Psychotherapy Research Interpersonal Narratives in ...

1 downloads 0 Views 1MB Size Report
Nov 25, 2010 - Stuart(1999) 'Interpersonal Narratives in Cognitive and Interpersonal Psychotherapies', Psychotherapy Research, 9: 1, 22. — 35. To link to this ...
This article was downloaded by: [University of Pennsylvania] On: 14 December 2010 Access details: Access Details: [subscription number 915031446] Publisher Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK

Psychotherapy Research

Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713663589

Interpersonal Narratives in Cognitive and Interpersonal Psychotherapies

Paul Crits-Christopha; Mary Connollya; Sandi Shappella; Irene Elkinb; Janice Krupnickc; Stuart Sotskyc a University of Pennsylvania, b University of Chicago, c George Washington University, Online publication date: 25 November 2010

To cite this Article Crits-Christoph, Paul , Connolly, Mary , Shappell, Sandi , Elkin, Irene , Krupnick, Janice and Sotsky,

Stuart(1999) 'Interpersonal Narratives in Cognitive and Interpersonal Psychotherapies', Psychotherapy Research, 9: 1, 22 — 35 To link to this Article: DOI: 10.1080/10503309912331332571 URL: http://dx.doi.org/10.1080/10503309912331332571

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

psychotherapy Research 9(1> 22-35, 1999

INTERPERSONAL NARRATIVES IN COGNITIVE AND INTERPERSONAL PSYCHOTHERAPIES Paul Crits-Christoph Mary Beth Connolly Sandi Shappell University of Pennsylvania

Irene Elkin

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

University of Chicago

Janice Krupnick Stuart Sotsky George Washington University

The purpose of the current investigation was to explore the predictors of the frequency and completeness of interpersonal narratives in psychotherapy. Narratives were extracted from 548 sessions of 72 patients who received either cognitive (CT) or interpersonal therapy (IPT) for depression. Consistent individual differences in narrative frequency, length, completeness, and number of therapist words per narrative were found. IPT sessions contained significantly more narratives than CT sessions, and CT sessions contained a higher proportion of therapist words per narrative. The alliance was positively related to the number of patient words per narrative, and patients with more involved interpersonal styles elicited more therapist words per narrative. Expectations about the causes of depression and helpfulness of focusing on interpersonal issues in therapy influenced narrative frequency and completeness more in CT than in IPT.

Over the past 20 years, a variety of methods have been developed for assessing maladaptive interpersonal themes in psychotherapy sessions. These methods include the Core Conflictual Relationship Theme (CCRT) method (Luborsky, 1977; Luborsky & Crits-Christoph, 19901, the FRAME method (Dahl & Teller, 19941, the Cyclical Maladaptive Pattern method (Schacht, Binder, & Strupp, 19841, and the configurational analysis (Horowitz, 19871, among others. Each of these methods examines the content of interpersonal processes, using psychotherapy transcripts as the source of information. The methods include the assumption that the content of patient interpersonal themes can be identified from patient discourse about their interactions with others, or from the actual interaction of patient and therapist. Although a number of studies have examined the content of interpersonal themes in psychotherapy (Connolly et al., 1996; Crits-Christoph, Demorest, Muenz, & Baranackie, 1994; McMullen & Conway, 1994; McMullen & Conway, 19971, there has been little attention directed towards the building blocks of understanding interpersonal Correspondence regarding this article should be addressed to Paul Crits-Christoph, PhD, Hospital of the University of Pennsylvania, 3600 Market Street, 7th Floor, Philadelphia PA 19104-2648.

22

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

INTERPERSONAL NARRATIVES

23

processes: the quantity and completeness (i.e., relative amount of detail) of patient speech about interpersonal events. One exception is Luborsky, Barber, and Diguer (1992), who summarize some data about the nature of “relationshipepisodes” used in the CCRT method. In particular, information on the frequency, typical length, time frame, and type of other people discussed in narratives is presented (Luborsky et al., 1992). N o studies to date, however, have examined the factors that might predict individual differences in the quantity or completeness of narratives about interpersonal interactions. However, individual differences in the completeness and quantity of narratives are likely to be potentially important influences on the material available to the therapist for formulation and intervention. The.relevant predictor of outcome is what the therapist does (i.e., “accuracy of intervention”) with the clinical material, not the narratives per se, although obtaining more and better quality material should lead to better interpersonal formulations, which in turn should increase the accuracy of therapist interpersonal interventions. Interpersonal accuracy has been shown to relate to improved outcome (Crits-Christoph et al., 1988; Norville, Sampson, & Weiss, 1996; Piper, Joyce, McCallum, & Azim, 19931, to better response to interpretations (Silberschatz, Fretter, & Curtis, 19861, and to the development of the therapeutic alliance over the course of treatment (Crits-Christoph, Barber, & Kurcias, 1993). Information about the determinants of frequency and completeness of interpersonal narration might also shed light on the use of psychotherapy research methods that utilized such narration for their scoring. Researchers might have a better sense of the kinds of treatments and patients to which such methods can be most fruitfully applied. The current study was designed to examine a number of hypotheses about the predictors of the frequency and completeness of interpersonal narratives in psychotherapy. Our overall hypothesis was that the occurrence of interpersonal narratives was multiply determined. More specifically, it was first hypothesized that mode of treatment would be related to frequency and completeness of such narratives, with a more interpersonally oriented therapy capable of eliciting more and better narratives. Second, the frequency and completeness of narratives was hypothesized to be related to the extent to which patients report interpersonal problems pretherapy. In general, patients who report high levels of interpersonal problems would be expected to discuss such problems more than patients who report few, if any, interpersonal problems (e.g., they focus on somatic symptoms rather than interpersonal problems). Of course, factors such as patient defensiveness might mitigate this relationship. For example, some patients might recount numerous detailed stories as a way of avoiding important topics. Patient expectations about psychotherapy is a third factor that might be crucial. It was hypothesized that patients who believed that the causes of their problems related to interpersonal issues, and also believed that discussing interpersonal relationships in therapy would be helpful, were likely to produce more examples and more complete descriptions of their interactions with others. Fourth, the interpersonal style of patients needs to be considered. Patients highly involved in interpersonal relationships, versus a detached style, would be more likely to have ongoing interactions with others that would serve as the basis for other interpersonal narrations, or simply might have greater interest in interpersonal processes. Finally, it was hypothesized that patient-therapist dyads characterized by stronger therapeutic alliances would yield greater frequency and completeness of relationship episodes. Within the context of a positive therapeutic alliance, it seemed likely that patients would feel less defensive and therefore be more comfortable discussing the troublesome

