Negative experiences in psychotherapy from clients’ perspective: A qualitative meta-analysis

ABSTRACT Objective: A considerable number of clients report adverse or unwanted effects of psychological treatments. This study aimed to synthesize the findings of qualitative studies focused on what clients perceive as negative experiences in psychotherapy. Method: A database search was conducted to find primary studies, and a qualitative meta-analysis was used to aggregate the findings on the kinds of negative experiences psychotherapy clients reported. Results: A total of 936 statements were extracted from 51 primary studies and categorized into 21 meta-categories, some of which were further divided into subcategories. These meta-categories covered clients’ experiences, which fell into four broad clusters: therapists’ misbehaviour, hindering aspects of the relationship, poor treatment fit, and negative impacts of treatment. Conclusion: Clients’ negative experiences of psychotherapy are a vast and heterogeneous area, the breadth of which is not captured by any single study. By synthesizing the findings of many primary studies, this meta-analysis represents the most comprehensive summary of these experiences to date.

Early in the history of quantitative psychotherapy research, Bergin (1966) observed increased variance in treatment outcomes compared to no treatment.He suggested that there was something unique about therapy that resulted in improvement for some and deterioration for others and that further research was urgently needed (Bergin, 1971).Since then, many studies have examined factors that contribute to improvement and deterioration.One essential line of research involves the study of clients' personal experiences with therapy (Bohart & Tallman, 2010;Ladmanová et al., 2022;Levitt et al., 2016).Although more attention has been given to clients' perception of good outcomes (e.g., Binder et al., 2010;Elliott & James, 1989;McElvaney & Timulak, 2013), there has been increased awareness of negative effects in psychotherapy among researchers in recent years (Rozental et al., 2018), including a call for the application of standardized approaches to assess and manage harm (Klatte et al., 2022).In one of the first studies of patients' negative experiences, Kepecs (1979) conducted a thematic analysis of 10-minute therapy segments.He identified both therapist errors (e.g., dominating or avoiding the patient) and patient expressions regarding the therapist's lack of warmth, connection, and genuine interest.Until recently, however, despite calls for exploratory studies on this topic (e.g., Elliott & James, 1989;Rennie, 1994), research on clients' firsthand experiences, especially negative experiences, has remained sporadic.
In the last two decades, the research landscape has changed substantially.Not only are there several dozen qualitative studies that explore clients' negative experiences, but there are also quantitative studies that use standardized measures of negative effects, such as the Negative Effects Questionnaire (NEQ; Rozental et al., 2016) and the Inventory for the Assessment of Negative Effects of Psychotherapy (INEP; Ladwig et al., 2014), as well as mixedmethod studies that use standardized coding systems to categorize open-ended descriptions of clients' experiences (e.g., Castonguay et al., 2010).As a result, we believe the time is right for a study that accumulates and aggregates these studies in an effort to identify dominant themes.Before turning to the methods for our meta-analysis, we reflect on the definition and nature of negative experiences in psychotherapy and the use of client first-person expressions of therapy in psychotherapy research.

Definition and Nature of Client Negative Experiences
We are aware of the complex nature of psychological care.It might be natural that clients report a range of positive and negative experiences during psychotherapy and/or termination.Clients' subjective reporting of negative impacts or feelings does not necessarily correspond to a deterioration of clients' mental health status; some negative feelings might be a natural side effect of a curative process or even part of a positive outcome.On the other hand, negative emotions can be detrimental to the psychotherapy process on their own.They may be perceived by clients as destructive to the therapeutic relationship (e.g., Haskayne et al., 2014).
Furthermore, clients evaluate their satisfaction with therapy depending on what they are looking for, what they need, and what they expect.Unmet expectations (Curran et al., 2019;Westra et al., 2010) can negatively impact therapy.As von Below and Werbart (2012) noted, "being dissatisfied with psychotherapy is a multifaceted phenomenon, where different hindering experiences interplay and reinforce each other, forming what we came to call a vicious circle of dissatisfaction" (p.220).
For our study, we choose the term "negative experience" as an existential phenomenon that covers both the widest and deepest human inner state based on feelings, thoughts, and relational and nonverbal perceptions.Terms used in several research and theoretical studies include perceptions of therapeutic ruptures (Haskayne et al., 2014), perceptions of hindering or unwanted events (Castonguay et al., 2010;Linden, 2013;von Below & Werbart, 2012), impeding factors (De Smet et al., 2019), detrimental impacts and deterioration (Rozental et al., 2018), or harmful (experiences of) therapy (Hardy et al., 2019).As Rozental et al. (2018) stressed, negative effects "lack a clear and coherent terminology" (…), [so] "no consensus currently exists as to what events need to be explored" (p.308).One possible response to this problem is to develop a taxonomy of client-reported negative experiences as they occurred in the many qualitative studies in the field.

