A 6-year-old boy kisses a girl in his first-grade class on the cheek. He is summarily suspended from school for sexual harassment. The media have a field day discussing the excesses of the sexual harassment construct. The school is embarrassed. The boy is reinstated. A fundamentally sound principle designed to prevent the encroachment of others on one’s personal space and to protect the individual from obnoxious and predatory behavior has been stretched to an absurd extreme.
This recent news story reminds us that when pendulums begin to swing, they commonly swing too far. A similar trend can be observed in the area of professional boundary violations within the practice of clinical psychiatry. In the last decade the mental health professions have experienced a progressive raising of consciousness about the issue of professional boundaries. A series of contributions from our own work (1—6) and that of others (7—9) has demonstrated how the violation of professional boundaries by psychiatrists and other mental health professionals may cause harm to patients while also undermining effective treatment. Amid this laudable interdisciplinary heightening of awareness, we have also observed a perhaps predictable overreaction to boundary concerns and an associated misapplication of the underlying principles. While such excesses are often proffered as indicia of patient protection, the perversion of boundary theory may place professionals at risk for undeserved sanctions and may potentially harm patients themselves by frightening the professionals into rigidity in therapeutic interactions. This extreme position is captured by a cartoon that shows a male patient putting forth his hand for a handshake with his female therapist: the latter looks disdainfully at the outstretched hand and sneers, "Nice try!"
Much of the current concern about professional boundaries has grown out of a wish to prevent sexual misconduct by clinicians by identifying milder nonsexual boundary violations as possible way stations on a "slippery slope" toward frank sexual contact. Clinical observations have taught us that many boundary transgressions that do not involve erotic touch nevertheless may be destructive to the treatment process and to the patient. Meanwhile, forensic practitioners have observed a shift in misconduct litigation from cases of overt sexual intercourse to cases of "pure" boundary violations: "pure" in the sense that the only violations that were alleged fell short of sexual intercourse (3). Plaintiffs’ attorneys have also noted that many malpractice insurance carriers will not provide coverage for sexual misconduct but may cover various types of transference or countertransference mismanagement. In addition to the civil litigation on these issues that alleges traditional malpractice, licensing and registration boards in medicine and the behavioral sciences, as well as ethics committees of professional organizations, have also become involved in patient complaints about alleged nonsexual boundary violations.
In this communication we will describe some of the significant misuses and misunderstandings that have occurred in the name of attention to professional boundary concerns. We will also propose a more rational conceptual framework within which to consider boundaries. Although we are primarily focusing on psychotherapy, we will also note appropriate modifications for other psychiatric treatment modalities. Our goal is to hasten the arrival of the pendulum to the midpoint between the extremes of boundaryless clinical practice and rigid overreaction.
A boundary may parsimoniously be defined as the "edge" of appropriate behavior (1). This edge resists sharp definition not only because of the diversity of psychotherapeutic approaches, but also because within any one school of thought, most clinicians would agree that the therapist must attempt to tailor the treatment strategy to the particular requirements of the individual patient. The establishment of clear boundaries is designed to create an atmosphere of safety and predictability within which the treatment can thrive. Professional boundaries involving such elements as the absence of sexual contact, the arrangements for the fee payment, the length of sessions, and the relative asymmetry of self-disclosure are not simply an arbitrary set of rules designed to make it easier to discipline deviant clinicians. Nor are they intended to create a remote and rigid way of relating between therapist and patient. On the contrary, external boundaries are established so that psychological boundaries can be crossed through a variety of mechanisms common to psychotherapy, including empathy, projection, introjection, identification, projective identification (5), and the interpretation of transference. Langs’s concept of the frame (10) and Winnicott’s notion of the "holding environment" (11) address similar concerns.
