Boundaries in Psychotherapy

home  » boundaries » boundaries in clinical practice

The Concept of Boundaries in Clinical Practice:
Theoretical and Risk-Management Dimensions

Thomas G. Gutheil & Glen 0. Gabbard

Abstract: The authors systematically examine the concept of boundaries and boundary violations in clinical practice, particularly as they relate to recent sexual misconduct litigation. They selectively review the literature on the subject and identify critical areas that require explication in terms of harmful versus nonharmful boundary issues short of sexual misconduct. These areas include role; time; place and space; money; gifts, services, and related matters; clothing; language; self-disclosure and related matters; and physical contact. While broad guidelines are helpful, the specific impact of a particular boundary crossing can only be assessed by careful attention to the clinical context. Heightened awareness of the concepts of boundaries, boundary crossings, and boundary violations will both improve patient care and contribute to effective risk management.

View citation and copyright.

NOTE: Please follow this link to related articles, research, & widely-used practice guides.

"Role boundaries may be crisp or flexible or fuzzy, depending on the role under consideration and on the cultural climate."
-Ingram (1)

The concept of boundaries, particularly in the sense of boundary crossings and boundary violations, has come under increased scrutiny in relation to the wave of sexual misconduct cases (2) arising in litigation, ethics committee hearings, and complaints to boards of licensure. Like many concepts in psychotherapy, such as "therapy," "transference," and "alliance," the term proves slippery on closer observation. The literature tends to focus on patient-therapist sexual misconduct (3) as an extreme violation and not on the wide variety of lesser and more complex boundary crossings, many of which are, at first glance, less obvious but pose difficulties of their own for clinicians.

Clinicians tend to feel that they understand the concept of boundaries instinctively, but using it in practice or explaining it to others is often challenging. This latter problem is rendered more difficult by the tendency of the legal system, particularly plaintiffs' attorneys, to apply it mechanistically: any boundary crossing is bad, wrong, and harmful. Empirical evidence suggests that boundary violations frequently accompany or precede sexual misconduct (2, 4, 5), but the violations themselves do not always constitute malpractice or misconduct or even bad technique. However, modern clinicians should be aware of three principles that govern the relationship among boundaries, boundary crossings, boundary violations, and sexual misconduct.

First, sexual misconduct usually begins with relatively minor boundary violations, which often show a crescendo pattern of increasing intrusion into the patient's space that culminates in sexual contact. A direct shift from talking to intercourse is quite rare; the "slippery slope" is the characteristic scenario. As Gabbard (4) and Simon (6) have pointed out, a common sequence involves a transition from last-name to first-name basis; then personal conversation intruding on the clinical work; then some body contact (e.g., pats on the shoulder, massages, progressing to hugs); then trips outside the office; then sessions during lunch, sometimes with alcoholic beverages; then dinner; then movies or other social events; and finally sexual intercourse. Second, not all boundary crossings or even boundary violations lead to or represent evidence of sexual misconduct.

A clear boundary violation from one ideological perspective may be standard professional practice from another. For example, the so-called "Christian psychiatry movement" might condone the therapist's attendance at a church service with one or more patients, and various group therapeutic approaches or therapeutic communities may involve inherent boundary violations, as when some behaviorist schools permit hiring patients in therapy to do work in the treatment setting. Bad training, sloppy practice, lapses of judgment, idiosyncratic treatment philosophies, regional variations, and social and cultural conditioning may all be reflected in behavior that violates boundaries but that may not necessarily lead to sexual misconduct, be harmful, or deviate from the relevant standard of care.

Third, despite this complexity, fact finders-civil or criminal juries, judges, ethics committees of professional organizations, or state licensing boards-often believe that the presence of boundary violations (or even crossings) is presumptive evidence of, or corroborates allegations of, sexual misconduct.

To summarize the foregoing more concisely, albeit metaphorically, smoke usually leads to fire; one can, however, find smoke where there is no fire, and yet fact finders may assume that where there's smoke, there's fire. This metaphor is not trivial. In a notorious Massachusetts case in which the doctor accused of sexual misconduct was eventually exonerated), the Board of Registration in Medicine, the state licensing authority, noted in the course of the process, "There was an undisputed level of intimacy between the two [patient and doctor] that supports the inference of sexual relations" (transcript of board proceedings, citation withheld). In its language here, the board clearly articulated its "inference" of fire from the "undisputed" presence of smoke. Moreover, recent court decisions suggest a trend toward findings of liability for boundary violations even in the absence of sexual contact (7). On this basis, the risk-management value of avoiding even the appearance of boundary violations should be self-evident.

This communication has three goals: 1 ) to review the subject in order to define, describe, and illustrate the range of boundary issues, 2) to demonstrate that crossing certain boundaries may at times be salutary, at times neutral, and at times harmful, and 3) to suggest preventive and reparative measures for clinicians dealing with boundary violations in themselves and their patients.


