Predictors of response to group cognitive-behavioral therapy in the treatment of obsessive-compulsive disorder

Available online at

www.sciencedirect.com European Psychiatry 24 (2009) 297e306

Original article

Predictors of response to group cognitive-behavioral therapy in the treatment of obsessive-compulsive disorder* Andre´a Litvin Raffin a,*, Jandyra Maria Guimar~aes Fachel b, Ygor Arzeno Ferr~ao c, Fernanda Pasquoto de Souza a, Aristides Volpato Cordioli d,e a

Graduate Program in Medical Sciences: Psychiatry, Universidade Federal do Rio Grande do Sul (UFRGS), Rua Ramiro Barcelos 2400, CEP 90035-003 Porto Alegre, RS, Brazil b Department of Statistics, Institute of Mathematics, Universidade Federal do Rio Grande do Sul (UFRGS), Av. Bento Gonc¸alves, 9500, Pre´dio 43-111, CEP 91509-900 Porto Alegre, RS, Brazil c Graduate Program, Centro Universita´rio Metodista, Instituto de Porto Alegre (IPA), Rua Cel. Joaquim Pedro Salgado, 80, CEP 90420-060, Porto Alegre, RS, Brazil d Anxiety Disorders Program, Department of Psychiatry, Universidade Federal do Rio Grande do Sul (UFRGS), Rua Ramiro Barcelos, 2400, CEP 90035-003, Porto Alegre, RS, Brazil e Hospital de Clı´nicas de Porto Alegre (HCPA), CEP 90430-180, Porto Alegre, RS, Brazil Received 18 July 2008; received in revised form 10 December 2008; accepted 10 December 2008 Available online 5 February 2009

Abstract Purpose. e To identify the presence of factors associated with treatment outcome in patients under group cognitive-behavioral therapy (GCBT) for obsessive-compulsive disorder (OCD). Subjects and methods. e This study evaluated 181 patients with OCD that attended a 12-session weekly GCBT program. Response criteria were: 35% reduction in Y-BOCS scores and global improvement score of the Clinical Global Impression (CGI)  2 at post-treatment evaluation. Sociodemographic data, OCD characteristics, and treatment data were studied. Results. e In the bivariate analysis, the following variables showed statistical significance ( p < 0.20) to enter the regression model: being woman ( p ¼ 0.074), greater insight ( p ¼ 0.017) and better quality of life (QOL) in all domains before treatment ( p ¼ 0.053), overall severity of disease according to the CGI ( p ¼ 0.007), number of associated comorbidities ( p ¼ 0.063), social phobia ( p ¼ 0.044), and dysthymia ( p ¼ 0.072). In the final regression model, these variables were associated with response to GCBT: female gender ( p ¼ 0.021); WHOQOLBREF psychological domain ( p ¼ 0.011); insight ( p ¼ 0.042); and global improvement score of the CGI severity-scale before therapy ( p ¼ 0.045). Conclusion. e Special attention should be paid to patients with poor insight, increasing the cognitive aspects of the therapy in an attempt to modify the rigidity and fixity of their beliefs. In addition, male patients should be more observed, since they showed lower chance of response to GCBT when compared to women. Patients with more severe global symptoms (CGI) are poorer responders to GCBT, which indicates that not only obsessive-compulsive symptoms (OCS) should be evaluated, since other symptoms, such as depression and anxiety, may affect the treatment; therefore, an attempt to reduce these symptoms, prior to the treatment of OCD, should be considered as an option in some cases. Ó 2008 Elsevier Masson SAS. All rights reserved. Keywords: Obsessive-compulsive disorder; Predictors of response; Cognitive-behavioral therapy; Group therapy

*

There are no conflicts of interest associated with the publication of this article. * Corresponding author. Servic¸o de Psiquiatria/Hospital de Clı´nicas de Porto Alegre, Rua Ramiro Barcelos, 2350, Sala 400N, CEP 90035-903 Porto Alegre, RS, Brazil. Tel.: þ55 51 2101 8294/þ55 51 3308 5161; fax: þ55 51 2101 8493/þ55 51 3308 5232. E-mail address: [email protected] (A.L. Raffin). 0924-9338/$ - see front matter Ó 2008 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.eurpsy.2008.12.001