We have suggested that boundary transgressions can be divided into two broad categories (1). The first is a "boundary crossing," a benign variant where the ultimate effect of the deviation from the usual verbal behavior may be to advance the therapy in a constructive way that does not harm the patient. (The use of the term "boundary crossing" in its technical sense was suggested by Axel Hoffer, M.D., and Douglas Ingram, M.D., in personal communications.) Some of these may be fully appropriate human responses to unusual events that might involve physical contact. A patient stumbled as she was leaving the office and fell to the floor. The therapist helped the patient up and made sure that she was all right. A patient entered her therapist’s office and announced that she had just received news that her son had died. The patient reached out to embrace the therapist, and the therapist accepted the embrace as the patient sobbed. A failure to respond in a human way in such situations would very likely devastate the patient and might even lead to a premature interruption of the therapy.
Another variant of a boundary crossing is the countertransference enactment that is largely unconscious and occurs spontaneously in the course of the therapy (12—15). While these may be departures from the usual mode of therapeutic interaction, they are subsequently discussed and explored in a way that is productive for the treatment process. For example, at the end of a long and frustrating session, a therapist felt as though nothing she had said had been helpful. She had been preoccupied with the fact that her own mother was ill, and she had not been as available to the patient as usual. As the therapist and patient got up from their chairs, the patient asked, "So where are you going next week?" Without thinking, the therapist responded, "I’m going to visit my mother in Colorado. She’s very ill." The patient said, "Oh, I’m sorry," and left the office looking worried. The therapist instantly recognized that she had revealed information that might have burdened the patient with her own problems. As she reflected on her enactment, she recognized that she had been feeling guilty about not having been very helpful because she was so preoccupied and that she had unconsciously tried to gain some sort of absolution or forgiveness from the patient by explaining the situation to him.
After her 1-week absence, the therapist brought up what had happened at the end of the session with the patient and explained that she probably should not have burdened him with the information she offered. She went on to suggest that they might beneficially explore his reactions to it. In the course of sharing some of his thoughts about what had happened, the patient noted that his mother had always confided her problems to him, and he thought something like that was repeating itself in the therapeutic relationship. In this instance no lasting harm was done to the patient, and the event turned out to be a useful focus for further exploration. An important factor that led this enactment to be constructive is that it was discussible by the therapist and patient.
The second broad category is the "boundary violation," when the transgression is clearly harmful to or exploitative of the patient. In contrast to the boundary crossing, the boundary violation is usually not productively discussed by the therapist and patient and may also be part of a repetitive practice. The harm may range from wasting time and therapeutic opportunity to inflicting severe trauma. Examples include the therapist who hugs the patient at the end of each session and tells him or her "I love you," the therapist who asks the patient to pick up his or her dry cleaning, the repeated disclosure of the therapist’s own personal problems in a way that burdens the patient, and of course, overt sexual contacts.
A significant challenge to forensic experts consulting on such matters is often the distinction between a boundary crossing and a boundary violation. In addition to the aforementioned issue of whether or not the transgression is discussible, another helpful differentiating factor is context.
Thinking about boundaries can lead one to an absurd end point, unless one understands the critical role of the context in which behavior occurs (1). The context may be constituted by the treater’s professional ideology, the presence or nature of informed consent by the patient, the point in the therapy at which behavior occurs, the respective cultures of the dyad, and such environmental factors as whether therapy occurs in a small town or in an urban center and whether public transportation is available. For example, a therapist who gives a patient a ride home in a blizzard might be judged differently, depending on whether the therapy occurs in a prairie town or a major city with a subway system, whether the patient feels coerced by the therapist to accept the ride, and whether the therapist also gives the patient rides when the weather is mild.
During the era of extended inpatient treatment, hospital therapists traditionally visited patients in the shops where therapeutic activities were going on and had conversations with their patients during leisurely walks on the hospital grounds. Many a shy patient with schizophrenia was far better able to communicate in the context of motion and open space than in the felt intimacy and imprisonment of an office. In most hospital settings today, the therapeutic culture is very different. Therapists and patients have only a short time to get acquainted (and thus to know each other’s motives and intentions). Hence, taking a walk with a patient will likely be seen as a potential boundary problem, and visiting an activity group as an invasion of privacy.