What is a boundary? Is it too amorphous, protean, and abstract to define at all? Should we take refuge by saying, as St. Augustine was supposed to have said about time, "Time? I know what time is, provided you do not ask me"?

Part of the difficulty encountered in defining appropriate boundaries can be related to the historical tradition that modern therapists have inherited. The great figures in the field gave out mixed messages on the issue. Freud, for example, used metaphors involving the opacity of a mirror and the dispassionate objectivity of a surgeon to describe the analyst's role, but his own behavior in the analytic setting did not necessarily reflect the abstinence and anonymity that he advocated in his writings. He sent patients postcards, lent them books, gave them gifts, corrected them when they spoke in a misinformed manner about his family members, provided them with extensive financial support in some cases, and on at least one occasion gave a patient a meal (8).

Moreover, the line between professional and personal relationships in Freud's analytic practice was difficult to pinpoint. During vacations he would analyze Ferenczi while walking through the countryside. In one of his letters to Ferenczi, which were often addressed "Dear Son," he indicated that during his holiday he planned to analyze him in two sessions a day but also invited him to share at least one meal with him each day (unpublished manuscript by A. Hoffer). For Freud the analytic relationship could be circumscribed by the time boundaries of the analytic sessions, and other relationships were possible outside the analytic hours.

The most striking illustration of this conception of boundaries is Freud's analysis of his own daughter, Anna. Freud was not alone in establishing ambiguous analytic boundaries. When Melanie Klein was analyzing Clifford Scott, she encouraged him to follow her to the Black Forest for her holiday. Each day during this Vacation, Scott underwent analysis for a 2-hour session while reclining on Klein's bed in her hotel room (9).

D.W. Winnicott, another therapist of considerable stature, occasionally took young patients into his home as part of his treatment of them (10). In Margaret Little's report of her analysis with Winnicott (1 1 ), she recalled how Winnicott held her hands clasped between his through many hours as she lay on the couch in a near psychotic state. On one occasion he told her about another patient of his who had committed suicide and went into considerable detail about his countertransference reactions to the patient. He also ended each session with coffee and biscuits.

These boundary transgressions by highly revered figures have occasionally been cited in ethics hearings as justification for unethical behavior. We wish to stress that these behaviors are no longer acceptable practice regardless of their place in the history of our field.

The problem of the contradiction between what the master therapists wrote and how they actually behaved in the clinical setting was compounded because psychoanalysis and psychotherapy are treatments that occur in a highly private context. The boundaries of the therapeutic relationship and the characteristics of acceptable technique were thus highly subjective and lacked standardization.

This lack of clarity was partially addressed by Eissler's classic 1953 paper (12) in which he suggested that in the ideal situation, the analyst's activity should be confined to interpretation. Any deviation from that model technique was defined by Eissler as a parameter. As examples of parameters, he cited Freud's setting a termination date for the Wolf Man and proposed a hypothetical situation in which an analyst might command a phobic patient to expose himself to the feared situation. By this standard of technique, Freud's own behavior, such as offering a meal to the Rat Man, has been regarded as indicative of an earlier technique that Freud subsequently abandoned (13) or a human failing rather than a technical recommendation (14).


Lipton (8) took a strikingly different view of Freud's apparently unorthodox behavior with the Rat Man. He insisted that Freud's providing a meal for the patient should not be considered part of his psychoanalytic technique. Instead, it should be regarded as part of the nontechnical personal relationship that Freud had with this patient. He pointed out that in every analysis, the analyst is called upon to offer assistance in a personal way from time to time. While the ramifications and fantasies produced by such behavior should be thoroughly analyzed, it would be erroneous, in Lipton's view, to expand the concept of technique to include all aspects of the analyst's relationship with the patient. Lipton expressed the following concern: "Modern technique tends to move from the position from which the analyst's technique is judged according to his purpose to one from which the analyst's technique is judged according to his behavior" (8, p. 262). He pointed out that following Eissler's model of analytic technique in its literal terms would cause any noninterpretive comment or action on the part of the analyst to be construed as a parameter.

Another major problem with any attempt to derive definitions of boundaries from psychoanalytic concepts of technique is that technique changes with treatments that are less expressive than analysis. As one moves along the expressive-supportive continuum of psychotherapy (15), one relies less on interpretation and more on alternative interventions such as clarification, confrontation, advice and praise, suggestion, and affirmation. Similarly, partial gratification of transference wishes is associated with supportive psychotherapy, whereas it is generally eschewed in psychoanalysis or highly expressive psychotherapy. Hence, there may be a built-in confusion between the notion of therapeutic boundaries and adjusting the technique to the ego organization of the patient.

Another approach to defining therapeutic boundaries is to conceptualize a therapeutic frame (16, 17), i.e., an envelope or membrane around the therapeutic role that defines the characteristics of the therapeutic relationship. The analyst or therapist constructs the elements of the frame partly consciously and partly unconsciously. These elements include the regular scheduling of appointments, the duration of the appointments, arrangements for payment of the fee, and the office setting itself.

