Clinicians’ Perceptions of Boundaries in Brazil and the United States
http://www.100md.com
《美国精神病学与法律学报》
Although there has been considerable discussion of boundary excursions in clinician-patient relationships, little empirical research exists. This pilot study adds to the existing theoretical discussion by comparing perceptions by mental health professionals in Brazil and the United States of what may constitute boundary violations. Participants rated each possible boundary violation according to its degree of harm and professional unacceptability. Three distinct groupings of boundary violations were found: (1) core, consisting of the most serious violations; (2) separation of therapist and client lives, involving encounters between therapists and clients outside of therapy; and (3) disclosure and greeting behavior, involving disclosure of information about the therapist and greeting behavior. The U.S. and Brazilian perceptions were found to be surprisingly similar, with only a few differences.
The issue of therapeutic boundary excursions, which includes everything from less serious crossings to more serious violations, has been extensively discussed since the 1990s in several publications.1–10 Strikingly, none of this recent literature involves an empirical examination of boundary violations, although some older literature has examined the prevalence of sexual infringements11–13 or harm to patients from such actions.14,15 Given the amount of theorizing about boundaries, more empirical work on the topic seems long overdue. Although several questions about boundaries and their violation might be asked, three will be focused on in this article: (1) How do clinicians classify different kinds of boundary excursions? (2) How might the cultural background of the therapist affect his or her perception of the seriousness of different excursions? and (3) In rating the seriousness of boundary excursions, do clinicians differentiate between the professional acceptability of a behavior and the harm a behavior might cause a client?
A boundary, according to the behavior-developmental view proposed herein,16,17 is a way of discriminating among environmental events and reacting in a way that may or may not lead to action. Actions taken in the context of some kinds of environmental events are seen as socially, legally, professionally, clinically, and morally acceptable. Such actions are also viewed as leading to improvement in the patient’s behavior, or at least as causing no harm. Actions in the presence of other kinds of situations, however, are not regarded as acceptable and may also be thought to cause harm. Because decisions on what actions to take in particular situations are often necessary, potential boundaries abound. Within the therapeutic context, however, some types of discriminations are clearly of greater interest than are others. Because this definition supports consideration of actions in context, it allows potentially different standards regarding what constitutes a boundary violation in different therapeutic and cultural contexts.
In much of the previous empirical work in this area, researchers have examined a limited set of possible boundary violations, particularly sexual violations. Empirically examining a large number of possible boundary transgressions may elucidate theoretical arguments about whether there are different kinds of boundary excursions. In some of the recent discussion about boundaries, for example, some boundary excursions have been said to represent more serious "violations" and others less serious "crossings."4–6,18,19 Can such statements by a few individual clinicians be confirmed by empirical work on a larger sample of clinicians? It is also possible that other useful ways of grouping boundary excursions will be found that may be helpful in thinking about boundaries in general.
Cultural Influences on Boundaries
Traditional notions of therapeutic boundaries in clinical practice have been said to ignore the broader social context, particularly culture (for example, Kroll6). Because culture, to a large extent, defines what is acceptable behavior, we agree with Kroll that boundaries should be seen in the context of larger groups, organizations, and societies (see Gutheil and Gabbard3,4 for additional discussion). While some kinds of behavior may be universally proscribed, others may differ in acceptability, depending on cultural norms.
Notions of psychological boundaries are most often considered with respect to intimacy. When and with whom would intimacies be allowed? Different cultures have widely varying notions of appropriateness20 on a large number of questions, including who one is allowed to talk to, visit with, or eat with; when and with whom various kinds of touching are appropriate; who can share property (and under what conditions); who is allowed to interact with children, or with women, and in what ways; with whom one is allowed to have sex, to name a few.
Brazilian and U.S. cultures, the two cultures from which the clinicians we studied originated, differ in a wide variety of ways. For example, Brazilian and other South American cultures are much closer to Mediterranean ones.21–23 In those cultures, physical contact in greeting and other interactions is normative. Standard greetings between people in Brazil include shaking hands, shaking hands with an arm clasp, kissing on both cheeks, and putting arms around people and hugging them. These greetings take place between people, even if they are just recent acquaintances. Touching, in general, is more frequent in such cultures than in Northern, Euro-American cultures. In the American culture, including in Boston, shaking hands on greeting is normative, but in some traditions even shaking hands might be avoided. Other kinds of touching during greeting and other types of physical interactions are usually reserved for people with whom one has a great deal of intimacy (family, close friends, romantic partners).
The two populations also differ with regard to payment for clinical services and gift-giving. Most people in Brazil do not have health insurance. If they have the money, they may pay out of pocket for health care services.24 Otherwise, state provision of services is common for those with little income. Patients may nevertheless feel obligated to compensate their therapists in some way, which can result in extensive gift-giving (Martins de Almeida K, personal communication, August 2001). Giving little gifts during visits is a valued part of this culture in most interpersonal relationships.
The economic realities of the two populations also differ in ways that may affect boundaries. The rates of poverty in Brazil are much higher than in the United States. For example, the top fifth of the population in Brazil receives 61.7 percent of the economy’s total income, and the bottom fifth receives only 3.3 percent.25 The U.S. Agency for International Development (USAID) reported in 2000 that about 20 percent of the Brazilian population lived below the poverty line (35 million/174.5 million total population).26 Because of the higher rates of poverty, the relatively low wages of mental health workers in Brazil, and the relatively high cost of drugs, "talk" therapy may be relatively more common than drug treatment compared with such treatments in the United States. One possible result of the frequency of exposure of patients to therapists in Brazil is that there is more opportunity for boundary violations to occur.
There are most likely many other differences between the two cultures as far as mental health practice goes. At this point, there is no single text or article that takes a wide variety of behavior in these two cultures and discusses the differences between them. Therefore, this discussion has had to rely on the relatively few studies that were cited earlier in the article.20–26 These studies suggest that as far as the general cultures are concerned, one would expect there to be differences in the behavior of Brazilians and North Americans that could have an impact on clinicians’ perceptions of what is appropriate within therapy.
It is important to add that another possible influence on clinicians’ behavior may be how those in the two cultures are trained and socialized. Clinicians in the Brazilian sample are heavily influenced by both European and American traditions, and particularly by psychoanalytic theories and approaches (Martins de Almeida K, personal communication, August 2001). The fact that training of clinicians in the two countries is similar may decrease cultural differences in the area of professional behavior.
Methodology
The current study empirically examines perceptions of boundary excursions in two groups of clinicians: one in the United States and one in Brazil. There have been a few empirical ratings of professionals’ perceptions of some number of hypothesized boundary crossings or violations,13 but these have not explicitly addressed the distinction between crossings and violations, as described by Gutheil and Gabbard.3,4 The present study goes beyond any earlier empirical rating studies, because there are few if any empirical data on cross-cultural differences in perceptions of boundary crossings and violations, despite arguments about the importance of culture in determining such perceptions.
The instrument used in our study asked professionals to rate the various examples on what we perceive as two different dimensions of boundary violations: (1) Was this practice judged to be professionally unacceptable, and (2) was this practice judged to be harmful to patients (both were rated on a seven-point scale). The distinction between professional unacceptability and harm is an important one. References to lack of professionalism are more abstract judgments about whether something is a norm for what constitutes professionally sound practice. That is, does a practice meet the standard of care? Judgments about harm may be based more on a practitioner’s experience with cases that are observed or heard about that resulted in actual harm to patients. Whether harm is experienced or not, therefore, seems to be more of an empirical matter. It is also possible that judgments of harm are more likely for more serious boundary violations, whereas lack of professionalism could relate to a wider range of such excursions. Of course, clinicians in the two cultures may perceive professional unacceptability and harm either in the same way or differently.