24

CRITS-CHRISTOPH ET AL.

and affect-laden interpersonal events within their worlds. For example, a depressed patient might be more likely to explicitly discuss past interactions with a lost loved one in the context of a positive therapeutic alliance, where it feels safe to experience the sadness that would be brought up by remembering such interactions. Before examining the predictors of the frequency and completeness of interpersonal narratives, we also examined the extent to which patients displayed consistent individual differences in such narrative qualities.

METHOD

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

PARTICIPANTS The sample for the current investigation involved participants in the National Institute of Mental Health (NIMH) Treatment of Depression Collaborative Research Program (TDCRP) (Elkin et al., 1985; Elkin et al., 1989). The TDCRP compared cognitive behavior therapy (CBT), interpersonal therapy (IPT), pharmacotherapy, and placebo for the treatment of major depressive disorder. Psychotherapy transcripts for patients who received either CBT or IPT from sites two and three were used in the current investigation. Of the 80 participants who entered treatment at sites two and three, we were able to use only 72 in this sample. Seven of the 80 had to be excluded because they did not have enough sessions, and one did not have enough audible tapes to be used for transcription. Each of the 72 Participants had four early sessions (two, three, four, and five whenever possible) available for transcription, and 65 of those participants each had four late sessions that were transcribed for this investigation. The late sessions were used only if there were at least three sessions between the last early session and the first late session. The late sessions also excluded the last two therapy sessions for each patient because the focus of those sessions was presumed to be the treatment termination. Given these guidelines, patients included in the current sample had a minimum of 14 sessions. The patients in the current sample were on average 34.5 years old (SD = 8.5). Twenty-four percent were male and 42% were married or living with a significant other. This patient sample was 93% European American and 7% African American, Asian, Hispanic, or other minority group. Sixty-nine percent of the sample had received some college education. TREATMENTS The treatments used in this study were manualized cognitive behavior therapy (CBT) (Beck, Rush, Shaw, and Emery, 1979, and interpersonal psychotherapy (IPT) (Herman, Weissman, Rounsaville, & Chevron, 1984). CBT uses techniques which focus on correcting the maladaptive beliefs that underlie a patient’s distorted, negative thoughts about life and the future. IPT is designed to help the patient understand hidher interpersonal problems and improve hidher social functioning (Elkin et al., 1989). THERAPISTS There were 12 therapists at the two sites used for our project (tapes from site one were not permitted to be released for further research at another location). Five

INTERPERSONAL NARRATIVES

25

therapists were in the CBT condition and seven therapists were in the IPT condition. Each of the therapists saw an average of six patients. These therapists averaged 13 years (SD = 6) of clinical experience, were on average 43 years old (SD = 8), and 67% were male.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

MEASURES

fdentifyingNarratiues. Therapy sessions were transcribed from audiotapes using the transcription rules described by Mergenthaler and Stinson (1992). The transcripts were then checked and corrected, during which time all identifying information in the sessions was changed to protect confidentiality. A total of 288 early sessions and 260 late sessions were transcribed. The judges used to identify the interpersonal narratives were one advanced undergraduate student and one recent college graduate research assistant, who were trained in the scoring of relationship episodes. Training consisted of work with 20 practice cases and extensive review of the principles for locating relationship episodes. Moreover, the two judges had periodic recalibration sessions throughout the period of working on the actual study materials. Relationship episodes are narratives told by the patient describing specific interactions between the patient and another person in his or her life (Luborsky & CritsChristoph, 1990). The two judges worked independently to identify all relationship episodes in all sessions. For each session, judges marked down where the narratives began and ended on relationship episode scoresheets.Judges used a one to five Likert scale to rate the relationship episodes on the “completeness” or level of elaboration of the CCRT components (wishes, responses of other, responses of self) (Luborsky & Crits-Christoph, 1990) within the relationship episode. Interjudge reliability assessed via the intraclass correlation coefficient (ICC (2,211 (Shrout & Fleiss, 1979) was .75 for completeness ratings (average of two judges). For each relationship episode that the judges scored, the two judges’ completeness ratings were averaged to get a final completeness rating. There were a total of 3245 episodes scored by both judges, and these episodes had an average completeness rating of 2.4 (SD = 0.9). Previous studies examining the content of relationship episodes, reviewed by Luborsky and Crits-Christoph(1990), have only utilized those episodes with an average completenessrating of 2.5 or greater. In the current data, there were 1302 such episodes. Of the episodes identified by judge one as having a completeness of 2.5 or greater, judge two located the same episode 95% of the time. Of the episodes identified by judge two as having a completeness rating of 2.5 or greater, judge one located the same episode 98% of the time. Of those episodes with an average completeness rating of 2.5 or greater, the median number of transcript lines of text different between the two judges on the beginnings of relationship episodes was zero, and, for 80% of the episodes, the two judges differed by seven lines or less on identifying relationship episode beginnings. For demarking the end of a relationship episode, the two judges were within 14 text lines of each other 80% of the time. A consensus judge was used to resolve discrepancies between the two judges regarding the start and end point of each relationship episode rated 2.5 or greater on average. For the current article, four dependent measures were used to represent the frequency and completeness of the interpersonal narratives in psychotherapy. The number of relationship episodes rated 2.5 or greater within each session was used as the measure of frequency of interpersonal narratives. Thus, there was one score

