Boundary violations in therapy: the patient's experience of harm

Summary

Harm in talking therapies, and in healthcare professionals’ relationships with patients generally, has received little attention in comparison with harm by medication and other treatments. There has been little research into causes, types and effects. Professionals behave as if it does not happen and tend to react defensively to complaints. We believe that it is essential for professionals to understand the potential for harm and evaluate their actions in order to make them safer. This article defines harm in the therapeutic context, discusses its prevalence and then focuses on adverse idealising transference: the adverse effects that may arise when a patient transfers idealising feelings onto the professional.

LEARNING OBJECTIVES

• • Develop a greater understanding of the problem of harm in psychotherapy

• • Be aware of adverse idealising transference and its possible harmful implications

• • Be aware of therapist actions that may encourage the development of an adverse idealising transference

DECLARATION OF INTEREST

None.

Reports of boundary violations – particularly violations of sexual boundaries – by people in positions of responsibility, including those in mental healthcare and other health professions, appear regularly in the media. Yet there is widespread ignorance and little acknowledgement of the problem among mental health professionals and healthcare regulators. The subject remains a taboo much as child sexual abuse used to be. There is inadequate training in the prevention of harm and the care and treatment of people who have experienced harm. Those who report concerns and seek help following abuse by a mental health professional frequently report a worsening in their symptoms as a result of a poor understanding of the matter and inadequate support. We contend that more action on prevention is needed, primarily through research, training and fostering a climate in which practitioners can be open about adverse events. There is also need for better support and treatment for victims who are brave enough to make their experience known.

This is the first of two articles in which we aim to encourage a dialogue on harm in therapy by sharing our experience of working, over many years, with patients and professionals caught up in the dynamics of harm. We define harm and discuss it prevalence, and explore the patient's general subjective experience of harm caused by boundary violations within the wider context of harmful practice. We devote much of this article to adverse idealising transferences (AITs) – the adverse effects that may arise when a patient transfers idealising feelings onto the professional – because, although we have found it to be a significant factor in most cases of harm, it is rarely discussed in the literature on harm. Our second article (Hook 2018) will focus specifically on sexual boundary violations – the assessment and management of victims and perpetrators and proposals for reducing risk.

Both articles derive principally from clinical work and research in psychotherapy, but most of what they contain is relevant to the practice of psychiatry and the caring professions more widely. Good practice in psychiatry is centred on forming a trusting relationship and an effective therapeutic alliance. Although concepts such as dependency and transference are embedded in the psychotherapeutic discourse, they are common to all professions with an inherent power imbalance, such as healthcare, social work, education and the police force.

Harm in psychological therapies

Definition

A central problem in the research to date is the lack of an accepted definition of harm. Parry et al (2016) suggest a definition comprising adverse events – significant episodes during or shortly after treatment, clinically significant deterioration following treatment, and lasting bad effects as described by the patient. The latter is of particular importance since our review of the literature suggests that the patient experience has often been undervalued and even dismissed as a relevant perspective on the course of therapy. We suggest that harm be defined as ‘any sustained negative consequence that the patient experiences as a result of engaging in a treatment’. We believe that ‘treatment’ should include any treatment or intervention that results in the subjective experience of harm, since such experiences result in deterioration and need to be better understood.

Causes

Differentiating categories of causes of harm is difficult because of overlap. In our experience, they fall into three principle categories: misconduct, poor skills and adverse patient reactions. Misconduct usually occurs when the professional fails to observe the boundaries of the professional relationship and exploits the patient sexually, financially or emotionally. Poor skills result from incompetence or negligence. Adverse reactions frequently occur because of an incompatibility between the patient and the treatment, with consequences ranging from anxiety to psychosis (Little 1958). These are comparable to adverse reactions that occur in drug therapy, except that information on adverse effects of drugs is freely available and routinely given, whereas information on the adverse effects of psychotherapy is not (Nutt 2008).

Side-effects of therapy

The side-effects of psychotherapy are not confined to AIT and include anxiety, depression, dependency, regression and depersonalisation. When there have been boundary violations it is common for patients to describe symptoms of post-traumatic stress disorder, suicidal ideation and suicide attempts; completed suicide also occurs (Resnik 2016). Other negative consequences range from ineffective use of time and money to relationship breakdown, as release of previously repressed affects and memories causes the patient to act out. Secondary harm may also be caused to the patient's family in such circumstances.