We specifically chose two samples that were relatively homogeneous within the cultures in which they existed. The clinicians sampled also had a relatively high degree of knowledge about therapeutic practice and possibly about boundary excursions in general. Our expectation is that these samples could then serve as comparison points for a similar survey to be administered to a much broader group of clinicians.
Participants
There were 60 participants: 28 Brazilians (20 women, 7 men, and 1 person who did not report gender) from the Institute of Psychiatry, Federal University of Rio de Janeiro; and 33 U.S. residents (18 women and 15 men) from the Program in Psychiatry and the Law, Department of Psychiatry, Harvard University, Massachusetts Mental Health Center, Boston. The U.S. sample was composed of attendees at and nonattending members of the Program in Psychiatry and the Law. The Program consists of a weekly meeting, where topics of all kinds relating to the practice of psychotherapy and to forensic psychiatry (including boundaries) are often discussed. The program conducts research on many of these subjects, and information is also exchanged on-line. Participants in Brazil were attendees at a regular weekly meeting of staff and members of the Institute of Psychiatry. During these weekly meetings participants either present and discuss cases or hear talks on a variety of professional topics. Staff members, postgraduate students, and other clinicians attend and participate in the meetings. During one particular meeting of the Institute, a talk stressing the ubiquity of cultural differences in boundaries was given by two U.S. investigators. This description of the meetings of the two groups suggests that both cultural groups interacted in discussion and theorizing about boundaries. It also suggests that within their own nations, both groups had a training, supervision, and information-sharing protocol for the participants. Both programs took place in major, prestigious universities in public hospitals. The exposure of the U.S. group to the subject matter of boundaries would have been greater than the exposure of the Brazilian group, according to feedback from that latter group. Further, the talk should have encouraged the Brazilian participants to report cultural differences if they existed, because it contained a strong message of acceptance of these differences.
All the participants from both countries were mental health professionals, although, as shown in Table 1, the distribution of types of degrees was somewhat different. Most respondents in the United States were psychiatrists, with a few having doctoral and/or master’s degrees. In Brazil, the sample was nearly evenly split between psychiatrists and pre-bachelor’s degree clinicians, who apparently do a great deal of the day-to-day work with patients. Pre-bachelor’s level clinicians in Brazil had obtained a two-year, post-high-school specialization degree in mental health practice. The one such individual in the United States was an undergraduate student who was attending the meetings due to an interest in forensic psychiatry. These differences partly reflect how mental health services are delivered in each country, as well as the membership of the Program in Psychiatry and the Law in Boston compared with the Institute of Psychiatry in Rio de Janeiro. While clearly there are differences in the distribution of different types of mental health workers in the two cultures, with the U.S. sample containing more MDs and PhDs, it was considered to be most important to have each sample reflect what was generally true in that context. In the context of the Institute of Psychiatry in Rio, the individuals in the sample realistically reflected those who work and deliver services to clients there. Trying to locate additional psychiatrists or psychologists who had credentials closer to those found in the U.S. would have resulted in a sample that did not realistically reflect this Brazilian context.
The study procedures, since they involved asking participants to fill out a questionnaire and participation was voluntary and anonymous, were judged to be exempt from review by the Institutional Review Board at the Massachusetts Mental Health Center, where the Program in Psychiatry and the Law was located at the time this study was conducted.
Instrument
An 87-item questionnaire was devised by consulting the literature on boundaries, from a large number of actual case examples, and from professional experiences shared within the Program in Psychiatry and the Law. The specific boundary excursions asked about are shown in the Appendix. Participants were asked to rate (on a six-point scale) the percentage of cases in which it would be harmful to the patient if a colleague behaved in the manner described, and concurrently, to rate in what percentage of cases this same behavior was professionally unacceptable. The rating scale asked participants to circle one of the following percentages: 0%, 2%, 16%, 50%, 84%, 98%, and 100%. This scale reflects a linear z-score scale that was converted into percentages, using the cumulative normal distribution function, with a probit transformation. This scale was used because (1) it represents the postulated distribution for most psychological characteristics, and (2) the differences at the ends of the scale are more important than the ones in the middle. The technique of asking about colleagues’ behavior, rather than the clinician’s own, was used to minimize defensive reactions that might occur if we asked individuals to rate how professionally unacceptable a behavior might be if engaged in by themselves.
All the questions and instructions were in English for the U.S. sample and in Portuguese for the Brazilian sample. The translation into Portuguese was the joint effort of two individuals. One was a born English speaker, a trained developmental psychologist in the English language, who had grown up in Brazil and speaks Portuguese fluently. The other was a born Portuguese speaker, a clinician trained in a Portuguese-speaking environment, who had worked as a clinician for several years in the United States with English- and Portuguese-speaking clients and colleagues. The native Portuguese speaker first translated the questionnaire into Portuguese. The native English speaker then compared the Portuguese phrases to the English version of the questionnaire, essentially back-translating it. In a few cases, the two speakers had to have further discussions about particular ideas that did not translate easily from one language to the other.
Procedure
Both groups were given the questionnaire. Almost all participants filled it out during a meeting of their respective professional groups. A few others completed it at another time and returned it to the investigators.
Results
Harm Versus Professional Acceptability
The questionnaire asked both about harm to patients and about the professional unacceptability of ever engaging in certain actions. The first step in the analysis, therefore, was to examine the extent to which ratings on these two types of questions were similar to each other. If responses to each were very different, then the analysis of the remainder of the results might have to consider each kind of rating separately.
The degree of similarity of the two ratings for each item, as measured by the correlation between harm and unacceptability,27 was r(86) = 0.769, p < .0005; the effect size28 for a correlation of this magnitude is 0.59, a large effect size. This suggests that in these respondents’ ratings, harm to a patient and lack of professionalism were seen as being highly related to each other. (Note that this correlation was determined in a three-step process. First, the correlation between the two ratings was determined for each of the 87 items (questions). Second, these 87 correlation coefficients were transformed using Fisher’s z transformation of r. Third, the two overall mean zr’s and the standard deviation of the zr’s (one for harm and one for lack of professionalism) were calculated. The correlation of these two means was then calculated, and it is this result that is reported here.)
Because the ratings on harm and on professional unacceptability had such a high degree of similarity, we could have decided to analyze one and not the other. There were two reasons to keep information from both kinds of ratings in the study. The empirical reason is that because they are not perfectly correlated, some variability in responses to each type of rating would be expected that then could be explored. Second, since information on both was collected as part of this study, we decided to keep both in future analyses.
Three Different Types of Boundary Excursions
With 87 questions and two ratings of each question (174 items in all), a natural question is whether we could group together questions that are related in some way. A useful technique for seeing whether questions might be related to one another is factor analysis. If questions were found to be statistically related, there would be a basis for grouping them together. Such groupings could help to distinguish different types of boundary excursions. Groupings of items would also make it easier to look at cultural differences, because fewer comparisons would have to be made.
An exploratory principal components factor analysis reduced 139 of the 174 responses to three distinguishable factors. The remaining 35 items did not relate strongly to any of these three factors. Each of these factors was named by the type of items that were most strongly represented within them. Both ratings of harm to the patient and ratings of the lack of professional acceptability were included.
The first factor, called core boundary violations, contained 85 items that generally constituted the most serious boundary violations (to simplify the presentation of the data, the specific items are not shown. More detailed data as well as copies of the questionnaires can be obtained from the authors). These violations were made up of several different types of therapist behavior. For example, 32.65 percent involved mixing therapy with personal or social considerations (e.g., telephoning the client to speak about matters besides therapy). The next largest number of items from this first factor (28.57 percent), involved sexual behavior (e.g., necking with client, or pretending sex is therapy). The other two types of violations that loaded with high coefficients on this factor were those involving financial matters, consisting of 18.37 percent of the items (e.g., selling non-therapy-related products to a client) and those involving actual physical or hostile aggression toward clients, 14.28 percent of the items (hitting the client). Both Americans and Brazilians indicated that Factor 1 items (core boundary violations) were the most serious, as indicated by an overall mean rating of those items of 6.04 ± 0.84 (equivalent to a rating of 98%), on a scale on which a mean rating of 7 would have indicated the most serious violation.
