This website uses cookies. By continuing to browse you are agreeing to our use of cookies, which you can read more about here.


Welcome to, an award-winning website published by the British Association for Counselling & Psychotherapy.

The site contains a searchable archive of over 500 articles published in Therapy Today since September 2005. Some articles are freely available, whereas others can be bought via the online purchasing system (log in to access more articles). To subscribe, click here.

Volume 18
Issue 2
March 2007


The evidence is that either we pay little attention to the risks posed by clients, or we find it impossible to put suitable measures in place. What is going on?

  • Personal safety: do counsellors care?

  • by

  • Sally Despenser
  • Personal safety is an important issue but one that appears to be neglected in counselling literature and in some trainings (and in conversations between therapists). The potential for serious risk posed by clients should not be underestimated. In 1996 a man with previous convictions for violence and rape turned up at the house of a woman who (a few years earlier) had been his support counsellor in prison, and then raped and strangled her. She had been a prison teacher and probation volunteer, and had counselling skills training. The trial report1 does not say how her murderer knew her private address, but it suggests that she had maintained contact with him without authorisation or supervision after his release from prison.

    The report underlines factors which therapists ignore at their peril: assessment of risk, competence, supervision, and the therapist’s omnipotence. The writer of her obituary suggests that ‘clients for whom we have the greatest understanding can also carry for us the potential seeds of our own difficulties’2.

    Risk assessment usually focuses only on risk to the client, but equal weight should be accorded to the safety of the therapist and the client. In this article I report the findings of a small survey which asked how some therapists handle the issue of their own safety in a variety of settings, and give an example of safety issues in a college setting, which were revealed in supervision. I will then discuss factors to consider in evaluating risk to the counsellor, and explore some of the dynamics (both conscious and unconscious) underlying the neglect of therapist safety and the personal vulnerability hinted at in the obituary quoted above.

    A survey: risk and safety in various settings
    During 2004, I conducted a small-scale postal survey amongst 14 of my colleagues. All those consulted worked both privately and in institutions. I asked: ‘In your practice as a therapist, what factors do you take into account when thinking about your own safety?’ Specific questions focused on the settings in which the therapists worked, their sense of ‘being alone’ or ‘in company’ in these various settings, the referral route for clients, their assessment system, and what other factors therapists took into account when asked to consider their own safety. The settings

    ■ working at home (‘alone’)
    ■ privately (‘alone’) in rooms in a multi-purpose building
    ■ in shared accommodation in a multi-purpose building, including some NHS settings and alternative health centres (‘in company’)
    ■ public buildings including education, NHS and staff support work, in a peripatetic fashion (the sense reported here was ‘alone’).

    Some respondents volunteered information about the danger of exits and entrances. In one case, fixed security bars on ground floor windows would block an emergency exit for the counsellor. In another case, where doors cannot be locked because of fire regulations, there is a risk of intruders. The risk of casual access from the street to a multi-purpose building was also reported.

    Several counsellors commented on their vulnerability working alone in any setting.

    ‘Being alone in the building is the main concern, sometimes up to mid evening.’ (Private practice – multi-purpose building)

    ‘There would at present be no recourse if threatened. Neighbours would take no notice and there is only one way out. There is no way of connecting an alarm to anyone who would respond.’ (Own consulting room)

    ‘If I see a client in the office in the evening, the only other people are out of hearing distance and the caretaker may be anywhere in the large building. I think I would be safer at home but the agency requires this.’ (Employee Assistance Programme (EAP) sessions – multipurpose building)

    Even if the therapist is technically ‘in company’, the presence of other people in shared premises is no guarantee of help:

    ‘Large rambling building with rooms tucked away at the end of long windy corridors. Calls for help would not be heard by anyone else in the building.’

    The next response also illustrates the particular risk attached to first sessions:

    ‘At the college I am in a room alone. Only one person knows I am there, and would not realise if anything happened as she only sees me at the beginning and end of the day. So if anything happened it could be a while [before anyone realised something was wrong]. However, if a client I had already seen and was worried about came back, I could ask my support person to ring me.’