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

26

CRITS-CHRISTOPH ET AL.

for each of the 548 sessions on this measure. Using this measure, the average number of episodes per session was 2.4 (SD = 1.9). The completeness rating of each episode was also included as a dependent variable, using all relationship episodes for which the two judges had agreed that an episode existed (3245 episodes in total). The number of patient words per episode was computed to represent the narrative length. We were also interested in the therapist contributions to relationship episodes. For this measure, we used the proportion of words within relationship episodes that were spoken by the therapist. The number of patient words and the proportion of therapist words were calculated, using only those episodes that were relatively complete (2.5 rating or greater). Expectations About Change. Items from the Patient Attitudes and Expectations form (Elkin et al., 1985) were used to assess patient beliefs regarding the causes of their problems and what would be helpful in psychotherapy. In regard to causes, each item was rated on a scale from one to seven, ranging from “has nothing to do with my problems” to “accounts fully for my problems.” Three items representing interpersonal causes were included: “never having really learned how to get along with people,” “having many arguments or conflicts with my family,” and “being alone and lonely too much of the time.” In regard to helpful aspects of therapy, each item was rated on a scale from one to seven representing helpfulness in alleviating depressive symptoms (one not helpful at all, seven extremely helpful). Three items relating to talking about interpersonal topics were used: “learning how to resolve my family conflicts,” “learning how to get along better with people,” and “learning how to reach out to other people.” The internal consistency (Cronbachs alpha) of the six item scale formed by averaging over the three “cause”items and three “helpful” items was .71. Interpersonal problems. Selected items from the Social Adjustment Scale (SAS) (Weissman & Pakel, 1974) modified for the TDCRP were used to assess problems in interpersonal functioning. We were interested in assessing the degree of interpersonal problems experienced in current relationships, yet did not want to confound our measure by including in our assessment items that capture a lack of interpersonal relationships. The 12 items selected, therefore, all reflected problems in ongoing relationships. They included two items from the social leisure scale: friction in relationships and hypersensitivity; four items from the extended family scale: friction in family relationships, worry, guilt, and resentment; and six items from the marital/scale: friction, submissive behavior, lack of affection, disinterest in sex, diminished sexual intercourse, and sexual problems. Each item was rated on a five point scale by a clinical evaluator on the basis of an intake interview. The final score used in the current investigation was the average rating across the available items for each patient (some patients did not have responses for the marital/partner items). The internal consistency of this 12 item scale, assessed via a split half reliability, was .65. Hamilton Rating ScaleforDepression. The current investigation used the rescreening scores o n the 17 item Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960) modified for the TDCRP to represent the participant’s pretreatment level of depressive symptomatology. Interpersonal Style Invento y .The Abbreviated Interpersonal Style Inventory (ISI) (Lorr & Youniss, 1973, Lorr & DeJong, 1986) was completed by participants at treatment intake. It consists of items designed to assess ways in which people relate or respond to one another (Lorr & DeJong, 1986). For this investigation, we used the second order factor labeled interpersonal involvement (Lorr & DeLong, 19841,which includes the sociable, help seeking, nurturant, and sensitive subscales. This factor

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

INTERPERSONAL NARRATIVES

27

consists of 40 items which are rated as true or false by the respondent. Internal consistencies of the component scale scores of this factor (ranging from .80 to .92), as well as data o n convergent and discriminant validity, are presented in Lorr and Manning (1978). Vanderbilt Therapeutic Alliance Scale. A modified version of the Vanderbilt Therapeutic Alliance Scale (VTAS) (Hartley & Strupp, 1983) was used to rate the strength of the therapeutic alliance. These ratings have been previously shown to predict outcome across all treatment conditions of the TDCRP (Krupnick et al. 1996). The modified VTAS consisted of 37 items. The total score of the 37 items rated on session three was employed for the current investigation. Raters for this measure were two MSW and one DSW clinical social workers and an advanced clinical psychology doctoral student, all of whom participated in an extensive rater training period before beginning the study. Interrater reliability for this total alliance score was reported to be .66 (Krupnick et al., 1996). Internal consistency (Cronbach’s alpha) of the total alliance score was .91.

RESULTS ARE THERE CONSISTENT INDIVIDUAL DIFFERENCES IN NARRATIVE FREQUENCY AND COMPLETENESS?