One additional characteristic of the items that loaded on Factor 1 is that 54 percent of these items, a slight majority, involved ratings that showed a perception that the item led to harm to patients, with 46 percent referring to lack of professional acceptability.
For the second factor, separation of therapist’s and client’s lives, the questionnaire contained 34 items that asked about situations in which the client and the therapist might encounter each other outside of the therapist’s office. These items differed from the items in Factor 1 that involved mixing therapy with personal and social matters, in that Factor 2 items involved a more distant mixing of therapy with other aspects. Only three (divulging one’s marital status or history of drug use, and offering refreshments during therapy) actually occurred in therapy. Some involved attending ceremonies involving the client (such as a wedding or funeral); others involved more inadvertent co-occurrences (such as attending an event at which the client was also present). Factor 2 items (separation of therapist’s and client’s lives) were given an overall mean rating that was just slightly above neutral (3.65 ± 0.95; SD), equivalent to a percentage halfway between 16% and 50%, suggesting that clinicians did not see these violations as being as serious as the core boundary violations. Slightly less than half of these items referred to harm to patients (47%), and slightly more than half (53%) to lack of professional acceptability.
Twenty items from the questionnaire were included in the third factor, disclosure and greeting items. In addition to asking about greeting behavior, such as shaking hands, kissing on the cheek, or hugging, these items asked about such behavior as displaying diplomas or awards or revealing one’s credentials. The disclosure and greeting items were given an overall rating between Factor 1, core boundary violations, and Factor 2, separation of therapist’s and client’s lives (4.35 ± 0.75; SD)—halfway between 50% and 84%. The preponderance of items that loaded on this factor (65%) referred to lack of professional acceptability, with only 35 percent referring to harm to patients.
For comparison purposes, we also calculated the overall rating of all items that did not load on the first three factors. These items received a rating that was just slightly below midpoint (mean, 3.89 ± 0.87; SD).
Cultural Differences
How did the U.S. and Brazilian clinicians differ on their ratings of harm and lack of professional acceptability of these items? To compare the two samples, an analysis of variance was performed on the overall item ratings for each of the three factors. Culture made a small but significant difference overall in the factor scores, with the Brazilians rating items in general more harmful and more unacceptable (F(1,166) = 5.708, p = .018, = .033). We also tested the effects of two other variables, gender and degree of training (doctoral level versus below that level). Only the items that loaded on Factor 2 (therapist-client separation) showed differences for these two variables. We found that the women tended to rate Factor 2 items somewhat higher in general than did the men (F(1,56) = 4.26, p < .044, = .071), a very small effect size. Nondoctoral level practitioners tended to rate Factor 2 items somewhat higher (F(1,54) = 6.23, p < .016, = .103), a small effect size.
Very few of the ratings of individual items showed significant cultural differences. The items included in the first factor showed almost complete uniformity in the two cultures. There were just two items that had significant differences and one item that was at the level of a statistical trend. (As suggested by Rosnow and Rosenthal,28 the raw p < .05 level was adjusted for the 139 comparisons by dividing .05 by 139. This Bonferroni-corrected value was p < .0036.) In all three cases, the differences were negative, meaning that the U.S. clinicians found these behaviors to be less harmful than did the Brazilians: Item 85a: "Percentage of cases where therapist’s calling client after office hours to talk about treatment is harmful to client" (Mean difference = –2.19, t(55) = –4.57, p < .0005); and Item 56a: "Percentage of cases where therapist’s sitting with a client in a café/cafeteria which both frequent is harmful to client, if done by colleagues" (Mean difference = –1.44, t(55) = –3.177, p < .002). The one item that was significant at the level of a statistical trend (mean difference = –2.94; t(55) = –2.84, p < .006) was "Therapist’s revealing his/her sexual orientation to client is harmful."
Within the second factor, there were only four significant cultural differences. In three cases, as with the items in Factor 1, Brazilian clinicians saw the harm or the lack of professional acceptability as higher for placing diplomas in one’s office and telling one’s own history of drug use. They saw as significantly less unacceptable attending a small event outside of therapy that was also attended by the client.
The items in Factor 3 showed a somewhat different pattern, with a larger number of them (9/20 or almost half) showing significant cultural differences. These differences were found in two different areas of therapist-client interactions. On the one hand, Brazilian therapists tended to rate behavior such as displaying diplomas, professional awards, and/or revealing one’s credentials or emotions, as more harmful and professionally unacceptable than did American therapists. On the other hand, they rated two greeting behaviors, shaking hands and kissing on the cheek (a very common way of greeting in Brazil), as less harmful and less professionally unacceptable.
Surprisingly few cultural differences appeared out of the possible 87 times 2 (174) items. By chance alone, one might expect 9 items to differ significantly at the .05 corrected level of chance; 15 were actually found to be different. While this is not evidence in favor of a strong cultural difference, we examined whether there were cultural differences in particular subsets of variables when these variables were combined in a different way. Some of the items involved: sexual violations, attempted business transactions between therapist and patient, how much information a therapist should disclose to a patient, situations in which therapist and patient may mix together outside the office, and routine touching such as shaking hands. We created five new variables based on these categories and then examined each category for whether there were cultural differences. The results are shown in Table 2.
Brazilian clinicians were likely to rate various self-disclosure items as significantly more harmful and professionally unacceptable than did U.S. clinicians, whereas they were likely to rate routine touching that takes place as part of greeting or comforting as significantly less harmful and professionally unacceptable.
Discussion
This study is the first empirical examination of a large range of boundary violations, beyond just sexual violations, and examines the perceptions of clinicians from two cultures. The results add several pieces of information to the general discussion of boundaries that has been taking place in the literature cited in the introduction. First, the results of the factor analysis suggest that there are three types of boundary violations, as represented by the three factors that were found. These types are distinguished partly by the fact that the clinicians surveyed rated some types of violations as being more serious and other types as less serious. The factors also grouped together items of different types.
The first type of boundary violation (the first factor in the factor analysis) we called core boundary violations. These are the most serious types of therapist misbehavior that can occur. They include not only the expected sexual violations, but also violations involving the mixing of therapy with personal or social considerations, violations involving financial matters (e.g., selling non-therapy-related products to a client) and those involving actual physical or hostile aggression toward clients. The boundary violations that made up this factor were ones that were particularly intrusive. Both U.S. and Brazilian clinicians rated the 87 violations that loaded on this factor as being very harmful to clients and as being highly unprofessional.
The second type of boundary excursion, separation of therapist’s and client’s lives, was covered by 34 items on the questionnaire that concerned situations in which the client and the therapist might encounter each other outside the therapist’s office. The items in this factor were rated as close to the neutral point by both U.S. and Brazilian clinicians. In other words, they were seen as less harmful and less professionally unacceptable than were the core boundary violations.
The third type of boundary excursion on the questionnaire consisted of disclosure and greeting behavior. Items from this factor were rated just above the neutral point, so they were seen as slightly less neutral than Factor 2 items.
The usefulness of these results is twofold. First, professionals classified some kinds of excursions as more serious than others. Although the basis for this classification needs further elucidation, these results at least give the field a preliminary empirical rubric with which to evaluate professionals’ behavior. They provide, in addition, some empirical validation of the notion of boundary crossings versus boundary violations. Second, by simply studying such a broad range of possible violations, we can educate professionals about the spectrum therapeutic behavior that may be similar in many ways to the few well-recognized types of boundary violations, such as sexual ones.