    Some therapists arrange for a chaperone to be on hand when they meet a client for the first time in a setting where they feel ‘alone’, a strategy that is widely recommended. However:

    ‘In theory it [chaperone arrangement] is in place, and there are others near in the daytime, but if a client had a knife, what good would that be?’

    One response acknowledged the flawed logic in arranging backup when meeting a male client but not a female client.

    In some settings there are phones or alarms in the counselling room (or workers have personal alarms). Bond3 argues that although alarms may not prevent assault, ‘they increase the possibility of getting assistance’. However, some people had little faith in the value of alarms:

    ‘The availability of an alarm system has never helped in past incidents where I have felt threatened.’

    Some sensible precautions to take in the counselling room were mentioned, including sitting with clear access to the door. Nobody mentioned equipment or furnishings as a source of danger.

    Referral routes fell into three groups:
    ■ where the client comes on the recommendation of another professional person (including EAP agency, GP, CPN, teacher, another therapist, supervisor) with  some prior screening (however minimal or informal);
    ■ where the client comes on the recommendation of a non-professional person known to the therapist or the therapist’s network (‘friend of a friend’, former client);
    ■ self-referral with no screening, and no connection to therapist’s own network (eg drop-in sessions, Yellow Pages, BACP online directory). However, where there was initial telephone contact, it was used as a form of screening, giving an opportunity for the therapist to reflect on the countertransference. It was noted that, in a drop-in service, extra care was needed, as there was no prior telephone contact. The value of a phone conversation used as ‘preliminary assessment’ is discussed by Bond.3 No participants reported using email for initial contact (nor was online counselling mentioned in any responses).

    I asked: ‘When assessing a client, what factors do you take into account with your safety specifically in mind?’ It was evident in the responses that relevant information was being gathered on a continuous basis, from before the first meeting. The factors mentioned fell into three categories:

    1. Signposts in client’s history: reported episodes of violence or aggression (ideas and actions), psychiatric and forensic history, self-harm, abuse, use of medication. Some of this information might be contained in a referral letter, otherwise it is gathered in first session/s.

    2. Signposts in client’s initial, current and ongoing state: including evidence at any time of erratic state or intoxication, body language, verbal clues, recent or anticipated changes in client’s situation or support network, ego strength.

    An example was given of a verbal clue that led to termination of the contract with a male client: ‘Aren’t you afraid, working down the end of this corridor at night?’

    Examples of body language included overt sexual posturing. Two participants, commenting on cases where they initially felt intimidated by a client’s physical size and dress, remarked that they would attempt not to judge the situation solely on appearances.

    3. Clues in countertransference feelings throughout the work, from first hearing the client’s voice. Terminology included: ‘gut reaction’, ‘all antennae active’, ‘instinct’. (These feelings, and the use made of them, were reported by most participants, regardless of their theoretical orientation.)

    A striking illustration of a therapist analysing her feelings appeared in one response: it shows the use of projective identification: ‘I still feel gut reaction is the most accurate indicator there is. If I feel fear, I note it, examine it and explore the reality of it. Verbal expressions of violence perceived as threats to myself could often be interpreted as powerful projections, inducing fear in others rather than experiencing the fear of their own inner rage.’

    Several therapists had declined to offer treatment after risks were highlighted at assessment or later on. Some of these clients were referred back to their GP. One counsellor in the voluntary sector had terminated a session prematurely to prevent escalation of threat from a female client. Another counsellor arranged for a security person to remain near the counselling room during sessions with a male client with a violent history: the contract was terminated after discussions in supervision, and ‘the client accepted this decision fairly well’. Some aspects of declining to work with a client are discussed later.

    During the writing of this paper, I invited the participants to respond to and comment on the foregoing summary of my survey. Two commented that they were becoming more observant of risks to themselves. One counsellor, wanting to draw attention to the dangers in her particular setting, had used the summary when debriefing with her NHS manager after a frightening incident. (This offered an unusual example of proper acknowledgment of risk by both the counsellor and the organisation, and contrasts with the example which I give below). A fourth reported that, in her GP setting, when she tried to discuss the issue of safety with counsellor colleagues, an astonished silence was followed by a chorus of dismissal, indicating that they did not want to think about it. Later, I will speculate about the reasons for this dismissal.

