Studying Change in Defensive Functioning in Psychotherapy Using the Defense Mechanism Rating Scales: Four Hypotheses, Four Cases

Abstract

Defense mechanisms are one of the original and the most durable theoretical contributions of psychoanalysis to dynamic psychology. Research has shown that there is a hierarchy of the general level of adaptation of defenses, divided into seven levels, which can be summarized as the level of overall defensive functioning (ODF). This chapter examines how the quantitative assessment of defense mechanisms can yield indicators of the progress and outcome in psychotherapy research. Four cases are presented with short- to long-term psychotherapy and one with very long follow-up. Each demonstrates how different aspects of defensive functioning change over different time periods and states (i.e., depressed vs. not depressed) exemplifying four hypotheses about how defenses change. The first is that as individuals change, they increase their overall level of defensive functioning, and at the same time, variability in defensive functioning tends to decrease, indicating increased resilience to stress. The second is that change in defense levels occurs in a stepwise fashion in which individuals trade off defenses lower on the hierarchy for those in the middle and only later developing those at the top of the hierarchy. The third is that individuals and groups have their own rates of change, which may vary across naturalistic and different treatment conditions, yet to be determined. Depressed states may be associated with initially large changes that then decelerate, whereas personality disorders (PDs) may have long initial periods of induction in the therapeutic process (“priming”), before change is initiated and becomes, more or less linear. Treatments that increase this rate of change are likely to be seen as more effective. Finally, in line with most of the research to date, as defensive functioning improves, symptoms will decrease and other aspects of functioning will improve. Although single cases do not prove a hypothesis, these cases offer some empirical support, while clearly demonstrating the value of research in this field. Furthermore, the identification of defenses in verbatim interviews and psychotherapy sessions permits the moment-to-moment analysis of the apparent effect of interventions on defensive functioning, also a topic worthy of further research.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

One of the earliest observations at the inception of psychoanalysis was that defense mechanisms inhibit, manage, and sometimes redirect the expression of motives and affects in symbolically meaningful ways when psychological conflicts are triggered. Beginning with the publication of The neuropsychoses of defence, Freud [1] described repression, reaction formation, displacement, and other defense mechanisms but he never became interested in their systematic study. Later, Anna Freud [2] began to systematize them. In further theoretical development, Waelder [3] posited that defenses could have multiple functions beyond guarding against forbidden wishes, which could include the gratification of wishes. Schafer [4] further expanded this idea suggesting that defenses also seek to increase gratification and to reduce pain, while still defending against forbidden wishes. Hartmann, et al. [5] suggested that some ego functions were neutralized or conflict-free, and not directed against forbidden wishes. They emphasized that defenses and some other ego functions (e.g., reality testing) could be directed either against motives or toward objects and events in the external world to improve the individual’s adaptation to reality factors. Norma Haan [6] extended this view by giving separate roles to defenses that deal with intrapsychic conflict and to related mechanisms that help adapt the individual’s motives to the demands and constraints of the external world. Thus she separated defense from related coping mechanisms. Subsequently, Lazarus and Folkman [7] rejected most of the views of the defense and drive perspective in favor of studying only coping as conscious mechanisms, which individuals employ to deal with external stress. A plethora of research on conscious coping has followed (see review [8]), while research on defense mechanisms has also proceeded at a slightly slower and less well-funded pace (see review [9]).

In psychoanalytic psychology, defense mechanisms are widely viewed as one constituent in the structure underlying personality or character. For instance, in his description of the difference between psychotic, borderline, and neurotic levels of personality organization, Kernberg [10] noted particular defenses along with identity formation and reality testing as the cornerstones of character. Recognizing the advances in assessing defenses (see below), defenses were added in Appendix B of DSM-IV [11] as a dynamic aspect of personality that could be coded separately from PDs. The presumption is that individuals tend to use the same repertoire of defenses, and these constitute the defensive structure of personality. Structural change over the course of treatment should then result in improvement in defensive functioning, i.e., a change in the defense repertoire toward more adaptive defenses. Advances in assessing defenses now allow us to track changes in defenses so that psychotherapists can see the effects of treatment on their patients during and after treatment. This chapter is devoted to this issue.

This chapter will examine one method of rating defense mechanisms, the Defense Mechanism Rating Scales (DMRS), and demonstrate how it may be used in the study of psychotherapy process and outcome. Using a longitudinal perspective on four cases, we will demonstrate how defenses can be viewed as a dynamic outcome variable, in addition to their usual function as mechanisms operating from moment to moment in real time. Although defensive and coping functioning are currently separated, if not divorced, we will also present some early data suggesting that a reconciliation may be possible based on the commonality of overall level of functioning, allowing defense and coping to relate to one another once again, if not actually remarry.

What are Defenses?

As noted elsewhere [9, 12–14], there is an approximate consensus among researchers on the following aspects of the defense mechanism construct.

• 1. A defense mechanism is the individual’s automatic psychological response to internal or external stressors or emotional conflict (see DSM-IV, Appendix B Defensive Functioning Scale, pp. 751–757). The action of a defense is triggered by the occurrence of what Freud [15] called signal anxiety, arising whenever internal wishes or drives conflict with internalized prohibitions or external reality constraints.

• 2. Defenses generally act automatically, without conscious effort. Often, the individual is totally unaware of the defensive operation, although in some instances, he or she may have partial awareness.

• 3. Defenses contribute to character traits that are in part made up of specific defenses that an individual tends to use repetitively in divers situations. Individuals tend to specialize, using a set or repertoire of defenses across various stressors, depending on the motives or conflicts active at the time. The defenses an individual shows at any point in time may vary in degree or specificity with the stressor, making some state effects expectable [16], while the average frequency of using the defense is dispositional. Whether there is specificity between type of stressor and choice of individual defense used is an empirical question.

• 4. One review tallied 42 different individual defense mechanisms described by various authors [17]. Although there is no clear rationale for selecting a definitive list of defenses [18], a process of consensus has favored those defenses with clear, nonoverlapping definitions, reliable application and demonstrated empirical findings. Although this process is advanced, it is not completed.

• 5. Defenses affect adaptation [14, 18]. Each defense may be highly adaptive in certain situations. Nonetheless, there is a clear hierarchy of defenses in relation to the overall adaptiveness of each individual defense. Defenses at the lower end are usually maladaptive, save in a few situations, while those at the higher end are adaptive in a broader array of circumstances.

• 6. When defenses are least adaptive, they protect the individual from awareness of the stressors, anxiety, and/or associated conflicts, but at the price of constricting awareness, freedom to choose ways of responding, and flexibility to maximize positive outcomes. When they are most adaptive, defenses maximize awareness of internal and external motives, stressors and constraints and thereby maximize the expression and gratification of wishes, minimize negative consequences, and enlarge the scope of choices and sense of control. Some individuals label the most adaptive defenses as coping mechanisms, after Haan [6], whereas many in the dynamic tradition retain the defense designation, because the so-called coping mechanisms share many characteristics with other defenses. This question of terminology, defense vs. coping, is a mixture of questions of preference, definition, and science.