298

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

1. Introduction Obsessive-compulsive disorder (OCD) is known to be a fairly common disease that will affect 1.6e3.1% of the population at some time in life [33]. It is a chronic disorder that usually affects young people at the end of adolescence, often persists for their whole lives, and has disabling consequences in about 10% of patients [36]. The reduction of obsessive-compulsive symptoms (OCS) with the use of antiobsessional medication is partial, of only about 30e40%, and only 20% of the patients achieve complete remission [19,42]. Behavioral therapy (BT) with exposure and ritual prevention (ERP) techniques has shown to be effective in about 70% of patients who comply with treatment [30,34]. The combination of cognitive and behavioral techniques has provided a better understanding of OCS and may result in a reduction of anxiety and a better compliance with ERP exercises [45]. Vogel et al. [53] concluded that patients receiving additional cognitive therapy usually show significantly low dropout rates. Eddy et al. [15], in a multidimensional meta-analysis of psychotherapy and pharmacotherapy for OCD, concluded that additional research is needed, since OCS persisted at moderate levels even following adequate and effective treatment and no replicable data are available on maintenance of gains for either form of treatment after one year. Because of the difficulties in accessing psychotherapeutic care and of the high cost of psychotherapy, group therapies have been proposed to provide treatment to a larger number of people. Some studies confirmed the efficacy of group cognitive-behavioral therapy (GCBT) [6,11,12,37]. Studies show that group therapy has an efficacy similar to individual therapy in reducing OCS [16,17]. Anderson and Rees [6] in a controlled study to compare an identical cognitive-behavioral therapy (CBT) protocol, delivered either individually or in a group, indicated that individual treatment was associated with a more rapid response; however, both treatments had equivalent rates of recovered participants by brief follow-up (one month). McLean et al. [37] concluded that group ERP was more effective than group CBT at the end of treatment and also after a three-month follow-up. In a recent study, Souza et al. [48] demonstrated the advantage of GCBT over medication, particularly in the reduction of compulsion symptoms and in the percentage of patients who achieve complete remission. However, these were acute effects of treatment e the therapeutic effect of medications could increase with time and, with a higher dose, some patients might have a different outcome. On the other hand, not all patients with OCD benefit from psychotherapy. Some of them, despite improvement, have residual symptoms after therapy completion and are at greater risk of relapse within two years of the end of treatment [8,21]. Raffin et al’s. [43] and Keeley et al’s. [31], proceeding systematic literature reviews focused on predictive factors of treatment results in OCD, found several studies in which factors associated with a better prognosis with BT or CBT in the treatment of OCD, either associated or not with selective

serotonin reuptake inhibitors (SSRI), such as: having a partner [10,49]; greater improvement by the end of treatment [8,21,41]; younger age [21]; being employed [9]; no history of prior treatment [9]; better treatment compliance [13,41]; therapeutic alliance [32]; greater patient motivation for treatment [14,32]; and complete remission of OCS with treatment [8]. Factors frequently associated with a poor prognosis were: sexual and religious obsessions [3]; hoarding [44,46]; psychiatric comorbidities [27,38]; poor insight [27,39]; early onset and chronic course of OCD [37,47]; longer illness [32]; lower income level [50]; greater baseline severity of OCS [7,14,27,32,37,51]; greater family dysfunction and negative family interactions [50]; male sex [7]; higher rates of depression [2,9,21,32,50] and severe comorbid depression [2,10,32]. Meier et al. [38] studied a sample of 45 CD patients who completed a 10-session symptom-focused outpatient GCBT. Only dependent personality traits were identified as risk factors for negative treatment outcome. The objective of the present study was to investigate possible predictors of response to GCBT in patients with OCD. Research involved the study of the following factors: sociodemographics; OCD intrinsic characteristics; cognitive factors; family history; and use of OCD medication during GCBT. 2. Subjects and methods 2.1. Study design A cohort of 181 patients with OCD treated with GCBT was conducted from October 1999 to December 2006 at the outpatient service of the Anxiety Disorders Program (Programa dos Transtornos de Ansiedade e PROTAN) of Hospital de Clı´nicas de Porto Alegre (HCPA), Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil. 2.2. Patients The patients were recruited among the general population by means of lectures, radio and TV interviews, and newspaper ads that offered GCBT to patients with OCD. Inclusion criteria were: diagnosis of OCD according to the DSM-IV-TR criteria [4]; not taking antiobsessional medications or taking an already stabilized dose for at least three months; age between 18 and 65 years; YaleeBrown Obsessive-Compulsive Scale (Y-BOCS) score  16. Exclusion criteria and corresponding number of patients excluded were: severe depression with risk of suicide (n ¼ 12); OCD onset after head trauma (n ¼ 1) or after rheumatoid fever (n ¼ 2); severe disorders that contraindicated group treatment, such as bipolar affective disorder (n ¼ 10), severe social phobia (n ¼ 7), severe eating disorder (n ¼ 4); psychotic disorders (n ¼ 3), severe personality disorders (n ¼ 7), mental retardation (n ¼ 3), or abuse of alcohol or other substances (n ¼ 7); Y-BOCS score < 16 (n ¼ 18); or taking antiobsessional medications at a dose not stabilized for at least three months (n ¼ 16).