    Risk in an institutional setting
    The events I now describe took place some years ago in a college struggling to survive in a climate of cutbacks. The counselling service provided open access and self-referral for all members of the college, both by appointment and via drop-in sessions. Nobody was turned away. The service was (as in many other colleges at the time) threatened with extinction. When I became their clinical supervisor, the counsellors told me that there had been occasions when they had been physically threatened by clients in the counselling room. They told me that although these incidents had been smoothed over by management, a panic alarm had been installed. However, it also transpired that there was no system (and there never had been) for responding to an alarm, although the counsellors had been trying for a long time to get one set up. When we actually tested the alarm one day, there was no response for 20 minutes. When two college managers arrived, they protested furiously that the alarm should not have been activated without prior arrangement. (This incident confirmed that an alarm system is only as effective as the detailed and dependable arrangements for responding to it.) In this setting, the promise of external help was illusory. When I left three years later, the situation was still not resolved, and I will look at some of its underlying dynamics in a moment.

    But first, I want to describe what seems to be an increasing level of risk, and suggest some reasons for this increase. I will explore how the counsellor’s feelings can be used to inform the assessment and any decision to exclude, before going on to examine the unconscious factors which, at both personal and institutional level, may make it difficult for counsellors to heed the  warning signs contained in these feelings.

    A growing risk
    The issue of physical safety for therapists was addressed in a BACP information sheet2 in 2003, for practitioners working in high-risk environments with high-risk clients. It emphasises the responsibility of both employers and practitioners to attend to the assessment of risk. Evaluation of risk and safety becomes even more pressing in the light of changes to mental health care in the NHS, and evidence that counsellors are now working with a client group that includes cases that are much more complex than they used to be.

    In the UK, the setting for therapy is changing. In the public sector, many of the changes are driven by cost. Clients who previously would have been treated in specialist settings are now seen in less costly, more dispersed counselling services. As a result there is a greater potential for borderline or psychotic patients to present (a) with no pre-screening, and (b) in settings that may be able to offer less containment.

    The Ethical Framework requires us to take responsibility for working within our competence. The diversification of training makes it much harder for outsiders to judge counsellor competence, with the result that difficult cases may be presented to therapists with insufficient skills, in settings where either assessment does not take place, or it fails to address the potential for acting out. In addition, as I discuss later, the unconscious fit between individuals and organisations may lead counsellors into working beyond their competence.

    What are the risks?
    There are four sources of risk:

    1. the isolation of the setting
    2. being alone with a stranger
    3. working in an organisation that (for various reasons) ignores or denies the existence of risk
    4. when the therapist ignores warning signs: this is discussed later.

    Several examples of risky situations were given in my survey, some of them similar to the circumstances in which the estate agent, Suzy Lamplugh, disappeared in 1986. Her memorial trust cautions against putting ourselves physically in such a situation4. As counsellors, we meet a stranger alone and with not much information about them (although as my survey showed, prior contact by telephone can provide an essential screening tool). For many counsellors, wherever we work, it is a risky business. But we can reduce the risk to ourselves by attending to warning signs in the client’s history and presentation, and, as I shall argue below, by attending to our own countertransference feelings. I will also argue that it is when we allow our own unconscious agenda (or that of our workplace) to overrule the information provided by these feelings that we seriously increase the risk to ourselves.

    Assessing risk: using countertransference feelings
    Thinking about risk does not take place only during an initial assessment; it is an ongoing process. The therapist’s fear or awareness of threat may grow quickly or slowly, and may occur at any time during the therapeutic contract. The table (Fig. 1) identifies signposts contained in the client’s history, in their initial, current and ongoing state, and in countertransference feelings throughout the work. In using these feelings (‘gut reaction’, ‘all antennae active’, ‘instinct’) to evaluate the threat, there are four interrelated questions:

    1. Are the therapist’s feelings of fear based in reality or in a projection by the client (in other words, a projective identification)?
    2. If projective identification: is it judged that the client will be able to take back the projections, or is their mental functioning borderline?
    3. If there is additional evidence in the client’s history and current state to confirm that the therapist’s fear is reality-based and that violence is a real possibility, what steps need to be taken to manage the risk or terminate the contract?
    4. Is the therapist properly trained, supervised and contained for this piece of work in this setting?

