Effect of Patient Gender on Outcome in Two Forms of Short-Term Individual Psychotherapy

Abstract

This study examined the relationship of patient gender and outcome for two forms (interpretive, supportive) of short-term, individual psychotherapy. Female and male patients (N=89) were randomly assigned to either interpretive or supportive therapy. Outcome was measured in the areas of depression, anxiety, and general symptomatic distress. A significant interaction effect between patient gender and form of therapy was found for measures of depression and general symptomatic distress at post-therapy. Male patients had better outcome in interpretive therapy than in supportive therapy. Female patients had better outcome in supportive therapy than in interpretive therapy. The findings suggest that patient gender may be differentially influential with different forms of short-term therapy.

We live in a gender-conscious society with norms about appropriate modes of behavior for men and women. Regardless of the extent to which we consciously accept or reject such norms, we often act and are often experienced by others in relation to them. This is not to deny that there is variability in the degree to which we are influenced by gender norms, as well as variability in personality and behavior among men and among women.

The effect of gender norms on the quality of the psychotherapy experience remains poorly understood, despite considerable interest reflected in the clinical and research literature. Much of the focus on gender has been directed at examining whether patient gender or therapist gender has an important impact on the outcome of therapy. Some studies have indicated that female patients tended to derive more benefit from therapy than male patients.^1,2 Others have found that patients of both genders benefited more from treatment when it was provided by female therapists.^1,3 Much of the evidence, however, suggests that the association between patient or therapist gender and treatment outcome is weak.^4–7

Others have argued that there may be interaction effects between patient gender and therapist gender, and that focusing solely on one party may not provide meaningful answers.⁸ However, attempts to identify an optimal fit have failed.^2,5,9 Methodological limitations of past research on gender effects in therapy may have contributed to the preponderance of nonsignificant findings. Limitations include use of small sample sizes, use of only female patients, lack of valid and reliable measures of outcome, post hoc collection of outcome data, and failure to consider the form of therapy, such as interpretive or supportive.

Relatively absent from the psychotherapy literature on gender is the issue of whether male and female patients respond similarly to different forms of psychotherapy. No author has described which forms of therapy may be most suitable for male and for female patients. However, a number of writers have argued that male and female patients may prefer or benefit more from different aspects of psychotherapy. For example, Kaplan¹⁰ and Stiver¹¹ have argued that female patients prefer to be listened to and understood in a way that precludes the kind of distancing that may occur in more traditional interpretive models of therapy. These authors posited that female patients prefer to participate in a relationship that is characterized by empathy, affiliation, and affective expressiveness on the parts of both participants. These are qualities that tend to be more characteristic of the patient–therapist relationship in supportive therapies.¹²

Others^13,14 have argued that female patients benefit more from an approach that considers external pressures—such as societal pressures for women to be both homemakers and income earners—and thus permits an understanding of many female patients' sense of inadequacy in the face of these pressures. Such consideration counters the tendency to place responsibility for their problems on themselves. Diminished self-blame can, in turn, free them for more effective problem-solving.¹⁵ This argument suggests that a supportive form of therapy may be more beneficial to female patients because it focuses on external circumstances, encourages problem-solving, and makes use of praise and gratification.

Relevant to male patients are the contentions of Hare-Mustin and Marecek,¹⁶ Kaplan,¹⁷ and O'Neil¹⁸ that a man's sense of self tends to be more reliant on the experience of independence, distinction, and separation from others. Thus, male patients may prefer a form of treatment that provides them with a relationship that allows them to maintain some emotional distance and sense of independence. Such a relationship tends to be more characteristic of interpretive (expressive) therapies.

Male patients may benefit more from certain aspects of therapy. Stiver¹¹ has suggested that the very factors involved in rearing men for independence may lead to an underdevelopment of affective awareness and expressiveness. Male patients typically use coping strategies that involve suppression or denial of their emotions. Thus, interventions that enable them to examine their emotions may be more beneficial in facilitating change.^19,20 Interpretive therapy, with its focus on uncomfortable emotions and intrapsychic conflicts, is more likely to provide male patients with new methods for dealing with their problems and new experiences of expressing and examining their emotions.

Although interesting, this set of ideas concerning what male and female patients may differentially prefer or benefit from in psychotherapy is not based on strong empirical research, but rather on the clinical experience of the various authors. Thus, they must be regarded as speculative.