A second question addressed by this study was whether differences between cultures would be apparent. In fact, with these samples, we found surprising uniformity across the two cultures. Very few of the ratings of individual items, for example, showed significant cultural differences. This was particularly true of the core boundary violations (only 2/87 of these items were significantly different). The other two factors of boundary excursions studied showed somewhat more differences in the ratings by Brazilian versus U.S. clinicians. For example, on the one hand, Brazilian clinicians tended to rate aspects of self-disclosure as more professionally unacceptable and harmful. On the other hand, they tended to see certain routine greeting behavior that is more common in Brazil as more acceptable and less harmful. These findings were further confirmed by a related analysis in which we collapsed together items according to their content (e.g., sexual behavior, other touching, business- or money-related disclosure).
When there were cultural differences in ratings of items, the Brazilian therapists tended almost always to rate those items as more harmful and/or more professionally unacceptable than did U.S. therapists. It may be (although we have no way of confirming this notion at present) that when individuals are less experienced in an area they tend to perceive rules as being less flexible. They do not yet have enough exemplars to see that following a rule too strictly is not always of benefit. The only exception to the tendency of the Brazilian therapists to rate all the boundary excursions as being more harmful and more unprofessional, was their rating of routine touching, such as shaking hands or kissing on the cheek, that may take place as part of greetings, which they rated as somewhat less harmful and less unacceptable. In any case, understanding the bases for such differences requires further empirical work, both to confirm that these differences are generally cultural and to understand the culturally-based interaction patterns that may underlie them.
Why were there not more cultural differences in beliefs? One reason may be that therapists in both samples were trained based on similar models. Training of mental health personnel in these cases seemed to emphasize to a greater or lesser extent the psychodynamic viewpoint and its associated views of the proper role for a therapist. While only a minority of the sample in Rio de Janeiro were psychiatrists, the other therapists worked alongside those psychiatrists and were most likely imbued with much of the same perspective. Perhaps professional training of this kind comes to supercede cultural differences in interactions that might be seen in other, more informal contexts. We also cannot eliminate the idea that a social-desirability bias may have been operating, and perhaps more so for the Brazilian therapists. The majority would have been well-enough educated in what is supposed to be appropriate conduct from the point of view of U.S. or European therapies, and may have answered as they thought they should answer. To answer otherwise might shed a negative light on therapeutic practice in Brazil. Finally, it is also possible that Brazilian and U.S. health professional cultures are just not that different, particularly when one examines professionally circumscribed behavior, except for a few small and relatively superficial aspects.
This is not the same as saying that there would not be any differences in the practice of therapy in the two cultures, as far as boundary violations are concerned. It is possible that in behavior, one would see greater differences. Anecdotal reports from some of the participants and from other Brazilian therapists (Martins de Almeida K, personal communication, August 2001) mentioned that there was little regulatory action surrounding boundary violations in Brazil. There were no lawsuits, no firings, no licensure or ethics complaints that these sources were familiar with. These same sources also noted that it was their perception that there might be a greater prevalence of major boundary violations in Brazil than in the United States.
A third finding is the surprising uniformity of ratings of harm and of lack of professional acceptability. We had predicted that the frequency of judgments of harm might be higher for more serious boundary excursions. Ratings of harm and lack of professional acceptability correlated relatively highly. We argued that because the correlation was not perfect, we might still see evidence of some difference between the two. In fact, it was found that in Factor 1, which consisted of the core boundary violations, the majority of items included referred to the harm to clients that a behavior might cause. In the other two factors, particularly Factor 3, more of the items referred to lack of professional acceptability rather than to harm. This is some evidence that the clinicians surveyed think of these two categories in slightly different ways. Some behaviors, in and of themselves, may be seen as lacking professional acceptability but as not being all that harmful to clients.
The samples used in this preliminary study were small and not very diverse. Because both groups, however, came from prestigious training institutions in major cities of their respective countries, they represent a kind of establishment view of boundary violations. As such, the results represent a reasonable place to begin to improve our understanding of boundary excursions and in fact supply a comparison point for further study. The uniformity of opinion, even across cultures, suggests that, despite the relatively small samples, the findings merit consideration.
In addition to studying professionals’ perceptions of boundaries within larger and more diverse samples (particularly in terms of therapeutic orientation), another important direction for such research is to try to document in a more complete fashion: (1) specifically how often the different kinds of excursions take place within therapy, and (2) what actual harms stemming from different kinds of excursions might exist.
Appendix: Boundary Excursions Asked About in the Questionnaire, in Order of Their Appearance
References
Duckworth K, Kahn M, Gutheil TG: Roles, quandaries and remedies: teaching professional boundaries to medical students. Harv Rev Psychiatry 1:266–70, 1994
Garfinkel PE, Dorian B, Sadavoy J, et al: Boundary violations and departments of psychiatry. Can J Psychiatry 42:764–70, 1997
Gutheil TG, Gabbard GO: The concept of boundaries in clinical practice: theoretical and risk management dimensions. Am J Psychiatry 150:188–96, 1993
Gutheil TG, Gabbard GO: Misuses and misunderstandings of boundary theory in clinical and regulatory settings. Am J Psychiatry 155:409–14, 1998
Gutheil TG, Simon RI: Non-sexual boundary crossings and boundary violations: the ethical dimension. Psychiatr Clin North Am 25:585–92, 2002
Kroll J: Boundary violations: a culture-bound syndrome. J Am Acad Psychiatry Law 29:274–83, 2001
Norris DM, Gutheil TG, Strasburger LH: "This couldn’t happen to me": boundary problems and sexual misconduct in the psychotherapy relationship. Psychiatr Serv 54:517–22, 2003
Simon RI: Sexual exploitation of patients: how it begins before it happens. Psychiatr Ann 6:670–7, 1989
Simon RI: Psychological injury caused by boundary violation: precursors to therapist-patient sex. Psychiatr Ann 2:614–19, 1991
Strasburger LS, Jorgensen L, Sutherland, P: The prevention of psychotherapist sexual misconduct: avoiding the slippery slope. Am J Psychother 46:544–55, 1992
Gartrell N, Herman J, Olarte S, Feldstein M, Localio R: Psychiatrist-patient sexual contact: results of a national survey, I: prevalence. Am J Psychiatry 143:1126–31, 1986
Kardener SH, Fuller M, Mensh IN: A survey of physicians’ attitudes and practices regarding erotic and nonerotic contact with patients. Am J Psychiatry 130:1077–81, 1973
Epstein RS, Simon RI, Kay GG: Assessing boundary violations in psychotherapy: survey results with The Exploitation Index. Bull Menninger Clin 56:150–66, 1992
Butler S, Zellen S: Sexual intimacies between therapists and patients. Prof Psychol 14:139–45, 1977
Pope KS, Tabachnick BG: Therapists as patients: a national survey of psychologists’ experiences, problems, and beliefs. Prof Psychol 25:247–58, 1991
Commons ML, Miller PM: A quantitative behavior-analytic theory of development. Mexican J Exp Anal Behav 24:153–80, 1998
Commons ML, Miller PM: A quantitative behavioral model of developmental stage based upon hierarchical complexity theory. Behav Analyst Today 2:222–40, 2001
Martinez R: A model for boundary dilemmas: ethical decision-making in the patient-professional relationship. Ethic Hum Sci Serv 2:43–61, 2000
Lazarus AA: How certain boundaries and ethics diminish therapeutic effectiveness. Ethics Behav 4:255–61, 1994
Hall ET: The Silent Language. Garden City, NY: Doubleday, 1959
Sussman NM, Rosenfeld HM: Influence of culture, language and sex on conversational distance. J Pers Soc Psychol 42:66–74, 1982
Regan PC, Jerry D, Narvaez M, Johnson D: Public displays of affection among Asian and Latino heterosexual couples. Psychol Rep 84:1201–02, 1999
Remland MS, Jones TS, Brinkman H: Interpersonal distance, body orientation, and touch: effects of culture, gender and age. J Soc Psychol 135:281–97, 1995(Patrice Marie Miller, EdD)
The issue of therapeutic boundary excursions, which includes everything from less serious crossings to more serious violations, has been extensively discussed since the 1990s in several publications.1–10 Strikingly, none of this recent literature involves an empirical examination of boundary violations, although some older literature has examined the prevalence of sexual infringements11–13 or harm to patients from such actions.14,15 Given the amount of theorizing about boundaries, more empirical work on the topic seems long overdue. Although several questions about boundaries and their violation might be asked, three will be focused on in this article: (1) How do clinicians classify different kinds of boundary excursions? (2) How might the cultural background of the therapist affect his or her perception of the seriousness of different excursions? and (3) In rating the seriousness of boundary excursions, do clinicians differentiate between the professional acceptability of a behavior and the harm a behavior might cause a client?