To assess whether there were consistent individual differences in relationship episode frequency and completeness, intraclass correlation coefficients (ICC) were computed on the number of relationship episodes per session, completeness ratings, number of patient words per relationship episode, and proportion of therapist words per relationship episode. Because of interest in predicting change in these aspects of relationship episodes from early to later in treatment sessions, we initially calculated ICCs o n the early and later sessions separately, a s well as using both early and late sessions together. The ICC assessed the extent to which within patient variability was less than between patient variability. Shrout and Fleiss’s (1979) ICC (1, k) was used, where k represents the number of sessions (for the measure of frequency of relationship episodes) or relationship episodes (for the measures of completeness, patient words, and therapist words) per patient. The results are displayed in Table 1. As seen in Table 1, ICCs ranged from .41 to .88. These coefficients can be interpreted as providing evidence that moderate to strong individual differences in narrative quantity and completeness exist (i.e., that there is relatively greater variability between patients compared to within patients), provided an adequate sampling of episodes is made. When the early and late sessions were examined separately, ICCs were somewhat lower. These ICCs justified examining the predictors of such individual differences. Final scores were created by aggregating across all sessions or episodes. Exploratory analyses examining early and late scores separately (i.e., assessing change from early to late) were also performed. The final aggregated measures of frequency and completeness of relationship episodes were relatively uncorrelated with each other (see Table 21, with two exceptions: the number of relationship episodes per session was moderately positively correlated with the average completeness rating ( r = .42,p < .OOl), and moderately negatively correlated with the proportion of words in an episode spoken by the thera-

28

CRITS-CHRISTOPH ET AL.

TABLE 1. Individual Differences in Relationship Episode Frequency and Completeness ICC

Early

Completeness rating Patient words per episode Therapist words per episode Frequency of episodes per session

M

SD

2.35 345

.87 258 14.7 1.9

15.3 2.4

Sessions

Later Sessions .65 .49 .78 .54

.57

.65 .87 .41

All

Sessions .70 .70 .88

.63

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

Note. Coefficients are intraclass correlations comparing within patient variability to between patient variability. 548 sessions for 72 patients were used.

pist ( r = -.47,p c .OOl). This latter result suggests that the more a therapist talks, the less opportunity a patient has to tell additional narratives. PREDICTION OF EPISODE FREQUENCY AND COMPLETENESS

PreliminaryAnalyses. Prior to relating our predictor variables to the dependent variables, we examined the degree of overlap among the predictors (treatment mode, expectations, interpersonal style, social problems, and alliance). These intercorrelations are given in Table 3. The Interpersonal Style Inventory correlated significantly with the measure of expectations about the helpfulness of talking about interpersonal issues in therapy(r = .29, p = .014), with a more involved interpersonal style being associated with greater expectations that talking about interpersonal issues would be helpful in therapy. The measure of expectation was also significantly correlated ( r = .32,p = .007) with social problems from the selected Social Adjustment Scale items, again in the direction that more social problems was associated with a greater expectation that talking about interpersonal issues would be helpful. There was also a significant ( r = -.26, p = .05) correlation between Treatment Mode and Social Problems, indicating that patients assigned to CBT began treatment on average with more social problems than patients assigned to IPT. This correlation, however, was a function of the reported imbalance (Elkin et al., 1989) of the treatment conditions in the TDCRP in regard to marital status. Controlling for marital status, the correlation between Treatment Mode and Social Problems was reduced ( p r = -.11, ns). All of the other correlations were near zero and nonsignificant, suggesting little overlap among the predictor variables. In preliminary analyses, we also examined the relationship between demographic variables (age, ethnicity, marital status, gender, education) and the measures of epi-

TABLE 2. Intercorrelationsof Dependent Variables Number of episodes Number of episodes Completeness rating Patient words Proportion of therapist words Note.

*p < ,001.N = 72 patients

Completeness

Patient words

.42 .01 -.47 *

-

.27

-

-.03

-.22

I NTERPE RS0NAL NARRATIVES

29

TABLE 3. Intercorrelations of Predictor Variables

SAS-social problems SAS-social problems Expectations Treatment mode Interpersonal style inventory

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

Treatment modality

Interpersonal style inventory

-

-. 17

.32 ** -.25 *

=

CBT, 2

-

.29 *

-.01 -.13

Alliance Note. Treatment Mode is coded as 1

Expectations

.14 =

-.03

-.01 .02

IPT. *p< .05; **p< .01; N

=

7 2 patients.

sode frequency and completeness. In addition, initial depression severity, as measured by the Hamilton Rating Scale, was also examined because of the importance of this variable within the TDCRP results (Elkin et al.; 1995). There was a trend for marital status (coded as currently living with a partner versus not living with a partner) to be significantly related to proportion of therapist words during early sessions ( r = -.21,p = .065). The direction of this relationship was that patients not currently living with a partner had therapists who spoke more often during relationship episodes. There was also a trend for intake Hamilton Depression scores to be related to the number of relationship episodes per patient ( r = .19, p = .lo), with the more severely depressed patients having relatively more frequent relationship episodes. Both marital status and intake Hamilton Depression scores were used as covariates for all predictive analyses presented below. Prediction of Relationship Episode Quantity and Completeness.The primary analyses consisted of multiple regressions relating the five predictor variables to each of the four dependent variables. Results predicting patient words per episode, proportion of words spoken by the therapist per episode, completeness rating, and frequency of episodes per session are given in Table 4. Our hypothesis that interpersonal psychotherapy would contain a greater number of narratives about interactions with others was strongly supported. In addition, cognitive therapy was characterized by higher proportions of therapist words within relationship episodes. Measures of social problems and expectations, however, did not relate to frequency, patient words, completeness, or therapist words. A statistically significant relationship between the quality of the therapeutic alliance and number of patient words per episode did emerge. Finally, an unpredicted relationship between interpersonal style and therapist words was found. Patients TABLE 4. Prediction of Relationship Episode Frequency and Completeness

