1; follow-up 2] Attrition – defined as termination of therapy before completion of 10th session. | Readiness to change was not significantly associated with attrition from treatment. Readiness to change did not interact with time to predict change in EDE Global score across treatment. |
14 | Solem et al. (2016) Norway | Outpatients with a diagnosis of OCD Naturalistic pre-post study | N = 121 (n = 65 group treatment; n = 56 individual treatment) Mean Age = 33.8 (12) Gender: Not Reported | Group ERP – 12 × 2.5 group sessions /week adapted from manual by Krone, Himle and Nesse (1991). Individual ERP – 15 × 90-minute sessions based on manual by Kozak and Foa (1997). | URICA | YBOCS Dropout from treatment (definition not provided) [Assessed at end of treatment] | The 8 treatment dropouts scored significantly lower than 113 treatment completers on the URICA contemplation subscale. No differences between treatment dropouts and treatment completers on other URICA subscales. URICA subscales were not significantly associated with end of treatment YBOCS score (controlling for baseline symptoms). |
15 | Treasure et al. (1999) UK | Women with Bulimia Nervosa Secondary analysis of RCT | N = 125 (MET n = 87; CBT n = 38) of which n = 91 completed the URICA Mean Age: MET = 28.8 (7.8) CBT = 28.5 (7.2) Gender: 100% Female | CBT: Bulimia Nervosa: A guide to recovery (Cooper, 1993). MET: first four chapters of clinician handbook of Getting Better bit(e) by bit(e) Individuals randomised to 4wks MET + 8 weeks Individual CBT; 4wks MET followed by 8 weeks group CBT; 4 weeks individual CBT followed by 8 weeks group CBT. Groups allocated to MET combined for current study to give n = 87 to MET and n = 38 to CBT. | URICA | Binge frequency, vomit frequency and frequency of use of laxatives at end of treatment phase 1 [Assessed 4 weeks into treatment – end of treatment phase 1] | Individuals classified as in the contemplation stage on the URICA showed less improvement in binge eating frequency than those in the action stage. Improvement in binge eating was moderated by stage of change on the URICA. No time x stage of change interaction for vomiting or laxative abuse. No treatment x stage of change interaction for any outcome. |
16 | Vogel et al. (2006) Norway | Consecutive referrals with OCD to city psychiatric outpatient clinic in Norway from 1993–1998 Naturalistic pre-post study. | N = 39 with N = 37 providing study data. Mean Age: 35.1 (12.1) Gender: 73% Female | Random assignment to one of two individual ERP treatments based on Foa (1991). Structurally identical for 75% of session time. Differed in including either additional CBT strategies or relaxation techniques. 120 min twice a week for 6 × 120-minute sessions. 60- minute follow-up at 3, 6, 9, 12 months. | URICA | YBOCS [Assessed at end of treatment and 12 months follow-up] | URICA score not significantly associated with treatment response at end of treatment or 12 month follow-up. |
17 | Wagner et al. (2015) Austria | Participants aged 18–35 with Bulimia Nervosa or EDNOS Secondary analysis of RCT | N = 126 (70 INT-GSH; 56 BIB-GSH) Mean Age: 24.17 (4.46) INT-GSH 25.02 (3.84) BIB-GSH Gender: 100% Female | CBT internet-based guided self-help with weekly email therapist support (SALUT-BN, 7 modules) CBT Bibliotherapy with weekly email therapist support (15 chapters- Getting Better Bit(e) by Bit(e), by Schmidt, Treasure & Alexander, 1993). | BNSCQ | Diagnostic Status at end of treatment (participant did or did not fulfil DSM-IV criteria based on previous month’s binging / purging behaviour). Drop-out during treatment (participating for less than 2 months or completing less than 3 modules of INT-GSH or 6 chapters of BIB-GSH). | Higher BNSOC score significantly associated with greater likelihood of remission at end of treatment in model exploring role of personality factors BNSOC not significantly associated with dropout from treatment. |