Any discussion of harm in psychotherapy needs to be seen in the context of an increasing evidence base for psychotherapy's effectiveness. There has been a tendency for mention of harm to be viewed as an attack on therapy. This reflects both on the uncertainties of the process, where every therapeutic relationship begins anew, and on an increasingly threatened profession. Parry et al (2016) comment, ‘patient safety has not been a priority for psychotherapy researchers’. One might add that this has been true for the profession as a whole. Clinical trials of psychotherapy are unlikely to describe adverse effects and drop-out rates may not be included. Scott & Young (2016) argue for a system of monitoring that goes beyond supervision: ‘Every branch of medicine learns from its mistakes; the same must surely be true for psychotherapy’. A prerequisite for learning from mistakes is creating a safe environment in which adverse events can be explored without fear or blame. Industries such as the airline industry have achieved spectacular results in this way (Syed 2016). Psychotherapy has barely begun such a process.

Prevalence of harm in psychological therapies

It is difficult to obtain prevalence data on harm from psychological therapies and there has been an unfortunate trend to equate lack of data with the assumption that harm is rare. There has also been a tendency to associate harm with inadequately qualified therapists, despite evidence that harm occurs disproportionately more often with more qualified, experienced professionals (Casemore 2001). This is certainly our experience. In 8 years of dealing with people who have been harmed by professionals, very few of the professionals were newly qualified or inexperienced; most were experienced, and some had served on ethics committees and/or had written about ethics.

Research studies show that a significant minority of psychotherapy patients experience harm. Estimates are reported as being between 3 and 10% (Mohr 1994; Lillenfeld 2007), with occasional studies showing higher rates. Buckley et al (1981) reported that over 20% of mental health professionals who had engaged in personal psychotherapy felt it had caused them some lasting harm. Crawford et al’s (2016) study of National Health Service (NHS) patients in England and Wales, with over 14 500 respondents, reported that around 5% experienced lasting bad effects. Younger adults and sexual and ethnic minorities reported significantly higher numbers of adverse events. Rates for specific modalities were 4% for cognitive–behavioural therapy and 9% for psychodynamic psychotherapy. It is puzzling that such a large study makes no mention of sexual boundary violations as a cause of harm.

Blaming the patient

During training, an impression that everything that emanates from the patient is pathological can be created. Although it is important for trainees to understand pathological processes, the idea of a continuum along which we all move avoids the impression of a split between the ‘deficit’ patient and the ‘functioning’ therapist. Normalising emotions that cause distress and acknowledging healthy aspects of the patient's mental functioning can reduce shame and support their judgement of themselves.

In psychotherapy, patients are usually seen as having been victims of neglect or abuse and deserving of help. This can rapidly change in the professional's mind when the patient complains. At a recent workshop on learning from patients’ complaints, organised by a national psychotherapy regulator, an ethics committee member asserted that patients who make complaints have borderline personality disorder. This appeared to be eagerly believed by other delegates, despite an absence of research confirming the statement. Similarly, a delegate, with an apparent grievance, asked that complaints be analysed within the therapy, implying that therapists should not have to defend their actions. Although analysing complaints in therapy is desirable, formal complaints usually arise when the therapist fails to hear the complaint and acknowledge any contributory behaviour. A common instance of this is when the therapist becomes overinvolved in the patient's life and encourages dependency. When the patient responds with frustration at the constraints of the therapeutic relationship, the therapist attributes the problem to borderline personality pathology in the patient, without acknowledging their own contribution. Work with personality disorder in particular requires rifts in the working alliance to be addressed as a crucial aspect of the success of therapy.

Another common way in which therapists side-step responsibility is by insisting that patients' complaints are re-enactments of childhood trauma rather than a here-and-now response to unsatisfactory therapist actions. We know of many instances where the therapist's failure to acknowledge a mistake has led to an escalation and complaint: a patient who expressed frustration because the therapist fell asleep during a session was deemed to be experiencing transference anger due to her mother's chronic inattention; a therapist who took a phone call during a session interpreted the patient's anger as envy because she did not want the therapist to have ‘any other children’. In such cases, the patient needed a simple acknowledgment of error before they could consider transference implications, but in each case the therapist refused, even when a direct request was made.

