Termination of treatment with a psychiatric resident because of the resident's graduation from the training program is a recurring event for patients who choose to—or have to—obtain treatment in teaching hospitals (+1,+2). Such a treatment relationship may also end when the resident must downsize his or her outpatient load in order to take on other responsibilities, such as chief residencies or research activities. Termination may also occur as a consequence of the time-limited nature of some clinical rotations; for example, in rotations that take place in outlying community clinics, or in subspecialties such as child psychiatry or group therapy, which may last 1 year or less for most residents. These terminations are thought to be qualitatively different from terminations that occur by mutual agreement (e.g., when treatment goals have been reached), or because of circumstances on the part of the patient, as may occur when the patient relocates to another city, runs out of funds or insurance coverage, or acts out by leaving treatment (+1,+2).

Patients may experience the news of the upcoming end of treatment as abandonment and loss; a mourning process may follow (+3,+4), and treatment outcomes may be negatively affected. Our experience suggests that common transference reactions to termination may also include a triad of 1) hostility toward the therapist and/or toward the new treater; 2) regression, as seen by worsening symptoms, acting out, or self-harming behavior; 3) new "material," as in new symptoms or disclosures on the patient's part.

Termination has been studied primarily in the psychotherapeutic or psychoanalytic setting and is presented mostly in the form of vignettes and case reports with individual patients. Quality data are limited regarding termination, and there are few articles addressing cases of forced termination (+5—+11). In one previous study of termination (+12), 101 residents were asked to answer questions about a representative case from their individual psychotherapy practice, including the resident's reaction to termination with the patient, supervision around termination, and other related issues. Not surprisingly, the study revealed that change of setting or end of training were the most common causes of termination. Almost half of the 54 residents who responded felt that therapy work was left incomplete, and fewer than 20% of patients were thought to be "ready" for termination. Interestingly, the least intensive supervision occurred in cases of "forced" or "premature" termination. The authors speculated that this dearth of supervision may have resulted from residents' and supervisors' avoiding the issue as a consequence of countertransference; indeed, other studies (+13—+15) have documented several different resident countertransference reactions that can emerge in such a setting.

Despite careful searches, we found no published studies on termination with psychopharmacology patients in particular. Because psychopharmacologic management is qualitatively and quantitatively different from psychotherapy, it is of interest to ask about what issues emerge in the setting of termination with a psychopharmacologist. We therefore wished to address whether residents' termination with psychopharmacology patients would have a negative impact on the patients' condition and/or treatment outcome and whether there might be ways to improve outcomes when change of treaters is inevitable. Using principles adapted from the lore of psychoanalysis and psychodynamic psychotherapy, we hypothesized that termination with psychopharmacology patients would result in a triad of hostility (acting out, noncompliance), regression (worsening of symptoms), and new "material" (new symptoms or diagnoses). As an ancillary hypothesis, we proposed that earlier announcement of termination would result in a better outcome.

A questionnaire (available from the authors upon request) was distributed to 38 adult psychiatry residents (PGY-2, -3, -4, and the "PGY-5" group of June 1997 graduates) from the Massachusetts General Hospital (MGH), in October of 1997. MGH psychiatry residents who wish to transfer psychopharmacology patients at the end of the academic year (usually graduating residents and those entering the PGY-4 year) arrange this through the chief resident in psychopharmacology, by providing a list of their patients, along with brief clinical information about each patient. The chief resident then assigns these patients to the incoming PGY-2 and PGY-3 residents. On the basis of our review of the chief resident's records, residents were identified on the basis of having received new patients or having transferred patients. Residents who did not transfer any patients and did not receive any new patients at the change of academic year were not asked to participate (only six residents, all from the PGY-4 class, fit this category).

Respondents were classified as 1) residents who had transferred patients at the end of their PGY-3 or PGY-4 years (June 1997); and 2) residents who had received patients at the beginning of the PGY-2 or PGY-3 years (July 1997). Residents who transferred patients were asked questions about the patients they transferred; these included 1) when they first announced the upcoming change in treaters; how many of their transferred patients were in therapy (and whether they were among the ones whose condition worsened); 2) how many patients worsened after the announcement, remained the same, or improved; how many considered it a major disruption in treatment; how many developed new symptoms, required changes in medications, required more frequent visits, expressed hostility about the change, or became noncompliant (defined as not taking medications as directed, not showing up for appointments, or dropping out of treatment). Residents who received these patients in July of 1997 (PGY-2/3) were asked the questions from Part 2 of the PGY-4/5 questionnaire (and also whether the patients were in therapy during that period) to provide an assessment of the course of their new patients during the first 4 months of care. Resident assessments of patients were based exclusively on their own clinical observations; no diagnostic instruments were used.

Specific data about the demographics and diagnoses of the patients were not requested, so as to keep the questionnaire simple, and also to focus on general trends among the patient population as a whole. A "typical" psychiatric resident outpatient load at the MGH consists primarily of patients with depression and anxiety disorders, relatively fewer patients with bipolar disorder, and very few patients with chronic psychotic disorders such as schizophrenia or schizoaffective disorder. Psychiatric outpatients at the MGH tend to be chronically ill, and their level of function may range from fully employed to completely disabled.

