A board-certified psychiatrist saw a woman in individual psychotherapy for ten years. During the course of the therapeutic relationship, he negotiated with her to sell her two of his boats, sight unseen. Additional transactions involved sales of her personal property to him: Waterford crystal, china, and a silver service, the last of which was appraised at $1,600 but was purchased by the psychiatrist for $200. In the same year he accepted a refrigerator and a dining table with six chairs as gifts. During the course of these commercial transactions, the patient had run up a significant bill with the psychiatrist. She sold her father's coin collection to the psychiatrist for $1,000 as a means of getting one of the boats into the water. Within a year, the bank repossessed the boat and the patient declared bankruptcy.
This vignette is just one among many cases in the spectrum of behaviors in boundary problems with patients. The last quarter century has produced an extensive literature aimed at clarifying the nature of the psychotherapy relationship and the variety and complexity of possible boundary difficulties in the therapeutic dyad (1,2,3,4,5,6,7,8,9,10,11,12). Television and movie dramas have portrayed boundary dilemmas in various ways, humorous and straight; consider the television program The Sopranos and the film Analyze This. Despite broad agreement in psychiatry that sexual misconduct and other boundary violations can cause notable harm to patients, some of our most senior and accomplished practitioners and teachers continue to find themselves embroiled in these difficulties. Next to suicide, boundary problems and sexual misconduct rank highest as causes of malpractice actions against mental health providers. Nevertheless, psychiatric training about boundary issues has continued to be ineffective despite today's wider awareness of these caveats, increased recognition of the severe dangers to patients, threats to psychiatrists' licensure from complaints to boards of registration, and professional ostracism. We speculate that these deficits of modern psychiatric training and practice may reflect the additional pressures of managed care—fostering a paradigm shift in psychiatry away from psychotherapy and toward pharmacology and excessively brief psychotherapies—but that the result is the same: We continue to see a steady stream of boundary violations, both sexual and nonsexual, in all psychiatric contexts (13).
We believe that, despite wide publicity, denial— "This couldn't happen to me"—must also play a significant role in the persistence of the problem.
Intensifying the complexity of boundary dilemmas is a form of backlash, expressed as strong criticism of boundary theory by a few scholars (14,15,16). These authors chide proponents of boundary theory for promoting technical therapeutic rigidity, excessive concerns with risk management, stultification of flexible innovation, and proposals of procrustean prescriptions of proper performance. Nonetheless, among the vast majority of practitioners, sensitivity to boundary issues remains an essential element of good clinical work that merits attention in training and practice.
Despite growing awareness of boundary issues and possible harms of boundary violations, the problem of maintaining boundaries obviously continues, as shown by the case vignette above. In this overview of the topic we aim to clarify, particularly for novices in clinical work, the theory behind boundary maintenance and the related pitfalls commonly encountered in the therapist-patient relationship. We examine therapist factors that make the development of boundary problems more likely, patient factors that make responding to the problems more difficult, and common issues in the dyad. We also provide suggestions for resolving these conflictual issues while preserving the efficacy of the therapy.
A boundary is the edge of appropriate professional behavior, a structure influenced by therapeutic ideology, contract, consent, and, most of all, context (1). Additional information about boundary theory may be found elsewhere (2,3,5,17). Boundary violations differ from boundary crossings, which are harmless deviations from traditional clinical practice, behavior, or demeanor. Examples of crossings include helping up a patient who has fallen, giving a patient an emergency taxi fare in a snowstorm, or accepting an invitation to attend a wedding. Neither harm nor exploitation is involved. Boundary violations, in contrast, are typically harmful and are usually exploitive of patients' needs—erotic, affiliative, financial, dependency, or authority. Examples include having sex or sexualized relations with patients, exploiting patients to perform menial services for the treater, exploiting patients for money or for financial demands beyond the fee, and generally using patients to feed the treater's narcissistic, dependent, pathologic, or sexual needs (1).