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

Therefore, 285 patients were initially evaluated; 90 were excluded for the reasons described above, and 8, although they met inclusion criteria, refused treatment (Fig. 1). A total of 187 patients started GCBT and 9 (4.8%) dropped out: 5 because they felt too much discomfort at the practice of ERP exercises and 4 because of conflict of schedule. Of the 9 dropouts, 3 had at least one evaluation after the beginning of therapy, and data obtained in their last evaluation were included in the end-of-treatment analysis. The other 6 were excluded, and the total number of participants in the study was, therefore, 181. Other studies with parts of the sample of this work have been previously published by our research group of the Anxiety Disorders Program [8,11,12,48]. This study was approved by the Ethics Committee of HCPA under protocol number 04-235, and all participants signed an informed consent term before their participation in GCBT. 2.3. Treatments 2.3.1. Group cognitive-behavioral therapy (intervention) GCBT consisted of 12 two-hour weekly sessions conducted according to a previously developed and standardized protocol, called ‘Therapy Manual’, described in detail in Cordioli et al., 2002 [11]. In the first sessions, patients received information on OCD and were taught to identify its symptoms. In these sessions, patients were provided live exhibitions of ERP. From the fourth session onwards, the use of ERP techniques associated with cognitive techniques is emphasized. In the final sessions, patients were taught selfmonitoring and strategies to prevent relapse. Sessions started with the definition of the agenda, followed by the review of

Fig. 1. Patient flowchart: from recruitment to study completion.

299

individual home tasks, a brief explanation by the coordinator of a topic related to OCD or to CBT, the personalized determination of the new home tasks, and ended with the evaluation of the session by all participants. During the sessions, all participants were stimulated to participate, to exchange experiences, information and suggestions, and to help each other to do the tasks. From October 1999 to December 2006, 25 successive groups were carried out with 5e8 participants each. Each group was coordinated by a different psychotherapist assisted by a cotherapist, at a total of four pairs of psychotherapists (each pair working at different time points). The percentage of positive responses obtained by the therapists was assessed and no significant difference was observed between therapists (c2 ¼ 0.35, DF 2 and p ¼ 0.837). Response criteria were: 35% reduction in Y-BOCS scores and global improvement score of the Clinical Global Impression (CGI)  2. Patients who responded to GCBTand had Y-BOCS scores < 8 and CGI < 2 at the final GCBT evaluation were classified as having achieved complete remission of OCS. Those who had Y-BOCS  8 and CGI ¼ 2 were classified as having achieved partial remission of OCS, but were also classified as responders to GCBT. 2.4. Evaluation and measures Baseline patient evaluation was conducted by an independent psychiatrist and three independent psychologists, who underwent training for the application of scales. Evaluation consisted of a structured clinical interview during which a form was filled out according to a questionnaire about the patient’s symptoms, disease history, previous treatments, and OCD diagnosis according to the DSM-IV-TR [4]. Demographic and socioeconomic data, occupational status, use of medication, and criteria for inclusion in the study were also recorded. The interview was complemented by the application of the Brazilian version of the MINI International Neuropsychiatric Interview [5] to diagnose comorbidities. The following instruments were applied before the beginning of treatment, after the fourth and eighth sessions, and after the conclusion of GCBT (12th session) by four independent, previously trained evaluators. A given patient may have been evaluated by different interviewers. However, a good level of agreement could be observed among all evaluators during training with the instruments used to evaluate the patients in our study.  Y-BOCS [24,25]: a semi-structured interview to evaluate the intensity of OCS. It comprises two subscales: (1) items 1e5: obsessions; (2) items 6e10: compulsions. Each item ranges from 0 (absent) to 4 (extreme), and the total score may range from 0 to 40. We used item 11 to evaluate patient’s insight.  Y-BOCS Symptom Checklist [24,25]: used to identify symptoms to be further evaluated. It provides a list of examples of the most common obsessions and rituals organized in 15 categories or themes (checking, washing, aggressiveness, etc.).