Our review of the literature found only one study⁶ that examined the interaction between patient gender and form of treatment. The study used follow-up data provided by the National Institute of Mental Health Treatment of Depression Collaborative Research Program. The treatments studied were cognitive-behavioral therapy, interpersonal therapy, imipramine plus clinical management, and placebo plus clinical management. No significant gender effects were found.

The present study used data from a recently completed comparative clinical trial that investigated the efficacy of interpretive and supportive forms of psychotherapy.²¹ The trial was not designed to examine gender effects, and the previous reports of the trial's outcome findings^21,22 have not included investigations of the effect of gender. However, the methodological structure and strengths of the trial provided a good opportunity to examine the influence of gender on treatment outcome. The trial involved random assignment of female and male patients to treatment conditions and balance of patient gender between treatment conditions and among therapists. In addition, the two forms of therapy differed considerably on aspects that the literature suggests may be differentially preferred by and beneficial to male and female patients.

Supportive therapy involved education, advice, praise, and an emphasis on strengths and talents. The supportive therapist actively directed therapy, focused on external circumstances related to the patient's difficulties, encouraged adaptive functioning, and facilitated problem-solving. In contrast, interpretive therapy involved ongoing pressure on the patient to talk, exploration of uncomfortable emotions, and interpretation of internal conflicts. The therapist abstained from providing direct praise and gratification. The patient was responsible for beginning each session and deciding what followed.

Given these differences between the two therapies, the study's methodology, and the suggestions from the literature about different preferences and benefits, there was an opportunity to test two gender hypotheses. They were: 1) male patients will benefit more in interpretive therapy than in supportive therapy, and 2) female patients will benefit more in supportive therapy than in interpretive therapy. Considering the two hypotheses together, the present study investigated whether there was a significant interaction effect between patient gender and form of therapy on treatment outcome.

In addition, the study examined the effect of the interaction of patient gender and form of therapy on the therapeutic alliance. The alliance is a commonly investigated variable in studies that examine the relationship between the process and outcome of treatment. We examined the alliance in the present study as a variable that may possibly mediate the effect of gender on outcome within each form of therapy. That is, do males in interpretive therapy and females in supportive therapy have stronger alliances with their therapists that may account for their better outcomes?

METHODS

Patients

A detailed description of the design and methodology of the comparative trial is presented by Piper et al.²¹ Patients were referred for psychotherapy from a large psychiatric outpatient clinic of a university hospital. After complete description of the study to the subjects, written informed consent was obtained. Patients participated in interview and questionnaire assessments of predictor, demographic, diagnostic, and outcome variables. Patients were matched on personality variables, use of medication, age, and gender and were assigned randomly to interpretive or supportive therapy and to one of eight therapists.

The sample for the present study consisted of 89 patients who completed treatment and provided outcome data at our three assessment times (pre-therapy, post-therapy, 12-month follow-up). Patients in both therapies experienced substantial improvement and did not differ significantly from each other with regard to outcome. Treatment gains were maintained across a 12-month follow-up period. In interpretive therapy, there were 29 females and 13 males. In supportive therapy, there were 29 females and 18 males. As indicated in a previous article,²³ patient gender was not associated with dropping out of treatment.

The 89 patients received diagnoses according to the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised.²⁴ Axis I diagnoses were identified by the computer-administered Mini-SCID²⁵ and validated by an independent clinical diagnosis assigned jointly by the intake assessor and a staff psychiatrist, both of whom saw the patient on the day of intake. Axis II diagnoses were determined by the computer-administered SCID-II Patient Questionnaire and AutoSCID II.²⁶ Rater reliability for Axis II diagnoses (for the interview portion of the SCID-II PQ) was calculated for 10 randomly selected cases and 5 raters. A kappa was calculated for each pair of raters for each disorder. The mean kappa for all pairs and disorders was 0.70.