Does the patient's role have a boundary? Spruiell (17) has noted that although the frame is deliberately unbalanced, the patient invariably joins the analyst in elaborating the frame. Most clinicians would agree, basing this answer on recollected violations they have witnessed, such as the patient who refers to the therapist as "Shrinkie" or springs from the chair and tries without warning to sit on the therapist's lap. It is clear, however, that the patient's boundary is a more forgiving and flexible one. The patient cannot be stopped from calling the therapist names, and that is part of the therapeutic process. The patient can be late and that can be discussed, but the therapist should not be late, and so on. In any case our focus here is on the clinician's boundary.

Let us also agree that the role of therapist embraces the structural aspects of therapy in addition to the content; these include time, place, and money, which may, together with other aspects discussed below, represent possible sites for boundary crossings or violations to occur. If this exploration is to be useful, we should adopt the convention that "boundary crossing" in this article is a descriptive term, neither laudatory nor pejorative.

An assessor could then determine the impact of a boundary crossing on a case-by-case basis that takes into account the context and situation-specific facts, such as the possible harmfulness of this crossing to this patient. A violation, then, represents a harmful crossing, a transgression, of a boundary. An example might be the case of a patient who had experienced severe or traumatic boundary violations in childhood and who might consequently be highly sensitive to later violations, even those usually considered benign. Note also that the difference between a harmful and a nonharmful boundary crossing may lie in whether it is discussed or discussable; clinical exploration of a violation often defuses its potential for harm.

To organize the discussion we consider the matter of boundaries, boundary crossings, and boundary violations under a series of headings: role; time; place and space; money; gifts, services, and related matters; clothing; language; self-disclosure and related matters; and physical contact. Sexual misconduct as an extreme boundary violation has been extensively addressed elsewhere (2, 4, 6, 18) and is not separately reviewed here.

We should also point out that in addition to serving as antecedents to sexual misconduct, some of these areas of boundary crossing may represent ethical violations in and of themselves.


Role boundaries constitute the essential boundary issue. To conceptualize this entity, one might ask, "Is this what a therapist does?" Although subject to ideological variations, this touchstone question not only identifies the question of clinical role but serves as a useful orienting device for avoiding the pitfalls of role violations.

A middle-aged borderline patient, attempting to convey how deeply distressed she felt about her situation, leaped from her chair in the therapist's office and threw herself to her knees at the therapist's feet, clasping his hand in both of her own and crying, "Do you understand how awful it's been for me?" The therapist said gently, "You know, this is really interesting, what's happening here-but it isn't therapy; please go back to your chair." The patient did so, and the incident was explored verbally.

Although such limit setting may appear brusque to some clinicians, it may be the only appropriate response to halt boundary-violating "acting in" (especially of the impulsive or precipitous kind) and to make the behavior available for analysis as part of the therapy.

Almost all patients who enter into a psychotherapeutic process struggle with the unconscious wish to view the therapist as the ideal parent who, unlike the real parents, will gratify all their childhood wishes (19). As a result of the longings stirred up by the basic transference situation of psychotherapy or psychoanalysis, it is imperative that some degree of abstinence be maintained (20). However, strict abstinence is neither desirable nor possible, and total frustration of all the patient's wishes creates a powerful influence on the patient in its own right (8, 19).

In attempting to delineate the appropriate role for the therapist vis-a-vis the patient's wishes and longings to be loved and held, it is useful to differentiate between "libidinal demands," which cannot be gratified without entering into ethical transgressions and damaging enactments, and "growth needs," which prevent growth if not gratified to some extent (21 ). Greenson (22) made a similar distinction when he noted that the rule of abstinence was constructed to avoid the gratification of a patient's neurotic and infantile wishes, not to lead to a sterile form of treatment in which all the patient's wishes are frustrated.

Efforts to delineate the two varieties of needs often lead to problems in the area of defining the appropriate role for the therapist. Certainly, the patient may have legitimate wishes to be empathically understood, but when the therapist goes too far in the direction of trying to provide parental functions that were not supplied by the original parents, the patient may experience the therapist as making false promises.

Casement (21 ) expressed reservations about Freud's providing a meal to the Rat Man because of the possibility that the patient may have experienced Freud's taking responsibility for a particular part of his life as an implicit promise that Freud was prepared to take over responsibility for other areas of the patient's life as well. Clearly, a therapist cannot become the "good mother" or "good father" in a literal sense and attempt to make up for all the deprivations of childhood.

Even when therapists feel as though they are being coerced into a parental role by their patients, they must strive not to conform to the patients' expectations. Spruiell (1 7) made the following observation: "It is as disastrous for analysts to actually treat their patients like children as it is for analysts to treat their own children as patients" (p. 12).