18 | Wolk and Devlin (2000) USA | Women with Bulimia Nervosa Secondary analysis of data from a single site of RCT | N = 110 randomised and n = 66 completing 19 sessions of treatment. Mean Age: for 110 randomised CBT: M = 28.2 (6.6) IPT: 25.1 (5.4) Gender: 100% Female | CBT (n = 32) 19 sessions - Fairburn Marcus and Wilson (1993) IPT (n = 34) 19 sessions (Fairburn et al., 1993) | SCS | EDE used to categorise participants into: full remission, partial remission/sub-threshold or still meeting diagnostic criteria [Assessed at end of treatment] | Overall stage of change associated with improved outcome (full remission/partial remission/still meeting criteria): None of the participants in the pre-contemplation stage at treatment entry achieved remission. Association between stage of change and outcome was not significant for CBT but was significant for IPT. |
19 | Lewis et al. (2012) USA | Participants presenting to a university training clinic for treatment of psychological problems. Naturalistic pre-post study | N = 173 Age: 27.94 (11.42) Gender: 66.5% Female | Case formulation driven CBT as in routine clinical practice. | SOCS | BDI BAI Assessed weekly until end of treatment. | Description of analysis and results inconsistent and impossible to interpret. Communication with author did not lead to resolution. Included as eligible but results not reported. |
20 | Wade et al. (2021) Australia/UK | Consecutive referrals to a University Eating Disorder Service RCT comparing two active CBT interventions | N = 98 Age: CBT-T = 26.91 (SD = 10.88) CBT-M = 25.77 (SD = 7.45) Gender: CBT-T – 91% Female CBT-M – 94% Female | CBT-T – manualised transdiagnostic CBT for eating disorders (Waller et al., 2019) CBT-M – Individually delivered CBT supported by two self-help books (Cooper, 1993; Cooper et al., 2001) | 100 point likert scale assessing readiness to change | EDE– Global Score Assessed at 4 weeks, 10 weeks, 14 weeks, 22 weeks | Main effect of readiness to change on symptom levels – Lower readiness to change at baseline was associated with a greater symptom burden over treatment. Significant readiness x time x group interaction: participants with low motivation to change in CBT-T showed lower rates of improvement at one-month and three-month follow-up than those with high motivation randomised to CBT-T and those with both low and high motivation randomised to CBT-M. |
21 | Ciesinski et al. (2024) | Individuals with a diagnosis of Intermittent Explosive Disorder who completed treatment across three study trials. | N = 64 Age:37.20 (SD = 9.87) Gender 34% Female | Multicomponent 12 session CBT protocol for IED (adapted from Deffenbacher & McKay 2000) N = 12 received protocol in 12 × 75 min group sessions. N = 52 received protocol in 12 × 50 min individual sessions | SCQ – 32 item version | Overt Aggression Scale – Modified (OAS-M) – past week ratings. Assessed at pre-treatment and post-treatment | No correlation between initial SOC score and change in OAS-M from pre to post treatment. Initial SOC score not predictive of IED diagnostic status at end of treatment. |
22 | Keegan and Wade (2024) | Individuals referred to a university outpatient eating disorder service. Subset of participants in a larger trial who met criteria for atypical AN or BN | N = 67 (AN = 33; BN = 34) Age: 25.17 (SD = 9.38) Gender: 95% female | CBT -T manualised transdiagnostic CBT for eating disorders (Waller et al., 2001) | Readiness assessed on 10-point likert scales: “How ready are you to change?” | EDE global score assessed at baseline, session 1, session 4, session 10, 1 month and 3 month follow-up. | Significant interaction between baseline readiness and change in EDE score over time – those with higher baseline readiness showed greater change over time. Three way interaction between baseline readiness, change over time and BN/AN diagnosis. Moderating effect of readiness on change in EDE score over time present for those with BN only. |
Location and Design
Studies were conducted in North America (n = 13), Western Europe (n = 7) or Australia (n = 2) and comprised primary (n = 1, 20) or secondary analyses of data from randomised trials (n = 6: 5, 6, 10, 15, 17, 18, 22), secondary analyses collapsing data across studies or trial arms (n = 3: 2, 8, 21), naturalistic pre-post studies in clinical settings with routine referrals (n = 9: 1, 3, 4, 9, 12, 13, 14, 16, 19), and quasi-experimental studies utilising a pre-post design (n = 2: 7, 11). Nineteen studies were published after 2009, with the earliest published in 1999.