A boundary, according to the behavior-developmental view proposed herein,16,17 is a way of discriminating among environmental events and reacting in a way that may or may not lead to action. Actions taken in the context of some kinds of environmental events are seen as socially, legally, professionally, clinically, and morally acceptable. Such actions are also viewed as leading to improvement in the patient’s behavior, or at least as causing no harm. Actions in the presence of other kinds of situations, however, are not regarded as acceptable and may also be thought to cause harm. Because decisions on what actions to take in particular situations are often necessary, potential boundaries abound. Within the therapeutic context, however, some types of discriminations are clearly of greater interest than are others. Because this definition supports consideration of actions in context, it allows potentially different standards regarding what constitutes a boundary violation in different therapeutic and cultural contexts.
In much of the previous empirical work in this area, researchers have examined a limited set of possible boundary violations, particularly sexual violations. Empirically examining a large number of possible boundary transgressions may elucidate theoretical arguments about whether there are different kinds of boundary excursions. In some of the recent discussion about boundaries, for example, some boundary excursions have been said to represent more serious "violations" and others less serious "crossings."4–6,18,19 Can such statements by a few individual clinicians be confirmed by empirical work on a larger sample of clinicians? It is also possible that other useful ways of grouping boundary excursions will be found that may be helpful in thinking about boundaries in general.
Cultural Influences on Boundaries
Traditional notions of therapeutic boundaries in clinical practice have been said to ignore the broader social context, particularly culture (for example, Kroll6). Because culture, to a large extent, defines what is acceptable behavior, we agree with Kroll that boundaries should be seen in the context of larger groups, organizations, and societies (see Gutheil and Gabbard3,4 for additional discussion). While some kinds of behavior may be universally proscribed, others may differ in acceptability, depending on cultural norms.
Notions of psychological boundaries are most often considered with respect to intimacy. When and with whom would intimacies be allowed? Different cultures have widely varying notions of appropriateness20 on a large number of questions, including who one is allowed to talk to, visit with, or eat with; when and with whom various kinds of touching are appropriate; who can share property (and under what conditions); who is allowed to interact with children, or with women, and in what ways; with whom one is allowed to have sex, to name a few.
Brazilian and U.S. cultures, the two cultures from which the clinicians we studied originated, differ in a wide variety of ways. For example, Brazilian and other South American cultures are much closer to Mediterranean ones.21–23 In those cultures, physical contact in greeting and other interactions is normative. Standard greetings between people in Brazil include shaking hands, shaking hands with an arm clasp, kissing on both cheeks, and putting arms around people and hugging them. These greetings take place between people, even if they are just recent acquaintances. Touching, in general, is more frequent in such cultures than in Northern, Euro-American cultures. In the American culture, including in Boston, shaking hands on greeting is normative, but in some traditions even shaking hands might be avoided. Other kinds of touching during greeting and other types of physical interactions are usually reserved for people with whom one has a great deal of intimacy (family, close friends, romantic partners).
The two populations also differ with regard to payment for clinical services and gift-giving. Most people in Brazil do not have health insurance. If they have the money, they may pay out of pocket for health care services.24 Otherwise, state provision of services is common for those with little income. Patients may nevertheless feel obligated to compensate their therapists in some way, which can result in extensive gift-giving (Martins de Almeida K, personal communication, August 2001). Giving little gifts during visits is a valued part of this culture in most interpersonal relationships.
The economic realities of the two populations also differ in ways that may affect boundaries. The rates of poverty in Brazil are much higher than in the United States. For example, the top fifth of the population in Brazil receives 61.7 percent of the economy’s total income, and the bottom fifth receives only 3.3 percent.25 The U.S. Agency for International Development (USAID) reported in 2000 that about 20 percent of the Brazilian population lived below the poverty line (35 million/174.5 million total population).26 Because of the higher rates of poverty, the relatively low wages of mental health workers in Brazil, and the relatively high cost of drugs, "talk" therapy may be relatively more common than drug treatment compared with such treatments in the United States. One possible result of the frequency of exposure of patients to therapists in Brazil is that there is more opportunity for boundary violations to occur.
There are most likely many other differences between the two cultures as far as mental health practice goes. At this point, there is no single text or article that takes a wide variety of behavior in these two cultures and discusses the differences between them. Therefore, this discussion has had to rely on the relatively few studies that were cited earlier in the article.20–26 These studies suggest that as far as the general cultures are concerned, one would expect there to be differences in the behavior of Brazilians and North Americans that could have an impact on clinicians’ perceptions of what is appropriate within therapy.
It is important to add that another possible influence on clinicians’ behavior may be how those in the two cultures are trained and socialized. Clinicians in the Brazilian sample are heavily influenced by both European and American traditions, and particularly by psychoanalytic theories and approaches (Martins de Almeida K, personal communication, August 2001). The fact that training of clinicians in the two countries is similar may decrease cultural differences in the area of professional behavior.
Methodology
The current study empirically examines perceptions of boundary excursions in two groups of clinicians: one in the United States and one in Brazil. There have been a few empirical ratings of professionals’ perceptions of some number of hypothesized boundary crossings or violations,13 but these have not explicitly addressed the distinction between crossings and violations, as described by Gutheil and Gabbard.3,4 The present study goes beyond any earlier empirical rating studies, because there are few if any empirical data on cross-cultural differences in perceptions of boundary crossings and violations, despite arguments about the importance of culture in determining such perceptions.
The instrument used in our study asked professionals to rate the various examples on what we perceive as two different dimensions of boundary violations: (1) Was this practice judged to be professionally unacceptable, and (2) was this practice judged to be harmful to patients (both were rated on a seven-point scale). The distinction between professional unacceptability and harm is an important one. References to lack of professionalism are more abstract judgments about whether something is a norm for what constitutes professionally sound practice. That is, does a practice meet the standard of care? Judgments about harm may be based more on a practitioner’s experience with cases that are observed or heard about that resulted in actual harm to patients. Whether harm is experienced or not, therefore, seems to be more of an empirical matter. It is also possible that judgments of harm are more likely for more serious boundary violations, whereas lack of professionalism could relate to a wider range of such excursions. Of course, clinicians in the two cultures may perceive professional unacceptability and harm either in the same way or differently.
We specifically chose two samples that were relatively homogeneous within the cultures in which they existed. The clinicians sampled also had a relatively high degree of knowledge about therapeutic practice and possibly about boundary excursions in general. Our expectation is that these samples could then serve as comparison points for a similar survey to be administered to a much broader group of clinicians.