Clients' Perspective
Openness to patients' views of psychotherapy has been highlighted for many years (Bohart & Tallman, 2010;Llewelyn, 1988;Llewelyn & Hume, 1979, 1988;Strupp et al., 1969).Strupp et al. (1969) were among the first to make the case for listening to patients' views of therapy, even though some suggested that patients' reports may be distorted by self-deception, biases, and wishful thinking, inaccurate when judged against objective criteria, or reflect unresolved entanglements with the therapist or other significant people in the person's life.Importantly, it was emphasized that both participants in psychotherapy bring into treatment their patterns of behaviour, their reactions, transference and countertransference, which could lead to negative interpersonal and emotional experiences (Binder & Strupp, 1997).Rozental et al. (2018) cautioned, however, that "determining whether negative effects have occurred during treatment largely depends on what perspective is endorsed.Incidents might, for instance, be considered adverse and unwanted by a patient, but deemed an inevitable part of certain interventions by a clinician" (p.309).As a result, it is important to consider a variety of perspectives when gathering data about the effects of therapy (Ogles, 2013).In this study, we focus specifically on the client's view of negative experiences in psychotherapy.
Obtaining a detailed client's perspective is not easy.Clients may feel shame associated with talking about their therapy (Grafanaki & McLeod, 1999) and do not easily talk about difficult experiences in therapy (Henkelman & Paulson, 2006).On the other hand, De Smet et al. ( 2019) stressed, that "patients tend to be more critical about therapy during interviews" compared with findings from quantitative studies (p.2).
In spite of the difficulties and the potential issues with a single perspective, the study of client experiences has become more common in recent years.These studies differ in their breadth of scope.While some studies have investigated clients' psychotherapy experience in general and reported both positive (helpful) and negative (hindering) categories (e.g., Bury et al., 2007;Dale et al., 1998), other studies have focused specifically on negative experiences (e.g., De Smet et al., 2019;Hardy et al., 2019).Yet another group of studies has focused on more narrowly defined types of experience, such as nonimprovement (Bear et al., 2022), misunderstanding events (Rhodes et al., 1994), and clients' deference (Rennie, 1994).

Systematic Reviews with a Similar Focus
We found two prior reviews that covered similar material with different purposes and/or methods.Curran et al. (2019) adapted the task analysis methodology to identify factors leading to harmful impacts of psychotherapy.Using experts, they first developed a rational-empirical model based on identifying key change events and factors leading to negative effects.Clients' testimonies were then obtained from a review of published articles along with blogs, discussion boards, and book chapters.They selected only studies that reported adverse processes and effects in the greatest detail and depth and that facilitated the task analysis.The central findings showed how concrete contextual issues, negative processes, relationship patterns and therapist behaviours with clients' unmet expectations are feeding adverse or harmful effects of psychotherapy.
In their qualitative meta-analytic study of clientidentified impacts of helpful and hindering events, Ladmanová et al. (2022) identified 12 helpful and eight hindering meta-categories regarding the impact of psychotherapy.Hindering impacts identified in their analysis were the most relevant to our study and included feeling disconnected from the therapist, having difficulty disclosing, feeling emotionally overwhelmed, having difficulty accomplishing the therapeutic task, lacking guidance from the therapist, feeling pressured, feeling aggrieved/ hurt, and feeling confused by the therapist (Ladmanová et al., 2022).

Aim of the Study
Given the consistent finding of some negative effects of treatment, the importance of considering the client's perspective of therapy, and the recent accumulation of qualitative studies about client negative experiences, we conducted a qualitative meta-analysis to aggregate and synthesize findings from primary qualitative studies.This study aimed to develop a comprehensive taxonomy of clientreported negative experiences in psychotherapy.To achieve this aim, we conducted a qualitative metaanalysis of primary studies that either focused solely on clients' negative experiences or contained descriptions of such experiences as a part of a broader scope (e.g., psychotherapy experiences in general).

Selection of Studies
We conducted a database search to collect primary studies on clients' negative psychotherapy-related experiences using the following databases: PsycArticles, PsycInfo, Academic Search Ultimate, Academic Search Complete, Web of Science, and Scopus.See Supplement 1 for the search string.The search was conducted in June 2021 and was limited to (a) peer-reviewed articles (b) classified in the field of psychology and (c) written in English, Spanish, Czech, or Slovak.
After removing duplicates, we screened titles and abstracts of the primary studies to identify those that (a) focused on psychotherapy or counselling (i.e., we removed studies that investigated other forms of intervention, such as alternative therapies and medical interventions), (b) were conducted from the client's perspective (i.e., we excluded studies that employed only therapists' perspectives), and (c) reported findings of a qualitative or a mixed-method study (i.e., we excluded quantitative studies, reviews, and meta-analyses).
Subsequently, we sought full texts of the identified articles and screened them for the following criteria: (a) the study focused on face-to-face individual psychotherapy (i.e., we excluded other therapeutic settings, such as group and couple therapy, other Psychotherapy Research 281 modes of delivery, such as online and telephone, and other kinds of treatment, such as art and music therapy), (b) the study explored clients' negative psychotherapy experiences or dissatisfaction (i.e., we excluded studies focused exclusively on positive and/or neutral experience with psychotherapy), (c) the study proceeded inductively in the identification of types of experience (i.e., we excluded studies using predefined categories to code clients' negative experience), and (d) the study was based on more than one participant (i.e., we excluded single-case studies because, typically, they do not systematically distinguish clients', therapists', and researchers' perspectives and are selective in reporting clients' firsthand experience).