Population Characteristics
Most studies included adults (n = 19) whilst three used adolescent samples (3, 6, 10). The proportion of Caucasian participants ranged from 58.9 to 92.5% for the n = 13 studies reporting ethnicity. Females were in a majority in all 21 studies reporting gender and in 11 more than 90% of the sample were female.
Three studies included participants with varied diagnoses (1, 4, 19), but the majority focused on single conditions: anxiety disorders (2, 5, 11), OCD or BDD (12, 7, 14, 16), eating disorders (3, 6, 9, 13, 17, 18, 20, 22), depression (10), somatoform disorders (8) or intermittent explosive disorder (21). Likewise, three studies had inpatient samples (1, 4, 12) or mixed inpatient/outpatient samples (3), whilst the majority included outpatients (n = 17), and those receiving therapist-supported guided self-help (17). Most studies reported that participants met DSM-III/IV/V diagnostic criteria for their stated condition (American Psychiatric Association, 1980, 2000, 2013), whereas three focused on inpatient samples (1, 4, 12) and reported diagnoses but not the diagnostic criteria employed (see Table 4).
Table 4
Clinical problems, sample description and outcome measures
Disorder/Problem/ Population | Study ID | First Author and Year | Description of Sample | Ethnicity | Outcome Measure |
|---|---|---|---|---|---|
Mixed Disorders | 1 | Berry, 2010 | Women admitted to inpatient unit. Diagnoses: 50.3% Major Depression; 23.3% Bipolar Disorder; 9.8% Psychotic Disorder or Schizophrenia; 10.4% Eating Disorder; 6.1% Other mental disorder. | 58.9% Caucasian; 19% Hispanic; 6.7% African American; 6.1% Mixed Ethnicity; 3.1% Asian; 5.5.% other | Length of inpatient stay (days) |
4 | Clarke et al. (2013) | Women with and without a substance use disorder admitted in a psychiatric inpatient unit. Depression 50%; Bipolar Disorder 24%; Schizophrenia 10%; Anxiety 2%; Eating Disorder 12%. | PwSA (n = 50) 68% Caucasian; 18% Hispanic, 8% Mixed ethnicity, 2% Asian, race; 2% African American. PwoSA − 70% Caucasian, 17.9% Hispanic, 3% African/American, 1.5% Asian, 1.5% Mixed ethnicity. | Outcome Questionnaire (OQ) | |
19 | Lewis, 2012* *This study was not included in the narrative synthesis due to inconsistencies in reporting of results. | Adults seeking treatment from a university outpatient clinic. Participants assessed using structured clinical interview for DSM-IV (SCID-I First, Spitzer, Gibbon & Wiliams, 1996) Reported Diagnoses: Depressive disorders – 56.1% Anxiety Disorders – 32% Other Disorders (e.g. dissociative disorder, bipolar disorder) – 12% Participants excluded if BDI score indicated less than mild symptoms. | 92.49% Caucasian | Beck Depression Inventory Beck Anxiety Inventory | |
Anxiety Disorder | 2 | Boswell, 2012 | Adults meeting diagnostic criteria for one or more anxiety disorders according to Anxiety Disorders Interview Schedule for DSM-IV-Lifetime version (ADIS-IV-L) | 94.5% Caucasian | Hamilton Anxiety Rating Scale (HARS) |
5 | Goodwin, 2019 | Adults meeting DSM-IV-Tr and DSM-V diagnostic criteria for Generalised Anxiety Disorder according to SCID-I (First et al., 1996) with a Penn state worry questionnaire score > = 68. | MI: 74% Caucasian; 14% Asian; 5% African-Canadian; 2% Hispanic; 5% Multiracial. CBT: 74% Caucasian; 12% Asian; 0% African-Canadian; 5% Hispanic; 7% Multiracial. | Penn State Worry Questionnaire (PSWQ) | |
11 | Lombardi, 2014 | Community adults with a principal diagnosis of Generalised Anxiety Disorder based on Anxiety Disorders Interview Schedule for DSM-IV (ADIS-IV; Brown, Barlow & Di Nardo,1994). | 59% Caucasian 22% Asian 11% Hispanic 8% African Canadian | Penn State Worry Questionnaire (PSWQ) | |
OCD/BDD | 12 | Monaghan, 2015 | Adults admitted to a residential intensive inpatient programme for OCD with a primary diagnosis of OCD determined by clinical interview. Participants’ mean YBOCS score in severe range. | 84.2% White; 4.5% Asian 0.9% Black or African American 0.5% American Indian or Native Alaskan 0.5% Native Hawaiian or Pacific Islander and 0.7% Other ethnicity. | Yale Brown Obsessive Compulsive Scale (YBOCS) |