Frequency Treatment mode Social problems Expectations Interpersonal style Alliance

.44*' .07 .08 -.14

.05

Completeness

Patient words

.09

.01

-.56"

.08

.14

-.02

.16

.10

-.05

-.07

.04

Therapist words

.07

.23'

.29'

-.18

Note. Coefficients are semi-partial correlations, controlling for other predictors and control variables (Hamilton Rating Scale for Depression, MaritaVPartner Status). Treatment Mode is coded as 1 = CBT, 2 = IPT. N = 7 2 patients; * p < .01; * p < ,001.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

30

CRITS-CHRISTOPH ET AL.

with a more involved interpersonal style elicited more therapist words during relationship episodes. On an exploratory basis, we examined interactions between predictors (social problems, expectations, alIiance, interpersonal style) and treatment modality in relation to characteristics of relationship episodes. These interactions were examined individually via cross-product terms entered into multiple regression analyses after the main effect predictors had been entered. Two findings emerged. Expectations interacted with treatment condition in predicting completeness ratings (cross-product term semi-partial r = -.30, p = .012), with a significant positive relationship in CBT (semi-partial r = .51, p = .OO3) and a nonsignificant relationship within IPT (semipartial r = -.21, ns). Expectations also interacted with treatment condition in predicting the number of episodes per session (cross-product term semi-partial r = -.21, p = .05), with a positive association in CBT (semi-partial r = .38,p = .027) and a nonsignificant association in IPT (semi-partial r = -.02, ns). We also explored whether nonlinear relationships were apparent. Examination of the scatterplots of the relationships between each predictor and each dependent variable revealed no evidence of nonlinear relationships. Finally, although none of our dependent variables evidenced statistically significant change from the early to later in-treatment sessions, we explored the relationship of the predictor variables to change in frequency and completeness of relationship episodes over the course of treatment. Multiple regression analyses were performed for each of the four measures, using later in-treatment scores as the dependent variable, and the corresponding early scores as covariates, before entering the five predictor variables. The results indicated that the therapeutic alliance during the early sessions was related to an increase in the number of patient words from early to late in-treatment (semi-partial r = .18,p < .05>.N o other predictors achieved statistical significance.

DISCUSSION

The present investigation documents that there are consistent individual differences in the number and completeness of interpersonal narratives told by patients in psychotherapy, as well as in therapist activity during such narratives. As such, these data add to the previous suggestions (Luborsky, Barber, & Diguer, 1992) that interpersonal narratives may be a n important aspect of psychotherapy process and a useful unit of analysis for process research. Given the relatively robust individual differences in narrative frequency and completeness that were found, research oriented towards understanding the determinants of such individual differences is warranted. The strongest predictive finding within our data was that treatment modality was associated with frequency of relationship episodes, with IPT demonstrating more episodes than CBT. The discrimination of IPT and CBT based upon greater interpersonal content is not surprising, given previous findings regarding the discrimination of these treatments within the TDCRP in terms of therapist adherence (Hill, O’Grady, & Elkin, 1992). In addition to having fewer episodes, CBT was also characterized by more therapist talk during patient narratives. This might simply reflect the overall greater amount of therapist directiveness within CBT. The fact that IPT yields more narratives with fewer therapist interruptions than CBT suggests that process studies that rely upon narratives as the unit of analysis can be more fruitfully performed within the context of IPT, rather than in CBT.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

INTERPERSONAL NARRATIVES

31

The extent to which CBT, however, might profit from a greater emphasis on interpersonal narratives is an open question. The TDCRP employed the original Beck et al. (1979) cognitive therapy for depression treatment manual, which predated the stronger emphasis on schemas and interpersonal processes in cognitive therapy (Safran & Siegal, 1990; Young, Beck, & Weinberger, 1993). A schema-focused cognitive therapy might contain more of such interpersonal narratives, in an attempt to elicit material that would aid the therapist in formulating interpersonal schemas. The possibility that more of an interpersonal focus in cognitive therapy would be beneficial is suggested by the studies by Hayes, Castonguay, and Goldfried (1996), who found that interventions that addressed the interpersonal and developmental domains were associated with improvement in cognitive therapy, and by Jones and Pulos (1993), who found that more use of psychodynamic techniques in cognitive therapy predicted outcome. Somewhat surprising was that degree of social problems, and expectations regarding the causes of depression and the helpfulness of an interpersonal focus in therapy, did not predict greater frequency or completeness of interpersonal narratives across the two treatment modalities. It may be that expectations regarding psychotherapy change rather rapidly as a patient is socialized into the process of treatment by the therapist, or that patients who have expectations that do not match the treatment focus are more likely to drop out early. Alternatively, more sensitive measures of these constructs might be necessary to uncover such a relationship. However, our exploratory analyses of interactions with treatment modality revealed that expectations concerning causes and the helpfulness of talking about interpersonal topics were an important determinant of the number and completeness of relationship episodes in CBT, but not IPT. One possible explanation of this difference is that IPT therapists attempted to focus on interpersonal topics regardless of whether a given patient felt this was a helpful direction, whereas CBT therapists allowed patients who were interested in interpersonal topics to elaborate on these topics, but didn’t particularly encourage the discussion of interpersonal topics for those patients who did not appear interested in talking in detail about interpersonal topics. This suggests that CBT therapists were tailoring treatment ,to the needs or expectations of patients, by allowing more detailed discussion of interpersonal topics, for those patients who felt it was more relevant to d o so, consistent with recent conceptualizations of psychotherapy process as inherently being characterized to some degree by therapist responsiveness (Stiles & Shapiro, 1994). The extent to which treatment manuals permit versus undermine such therapist responsiveness needs to be further understood. At the least, CBT therapists might want to consider systematically asking patients about their expectations regarding the helpfulness of talking about interpersonal issues. This assessment might help shape the process of therapy. Particularly interesting in the current study was the finding that the strength of the therapeutic alliance was associated both with the length of episodes (i.e., number of patient words per episode) and relative change in the length of episodes from early sessions to later sessions. In addition, there was a trend for the alliance to interact with treatment modality in predicting the number of relationship episodes within a session, with a positive association in IPT but a negative association in CBT. These findings add to the growing literature o n the importance of the alliance (Horvath & Symonds,l991). Within IPT, a positive alliance may help foster the discussion of the patient interpersonal problems, while in CBT a positive therapeutic alliance may help the patient and therapist focus on the cognitive model rather than on interpersonal exploration. We can speculate that if a CBT therapist focused more on interpersonal