• 7. Whether defenses emerge in a certain developmental sequence is an empirically open issue. The usage of developmental terms to describe groups of defenses, such as immature or mature, is done for reasons of history and convenience only. In fact, preverbal toddlers may use high proportions of mature or high adaptive-level defenses [19], which suggests that much more is needed to understand why some adults come to use the lower-level defenses associated with PDs. However, during adulthood, there appear to be sequences in which certain lower-level defenses progress to higher-level defenses on the continuum of adaptiveness [14]. For instance, acting out in early life (e.g., rebelling against authority) may evolve later to reaction formation (taking the side of authority) and eventually to altruism (helping the less powerful obtain fair responses from authority). These sequences deserve study because they may hold important implications for patterns of therapeutic change in personality, both as a patient “trades up” within a therapy session and how he or she develops over time. Our clinical examples address this in particular.

• 8. Identifying and understanding the function of defenses serves as an aid to understanding the problems and therapeutic challenges in treatment. For instance, defenses in the lower half of the defense hierarchy mediate many of the most maladaptive ways of handling stress and conflict in PDs and depression. Specific defense levels may be associated with certain individual disorders or clusters, which relate to core psychopathology, such as splitting and projective identification in borderline personality [10, 20, 21], or omnipotence and devaluation in antisocial and narcissistic personality disorders (NPDs) [22]. Whenever used, these same defenses serve as instantaneous markers alerting the clinician that a core issue for a given person may be operating.

The Hierarchy of Defenses

Defenses have been hierarchically ordered based on the empirical relationship to general measures of adaptiveness or psychological health. A number of studies have led to a generally accepted hierarchy [12, 14, 18, 23, 24]. Examples include the Defensive Functioning Scale in DSM-IV [11, 25], and the DMRS hierarchy (Table 6.1). Defenses that have common aims are grouped together in one of the eight levels. For instance, among the so-called immature defenses, the disavowal level includes three defenses – denial, rationalization, and projection. These have a common function of disavowing certain affects, actions, ideas, or motives, which others can identify in the person using them. However, each defense differs in method of handling the disavowed material. Denial actively avoids it altogether, rationalization avoids it by covering it up with something deemed more socially acceptable, while projection avoids it by misattributing it to others and thereby maintaining an interest in it, but at a distance. The defense levels range from the lowest level of defensive dysregulation (so-called psychotic defenses) to the high adaptive level (so-called mature defenses). Healthier individuals use a larger proportion of highly adaptive defenses and a lower proportion of defenses at the low end of the hierarchy. The DMRS does not yet have a complete section of psychotic-level defenses, such as psychotic denial, distortion, and delusional projection [18], as to date the method has mostly been used on nonpsychotic samples.

Table 6.1 DMRS hierarchy of defense levels and individual defense mechanisms

The Defense Mechanism Rating Scales Quantitative Rating Method

We currently score defenses according to the quantitative directions of the DMRS, fifth edition [26]. The DMRS is a quantitative, observer-rated method [9], which is almost identical to the qualitative Provisional Defense Axis in Appendix B of DSM-IV [11, 27]. Each of 30 defenses is identified whenever it occurs in the session. This method differs from other observer-rated methods that are qualitative or semiquantitative ratings (e.g., most prominent defenses as in Vaillant [14]), which yield global ratings for the whole interview [9].

Three levels of scoring are used, all of which are continuous, ratio scales.

• 1. Individual defense score. A proportional or percentage score is calculated by dividing the number of times each defense was identified by the total instances of all defenses for the session.

• 2. Defense-level score. The defenses are arranged into seven defense levels hierarchically arranged by their general level of adaptiveness. Each defense level is represented by a proportional or percentage score.

• 3. Overall defensive functioning. The ODF score is obtained by taking the average of each defense level score, weighted by its order in the hierarchy, yielding a number between 1 (lowest) and 7 (highest).

In addition, the defense-level scores can be divided into several superordinate levels: mature, neurotic, immature, and psychotic, although in most publications using the DMRS, the fourth is not included.

There is good convergent and discriminant validation for the overall hierarchy vis-a-vis other functioning and symptom measures [18, 24, 27–29]. Interrater reliability can be quite high at the level of identifying the number of defenses and ODF with intraclass R values above 0.80, while the defense levels and individual defenses are generally somewhat lower (see review [30]). Short-term stability was determined by rating five consecutive weekly psychotherapy sessions [16], which indicated that about half the variance is due to trait, while the other half of variance was due to state effects that varied from session to session. The stability of ODF was intraclass R (I _R) = 0.48, while the stabilities of the individual defense levels had a median of 0.47, range 0.08–0.73. By contrast, the number of defenses used per session varied greatly (stability I _R = .18), indicating that the rate of using defenses (e.g. number per 50-min session) is highly state-dependent.

Studies using the DMRS have indicated the range of defensive functioning from healthy-neurotic to personality-disordered. A community sample of mothers with no recent stressors and no psychiatric-treatment history gave Relationship Anecdote Paradigm (RAP) interviews, which were rated for defenses (unpublished data of the first author). Most healthy individuals had ODF scores ranging from 5.0 to 6.4 with a mean above 5.6. Perry and Hoglend [27] found that depressed individuals had a mean ODF of 4.68 at intake which with treatment rose to 5.11, indicating that they returned to a neurotic level of functioning as the episode diminished. The psychotherapy pilot study [16] indicated that PDs had ODF scores below 5.0 with a mean of 4.32 (range 3.31–4.97) in the early months of psychotherapy. Furthermore within the PD group, those with borderline personality disorder (BPD) had significantly lower ODF than those with non-BPD group (4.07 vs. 4.62), although the samples were small. These data indicate that there is good convergence between an ordering of diagnostic groups and a normative group of nonill women and the level of ODF. This suggests that ODF and possibly specific levels may serve as a dynamic measure of adaptive functioning.

Based on previous work examining defenses common in individuals with depression (both major depression and dysthymic disorder), Hoglend and Perry [31] examined a group of eight so-called depressive defenses, a subgroup immature defenses. These include acting out, passive aggression, help-rejecting complaining, splitting of self-images, splitting of others’ images, projective identification, projection, and devaluation. Depressed individuals with major depression who had a greater proportion of these defenses had poorer responses to treatment by 6 months compared to those with fewer of these defenses. This suggested that these defenses may play a particular role in the onset or maintenance of depression, suggesting that they may also serve as a marker of the underlying dynamic vulnerability to depression. This idea is explored in several of our cases below.