7 | Greenberg, 2019 | Adults with a DSM-IV primary diagnosis of Body Dysmorphic Disorder (BDD) or its delusional variant based on Structured Clinical Interview for DSM-IV patient version (SCID I/P First, Spitzer, Gibbon & Williams, 1996) and BDD-YBOCS score 24 or greater. | White 88% Black 7%, Other ethnicity 5% | Body Dysmorphic Disorder- Yale Brown Obsessive Compulsive Scale (BDD-YBOCS) | |
14 | Solem, 2016 | Adult outpatients with a DSM-IV diagnosis of OCD | Not Reported | Yale Brown Obsessive Compulsive Scale (YBOCS) | |
16 | Vogel, 2006 | Adults meeting DSM-II-R criteria for primary OCD with overt compulsions assessed via Anxiety Disorders Interview Schedule Revised (ADIS-R). | Not Reported | Yale Brown Obsessive Compulsive Scale (YBOCS) | |
Eating Disorder | 3 | Castro Fornieles et al., 2010 | Children and adolescents meeting DSM-IV diagnostic criteria for Bulimia Nervosa assessed by clinical interview. | Not Reported | Eating Disorder Inventory-2 Number of binges/vomits per week |
6 | Gorrell, 2019 | Adolescents meeting DSM-IV criteria for Bulimia Nervosa as assessed by Eating Disorder Examination. | Not Reported | Eating Disorder Examination Global Score | |
9 | Hudson, 2020 | Adults Meeting DSM-5 criteria for AN, BN or OSFED based on structured clinical interview with psychiatrist | Not Reported | Number of CBT sessions attended. Engagement, Drop out. | |
13 | Pellizer, 2017 | Adults meeting DSM-V diagnostic criteria for an eating disorder. | 88.5% Caucasian | Eating Disorder Examination Global Score | |
17 | Wagner et al. (2015) | Adults aged 18–35 meeting DSM-IV-TR criteria for BN (90%) or EDNOS (10%) with BMI > 18. | Not Reported | Diagnostic Status at End Treatment | |
18 | Wolk, 2000 | Adult women meeting DSM-III criteria for bulimia nervosa | For n = 110 Randomised: CBT: White 78% Hispanic 9% African American 7% Asian 6% American Indian 0% IPT: White 66% Hispanic 18% African American 9% Asian 7% American Indian 0% | Eating Disorder Examination – women categorised as: full remission, partial remission/ sub threshold, still meeting criteria | |
15 | Treasure, 1999 | Adult women meeting DSM-IV criteria for bulimia nervosa. | Not Reported | Binge Frequency, Vomit Frequency, Laxative Frequency | |
20 | Wade, 2021 | Adults meeting DSM-V criteria for an eating disorder:: Anorexia, n = 5, 5%; Bulimia, n = 68, 69%; Binge Eating Disorder, n = 5, 5%; Other Specified Feeding or Eating Disorder, n = 20, 21% | Not Reported | Eating Disorder Examination – Global Eating Disorder Psychopathology Score. | |
22 | Keegan and Wade (2024) | People age 15 or above with a DSM-V diagnosis of atypical Anorexia Nervosa (n = 33) or Bulimia Nervosa (n = 34) | White 88% | Eating Disorder Examination – Global Eating Disorder Psychopathology Score. | |
Depression | 10 | Lewis, 2009 | Adolescents with a DSM-IV primary diagnosis of major depressive disorder | 75.9% White; 10.2% African American; 9.3% Hispanic; 0.3% Pacific Islander | Children’s Depression Rating Scale-Revised (CDRS-R) |
Somatoform Disorders | 8 | Heider, 2018 | Adults with various somatoform disorders according to DSM-IV-TR criteria assessed using the International diagnostic checklist for DSM-IV (IDCL: Hiller, Zaudog & Mombour, 1996) | Not Reported | Screening for Somatoform Disorders (SOMS-7) |
Intermittent Explosive Disorder | 22 | Ciesinski, 2024 | Adults meeting current integrated research criteria for Intermittent Explosive Disorder (consistent with DSM-V criteria) | 31.3% Black/African American 6.3% Asian/Native American 62.5% White | Overt Aggression Scale – Modified (OAS-M) |
Intervention Characteristics
Studies explored various CBT formats: group (n = 2), individual (n = 14) combined individual and group (n = 5) and guided self-help (n = 1). In some naturalistic settings CBT was augmented with medical management and/or adjunctive therapies addressing comorbidities (e.g., substance use education) or general psychosocial support (e.g., milieu experiences in inpatient facilities). One study reported on the first four sessions of CBT, compared to a four-session motivational interviewing module (15). The remainder examined outcomes of courses of CBT treatment, with length depending on treatment protocol and setting (see Table 3).