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

32

CRITS-CHRISTOPH ET AL.

topics, a positive alliance might also be associated with more interpersonal narration on the part of the patient. Our findings raise the possibility that the alliance may be important to psychotherapy in a number of ways. Whereas it is often assumed that the alliance influences outcome in part by providing a “corrective emotional experience” for patients, it may also be the case that a positive alliance provides a safe environment, encouraging patients to talk about difficult emotions associated with interpersonal interactions. These multiple ways in which the alliance (therapeutic bond) might influence both session and treatment outcome have been described by Orlinsky, Grawe, and Parks (1994). Taken together with previous data suggesting that accurate interpersonal interventions improve the alliance (Crits-Christoph, Barber, & Kurcias, 19931, the current data are consistent with a model of reciprocal causation between the alliance and interpersonal techniques: the alliance may foster the technical work of interpersonally oriented therapies, and accurate interpersonal interventions foster the development or maintenance of the alliance. Gaston and Marmar (1994) explain that “as much as the alliance interacts with different types of interventions in promoting change, the alliance is also likely to be influenced by interventions” (p. 103). The authors suggest that creative research strategies are needed to further understand the reciprocal relation between the alliance and therapeutic techniques. An unpredicted finding of the current study was the relation between interpersonal style (involved vs. detached) and proportion of words spoken during relationship episodes by the therapist. This result suggests that therapists are responding to patient preferences for greater interpersonal involvement by increasing their own involvement, and is consistent again with the notion that therapists are responsive to patient needs. It is hot clear, however, whether the relatively greater therapist involvement (i.e., more therapist words during patient narratives) has a positive or negative impact on eventual treatment outcome. Although therapist verbal activity level has often been positively associated with treatment outcome (Orlinsky, Grawe, & Parks, 1994), it may be that within the narrow context of degree of therapist talk during patient narratives, the positive benefit of therapists displaying more involvement is offset by therapists not allowing patients to fully develop their narrations, with consequently more sparse interpersonal material being available for formulating and addressing interpersonal issues in treatment. Although the current investigation found some predictors of individual differences in narrative quantity and completeness, much remains to be learned about the determinants of narratives in diverse psychotherapies. There may be other patient factors not assessed in the current study that may influence the production of narratives during psychotherapy. Sheer verbal ability may be one factor. Another might be the presence of focal interpersonal conflict (e.g., marital distress), rather than general level of social problems or general interpersonal style. However, it may also be the case that patient factors are only limited determinants of narratives. An important question will be the extent to which therapist interventions can facilitate the expression of such narratives, thereby increasing the amount of clinical material on which to formulate patient-specific interpersonal themes. Because we found no evidence that narrative frequency and completeness changed significantlyover the course of therapy, it seems likely that the therapist influence on narratives is largely within the first several sessions, which set the tone for the rest of treatment. Nevertheless, the widespread interest in interpersonal processes across interpersonal, psychodynamic, cognitive behavioral, and experiential therapies would suggest that greater research attention to the processes through which interpersonal material is elicited in therapy is warranted.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

INTERPERSONAL NARRATIVES

33

Beyond studies of the determinants of narratives, it will be important to examine whether high frequency and completeness of narratives in fact facilitates therapist accuracy of interpersonal intervention. Some patients may tell many detailed but irrelevant stories as a way of avoiding the “real”issues or of avoiding direct interaction with the therapist. This suggests that the relation between narrative frequency and completeness and therapist accuracy may not be a simple linear one. Ultimately, the content of narratives, rather than their simple frequency or amount of detail (completeness), affects the clinical process. Future studies might examine the predictors of different types and contents of interpersonal narratives, rather than their frequency. Studies are also needed to determine whether therapist accuracy at addressing the interpersonal themes derived from narratives ultimately relates to the outcome of CBT and IPT, as was found in studies of psychodynamic therapy (CritsChristoph et al., 1988; Norville et al., 1996; Piper et al., 1993). A study testing this hypothesis is currently in progress.