Participants
There were 60 participants: 28 Brazilians (20 women, 7 men, and 1 person who did not report gender) from the Institute of Psychiatry, Federal University of Rio de Janeiro; and 33 U.S. residents (18 women and 15 men) from the Program in Psychiatry and the Law, Department of Psychiatry, Harvard University, Massachusetts Mental Health Center, Boston. The U.S. sample was composed of attendees at and nonattending members of the Program in Psychiatry and the Law. The Program consists of a weekly meeting, where topics of all kinds relating to the practice of psychotherapy and to forensic psychiatry (including boundaries) are often discussed. The program conducts research on many of these subjects, and information is also exchanged on-line. Participants in Brazil were attendees at a regular weekly meeting of staff and members of the Institute of Psychiatry. During these weekly meetings participants either present and discuss cases or hear talks on a variety of professional topics. Staff members, postgraduate students, and other clinicians attend and participate in the meetings. During one particular meeting of the Institute, a talk stressing the ubiquity of cultural differences in boundaries was given by two U.S. investigators. This description of the meetings of the two groups suggests that both cultural groups interacted in discussion and theorizing about boundaries. It also suggests that within their own nations, both groups had a training, supervision, and information-sharing protocol for the participants. Both programs took place in major, prestigious universities in public hospitals. The exposure of the U.S. group to the subject matter of boundaries would have been greater than the exposure of the Brazilian group, according to feedback from that latter group. Further, the talk should have encouraged the Brazilian participants to report cultural differences if they existed, because it contained a strong message of acceptance of these differences.
All the participants from both countries were mental health professionals, although, as shown in Table 1, the distribution of types of degrees was somewhat different. Most respondents in the United States were psychiatrists, with a few having doctoral and/or master’s degrees. In Brazil, the sample was nearly evenly split between psychiatrists and pre-bachelor’s degree clinicians, who apparently do a great deal of the day-to-day work with patients. Pre-bachelor’s level clinicians in Brazil had obtained a two-year, post-high-school specialization degree in mental health practice. The one such individual in the United States was an undergraduate student who was attending the meetings due to an interest in forensic psychiatry. These differences partly reflect how mental health services are delivered in each country, as well as the membership of the Program in Psychiatry and the Law in Boston compared with the Institute of Psychiatry in Rio de Janeiro. While clearly there are differences in the distribution of different types of mental health workers in the two cultures, with the U.S. sample containing more MDs and PhDs, it was considered to be most important to have each sample reflect what was generally true in that context. In the context of the Institute of Psychiatry in Rio, the individuals in the sample realistically reflected those who work and deliver services to clients there. Trying to locate additional psychiatrists or psychologists who had credentials closer to those found in the U.S. would have resulted in a sample that did not realistically reflect this Brazilian context.
The study procedures, since they involved asking participants to fill out a questionnaire and participation was voluntary and anonymous, were judged to be exempt from review by the Institutional Review Board at the Massachusetts Mental Health Center, where the Program in Psychiatry and the Law was located at the time this study was conducted.
Instrument
An 87-item questionnaire was devised by consulting the literature on boundaries, from a large number of actual case examples, and from professional experiences shared within the Program in Psychiatry and the Law. The specific boundary excursions asked about are shown in the Appendix. Participants were asked to rate (on a six-point scale) the percentage of cases in which it would be harmful to the patient if a colleague behaved in the manner described, and concurrently, to rate in what percentage of cases this same behavior was professionally unacceptable. The rating scale asked participants to circle one of the following percentages: 0%, 2%, 16%, 50%, 84%, 98%, and 100%. This scale reflects a linear z-score scale that was converted into percentages, using the cumulative normal distribution function, with a probit transformation. This scale was used because (1) it represents the postulated distribution for most psychological characteristics, and (2) the differences at the ends of the scale are more important than the ones in the middle. The technique of asking about colleagues’ behavior, rather than the clinician’s own, was used to minimize defensive reactions that might occur if we asked individuals to rate how professionally unacceptable a behavior might be if engaged in by themselves.
All the questions and instructions were in English for the U.S. sample and in Portuguese for the Brazilian sample. The translation into Portuguese was the joint effort of two individuals. One was a born English speaker, a trained developmental psychologist in the English language, who had grown up in Brazil and speaks Portuguese fluently. The other was a born Portuguese speaker, a clinician trained in a Portuguese-speaking environment, who had worked as a clinician for several years in the United States with English- and Portuguese-speaking clients and colleagues. The native Portuguese speaker first translated the questionnaire into Portuguese. The native English speaker then compared the Portuguese phrases to the English version of the questionnaire, essentially back-translating it. In a few cases, the two speakers had to have further discussions about particular ideas that did not translate easily from one language to the other.
Procedure
Both groups were given the questionnaire. Almost all participants filled it out during a meeting of their respective professional groups. A few others completed it at another time and returned it to the investigators.
Results
Harm Versus Professional Acceptability
The questionnaire asked both about harm to patients and about the professional unacceptability of ever engaging in certain actions. The first step in the analysis, therefore, was to examine the extent to which ratings on these two types of questions were similar to each other. If responses to each were very different, then the analysis of the remainder of the results might have to consider each kind of rating separately.
The degree of similarity of the two ratings for each item, as measured by the correlation between harm and unacceptability,27 was r(86) = 0.769, p < .0005; the effect size28 for a correlation of this magnitude is 0.59, a large effect size. This suggests that in these respondents’ ratings, harm to a patient and lack of professionalism were seen as being highly related to each other. (Note that this correlation was determined in a three-step process. First, the correlation between the two ratings was determined for each of the 87 items (questions). Second, these 87 correlation coefficients were transformed using Fisher’s z transformation of r. Third, the two overall mean zr’s and the standard deviation of the zr’s (one for harm and one for lack of professionalism) were calculated. The correlation of these two means was then calculated, and it is this result that is reported here.)
Because the ratings on harm and on professional unacceptability had such a high degree of similarity, we could have decided to analyze one and not the other. There were two reasons to keep information from both kinds of ratings in the study. The empirical reason is that because they are not perfectly correlated, some variability in responses to each type of rating would be expected that then could be explored. Second, since information on both was collected as part of this study, we decided to keep both in future analyses.
Three Different Types of Boundary Excursions
With 87 questions and two ratings of each question (174 items in all), a natural question is whether we could group together questions that are related in some way. A useful technique for seeing whether questions might be related to one another is factor analysis. If questions were found to be statistically related, there would be a basis for grouping them together. Such groupings could help to distinguish different types of boundary excursions. Groupings of items would also make it easier to look at cultural differences, because fewer comparisons would have to be made.
An exploratory principal components factor analysis reduced 139 of the 174 responses to three distinguishable factors. The remaining 35 items did not relate strongly to any of these three factors. Each of these factors was named by the type of items that were most strongly represented within them. Both ratings of harm to the patient and ratings of the lack of professional acceptability were included.
The first factor, called core boundary violations, contained 85 items that generally constituted the most serious boundary violations (to simplify the presentation of the data, the specific items are not shown. More detailed data as well as copies of the questionnaires can be obtained from the authors). These violations were made up of several different types of therapist behavior. For example, 32.65 percent involved mixing therapy with personal or social considerations (e.g., telephoning the client to speak about matters besides therapy). The next largest number of items from this first factor (28.57 percent), involved sexual behavior (e.g., necking with client, or pretending sex is therapy). The other two types of violations that loaded with high coefficients on this factor were those involving financial matters, consisting of 18.37 percent of the items (e.g., selling non-therapy-related products to a client) and those involving actual physical or hostile aggression toward clients, 14.28 percent of the items (hitting the client). Both Americans and Brazilians indicated that Factor 1 items (core boundary violations) were the most serious, as indicated by an overall mean rating of those items of 6.04 ± 0.84 (equivalent to a rating of 98%), on a scale on which a mean rating of 7 would have indicated the most serious violation.