The therapist's role is subject to some variation, of course. While most therapy is talking, there may be times when it is appropriate, for example, to write a letter on a patient's behalf. Under some circumstances, such a "breach of the frame" (16) might constitute a boundary violation, as when the therapist attempts to intervene in some extra-therapeutic realm of the patient's life (e.g., a therapist wrote a stern letter to a patient's employer rebuking the latter for giving the patient excessively burdensome tasks on the job). In addition, since different modalities of therapy are commonly combined, the "talking therapist" might appropriately give medications, conceivably by injection at times-a clear boundary crossing but presumably therapeutic and benign.


Time is, of course, a boundary, defining the limits of the session itself while providing structure and even containment for many patients, who derive reassurance because they will have to experience the various stresses of reminiscing, reliving, and so forth for a set time only. The beginnings and endings of sessions-starting or stopping late or early-are both susceptible to crossings of this boundary. Such crossings may be subtle or stark.

A male psychiatrist came in to the hospital to see his female inpatient for marathon sessions at odd times, such as from 2:00 to 6:00 in the morning, rationalizing that this procedure was dictated by scheduling problems. This relationship eventually became overtly sexual.

An interesting prejudice about violating the boundary of time has evolved in sexual misconduct cases, a prejudice deriving from the fact that a clinician interested in having a sexual relationship with a patient might well schedule that patient for the last hour of the day (although, of course, after-work time slots have always been popular). In the fog of uncertainty surrounding sexual misconduct (usually a conflict of credibilities without witnesses), this factor has gleamed with so illusory a brightness that some attorneys seem to presume that because the patient had the last appointment of the day, sexual misconduct occurred! Short of seeing patients straight through the night, this problem does not seem to have a clear solution. Admittedly, however, from a risk-management standpoint, a patient in the midst of an intense erotic transference to the therapist might best be seen, when possible, during high-traffic times when other people (e.g., secretaries, receptionists, and even other patients) are around.

Langs (23) noted that the boundary of time may be psychologically violated when the therapist brings up material from a previous session. Some patients, indeed, feel that this practice is disruptive and is a departure by the therapist from the here and now. However, most clinicians would regard this view as extreme, since effective therapy depends on continuity from session to session.

The issue of the appropriateness of phone calls between psychotherapy sessions is a controversial one, particularly when the patient suffers from borderline personality disorder. Some therapists view such phone calls as necessary and expectable in light of the borderline patient's difficulties with evocative memory (24, 25). In other words, the patient's inability to evoke a holding, soothing introject causes anxiety of catastrophic proportions related to the fear that the therapist has disappeared. Phone calls are a way of reestablishing contact with the therapist and soothing this anxiety, which might otherwise lead to ill-advised self-destructive behavior. On the other hand, other therapists view such calls as unnecessary and counter-therapeutic (26, 27). These therapists go to great lengths in the initial contractual period, at the beginning of therapy, to extract an agreement from the patient that phone calls will be used only in emergency situations.

This controversy reflects how a boundary violation may be defined according to the extent to which the appropriate treatment is viewed as having an expressive versus a supportive emphasis.


The therapist's office or a room on a hospital unit is obviously the locale for almost all therapy; some exceptions are noted in the next section. Exceptions usually constitute boundary crossings but are not always harmful. Some examples include accompanying a patient to court for a hearing, visiting a patient at home, and seeing a patient in the intensive care unit after an overdose or in jail after an arrest.

Some boundary crossings of place can have a constructive effect. As with medication, the timing and dosage are critical.

After initially agreeing to attend his analysand's wedding, the analyst later declined, reasoning that his presence would be inappropriately distracting. Later, after the death of the analysand's first child, he attended the funeral service. Both his absence at the first occasion and his presence at the second were felt as helpful and supportive by the analysand. They both agreed later that the initial plan to attend the wedding was an error.

A relevant lesson from this example is that boundary violations can be reversed or undone with further consideration and discussion. At times, an apology by the therapist is appropriate and even necessary.

Some sexual misconduct cases reveal space violations that seem to manifest wishes for fusion on the part of the therapist, as in the following case.

A lesbian therapist treating a female patient would contrive to use the bathroom at the clinic whenever the patient did so and, entering the adjoining stall, would attempt to continue the conversation. The relationship became overtly and exploitatively sexual, with the therapist often wearing the patient's clothes to work the next day after they had spent the night together.

Sorties out of the office usually merit special scrutiny. While home visits were a central component of the community psychiatry movement, the shift in the professional climate is such that the modern clinician is best advised to perform this valuable service with an opposite-sex chaperon and to document the event in some detail.

Sessions during lunch are an extremely common form of boundary violation. This event appears to be a common way station along the path of increasing boundary crossings culminating in sexual misconduct. Although clinicians often advance the claim that therapy is going on, so, inevitably, is much purely social behavior; it does not look like therapy, at least to a jury. Lunch sessions are not uncommonly followed by sessions during dinner, then just dinners, then other dating behavior, eventually including intercourse.

Sessions in cars represent another violation of place. Typically, the clinician gives the patient a ride home under various circumstances. Clinician and patient then park (e.g., in front of the patient's house) and finish up the presumably therapeutic conversation. From a fact finder's viewpoint, many exciting things happen in cars, but therapy is usually not one of them.