Or consider this example in which cultural differences, timing, and transference combined to create a context for an action felt as a boundary transgression by the patient:
In his native Vienna an analyst had courteously helped female patients put on and take off their overcoats. This was part of everyday politeness, and he would have been seen as rude had he abstained from this ritual. Upon moving to the Midwest he began the analysis of a young woman in early September. By the time cold weather began, the patient, a native Midwesterner, had developed erotic feelings toward the analyst, felt exquisitely embarrassed about them, and had not yet plucked up the courage to reveal them. On the first cold day of winter she experienced his swift advance to help her take off her coat as if he were about to disrobe her. Her sullen awkwardness signaled trouble to the analyst, who asked her what had disturbed her. She mumbled that she was not used to men taking her coat off. While it would still be a while before a full examination of her fear of (and wish for) intimacy would become possible, he elected in the meantime not to repeat his polite action. The analysis proceeded, but in today’s world and with a patient more ill and contentious than this one, this analyst’s behavior might well be misconstrued by overzealous fact finders as an inappropriate violation of a boundary.
In this situation a completely benign act of politeness was perceived subjectively by the patient as a potentially exploitative boundary violation, in part because of the stage of the treatment. This latter issue, as noted, represents a relevant context to aid in distinguishing the impact of a behavior from its intent.
The matter of context is all too often disregarded by fact finders and decision makers in this area, although it is essential to determine, inter alia, whether a specific behavior represents a boundary crossing or a boundary violation; indeed, the identical behavior may constitute either a boundary crossing or a boundary violation, depending entirely on the context in which it occurs.
In the following example, the context omitted from consideration was the treater’s task and discipline:
A female case manager (a community mental health center staff member given the responsibility of coordinating care and rendering practical assistance to patients) took many trips with a female patient to necessary appointments, accompanied the patient’s family to the beach with her own family to encourage socialization and to model parenting, and performed many other out-of-office activities. The patient, a paranoid young woman, formed a strong attachment to the case manager and became furious when the relationship had to end because of the latter’s pregnancy. At that point she persuaded her parents to bring an ethics violation complaint against her former helper. In subsequent litigation, no evidence emerged that the patient had been exploited or harmed in any way. But the regulating board hearing the case held the case manager to the standard of a psychoanalyst as to what constituted professional boundaries.
Since case management has always been a far more hands-on role than classical psychoanalysis or even psychotherapy, the preceding use of the wrong yardstick appears to reflect a fixed and mechanistic view of boundary questions; yet such misapplications are not uncommon in civil litigation, where the idea of an absolute boundary is appealing to plaintiffs’ attorneys seeking a clear standard.
A therapist faced with deciding whether to call or visit a patient with a serious medical illness confined to a general hospital must use carefully considered clinical judgment. He or she would have to weigh the humanitarian benefits of a visit to the hospital against the potential for the patient to react with feelings of exposure and intrusion. Here the context might depend on the patient’s history, the seriousness of the illness, the nature of the therapeutic alliance, and the patient’s preferences. Obviously, a frank discussion between therapist and patient about the pros and cons of such a visit would be of considerable assistance in resolving the therapist’s dilemma.