Furthermore, we examined the reference lists of all eligible studies to identify other potentially eligible primary studies.Those that could be retrieved in full text were then screened for eligibility using the abovementioned criteria.See Supplement 2 for the flow diagram.

Data Preparation and Analysis
We followed the procedures and recommendations delineated by Timulak (2009;2013;Timulak & Creaner, 2022) and Levitt (2018;Levitt et al., 2018).Qualitative meta-analysis (QMA) offers an empirically-founded, systematic approach to synthesizing, interpreting, and drawing conclusions from the findings available across individual primary qualitative studies (Timulak, 2009(Timulak, , 2013;;Timulak & Creaner, 2022).To represent the clients' perspective as faithfully as possible, we chose a more descriptive approach to QMA, remaining at a semantic (explicit) level as much as possible and not speculating at a latent or interpretative level (Braun & Clarke, 2006).
First, we extracted the relevant parts of the results sections of the primary studies.If a study contained parts unrelated to our research question (e.g., clients' positive experiences of psychotherapy or therapists' experiences), we disregarded them.The meaning units that we extracted from the primary qualitative studies included prior researcher-derived themes (categories) and their descriptions, as well as direct quotations from interviews used by the primary studies' authors to illustrate the themes.Second, we divided the data into meaning units (i.e., extracts containing a discernible idea related to the research question) and numbered them.While meaning units varied in length, most of them were short phrases (e.g., "I've got some kind of performance anxiety that I never had before"; Werbart et al., 2015, p. 554).Third, two authors (Z.V. and B. O.) proceeded independently with the following steps: (1) they read all meaning units and attached a code to each of them capturing the essence of the meaning, and (2) they sorted the meaning units into provisional meta-categories.The analysts proceeded in a bottom-up manner in categorizing the meaning units and refrained from using any preexisting categorization scheme.Fourth, all authors met twice to discuss the coding and categorizations and developed a consensual classification taxonomy based on a combination of the two individual classifications as well as the input of the remaining two coauthors (B.U., T. R.).Fifth, Z. V. went through the list of meaning units and recategorized them based on the new system of meta-categories.B. O. then reviewed those ratings and identified 64 meaning units where the two analysts disagreed.These were resolved through comparison and discussion, and when necessary, T. R. helped to resolve discrepancies by providing his opinion on the suggested classifications and/or facilitating the discussion.Repeated team discussions served as a means to enhance the analysts' reflexivity and allowed them to be aware of their personal perspectives that entered the analytic process.

Reflection on Analysts' Background
(1) Z. V. is a psychologist with 14 years of fulltime or part-time practice in psychotherapy and a long time as a faculty member.His training was integrative and based on psychodynamic group therapy.He authored and coauthored articles on the negative effects of psychotherapy using the NEQ or single case-study method.He is also the author or coauthor of two theoretical articles and book chapters on risky approaches in psychotherapy.
(2) B. O. is a faculty member who teaches and conducts research about psychotherapy and has a very limited private practice.His training was rather eclectic, with a primary emphasis on the humanistic foundation for the therapeutic relationship and CBTbased interventions when appropriate.He has a strong interest in common factors as a route to client change.
(3) T. R. is a psychologist and therapist with 15 years of part-time therapeutic practice and interest in both qualitative and quantitative research.He was initially trained in Gestalt therapy and was considerably influenced by the psychotherapy integration movement.
He is a coauthor of several studies on helpful and hindering events in psychotherapy.
(4) B. U. and V. G. were graduate students of psychology and had no experience working with clients.

Description of Primary Studies
We extracted data from 51 primary studies based on 48 independent datasets (for the purpose of this description, we treated studies based on the same dataset as a single study).The studies were published between 1986 and 2022 (half of them in the last decade).Most studies were localized in the UK (n = 17), followed by the US (n = 10), Sweden (n = 6), Canada (n = 6), Belgium (n = 2), and Chile (n = 2); other countries (Argentina, Colombia, Czech Republic, Denmark, Germany, Ireland, Israel, New Zealand and South Africa) were represented by no more than one study.Altogether, they were based on a sample N = 1267 clients.The sample sizes of the individual studies varied between 4 and 185, with a median of 15 clients.Most studies were based on client reports, but four included therapists reporting on their experiences as clients, and one included reports by clients' caregivers or other family members.In terms of treatment approach, most studies reported a mixture of treatments (n = 17) or did not report the treatment approach (n = 16).Studies focused on a single treatment approach included cognitive-behavioral (n = 6), psychoanalytic (n = 5), psychodynamic (n = 3), and integrative (n = 1) approaches.