The complexity of the matter increases, however, when we consider other therapeutic ideologies. For example, it would not be a boundary violation for a behaviorist, under certain circumstances, to accompany a patient in a car, to an elevator, to an airplane, or even to a public restroom (in the treatment of paruresis, the fear of urinating in a public restroom) as part of the treatment plan for a particular phobia. The existence of a body of professional literature, a clinical rationale, and risk-benefit documentation will be useful in protecting the clinician in such a situation from misconstruction of the therapeutic efforts.


Money is a boundary in the sense of defining the business nature of the therapeutic relationship. This is not love, it's work. Indeed, some would argue that the fee received by the therapist is the only appropriate and allowable material gratification to be derived from clinical work (28). Patient and clinician may each have conflicts about this distinction (29), but consultative experience makes clear that trouble begins precisely when the therapist stops thinking of therapy as work.

On the other hand, most clinicians learned their trade by working with indigent patients and feel that some attempt should be made to pay back this debt by seeing some patients for free-a form of "tithing," if you will. Note that this decision-to see a patient for free and to discuss that with the patient-is quite different from simply letting the billing lapse or allowing the debt to mount. The latter examples are boundary crossings, perhaps violations.

Consultative experience also suggests that the usual problem underlying a patient's mounting debt is the clinician's conflict about money and its dynamic meanings. Initially reluctant to bring up the unpaid bill, the clinician may soon become too angry to discuss it. Explorations of the dynamic meaning of the bill are more convincing when they do not take place through clenched teeth. A clinician stuck at this countertransference point may simply let it slide. In the minds of fact finders, this raises a question: "The clinician seems curiously indifferent to making a living; could the patient be paying in some other currency?"-a line of speculation one does not wish to foster.

In rural areas even today, payments to physicians may take the form of barter: when the doctor delivers your child, you pay with two chickens and the new calf. For the dynamic therapist this practice poses some problems, because it blurs the boundary between payment and gift (covered in the next section). The clinician should take a case at a reasonable fee or make a decision to see the patient for a low fee (e.g., one dollar) or none. Barter is confusing and probably ill-advised today. Of course, all such decisions require documentation.


A client became very upset during an interview with her therapist and began to cry. The therapist, proffering a tissue, held out a hand-tooled Florentine leather case in which a pocket pack of tissues had been placed. After the patient had withdrawn a tissue, the therapist impulsively said, "Why don't you keep the case?" In subsequent supervision the therapist came to understand that this "gift" to the patient was an unconscious bribe designed to avert the anger that the therapist sensed just below the surface of the patient's sorrow.

This gift was also a boundary violation, placing unidentified obligations on the patient and constituting a form of impulsive acting in. A related boundary violation is the use of favors or services from the patient for the benefit of the therapist, as Simon's startling vignette illustrates:

Within a few months of starting . . . psychotherapy, the patient was returning the therapist's library books for him "as a favor." . . . The patient began having trouble paying her treatment bill, so she agreed--at the therapist's suggestion--to clean the therapist's office once a week in partial payment . . . . The patient also agreed to get the therapist's lunch at a nearby delicatessen before each session. (6, p. 106)

The obvious exploitive nature of these boundary violations destroys even the semblance of therapy for the patient's benefit.

When Freud heard that one of his patients was planning to buy a set of his complete works, he gave the patient the set as a gift (30). Immediately following the receipt of this gift, Freud's patient found that he was unable to use his dreams productively in the analysis as he had before. Freud related this "drying up" to the gift and noted, "You will see from this what difficulties gifts in analysis always make" (p. 42).

Other boundary crossings can be relatively minor but can promote a chain of subsequent crossings, as in this example:

A patient walked into the room while her therapist was pouring coffee from a carafe. He later described how he had felt socially incapable of not offering some coffee to the patient and had indeed offered some. At the next session, the patient brought doughnuts.

As the vignette shows, many boundary problems may arise at the interface between manners and technique. In contrast to the potentially harmful or at least confusing effects of the preceding examples, compare the practice (not uncommon among psychopharmacologists) of giving patients, as part of treatment, educational texts designed for laypersons (e.g., giving Moodswing [31] to a patient with bipolar disorder). Such a boundary crossing may foster mastery of the illness through information--a positive result. A similar point might be made for judicious "gifts" of medication samples for indigent patients. These two instances represent clear boundary crossings that have some justification. Ideally, even these should be discussed with an ear to any possible negative effects.

A patient in long-term therapy had struggled for years with apparent infertility and eventually, with great difficulty, arranged for adoption of a child. Two years later she unexpectedly conceived and finally gave birth. Her therapist, appreciating the power and meaning of this event, sent congratulatory flowers to the hospital.

In this case, the therapist followed social convention in a way that--though technically a boundary crossing--represented a response appropriate to the real relationship.

Offering a tissue to a crying patient and expressing condolences to a bereaved one are similar examples of appropriate responses outside the classic boundaries of the therapeutic relationship.


