To download a pdf version of this article, click here.

How Useful Are Clinical Reports Concerning the Consequences of Therapist-Patient Sexual Involvement?

Department of Psychiatry
Kaiser-Permanente Medical Center
Santa Clara, California

1995, American Journal of Psychotherapy, 49 (2),237-243.


Psychotherapist-patient sexual involvement has been a difficult topic for systematic research.  Systematic studies of harmfulness to patients or prevalence of this unethical practice have been limited by a variety of insurmountable methodological problems.  Nevertheless, a pattern such that patients have been harmed by pre- and post-termination sexual involvement with therapists has been clinically observed.  The present paper reviews the inferential limitations on clinical observations in general, and with respect to therapist-patient sex in particular, and concludes that clinical observation cannot establish a valid cause-effect relationship between therapist-patient sexual involvement and adverse consequences.

How Useful Are Clinical Reports Concerning the Consequences of Therapist-Patient Sexual Involvement?

Inferential problems inherent in systematic research on the harmfulness and prevalence of therapist-patient sexual involvement have been discussed elsewhere, e.g., (Williams, 1992a, 1992b).  Sampling bias and resultant limitations on the generalizability of findings have been identified as significant issues.  Because of these inferential problems, conclusions regarding the degree of harm generally caused by therapist-patient sexual involvement may be surprisingly difficult to draw.  In fact, despite one’s personal convictions to the contrary, it remains empirically unsubstantiated that sexual involvement with one’s therapist generally causes damage beyond that which would have been manifested due to the patient’s pre-existing condition, or that such sexual involvement causes more harm than that caused by mundane sexual involvement, i.e., between the same patient and any individual who has never been that person’s therapist (Williams, 1992a).   Three familiar phenomena contribute to inferential difficulties concerning the damage caused by therapist-patient sex, either before or after termination:  (1) Individuals, including those with psychological disorders, sometimes change unpredictably over time in the degree to which their symptomatology is experienced or manifested, rendering it difficult, in many cases, to establish that a particular event was a causal precursor to subsequent poorer psychological functioning; (2) Individuals may misattribute their adverse psychological conditions to precursors which were not, in fact, the causal agents, and such misattribution may occur in a reliable, and hence, especially misleading, fashion; and (3) Mundane sexual relationships, i.e., those not involving patients and their therapists, may be a significant cause of subjectively perceived and objectively measured psychological harm, leading to questions concerning how much harmfulness constitutes a baseline, average, normal or expectable aspect of any sexual relationship.  These three phenomena may cause insurmountable inferential difficulties which would preclude the valid attribution of harm to prior sexual involvement with a therapist based on subsequent interviews with the patient by therapists or researchers (Williams, 1992a).

Despite this, Williams (1992a) and others have asserted that a clinical picture of harmed individuals does exist.  However, a recent and unique publication by Brown et al.(1992) has led me to reconsider the question of the extent to which one can legitimately call upon clinical observations to substantiate what are believed to be typical cause-effect patterns of behavior.  The Brown et al. (1992) article is unique by virtue of its having been co-authored by nearly every significant contributor to the empirical literature regarding therapist-patient sexual involvement.  Consequently, any positions which are espoused regarding inference are especially noteworthy, as they can be taken as indications of the underlying philosophy which guides these authors in their approach to the problem of determining the harmfulness of therapist-patient sex.  The article includes the following claim that clinical findings have substantiated what more systematic empirical studies have not adequately done: established that post-termination sexual relationships are harmful to patients.  Brown et al. state the following regarding Applebaum and Jorgenson’s (1991) assertion that post termination harm has not been substantiated by valid empirical studies:

Dr. Applebaum and Ms. Jorgenson also argued that there is no strong evidence for harm as a result of such relationships.  While we would agree that the published literature is short on such information, and consider ourselves partly responsible for not having collected and published the information we have gathered in the course of our clinical work, our impressions are otherwise.  That is, we have seen the same type and severity of harm devolving from posttermination relationships as from those initiated by therapists with less effective impulse control. (Brown et al., 1992, p. 980)

This quotation leads to the inferential question of what difference it would have made had these authors done what they chided themselves for having failed to do: gathered together and published all instances of harmed patients which they had observed in the course of their clinical work.  I propose that such publication would not, and should not, affect general conclusions concerning the consequences of post-termination therapist-patient sexual involvement.  The inferential reasons for this proposal may seem self-evident, yet they require review here because, perhaps especially in the area of therapist-patient sexual involvement, concern for victimization of patients may have clouded otherwise clearly recognized limitations on inferences which may be drawn from clinical observation.