One additional characteristic of the items that loaded on Factor 1 is that 54 percent of these items, a slight majority, involved ratings that showed a perception that the item led to harm to patients, with 46 percent referring to lack of professional acceptability.
For the second factor, separation of therapist’s and client’s lives, the questionnaire contained 34 items that asked about situations in which the client and the therapist might encounter each other outside of the therapist’s office. These items differed from the items in Factor 1 that involved mixing therapy with personal and social matters, in that Factor 2 items involved a more distant mixing of therapy with other aspects. Only three (divulging one’s marital status or history of drug use, and offering refreshments during therapy) actually occurred in therapy. Some involved attending ceremonies involving the client (such as a wedding or funeral); others involved more inadvertent co-occurrences (such as attending an event at which the client was also present). Factor 2 items (separation of therapist’s and client’s lives) were given an overall mean rating that was just slightly above neutral (3.65 ± 0.95; SD), equivalent to a percentage halfway between 16% and 50%, suggesting that clinicians did not see these violations as being as serious as the core boundary violations. Slightly less than half of these items referred to harm to patients (47%), and slightly more than half (53%) to lack of professional acceptability.
Twenty items from the questionnaire were included in the third factor, disclosure and greeting items. In addition to asking about greeting behavior, such as shaking hands, kissing on the cheek, or hugging, these items asked about such behavior as displaying diplomas or awards or revealing one’s credentials. The disclosure and greeting items were given an overall rating between Factor 1, core boundary violations, and Factor 2, separation of therapist’s and client’s lives (4.35 ± 0.75; SD)—halfway between 50% and 84%. The preponderance of items that loaded on this factor (65%) referred to lack of professional acceptability, with only 35 percent referring to harm to patients.
For comparison purposes, we also calculated the overall rating of all items that did not load on the first three factors. These items received a rating that was just slightly below midpoint (mean, 3.89 ± 0.87; SD).
Cultural Differences
How did the U.S. and Brazilian clinicians differ on their ratings of harm and lack of professional acceptability of these items? To compare the two samples, an analysis of variance was performed on the overall item ratings for each of the three factors. Culture made a small but significant difference overall in the factor scores, with the Brazilians rating items in general more harmful and more unacceptable (F(1,166) = 5.708, p = .018, = .033). We also tested the effects of two other variables, gender and degree of training (doctoral level versus below that level). Only the items that loaded on Factor 2 (therapist-client separation) showed differences for these two variables. We found that the women tended to rate Factor 2 items somewhat higher in general than did the men (F(1,56) = 4.26, p < .044, = .071), a very small effect size. Nondoctoral level practitioners tended to rate Factor 2 items somewhat higher (F(1,54) = 6.23, p < .016, = .103), a small effect size.
Very few of the ratings of individual items showed significant cultural differences. The items included in the first factor showed almost complete uniformity in the two cultures. There were just two items that had significant differences and one item that was at the level of a statistical trend. (As suggested by Rosnow and Rosenthal,28 the raw p < .05 level was adjusted for the 139 comparisons by dividing .05 by 139. This Bonferroni-corrected value was p < .0036.) In all three cases, the differences were negative, meaning that the U.S. clinicians found these behaviors to be less harmful than did the Brazilians: Item 85a: "Percentage of cases where therapist’s calling client after office hours to talk about treatment is harmful to client" (Mean difference = –2.19, t(55) = –4.57, p < .0005); and Item 56a: "Percentage of cases where therapist’s sitting with a client in a café/cafeteria which both frequent is harmful to client, if done by colleagues" (Mean difference = –1.44, t(55) = –3.177, p < .002). The one item that was significant at the level of a statistical trend (mean difference = –2.94; t(55) = –2.84, p < .006) was "Therapist’s revealing his/her sexual orientation to client is harmful."
Within the second factor, there were only four significant cultural differences. In three cases, as with the items in Factor 1, Brazilian clinicians saw the harm or the lack of professional acceptability as higher for placing diplomas in one’s office and telling one’s own history of drug use. They saw as significantly less unacceptable attending a small event outside of therapy that was also attended by the client.
The items in Factor 3 showed a somewhat different pattern, with a larger number of them (9/20 or almost half) showing significant cultural differences. These differences were found in two different areas of therapist-client interactions. On the one hand, Brazilian therapists tended to rate behavior such as displaying diplomas, professional awards, and/or revealing one’s credentials or emotions, as more harmful and professionally unacceptable than did American therapists. On the other hand, they rated two greeting behaviors, shaking hands and kissing on the cheek (a very common way of greeting in Brazil), as less harmful and less professionally unacceptable.
Surprisingly few cultural differences appeared out of the possible 87 times 2 (174) items. By chance alone, one might expect 9 items to differ significantly at the .05 corrected level of chance; 15 were actually found to be different. While this is not evidence in favor of a strong cultural difference, we examined whether there were cultural differences in particular subsets of variables when these variables were combined in a different way. Some of the items involved: sexual violations, attempted business transactions between therapist and patient, how much information a therapist should disclose to a patient, situations in which therapist and patient may mix together outside the office, and routine touching such as shaking hands. We created five new variables based on these categories and then examined each category for whether there were cultural differences. The results are shown in Table 2.
Brazilian clinicians were likely to rate various self-disclosure items as significantly more harmful and professionally unacceptable than did U.S. clinicians, whereas they were likely to rate routine touching that takes place as part of greeting or comforting as significantly less harmful and professionally unacceptable.
Discussion
This study is the first empirical examination of a large range of boundary violations, beyond just sexual violations, and examines the perceptions of clinicians from two cultures. The results add several pieces of information to the general discussion of boundaries that has been taking place in the literature cited in the introduction. First, the results of the factor analysis suggest that there are three types of boundary violations, as represented by the three factors that were found. These types are distinguished partly by the fact that the clinicians surveyed rated some types of violations as being more serious and other types as less serious. The factors also grouped together items of different types.
The first type of boundary violation (the first factor in the factor analysis) we called core boundary violations. These are the most serious types of therapist misbehavior that can occur. They include not only the expected sexual violations, but also violations involving the mixing of therapy with personal or social considerations, violations involving financial matters (e.g., selling non-therapy-related products to a client) and those involving actual physical or hostile aggression toward clients. The boundary violations that made up this factor were ones that were particularly intrusive. Both U.S. and Brazilian clinicians rated the 87 violations that loaded on this factor as being very harmful to clients and as being highly unprofessional.
The second type of boundary excursion, separation of therapist’s and client’s lives, was covered by 34 items on the questionnaire that concerned situations in which the client and the therapist might encounter each other outside the therapist’s office. The items in this factor were rated as close to the neutral point by both U.S. and Brazilian clinicians. In other words, they were seen as less harmful and less professionally unacceptable than were the core boundary violations.
The third type of boundary excursion on the questionnaire consisted of disclosure and greeting behavior. Items from this factor were rated just above the neutral point, so they were seen as slightly less neutral than Factor 2 items.
The usefulness of these results is twofold. First, professionals classified some kinds of excursions as more serious than others. Although the basis for this classification needs further elucidation, these results at least give the field a preliminary empirical rubric with which to evaluate professionals’ behavior. They provide, in addition, some empirical validation of the notion of boundary crossings versus boundary violations. Second, by simply studying such a broad range of possible violations, we can educate professionals about the spectrum therapeutic behavior that may be similar in many ways to the few well-recognized types of boundary violations, such as sexual ones.