Hypotheses for Psychotherapy and Long-Term Change

• 1. Individuals have a general level and range of day-to-day defensive functioning, representing their level of defensive adaptation. Personality disorders in particular have a general level of defensive functioning lower than neurotic and healthy-neurotic individuals with higher reactivity to stressors. Although most individuals with PDs have a defense repertoire that spans the hierarchy of defenses, depending on occasion, salient stressors may interact with or trigger some of their psychological conflicts, which, in turn, pull for defenses from particular levels. For instance, we previously published [16] a case of a man with passive-aggressive PD who in the early years of treatment had a generally high level of action defenses but with substantial session-to-session variability. The spikes upward in using action and other immature defenses were usually related to issues of authority and invalidation of his point of view. However, whenever these issues were absent, the subject functioned more at a neurotic level. After the third year of treatment, not only had ODF begun to rise significantly, but the week-to-week variability decreased. This stabilizing of the defensive structure, or increased resilience in the face of stress, indicated that defensive functioning was becoming less state-dependent and more trait-like, alongside improvement in the general overall level of functioning.

  Thus we hypothesize that improvement in defensive functioning involves both an increase in the overall level and a decrease in the amount of variability, indicating increased resilience to stress.

• 2. Improvement occurs in stages more often than across all levels of defensive functioning. In the above case example, as the patient improved, the action level defenses decreased linearly throughout the treatment, while the minor-image-distorting defenses (e.g., idealization) temporarily increased before beginning to decrease after 3 years. As these lower-level defenses decreased, there was an increased reliance on neurotic-level defenses – repression, reaction formation, and displacement – while the already high level of obsessional defenses remained stable. Finally, there was a nonsignificant increase in high adaptive-level defenses. This case illustrates Vaillant’s finding that as young adults age, they first trade off immature for neurotic-level defenses, and somewhat later they trade off neurotic for mature-level defenses [18].

  Thus we hypothesize that individuals whose defensive functioning involves a large proportion of immature defenses will improve in stepwise fashion with respect to the proportion of defenses at each level. The lowest-level defenses, such as action and major-image-distorting defenses, will begin to decrease earliest, other immature defenses – disavowal and minor-image-distorting defenses – will begin to decrease next – perhaps after a small temporary increase – while the neurotic-level defenses will tend to increase. Finally, after the above improvements are well under way, high adaptive-level defenses will begin to show significant improvement. Although this is the hypothesized complete development cycle, it is of course possible that developmental progression may slow or cease at any point for various reasons.

  Individuals who initially have a very high level of neurotic defenses may follow the same general stepwise movement, although the therapy literature suggests that there may be periods of regression in defensive functioning before there is any significant increase in high adaptive-level defenses.

• 3. Each person and each class of persons (e.g., a specific disorder) may have unique rates of change under natural history or specific treatment conditions. For instance, individuals who are in an episode of major depression will show a different rate of change from ill to remission, which may be faster than from the early period of remission until a subsequent follow-up period. Individuals and disorders can be classed by their rates of change, which may be informative about the type and goals of therapy that may be effective. Additional research may then examine differential predictors of these rates of change, such as childhood neglect or abuse, linking developmental experiences with pathology and the natural course or response to treatment.

• 4. As defensive functioning improves, symptoms decrease and social role adaptation also improves. This hypothesis is a general version of the finding of cross-sectional associations between defensive functioning and measures of adaptation (see review [12]). As a general hypothesis, it should hold across a wide variety of measures of symptoms and functioning, indicative of the general underlying psychological role that defenses play in psychopathology and adaptation. As a result, a period of sustained, healthy, defensive functioning should presage sustained improvement in other aspects of psychological functioning. Whether change in defenses mediate the other improvements, or is merely correlated with them, will require further study.

Case Illustrations

In the remainder of the chapter, we describe four case reports focusing on findings that illustrate how defensive functioning changes over time. These are illustrated graphically by figures of quantitative ratings of dynamic or RAP interviews or psychotherapy sessions.

Case A. A Woman with Recurrent Major Depression

The first case is taken from a pilot study (n = 12) of adults with acute, recurrent major depression who entered a comparative treatment trial of antidepressive medications (ADM) and 20 sessions of either cognitive–behavioral therapy or dynamic psychotherapy. The aim of the study was to estimate the proportion of subjects who attained a full recovery from depression, defined as greater than 8 weeks with a Hamilton Rating Scale for Depression-17 (HRSD-17) score below 6, and remained well at 1-year follow-up. A related aim was to examine whether potential psychological mediators of depression – especially defenses – would normalize by termination and remain at healthy levels at 1 year. This explicitly examined whether initially elevated depressive defenses would decrease to below a healthy cutoff (< about 7%), and ODF would reach healthy-neurotic levels (> about 5.64). If short-term psychotherapy and ADM were generally curative, then the median figures for the sample should lie in the recovered ranges at termination and/or follow-up.

The patient was a 29-year-old married special education teacher who sought treatment for a depression following a miscarriage. She felt deeply ashamed of being unable to control her feelings, of which disappointment was the most salient. She did not express anger over the miscarriage but expressed fear that she was not going to be able to have a baby, possibly due to a previous abortion. She had been an only child, often lonely, for whom self-sufficiency and not being a bother to her parents became a hallmark. She had married a very supportive and nice man whom she initially had not been attracted to but came to like as she realized how good he made her feel. After joining the research project, she was randomly assigned to 20 sessions with a senior cognitive–behavioral therapy (CBT) therapist. This therapy has been described in detail elsewhere [32] and was scored on the Analytic Process Scales. Although working in the cognitive–behavioral model, the treatment addressed the patient’s reluctance to express her own feelings, clarified some of them, and helped her explore her own emotional life as well as think through her responses. Treatment was considered very successful and the patient returned to work, became pregnant, and worked through her fear that the pregnancy would end badly. Her HRSD-17 score dropped from 24 at intake to 9 (mild depression) at termination and 6 (the cutoff for nondepressed) at 1-year follow-up. Similarly the Beck Depression Inventory (BDI) went from 12 (above the depressed range) to 7 (nondepressed) to 0 at 1 year.