A total of 67% of the patients received an Axis I diagnosis. The most frequent disorders were current major depression (64%), adjustment disorder (8%), dysthymia (7%), and panic disorder (7%). A total of 60% of the patients received an Axis II diagnosis. The most frequent Axis II disorders were avoidant (18%), obsessive-compulsive (16%), paranoid (14%), dependent (11%), and borderline (10%). A total of 40% of the patients received both Axis I and Axis II diagnoses, and 13% of the patients did not receive either an Axis I or Axis II diagnosis. These patients' clinical presentations and histories did not meet the full criteria for a clinical syndrome or disorder; however, most of the patients received a V-code on Axis I or evidenced traits of an Axis II disorder. Patients with primary problems related to psychosis, substance abuse, or sociopathic behavior were excluded. The patients' presenting problems were consistent with the above diagnostic profile and representative of an outpatient psychotherapy population, comprising difficulties with depression, anxiety, low self-esteem, and interpersonal conflict.

The average age of the patients was 35.1 years (SD=9.9; range 19–62 years). There were 58 women and 31 men. Forty-four percent were married or living with a partner, 20% were separated or divorced, and 36% had never been married. Sixty-four percent were educated beyond high school, and 74% were employed. The racial composition was Caucasian, 96%; East Indian, 2%; Asian, 1%; and Métis (mixed Native Canadian and Caucasian), 1%. Many (76%) reported receiving previous psychiatric treatment, but few (9%) reported a history of psychiatric hospitalization. Female and male patients were not found to differ on diagnostic, demographic, or initial disturbance variables. The absence of a gender effect on some diagnoses reported in the literature to be more prevalent among males or females, such as antisocial and borderline personality disorders, may be the result of the numbers of patients with these disorders being insufficient to detect a gender difference.

Therapists and Therapies

There were eight therapists (three psychologists, two social workers, two occupational therapists, one psychiatrist). Seven were white and one East Indian. Five were female. The therapists' average age was 43.6 years (SD=6.1; range 37–52), and their average experience practicing individual psychotherapy was 11.8 years (SD=4.9; range 3–19). Each therapist treated a similar number of interpretive therapy patients and supportive therapy patients and a similar number of male and female patients. Although we believe that therapist gender may also affect the outcome of therapy, the small number of therapists (particularly males) in our study precluded inclusion of therapist gender as a variable in our statistical analyses.

Each patient received a manualized form of psychotherapy that emphasized interpretive or supportive features. They were labeled interpretive therapy and supportive therapy, respectively. The patient was scheduled for 20 weekly 50-minute sessions at a regular prearranged time. Punctual attendance was emphasized, and missed sessions were not rescheduled. The therapist was paid by a third party (Canadian Healthcare System). Apart from these similarities, the two forms of therapy were quite different.

In interpretive therapy, the primary objective is to enhance the patient's insight about repetitive conflicts (intrapsychic and interpersonal) and trauma that serve to underlie and sustain the patient's problems. The therapist makes use of the here-and-now relationship and attends to linkages with past significant relationships. Relative to supportive therapy, the interpretive therapy situation is more demanding, depriving, and anxiety-arousing. The therapist encourages the patient to explore uncomfortable emotions and withholds immediate praise and gratification. Overall, the therapist is moderately active, interpretive, and transference-focused.

In supportive therapy, the primary objective is to improve the patient's immediate adaptation to his or her life situation. Relative to interpretive therapy, the supportive therapy climate is more relaxing, gratifying, and comforting. The therapist attempts to minimize anxiety and regression in the session, focuses on external circumstances related to the patient's difficulties, and provides praise and immediate gratification. Overall, the therapist is active, noninterpretive, and other-focused (i.e., addresses current external relationships).

Although the therapists were experienced in providing a variety of interpretive and supportive therapies in the clinic, they participated in a six-month training seminar prior to taking cases in the trial. This phase included treating pilot cases and attending a weekly training session where technical principles were covered and cases were presented. The weekly seminar continued throughout the trial. The therapists followed a two-part technical manual that described, illustrated, and compared the technical emphases associated with the two forms of therapy. Therapist compliance with the treatment manual guidelines was monitored with the Interpretive and Supportive Technique Scale,²⁷ which was completed by external observers. The two therapies were well differentiated, as intended.

Outcome Variables

Outcome in three areas was assessed: depression, using the Beck Depression Inventory;²⁸ anxiety, using the Spielberger Trait Anxiety Scale;²⁹ and general symptomatic distress, using the Global Severity Index of the Symptom Checklist-90-Revised.³⁰ These three instruments were chosen because of their wide use in psychotherapy research, their strong psychometric properties, and the availability of substantial normative data, which allowed for the determination of the clinical significance of observed changes. Patients were assessed on each of these measures at pre-therapy, post-therapy, and 12-month follow-up. Scores on each of these measures at each assessment time are provided in Table 1.