Factors Which Limit the Usefulness Of Clinical Reports

The following three factors limit the usefulness of clinical reports of harm befalling patients as a means of drawing general conclusions regarding the harmfulness of therapist-patient sex:

1. The problem of establishing a rate of harm.  Within reasonable limits, clinical reports are unable to establish general trends in populations of patients.  Unless some unlikely hypothetical critical mass were reached, such that nearly all therapists report seeing large numbers of similar patients—-with a remarkably large number of therapists sharing identical observations—-reports of any clinical condition are of very limited utility because they lack a denominator.  Even if a given psychotherapist interviews one, ten, or a hundred harmed survivors of therapist-patient sexual involvement, this therapist is still be unable to ascertain whether the observed harm is a common or uncommon result of the sexual involvement.  Such observations lack the following denominators: number of survivors of therapist-patient sexual involvement in the population, and the number of these survivors who are unharmed.  Brown et al. (1992) make clear that they believe that harm is a likely result of therapist-patient sexual involvement.  Thus, for each clinical case these authors encounter, one would expect them subjectively to minimize the denominator of unharmed survivors and to view, with questionable empirical basis, instances of harm as typical.  Although many, including myself, agree with Brown et al’s. conclusion, this conclusion may bolstered more by faith than by science.

This is not to say that case reports are useless.  They are useful for providing in-depth studies of particular conditions, and they can bring to light clinical patterns which were not previously known to exist.  For example, they have brought to light that therapist-patient sexual involvement does occur, i.e., occurs with some frequency greater than zero, and they can document observed patterns of harm.  However, despite recent demonstrations of the validity of “softer,” self-report measures of behavior (Howard, 1993), case reports cannot analyze population trends.  Even pooled clinical case reports cannot, by themselves, distinguish exceedingly rare population occurrences from the commonplace.  Something that is common in one therapist’s practice, or in the practice of some group of therapists, may be rare elsewhere.  One recalls that the purported efficacy of formerly accepted, but now debunked, medical treatments such as cupping, or the therapeutic use of poisons like arsenic or strychnine, had been promulgated by means of the case study methodology.

2. The problem of determining the cause of the observed harm.   Therapists who treat survivors of prior therapist-patient sexual involvement may incorrectly conclude that the subsequent adverse clinical picture is the result of the inappropriate sexual relationship which preceded it.  When a psychotherapist treats a survivor of therapist-patient sexual involvement, that therapist must make a post facto assessment of the patient’s condition as it would have existed prior to the sexual abuse, and must posit what the current condition would have been had no sexual abuse occurred.  It is only with respect to this posited prior state, and alternative subsequent state, that one determines the harm that resulted from the  sexual abuse.  This assessment is typically based only on history provided by the patient.  Although in some instances prior treatment records, which predate the episode of abuse, are available and are forwarded to the new therapist, this is not typical.  Generally, the therapist must make a set of assumptions regarding what the current condition of the patient would have been had the sexual involvement not occurred. 

Campbell and Stanley (1963) have labeled research designs in which a group is studied only once the “one-shot case study,” and they have noted that a major inferential problem in such studies is the inability to establish what the subjects’ condition would have been had they not undergone the treatment in question--in this case, the “treatment” is therapist-patient sexual involvement.  This problem is insurmountable in clinical reports.  The therapist may deeply believe that the patient’s:

ambivalence; guilt; feelings of isolation; feelings of emptiness; cognitive dysfunction...; identity and boundary disturbance; inability to trust (often focused on conflicts about dependence, control and power); sexual confusion; lability of mood; suppressed rage; and increased suicidal risk (Pope and Bouhoutsos, 1986, p. 64)

are the result of the patient having experienced sexual involvement with a prior therapist.  The therapist may further believe that the patient’s current condition would have been far better had the sexual involvement not occurred.  Pope and Bouhoutsos (1986) have suggested that the above list of clinical conditions may be the result of therapist-patient sexual involvement, and may constitute the Therapist-Patient Sex Syndrome. 

