LONGITUDINAL STUDY
Rosemary Cogan and John H. Porcerelli
Rosemary Cogan, Ph.D., retired recently after many years in the Department of Psychological Sciences at Texas Tech University where she continues as a professor emeritus. She retired from clinical practice several years ago.
John H. Porcerelli, Ph.D., is on the faculty in the Department of Psychology at the University of Detroit–Mercy and is in private practice in Birmingham, Michigan.
There is a robust and interesting literature on psychoanalytic outcomes. Many readers will be familiar with the Psychotherapy Research Project at the Menninger where changes in 42 patients at Menninger Hospital, 22 in psychoanalysis and 20 in psychotherapy, were compared in a series of studies. Other readers may know about the more recent European studies such as those in Heidelberg-Berlin, Helsinki, Munich, the Netherlands and Stockholm. Although some of the very early studies involve only reports of the treating analyst, most have involved reports by or interviews of analysands as well.
We honored this important literature, but we recognized also that very little of the research is longitudinal. We wanted to know two kinds of information. First, what are different types of analytic outcomes really like? For instance, what exactly do analysts mean when they talk about analyses ending with maximum benefits? Second, what at the beginning of analysis predicts the outcome of analysis? It seemed to us this would be extremely helpful in recognizing problems before they became full-blown.
Our interest as psychoanalyst-researchers was truly piqued when we began to read of the work of Drew Westen and Jonathan Shedler who developed a remarkable clinician report measure of the personality and functioning of clinical patients. Shedler and Westen noted that clinicians are highly trained to observe and consider what happens in the clinical situation. They spent some years developing a measure. The Shedler-Westen Assessment Procedure (SWAP) includes 200 items that might describe a patient. The items are not theoretically oriented and are written in sensible language. The treating clinician sorts these items into eight categories ranging from those that describe the patient very well to those which do not apply to the patient or about which the clinician has no information. The items must be sorted into a fixed number in each category, ranging from eight items that describe the patient very well to 100 that do not apply. The SWAP has been used in hundreds of studies of psychotherapy and psychoanalysis.
Here we plan to introduce our longitudinal study of change during psychoanalysis using the SWAP-200. Our intent is not to provide a detailed review of the literature or an exhaustive report of our methods, statistical work or results. These are available in our book describing the complete project: Psychoanalytic Treatment in Adults: A Longitudinal Study of Change, published in 2016 by Routledge. We want to share with you some of our findings.
Our Background
Researchers and clinicians, we met during a training year at the Detroit Psychiatric Institute, a Michigan public hospital now closed. As our psychoanalytic training progressed in Dallas and Michigan respectively, we continued to work on research together. We discovered and read with enthusiasm the early papers about the development of the clinician procedure for measuring personality, the Shedler-Westen Assessment Procedure, reported in papers in 1998 and 1999. From the analyst’s sorting of the 200 items, scales could be scored. “Pictures” of the patient could be put together from the items directly and from the scoring of SWAP scales, which include adaptive functioning, personality and trait scales. We were fascinated and carried out several studies to see how well the results of the SWAP fit areas of our separate clinical work and our collaborative research. We set to work and organized a longitudinal study of change during psychoanalysis. Shedler and Westen were very helpful. The International Psychoanalytic Association funded the costs of the beginning of the work.
The Research Plan
We put short recruiting posts on APsaA’s listserv once a month for six months, inviting analysts who were beginning an analysis with a new patient to participate by completing the SWAP-200 and a short questionnaire every six months from the beginning to the end of the analysis. The questionnaires included an item on which the analyst, at the end of the analysis, indicated the outcome of the analysis. After a dozen years, we had data from 60 completed analyses. Our heartfelt thanks go to the analysts who so patiently completed this work over quite a few years.
We began by considering the characteristics of patients and the characteristics of the participating analysts as the analyses began. As the analyses ended, the analysts responded to a question about the nature of the outcome. We grouped the information into five outcome groups: negative therapeutic reactions, dropouts, analyses ending because of external factors, and analyses ending by mutual agreement between analyst and patient either with or without maximum benefit.
Major Findings
Below are some of our findings in three areas: Prevalence of the five outcome groups and characteristics of each outcome group, characteristics at the beginning of analysis that were predictors of the five outcomes, and the role of insight in changes during analysis.
We learned that, interestingly, in each case, our findings were similar to the relevant reports of the outcomes of psychoanalysis and psychotherapy in the published literature.
Patients whose analyses ended with maximum benefits had lower scores than others on the Paranoid Personality Disorder SWAP scale and higher scores on the Global Assessment of Functioning and Insight scales.
At the beginning of each analysis:
Analyses ending with maximum benefit ranged from 28 to 121 months. We invite the reader to pause for a moment and consider the problems this range presented for considering the role of insight in change statistically. In reality, psychoanalysis and psychotherapy really do range in length. We believe that artificial constraints on the duration of treatment, quite common in some types of outcome research, can seriously distort research conclusions. We consulted with an applied mathematician, N.G. Cogan, who is quite accustomed to dealing with relevant research issues. He developed a mathematical model of changes in the SWAP Insight scale and changes in Psychological Health, Paranoid Personality Disorder and Dysphoria. Two useful conclusions came from this. First, the model showed that Insight changed before, for instance, changes in Dysphoria. Second, the model also indicated that there are three phases of analysis: a short early phase involving about 10 percent of the duration of analysis, a long mid-phase, and an end phase involving about 25 percent of the time in analysis. He also included material about finding and working with an applied mathematician likely to be helpful. We are glad of and grateful for this collaboration.
Finally, we have joked—wishfully—about replicating this large study. More seriously, we would encourage working groups to consider what might be practical. Group research has real challenges. Perhaps at a practical level, a training center, might arrange to have candidates and perhaps other members use the SWAP to describe patients beginning analysis and then record the outcome group at the end of analysis. A good number of our participants, all analysts, told us they found the SWAP interesting and helpful in their thinking about their patients. We would be happy to visit with any readers about possible projects and share our questionnaires.