While most studies focused on clients' experiences in general or were framed as studies on both positive (helpful) and negative (hindering) experiences (n = 24), some studies focused exclusively on the negative side of the spectrum (n = 14), and the remaining studies focused on more narrowly specified topics, such as client deference, barriers to adherence, treatment termination, and therapist disclosure.In terms of data collection, most studies relied on interviews (n = 37), some collected written open-ended responses (n = 6), and the rest used a combination of methods (e.g., interviews plus session transcripts).Typically, the data were collected at some interval after treatment termination (n = 26), but some studies collected data within treatment (n = 4) or combined the two strategies (n = 1).However, many studies did not provide this information (n = 17).The time that elapsed from the treatment termination varied between 0 and 42 years (Mdn = 1.5 years).In terms of data analytic methods, most studies reported using grounded theory (n = 11), some form of phenomenological analysis (n = 10), thematic analysis (n = 9), consensual qualitative research (n = 4), and, less frequently, other methods (e.g., discourse, framework, heuristic analysis or nonspecified categorization/coding).See Supplement 3 for the list of primary studies.

Meta-Categories
We extracted 936 statements from the primary studies (see Supplement 4).Based on a qualitative meta-analysis, we formulated 21 meta-categories grouped into four clusters (see Supplement 5).Some of the meta-categories were divided into subcategories.In naming and describing the meta-categories and subcategories, we strived to remain as close to the clients' perspective as possible.
Cluster I: therapists' misbehavior.Therapist Not Listening (n = 9, 17%).Some clients expressed doubts about whether their therapist was listening: "My God, is she at all listening to what I'm saying?" (Nilsson et al., 2007, p. 562).Such an experience of feeling unheard may result in clients thinking the therapist does not perceive them as significant human beings.Other clients used expressions such as an "absence of genuine interest" (Bowie et al., 2016, p. 82) and felt that the "counsellor didn't really care" (Paulson et al., 2001, p. 58).
Therapist Not Understanding (n = 19, 37%).This meta-category describes instances of feeling misunderstood even though the therapist seemed to be paying attention.For instance, Paulson et al. (2001) reported an experience of "saying something and having the counsellor summarize it differently than I want" (p.57), and Rhodes et al. (1994) reported clients' feelings that the therapist's interpretation was off.In more extreme cases, clients may feel they "hit a wall" when their therapist does not seem to understand the situation they are explaining (Knox et al., 2011, p. 158).
Therapist Perceived as Incompetent (n = 19, 37%).Clients experienced discontent when they perceived their therapist as incompetent or inexperienced: "You couldn't get a sense that she could handle it really" (Hardy et al., 2019, p. 410).The perception of incompetence may emerge a priori or during treatment due to the therapist's response to the client's nationality, gender, religion, or sexual identity.For instance, therapists' lack of cultural competency or sensitivity was explicitly mentioned.In one instance, the therapist was perceived as having been "trained" in cultural competencies only by stereotypes shown on TV: "She had made up her mind from the way we are usually shown on TV as Arab or Muslim men, as basically being female oppressors; […] she just simply seemed to Psychotherapy Research 283 have this negative view" (Al-Roubaiy et al., 2017, p. 467).
Therapist Devaluing the Client (n = 29, 56%).Several types of therapist behaviours were viewed by the client as devaluing.First, clients reported feelings of rejection when therapists either disregarded clients' issues or rejected them (e.g., Bess & Stabb, 2009;von Below & Werbart, 2012;Israel et al., 2008).Second, clients reported disrespect.They felt denigrated, admonished, or perceived the therapist as hostile.This was attributed to a therapist's characteristic or emotional state such as appearing annoyed, resentful, smug, or offended.As one client said, "I had challenged him and he made me feel really small" (Bowie et al., 2016, p. 83).Third, in four studies, clients referred explicitly to feeling humiliated.For some clients, their typical experience was being viewed more as a thing than a person (Nilsson et al., 2007) or "feeling like just another statistic to the counsellor" (Paulson et al., 2001, p. 57).Some clients experienced the counsellor as violating (Koehn, 2007).Fourth, clients felt that therapists imposed their values on them.Clients described therapists who trivialized clients' issues and felt that their problems were discounted.One client described the therapist as having "an air of superiority" (Bowie et al., 2016, p. 83).Perceived devaluing can be tailored into a complex suit.As one client said, the therapist was "talking at me, judging me, and not respecting me" (Cragun & Friedlander, 2012, p. 385), and "it was a negative turn off for me.It was like, 'I can't help you' kind of thing" (p.386).
Therapist Judging (n = 17, 33%).Some clients perceived the therapist as judgmental.In nine studies, the perception of judgment was specifically associated with the client's religion, sexual preference, or sexual identity.Bess and Stabb (2009) mentioned that transgendered clients experienced some previous therapists as heterosexist, sexist, having pathologizing biases, or providing hostile responses.Dismissive judgments impacted client's well-being in therapy, as in the following example: "Therapist told the client that she was 'too Catholic' which made the client feel bad" (Knox et al., 2005, p. 296).In other cases, biases and judgments led to invalidation of the client's identity (Israel et al., 2008) or disaffirmed the client's sexual orientation (Victor & Nel, 2016).