This case was selected because her change in defensive functioning was close to the median of the sample, hence illustrative of the average response of defensive functioning to a 20-session treatment. The patient’s dynamic interviews at intake, termination, and 1-year follow-up were scored for defenses, blind to time, and other data. At intake, the patient had 17% immature defenses of which the majority, 9%, were the eight so-called depressive defenses. Her neurotic defenses constituted the bulk of her defensive functioning at 78%, while high adaptive-level defenses were 6%. Illustrated in Fig. 6.1, her depressive defenses dipped to 2% at termination, then rose somewhat to 5% by 1 year. At the same time, neurotic defenses rose to 89% at termination but at one year follow-up they decreased to 78%, about where they started at intake. Finally, Fig. 6.1 also shows that her high adaptive (mature) defenses were largely unchanged, at 7%, at termination, but rose somewhat to 12% by 1 year. Figure 6.2 displays ODF across the three time periods. She began with an ODF of 4.97 in the depressive range just below the neurotic range. It rose to 5.33 at termination and to 5.38 by 1 year, both in the middle of the neurotic range. This constituted a within-condition effect size (ES) of 1.13 equal to the mean and close to the median for the sample (ES = 0.99)

Fig. 6.1

Depressive defenses decrease to normative levels after 20 sessions, regressing slightly at 1 year. High adaptive (mature)-level defenses begin to improve only after termination at the 1-year follow-up, attaining about one-third of normative levels

Fig. 6.2

After 20 sessions of treatment, overall defensive functioning (ODF) rises from depressive to neurotic levels, shy of healthy-neurotic levels at 1-year follow-up

This case improved in line with the above hypotheses. She began in the depressive range and moved to the neurotic range, trading off depressive for neurotic-level defenses. Only at 1-year did she begin to increase her high adaptive defenses somewhat, at the expense of a decrease in her neurotic-level defenses. Most of her positive changes occurred while she was in treatment and were concurrent with the large decrease in depressive symptoms. The exception was the increase in high adaptive-level defenses at 1 year, which coincided with a further reduction in depressive symptoms. From a defense point of view, the patient was characterologically mildly depressive (depressive defenses) and more strongly obsessional and counter-dependent (reaction formation, displacement, undoing, intellectualization, and isolation). At termination and 1 year, the neurotic defense constellation remained the largest part of her defense repertoire. This is consistent with the conclusion that treatment was very successful for treating the depressive part of her character but, in all likelihood, returned her to the status quo ante of those neurotic-level characterological defenses in place prior to the depressive episode. In the time frame of 1 year, she improved meaningfully but, as was generally true for the sample, she did not attain a healthy-neurotic level of functioning.

Case B. A Two-and-Half-Year Therapy Episode in a Patient with Borderline Personality Disorder–Narcissistic Personality Disorder

In this case, we will fully describe the changes in defensive functioning across the hierarchy. This was a 24-year-old single woman who was referred after terminating a therapy during which she had become very regressed, suicidal, accompanied by acting out, which required extensive use of hospitalization. At that termination, she sent a poisonous object in the mail to her former therapist’s office and was charged with assault. Following this event and subsequent legal proceedings, she was referred elsewhere. By agreement she was seen once weekly in dynamic psychotherapy for a little over 2 years, until the therapist had to terminate due to a change in job. She also had adjunctive CBT during the first year only. At termination, she was referred for further dynamic psychotherapy to another therapist. At intake, her Axis I diagnoses included dysthymic disorder, with a past history of bulimia, while on Axis II she had both BPD and NPD. She was on no medication.

This case was chosen to demonstrate the process of change in a patient with BPD and NPD who, while starting with a high proportion of immature defenses, had strong motivation to get better and then made great strides over several years. At the outset of the current therapy, the patient was highly motivated to avoid repeating her previous regressive experience, to renew her studies, to live in her own apartment, and eventually to be able to tolerate close relationships. There had been a number of experiences in her personal history, which she was aware were important, but she strongly preferred keeping the initial focus on staying in school and avoiding any further hospitalizations.

In her first year of therapy, she attended regularly. The focus was on external current issues, and transference was addressed more to manage it, and understand its immediate impact, than to explore it at deeper genetic levels. There were a number of regressive episodes in the first year, such as a disappointment in an interaction with a teacher following a test. Typically, these disappointments were dealt with by acting out (e.g., getting drunk), passive aggression (e.g., staying in her car overnight at school), projective identification (e.g., blaming the teacher and seeing him as the cause of her actions), and splitting (e.g., seeing herself as all bad when subsequently discussing it). She responded quite well when the therapist and she agreed upon limits to regressive wishes, which sometimes included the threat of increasing the frequency of sessions! Therapy focused on helping her trade up to utilize other defenses already in her repertoire, such as rationalization (e.g., “the test was hard and no one did well”) or omnipotence (e.g., “I think I can do better than the others in the class, anyway”). The result was that there were quite wide swings in defensive functioning from session to session in the first year. These are illustrated in Figs 6.3–6.5.

Fig. 6.3

High levels of both action and major image-distorting (borderline)-level defenses decrease steadily over 79 sessions (2+ years), attaining levels consistent with no longer having definite borderline personality disorder (BPD)

Fig. 6.4

Disavowal-level defenses (e.g., rationalization) increase slightly, while minor image-distorting (narcissistic) defenses increase, then begin to decrease over 79 sessions. As BPD diminishes, the patient relies temporarily on defenses associated with narcissistic personality disorder (NPD)

Fig. 6.5

Over the 79 sessions, hysterical (mostly repression) and other neurotic defenses (displacement and reaction formation) increase, reflecting a shift toward neurotic away from personality-disordered-level functioning

For each defense level, we ran simple linear regression models with time (session number) as the independent variable. Table 6.2 shows the results of individual linear regression models of her defense-level scores over times. Her action and major image-distorting (borderline) defense levels were initially around 13–14% each of her total defensive functioning, while both decreased highly significantly by the end of the 2.5 years of treatment, each ending at less than 2% of defensive functioning (also see Fig. 6.3). The other two immature defense levels, next highest on the defense hierarchy, behaved slightly differently. Disavowal defenses were almost one-quarter of her defense repertoire and increased nonsignificantly, while minor image-distorting (narcissistic) defenses initially increased then decreased (nonsignificantly) toward the end of the second year (see Fig. 6.4). Meanwhile the subject showed increases in three neurotic defense levels, significantly with displacement/reaction formation and nonsignificant trends with repression/dissociation (hysterical) and obsessional defenses (see Fig. 6.5). Finally, the correlated changes in the tripartite defense levels are shown in Fig. 6.6. As the immature defenses decreased highly significantly, the neurotic defenses increased significantly, while the high adaptive (mature) defense level increased nonsignificantly. As a result, ODF improved highly significantly, beginning at an ODF estimated by the model at 3.55, increasing 1.09 points to 4.64 estimated at session 79. In addition, the variability in ODF, as represented by the range, decreased from a high of 1.21 points after intake to 1.00 after 1 year and 0.72 points after the second year. The final result was that by the foreshortened end of the treatment, the subject had reached a level of ODF consistent with narcissistic personality, still somewhat below the level of neurotic character functioning. This indicates that the goal of stabilizing the patient had been achieved with an extensive reduction in the lowest-level defenses, trading up to higher immature and neurotic levels, but not yet to high adaptive defenses within the time frame of the therapy.