This is similar to the situation that exists in psychiatry concerning side-effects, and particularly withdrawal effects, of psychiatric medication. Reading patients' accounts on online forums makes it clear that they lose trust in their psychiatrists because they are not listened to or believed. They often feel left to cope with debilitating symptoms by themselves and are frequently diagnosed as suffering relapse of the original condition or are diagnosed with another condition requiring further medication.

The reluctance of the professions to engage with patients' perspectives is disappointing because patients have been publishing detailed accounts of harm for decades. Professionals' responses to such accounts are frequently dismissive, disrespectful and frankly abusive (Devereux 2010). Patients who make complaints about sexual boundary violations similarly find themselves disbelieved or diagnosed with new conditions such as borderline personality disorder or erotomania. Indeed, it is not uncommon for them to subsequently find that the events are described in their notes as ‘delusional’ and that they are referred to as ‘serial complainers’.

Adverse idealising transference (AIT)

Although most psychotherapists encounter the occasional patient with a previous adverse experience of psychotherapy, one of us (D.D.) has worked almost exclusively with this patient group over the past 8 years and has built up considerable expertise in this area. In the course of this work it has become apparent that extreme feelings of idealisation, by the patient for the professional, play a significant part in the majority of cases of harm. This is not only in psychotherapy, where the idealising transference is a recognised part of the therapeutic process, but in other professional relationships where the notion of transference may not be understood or recognised.

A temporary state of idealisation is common where dynamics of failed dependency through neglect and trauma are prominent. If this is understood and worked through it can be a necessary and productive aspect of the therapeutic process. If it is ignored, encouraged to persist or treated defensively this opportunity is likely to be lost. Either or both parties may mistake idealisation for the patient's love of the professional – these two states may overlap, but are not synonymous. This is potentially problematic as key aspects of the phenomena of idealisation may be left unnoticed and unanalysed. Specifically, the professional fails to address the fact that the patient is in a powerless state and is relatively unable to make use of their communications. One of us (J.H.) experienced an idealising transference in personal analysis, which was unacknowledged. Although this did not lead to serious boundary violation it persisted for several years after the therapy ended and required considerable further therapeutic work to elaborate and repair the effects. Many therapists have described, in personal communications, a similar situation in their training analysis, making it all the more surprising that the phenomenon is not more directly associated with therapeutic failure and harm.

In an attempt to encourage idealising transferences to be recognised as a potentially serious cause of harm, the term ‘adverse idealising transference’ (AIT) has been coined (Devereux 2016). It is defined as a chronic idealising transference reaction that adversely affects a person's mental capacity and psychological well-being, to the extent that they are unable to function in their usual way over a sustained period. Patients often feel deeply ashamed of such feelings and hide them from the professional, allowing them to flourish in silence. Most cases of AIT arise out of a predisposition in the patient and the technique of the professional and can be avoided through appropriate technique. This includes avoiding actions that breach professional boundaries, encourage dependency and result in the patient feeling special.

Here we bring the experience of working with people who have experienced AIT, together with the patient and professional literatures on AITs, with the aim of increasing knowledge of the phenomenon. To preserve the confidentiality of our former patients, all quotations are taken from the published literature.

Patients' accounts of AIT

Patients' accounts of ordinary idealising transferences are generally positive; when the feeling is not excessive most perceive the transference to be a motivating factor in the therapy. Descriptions of AIT are quite different. Patients describe intense confusion and loss of agency and compare the experience to being drugged or hypnotised. They also describe how it interferes with their mental capacity:

‘Feelings of extreme dependency are compounded by a regression to an infantile state with the overall result that the client becomes more or less detached from reality. The effect is similar in some ways to that produced by LSD’ (Alexander 2003: p. 295).

‘A magic trick had been performed on me: in just a few hours of sitting alone in a room with Paul, a large part of my mind had effectively been taken over, leaving me with little left to expend on my work, social life and other parts of normal life’ (Simpson 2006: p. 91).