Therapist Using Client for Own Benefit (n = 14, 27%).In five studies, clients reported therapists' sexual involvement.They experienced unwanted compliments or touch or felt like the therapist was leering.Second, clients reported violations of nonsexual boundaries, including dual relationships (Grunebaum, 1986;Knox et al., 2011).They reported that therapists revealed intimate details from their sexual lives or felt used in a nonsexual way, such as the therapist trying to become the client's friend (Paulson et al., 2001).
Third, clients reported that confidentiality was broken when boundaries were minimized or violated within sessions (Dale et al., 1998;Koehn, 2007).Fourth, clients described therapists' self-disclosure for their own benefit.In such cases, the clients perceived their therapist as using the therapeutic situation to meet their own needs or satisfy their own desires.Such therapists neglected what clients needed in the moment.However, an opposite type of experience was also reported: "[Patients] mentioned that they missed knowing something about the therapist's personal life, that the therapist was not allowed to be seen as a person" (Poulsen et al., 2010, p. 475).
Cluster II: hindering aspects of the relationship.Experiencing Distance and/or Lack of Empathy (n = 23, 44%).This meta-category captures the client's sense that the therapist was detached or disconnected from the therapy and the client.Perceived distance from the client's perspective was created in several ways, such as the use of pathologizing diagnostic labels (Bess & Stabb, 2009), needing more sensitivity (Nilsson et al., 2007), and perceiving the therapist as a person who was not collaborating with them (Bowie et al., 2016).For example, one client expressed, "I was disappointed that he didn't [and] wasn't able to relate to me!" (Pope-Davis et al., 2002, p. 381).
Experiencing Insecurity or Distrust (n = 11, 21%).Trust was perceived as a fundamental basis for therapy; consequently, distrust was framed as one of the key obstacles.As one client said, "In fairness, there was very little trust between me and the psychologist, and I don't think they really stressed the confidential nature of what they were doing; by this time, I was quite paranoid and you know I was petrified" (Hardy et al., 2019, p. 411).Distrust also arose when the client felt pushed to explore or disclose risky topics (Israel et al., 2008).Sometimes clients simply felt "general" distrust: "I don't trust him anymore in fact I feel really uncomfortable with him" (Bowie et al., 2016, p. 83).
Experiencing Confusion or Uncertainty (n = 12, 23%).Some clients felt confused, annoyed, or upset.As one client said, "I didn't understand the game.I didn't understand the rules" (Dale et al., 1998, p. 150).Some clients felt bad or uncomfortable without knowing why or how.Confusion and uncertainty were also sometimes experienced between sessions: "At the end of most of the sessions, I was left with questions … the whole week I'm thinking what does she mean by asking me this?Why did she pick on that?" (Wilson & Sperlinger, 2004, p. 227).
Experiencing Poor Interpersonal Match with the Therapist (n = 13, 25%).Sometimes, both the therapist and the client became stuck in an "negative interactional pattern" from which change seemed impossible (Hardy et al., 2019, pp. 411-412).For this type of situation, clients used metaphors such as "lack of chemistry" (Roe et al., 2006b, p. 533) or the therapist being "not at the same plane as [the] patient" (Knox et al., 2011, p. 158).The reflection of a poor match was also formulated in such a way that the therapist may have been effective for other clients, though perhaps the wrong therapist for the specific client.
Cluster III: poor treatment fit.Negative Evaluation of Practical Aspects of Therapy (n = 17, 33%).Clients expressed objections to procedural or practical aspects of treatment, such as the length of treatment.Other examples included raising the fee during therapy (Grunebaum, 1986); late therapist arrivals (Goldman et al., 2016); the counsellor not accommodating the client's work hours (Paulson et al., 2001); receiving treatment they did not want (Hardy et al., 2019); and therapists taking phone calls during the session (Hardy et al.).
Unmet Expectations (n = 17, 33%).Some clients felt frustrated or disillusioned because their expectations about therapy were not met or were violated by the therapist (Westra et al., 2010).Clients stated that their problems had not been adequately resolved: "I found the therapy too focused on food […] as opposed to exploring feeling and behaviours.I would have liked to explore why I do this to myself" (Zainal et al., 2016, p. 6).Some clients felt that their therapy did not address their childhood issues that were at the core of their difficulties (Omylinska-Thurston et al., 2019) or did not "dig" deep enough (De Smet et al., 2021).Unmet expectations can have a long-term impact.Sometimes, however, clients made excuses for their therapist even when disappointed: "But it's no one's fault.But it's not the therapist's fault" (Westra et al., 2010, p. 7).