Fig. 6.6

With all defense levels divided into three categories, as immature defenses decrease, neurotic defenses increase at very similar rates, while high adaptive (mature) defenses increase more slowly

Table 6.2 Case A: change in defense levels and ODF over 79 sessions (2 years)

These results are consistent with the hypotheses. As the subject improved, her ODF increased and the session-to-session variability in ODF decreased. The patient had consolidated her defensive functioning to a level common with narcissistic personality but no longer borderline. Further work was still necessary to develop healthy-neurotic functioning. As to her life goals, by the end of therapy, she had completed her program of studies and was contemplating her next career move. She had successfully kept a job and had begun a somewhat superficial relationship with a boyfriend, albeit one whom she found somewhat disappointing. She did not have any significant regressions in the year prior to termination, nor had there been any emergency room visits or hospitalizations over the treatment course. She was uncertain as to her next career move, as she was concerned about parental pressures on her to follow a certain career path. No regression in functioning occurred at termination as she transferred treatment to a subsequent therapist. Thus she exemplified the third hypothesis as well that improvement in ODF is associated with improvement in other measures of functioning and social role adaptation.

Case C. A Woman Entering Residential Treatment and Followed for 13.25 Years

This woman in her early twenties entered residential, intensive dynamic treatment at the Austen Riggs Center following a series of life stressors leading to a suicide attempt. She had recently married someone about whom she was highly ambivalent while on the rebound from a previous relationship. She was working and going to college part-time and became highly symptomatic. On Axis I she had major depressive disorder, dysthymic disorder, a recent history of cocaine and alcohol abuse, panic disorder with agoraphobia, and a history of bulimia nervosa. On Axis II she had borderline, dependent, and self-defeating PDs. She had been in psychotherapy twice weekly with a very supportive therapist for several years prior to hospitalization with whom she continued after discharge. She had a series of three hospitalizations over 3 years, each precipitated by suicidal episodes and intense guilt over wishes to become more independent and to resist familial pressure to make her marriage work and remain a dutiful daughter. A breakthrough came during her third hospitalization, one result of which was a decision to end the marriage, which had remained unconsummated. Following this, her regressive episodes ceased and she resumed her college career path, steering clear of intimate relationships for a period of several years until she established her independent adult identity more. She continued to see her therapist, gradually cutting down to once weekly, then every other week by about 7 years for several years longer. She obtained a good position in her chosen career and began to date, eventually marrying and having a child. By the 14th year of follow-up, she was recovered from all of her Axis I disorders and was functioning at a healthy adult level by global assessment of functioning (GAF) (>71) and in other social roles as well.

This case was selected because it demonstrates two points: how long it can take until someone with treatment-refractory disorders develops a steady rate of change in defensive functioning and how long it takes to develop healthy-neurotic functioning. Over the 13.25 years of follow-up, her defenses were assessed at 17 points in time from 7 dynamic and 10 RAP interviews. Given the length of follow-up, we explored the question of determining the time at which her long-term trends in defensive functioning became evident or stabilized. To do this, we present four figures, each of which displays the linear trends for immature, neurotic, and high adaptive-level defenses using observations up to four increasing lengths of follow-up: 3,5, 7, and the total 13.25 years. For ease of visual comparison across all four figures, the linear trends of the first three figures are extrapolated out to 14 years.

Figure 6.7 examines the change in defenses over the first 3 years of follow-up, prior to her decision to end her marriage. That this was a highly turbulent period is demonstrated by the increasing levels of immature defenses and accompanying wide variation. Neurotic levels actually decreased, while, unexpectedly, high adaptive-level defenses increased. The intercepts reflect the modeled level of defenses at intake, along with the actual change rates, the change in percentage of overall functioning per year. These are displayed below the graphic presentations for each tripartite defense level. Figure 6.8 displays the trends based on the first 5 years of data, including the early period of stabilization after deciding to end her marriage. By 5 years, different trends have emerged. Immature defenses decreased by 3% per year, high adaptive defenses increased at a slightly slower rate, and neurotic defenses increased very slowly (0.25% per year). Figure 6.9 indicates that the trends are similar at 7 years with the trends for high adaptive and immature defenses in the same direction but at slightly steeper rates than at 5 years. However, the direction of change for the neurotic-level defenses shifted toward decreasing, again at a slow rate (–0.35% per year). Finally, Fig. 6.10 displays the data using all 13.25 years. Over this long time frame, immature defenses have decreased to less than 15% of total defensive functioning, while neurotic defenses have increased to almost 40% and mature defense to almost 50% of total defensive functioning. Her ODF – not shown – was scored at 4.38 and 4.53 in her first two interviews – consistent with severe BPD – yet were scored at 6.00 and 5.68 in her last two interviews, indicating that she had achieved a healthy neurotic level of functioning. A simple linear regression model using all observations over the 13+ years estimated that her initial ODF was 4.41, significantly increasing 1.65 points by the end of follow-up to an estimated final ODF of 6.06 (df = 1.15, F = 20.01, p = 0.0004). Variability was much higher during the first half of the follow-up, intake to 6.5 years, than in the second half of the follow-up, intake to 6.5–13.25 years (ODF range 3.79 vs. 1.62 points).

Fig. 6.7

After 3 years, the patient appears somewhat unstable and regressed, with immature defenses increasing and neurotic level defenses decreasing. However, high adaptive-level defenses increase

Fig. 6.8

After 5 years of follow-up, immature defenses decrease, predicted to reach normative levels by 12 years, while neurotic and high adaptive defenses increase

Fig. 6.9

After 7 years, immature defenses decrease at a rate established by 5 years, while neurotic defenses decrease slightly. High adaptive-level defenses increase still slightly faster than at 5 years

Fig. 6.10

By over 13 years, immature, neurotic, and high adaptive-level defenses have reached normative levels, having continued on trajectories established by 5 years

The graphical presentations indicate clearly that as time progressed, the subject’s variability tended to decrease leading to relatively stable trends by about 5 years. Graphically this is evident by the similarities of the trend lines at 5, 7, and 13.25 years. Her ODF also clearly was increasing by this time and continued to increase relatively linearly by the end of the 13.25 years. These data are consistent with our first hypothesis. Also, she developed solid healthy functioning in other areas of her life, including work and close relationships, consistent with our third hypothesis. Finally, she did develop rather consistent rates of change between 3 and 5 years, consistent with expectations of individuals as noted in the fourth hypothesis.