Clothing represents a social boundary the transgression of which is usually inappropriate to the therapeutic situation, yet a patient may appropriately be asked to roll up a sleeve to permit measurement of blood pressure . Excessively revealing or frankly seductive clothing worn by the therapist may represent a boundary violation with potentially harmful effects to patients, but the issue can also be overdone, as in the following case.

A patient in a western state, as part of a sexual misconduct allegation that a jury later found to be false, accused the therapist (among other things) of conducting therapy sessions with the top two buttons of his shirt undone. While such a phenomenon might conceivably represent a violation for a very sensitive patient, evidence was introduced that revealed the exaggerated nature of this claim in this case.

Berne (32) noted the technical error of the male clinician who, confronting a patient whose skirt was pulled up high, began to explain to the patient his sexual fantasies in response to this event. Berne suggested instead saying to the patient, "Pull your skirt down." Similar directness of limit setting appears to be suited to the patient who--either from psychosis or the wish to provoke--
begins to take off her clothes in the office. As before, the comment, "This behavior is inappropriate, and it isn't therapy; please put your clothes back on," said in a calm voice, is a reasonable response.


As part of the otherwise laudable efforts to humanize and demystify psychiatry a few decades back, the use of a patient's first name was very much in vogue. While this may indeed convey greater warmth and closeness, such usage is a two-edged sword. There is always the possibility that patients may experience the use of first names as misrepresenting the professional relationship as a social friendship (28). There may well be instances when using first names is appropriate, but therapists must carefully consider whether they are creating a false sense of intimacy that may subsequently backfire.

A middle-aged woman tried for more than a year to get her therapist to use her first name, but the requests were denied, and exploration of the issue took place instead. After some time the patient recovered memories of previously repressed material, in part because of increased trust in the therapist. The patient spontaneously related her trust to the use of last names as a boundary issue; boundaries had been badly blurred in her family and this had included sexual abuse.

There are distinct advantages to addressing the adult in the patient, in terms of fostering the adult observing ego for the alliance. Trainees often do not see the paradox of expecting adult behavior on the ward from someone they themselves call "Jimmy," which is what people called the patient when he was much younger. Last names also emphasize that this process is work or business, an atmosphere which may promote a valuable mature perspective and minimize acting out. In addition, calling someone by the name used by primary objects may foster transference perceptions of the therapist when they are not desirable, as with a borderline patient prone to forming severe psychotic transferences.

For balance, however, recall that use of last names may also sound excessively distant, formal, and aloof.

Tone is also a part of language. A patient won a settlement in an allegation of sexual misconduct when the tape recording she had made of a phone call from her therapist revealed his intimate, seductive tone. The therapist's attorney urged the settlement for fear that the jury would hear the intimate tone as evidence of a sexual relationship.

Word choice can also be violative, as when the therapist inquires, "What are you feeling now in your vagina?" Note that this inquiry might be proper in analytic therapy after appropriate preparation. Clinical utility aside, the way in which such explorations may violate boundaries should be kept in mind.

Finally, psychotherapy may be a forum for sadomasochistic enactments in which aggressive verbal abuse grows out of countertransference sadism. Cruel and contemptuous comments by the therapist may be rationalized as therapeutic confrontation.



Few clinicians would argue that the therapist's self-disclosure is always a boundary crossing. Psychoanalysis and intensive psychotherapy involve intense personal relationships. A useful therapeutic alliance may be forged by the therapist's willingness to acknowledge that a painful experience of the patient is familiar to himself (19). However, when a therapist begins to indulge in even mild forms of self-disclosure, it is an indication for careful self-scrutiny regarding the motivations for departure from the usual therapeutic stance.

Gorkin (33) observed that many therapists harbor a wish to be known by their patients as a "real person," especially as the termination of the therapy approaches. While it may be technically correct for a therapist to become more spontaneous at the end of the therapeutic process, therapists who become more self-disclosing as the therapy ends must be sure that their reasons for doing so are not related to their own unfulfilled needs in their private lives but, rather, are based on an objective assessment that increased focus on the real relationship is useful for the patient in the termination process. Self-disclosure, however, represents a complex issue.

Clearly, therapists may occasionally use a neutral example from their own lives to illustrate a point. Sharing the impact of a borderline patient's behavior on the therapist may also be useful. The therapist's self-revelation, however, of personal fantasies or dreams; of social, sexual, or financial details; of specific vacation plans; or of expected births or deaths in the family is usually burdening the patient with information, whereas it is the patient's fantasies that might best be explored.

The issue is somewhat controversial: a number of patients (and, surprisingly, some therapists) believe that the patient is somehow entitled to this kind of information. In any case, it is a boundary violation and as such may be used by the legal system to advance or support a claim of sexual misconduct. The reasoning is that the patient knows so much about the therapist's personal life that they must have been intimate compare the remark by a board of registration, quoted earlier).