Although intentional breaches of boundaries clearly focus on exploiting the patient, violators are often not aware that any exploitive action has occurred—for example, employing a student patient, rationalized as helping the patient with the cost of the therapy. However, this boundary violation creates dual roles, and thus confusion, in the relationship. Patients are governed by no professional code; therefore, maintenance of boundaries is always the responsibility of the clinician. Thus if a patient requests, demands, provokes, or initiates a boundary violation—as many do—the clinician must refuse to participate in that behavior and then must explore the underlying issues, aided by consultation as indicated. Repeated demands to breach boundaries should prompt personal and consultative review about the viability of the treatment relationship, especially if the boundary issues become the only subject the patient can discuss. As always, documentation in clinical notes of the patient's refusal to discuss other subjects—coupled with the therapist's seeking consultation before, during, and after taking any action that impinges on boundaries—is the best protection against even inadvertent harm to the patient and against liability resulting from one's interventions (17).
This discussion of therapist factors is accompanied by three caveats. First, a therapist's personal problems do not mean a release from responsibility for setting and maintaining therapeutic boundaries; the therapist always bears the professional burden in this regard. Second, discussions of boundary problems sometimes focus on the "bad apple" model: boundary problems and sexual misconduct occur only with a few bad apples, and the simple solution is to kick those persons out of the field (18). This simplistic view misses a central point of our discussion: boundary issues arise in all therapies and for all clinicians, apparently irrespective of the number of years of experience, and even for those practicing only psychopharmacology. The relevant question is whether the difficulties can be successfully surmounted. Third, repeated boundary challenges by a particular patient should lead to a review of whether the treatment relationship is a wise one.
Therapists must learn to recognize the following trouble spots as risk factors for developing boundary difficulties.
Empirically, midlife and late-life crises in therapists' development appear repeatedly as common precipitants of boundary problems with patients, although early-career practitioners are not immune from boundary difficulties. For this last group, the challenges include difficulty establishing a practice; an excessive need to please patients, associated with filling empty hours in the schedule; and balancing the demands of family and professional life. For therapists in general, the effects of aging, career disappointments or unfulfilled hopes, marital conflict or disaffection, and similar common stress points are often associated with a therapist's turning to a patient for solace, gratification, or excitement (3,18).
Retirement, job loss, job change—even promotion—or job transfer may produce predictable anomie that makes a therapist susceptible to crossing the line with patients. In addition, financial reversals, working with managed care, stock market declines, envy of patient wealth, greed, and other factors may increase clinicians' susceptibility to fiscal exploitation. Indeed, the authors' consultations with other clinicians suggest that some cases of financial exploitation outnumber the sexual ones.
Therapists' illness appears to increase their vulnerability to turning inappropriately to a patient for solace and support, although this topic has been relatively underexplored (personal communication, Duckworth K, April 1990). The authors' consultations with other clinicians suggest that death anxiety and fears of mortality play a role in a therapist's turning to a patient for comfort.
Loneliness and the impulse to confide
A therapist encountering some life difficulty and seeking a "sympathetic" ear may struggle with the need to confide in a patient about financial reversals, marital or sexual problems, professional setbacks, problems with his or her children, and the like. This lapse may precipitate a role reversal in which the patient takes care of the therapist. Indeed, many patients have been someone's "therapist" in their family of origin and may slip with familiar ease into this inappropriate role. In other cases the otherwise laudable desire to find common ground with a patient may miscarry if the therapist indulges a need to self-disclose.
Self-disclosure, one of the most controversial boundary issues, is an issue that often leads to confusion and uncertainty among therapists, ethics committees, and boards of registration (19). Demanding patients at times insist that the therapist disclose personal data to restore symmetry to the therapeutic situation, to demonstrate the therapist's commitment to the patient and the therapy, or to dispel or confirm a fantasy about the therapist that is so preoccupying to the patient that it interferes with treatment. In part, therapists' uncertainty stems from the empirical observation that self-disclosure is often the final boundary excursion before sexual relations, even though self-disclosure does not in itself lead inevitably to that outcome. Furthering the confusion are the different approaches that different ideologies assign to the role of self-disclosure in clinical work; for example, it may be common in reality therapy but eschewed in psychoanalysis. In the name of "honesty," therapists may slip into countertransference-based interventions, such as "When you say such things, I become sexually aroused; how can we understand that?"