Even among professionals of similar disciplines, the task may vary widely. As we have noted previously (1), a behavior therapist assisting an agoraphobic patient through the use of in vivo exposure may drive the patient to a shopping mall to encounter the feared situation. A clinician engaged in psychoanalytic psychotherapy would be in serious need of supervision if similar activities were going on. Even within psychodynamic therapy, however, the appropriate frame varies from patient to patient (5, 16, 17). Certain patients require greater degrees of verbal activity or therapeutic self-disclosure for them to feel engaged in a treatment process. Also, some therapists may have a more self-revelatory style than others. The therapeutic frame is necessarily constructed by the two subjectivities that meet in the consulting room. Hence, it is a collaborative construction based on the particular characteristics of a specific dyad. A senior male therapist treating a 19-year-old young man who finds it difficult to talk may use conversations about sports to engage the patient at a level where he feels comfortable about relating to the therapist. In the course of this conversation, the therapist may reveal a great deal about his interests in sports, his preferences for a certain team over another, and his activities during the past weekend (if he happened to attend a sporting event). While he is self-disclosing, he is also adapting the frame to connect with the patient in a way that facilitates a therapeutic alliance.
One of the most frequent misuses of the concept of boundaries revolves around self-disclosure. Contributions in the literature (18—20) have stressed that therapists are disclosing aspects of themselves every time they open their mouths. Indeed, the visual cues available in the therapist’s office speak abundantly of the character of the occupant through the order or disorder of the office contents, the expression of taste (or lack of it) in the decor, and the comfort or discomfort of the furnishings. While these disclosures affect first impressions and generally fade in importance under the influence of getting better acquainted and of transference, their impact illustrates the impossibility of avoiding being known.
The boundary issue is not whether self-disclosure occurs or does not occur. Rather, the key issue is what the therapist self-discloses and whether the therapist burdens the patient with personal problems in a manner that reverses the roles in the dyad. For the therapist to acknowledge his or her anger at the patient can be useful in advancing the therapy. For example, a hospitalized patient lied to her therapist and the nursing staff about possessing a razor blade and delighted in having won their trust only to flaunt her power to deceive them. The therapist felt betrayed and hopeless as to his ability to help someone so deceitful and mean. He was encouraged by his supervisor to see his patient briefly at their next scheduled meeting to inform her that he was too angry to be helpful to her for the moment and that he would need some time to think about what had happened and whether he would be able to go on being her therapist. Such a period of reflection allowed him to explore his emotional reaction, while affording the patient an opportunity (with the help of the skilled nursing staff) to consider the consequences of her behavior.
There are, of course, other countertransference feelings that therapists should not disclose to their patients. It is rarely useful to say to a patient, "I have sexual feelings for you" (21). Saying "I hate you" is also unlikely to be helpful. Nevertheless, the therapist might well examine with the patient certain interactions that are likely to lead to hateful or sexual feelings. Here again, context and the manner of disclosure are critically important.
Although ostensibly concerned with general fitness to practice, boards of registration are often recipients of specific complaints couched in terms of boundary issues. Those issues can be remarkably trivial and curious.
During a visit with her internist, a patient reported having had a number of recent losses through deaths of close family members. Discussion of these losses led her to burst into intense sobbing. Later that evening the internist called the patient at home to see if she was all right. She reassured him that she was, but she later reported him to her state licensing board for the alleged boundary violation of calling her at home.
Boards themselves vary to a striking degree in their rigor, flexibility, and, regrettably, punitive attitudes toward the clinicians they license. The following excerpt of sworn testimony at a licensing board hearing in a Midwestern state captures another variation of an extreme position on nonsexual boundaries:
Q [board prosecutor]: Did you also find in the testimony and the records that [the patient] attended the same church [the accused doctor] did?
A [psychiatric expert witness:]: Yes.
Q: Do you recall how [the patient] became acquainted with that church?
A: Yes, I believe [the doctor] gave the patient a list of churches, including his own [emphasis added; other testimony revealed that the patient had asked for referral to "noncoercive" local churches].
Q: If [the doctor] provided a list of four or five churches to [the patient], one of them was his own and the patient knew that, which church would this patient likely choose? Defense attorney: Object, calls for speculation.
Judge: I will sustain that. You can ask him if he has got an opinion on that.