TABLE 1. Outcome scores on measures of depression, general distress, and anxiety for males and females in interpretive therapy and supportive therapy.

Therapeutic Alliance

Therapeutic alliance was defined as the working relationship between the patient and therapist. It was assessed by soliciting brief ratings from the patient and the therapist after each session. The patient and therapist each rated six items on a 7-point, Likert-type scale that ranged from “very little” to “very much.” The first four items focus on whether the patient 1) had talked about private important material, 2) felt understood by the therapist, 3) understood and worked with what the therapist said, and 4) felt that the session enhanced understanding. The remaining two items focus on 5) whether the therapist was helpful and 6) whether the therapist and patient worked well together. The six items were averaged across their respective assessments. Principal components analyses of each set of items (patient-rated, therapist-rated) resulted in one patient-rated factor and one therapist-rated factor. Examination of the internal consistency of the two item sets revealed high coefficient alphas for each (0.97 for patient-rated, 0.96 for therapist-rated). An overall alliance score was devised by calculating the average of the six items. Thus, two scores (patient, therapist) served as summary measures of the therapeutic alliance over the entire course of therapy.

RESULTS

All analyses were carried out by using the Statistical Package for the Social Sciences for Windows (version 8). The statistical approach used for the analyses of the data was repeated-measures analysis of variance (ANOVA). For each outcome variable, the interaction between patient gender and form of therapy was examined across all three assessment times (pre-therapy, post-therapy, 12-month follow-up). This three-way interaction effect (gender×treatment×time) was significant for two of the three variables: depression (F=3.60, df=2,83, P=0.032) and general symptomatic distress (F=3.42, df=2,84, P=0.037). The interaction effect accounted for 8% of the variance in depression scores and 7.5% of the variance in general symptom scores. The three-way interaction effect for depression is illustrated in Figure 1. The three-way interaction effect for general symptomatic distress is illustrated in Figure 2.

FIGURE 1.

Interaction between gender and treatment for the Beck Depression Inventory (BDI)Time 1=pre-therapy; Time 2=post-therapy; Time 3=12-month follow-up.

FIGURE 2.

Interaction between gender and treatment for the Global Severity Index (GSI)Time 1=pre-therapy; Time 2=post-therapy; Time 3=12-month follow-up.

In order to more clearly understand the nature of these three-way interaction effects across all assessment times, we examined similar three-way interactions across the two assessment times of the treatment period and then across the two assessment times of the follow-up period. For the treatment period, the test of the interaction for depression produced the result F=6.80, df=1,84, P=0.011, and the test of the interaction for general symptomatic distress produced the result F=5.71, df=1,85, P=0.019. For the treatment period, the findings indicated that male patients improved more in interpretive therapy than in supportive therapy. Conversely, female patients improved more in supportive therapy than in interpretive therapy. For the follow-up period, the P-values for the interaction F tests for depression and general symptomatic distress were 0.177 and 0.076, respectively. Thus, an interaction effect for gender and form of therapy was not found for the follow-up period. As well, no main effects for patient gender or treatment were found.

We also examined the interactions for the period between pre-therapy and 12-month follow-up, disregarding the post-therapy assessment. The P-values for the interaction F-tests for depression and general symptomatic distress were 0.204 and 0.324, respectively. Thus, an interaction effect for gender and form of therapy was not found when the post-therapy assessment was ignored. In addition, no main effects for patient gender or form of therapy were found.

Post hoc analyses were conducted in order to determine whether patients in each of the four groups (males-interpretive; males-supportive; females-interpretive; females-supportive) showed significant change in depression and general distress scores across the treatment period and across the follow-up period. To determine whether change was significant, we conducted paired sample t-tests. To account for multiple comparisons, a Bonferroni-corrected alpha level of 0.006 (0.05/ 8) was used. We found that over the treatment period, results on 6 of the 8 tests were significant, indicating favorable change. Over the follow-up period, none of the test results was significant. The overall pattern indicated maintenance of treatment gains over follow-up.