Subtler variations on the information theme may occur, as when a therapist sees members of a couple in parallel treatment but separately alludes in one member's session to material from the other's. Sensitivity in this area may run quite high.

A patient had a dream involving Nazis. In the interpretation the therapist suggested that this detail referred to himself. The patient seemed doubtful. The therapist noted that the interpretation was based in part on the fact that other patients of his had dreamed of Nazis in response to the therapist's German last name. The patient's mood changed; only later was she able to tell the therapist how violated she had felt at his "intruding" other patients into the session.

Although the intrusion was at a verbal level only, the impact was clear for this patient, who had been subjected to some disregard of her boundaries in previous therapy.

Finally, the boundary can be violated from the other side. An example would be the therapist's using data from the therapy session for personal gain, such as insider information on stock trading, huge profits to be made in real estate, and the like.


To place the issue of physical contact in context, it should be noted that psychiatrists traditionally performed their own physical examinations. This practice has declined so markedly that a senior psychiatrist recently wrote about examining a patient's bruised leg as a major return to the past. Hospitals commonly use internists for this purpose. Psychiatric residents still do their own physical examinations but commonly maintain distance by examining each other's patients. Abnormal Involuntary Movement Scale examinations for tardive dyskinesia are often the only routine physical contact.

There is room here for regrets. Physicians working with a patient with AIDS or HIV seropositivity often describe wishing to touch the patient in some benign manner (pat the back, squeeze an arm, pat a hand) in every session. They reason that such patients feel like lepers, and therapeutic touch is called for in these cases.

But even such humane interventions must be scrutinized and, indeed, be documented to prevent their misconstruction in today's climate.

From the viewpoint of current risk-management principles, a handshake is about the limit of social physical contact at this time. Of course, a patient who attempts a hug in the last session after 7 years of intense, intensive, and successful therapy should probably not be hurled across the room. However, most hugs from patients should be discouraged in tactful, gentle ways by words, body language, positioning, and so forth. Patients who deliberately or provocatively throw their arms around the therapist despite repeated efforts at discouragement should be stopped. An appropriate response is to step back, catch both wrists in your hands, cross the patient's wrists in front of you, so that the crossed arms form a barrier between bodies, and say firmly, "Therapy is a talking relationship; please sit down so we can discuss your not doing this any more." If the work degenerates into grabbing, consider seriously termination and referral, perhaps to a therapist of a different gender.

What is one to make of the brands of therapy that include physical contact, such as Rolfing? Presumably, the boundary extends to that limited physical contact, and the patient expects it and grants consent; thus, no actual violation occurs. Massage therapists may struggle with similar issues, however. In other ideologies the issue may again be the impact of the appearance of a violation:

A therapist-who claimed that her school of practice involved hugging her female patient at the beginning and end of every session, without apparent harm-eventually had to terminate therapy with the patient for noncompliance with the therapeutic plan The enraged patient filed a sexual misconduct claim against the therapist. Despite the evidence showing that this claim was probably false (a specious suit triggered by rage at the therapist), the insurer settled because of the likelihood that a jury would not accept the principle of "hug at the start and hug at the end but no hugs in between." If the claim was indeed false, this is a settlement based on boundary violations alone.

At another level this vignette nicely suggests how nonsexual boundary violations may be harmful to a patient in much the same way that actual sexual misconduct is. Instead of engaging the patient in a mourning process to deal with the resentment and grief about the deprivations of her childhood, the therapist who hugs a patient is often attempting to provide the physical contact normally offered by a parent. The patient then feels entitled to more demonstrations of caring and assumes that if gratification in the form of hugs is available, other wishes will be granted as well (compare Smith's concept of the "golden fantasy" that all needs will be met by therapy [34]). When actual physical contact occurs, the crucial psychotherapeutic distinction between the symbolic and the concrete is lost (21 ), and the patient may feel that powerful infantile longings within will finally be satisfied.


Boundary crossings may be benign or harmful, may take many forms, and may pose problems related to both treatment and potential liability. The differences in impact may depend on whether clinical judgment has been used to make the decision, whether adequate discussion and exploration have taken place, and whether documentation adequately records the details.

The complexity of the subject and the variability of results from case-by-case analysis merit empirical study. Educational materials are available through the Office of Public Affairs of the American Psychiatric Association. Heightened awareness of the concepts of boundaries, boundary crossings, and boundary violations will both improve patient care and contribute to effective risk management.

In an effort to prevent more serious boundary violations of a sexual nature, Epstein and Simon (28) have developed an exploitation index which comprises a list of questions that therapists can ask themselves about their current behavior with patients. In this manner these authors have attempted to provide an ongoing self-monitoring system. While such approaches may be useful for some clinicians, we must acknowledge that considerable personal variation exists in our field. The relationships between therapist and patient vary from one therapist to another, and there are even variations across patients in the practice of one therapist.

As Lipton (8) observed, it is ultimately impossible to codify or prescribe a personal relationship between therapist and patient in a precise manner.