300

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

 CGI [26]: it comprises two scales from 1 to 7 points to be scored by the interviewer. The ranges of the scales are: (1) Global Severity of Disease e from 1 (normal, no symptoms) to 7 (extremely severe symptoms); (2) Global Improvement e from 1 (extremely better) to 7 (extremely worse).  World Health Organization Quality of Life Assessment e Abbreviated Version (WHOQOL-BREF): to evaluate quality of life (QOL) developed by the WHO. It comprises 26 questions, each with a 5-point scale to evaluate four domains: physical, psychological, social, and environmental. It was translated and validated for Portuguese [20]. Of the 26 questions, two are general questions about quality of life, and the other 24 questions of the instrument represent: Physical Domain (seven questions focused on physical well-being); Psychological Domain (six questions: positive feelings; thinking, learning, memory and concentration; self-esteem; body image and appearance; negative feelings; meaningful life/religion/personal beliefs); Social Relations Domain (three questions focused on relationships); Environmental Domain (eight questions focused on comfort, safety, leisure, protection, among others). 2.5. Statistical analysis Quantitative variables were described as means and standard deviations, and categorical variables, as absolute and relative frequencies. The Pearson chi-square test was used to evaluate the association between categorical variables and response to

treatment. The Student’s t-test for independent samples was used to evaluate quantitative variables in relation to categories of response to treatment. Variables with a p-value lower than 0.25 in bivariate analysis were included in the initial logistic regression model [29]. The purpose of this analysis was to evaluate predictors of response to treatment and control for potentially confounding factors. Crude and adjusted odds ratios (OR) and 95% confidence intervals were estimated for variables that remained in the model. Significance was established at 5%, and variables that met this criterion were retained in the final logistic model. All analyses were performed with the software Statistical Package for the Social Sciences 12.0. Risk factors were checked for confounding and colinearity. Confounding variables were included in the multivariable models if covariate inclusion changed the coefficient of any statistically significant variable in the logistic regression model by >10%. Logistic regression model for multivariable analysis used backward deletion method. 3. Results 3.1. Sociodemographic characteristics of sample Of all patients in the sample (n ¼ 181), 133 (73.5%) responded to treatment: 49 (27.1%) achieved complete remission of symptoms and 84 (46.4%), partial remission. The other 48 patients (26.5%) did not respond to GCBT. Age at beginning of treatment, disease duration, marital and occupational status, and education did not have any significant effect on treatment results (Table 1).