While people are in the midst of AIT it is clear that the locus of distortion is their conviction that psychological transcendence will occur as a consequence of their relationship with the professional. Importantly, the idea of transcendence is not consequent on the therapeutic process, but rather on the notion of an identity merger with the professional, which may be entirely unconscious. The import here is that that it helps to explain the tenacity of the attachment and how it predisposes to exploitation. The exploitation that arises falls broadly into three categories: sexual, psychological and financial. The following patient's quotations give an example of each:

‘He'd been my GP for 5 years and my feelings for him were immense. The day I disclosed childhood sexual abuse he put his hand on my knee and looked at me intently […]. My desire to be connected to him was so intense that the offer of sordid and selfish sex was irresistible […]. When I reported it to the police they described it as an affair; it was not, I was incredibly vulnerable’ (Rooks 2002: p. 2).

‘His interests and worldview became a source of huge fascination and I devoted myself to them, reading everything I could in order to be of interest to him. He encouraged this, never questioning my motivation’ (Pearson 2002: p. 4).

‘I saved enough for eight sessions but became so addicted to her that her suggestion that I use my house deposit to pay for therapy seemed entirely reasonable’ (Nash 2002: p. 6).

Although there is an assumption in the literature that such transferences resolve, our experience is that they can persist and, in extreme cases, last for decades.

Professional accounts of AIT

Kohut (1968) first used the term ‘idealising transference’ to describe a type of transference in which the therapist's character is distorted and imbued with idealised attributes that reflect the patient's unmet developmental longings. He postulates that this arises if a mother is unable to attune to the needs of her baby and the baby is unable to internalise a sufficiently idealised mental image of the mother. Such transferences are a pervasive part of the therapeutic process and form a continuum ranging from mild admiration to pathological obsession with the therapist. It is widely recognised that transferences at the milder end of the continuum are useful both in helping the patient to engage with therapy and in providing insight into the patient's developmental history. Although Kohut recognised that erotic elements are often present, his particular contribution was to emphasise the central importance of the idealising aspect of the transference.

The professional literature recognises that idealising transference reactions can be difficult to manage, but usually describes them from the perspective of the therapist and not the patient. Freud (1915) draws an analogy between an analyst handling the transference and a chemist handling highly explosive materials. Such an analogy communicates the intensity of the transference and the difficulties in managing the patient, but it does not make clear the intractable harm described by patients. The psychoanalyst Margaret Little (1958), who experienced such a transference herself, articulates this, describing such transferences as ‘terrifying’ and ‘annihilating’, although she too seems to assume that the experience will resolve satisfactorily. It is difficult to find anything in the professional literature that acknowledges that idealising transferences do not always resolve.

There is a consensus in the literature that psychotic (Little 1958) transferences are particularly difficult to treat. In a psychotic transference a person who has never experienced psychosis will experience delusions about the therapist. Hedges (1994) and Frayn (1990) contend that they stem from infancy, resulting in difficulties with verbalisation and a subsequent tendency to act out. This concurs with our experience. We would also agree with the suggestion that non-facilitating, intractable transferences, which are not primarily induced by poor technique, are frequently sadomasochistic re-enactments and pathological attempts at regulation of self-esteem (Frayn 1986). Hedges (1994) emphasises that primitive processes are in play and warns therapists that work with such patients may lead to false allegations of malpractice.

Personality structure of people who develop AIT

The literature associates intense idealising transferences with narcissistic personality organisation (Kohut 1971; Frayn 1990). Kernberg (1995) associates intense manifestations of the phenomenon with borderline personality organisation. Frayn (1990) suggests that idealisation is used to maintain narcissistic fusion against feelings of emptiness and powerlessness and may result in a need to seek approval from parental figures and a deep need for attunement. Klein (1957) believed that people who idealise are predisposed to feel envy and have difficulties with separateness and separation.