Lack of Fit with the Intervention (n = 34, 65%).Clients may view the psychotherapy process either as too structured or too unstructured.On the "too structured" side (in ten studies involving various types of therapies), clients described the perceived inflexibility of some approaches, unwanted directivity from the therapist, and a treatment structure imposed with pressure or in a confrontational manner (Poulsen et al., 2010).On the "too unstructured" side (in fifteen studies based mostly on psychoanalytic therapy), clients complained about too much silence and therapists' passivity.Some patients wanted more "concrete advice" on how to handle a troublesome situation (Lilliengren & Werbart, 2005) and lacked a distinct and problem-oriented focus (Werbart et al., 2015).Some patients in psychodynamic psychotherapy were displeased that they had not gained the right tools or were not given any directions (De Smet et al., 2019).As another client said, "I needed the structure.I needed someone to say: 'Don't do this.Do that!'" (Wilson & Sperlinger, 2004, p. 227).
Dissatisfaction with Therapy Ending (n = 26, 50%).Many clients (in 26 studies) perceived treatment termination as a vulnerable moment that posed a threat in several aspects.First, a range of factors may cause inappropriate and misprocessed endings, e.g., organizational pressure (Bear et al., 2022).Second, some clients felt the treatment termination was premature, leaving them with a "fear that I will not be able to spread my wings and fly without my therapist" (Roe et al., 2006a, p. 77).Third, clients reported feelings of loss and abandonment, including feelings of pain, sorrow, deep sadness, or disappointment.
Increased Problems after Therapy (n = 19, 37%).In 19 studies, therapy resulted in deterioration of mental health or increased clients' interpersonal problems.Typically, clients stated that while they wanted to get better, they got worse in the end.Examples included statements such as "I became much more depressed" (Grunebaum, 1986, p. 168).This experience sometimes made clients feel confused: "It can't be possible that I voluntarily wanted to get better and I got even worse in the end" (Polakovská & Vybíral, 2018, p. 122).
Feeling Fear of the Therapy Process (n = 26, 50%).In 26 studies, clients experienced fear of the therapy process and felt threatened by a loss of Psychotherapy Research 285 control.Consequently, this made them less willing to self-disclose in psychotherapy and hindered their trust in the treatment (Hoff & Sprott, 2009).
Loss of Motivation or Hope; Resignation (n = 12, 23%).As a consequence of unsuccessful therapy, some clients were less motivated to continue their treatment (De Smet et al., 2019;Hardy et al., 2019) and hesitated to seek psychological help again (as mentioned explicitly in 12 studies).Some clients were resigned: "To be honest with you, I could not sit down now and go through all what I went through again" (Hardy et al., 2019, p. 408).As one LGB person said, "I'll never go back to a straight counsellor" (Victor & Nel, 2016, p. 358).
Unpleasant Feelings During Therapy (n = 31, 60%).This meta-category included five subcategories.First, arousal without resolution and the state of being overwhelmed were reported when clients "had no skills to deal with" apparent emotions (Hardy et al., 2019, p. 410).Second, clients mentioned feeling hurt (e.g., desperate, retraumatized, broke), including various impacts: "I […] came out and it was where I had to put my dark glasses on because I had been crying that much" (Goldman et al., 2016, p. 294).Third, some clients felt angry or resentful toward their therapist and sometimes also at themselves (Krause et al., 2018).Fourth, some clients felt dependent on therapy or therapists or experienced a variety of other unpleasant feelings (e.g., being cheated, disgust, hate, feeling quite paranoid).Clients very often (in 22 studies) described various emotional discomfort caused by the therapist's behaviour, such as feeling ashamed, frustrated, annoyed, "absolutely knackered" (Hardy et al., 2019, p. 411) or "became numb" (Koehn, 2007, p. 48).
Negative Cognitions Aroused in Therapy (n = 18, 35%).Clients reported various negative cognitions attributed to their treatment.The most prominent group of these cognitions included selfaccusation and self-blaming.Some clients blamed themselves for the failure of therapy (Radcliffe et al., 2018) and thought that perhaps they had said or done something wrong (Koehn, 2007).Across 13 studies, we found echoes of this self-critical inner voice of "being guilty" simultaneously justifying the therapist's behaviour or unhelpful therapy: "It could have been a lot better if I'd tried harder" (Omylinska-Thurston et al., 2019, p. 459).Sometimes, blaming both the therapist and the self were present (Pope-Davis et al., 2002).

Discussion
This study aimed to synthesize existing qualitative research on clients' negative experiences in psychotherapy.The breadth of the categories in our findings connect with many extant areas of psychotherapy research and cannot all be addressed within the limits of one paper.As a result, we discuss selected important conceptual tendencies that arose during the qualitative meta-analysis and their consistency with or extension of previous research.
Our inductive process led to the development of 21 meta-categories grouped into four higher-order clusters.Therapists' misbehavior (Cluster I) contained meta-categories capturing clients' negative evaluations of therapists' personal qualities and behaviours, either in terms of therapists' competence or in the manner in which they approached the client (not listening/understanding, being disrespectful, judging, and abusing the client).Although some of these categories could also be framed as relational (see Cluster II), clients often phrased them as therapists' qualities.Dating back to Rogers' (1957) seminal article, clients' perception of unconditional positive regard and empathic understanding from the therapist have been considered necessary conditions for successful treatment.Recent meta-analyses have confirmed therapists' empathy (Elliott et al., 2018), positive regard (Farber et al., 2018), and genuineness (Kolden et al., 2018) as consistent predictors of outcomes.The fact that clients explicitly voice the lack of these qualities as a source of dissatisfaction is consistent with this body of research.