Perhaps the best risk management involves careful consideration of any departures from one's usual practice accompanied by careful documentation of the reasons for the departure.

Finally, the value of consultation with a respected colleague should be a built-in part of every practitioner's risk-management program.


1. Ingram DH: Intimacy in the psychoanalytic relationship: a preliminary sketch. Am J Psychoanal 1991; 51:403-411

2. Gutheil TG: Borderline personality disorder, boundary violations, and patient-therapist sex: medicolegal pitfalls. Am J Psychiatry 1989; 146:597-602

3. Stone MH: Boundary violations between therapist and patient. Psychiatr Annals I 976; 6:670-677

4. Gabbard GO (ed): Sexual Exploitation in Professional Relationships. Washington, DC, American Psychiatric Press, 1989

5. Borys DS, Pope KS: Dual relationships between therapist and client: a national study of psychologists, psychiatrists, and social workers. Professional Psychology: Research and Practice 1989; 20:
280- 293

6. Simon RI: Sexual exploitation of patients: how it begins before it happens. Psychiatr Annals 1989; 19:104-122

7. Jorgenson L: Consultation forum. Psychodynamic Letter 1991;

8. Lipton SD: The advantages of Freud's technique as shown in his analysis of the Rat Man. Int J Psychoanal 1977; S8:255-273

9. Grosskurth P: Melanie Klein: Her World and Her Work. New York, Alfred A Knopf, 1986

10. Winnicort DW: Hate in the counter-transference. Int J Psychoanal 1949; 30:69-74

11. Little MI: Psychotic Anxieties and Containment: A Personal Record of an Analysis With Winnicott. Northvale, NJ, Jason Aronson, 1990

12. Eissler KR: The effect of the structure of the ego on psychoanalyric technique. J Am Psychoanal Assoc 1953; 1:104-143

13. Zetzel ER: 1965: Additional notes upon a case of obsessional neurosis: Freud 1909. Int J Psychoanal 1966; 47:123-129

14. Shapiro T: On neutrality. J Am Psychoanal Assoc 1984; 32:269- 282

15. Gabbard GO: Psychodynamic Psychiatry in Clinical Practice. Washington, DC, American Psychiatric Press, 1990

16. Langs R: The Bipersonal Field. New York,Jason Aronson, 1976 1 7. Spruiell V: The rules and frames of the psychoanalytic situation. Psychoanal Q 1983; 52:1-33

18. Appelbaum PS, Gutheil TG: Clinical Handbook of Psychiatry and the Law, 2nd ed. Baltimore, Williams & Wilkins, 1991

19. Viederman M: The real person of the analyst and his role in the process of psychoanalytic cure. J Am Psychoanal Assoc 1991; 39:451-489

20. Novey R: The abstinence of the psychoanalyst. Bull Menninger Clin 1991; 55:344-362

21. Casement PJ: The meeting of needs in psychoanalysis. Psychoanal Inquiry 1990; 10:325-346

22. Greenson RR: The Technique of Psychoanalysis, vol 1 . New York, International Universities Press, 1967

23. Langs R: Psychotherapy: A Basic Text. New York, Jason Aronson, 1982

24. Adler G: Borderline Psychopathology and Its Treatment. New York, Jason Aronson, I 985

25. Adler G, Buie DH Jr: Aloneness and borderline psychopathology:
the possible relevance of child development issues. Int J Psychoanal 1979; 60:83-96

26. Kernberg OF: Severe Personality Disorders: Psychotherapeutic Strategies. New Haven, Conn, Yale University Press, 1984

27. Kernberg OF, SeIzer MA, Koenigsberg HW, Carr AC, Appelbaum HA: Psychodynamic Psychotherapy of Borderline Patients. New York, Basic Books, 1989

28. Epstein RS, Simon RI: The exploitation index: an early warning indicator of boundary violations in psychotherapy. Bull Menninger Clin 1990; 54:450-465

29. Krueger DW (ed): The Last Taboo: Money As Symbol and Reality in Psychotherapy and Psychoanalysis. New York, Brunner/ Mazel, 1986

30. Blanton 5: Diary of My Analysis With Sigmund Freud. New York, Hawthorn Books, 1971

31. Fieve R: Moodswing: The Third Revolution in Psychiatry. New York, Bantam Books, 1976

32. Berne E: What Do You Say After You Say Hello? The Psychology of Human Destiny. New York, Grove Press, 1972

33. Gorkin M: The Uses of Countertransference. Northvale, NJ, Jason Aronson, 1987

34. Smith 5: The golden fantasy: a regressive reaction to separation anxiety. IntJ Psychoanal 1977; 58:311-324

NOTE: THOSE WHO WOULD LIKE A REPRINT MAY SEND A REQUEST TO: Thomas G. Gutheil, Professor of Psychiatry, Harvard Medical School, 6 Wellman St., Brookline MA 02446; email: <>.

Copyright (c) 1993 American Psychiatric Association.


[Back to Top]