Reports varied in the detail provided about CBT treatment. For six (6, 7, 9, 10, 11, 18) further information was obtained from primary outcome studies. Five provided no information about therapist training (3, 8, 10, 14, 22). Seven reported that interventions were delivered in whole or part by trainee psychologists (2, 4, 5, 11, 13, 19, 20), whilst the remainder (n = 10) reported delivery by a range of professionals with varying degrees of post-qualification experience. Eleven either referenced adherence checks, weekly supervision and/or measures to establish competency prior to delivery of therapy (2, 4, 5, 6, 7, 11, 16, 18, 19, 20, 21).
Outcome Measures
Most studies described the relationship between readiness and change on a validated symptom measure corresponding to the primary treatment target (see Table 4), usually assessed on a continuous measure of symptom severity. Single studies examined abstinence from ED behaviours (6), diagnostic status at end of treatment (17, 21), meeting recovery criteria (11) or demonstrating clinically significant response (7). Five examined the relationship between readiness and treatment engagement: length of inpatient stay & number of CBT sessions attended (1); number of CBT sessions attended, engagement with treatment, dropout (9); attrition from treatment (13), or dropout (14, 17).
Measurement of Readiness
Eight studies used the URICA (McConnaughy et al., 1983), either the 32-item version (1, 2, 4, 7, 12, 16), the 24-item version (15) or an unspecified version (14). Five derived a total score conceptualising readiness as a continuum (1, 2, 4, 7, 16), corresponding to the sum of the raw/mean contemplation, action and maintenance item scores, with the sum of the precontemplation score subtracted (‘Readiness Index’, Carbonari et al., 1994). Two studies used subscale scores (12, 14), one (15) used participants’ highest subscale score to categorise their stage of change and two (3, 17) used the BNSCS total score (Martinez et al., 2007). Two used the Motivational Interviewing Skills Code 1.1 (Glynn & Moyes, 2009) to derive measures of change talk and counter change talk (5, 11). One of these derived all ratings in session 1 (5). A second derived ratings from session 1 (78%) or 2 (22%) but was included as it also employed another measure of motivational readiness, obtained prior to the start of treatment (11). Two study used a Readiness Ruler (10-point rating of readiness to change; 9, 22) and two used a single item rating on a scale from 0 to 100 (13, 20). A range of other measures were used by one or two studies (FF-STABS: 8; Stages of Change Schedule: [19]; Stages of Change Questionnaire: 10, 21; Change Questionnaire: 11; Stage of Change Scale: 18).
Ten studies reported data on the internal consistency of the measure used (1, 2, 4, 5, 7, 8, 10, 17, 19, 21) and in most cases, reliability was adequate to good. More than half the studies provided no information on the internal consistency of the measure within the sample.
Association Between Readiness and Symptom Change
Nineteen studies (2, 3, 4, 12, 5, 6, 8, 10, 11, 13, 7, 14, 15, 16, 17, 18, 20, 21, 22) examined the relationship between readiness and outcome following CBT/ERP1, with a majority (n = 13, 68%) identifying at least one significant association.