The patient deviates somewhat, however, from our second hypothesis. During the first 3 years, the patient had unstable trends for both immature- and neurotic-level defenses, with only the high adaptive-level defenses showing a trend that would be upheld across the four time periods. Thus at 3 years, the patient began improving first in the one area, which we hypothesized would be the last to show improvement. By 5 years, the improvement pattern was largely in line with our second hypothesis, although improvement was greater in high adaptive than in neurotic levels, which is not typical. These trends remained more or less throughout the remainder of follow-up. Our explanation is post hoc but several things are evidently different in this patient’s treatment than in our other cases. First, she was well into an ongoing intensive outpatient treatment prior to her first hospitalization. Second, during the first 3 years, she had three hospitalizations of several months duration, each of which included intensive psychotherapy (four sessions per week) with the same therapist at the Austen Riggs Center. Third, her turbulence decreased when she made her important life decision to end her marriage and to strike a more independent position as an adult. We believe that we captured this patient with two distinct trends going on in parallel. The first trend reflects the common experience of the borderline patient in repeated crises related to anaclitic close relationships, which took between 3 and 5 years to stabilize. The second is the self-reflective process, which she was already engaged in and intensified periodically during her residential stays. Thus we feel she was developing insight and some related healthy defenses all along, necessary for her to make her life-altering decision. Had she begun treatment only at intake into the study, it is possible that she might have taken a more typical course, perhaps not showing significant changes in high adaptive-level defenses until some time after her immature defenses began to decrease, but this cannot be known for sure.

Case D. A woman treated for recurrent depression with 18 months of psychotherapy

This case is that of a 38-year-old woman with one child and a partner of many years, who sought treatment for an episode of recurrent major depression on top of dysthymic disorder. She had some depressive and self-defeating personality traits but did not meet full criteria for a PD. Although she had her first depressive episode in her late teens during university, she had remained largely free of episodes until the previous year during which she experienced a number of serious stressful events in her work and family life, including several serious illnesses of relatives. Everyone who knew her saw her as competent and noncomplaining. Although she handled the recent stressors well from an external point of view, the problems left her with a feeling that she was holding too much in and a sense that disaster was around the corner. Several months into the episode, she saw her general practitioner who prescribed an antidepressant, which helped somewhat, but after several months without greater improvement, she sought psychotherapy.

At intake, she was still moderately depressed with an HRSD-17 score of 21 and a BDI-II score of 19, both indicating moderately severe depression. She received 18 months of dynamic psychotherapy with an experienced clinician initially once but eventually twice weekly. Both patient and clinician described the therapy as very helpful; however, the therapist communicated that this was really a case requiring longer-term treatment. Termination occurred at a time when the patient was beginning to address deeper characterological issues, and thus was experienced as premature by both patient and therapist. Over the course of the therapy, she slowly remitted, attaining recovery from the index episode by month 12 of treatment. Over the last 6 months of treatment and subsequent 24 months of follow-up (30 months total), she had four of eight HRSD-17 scores above 6, indicating a mixed period of recovery interspersed with mild symptoms. Although there were no recurrences of a depressive episode meeting full criteria, this indicates that some vulnerability to depression remained. Figure 6.11 shows the changes in the HRSD-17 score over this time period.

Fig. 6.11

The Hamilton Rating Scale for Depression –17 (HRSD-17) decreases in severity, indicating the patient is fully recovered (<6) after almost 2 years. However, half the observations from 2.5 years onward are in the mild range, suggesting residual symptoms remain

This case demonstrates that characterological issues may lead to some regression in defensive functioning even while the patient may improve symptomatically and in other areas of coping. If therapy terminates before the deeper work is done, some vulnerability in defensive functioning may remain, which may be associated with residual symptoms.

Defenses

The significant characterological issues in this case are reflected in how defenses changed over the course of therapy. Overall defensive functioning changed very slowly but nonsignificantly up to session 83, the last one rated, 2 months from termination. Combining ratings of all three data types (dynamic and RAP interviews and therapy sessions), a linear regression model indicated that her initial ODF (intercept) was 5.29 with a rate of change (slope) of 0.026 points per year, statistically nonsignificant, estimating a final ODF of 5.33 at about 16 months. This indicates that she began and ended about in the midpoint of the range of neurotic functioning. Figure 6.12, which breaks down the change in ODF by interview type, displays three different patterns of change. According to the dynamic interviews, her ODF remains about the same, whereas the RAP interview, which reflects interpersonal interactions, indicates an increase of 0.243 points per year. However, the psychotherapy sessions displayed an interesting pattern: ODF initially displayed a lot of variability, which became more stable by the end of the 18 months of treatment, albeit at a lower level (ODF rate of change = –0.194 points per year). In therapy, as the patient stabilized, there was a mild degree of regression in functioning, consistent with an intensive process of dynamic exploration wherein the patient allows neurotic issues and related defenses to emerge for exploration of affective meaning and development of insight. Had therapy proceeded into a longer-term therapy, once hitting a plateau, ODF would have been expected to rise as treatment worked through the underlying characterological issues. Further examination of the details can help support whether this interpretation is correct, that real change is occurring, but the process was not brought to completion in the 18-month time frame.

Fig. 6.12

Over the 18 months of treatment, overall defensive functioning (ODF) remains the same in the dynamic interviews, but improves in interpersonal vignettes (RAP interviews), while showing some stabilization and regression in the therapy sessions

The above differences in ODF across the three interview types correlate with the degree of structure in each. Beck and Perry [33] found that RAP interviews are more structured than dynamic interviews, while sessions are the least structured. They hypothesized that less-structured interviews would be associated with a diminution in the level of defensive functioning, which is true in this case.

We examined variability comparing the range of ODF in the first half to that in the second half of the period over which defenses were rated (0.93 vs. 0.45 points). This indicated that the variability was decreasing, one indicator of improvement, even though ODF did not yet break significantly higher.

Figure 6.13 examines three groups of defenses. Her immature defenses, which include the depressive defenses, decreased nearly significantly ( p = 0.07), while her use of repression increased markedly but nonsignificantly ( p = .14). This indicates that as she traded off her lower-level defenses, she increasingly explored material that had heretofore been more deeply repressed. The increase in repression indicates that the material is closer to consciousness and within the reach of some exploration. At the same time, there was a slight increase in obsessional defenses as she attempted to understand what she is conflicted about, but was still not able to tolerate the associated affects fully. The mature defenses did not change significantly yet, again indicative that at the time of termination, she was only midway through the process of getting to healthy-neurotic. Were the patient to have continued and completed the work of a longer-term therapy, we would predict that the immature defenses would continue to fall, and in particular the depressive defenses would fall below 7%, the cutoff for individuals without depression. Furthermore, the subject would trade off her hysterical and obsessional defenses for an increase in mature-level defenses.