REFERENCES Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive Therapy of Depression. New York, NY:Guilford Press. Connolly, M. B., Crits-Christoph, P, Demorest, A,, Azarian, K., Muenz, L., & Chittams, J. (1996). The varieties of transference patterns in psychotherapy.Journal of Consulting and ClinicalPsychology, 64, 1213-1221. Crits-Christoph, P., Barber, J., & Kurcias, J. (1993). The accuracy of therapists’ interpretations and the development of the therapeutic alliance. Psychotherapy Research, 3, 25-35. Crits-Christoph, P., Cooper, A,, & Luborsky, L. (1988). The accuracy of therapists’ interpretations and the outcome of dynamic psychotherapy. Journal of Consulting and Clinical Psychology, 56, 490-495. Crits-Christoph, P., Demorest, A,, Muenz, L., & Baranackie, K. (1994). Consistency of interpersonal themes for patients in psychotherapy. Journal of Personality, 62, 499-526. Dahl, H., & Teller, V. (1994). The characteristics, identification, and applications of FRAMES. Psychotherapy Research, 4, 253-276. Elkin, I., Gibbons, R. D., Shea, M. T., & Sotsky, S. M., Watkins, J. T., Pilkonis, P. A., & Hedeker, D. (1995). Initial severity and differential treatment outcome in the National Institute of Mental Health treatment of depression collaborative research program. Journal of Consulting and Clinical Psychology, 63, 841-847 Elkin, I., Parloff, M. B., Hadley, S. W., & Autry, J. H. (1985). National Institute of Mental Health treatment of depression collaborative research program: Background and research plan. Archives of General Psychiaty, 42, 305-316. Elkin, I., Shea, T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins, J. F., Glass, D. R., Pilkonis,

P. A,, Leber, W. R., Docherty, J. P . , Piester, S. J., & Parloff, M. B. (1989). National Institute of Mental Health treatment of depression collaborative research program: General effectiveness of treatments. Archives ofGeneral Psychiatry, 46,971-982. Gaston, L. & Marmar, C. R. (1994). The California psychotherapy alliance scales. In A. 0. Horvath & L. S. Greenberg (Eds.), The working alliance: Theoty, research, and practice (pp. 85-108). New York: John Wiley & Sons, Inc. Hamilton, M. A. (1960). A rating scale for depression. Journal of Neurology and Neurosurgical Psychiaty, 23, 5 6 6 2 . Hartley, D. E. & Strupp, H. H. (1983). The therapeutic alliance: Its relationship to outcome in brief psychotherapy. In J. Masling (Ed.), Empirical Studies of Psychoanalytical Theories. (Vol. l), Hillsdale, NJ: Erlbaum. Hayes, A. M., Castonguay, L. G., & Goldfried, M. R. (1996). Effectiveness of targeting the vulnerability factors of depression in cognitive therapy.Journal of Consulting and ClinicalPsychology, 64, 623-627. Hill, C. E., O’Grady, K. E., & Elkin, I. (1992). Applying the collaborative study psychotherapy rating scale t o rate therapist adherence in cognitive-behavior therapy, interpersonal therapy, and clinical management. Journal of Consulting and ClinicalPsychology, 60, 7379. Horowitz, M. J. (1987). States of mind: Confgurational analysis of individualpsychology. (2nd ed.). New York: Plenum Press. Horvath, A. O., & Symonds, D. B. (1991). Relationship between working alliance and outcome in psychotherapy: A meta-analysis.Journal of Counseling Psychology, 38, 139-149.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

34 Jones, E. E. & Pulos, S. M. (1993). Comparing the process in psychodynamic and cognitivebehavioral therapies.Journa6 of Consulting and Clinical Psychology, GI, 306-316. Merman, G. L., Weissman, M. M., Rounsaville, B. J., & Chevron, E. S. (1984). Interpersonal Psychotherapj ofDqwssion. New York: Basic Books Inc. Krupnick, J. L., Sotsky, S. M., Simmons, S., Moyer, J., Elkin, I., Watkins, J., & Pilkonis, P. A. (1996). The role of the therapeutic alliance in psychotherapy and pharmacotherapy outcome: Findings in the National Institute of Mental Health treatment of depression collaborative research program. Journal of Consulting and Clinical PSycholo~,64, 532-539. Lorr, M. & DeLong, J. (1986). A short form of the interpersonal style inventory (ISI). Journal of Clinical Psychology, 42, 466469. Lorr, M. & DeLong, J. (1984). Second-order factors defined by the ISI. Journal of Clinical Psychology, 40, 1378-1382. Lorr, M. & Manning, T. T. (1978). Higher-order personality factors of the ISI. Multivariate Behavioral Research, 13, 3-7. Lorr, M. & Youniss, R. P. (1973). An inventory of interpersonal style. Journal of Personality Assessment, 32, 165-173. Luborsky, L. (1977). Measuring a pervasive psychic structure in psychotherapy: The core conflictual relationship theme. In N. Freedman & S. Grand (Eds). Communicative structures and psychic structures (pp. 367-3951, New York: Plenum. Luborsky, L., Barber, J. P., & Diguer, L. (1992). The meanings of narratives told during psychotherapy: The fruits of a new observational unit. Psychotherapy Research, 2, 277-290. Luborsky, L. & Crits-Christoph, P. (1990). Onderstanding transference: The CCRT method. New York: Basic Books. McMullen, L. M., & Conway, J. B. (1994). Dominance and nurturance in the figurative expressions of psychotherapy clients. Psychotherapy Research, 4, 43-57. McMullen, L. M., & Conway,J. B. (1337).Dominance and nurturance in the narratives told by clients