In addition to exploring whether the gender-by-treatment effects were statistically significant, we examined whether these effects were also clinically important. In treatment research, clinical significance and reliable change have emerged as two additional methods to evaluate the importance of change. Clinically significant change refers to whether the change was clinically important—that is, whether the patient moved into a normal range of functioning. Reliable change refers to whether the change made by a patient exceeds measurement error for the particular assessment tool.

We examined clinically significant change and reliable change for each of the three outcome variables (depression, general symptomatic distress, anxiety) for the treatment period. Procedures for calculating the specific criteria, developed by Jacobson and Truax³¹ and refined by Tingey and colleagues,³² are described in our previous article.²¹ The clinically significant change cutoff criteria were 12.1 for depression, 40.1 for anxiety, and 0.61 for distress. The reliable change criteria were 8.7 for depression, 11.0 for anxiety, and 0.55 for distress. For example, to achieve clinically significant improvement for depression, a patient who was previously above the criterion of 12.1 had to move below it. To achieve reliable change improvement, a patient had to change by 8.7 scale points or more. Using the pre-therapy scores as the base, we determined which patients achieved both clinical and reliable change at post-therapy for each therapy for each outcome variable.

As shown in Table 2, a greater proportion of males made clinically significant and reliable change in interpretive therapy than in supportive therapy. The opposite was true for female patients. A greater proportion of females made clinically significant and reliable change in supportive therapy than in interpretive therapy. The differences in proportions were the greatest for depression and general symptomatic distress. Overall, the findings supported the two gender hypotheses for the treatment period, but not for the follow-up period.

TABLE 2. Proportion of patients who achieved clinically significant and reliable change on each of the three outcome variables at post-therapy.

The effect of the interaction of gender and treatment on the alliance was investigated by using two factorial analyses of variance. The dependent variables were patient-rated alliance and therapist-rated alliance. The predictors were form of therapy, patient gender, and their interaction. No main or interaction effects were found for either alliance variable.

DISCUSSION

The present study revealed a significant interaction between patient gender and form of therapy during treatment for outcome on measures of depression and general symptomatic distress. During the treatment period, male patients in both forms of therapy improved; however, those in interpretive therapy made significantly larger gains. Female patients in both forms of therapy also improved; however, those who received supportive therapy made significantly larger gains. The differential treatment effects for male and female patients during the treatment period were further illustrated by examining the proportion of patients who achieved clinically significant and reliable change on each of the outcome measures.

The interaction effect between patient gender and form of therapy was not found for the follow-up period. During the follow-up period, male patients who received supportive therapy continued to make improvements, eventually reaching nearly the same outcome levels achieved by males in interpretive therapy at post-treatment. In contrast, males who received interpretive therapy did not continue to improve; rather, they essentially maintained their post-therapy outcome levels. During follow-up, female patients from each form of therapy also tended to maintain their post-therapy outcome levels. Thus, female patients who received supportive therapy continued to maintain a more favorable level of outcome.

An attempt to identify a mediator of the gender effect by examining the therapeutic alliance did not produce any significant findings. Clearly, the alliance is not synonymous with patient preferences or measurable benefits.

The psychotherapy field has devoted little attention to the issue of whether male patients and female patients respond similarly to different forms of therapy. Thus, there is no research evidence and only suggestive clinical reports in the literature to help explain the interaction findings of the present study. Nevertheless, some of the ideas concerning how male and female patients differentially prefer or benefit from different aspects of therapy may provide explanations for the findings of the current study.

As reviewed in the introduction, some authors have suggested that female patients may prefer a more collaborative and personal relationship with the therapist and may benefit more from problem-solving and interventions that underscore the influence of external circumstances for current difficulties. These are characteristics consistent with the supportive form of therapy that was provided in our study. For male patients, it has been suggested that a more neutral relationship between the patient and therapist may be preferred and that males may benefit more from interventions that encourage introspection and examination of uncomfortable emotions. These are qualities more consistent with the interpretive form of therapy in our study.

In summary, the writings of these various authors suggest that providing patient–therapist relationships that are consistent with female and male patients' preferences can be expected to facilitate trust and willingness to work. Patients may then work on difficult topics and engage in new coping strategies that otherwise would have been avoided. For female patients, this may involve a greater focus on external problem-solving to counter a ruminative response style that amplifies vulnerability to depression.¹⁵ For male patients, this may involve introspective examination to facilitate greater affective awareness. The result of such work would likely be greater benefit from treatment.