Fig. 6.13

During 18 months of therapy, immature defenses decrease while both hysterical and obsessional levels rise, indicating increasing reliance on neurotic mechanisms

Coping

We also assessed coping mechanisms in this case. As briefly described in the introduction, Haan’s [6] distinction between defense mechanisms that respond to intrapsychic conflict and coping mechanisms that deal with adaptation to external reality led the way for Lazarus and Folkman’s [7] complete separation of coping mechanisms from defenses. Coping mechanisms were thus conceptualized as largely conscious processes that individuals employ in response to stressors or problems encountered in their environment. A recent review of empirical methods for studying coping [8] concluded that coping has a hierarchical structure in which specific instances of coping action patterns (CAPs) can be classified into 12 categories, and these 12 categories can be further grouped into three broader families of coping that are grouped together based on shared objectives. The competence family seeks to coordinate actions in the environment, the relatedness family seeks to coordinate self-reliance and social resources, and the autonomy family seeks to coordinate individual preferences and available options. Each family consists of two more-adaptive coping mechanisms and two less-adaptive coping mechanisms (adaptive processes are listed first). Competence processes include problem-solving, information seeking, helplessness, and escape; relatedness processes include self-reliance, support seeking, delegation, and isolation; autonomy processes include accommodation, negotiation, submission, and opposition. Thus, much like defenses, coping mechanisms are organized into a hierarchy, and it is expected that more-adaptive forms of coping would be related to better psychological and social functioning.

In the present case example, we used the hierarchical structure of coping to create a manual for coding verbatim text from patient interviews and psychotherapy sessions (Perry et al., ms in preparation). Coping was divided into positive CAPs (more-adaptive coping) negative CAPs (less-adaptive coping). Furthermore, a general score of overall coping functioning (OCF) was constructed as the overall proportion of adaptive coping used by the patient in each transcript (range 0–1.0). Figure 6.14 displays the OCF scores for the interviews and sessions. At intake, the subject’s score was less than 0.50, which is the mean for subjects with acute depression, indicating that coping is divided equally between positive and negative categories. Simple linear regression found this rate of change (0.129 per year) to be significant ( p = .007). By the end of treatment, OCF was estimated to have increased to 0.687, indicating that two-thirds of her coping was positive, although this is still below the mean of 0.75 for a community sample of nonill women who were mothers (data still being rated by the authors). Thus she improved measurably but may not have fully reached healthy norms.

Fig. 6.14

Overall coping functioning (OCF, the proportion of coping mechanisms that are positive) rises over the 18 months of therapy, but remains shy of normative levels

We were interested in the relationship between coping and defenses as markers of change of functioning over psychotherapy. The median number of defenses was 58.5, while the median number of CAPs was 24, indicating a ratio of 2.4 defenses to every CAP. Furthermore, across the 12 interviews, the numbers scored for each were highly correlated (Spearman correlation or r _s = 0.73, n = 12, p = 0.007). Not surprisingly, as ODF did not change overall, there was little correlation with OCF across the sessions (r _s = .16, n.s.), whereas there was a significant correlation between the immature defense score and OCF (r _s = –0.72, p = 0.008), as the immature defenses decreased OCF improved. The trade off was largely to neurotic not to mature defenses, with the former correlating at a trend level with OCF (r _s = 0.51, p = 0.09). Examining some of the individual defenses noted in Fig. 6.13, the increase in repression was correlated with OCF (r _s = 0.62, p = 0.03), suggesting that as warded off issues became more evident, even though still somewhat repressed, her overall coping improved. Similarly the levels of intellectualization and isolation correlated with OCF (r _s = 0.58, p = 0.05 and r _s = 0.48, p = 0.12, respectively). As we posited earlier that her defensive functioning returned to a neurotic level of functioning, as therapy progressed and depression lifted, this suggests that some of her improvement in OCF may have been a return to previous levels of coping available to her as well. One test of this would be the durability of change in both defensive and coping functioning over ensuing years of follow-up. Only time can sort out the state changes from the more resilient trait improvements. In any case, in this subject, the immature defenses were improving and were highly correlated with improved coping. Unfortunately, the termination before changes were seen in high adaptive-level defenses precluded determining whether such changes would have heralded a lift of OCF into healthy normative levels. Overall then, we have evidence of both some convergence and discrimination of defensive and coping functioning.

Discussion

The four cases presented are relevant to all or most of the hypotheses about how defenses change. Each case was selected to represent a different type of disorder or therapeutic approach, resulting in too few cases to test the validity of the hypotheses. Nonetheless, we can reasonably discuss whether the data from each case support the hypotheses or not, offering very preliminary evidence of validity. Of course, full validation would require larger samples and different treatment conditions and designs, especially for the fourth hypothesis. Table 6.3 presents the findings from each case for each of the hypotheses discussed.

Table 6.3 Confirmation or disconfirmation of hypotheses by case

Hypothesis 1 stated that as patients improve, ODF rises. Cases A–C clearly demonstrated this, while case D demonstrated more limited changes in defensive functioning that resulted in only a small rise in the overall level (ODF). Case A was assessed at only three points in time over a year and showed a move from depressive-immature to a neurotic level by termination, maintained at 1-year follow-up. Although meaningful, it is not clear whether this was a state change, back to the status quo prior to the depressive episode, or whether it represented a move to a new, higher level of defensive functioning. Cases B and C are different in that each patient clearly moved to and remained at a significantly higher ODF over time, consistent with true trait changes. Alongside the rise in ODF was the reduction of variability in defensive functioning as patients improved. Although this is harder to estimate in case A with only three ratings, clearly Cases B–D all showed that over time the range from lowest to highest ODF diminished by half or more. This supports the subhypothesis that as defenses improve, they become more resilient, as indicated by less deviation from their usual level of functioning. Stress does not result in as wide swings in defensive functioning as was evident prior to and during the earlier phases of treatment.

Hypothesis 2 stated that improvement occurs stepwise up the hierarchy, first trading off lower-level defenses, e.g., action or borderline, for defenses somewhat higher on the hierarchy. After this is well under way, then midlevel defenses, such as higher-level immature, e.g., narcissistic defenses, or neurotic defenses diminish as mature defenses increase. Thus the whole hierarchy does not shift at once, but in temporal phases affecting lower defense levels first, followed by mid level, then high level. All four cases demonstrated this in part, and three in full. Each case demonstrated substantial decreases in immature defenses with simultaneous increases in mid level, but not high adaptive-level defenses. Case C was atypical in that high adaptive-level defenses began to increase from intake onward. However, that case was also atypical in that the subject’s therapy had been ongoing for several years prior to intake into a naturalistic study, and thus may have been in a later phase of treatment with her therapist than the other three cases in which the total therapy was observed. Cases A and C developed some or a sizable proportion of mature defenses, in the case of the latter as a result of a very long period of treatment and follow-up. Cases B and D improved shy of developing mature defenses, because the treatment in both cases appeared to terminate before completing the therapeutic tasks as reported by both therapists. This relates to the fourth hypothesis as well, regarding the role of patient factors that moderate change. Thus all four cases support the idea that defenses change in stages, but treatment to a healthy level of functioning was not carried out in all cases, which would have provided a more complete picture across the hierarchy.

Hypothesis 3 posited that different patients or classes of patients and different treatment conditions should be associated with different rates of change in defensive functioning. Four cases can illuminate only some factors that might plausibly be related to the rate of change, or the amount of change over a given time period. Although common experience alone would support this, it remains to define and validate these moderators (i.e., patient characteristics) and mediators (e.g., type of treatment, frequency of sessions, therapeutic alliance, length of treatment, or follow-up).