Hindering aspects of the relationship (Cluster II) largely corresponded to aspects of research about the therapeutic alliance and alliance rupture (Eubanks et al., 2023), including the emotional bond and goal/task consensus.The working alliance has been confirmed as a robust predictor of outcomes in several meta-analyses (Flückiger et al., 2018;Horvath et al., 2011).Clients in the primary studies voiced their experiences of suboptimal therapeutic relationships in terms of distance, lack of support, insecurity, and confusion.
Poor treatment fit (Cluster III) contained experiences of the therapist or therapeutic process not meeting clients' expectations or needs during the therapy.This corresponds to the finding that accommodating clients' treatment preferences is connected with better outcomes and lower dropout rates (Swift et al., 2018).Indeed, treatment personalization at the level of treatment prescription (Beutler & Harwood, 2000), accommodating clients' preferences (Norcross & Cooper, 2021), or therapists' moment-by moment responsiveness (Watson & Wiseman, 2021;Wu & Levitt, 2020), represents the hallmark of state-of-the-art psychotherapy.Furthermore, high sensitivity of the ending of therapy emerged, stressing the need to pay more attention to the termination of treatment as a loss of a meaningful relationship (Roe et al., 2006a), which sometimes causes a long-lasting impact on well-being.
Negative impacts of treatment (Cluster IV) included references to dissatisfactory outcomes (i.e., lack of change and deterioration) as well as clients' more immediate reactions to the treatment.This finding supports the need for routine outcome monitoring so that therapists can know when clients are not improving as expected and take actions to address concerns (De Jong et al., 2021;Lambert et al., 2019).In addition, when comparing to a recent qualitative meta-analysis on client-reported hindering impacts (Ladmanová et al., 2022), our study identified several additional reactions that clients considered unwanted.These included loss of motivation or hope, feeling angry, feeling dependent on the therapy or the therapist, and self-accusation/selfblaming.Herzog et al. (2019) reviewed quantitative measures of the negative effects of psychotherapy to determine what kinds of clients' experiences were covered by these instruments.While there is a great overlap between our results and the domains of negative experiences identified by Herzog et al., our study identified some new categories, including the perception of the therapist as incompetent, dissatisfaction with therapy ending, loss of motivation, and selfaccusation/self-blaming.From this perspective, a revision of the existing quantitative instruments may be warranted.
In studies that explored both helpful and hindering experiences, negative experiences were mentioned less often than positive experiences (Castonguay et al., 2010;Knox et al., 2005;Krause et al., 2018;Manthei, 2007).While this may mean that positive experiences generally prevail in therapy, clients may also hesitate to report their negative experiences.As Rennie (1994) wrote, "Clients may feel that it is not their place to question their therapist's approach because they view themselves as naive laypersons and the therapist as an expert who probably knows what he or she is doing" (p.430).Another potential obstacle could be related to language and the ability to find words for subtle feelings of dissatisfaction and attribute them to a single cause (Hardy et al., 2019;Suárez-Delucchi et al., 2022).
Furthermore, there seems to be a continuum of therapy harmfulness (Bowie et al., 2016) that does not always enable clients to distinguish negative experiences from positive experiences.Nevertheless, the primary studies collected in our meta-analysis show that when client is allowed to become a "story-making author", clients describe a rich variety of negative experiences.
More attention should be given to "understanding how therapy participants construct these […] events through different kinds of story making performance" (Grafanaki & McLeod, 1999, p. 291).In principle, we deal with knowledge that relies on the client's meta-perceptions (Castonguay et al., 2010).It is notable that some words might have different meanings for clients and researchers.For instance, in the phrase "lack of collaboration", the term "collaboration" seemed to have different meanings in diverse studies.In Bowie et al.'s (2016) study it referred to the therapist being cooperative within the session, while in Lindgren and Rozental's (2021) study it indicated inflexibility in terms of scheduling appointments and other technical aspects.
Clients' negative perceptions did not automatically mean that the therapies or sessions led to harm and deterioration.As von Below (2020) concluded, "Even a very well-functioning therapy contains many instances of misunderstandings, difficult emotions the patient defends against, and disagreements" (p.9).Conversely, the deterioration of outcome measures does not automatically imply that the patient had a negative experience of therapy (Hardy et al., 2019).Psychotherapy may lead to ambivalent outcomes (Poulsen et al., 2010) and may be beneficial in spite of clients' negative experiences.Identifying patients who were dissatisfied but showed improvement led De Smet et al. (2021) to broader reflection on discrepant or contradictory signals coming from quantitative outcome evaluation and qualitative data.