Self-disclosure may cause no problems in the therapy, but even in response to seemingly innocent queries, it may intrude on the patient's psychic space or replace a patient's rich and clinically useful fantasy with dry fact, stripped of meaningful affect. To use a perhaps extreme example, a patient who hears that a therapist is Catholic may have greater trouble or discomfort discussing her abortion. Similarly, "Are you married?" may stand in for "Are you gay? Are you available? Have you failed in past relationships?" The point here is that the therapist's inner awareness of longing to self-disclose to or confide in a particular patient may serve as an alert to potential boundary difficulty to come.
Idealization and the "special patient"
Therapists must be alert for early harbingers of trouble in certain of their own countertransference attitudes toward patients (15,20). Typical views commonly associated with problems in maintaining boundaries include viewing the patient as "special"—for example, because of beauty, youth, intellect, artistic creativity, fame or status in the community, or therapeutic challenge. Unsophisticated therapists experiencing erotic feelings toward a patient may find these common, if not universal, feelings highly threatening, creating anxiety that may distort clinical judgment.
Such feelings are an excellent stimulus to seeking consultation or supervision but not to terminating therapy with patients or abandoning them, as some clinicians seem to believe. Sexual feelings, hostile feelings, and boredom are all responses to patients that therapists must handle within the process of treatment unless these reactions become unmanageable or are unresponsive to supervision and consultation. Clues to these attitudes may lie in the therapist's tendency to treat the patient as an exception to the usual rules of the therapist's practice: scheduling excessive or excessively long sessions, especially at the end of the day; giving permission to run up a high unpaid balance; making special allowances for the patient; and having nonemergency meetings outside the office. Therapists seeking consultation on such cases often begin the request with "I don't usually do this with my patients, but in this case. …"
Therapists with intact self-esteem systems are entitled to take pride in their work, but self-esteem—like all traits—can miscarry through excess and denial: "This couldn't happen to me." One would think that this problem is an especially common one for the younger, inexperienced therapist, and this is often the case. However, a pitfall that is especially relevant to very senior therapists, who are often sought out for consultation, is their inclination to brush aside the need to seek consultation themselves. Seasoned practitioners may believe that, given their level of experience, they can take risks in this area: "I have good control and I know what I'm doing." We knew of one such therapist, who resisted undergoing such a review on the grounds that he knew the consultant would tell him the relationship with the patient was wrong and should be terminated. In its extreme form, this narcissistic difficulty supports the belief that one is above the law and that the usual rules do not apply.
Problems with limit setting
Some patients—who cannot be blamed for the impulses—tend to press for boundary breaches for a variety of psychological reasons (21). The question then becomes, Can the therapist set appropriate limits on this intrusion? A common barrier to appropriate limit setting is the therapist's countertransference conflicts about aggression or sadism when the prospect of the patient's expected distress, discomfort, or frustration at being told "no" is intolerable to the therapist. When caught in such conflicts, therapists often feel that they cannot refuse patients' requests to violate a boundary. These therapists report feeling pressed or intimidated by patients' unrestrained rage.
Closed communities pose another sort of boundary problem. They may be small towns; isolated institutions like schools, convents, and communes; or subcultures with a restricted social compass, such as some gay or lesbian subcultures in urban settings. When there is only one store (or gym, pool, or post office) in town, one cannot avoid the possibility of encountering patients outside the office in nonprofessional settings. Such conditions require more circumspection and care about boundaries, not less. Simon and Williams (22) have provided an excellent discussion of this issue.
Finally, denial about early problematic situations, which can lead to their evolving into full-fledged boundary disasters, is another common factor in clinical misadventures—particularly with more seasoned and experienced therapists. Evasion, externalization, and rationalization may be used by the therapist to help maintain the pretense that boundary problems are not serious, not harmful, or even not occurring at all. Here, consultation can be extremely useful in gaining perspective, but all too often the need for a consultation is also rationalized away.