Q: In your opinion, [doctor serving as expert witness], as a psychiatrist, if a psychiatrist provided a list of four or five churches to a patient and that patient was having idealizing transference with that psychiatrist, one of those churches was the psychiatrist’s and the patient knew that, in your opinion, which church would the patient choose?
A: In my opinion, it would be likely that the patient would be influenced to go to the church that the psychiatrist recommended, that the psychiatrist was going to. [For these and all the following subjunctives, emphasis was added.]
Q: And why would a patient do that?
A: Because the patient would trust in the psychiatrist’s judgment, would want to be close to the psychiatrist, would want to do what the psychiatrist does or recommends.
Q: Was it within the minimum standard of care in the late 1980s for [the doctor] to recommend his own church to [the patient]? [Note that in this jurisdiction the licensing board standard is the extreme position, "minimum standard of care," rather than the more familiar "average reasonable care" of civil litigation.]
A: Well, first of all, I don’t think the psychiatrist’s job or duty is to recommend churches to patients. There are others who can do that. [Note here that the situation is one where the patient asked and the doctor responded, rather than whether recommending churches is a doctor’s duty, as this witness’s testimony implies.] Second of all, it would increase the likelihood that the patient would be in a social interaction with the psychiatrist and would provide an opportunity that would be ripe for all kinds of boundary problems, boundary blurrings, and boundary violations.
Q: Can you describe some of those boundary blurrings that occur?
A: They could be sitting next to each other in the church. They could be involved in church activities together. The psychiatrist and his own family might be involved with the patient and his family. There would just be an increase in the likelihood of significant social interaction between them.
None of the preceding colloquy establishes how the patient would be realistically exploited by the doctor who is answering the patient’s request for the names of churches. Instead, all of the preceding subjunctives posit a domino theory or slippery slope of possibilities, as if the church referral were a kind of quicksand within which the inextricably trapped doctor and patient would be forced to cross boundaries, unable to make other choices (e.g., going to another of the five churches). Disregarding the fact that no literature supports the claim that such church referral is below the minimum standard of care, we note that the late 1980s, the period referred to in the preceding testimony, was indeed the time that professional awareness of nonsexual boundary issues was first appearing in professional journals. Ironically, that same era marked psychiatry’s dawning awareness of the importance of religion and spirituality in patients’ lives. One "might" conclude that the speculative hypotheticals in the preceding excerpt are too gossamer a footing on which to place the weight of a physician’s licensure and future career in a licensing board complaint.
This testimony also reflects the common confusion between boundary violations and chance extratherapeutic contacts. Therapists who have consulted us have expressed concern that they will be accused of violating boundaries if they run into their patients in a club, at a theater production, at the gym, or in a social gathering. Such chance meetings are, of course, inevitable, especially in small towns (although even in large cities certain therapists and patients find that their paths cross frequently because of common interests). The only significant risk of a boundary violation in such cases is if the therapist openly acknowledges that he or she is treating the patient and thereby breaks confidentiality. In most cases, the therapist can handle the situation appropriately by following Karl Menninger’s dictum "When in doubt, be human." These encounters can be discussed later in the therapy in terms of their meaning to the patient; indeed, it is just such discussion that preserves the therapeutic process despite the encounter. Only planned meetings in a motel, a restaurant, or similar settings for purposes that in no way relate to the therapy are true boundary violations.