When comparing our taxonomy and Curran et al.'s (2019) final model, many similarities can be identified.Several of our meta-categories correspond directly with Curran et al.'s eight domains, and many of our themes match their themes.For example, our Clusters II and IV correspond directly with their "relationship" and "therapy processes" domains, respectively.Similarly, our Cluster I corresponds with two of Curran et al.'s domains, namely, therapist factors and therapist behaviour.Three of Curran et al.'s domains, namely, contextual factors, pretherapy factors and endings, include many themes that match our Clusters II and IV.Despite fundamental differences in the method, there is a substantial overlap between our study and Curran et al.'s study.In addition, the two qualitative meta-analyses present findings in semantically different ways and describe patients' experiences using diverse labels.This comparison indicates how terms used for adverse impacts and subjective negative experiences vary in the field and indicates the need for precise and nuanced naming.

Limitations
We strived to include the broadest collection of primary studies possible.However, most of them were conducted in three countries (i.e., the UK, the US, and Sweden), leaving other countries, and even whole continents, missing or underrepresented.In terms of treatment approach, primary studies often did not specify the theoretical orientation of the therapists or included a mixture of orientations without allowing readers to link theoretical orientations to particular types of experience.Especially underrepresented were humanistic/experiential and family/systemic orientations.
We limited our search to primary studies on faceto-face individual psychotherapy.Including other settings (e.g., couple, family, and group therapy), modes of delivery (e.g., online and telephone therapy) and, possibly, other types of treatment (e.g., art and music therapy) might reveal additional types of negative experiences not covered on our study.We also programmatically excluded case studies.However, this only led to the exclusion of one study and, therefore, it did not bias our findings.
While most primary studies were broadly focused, some had a narrow focus on a specific topic (e.g., clients' deference), which could have introduced some bias, especially in terms of the meta-category frequencies.However, narrowly focused studies are also needed to understand specific areas of negative experiences in more depth.For instance, studies that focus on cultural, religious, and race/ethnicity issues are largely missing.Furthermore, an exploration of disorder-specific unhelpful factors was recommended by Omylinska-Thurston et al. (2019).
The primary studies varied in the data collection and analysis methods.While the differences in the method of data analysis may not be crucial (most methods of qualitative data analysis result in some form of the categorization of meanings that are comparable across studies), the data collection method determines the depth and richness of the data acquired in the primary study.Studies that relied on clients' written responses were threatened by relatively poorer data and a lack of context, which made data interpretation more difficult both in the context of the primary study itself and in the context of the meta-analysis.
The primary studies also differed markedly in the amount of time that elapsed between the treatment and the data collection.Arguably, clients may perceive events and their impacts differently based on the level of retrospectivity.However, since 35% of the primary studies did not report this information, we were unable to factor it into our analysis.Studies that allow for a better differentiation between temporary and lasting negative impacts are needed.
We found 21 meta-categories that captured many kinds of negative experiences clients reported after or during their treatment.However, the level of convergence of the primary studies was relatively lower compared to other qualitative meta-analyses (e.g., Polakovská & R ̌ihácěk, 2022).None of the meta-categories was present in more than 65% of the primary studies; the range was 17% to 65% with a median of 37% (compared to 55% to 85% with a median of 73% in Polakovská & R ̌ihácěk).Thus, despite the breadth of experiences described in the primary studies, the theoretical saturation seems to require more primary studies in this case, and there may be some types of experience not captured in our study.
It should also be noted that the scope of negative effects would probably be broader than the list described in our systematic review if we included the negative impacts of various archaic, risky and potentially harmful approaches, such as some forms of applied behaviour analysis (Sandoval-Norton & Shkedy, 2019) and other fringe and risky approaches (Lilienfeld, 2007).In this context, negative effects might be more severe compared to those we described.

Conclusions and Future Directions
We synthesized the existing qualitative evidence about clients' negative experiences in psychotherapy.Our study brought to the fore that despite some overlap among the primary studies, there was huge diversity in the reported findings, as a typical primary study captured less than half of the meta-categories we developed.This provides justification for conducting qualitative meta-analyses, especially in hard-to-reach areas of experience.Since negative experiences are less often reported compared to positive ones (Ladmanová et al., 2022), more primary studies may be needed to cover the breadth and varieties of these experiences.
Both qualitative and quantitative studies are still needed on this subject.Qualitative studies should focus on exploring negative effects in specific populations and under specific conditions (e.g., specific treatments, which may uncover other, yet unknown categories).This would than lead to a refinement of existing measures that, in turn, would allow researchers to address questions of quantitative nature, such as the prevalence of various types of negative experiences in different treatments and settings and the persistence of these experiences after treatment termination.A crucial question that remains to be answered is the relationship between client-reported negative experiences and the outcome of psychotherapy.Although Pourová et al. (2022) indicated a negative relationship between clients' overall improvements and various areas of negative effects, any causal explanations should be taken very carefully (Maxwell & Levitt, 2023).Further empirical studies are needed to understand the nature of this relationship.Altogether, these findings will hopefully lead to a better conceptual understanding of the nature of negative experiences and their role in the treatment.