Another therapist, however, might attribute the present set of adverse clinical conditions to childhood sexual abuse while positing that the sexual relationship with the therapist hardly changed a picture of chronic distress and re-victimization.  Still another therapist might assert that the patient suffers from a personality disorder and might discount any contribution to the patient’s suffering from such a proximal cause as any particular sexual relationship during adulthood.  Finally, a therapist might believe that the prior therapist-patient sexual relationship had caused the patient’s present distress, but this therapist might believe also that the patient would have suffered equally, because of a personality disorder, from any interpersonal intimacy—whether with a therapist or not.  Indeed, Pope and Bouhoustos (1986) have cautioned

For many patients there may be no data deriving from formal testing and assessments performed prior to the sexual involvement with the therapist.  Without such baseline data, the assessment of the damage that was due to the sexual involvement becomes more complex and difficult.  (p. 64 ) [italics added]

Ultimately, the therapist’s attribution of the cause of the patient’s present constellation of symptoms may have less to do with the symptoms themselves, and less to do with the patient’s reported history, than with the therapist’s own enduring beliefs about life history antecedents of observed symptoms.  Tavris (1993) has recently criticized the current trend such that therapists attribute a wide variety of psychological symptoms to the prior experience of childhood sexual abuse, often without any basis for this attribution other than the therapist’s deepest beliefs.  Tavris notes that therapists sometimes reason backwards, taking the presence of certain very common symptoms to constitute nothing less than proof of an earlier experience of incest.  It remains questionable, in many cases, whether childhood incest actually occurred or whether it had been suggested to the patient by a therapist in adulthood, creating a false memory (Loftus, 1993).  It remains questionable, too, in some cases how much of subsequent psychological harm was caused by even well substantiated episodes of childhood incest.  For example, in very dysfunctional, sadistic and brutally abusive families, how much is an overall picture of overwhelming emotional damage altered by the presence or absence of sexual abuse?  

These observations regarding childhood incest may apply as well to the method by which some therapists attribute subsequent damage to a prior episode of therapist-patient sex.  The causal link is often completely without substantiation.  A therapist interviews a woman who states that she arrived at her current distressed and dysfunctional state by virtue of sexual abuse by a prior therapist, and her report is may be taken at face value.  In a prior era, such reports were routinely met with excessive skepticism, e.g., Wright (1985), but today the pendulum may have swung too far in the opposite direction.  Mistaken causal links, or causal connections more complicated than those posited by the patient, my not be suspected by the new therapist.  This is not to invalidate the posited causal connection, i.e., therapist-patient sex leading to the Therapist-Patient Sex Syndrome.  It is only to argue that case studies may not be able to supply the needed validation.

3. The natural history of the borderline and histrionic personality disorders.  One reason to question that therapist-patient sexual involvement has actually caused a given instance in the which patient appears to manifest the Therapist-Patient Sex Syndrome is the natural history of the borderline and histrionic personality disorders.  Pope and Bouhoutsos (1986) have noted that individuals suffering from these disorders are at greatest risk for therapist-patient sexual involvement.  However, this is confounded by the following:  Individuals who suffer from borderline or histrionic disorders tend, at any given time, to manifest symptoms which greatly resemble those of the  Therapist-Patient Sex Syndrome.  Hence, such individuals would seemingly show the sequellae of therapist-patient sexual involvement whether they had experienced such involvement or not.  Further, if such individuals are in fact at greater risk for therapist-patient sexual involvement, then this would create a reliable sampling bias such that many or most survivors of therapist-patient sex would manifest the Syndrome of harm, when, in fact, the Syndrome may have predated the sexual abuse, leaving the causal question of which symptoms are reliably caused by therapist-sexual abuse very difficult to address.