Patients generally expect that physicians will treat them with respect and act in their best interests. Despite widespread coverage of therapist misconduct in various media, sexual misconduct and its common precursor, boundary violations, are sometimes hard for patients to recognize, or to report if recognized. Moreover, patients with some disorders appear to have more trouble with boundaries than those with other disorders (21). A number of factors may account for these vulnerabilities.
Patients in psychotherapy may seek dependency rather than autonomy. With some patients, an intensely enmeshed, symbiotic relatedness may result, making it difficult for the patient either to break away or, later, to report the matter. In one case we know of, a patient described feeling tied to a former therapist in part because she was aware there was "something wrong" with what was going on in therapy, but she also felt that it was a basis for closeness with the therapist: "We were in this together." This condition may lead to the patient's clinging tighter to the image of the protective therapist.
Changing roles: from victim to actor
Initially a patient comes to treatment seeking help and, in part through transference, imbues the therapist with healing powers and intent. The patient may then seek a dependent position that precludes questioning or challenging the therapist's decisions or actions. To challenge the therapist would mean altering one's role and identifying with the therapist's aggression to become more aggressive and less dependent. In one case, a patient sought therapy after the traumatic loss of her husband and became involved in an exploitive relationship with a therapist. She was unable to report this relationship to professional boards for some time after the therapy—and the relationship—ended. Once she did so, however, she became more assertive and instituted support groups for abused patients.
Some patients enter therapy to deal with the effects of previous, often childhood, trauma. For such patients, boundary violations and even outright abuse by the therapist may recapitulate this early experience, including felt helplessness to enact any escape or remedy. The familiarity of the victim role may increase the likelihood of repetition, a condition described by one clinician as the sitting duck syndrome (23,24). Tragically, such repetition of childhood patterns may recur with a subsequent treater.
Patients involved in boundary violations or sexual misconduct often struggle with self-blame, accusing themselves of failure to know better, failure to recognize abuse, having made foolish choices, and so on. Others fault themselves for causing the therapist to lose control or cross the line or for being "too seductive" or believe they bear full responsibility for the misconduct. None of these views, of course, captures the true picture.
Though perhaps owing much of its force to the transference, intense feeling can develop in therapy, if only because of the inherent intimacy of the situation, especially if the patient has few or no other relationships on which to draw. The relationship with the therapist may appear the only or the last chance for "true love" in the patient's sphere. Indeed, a small percentage of patients enter treatment specifically to have an intense emotional experience in a relationship of some kind, even a paid one. In one legal case, a patient tearfully told an expert witness that, although she knew that the misconduct was wrong and that she had been taken advantage of, she despaired of ever having such an exciting relationship again.
Most boundary violations occur in the context of a helping relationship the patient depends on. It is difficult for the therapist to discern what is help and what is overinvolvement, and it can be very difficult for the patient to give up the relationship. Psychoanalyst Peter Fleming, speaking on a panel on boundaries, noted that a long-term patient who had entered a nursing home began to call him "honey" and "dear" rather than "Dr. Fleming" and to touch him a good deal when he got up to leave her room. He became concerned and raised his concern with her, which led to the patient's sobbing that she had lost her memory and could not recall his name. Had he not dealt with this boundary change, he would not have discovered the problem.
One stimulus for this article was our increasing awareness that the changing focus on the economics of health care delivery in this country has altered the nature and content of training for mental health professionals. Psychodynamic theory, with its central discussion of the role of transference, is now taught less and, not surprisingly, requested more by trainees to enhance their understanding of the psychotherapeutic relationship. A psychodynamically naive therapist who becomes the focus of idealization by a patient or who is placed on the positive side of a good doctor-bad doctor split by a patient with borderline personality disorder may feel that the patient is experiencing true love—a situation that must be acted upon. Whatever the current attitudes toward psychoanalysis, our professional schools have an obligation to teach trainees about transference and boundary issues.