    In medical settings, when danger (to patient, therapist, others) becomes apparent, treatment may continue after additional strategies to support and contain are put in place (the container as safety net). Not all settings or therapists are in a position to offer this degree of containment, and organisations and individual counsellors therefore need to develop their own rationale and guidelines for declining certain referrals.

    For therapists in some settings, some difficult behaviours (eg. intoxication, refusal to leave the room, refusal to pay) might be manageable. However, by definition, the practice of assessment admits the possibility of exclusion from treatment. What reasons might be given for refusing therapy? First, realistic danger to the therapist. ‘No treatment is offered to patients who might present a danger to the therapist.’5 Second, even if the therapist’s feelings are derived more from projection than reality: ‘Whatever the cause of the sense of the danger, there is no ethical requirement that the counsellor should continue to see clients who evoke such feelings.’3

    Third, if the setting cannot provide adequate containment, a good referral is desirable (though in many areas of the UK this can be difficult to achieve). Fourth, if the counsellor does not have the necessary skills and experience (and in the absence of objective measures of competence, this judgement may also depend on the supervisor), again a good referral is recommended. However, as I shall discuss below, the grandiosity of the therapist might obstruct this option.

    If a decision is made to exclude such clients on the grounds that they pose a risk to the therapist (or to a third party), is there a duty to warn other therapists to whom a referral might be made? Do agencies and individuals have a clear policy that balances risk, confidentiality and respect for the therapeutic frame? While therapists are trained and supported to contain this tension, and to recognise the importance of tolerating ambivalence and uncertainty implicit in ethical reasoning, this approach is at odds with a public (‘paranoid-schizoid’) demand for certainty, with the splitting and projection that this involves.

    Some unconscious dynamics
    Personal needs
    In all the settings covered in my survey, the identified risks faced by counsellors are striking: physical isolation, dangerous premises, some clients being seen without pre-screening, neglect of safety by counsellors and managers.

    Why do counsellors accept this level of risk? What unconscious need is being met? Why is the increasing level of risk denied?

    The unconscious motives for choosing to be a therapist are well rehearsed. It is posited that the counsellor has an unconscious need to ‘come to terms with unresolved issues from our past6, or to re-enact old roles, drawing on skills once developed to meet a parent’s narcissistic needs. The roles include playing saviour, trying to love people better, taking excessive responsibility, denial of own needs, and working beyond competence (‘the heroic saviour’). The skills include special sensitivity to the unconscious signals and needs of others, accompanied by an unconscious predisposition to receive projections familiar from early life. The individual is drawn to work and settings that replicate this unconscious pattern. The traps include refusal to give up, an omnipotent expectation of one’s capacity to care which may also blind us to the client’s rage7, and a fantasy in which the damaged client becomes confused with the therapist’s own damaged parent. In the context of therapist safety, these fantasies pose serious dangers.

    There is also a danger in the unconscious fit between the narcissism of both parties to the therapy. The patient ‘craves the omnipotent parent of early infancy’8 and in the analyst ‘sees a person endowed with special power, special intelligence and wisdom … partaking in the omnipotence which the child attributes to the parents’8. A transference infused with flattery and idealisation fuels the therapist’s grandiosity and self deception, and reinforces denial of risk. A dangerous self-delusion results when the therapist colludes with the client’s demand for a saviour.

    There is another set of needs: to be admired or to be seen as infallible, to be right and omniscient, to have certainty. These serve to protect us against feelings of worthlessness and the fear of failure.