Our experience is that there is an association between AIT and behaviours related to borderline personality structures at the most severe end of the spectrum, particularly in terms of patients' need to control the therapist and seek concrete expressions of care. It is not clear that this is causative and at this stage can only be regarded as an association, since we also have experience of working with patients who develop AIT and do not have a borderline personality structure. Patients' histories more frequently confirm Kohut's (1979) contention that the presence, even fleetingly, of adults who provide restorative experiences can moderate the damage to the child resulting in transferences in which only part of the personality is exposed to AIT, leaving another part to function reasonably well. Many of our patients have been able to articulate in an honest and impressive way how they are drawn to the experience of idealisation, giving a meta-commentary on their thinking while working collaboratively to overcome it.

The phenomenon affects people from all backgrounds. Intellectual and social success is no bar. We have found that it occurs most commonly in female-patient–male-professional dyads, although it is also common in all-female dyads and in all-male dyads where the patient is homosexual. It is generally less common in men.

Therapeutic technique and the idealising transference

Background

Freud (1915) believed that idealising transferences could act as a motor to the therapy, but he saw them as a resistance to treatment and an attempt to seek cure thorough a new relationship. He was clear that such transferences were to be analysed and not reciprocated. In our experience, reciprocation encourages the development of AIT, particularly disclosing emotional feelings about the patient and disclosures that make the patient feel special. A psychiatrist writing about her own experience of AIT illustrates this:

‘He sometimes told me vignettes from his life. I felt special, as if I knew things about him that others did not […]. I would never talk about him to anyone outside analysis, never reveal the things he told me. That made them all the more precious; furthermore, it made me feel special and secretly loyal to him. He and I had a little secret life’ (Gabbard 1995: p. 132).

Although Kleinians cautioned against reciprocation, their particular contribution was to suggest the need to interpret the aggressive aspects of the transference. In doing so they emphasised that idealisation frequently involves complex negative feelings, particularly in relation to envy of the therapist. Making these feelings explicit through interpretations clearly depends on the patient's ability to tolerate such interpretations. We would also stress that, although some patients develop destructive, envious feelings towards the therapist, the majority do not.

Ideas about technique changed with Kohut's (1971) belief that the idealising transference should be facilitated in order to encourage an empathic atmosphere. Kohut did, however, also recognise the need for restraint because he states that in the early stages of therapy there is a need for a non-intrusive, non-seductive atmosphere. He describes how easy it is to unwittingly use language that is overstimulating with potentially catastrophic consequences. In our view, restraint should continue beyond the initial stages of therapy.

The idea of encouraging the erotic aspects of the transference gained ground in the decades following Kohut, when some therapists recommended erotic bonding. This was envisaged as erotic feelings ‘forged at a deep emotional level’ which bound the analytic couple together in fantasy (Mann 1999). In a similar spirit, Samuels (1999: pp. 153–4) argued against ‘safe’ analysis, stating the impossibility of exploring sexual emotions ‘without there being something literal, actual, concrete, corporeal, real, experiential in either or both of the participants’. More recently, Haule (2015) has compared the relationship between patient and therapist to a deep, erotic, mystical union with God. Although he acknowledges that this may make other important relationships appear mundane, he does not consider the disastrous effect it could have on the patient's personal life. Many patients describe irreparable damage to personal relationships because they compare the intimacy of a non-mutual therapy relationship to that of a real relationship and find their partners wanting.

We have helped many people who have experienced AIT in relationships with non-psychotherapist professionals, particularly general practitioners and psychiatrists. Because such relationships tend to focus primarily on issues other than the patient's inner feelings, the patient will typically find it highly embarrassing and inappropriate to reveal their feelings. Professionals often pick up on the patient's feelings, and if they are in a vulnerable position themselves (with difficulties at work, relationship problems), may slip into behaviours that exacerbate the problem. Such behaviours include making appointments more often than necessary, booking the patient at the end of the clinic to allow for a longer appointment, giving personal information, especially information relating to work or relationship difficulties, becoming overinvolved in the patient's life and giving the patient their private mobile phone number in order to bypass the usual system for appointments.