Case A was characterized by a discrete short-duration depressive episode and good pre-episode personality functioning. The rise of ODF to neurotic levels was accomplished over 20 weeks, concurrent with cessation of depression. However, depressive defenses were still somewhat above healthy norms, and high adaptive-level defenses were just beginning to rise, suggesting that the rise in ODF was at least somewhat state-dependent, and further change after termination would occur at a slower rate. This was in line with the findings of Hoglend and Perry [31], who found that their depressed patients returned to a neurotic level (mean ODF = 5.11) after 6 months. Case D exemplified this more clearly. Her depression was of longer duration and there were more characterological issues, which a longer but not open-ended therapy – limited to 18 months – tried to address. This process included a period of regression in the therapy as deeply repressed issues emerged. When therapy stopped at 18 months per protocol, both patient and therapist felt still “in the middle” of the treatment. As a result, lower-level defenses decreased, neurotic-level defenses increased, but high adaptive defenses did not emerge yet. Thus the patterns of change in depression and defenses were not the same, with the characterological work producing a countertrend in defensive functioning, in the therapy itself, whereas defenses in relationship vignettes (the RAP) did appear to improve.

Both cases B and C had BPD and entered their respective studies in a regressed state. The rate of change was slow but after a period of time became quite steady, after only a year in case B. A rate of change which help-up remarkably until year 13-plus. Notably, over most of this time, she remained with the same therapist. Thus we can see differences between these two cases with PDs – both had BPD – and the two cases with discrete depressive episodes. In the latter, the early years showed unstable defensive functioning, which included periods of regression, before stable rates of change developed. In the two depressive cases, return of functioning was quicker in the presence of mild characterological issues. In case D with more characterological issues, there was improvement in the immature defenses while some regression to more neurotic levels as the deeper work proceeded. In this case, regression was much more limited in depth than in the cases with PDs.

Hypothesis 4 posited that improvement in defensive functioning should be associated with improvement in symptoms and other aspects of functioning. All four cases followed this pattern. What remains to be determined is whether attaining healthy-neurotic levels of defensive functioning (e.g., ODF > 5.64) confers protection from major psychiatric conditions, such as recurrence of major depression, and disruption of generally high levels of functioning, in the face of difficult environmental stressors, such as death of a loved one and loss of a job. Future studies should examine whether individuals who have developed healthy functioning are more likely to remain so if their defensive functioning has also risen to healthy-neurotic levels.

Some Future Directions

These four cases suggest several further related hypotheses, which include potential interactions between moderators and mediators of defensive change.

• A. Episodic disorders, such as major depression, may be associated with rates of change related to the phase of the disorder. The rate of change during the depressive phase through remission may be faster than the subsequent rate of change once remission has occurred while treatment or follow-up continues.

• B. Treatment length may paradoxically affect the rate of change over certain periods of time. Short-term treatments (e.g., 20 sessions or less) may be associated with a faster rate of change than is found in longer-term treatments. The faster rate of change may be due to the experience of state effects associated with starting treatment, which wash out in longer-term treatments as characterological work gradually progresses. However, further change may not occur or may occur at a slower pace (viz., point A above).

• C. Some patient factors may attenuate the rate of change in both treatment and naturalistic conditions. Likely candidates include having a PD, especially BPD, high levels of Axis I comorbidity, a longer duration of illness, early age of onset, and greater number of prior episodes. As these are validated, they will have implications for the duration, and perhaps the frequency (intensity) of treatment, noted next.

• D. The optimal duration of any treatment, which has a goal of improving defensive functioning, will be a function of the patient factors moderating the rate of change. For instance, case B improved markedly but clearly required additional treatment beyond the two plus years in the index therapy, truncated when the therapist had to terminate. Case C required a very long-term treatment (about 10 years, although intensity diminished over time).

• E. Related to points C and D is the hypothesis that for more disturbed patients, the beginning of therapy will be followed by a period of induction. In this early phase, the patient and the therapist form an alliance and deal with repeated crises and disruptions to the work of therapy, before engaging in the more steady, fundamental characterological work. During this induction phase, there may appear to be no consistent change in defensive functioning, i.e., a negligible rate of change. Case B, who was highly motivated by the painful regressive experience in the preceding failed therapy, finished the induction phase by the second year, while case C, who was far more dysfunctional, required between 3 and 5 years to achieve a steady rate of change. Following the induction phase, when the patient is thoroughly engaged in the fundamental therapeutic tasks, the patient will develop a faster, more or less linear rate of change over a long time frame, assuming that therapy continues. Of course, each patient under optimal circumstances will still reach a ceiling – healthy-neurotic functioning – at which point the rate of change diminishes toward zero.

• F. Any effective treatment should increase the rate of change in defensive functioning when compared to natural history. A related idea is that sufficient treatment may bring the patient to the point where, after termination, the rate of change continues to be positive, even though healthy-neurotic functioning was not attained by termination. This is often called a delayed treatment effect. The patient takes away the ability to continue the work of therapy on his or her own. What remains to be determined is what type of treatment is effective, at what intensity, for which class of patients, after requiring what duration of a treatment induction phase. It is plausible that all treatments including supportive, experiential, interpersonal, cognitive–behavioral, dynamic, and psychoanalysis proper may affect defensive functioning but that the rates of change may differ among them.

Conclusion

This chapter has examined change in defensive functioning as an outcome in itself as well as in relationship to other outcomes. The issue of how to change defenses is the province of studies of the therapeutic process and has been only touched upon in passing in this chapter. That too is worthy of its own exploration, as we have discussed elsewhere [13, 34]. Ultimately, one could examine defenses in response to the therapist’s interventions, with the expectation that a given level of defensive functioning would interact with certain interventions to lead to differences in outcome. This requires that both defenses and interventions intended to affect them be examined together, a step that has begun using several instruments to identify relevant therapeutic interventions [35, 36]. For example, in a study of a four-session brief psychodynamic investigation, in the initial session, a high level of therapist attunement to the level of defensive functioning leads to positive changes in the therapeutic alliance across the following three sessions [37]. A result of studies of this type should be that defenses could be examined at points in treatment as an indicator of the instantaneous effect of the therapeutic process, while they could also be summed, compared across sessions or interviews within or outside of therapy to provide an overall indicator of outcome. Defenses are a robust measure of how personality structure is functioning at any moment in time, which measured over time can reveal whether that structure is changing. Thus we can note the defenses arising in psychotherapy to attune to a patient’s functioning, and we can assess defenses over time to measure structural change. This makes the study of defenses one of the most versatile psychodynamic endeavors spanning theory, research, and clinical work. Finally, this chapter described a research agenda that will further bridge process and outcome studies to advance our future understanding of how our treatments work.