Table 1 Sociodemographic characteristics according to response to treatment. Characteristics of sample

Total (n ¼ 181)

Responders (n ¼ 133)

Non-responders (n ¼ 48)

p

Age, M  SD

37.44  12.52

37.86  12.61

36.25  12.30

0.445a

Gender, n (%) Male Female

50 (27.6) 131 (72.4)

32 (24.1) 101 (75.9)

18 (37.5) 30 (62.5)

0.074b

Marital status, n (%) With a partner Without a partner

87 (48.1) 94 (51.9)

65 (48.9) 68 (51.1)

22 (45.8) 26 (54.2)

0.847b

Education, n (%) Primary Secondary Higher

28 (15.5) 91 (50.3) 62 (34.3)

21 (15.8) 62 (46.6) 50 (37.6)

7 (14.6) 29 (60.4) 12 (25.0)

0.223c

Occupation, n (%) Employed Unemployed

88 (48.6) 93 (52.4)

69 (51.9) 64 (48.1)

19 (39.6) 29 (60.4)

0.196b

WHOQOL Physical domain, M  SD Psychological, M  SD Social relations, M  SD Environmental, M  SD

59.57  18.30 54.45  15.75 55.71  19.35 60.95  15.92

61.21  17.74 56.85  15.05 57.33  18.96 62.38  15.08

55.05  17.08 47.78  13.93 51.22  17.58 57.00  15.85

0.039a <0.001a 0.053a 0.038a

M ¼ mean; SD ¼ standard deviation; and WHOQOL ¼ quality of life before beginning of treatment. a Student’s t-test. b Chi-square test with Yates correction. c Pearson’s chi-square test.

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

301

3.2. Severity of OCD symptoms

3.3. Insight

Severity of symptoms was evaluated with the Y-BOCS and global severity score of the CGI. After treatment, overall group mean score in Y-BOCS was 12.06 (7.48), and 2.19 (1.23) in CGI, a decrease of 53.7% in mean Y-BOCS scores and 51.9% in CGI scores. However, only severity of symptoms according to the CGI showed a statistically significant association with response to treatment ( p ¼ 0.007), i.e., patients with lower scores, or whose disease was less severe, had more chances of achieving success in the treatment (Table 2).

Table 2 shows that insight was significantly associated with treatment results ( p ¼ 0.017). Patients with good, very good, or excellent insights were classified as having insight and patients with poor or no insight as ‘‘no insight’’. 3.4. Comorbidities A total of 124 patients (68.5%) had at least one comorbidity associated with OCD. As seen in Table 2, of the predominant

Table 2 Clinical characteristics of patients at baseline and response to treatment. Clinical characteristics

Total (n ¼ 181)

Responders (n ¼ 133)

Non-responders (n ¼ 48)

p

Total Y-BOCS, M  SD Y-BOCS obsession, M  SD Y-BOCS compulsion, M  SD CGI, M  SD

26.07  5.42 12.70  3.12 13.38  2.96 4.55  0.81

26.19  5.13 12.68  3.07 13.44  2.75 4.45  0.72

25.75  6.20 12.75  3.30 13.21  3.49 4.81  0.96

0.695a 0.901a 0.684a 0.007a

Number of comorbidities None One Two or more

57 (31.5) 48 (26.5) 76 (42.0)

45 (33.8) 39 (29.3) 49 (36.8)

12 (25.0) 9 (18.8) 27 (56.3)

0.063b

Comorbidities Dysthymia Social phobia GAD MDE Agoraphobia Panic disorder

35 43 48 43 29 16

14 26 35 29 21 14

21 (43.7) 17 (35.4) 13 (27.0) 14 (29.1) 8 (16.6) 2 (4.16)

0.072c 0.044c 1.000c 0.407c 1.000c 0.243c

Categories of symptomsd Obsessions Aggressive Contamination Symmetry Doubts Somatic Hoarding Magic Religious Sexual

59 (32.6) 142 (78.5) 133 (73.5) 111 (61.3) 92 (50.8) 88 (48.6) 72 (39.8) 59 (32.6) 29 (16.0)

47 (35.3) 103 (77.4) 100 (75.1) 80 (60.1) 70 (52.6) 64 (48.1) 55 (41.3) 43 (32.3) 20 (15.0)

12 (25.0) 39 (81.2) 33 (68.7) 31 (64.5) 22 (45.8) 24 (50.0) 17 (35.4) 16 (33.3) 9 (18.7)

0.258c 0.730c 0.499c 0.713c 0.523c 0.956c 0.583c 1.000c 0.710c

Compulsions Repetition Checking Cleaning Ordering Voidness Slowness Accumulation Counting Mental rituals Avolition Various