A study of the relationship between readiness and outcome following two formats of guided self-help CBT for bulimia (17), found that those with higher pre-treatment readiness were more likely to achieve remission irrespective of treatment format. A study of CBT for eating disorders (20) reported an interaction between readiness, treatment condition (CBT-T or CBT-M) and outcome, such that participants with low readiness randomised to CBT-T, which did not address motivation for change, showed higher symptom burden at 1 and 3 month follow-up than either those with high readiness allocated to CBT-T, or those with low or high readiness allocated to CBT-M (which addressed motivation in early sessions). Another study of CBT-T showed that patients with atypical bulimia nervosa showed an interaction between baseline readiness to change and symptom reduction over time (higher readiness associated with greater symptom reduction) but that this effect was not present for those with atypical anorexia (22). In a study of transdiagnostic CBT for anxiety disorders, those with higher readiness showed greater improvements on measures of anxiety when baseline symptom severity was also high (2). Finally, an interactive effect was also observed in CBT for somatoform disorders (8): both precontemplation and action subscale scores interacted with symptom levels at baseline to predict post-treatment symptoms, with a range of other conditional effects observed (see Table 3).
Some studies adopted an exploratory approach. For example, a study of people receiving CBT for BDD (7) reported that baseline readiness was positively associated with treatment response. This was one of a total of three variables (of 14 examined) that showed a statistically significant associated with treatment response. A second study (3) found that higher baseline readiness was associated with a greater reduction in binges, and greater change on two of the 11 subscales of the EDI, with readiness being the only variable associated with clinical change in binges and decreases in the EDI bulimia subscale, after controlling for known confounders. Two other studies reported mixed findings. The first (6) identified an association between higher baseline readiness and a lower score on the EDE at end of treatment in adolescents with bulimia, but no association between readiness and abstinence from ED behaviours. The second (15) found that (a more advanced) stage of change was associated with greater change in binge frequency at end of treatment, but not frequency of vomiting or laxative use.
The above studies employed self-report measures. However, others captured participant in-session behaviour. One study (11) focused on anxiety disorders demonstrated that participants with higher readiness at pre-treatment had lower levels of post-treatment worry. Additionally, after accounting for self-reported readiness, increased counter-change talk in early treatment was associated with higher post treatment worry and a reduced chance of reliable symptom change. Extending these findings, a second (5) showed that after controlling for self-reported readiness, greater session one change talk was associated with an increased chance of clinically significant response, and greater counter-change talk, a decreased chance.
Six studies did not report any significant associations between readiness and clinical outcome. A study of people receiving ERP for OCD (14), identified no association between readiness and post-treatment symptoms, whilst another exploring inpatient CBT for OCD (12) identified no association between readiness and symptom change. A third (16) found no association between readiness and outcome in people with OCD receiving ERP augmented with CBT strategies and a fourth (18) found that stage of change was not associated with remission in people receiving CBT for bulimia nervosa (although it was associated with remission for the overall sample combining CBT and IPT, and for those receiving IPT only). A fifth found no association between initial stage of change and symptom reduction and post-treatment diagnostic status in people diagnosed with IED (21). Finally, a study of people receiving CBT for EDs found that readiness did not predict change in EDE-Global Score (13).
Association Between Readiness and Treatment Attendance
Five studies (1, 9, 13, 14, 17) explored treatment attendance and found only limited evidence of an association with readiness. One (1) identified a significant association between greater readiness at admission and longer inpatient stay, but no significant association with number of CBT sessions attended. A second (14) compared those who completed ERP for OCD with those who dropped out on the four URICA subscales (precontemplation, contemplation, action, maintenance) and found those who dropped out scored significantly lower on the contemplation subscale than those who remained in treatment but did not differ on the other subscales. Of the remaining studies, one (9) identified no significant association between readiness and number of sessions of CBT attended, initial engagement with CBT or dropout from treatment, whilst two (13, 17) found no association between baseline readiness and dropout.