Because the two resident samples transferring patients (PGY-4 and PGY-5) were considered independent of each other, an independent t-test was carried out to assess statistical significance for differences in the time of announcement of termination by the two resident classes. Chi-square analysis was carried out to determine if the presence of concomitant psychotherapy influenced whether or not patients worsened during the period of transfer of care (+16).

The return rate on the questionnaire was 38/38 (100%), a reflection of the small sample size, the ease of reminding participants to fill out and return the questionnaire, and the general conscientiousness and good will of the residents in our program. A total of 16 residents had transferred patients at the end of their PGY-3 or PGY-4 years (July 1997), and a total of 22 residents had received patients at the beginning of the PGY-2 or PGY-3 years (July 1997).

Eleven PGY-5 residents were surveyed. They had transferred a total of 138 patients and reported an average time of announcement of 4.4 months before the end of the academic year. The range of announcement time was from 1.5 to 6 months before the end. Although most patients remained unchanged, and a small number even improved, 22% of the patients were thought to have worsened clinically; of the 31 patients who worsened, only 12 (39%) had other mental health treaters, such as psychotherapists. Overall, 37% of the patients considered the change in treaters a major disruption in treatment (see +Table 1).

Five PGY-4 residents were surveyed. Typical reasons for termination of treatment with their psychopharmacology patients at the end of their PGY-3 year included the taking on of chief residencies, research commitments, short-tracking into fellowships, and plans to carry out clinical work in other outpatient settings. They transferred a total of 47 patients and reported an average time of announcement of 2.2 months before the end of the academic year. The range of announcement time was from 1.5 to 3 months before the end. The time of announcement was noticeably shorter than that reported by graduating residents and was found to be statistically significant by the independent t-test (P<0.01). This group of residents reported that 89% of their patients remained unchanged, and 11% worsened; of the five patients who worsened, only one (20%) had a psychotherapist or another mental health treater. Only 9% of the total patient group considered the change in treaters a major disruption in treatment (see +Table 1).

Pooled results for all residents who transferred patients show that 19% of all patients worsened after announcement of the upcoming change, and 30% of all patients considered the change a major disruption of treatment. Fewer than 10% expressed hostility, needed changes in medications (Note: decisions regarding need for medication changes were made by the residents, sometimes on their own and sometimes in consultation with supervisors), or developed new symptoms, and up to 9% became noncompliant (see +Table 1).

Twenty-two residents who received patients at the start of the new academic year were surveyed; of these, 10 residents were beginning their PGY-2 year; 11 residents were beginning their PGY-3 year; and 1 resident was beginning his PGY-4 year. These residents received a total of 180 patients. Residents were asked to assess the course of their patients 4 months after the change in treaters. The results discussed here are pooled for all residents, given that the circumstances of their receiving patients were essentially the same for all classes. These residents reported that 65% of the patients remained unchanged after the transfer; 18% had improved by 4 months; and 10% had worsened by 4 months; of the 18 patients who worsened, only 9 (50%) had psychotherapists or other mental health treaters. Overall, 28% of patients considered the change in treaters a major disruption in treatment. Also, 32% of all patients required medication changes after the change in treaters; 12% of all patients expressed hostility; 7% of all patients developed new symptoms; and up to 12% became noncompliant (see +Table 1). These findings suggest that transferred patients had a more difficult time after the transfer than before. Consequently, the residents who received patients seemed to have more difficulty managing the patients after the change than the residents who had transferred them.

We found that 62/185 patients (34%) were in psychotherapy during the transition time. Of these, 13/62 (21%) worsened during this period, and of the 123 who received no therapy, 23 (19%) worsened. Chi-square analysis demonstrated no statistically significant difference (χ²=0.135; NS).

The following three vignettes serve to illustrate the effect of termination on psychopharmacology patients:

It is generally agreed that termination, the end-phase of treatment, has qualitative differences from the rest of treatment (+14), and this phase has been the subject of much analysis in the psychotherapy literature. As mentioned earlier, transference and countertransference issues can arise in such situations (+15), and the various therapist and patient reactions seen in the context of change in treaters have been described as a "transfer syndrome" (+17).

It is worthwhile to ask why there have been no published studies of termination with psychopharmacology patients and why no residency programs, to our knowledge, include the topic of termination with psychopharmacology patients in the curriculum or in supervision. One reason may be limited supervision; residents tend to work more independently with psychopharmacology patients, and supervision tends to focus on diagnosis and treatment rather than on process. A typical psychopharmacology supervision session may occur as infrequently as once per month and may cover several patients, as opposed to psychotherapy supervision, which tends to focus on one patient at a time. These constraints may make it difficult for the resident and supervisor to explore termination with psychopharmacology patients and may result in avoidance of the subject altogether.

