Empirical Evidence and Psychotherapy: A Growing Scientific Base

We live in a society enamored of technology. Health care is delivered in the context of that society, of course, and reimbursement patterns for treatment reflect the bias towards “high-tech” medicine. It is not uncommon for insurance policies to pay in six figures for organ transplants but offer only nickels and dimes for psychotherapy. The idea that something therapeutic may occur when someone talks to a trained professional has always been controversial. In some quarters, psychotherapy is viewed as being akin to hand-holding.

To be sure, practitioners of psychotherapy have been slow to use rigorous empirical methods to demonstrate the usefulness of what they do. At long last, however, research data have been accumulating that suggest psychotherapy produces lasting improvements and even changes brain functioning (1–3). Because the mind is inextricably connected to the brain, these findings should not come as a surprise. Indeed, it is a remnant of persistent Cartesian thinking that has led many skeptics to think that psychotherapy may be nothing more than balm for the “worried well.” A 1999 report by Bateman and Fonagy (4) helped to put that view to rest by pointing out that even with seriously disturbed patients, such as those suffering from borderline personality disorder, psychoanalytically oriented psychotherapy in the context of a partial hospital program could effect far-reaching changes in suicidality, acts of self-harm, the need for hospitalization, and a host of psychiatric symptoms, including depression and anxiety.

In this issue of the Journal the same investigators report on their sample of 44 patients with borderline personality disorder after 18 months of follow-up. They demonstrate that the significant clinical improvement resulting from the original 18-month psychoanalytically oriented partial hospital program was maintained at follow-up a year and a half later. While this finding is encouraging, the news is even better. They also found that the patients who received the treatment continued to improve in measures of symptom distress, major clinical difficulties, need for hospitalization, and self-harm. The psychotherapy literature has suggested that psychoanalytic/psychodynamic treatments might be associated with continued improvement after termination of the therapy (5), and the Bateman and Fonagy study adds more evidence to that growing impression. However, even though the partial hospital program had ended, the patients who had been randomly assigned to the partial hospital group continued in psychoanalytic group therapy twice a week, so the impact of that treatment must be considered in terms of assessing the continued improvement.

It is noteworthy that while the investigators employed a rigorous experimental design, including random assignment to either the experimental group or the control group, they stressed that their study was an effectiveness trial rather than an efficacy trial. In recent years a backlash has developed against tightly controlled efficacy studies of psychotherapy in academic settings (6–10). Criticisms of this approach have included the fact that 80% to 90% of subjects are excluded, patients who respond to advertising may be different from patients seen in naturalistic settings, patients with complicated comorbidity are not included, and generalizability from the artificial research setting to the “real world” may be limited. To address these concerns, NIMH has specifically called for more effectiveness research (11). This contribution by Bateman and Fonagy is a model of effectiveness research, as ordinary clinical referrals were the subjects of the study, a bare minimum of exclusion criteria were used, and the psychotherapy was conducted in a naturalistic setting.

Psychotherapy is often thought to be prohibitively expensive. Insurance companies have had a long history of discriminating against the use of psychotherapy as a treatment for fear that it will “break the bank.” Although strict cost-effectiveness measures were not employed in the research design, the study by Bateman and Fonagy provides provocative data suggesting that in some cases we may actually save money by providing effective psychotherapy. If they are untreated, patients with borderline personality disorder often utilize expensive health care resources, such as inpatient treatment, intensive care, and emergency room services. The patients participating in the study intervention had fewer episodes of self-harm and fewer suicide attempts. They also dramatically decreased their use of inpatient treatment, while the control group used more of all types of services than the treatment group, supporting the view that psychotherapy may be highly cost-effective in treating patients with borderline personality disorder (12). If cost-effectiveness measures were systematically incorporated into future psychotherapy studies, the field might reach a point where it would be possible to identify which subgroups of patients warrant the investment of time, energy, and money associated with extended psychotherapy.

A second report in this issue of the Journal also makes admirable use of a naturalistic setting to study a complex aspect of the patient-provider relationship in the treatment of diabetes. Ciechanowski et al. used an ingeniously creative approach to examine the nemesis of both medical and psychiatric practice—noncompliance. It has long been known that lack of adherence to diabetic self-management regimens is linked to serious complications of diabetes, including peripheral vascular disease, renal involvement, retinopathy, and neuropathy. The investigators applied adult attachment theory in an attempt to better understand the anatomy of this noncompliance. They found that patients with a dismissing attachment style had significantly higher levels of glycosylated hemoglobin. In addition, among patients with a dismissing attachment style, those who perceived that they had poor communication with their health care provider had higher levels compared to those who felt the quality of communication was good.

Attachment theory, which is finally getting the attention it deserves in our field, is based on the seminal work of John Bowlby, who observed that internal working models are developed in every individual based on early experiences with parents and/or caregivers. These models are transposed onto adult relationships and profoundly affect how one relates to others throughout life (the basis of transference in psychoanalytic psychotherapy). Adults who have a dismissing attachment style generally have experienced caregivers or parents as consistently emotionally unresponsive. They become compulsively self-reliant as a result and try to avoid the kind of collaborative relationship necessary for treatment.

This contribution underscores the point that psychotherapeutic principles are essential in all provider-patient relationships. Krupnick et al. (13) demonstrated that in depressed patients, the therapeutic alliance is just as important in pharmacotherapy as it is in psychotherapy. Transference-based perceptions of the provider based on the internalized working models from early experience may impair the capacity of a patient to collaborate with a provider and participate in the therapeutic alliance. By making a special effort to reach such patients through sensitive and empathic communication, health care providers may help improve treatment adherence.

Ciechanowski et al. have opened the door to a fertile field for research. They are currently collecting data on the attachment style of treaters as well as patients. Such research may teach us sophisticated methods that allow us to assign patients with adherence problems to treaters who have a specific complementary attachment style. This research holds out the possibility of shedding light on the elusive construct of “therapist-patient match.”

Finally, I have long thought that in training psychiatrists, we make too much of a distinction between psychotherapy and pharmacotherapy. The doctor-patient relationship is inherently psychotherapeutic. Careful attention to concepts such as therapeutic alliance, doctor-patient collaboration, transference, and countertransference apply across the board in all of our interactions with patients. Even when we treat the brain with somatic treatments, we cannot bypass the mind.

Address reprint requests to Dr. Gabbard, The Menninger Clinic, P. O. Box 829, Topeka, KS 66601-0829; [email protected] (e-mail).