In sincere efforts to provide guidance for clinicians, some states have promulgated statutory or regulatory advisories on boundary issues. These are subject to the same mechanistic flaws as the previous contexts. Consider a recently passed Rhode Island civil law that defines "sexual contact" as various forms of intercourse, sustained kissing, fondling, and the following:
Exhibition by the MHP [mental health professional] in view of the patient or former patient of the MHP’s genital area, groin, inner thigh, buttocks, or breast; voyeurism by the MHP in the form of viewing the patient’s or former patient’s genital area, groin, inner thigh, buttocks, or breast. The Act provides a civil cause of action against any MHP who engages in conduct as defined in the chapter. (22)
This broadened definition of sexual contact appears to open a hornet’s nest of confusion. With women’s skirts routinely above the knee, any time a female therapist or female patient crosses her legs, she exposes her inner thigh to the other party. Also, many male and female patients appear at therapy wearing shorts. Therapists would be forced into maintaining rigid and unswerving eye contact in a way that would totally constrict their free-floating attention and free-floating responsiveness to subtle enactments within the patient-therapist field. For psychodynamic therapy to be effective, therapists must immerse themselves in the experience of the patient without feeling undue restrictions regarding where they may look or how they might feel or think. Patients who characteristically evoke erotic reactions in others because of the way they behave, look, dress, talk, or move will inevitably evoke similar reactions in the therapist. This countertransference identification may lead the therapist to think about the patient in sexual terms. Like any other transference-countertransference interaction, the meaning of this erotic tension must be explored and understood to help the patient in other relationships outside the therapy.
Freedom of thought is necessary for the therapist’s achievement of an optimal state of effectiveness. Restrictions on where the therapist looks, how the therapist feels, or what the therapist thinks are destructive to the process. When regulatory laws define "sexual contact" so broadly that they include looking at the patient’s thigh, things have gone too far.
A similar problem has arisen in regard to restrictions on verbal discourse. For example, a Minnesota medical licensing board regulation (which has the force of law) defines grounds for disciplinary action to include "Engaging in conduct with a patient which is sexual or may reasonably be interpreted by the patient as sexual, or in any verbal behavior which is seductive or sexually demeaning to a patient" [emphasis added] (23, p. 246).
This wording places in jeopardy such basic psychiatric and psychotherapeutic procedures as taking a sexual history, exploring the patient’s sexual fantasy life, or understanding erotic transference. If a clinician makes a neutral comment, for example, that is interpreted as sexual by the patient, this perception is an important component of the patient’s transference that needs to be explored. The problem with this kind of regulation is that a transference perception becomes reified into a fact for legal purposes. Casting this broad a net may create such anxiety in therapists that they systematically avoid discussions of a sexual nature with their patients. Many patients have problems in the area of sexuality, and most have some difficulty in bringing them up in treatment. If a patient senses that the therapist becomes distressed when sexual issues are raised, the patient may feel that his or her sexuality is indeed too disgusting or shameful for open discussion (24). The therapist may be realistically concerned that exploration of any sexual topic about which the patient is conflicted may result in that patient’s feeling demeaned and therefore violated (not to mention aroused).
On the basis of the foregoing, it is clear that any efforts that hasten the boundary pendulum’s arrival at a rational midpoint are welcome. In addition to openly discussing the issue as we have done, the following approaches might hasten a return to reason.
Statutes and regulations designed to protect patients should address the problem of the critical role of context in determining when a violation has occurred. Use of mechanistic checklists should be replaced by thoughtful examination of the boundary issues on a case-by-case basis—a procedure that courts and licensing boards have long performed to the point of complete familiarity.
Expert witnesses should familiarize themselves with the complexity of boundary issues in order to present ethical testimony that does justice to the variability and context dependence of the standard of care.
Educators and supervisors should explain the role of boundaries in terms of protection of patients and of the therapeutic frame, rather than presenting boundary lists as compendia of generically forbidden behavior. Ethics must be taught in the context of clinical concepts such as neutrality, therapeutic alliance, transference, countertransference, enactment, and so forth (25).
Ethics committees should similarly look with flexible standards at the critical question of context to assess the reality of exploitation or harm, rather than using lists. While concrete harm may be hard to demonstrate in all cases, guidelines aim at decreasing situations of potential harm.
Finally, we must recognize that clinical psychiatric practice is as much an art as a science. Our primary tools are ourselves. As such, we are imperfect and often need to flounder a bit before we strike the optimal balance between intimacy and distance that is favorable for a particular patient and a particular therapist at a particular time in treatment.