A second question addressed by this study was whether differences between cultures would be apparent. In fact, with these samples, we found surprising uniformity across the two cultures. Very few of the ratings of individual items, for example, showed significant cultural differences. This was particularly true of the core boundary violations (only 2/87 of these items were significantly different). The other two factors of boundary excursions studied showed somewhat more differences in the ratings by Brazilian versus U.S. clinicians. For example, on the one hand, Brazilian clinicians tended to rate aspects of self-disclosure as more professionally unacceptable and harmful. On the other hand, they tended to see certain routine greeting behavior that is more common in Brazil as more acceptable and less harmful. These findings were further confirmed by a related analysis in which we collapsed together items according to their content (e.g., sexual behavior, other touching, business- or money-related disclosure).
When there were cultural differences in ratings of items, the Brazilian therapists tended almost always to rate those items as more harmful and/or more professionally unacceptable than did U.S. therapists. It may be (although we have no way of confirming this notion at present) that when individuals are less experienced in an area they tend to perceive rules as being less flexible. They do not yet have enough exemplars to see that following a rule too strictly is not always of benefit. The only exception to the tendency of the Brazilian therapists to rate all the boundary excursions as being more harmful and more unprofessional, was their rating of routine touching, such as shaking hands or kissing on the cheek, that may take place as part of greetings, which they rated as somewhat less harmful and less unacceptable. In any case, understanding the bases for such differences requires further empirical work, both to confirm that these differences are generally cultural and to understand the culturally-based interaction patterns that may underlie them.
Why were there not more cultural differences in beliefs? One reason may be that therapists in both samples were trained based on similar models. Training of mental health personnel in these cases seemed to emphasize to a greater or lesser extent the psychodynamic viewpoint and its associated views of the proper role for a therapist. While only a minority of the sample in Rio de Janeiro were psychiatrists, the other therapists worked alongside those psychiatrists and were most likely imbued with much of the same perspective. Perhaps professional training of this kind comes to supercede cultural differences in interactions that might be seen in other, more informal contexts. We also cannot eliminate the idea that a social-desirability bias may have been operating, and perhaps more so for the Brazilian therapists. The majority would have been well-enough educated in what is supposed to be appropriate conduct from the point of view of U.S. or European therapies, and may have answered as they thought they should answer. To answer otherwise might shed a negative light on therapeutic practice in Brazil. Finally, it is also possible that Brazilian and U.S. health professional cultures are just not that different, particularly when one examines professionally circumscribed behavior, except for a few small and relatively superficial aspects.
This is not the same as saying that there would not be any differences in the practice of therapy in the two cultures, as far as boundary violations are concerned. It is possible that in behavior, one would see greater differences. Anecdotal reports from some of the participants and from other Brazilian therapists (Martins de Almeida K, personal communication, August 2001) mentioned that there was little regulatory action surrounding boundary violations in Brazil. There were no lawsuits, no firings, no licensure or ethics complaints that these sources were familiar with. These same sources also noted that it was their perception that there might be a greater prevalence of major boundary violations in Brazil than in the United States.
A third finding is the surprising uniformity of ratings of harm and of lack of professional acceptability. We had predicted that the frequency of judgments of harm might be higher for more serious boundary excursions. Ratings of harm and lack of professional acceptability correlated relatively highly. We argued that because the correlation was not perfect, we might still see evidence of some difference between the two. In fact, it was found that in Factor 1, which consisted of the core boundary violations, the majority of items included referred to the harm to clients that a behavior might cause. In the other two factors, particularly Factor 3, more of the items referred to lack of professional acceptability rather than to harm. This is some evidence that the clinicians surveyed think of these two categories in slightly different ways. Some behaviors, in and of themselves, may be seen as lacking professional acceptability but as not being all that harmful to clients.
The samples used in this preliminary study were small and not very diverse. Because both groups, however, came from prestigious training institutions in major cities of their respective countries, they represent a kind of establishment view of boundary violations. As such, the results represent a reasonable place to begin to improve our understanding of boundary excursions and in fact supply a comparison point for further study. The uniformity of opinion, even across cultures, suggests that, despite the relatively small samples, the findings merit consideration.
In addition to studying professionals’ perceptions of boundaries within larger and more diverse samples (particularly in terms of therapeutic orientation), another important direction for such research is to try to document in a more complete fashion: (1) specifically how often the different kinds of excursions take place within therapy, and (2) what actual harms stemming from different kinds of excursions might exist.
Appendix: Boundary Excursions Asked About in the Questionnaire, in Order of Their Appearance
References
Duckworth K, Kahn M, Gutheil TG: Roles, quandaries and remedies: teaching professional boundaries to medical students. Harv Rev Psychiatry 1:266–70, 1994
Garfinkel PE, Dorian B, Sadavoy J, et al: Boundary violations and departments of psychiatry. Can J Psychiatry 42:764–70, 1997
Gutheil TG, Gabbard GO: The concept of boundaries in clinical practice: theoretical and risk management dimensions. Am J Psychiatry 150:188–96, 1993
Gutheil TG, Gabbard GO: Misuses and misunderstandings of boundary theory in clinical and regulatory settings. Am J Psychiatry 155:409–14, 1998
Gutheil TG, Simon RI: Non-sexual boundary crossings and boundary violations: the ethical dimension. Psychiatr Clin North Am 25:585–92, 2002
Kroll J: Boundary violations: a culture-bound syndrome. J Am Acad Psychiatry Law 29:274–83, 2001
Norris DM, Gutheil TG, Strasburger LH: "This couldn’t happen to me": boundary problems and sexual misconduct in the psychotherapy relationship. Psychiatr Serv 54:517–22, 2003
Simon RI: Sexual exploitation of patients: how it begins before it happens. Psychiatr Ann 6:670–7, 1989
Simon RI: Psychological injury caused by boundary violation: precursors to therapist-patient sex. Psychiatr Ann 2:614–19, 1991
Strasburger LS, Jorgensen L, Sutherland, P: The prevention of psychotherapist sexual misconduct: avoiding the slippery slope. Am J Psychother 46:544–55, 1992
Gartrell N, Herman J, Olarte S, Feldstein M, Localio R: Psychiatrist-patient sexual contact: results of a national survey, I: prevalence. Am J Psychiatry 143:1126–31, 1986
Kardener SH, Fuller M, Mensh IN: A survey of physicians’ attitudes and practices regarding erotic and nonerotic contact with patients. Am J Psychiatry 130:1077–81, 1973
Epstein RS, Simon RI, Kay GG: Assessing boundary violations in psychotherapy: survey results with The Exploitation Index. Bull Menninger Clin 56:150–66, 1992
Butler S, Zellen S: Sexual intimacies between therapists and patients. Prof Psychol 14:139–45, 1977
Pope KS, Tabachnick BG: Therapists as patients: a national survey of psychologists’ experiences, problems, and beliefs. Prof Psychol 25:247–58, 1991
Commons ML, Miller PM: A quantitative behavior-analytic theory of development. Mexican J Exp Anal Behav 24:153–80, 1998
Commons ML, Miller PM: A quantitative behavioral model of developmental stage based upon hierarchical complexity theory. Behav Analyst Today 2:222–40, 2001
Martinez R: A model for boundary dilemmas: ethical decision-making in the patient-professional relationship. Ethic Hum Sci Serv 2:43–61, 2000
Lazarus AA: How certain boundaries and ethics diminish therapeutic effectiveness. Ethics Behav 4:255–61, 1994
Hall ET: The Silent Language. Garden City, NY: Doubleday, 1959
Sussman NM, Rosenfeld HM: Influence of culture, language and sex on conversational distance. J Pers Soc Psychol 42:66–74, 1982
Regan PC, Jerry D, Narvaez M, Johnson D: Public displays of affection among Asian and Latino heterosexual couples. Psychol Rep 84:1201–02, 1999
Remland MS, Jones TS, Brinkman H: Interpersonal distance, body orientation, and touch: effects of culture, gender and age. J Soc Psychol 135:281–97, 1995(Patrice Marie Miller, EdD)