Cases exist in which the behavior of the psychotherapist was uncommonly predatory and remarkably selfish, cases in which the psychotherapist engaged in fraud, manipulation, and even brainwashing or brute force in order to induce a patient to have sex, and cases of deception that offends any notion of even the most liberal professional ethics.  Our sense of outrage at such flagrant violations, not only of our professional codes of ethics but also of our notions of ordinary business ethics and human decency, may make us wish for punishment of the therapist and reparations for the client.  We may endorse the payment of large settlements in civil court to help rectify the injustice which was perpetrated on the client by virtue of the malpractice which occurred. 

All this notwithstanding, our degree of outrage at what was done must be separated from our assessment of the causal connection between the short and long term harm that resulted and the precursor of therapist-patient sexual involvement.  Two key conceptual questions remain:  (1) Regardless of what transpired, did lasting harm result, and (2) If harm was the result, how much did the sexual acting-out which took place contribute to the unfortunate result, and how much was attributable to other psychological and social events, such as the patient having being deprived of adequate psychotherapy when it was needed?  In the legal arena, the truth may be approximated only after careful examination of complete records, interviews with corroborative witnesses, and in-depth assessments by expert witnesses.  In contrast, the clinician compiling case records generally has little or no access to sources of independent validation.  How often do sympathetic therapists simply accept the patient’s understanding of what transpired in the past?  Some may argue that any skepticism towards such reports of victimization constitute “blaming the victim” and should be avoided on that basis.  

Plaintiff’s attorneys have created a sizable industry based on the pursuit of cases of therapist-patient sexual abuse.  This industry was born out of several remarkably egregious and notorious cases of patient abuse by therapists, e.g., Hartogs vs. Roy and Freeman vs. Zipkin (Pope and Bouhoutsos, 1986), and it matured with much help from Brown et al., who provided reliable, but not necessarily valid  research findings which seem to demonstrate that the presence of sexuality in dual relationships between therapist and patient leads to severe damage to the patient(Williams, 1992a).  Now it is time to step back and consider how much is really known about damaged patients. 

Proof of ethical violations by the therapist may be sufficient to lead to a substantial settlement for the patient.  Without doubt, this connection between violation and remuneration has its place in the courts.  However, the question of entitlement to remuneration must remain distinct from the scientific question concerning the cause-effect relationship between sexual behavior and psychological trauma.  Although sympathy towards victims is an appropriate antecedent to a large settlement and criminal law enforcement, it has no place in a clinical science which seeks to better understand human behavior.  To put this another way, the standards for validity adopted by those who author reports and, implicitly, by those who read them, should not be any less stringent simply because of mutual outrage regarding the issue under consideration.


Applebaum, P. & Jorgenson, L. (1991).  Psychotherapist-patient sexual contact after termination of treatment: an analysis and a proposal.  American Journal of Psychiatry, 148, 1466-1473.

Brown, L., Brodsky, A., Gartrell, N., Bouhoutsos, J., Davidson, V., Feldman-Summers, S., Gabbard, G., Herman, J., Olarte, S., Pope, K., Vasquez, M., Folman, R., George, J., Lerman, H., Olsson, P., & Sonne, J. (1992).  Psychotherapist-patient sexual contact after termination of treatment.  American Journal of Psychiatry, 149, 979-980.

Campbell, D.T., & Stanley, J.C. (1963).  Experimental and quasi-experimental designs for research.  Chicago:  Rand McNally College Publishing Company.

Howard, G.S. (1993).  I think I can! I think I can!  reconsidering the place for practice methodologies in psychological research.  Professional Psychology: Research and Practice, 24, 237-244.

Loftus, E. (1993).  The reality of repressed memories.  American Psychologist, 48, 518-537.

Pope, K.S. & Bouhoutsos, J.C. (1986).  Sexual intimacy between therapists and patients.  New York: Praeger.

Tavris, C. (Jan. 3, 1993).  Beware the incest-survivor machine.  New York Times Book Review,  1-17.

Williams, M.H. (1992a).  Exploitation and inference: mapping the damage from therapist-patient sexual involvement.  American Psychologist, 47, 412-421.

Williams, M.H. (1992b).  Correction to Williams.  American Psychologist, 47, 1114).

Wright, R.H. (1985).  Who needs enemies.  Psychotherapy in Private Practice, 3, 111-118.

To download a pdf version of this article, click here.