Especially for younger clinicians and trainees, concerns are often expressed about the possible stifling of novel, innovative approaches to treatment of a patient or treatment in general. In designing these new approaches, the clinician can avoid both the Scylla of too little attention to boundaries and the Charybdis of too rigid an approach to them by keeping in mind the critical issue of maintaining sensitivity without exploiting the patient. Training techniques using films and videotapes and including presentations by victims and offending psychiatrists provide for more innovative approaches (25,26).
Important aspects of the supervisory relationship are the dynamic learning opportunities for all participants, both trainees and supervisors. In another case, a senior forensic psychiatrist was asked to consult about the dangerousness of a former patient who was a possible stalker. Unraveled, the case proved to be one of a patient who began to experience erotic feelings for his female therapist—feelings that she did not know how to handle. Two successive layers of senior supervisors could not deal with this issue either, and the therapist terminated the psychotherapy on their recommendation. In reality, the baffled patient had taken to hanging about the clinic trying to get a straight answer about what had happened—hence, he was a "stalker."
This vignette underscores the importance of having supervisory resources able to handle dynamic issues at different points in the course of treatment. Supervision provides the ideal setting for emphasizing and clarifying to the trainee how boundary issues inevitably arise in clinical work and how they may be managed successfully. Boundary questions commonly evoke countertransference issues, which may also be profitably explored in the protected supervisory context, as well as in the clinician's personal therapy. Supervisors' openness to seeking consultation presents another learning opportunity for trainees about the complexities of the work.
In a grim paradox, consultation—which would often make possible the solution of a therapeutic boundary problem—is all too often scanted, for reasons deriving from the same therapeutic knot that first produced the boundary problem. As noted above, therapists should consistently maintain a low threshold for seeking consultation and should respond positively when a patient requests it and welcome the occasion for both clinical and risk-management reasons. Therapists may refuse consultation because they "know" the consultant would urge them to stop treatment and get out of the relationship—an outcome they could not tolerate. Obviously, this is an inappropriate view of consultation. This individual problem is heightened by denial and resistance on the part of training institutions, especially when the boundary-violating practitioner is a senior clinician who may have trained many in the professional community (27).
The apparent challenges for the field are two. First, lower the threshold for nonjudgmental consultation with peers or specialists at early stages of difficulty. Second, heighten clinicians' awareness about issues and stages of the work that particularly present boundary risks; thus, clinicians will be likely to seek consultation early enough to benefit.
What guideposts can therapists employ to identify the need for consultation? Some have already been discussed as therapist risk factors: illness or changing life circumstances, feelings of specialness about the patient or the tendency to make exceptions, early boundary incursions and crossings, and so on. Other signs of boundary problems may include the feeling of being solely responsible for the patient's life; the feeling of being unable to discuss the case with anyone because of guilt, shame, or the fear of having one's failings acknowledged; and the realization that one has let the patient take over the management of his or her own case. Finally, noting that a patient is provoking the therapist to cross boundaries would be an excellent trigger for consultation.
Boundary problems are universal concerns, not merely the character defects of bad apples in the professional barrel; nor are they relevant only to those doing psychoanalytically oriented psychotherapy; nor are they confined to the offices of the private practitioner. We have presented an overview of characteristics of the patient and of the psychotherapist that may predispose to serious disruptions of the therapy process. Clinicians young and old, and in all settings, must overcome their understandable but damaging reluctance to fully examine this topic in every setting—didactic, training, consultative, and supervisory. Here we have tried to initiate that process and provide an overview for trainees and early-career and senior practitioners. We have done so because history teaches the hard lesson that this matter must be reviewed and revisited at least as often as the Physicians' Desk Reference, and for the same reason: the welfare of the patient and the serious and often tragic consequences of missteps in this area for both patient and practitioner (28,29,30).
The authors are affiliated with the program in psychiatry and the law at the Massachusetts Mental Health Center of Harvard Medical School, 74 Fenwood Road, Boston, Massachusetts 02115 (e-mail, firstname.lastname@example.org).