However, it should be emphasized that not all female patients and not all male patients in the present study fit the general interaction pattern that emerged. That is, the interaction of patient gender and form of therapy did not perfectly predict treatment outcome. The preferences and needs of male and female patients as described above were certainly not shared by all males and all females in the study. In addition, other factors besides gender also likely affected therapy outcome.

The findings of the present study differ from those of Zlotnick et al.,⁶ who found that patient gender did not interact with form of therapy. A number of differences between the two studies may account for the differences in findings. First, the patient samples differed. The sample in the present study consisted of a heterogeneous group of outpatients, many of whom received a diagnosis of personality disorder. The patient sample in the Zlotnick study was a more homogeneous group of outpatients who presented with primary difficulties related to major depression. Second, the findings of the present study indicated that the interaction between patient gender and form of therapy was evident for outcome assessed at post-therapy. Zlotnick and colleagues did not examine the effect of the interaction at post-therapy; instead, they focused on follow-up. Thus, the two studies are actually consistent in not finding a significant gender-by-treatment interaction effect during follow-up. Finally, the studies differed on the types of therapy that were compared. The differences between the two therapies in the present study were consistent with suggestions in the literature concerning differences in what male and female patients prefer and find useful. The treatments in the study by Zlotnick et al. may not have differed to the same degree on therapy characteristics that men and women differentially prefer or find useful.

The different patterns of change for male and female patients in each form of therapy that was evident during the treatment period (i.e., the interaction effect) was not found during the follow-up period. One possible explanation for this is that most patients had improved enough to score near the lower end of the outcome scales at post-therapy, leaving little room for additional improvement during follow-up. Thus, there was limited opportunity to detect an interaction effect unless at least one of the groups (e.g., males in supportive therapy) deteriorated substantially. The patients did not deteriorate; rather, they tended to maintain their treatment gains or in some cases even to improve over the 12-month follow-up period.

Another possibility is that given the absence of the intensive treatment experience during follow-up, patients were not able to make further significant changes in their outcome levels from post-therapy to 12-month follow-up. Specifically, in this article we argue that patient trust and willingness to work are facilitated by providing patients with an interpersonal environment where they may work on difficult issues and learn new coping strategies. Without this arena where patients can reflect, receive input, and modify thought and behavior, they will be less likely to make additional substantial changes. The literature consistently demonstrates that during follow-up, patients are more likely to maintain their post-treatment outcome levels than to make further significant improvements.³³

We also found that the interaction effect was not evident when only the pre-therapy and 12-month follow-up assessments were considered. Although the final outcome (i.e., the 12-month outcome) is important, we believe that it is informative to consider the pattern of outcome across all assessments. For example, although male patients in interpretive and supportive therapy achieved similar levels of outcome by 12-month follow-up, patients who received interpretive therapy achieved that level immediately after therapy. It took patients who received supportive therapy another 12 months to achieve that level of outcome. With regard to female patients, those who received interpretive therapy were unable to achieve the outcome level of those who received supportive therapy, even after 12 months.

Several limitations associated with the present study should be acknowledged. With the exception of the therapeutic alliance measure, the study did not use process measures that could elucidate how the therapy process evolved differently for men and women. Such measures would have proved useful in attempting to identify which patient and therapist behaviors were responsible for the different outcomes. As suggested, the differences in outcome for men and women in the two forms of therapy may have been influenced by patients' preferences for a particular type of therapy. Unfortunately, such pre-therapy preferences were not assessed. Generalization is limited by the fact that the majority of the patients and therapists were Caucasian. Finally, the findings reported in this article are based on only one study, and the explanations offered are admittedly speculative. The assumptions made in the literature and in this article concerning preferences and benefits obviously do not apply to all female patients and all male patients. The reader is cautioned against making decisions solely on the basis of these assumptions. However, despite their speculative nature, the ideas provide plausible explanations of the findings. Further research is needed to substantiate or refute these explanations.

Success in psychotherapy is determined by many variables. The significant interaction effects of the current study identified patient gender as a potentially influential variable. The findings suggest that male and female patients may not benefit equally well from the same types of short-term therapy. We hope that our preliminary findings will stimulate further research by others along similar lines.