The nature of psychopharmacologic treatment may also contribute to the problem. There often are long intervals between patient visits, which may limit a thorough exploration of termination issues as well as appropriate length of notice for the patient. The time constraints during the shorter visits, which often run between 10 and 30 minutes, also contribute to the problem. Some residents may also assume that patients—particularly those who have been treated in teaching centers for many years and have changed treaters multiple times—"get used to" these changes. Finally, psychiatric residents, and psychopharmacologists in general, may underestimate the importance of the psychopharmacologist—patient relationship and the strength of attachment of psychopharmacology patients, believing that attachment occurs only in the psychotherapy setting. In reality, the attachment to the psychopharmacologist may be just as strong, but perhaps subject to less distortion than may be seen in psychotherapy and less overtly expressed.

In view of the above, we were interested in whether the observations made in the psychotherapy literature bore any similarity to our residents' experience with psychopharmacology patients. Our pilot study demonstrates that although the majority of patients manage to weather the change in psychopharmacologists well, a significant number considered the change a major disruption in treatment, and many worsened or required medication changes during the transition period.

Because we did not examine a comparison population (patients not undergoing changes in treaters), it is difficult to say how many of the patients from our survey would have worsened or required changes in medications had they not been subjected to a forced termination of treatment. This question will be addressed in future studies with larger populations, which may allow for an effective comparison of outcomes between patients undergoing change in treaters and those not undergoing change in treaters. The goal of this pilot study was to obtain the perceptions of residents involved in the exchange regarding the course of their patients, and the observations suggest that further research may be beneficial.

Another limitation of the study is the lack of direct inquiry—systematic assessment—of patients. In our survey, we have relied exclusively on the clinical impressions of residents, which, admittedly, may be biased and subject to countertransference feelings. Future studies will directly inquire of patients their impressions and reactions to change in treaters.

Given the fact that the patient population as a whole appeared to worsen after the transfer of care, our data must also be interpreted in light of the fact that patients were transferred from more-experienced to less-experienced residents. However, we do not believe that resident inexperience, in itself, would explain any significant worsening in patient condition. Junior/less-experienced residents generally receive more intensive psychopharmacology supervision than senior/more-experienced residents, which is expected to "level the field" vis-à-vis effectiveness of care. The worsening of patients is more likely due to other factors, such as delays in follow-up with the new treaters (which often occurs during the summer months), the general discomfort involved in establishing a relationship with a new treater, and the mourning of the loss of the past treater.

When residents transfer only a portion of their patient load (as occurs when residents entering the PGY-4 year downsize their patient load), they often choose to transfer their more "difficult" patients (+12), that is, those with severe Axis II issues, who may be especially needy and/or refractory. These patients may not handle changes as well as higher-functioning individuals, and could therefore bias the results of such a study. However, the results of our survey would appear to disagree. Patients selected for transfer by PGY-3 residents entering the PGY-4 year seemed to be less disrupted by the change, even though their time of notice was considerably shorter than that provided by graduating residents. Several explanations are possible. Perhaps these patients were selected for transfer because residents believed they would be better able to manage the change; perhaps they were selected for lack of attachment to the treater, in which case, the loss might not seem as great. The shorter duration of treatment, when compared with patients terminating with graduating residents, may also ameliorate the change, as there may be less emotional investment on the patient's part. Indeed, outcomes appeared worse for patients transferred by graduating residents, which suggests that a longer course of treatment may result in feelings of a greater loss when the treatment ends.

The relatively small number of patients in psychotherapy raised the question of whether psychotherapy might ameliorate feelings of loss in the setting of forced termination. We expected that having a therapist might facilitate the change in psychopharmacologists, but our results suggest that the presence of therapy did not significantly influence patient outcomes. It is possible that patients who were in therapy during the change in psychopharmacologists did not address their concerns about it during their therapy sessions, but rather continued to focus on the regular topics of their therapy. It is also possible that therapy might result in (or predispose to) more exploration of the feelings around the loss and hence worsen the reaction to the loss. Further investigation will be necessary to clarify this question.

Forced termination with psychopharmacology patients may result in issues similar to those experienced at termination in psychotherapy. Almost one-third of all patients in this sample appeared to be negatively affected by termination. We base this conclusion on the need for medication changes and the fact that many patients regarded the change as a major disruption in treatment. The "termination triad" of hostility, regression, and new material was most noted for regression in the psychopharmacology setting, and less for hostility and new material. Although most patients managed to weather the change, there clearly is room for improvement. Further investigation is necessary to better define and quantify the negative impact of termination on psychopharmacology patients, to help identify patient populations at highest risk for worsening, and to determine who may benefit from which interventions.

On the basis of the authors' experiences and previous reports (+2,+18,+19), we recommend the following interventions for ameliorating the change of psychopharmacologists:

We hope that the guidelines proposed in this article will be of use not only to psychiatry residents, but also to psychopharmacologists who practice in clinic settings that may be prone to high physician turnover.

The authors thank Michael Otto, Ph.D., for his assistance with statistical procedures.

This work was presented in poster form at the American Psychiatric Association Annual Meeting, Toronto, Ontario, Canada, June 1, 1998.