Association Between Readiness and Outcome in Randomised Controlled Trials
None of the seven studies that employed an RCT design (5, 6, 10, 15, 17, 18, 20) included a waitlist arm or no-treatment control arm. As a result, it was not possible to compare the association between readiness and outcome in those receiving CBT and those experiencing spontaneous symptom change (research question 2). Five of the seven RCTs compared CBT to another therapeutic intervention: Interpersonal Psychotherapy (18), Family-Based Therapy (6), Motivational Interviewing combined with CBT (5), Fluoxetine or Placebo (10), Motivational Enhancement Therapy (15), whilst one (17) compared two forms of guided self-help CBT, and one two forms of CBT (20). All five comparing CBT to another intervention identified at least one significant positive association between readiness and outcome across the whole study population. However, exploring research question 3, none of the four which tested for a significant readiness by treatment group interaction (5, 6, 10, 15) identified one, providing little evidence that readiness influences outcome differently in CBT and other forms of therapeutic intervention when examined in the same clinical population. The fifth study, (18) examined the association between stage of change and outcome separately for participants receiving IPT and CBT and identified a significant association only in the IPT group. The authors suggested that CBT may be better suited to addressing ambivalence to change than IPT and hence the association between readiness and outcome (which was observed for the sample overall) may have been attenuated in the CBT group. However, in the absence of a test of interaction it is not possible to conclude that the association was significantly stronger in the group receiving IPT than CBT, simply that it was observed to reach a conventional level of significance only for those receiving IPT.
Discussion
This review considered studies examining the relationship between pre-treatment RTC and subsequent attendance or outcome of CBT for mental health problems other than addiction. A systematic search identified 22 eligible studies, with 21 included in the narrative synthesis.
Nineteen studies explored the relationship between RTC and symptom outcome with 13 identifying a statistically significant relationship, either directly or in interaction with another variable, consistent with a positive relationship between readiness clinical outcome. A number also tested other associations between RTC and outcomes, which were non-significant. Few explored the association between RTC and treatment attendance and findings were inconclusive. It has been suggested that RTC may be more relevant when the problem in question has clear behavioural markers (Drieschner et al., 2004), and the TMSC has historically been applied most to addictions. However, there was no clear evidence to suggest that findings were more consistent for studies involving some clinical populations than others. Likewise, those studies which compared the association between RTC and outcome in people randomised to CBT versus to another non-CBT therapeutic intervention (5, 6, 10, 15, 18) found no clear evidence indicating that RTC had a differential impact on treatment response across a range of interventions, suggesting that any mechanisms are likely to be relatively non-specific.
Across the studies in the review, female participants were in the majority, with several studies recruiting only female patients. Likewise, where ethnicity data were reported, most participants were Caucasian and for a significant proportion of studies ethnicity data was not reported at all. The review was restricted to English language studies, with almost all originating in North America or Western Europe. It is established that both gender (e.g., Wendt & Shafer, 2016) and ethnicity (e.g., McGuire & Miranda, 2008) influence patterns of treatment seeking for mental health problems, and it is likely that these factors may also influence the processes through which individuals recognise and take steps towards change and interact with treatment, as well as the relative accessibility of services. None of the studies in the review explored the role of gender or ethnicity as they pertained to the relationship between RTC and treatment outcome, and this remains an important avenue for future work, in line with efforts to increase ethnic diversity in research populations (Farooqi et al., 2018).
The 21 studies included in the narrative synthesis employed 9 different measures of RTC. Only around half reported information on the internal reliability, and few reported convergent or discriminant validity. Some used methods, for example assigning participants to a stage of change based on an algorithm comprising yes/no questions (18) or according to the URICA subscale on which they scored highest (15), which although widely employed within TMSC research have subsequently had their validity questioned. For example, research suggests that only a minority of people fall into a single stage of change, and so such classifications may be quite arbitrary or reductive (e.g., Littell & Girvin, 2002). Likewise, the typical method for deriving the ‘readiness index’ involves adding maintenance scores to contemplation and action scores, before subtracting pre-contemplation scores (employed in studies 2, 4, 7, 16). It has been suggested that where individuals are scoring highly on the maintenance facet, the overall measure may not really be capturing ‘readiness’ to change, but also the process of making changes and protecting those that have been made. Within the context of this review, which focused on people with recognised mental health problems at the start of treatment this is likely to be less of a concern but remains a broader issue for the literature. Additionally, most studies relied on self-report, which is likely to be subject to demand effects. The two studies that employed analysis of client in-session speech (5, 11) provided evidence to suggest that this had incremental explanatory power. Although labour-intensive to code, it would be useful to conduct further research using these novel methods. Finally, all studies explored RTC in relation to a single problem area, whether specified (‘worrying too much’) or unspecified (e.g., ‘the problem’). It is possible that people may vary in their readiness to make changes in different problem areas, for example being ready to reduce purging, but not to gain weight. Exploring more nuanced aspects of RTC, including potential conflicts between different aspects of change, may be an interesting line of future work.