    Denial as part of grandiosity may explain why ‘personal security for therapists is often not taken seriously enough’9. It might explain the silence and dismissal reported by one of my colleagues when she tried to engage in a discussion about counsellor safety with her colleagues, and the absence from the counselling literature of discussion about safety. Where our personal safety might be threatened, this grandiosity may be the factor that puts our lives at risk. Grandiosity may also make it difficult to bring the therapy to an end at any stage. We deny that we are dispensable: ‘We behave as if immortal, indefinitely staving off extinction.’8

    In summary, the unacknowledged contents of our shadow may blind us to the reality of our powerlessness, make it difficult for us to become conscious of threats and realistic limitations, obscure the reality that not everyone can be helped, that we also may fail where others have failed before. In other words, the dangers of flattery, idealisation and the selfdeception of grandiosity are very real.

    Institutional anxieties and needs
    Earlier on, I gave an example of risk to counsellors in a college. Institutional refusal to attend to the safety of the counselling team over a long period, and failure to respond helpfully to their protest, appeared to be linked to institutional anxiety. There were several anxieties: about extinction, the presence of disturbing material, and some other painful realities. A container-as-dustbin was needed to receive different sets of projections: disowned fears of extinction, the splitting off of frightening material, and the denial of some other painful realities. This was the unconscious role assigned to the counsellors.

    1. Extinction
    It seems that institutional anxieties about cuts and redundancies were, at another level, about death and extinction. Such unbearable feelings were split off, projected into the counselling team and dumped as in a dustbin When the counsellors expressed their own vulnerability on the occasion cited, their protests threatened to make conscious the disowned institutional fears about extinction. The college replied with ‘an attack on disowned, projected vulnerability, an unconscious defence against the worst anxiety of all, oblivion and extinction of the self 10.

    2. Frightening material
    The vocal protest of the counsellors (and the unannounced alarm) also reminded the institution of an unwelcome presence in its midst of disturbing and frightening (client) material which could normally be denied and split off, since the work was usually kept out of sight and hearing, behind several sets of anonymous doors. The protest threatened to bring this frightening material to consciousness. This is linked to the third point.

    3. Other painful realities
    College cuts would have implications for students, particularly for the substantial number with learning difficulties. Before the cuts were threatened, a defensive idealisation of these students protected staff (and families) against despair and sadness, and kept some painful realities out of sight. When other disturbing material, normally hidden, was forced into view by the counsellors’ vocal protests, it was as if the larger group was in danger of being overwhelmed by the disowned contents of the container-dustbin.

    In this article I have argued that the safety of both the therapist and the client is equally important. I have summarised the factors to be considered during assessment: signposts in the client’s past history, their current situation, and the therapist’s countertransference. I discussed the consequences when, on the basis of assessment, it is decided that the potential for risk to the therapist makes it unsafe to offer therapy. I explored how the unconscious fit between the client and the counsellor may affect safety, and suggested that the most dangerous element in the counselling room could be the therapist’s own narcissism.

    I also explored the unconscious fit between the needs of the organisational setting and the counsellor, showing how a vulnerability to certain institutional projections may lead the counsellor to be drawn into a collusive grandiosity, and to neglect their own safety.

  • Sally Despenser is accredited as counsellor and supervisor, and works in Oxford.

    This is a shortened and revised version of an article previously published in Psychodynamic Practice November 2005; 11(4): 429-446.

  • References:

    1. The Guardian, 19 February 1997.
    2. Jackson H, revised Chaytor D. Personal safety for practitioners working in high-risk environments and with high risk clients. Information sheet G5. Rugby: BACP; 2003.
    3. Bond T. Standards and ethics for counselling in action.London: Sage; 1993.
    5. Burton M. Psychotherapy, counselling and primary mental health care. Chichester: Wiley; 1998.
    6. Roberts VZ. The self-assigned impossible task. In: Obholzer A, Roberts VZ. (eds) The unconscious at work. London: Routledge; 1994.
    7. Casement P. Learning from our mistakes. Hove: Brunner Routledge; 2002.
    8. Malcolm J. Psychoanalysis: the impossible profession. London: Granta Books; 2004.
    9. Jenkins P. Counselling, psychotherapy and the law. London: Sage; 1997.
    10. Nimmo S. Counselling in a hospice. Psychodynamic Counselling. 1997; 3(2):137.

<< Previous article Back to top ^ Next article >>