CRITS-CHRISTOPH ET AL. in psychotherapy. Psychotherapy Research, 7, 83-99. Mergenthaler, E. & Stinson, C. H. (1992). Psychotherapy transcription standards. Psychotherapy Research, 2, 125-142. Norville, R., Sampson, H., & Weiss, J. (1996). Accurate interpretations and brief psychotherapy outcome. Psychotherapy Research, G, 16-29. Orlinsky, D. E., Grawe, K., & Parks, B. K. (1334). Process and outcome in psychotherapy-Noch Einmal. In A.E. Bergin & S.L. Garfield (Eds.), Handbook ofpsychotherapyand behaviorchange (Fourth Edition). New York: John Wiley. Piper, W. E., Joyce, A. S., McCallum, M., & Azim, H. F. A. (1993). Concentration and correspondence of transference interpretations in shortterm psychotherapy.Journal of Consulting and Clinical Psychology, GI,586-595. Safran, J. D. & Segal, 2. V. (1990). Interpersonal process in cognitive therapy. New York: Basic Books. Schacht, T. E., Binder, J. L., & Strupp, H. H. (1984). The dynamic focus. In H. H. Strupp & J. L. Binder (Eds). Psychotherapy in a new key: A guide to trme-limited dynamic psychotherapy (pp. 65109). New York: Basic Books. Shrout, P. E. & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86 420-428. Silberschatz, G., Fretter, P. B., & Curtis,J. T. (1986). How do interpretations influence the process of psychotherapy? Journal of Consulting and Clinical Psychology, 54, 646652. Stiles, W. B. & Shapiro, D. A. (1994). Disabuse of the drug metaphor: Psychotherapy processoutcome correlations. Journal of Consulting and Clinical Psychologp, G2, 942-948. Weissman, M. M., & Paykel, E. S. (1974). Thedepressed woman: A study of social relationships. Chicago: University of Chicago Press. Young, J. E., Beck, A. T., & Weinberger, A. (1993). Depression. In D. H. Barlow (Ed.), Clinical handbook of psychological disorders: A step-by-step treatment manual (2nd ed., pp. 240-2771, New York: Guilford Press.

Zusammenfassung Ziel der hier beschriebenen Untersuchung war eine Uberprufung des pradiktiven Charakters der Haufigkeit und Vollstandigkeit interpersonaler Narrative in Psychotherapien. Diese Narrative wurden aus 548 Sitzungen von 72 Patienten extrahiert, die entweder an einer kognitiven (CT) oder an einer interpersonalen Therapie (IPT) wegen einer Depression teilnahmen. Es fanden sich konsistente individuelle Unterschiede in der Haufigkeit von Narrativen, ihrer Lange, ihrer Vollstandigkeit und im Hinblick auf die Menge der im Narrativ enthaltenen TherapeutenaulSerungen. Die IPT-Sitzungen enthielten signifikant mehr Narrative als die CT-Sitzungen, wohingegen die CT-Sitzungen einen hoheren Anteil an Therapeutenauserungen pro Narrativ enthielten. Die therapeutische Allianz stand in positiver Beziehung zur Anzahl der Patientenworte pro Narrativ; Patienten mit involvierteren interpersonalen Zielen riefen mehr TherapeutenauBerungen pro Narrativ hervor. Einstellungen im Hinblick auf die

INTERPERSONAL NARRATIVES

35

Ursachen der Depression und Erwartungen, wonach die Fokussierung interpersonaler Themen in der Therapie hilfreich sein konnte, beeinflugten sowohl die Haufigkeit als auch die Vollstandigkeit der Narrative in beiden Therapieformen.

Downloaded By: [University of Pennsylvania] At: 17:13 14 December 2010

Resume Le but de cette investigation a ete d’explorer les prkdicteurs d e la frequence et d e l’exhaustivite de narration interpersonnelles e n psychothkrapie. Les narrations ont ete extraites de 548 seances avec 72 patients ayant beneficie d’une therapie cognitive (TC) ou interpersonnelle (IPT) pour depression. Des differences individuelles consistantes ont kte trouvees concernant la frequence des narrations, leur longueur, leur exhaustivitk, ainsi que le nombre d e mots par narration enounces par le thkrapeute. Plus de narrations par seance ont ete trouvks en IPT, et une plus grande proportion de mots du thkrapeute par narration e n CT. L’alliance a ete positivement correlee avec le nombre d e mots du patient par narration, et les patients au style interpersonnel plus engage ont suscite plus de mots du cBtk du therapeute. Les attentes au sujet des causes de la depression et I’utilite du focus sur des sujets interpersonnels en therapie ont plus influence la frequence et l’exhaustivite des narrations en CT qu’en IPT.

Resumen El prop6sito de esta investigaci6n fue explorar 10s predictores de la frecuencia y la completud de la narrativa interpersonal e n la psicoterapia. Se obtuvieron narrativas d e 548 sesiones de 72 pacientes que recibieron terapia cognitiva (TC) o interpersonal (TIP) para la depresion. Se encontraron diferencias individuales consistentes e n la frecuencia de la narrativa, la longitud, la completud y la cantidad de palabras del terapeuta por narrativa. Las sesiones de TIP contuvieron, en forma significativa, mas narrativas que las sesiones d e TC, las q u e a su vez contuvieron una proporci6n mayor d e palabras del terapeuta por narrativa. La alianza mostr6 una relaci6n positiva con el ndmero d e palabras del paciente por narrativa y 10s pacientes con estilos interpersonales mas comprometidos provocaron mas palabras del terapeuta por narrativa. La bdsqueda de causas de la depresion y la centraci6n en cuestiones interpersonales en la terapia influyeron m i s sobre la frecuencia y la completud d e la narrativa e n la TC que en la TIP.

ReceivedJune 11,1997 Revision Received March 24,1998 Accepted February 23,1998