Studies were included on the basis that they delivered a CBT intervention (or for OCD spectrum disorders ERP/CBT). Many provided references to specific treatment protocols, or described session content, but the dose of CBT delivered varied considerably from study to study depending on the context in which the study was conducted. In most cases CBT was the only treatment offered, but particularly in inpatient settings and in eating disorder samples, it was typical for people to also receive medication management, milieu experiences or adjunctive therapies (studies 1, 3, 4, 9, 12). To support inclusion of more naturalistic samples, where there may be greater variation in RTC, and to broaden generalisability of review findings, such studies were included. However, it is important to bear in mind that in studies including adjunctive interventions, RTC will potentially have influenced engagement with these interventions and these interventions will potentially have influenced outcome. Likewise, where CBT therapists address motivation formally (e.g., as part of the Unified Protocol, or in some forms of CBT for eating disorders) or informally, the relationship between pre-treatment RTC and outcome may be attenuated (e.g., Wade et al., 2021).
Relatively few studies described who collected baseline data and it was often unclear whether therapists were blind to participant ratings. Having such knowledge may be advantageous – encouraging the therapist to adapt their approach to the client – or disadvantageous – if it sets up expectations about likely client outcome. Likewise, it was frequently unclear whether those gathering post-treatment data would have been blind to baseline measures of RTC. Whilst most studies employed either self-report questionnaires as outcomes or utilised data on session attendance, it remains possible that in the absence of blinding, bias was introduced.
Review Limitations
This review included studies which assessed RTC in ways aligned to the TMSC. Whilst the search terms and range of databases used were designed to create a broad initial search, it is nevertheless likely that some potentially eligible studies will have been missed. Likewise, other conceptualisations of readiness and motivation for treatment exist and review findings would have been altered significantly if a different definition had been employed. For pragmatic reasons, the review did not include 100% independent double rating for titles/abstracts, data extraction or study quality, although this would have enhanced methodological rigour. Finally, it was not possible to conduct a meta-analysis and create a forest plot and look for evidence of publication bias due the heterogeneity of the studies included. However, since many of the published studies comprised secondary analyses, it is highly likely that such a bias exists and that other exploratory analyses with non-significant findings remain in the ‘file draw’ (Rosenthal, 1979).
Clinical Implications
RTC is often conflated with treatment motivation (Carey et al., 1999) although these are distinct constructs. It is likely to be helpful for therapists to explore RTC, treatment motivation and treatment expectancy at the outset of a course of treatment, in order to identify potential barriers to engagement. The review also focused on people’s pre-treatment readiness rather than the impact of any interventions explicitly intended to increase client motivation, and it is unclear to what extent spontaneous RTC equates to readiness promoted through specific motivational techniques. Most of the studies reviewed here identified either no association or a positive association between RTC and outcome of CBT. However, positive associations were only observed in around two thirds of studies, and often only observed in relation to specific outcomes. This suggests that whilst greater RTC may be a positive prognostic indicator, lower RTC at the outset of therapy may not always be a cause for concern in CBT treatment of mental health problems other than addiction. Indeed, two recent studies which have examined in-session change talk and sustain talk across the course of metacognitive therapy for generalised anxiety disorder (Lassen et al., 2022) and depression (Lassen et al., 2024) suggest that it is increases in change talk over time, rather than session 1 change talk, that correlate with better end of therapy outcomes. Therapists may therefore wish to consider how clients are talking about change as therapy progresses, focusing on and promoting client recognition of their ability to change when treating common mental health problems.
Conclusions
The review identified some evidence of an association between RTC and outcome in CBT for common mental health problems other than addiction, but also highlighted methodological weaknesses and uncertainties affecting the research area. Despite critique of the TMSC, the concept of RTC is clearly still regarded as a useful heuristic for considering people’s orientation to their problems at entry to treatment. Further high-quality research in populations with common mental health problems other than addiction is required to establish whether that the construct of RTC is useful in predicting CBT treatment response within treatment seeking populations and through what mechanisms it might exert beneficial effects.
Declarations
Conflict of Interest
The authors have no conflicts of interest to declare.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