Appropriate technique

In our experience, appropriate technique is crucial to preventing and limiting AIT, beginning with consistent boundaries and a collaborative relationship that facilitates open discussion. A seductive, soulmate atmosphere is common in cases of AIT, but so is the converse: professionals who appear annoyed, embarrassed or defensive about the situation. This often arises when the professional has been seductive and becomes fearful following the patient's response. Although it may be necessary for the professional to state explicitly that there can never be a personal relationship with the patient, this should be done in a way that avoids rejection and emphasises the professional's commitment to working with the patient and exploring the transference. Professionals who respond to AIT by abruptly ending the therapeutic relationship (sometimes by email) will almost certainly exacerbate the problem and leave the patient with a harmful, difficult-to-resolve transference.

Deficiencies in technique usually arise from vulnerabilities in the professional and inadequate training. Personal vulnerabilities induce them (often unconsciously) to use the patient to meet their own psychological needs. These vulnerabilities may not come to light during training or supervision or a blind eye may be turned, perhaps on the grounds that in psychodynamic therapies at least they will be addressed in personal therapy (Freud 1937). We believe this to be a questionable assumption. More recent research suggests that training analyses may increase narcissism in the therapist (Welt 1990). This is significant, because professionals who operate from a narcissistic position have a propensity to use their patients for ego support.

Proper training of health professionals could help make the pitfalls of idealisation explicit. This is necessary because the individual actions that encourage idealisation may not be perceived as boundary breaches. Indeed, the professional may believe they are going ‘above and beyond’ in caring for the patient. Professionals should also be trained to carry out regular reviews in which they consider whether the treatment is addressing the patient's needs.

Informing patients of the risk of AIT and other side-effects

An accepted principle of medical ethics is that patients have a right to information on risk in order to make informed choices on treatment (Beauchamp 2013). Informed consent and discussion of side-effects are, however, uncommon in both psychotherapy and psychiatry, other than physical and pharmacological treatments. Patients who have experienced AIT frequently compare its incapacitating effects to the side-effects of a drug, observing that if a clinician had prescribed a drug with the same adverse potential it would be unethical not to inform the patient of the risks. We support this view, as do Nutt and Sharp, who also draw an analogy to drug therapy, stating that the side-effects of psychotherapy are in fact potentially greater and must be discussed (Nutt 2008: p. 5).

Discussions with psychotherapists and psychiatrists about informed consent suggest that the reluctance to discuss side-effects of psychotherapy stems primarily from the belief that patients will be alarmed by such a discussion. Professionals worry that discussion of the idealising transference will seem far-fetched or will interfere with psychoanalytic work ‘in the transference’. Clearly, the discussion needs to be tailored to the patient. In relation to speaking about the idealising transference, it is helpful to begin with something like: ‘It's important that you know that you may experience intense, unexpected emotions as a result of psychotherapy and that this is completely normal’. As well as giving information, the discussion is an opportunity to encourage patients to be open about any symptoms or emotions as they arise. This kind of conversation also helps to engage the patient in a collaborative relationship with the professional. Our experience of providing information has only ever been positive. On many occasions, patients have referred back to initial discussions when bringing up side-effects: ‘You know you said I might experience…’. These benefits are supported by the study we mentioned earlier, of over 14 500 cases of psychotherapy, which showed that informed consent improves outcome (Crawford 2016).

Conclusions

It is clear from patients' descriptions that insufficient attention is paid to harm in psychotherapy. This has resulted in lost opportunities to reduce harm by educating professionals and informing patients about risk. In this article we have focused on harm in general and AIT in particular and have shown how AIT usually arises from a combination of patient susceptibility and vulnerabilities in the professional. Although it is most closely associated with psychotherapy, AIT also arises in psychiatry and other professional relationships. AIT is potentially difficult to work with and requires active engagement on the part of the professional in order to guard against serious deleterious effects. This entails keeping appropriate boundaries and not encouraging dependency. If the professional suspects that an idealising transference is adversely affecting a patient, the matter should be addressed in an open and collaborative way. Professionals who end therapeutic relationships abruptly risk causing great harm.

Although the professional literature articulates the difficulty of the idealising transference, it does not sufficiently acknowledge the harm. It also fails to consider the effect of the phenomenon on a patient's mental capacity and how it may make them vulnerable to emotional, financial and sexual exploitation. Patients who have experienced AIT are clear that it should be seen as a potentially serious side-effect of psychotherapy and that there should be open discussion about this and other possible side-effects before patients embark on treatment.