120 (66.3) 137 (75.7) 147 (81.2) 128 (70.7) 121 (66.9) 103 (56.9) 84 (46.4) 79 (43.7) 78 (43.1) 75 (41.4) 72 (39.8)

93 (69.9) 104 (78.2) 108 (81.2) 95 (71.4) 86 (64.6) 75 (56.3) 62 (46.6) 58 (43.6) 60 (45.1) 55 (41.3) 53 (39.8)

27 33 39 33 35 28 22 21 18 20 19

(56.2) (68.7) (81.2) (68.7) (72.9) (58.3) (45.8) (43.7) (37.5) (41.6) (34.5)

0.104c 0.266c 1.000c 0.869c 0.388c 0.950c 1.000c 1.000c 0.457c 1.000c 1.000c

Insight No insight Insight

25 (13.8) 156 (86.2)

13 (9.8) 120 (90.2)

12 (25.0) 36 (75.0)

0.017c

M ¼ mean; and SD ¼ standard deviation. a Student’s t-test. b Pearson chi-square test. c Chi-square test with Yates correction. d Item 11 of the Y-BOCS.

(19.3) (23.8) (26.5) (23.8) (16.0) (8.8)

(10.5) (19.5) (26.3) (21.8) (15.7) (10.5)

302

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

comorbidities, only social phobia ( p ¼ 0.044) was significantly associated with therapy results. However, results of patients without any comorbidities showed a trend ( p ¼ 0.063) to better response to treatment. 3.5. Categories of symptoms OCD is characterized by its diversity of clinical presentations. Table 2 shows the categories of symptoms in our sample. Each patient might have different symptoms, and severity of each symptom was not recorded. Only repetition compulsion ( p ¼ 0.104) was included in the final logistic regression model. 3.6. Other clinical characteristics A total of 63 (34.8%) patients were taking antiobsessional medication at a stabilized dose for longer than three months. The use of antiobsessional medications during GCBT did not affect response to treatment. Although these patients had been using antiobsessional drugs at adequate doses for longer than three months, they had clinically relevant symptoms at the beginning of the treatment (mean Y-BOCS score ¼ 24.81  5.05), and intensity of symptoms was similar to that of patients who did not take antiobsessional medication (mean Y-BOCS score ¼ 26.73  5.51). The following variables were not significantly associated with response to GCBT: age at onset of symptoms: 54 patients (29.8%) had an early onset (before 10 years of age); type of onset of OCS (sudden or gradual); course of the disease (with or without deterioration); family history of OCD: 106 patients (58.6%) had at least one second degree relative with a positive history of OCD; and duration of disease: mean duration of 22.8 years (12.4). 3.7. Logistic regression analysis After modeling, the results of multivariate logistic regression analysis were obtained. The variables, OR and 95% confidence intervals are shown in Table 3. Predictors that retained association with response to treatment were: gender ( p ¼ 0.021); WHOQOL-BREF psychological domain ( p ¼ 0.011); insight ( p ¼ 0.042) and CGI severity before GCBT ( p ¼ 0.045). In our study, women had 2.58 times greater odds of responding to treatment than men, and showed a greater rate of complete symptom remission at the end of treatment (32.1% vs. 14.0%). However, no significant differences were found between men and women in severity of symptoms (25.3 vs. 26.4; p ¼ 0.257); frequency of repetition compulsions (74.0% vs. 63.4%; p ¼ 0.239); and early age of symptoms onset (26.0% vs. 31.3%; p ¼ 0.607); the latter variable was also not associated with response to GCBT. The variable severity of OCS measured using the CGI scale had a level of significance that warranted its inclusion in the regression model, which confirmed that it is a predictive variable of response to GCBT. For each one-point increase in

Table 3 Crude and adjusted odds ratio obtained in logistic regression analysis to evaluate predictors of response to GCBT. Variables

ORcrude (95% CI)

ORAdjusted (95% CI)

p

Gender Male Female

1.00 1.89 (0.93e3.84)

1.00 2.58 (1.15e5.80)

0.021

Occupation With Without

1.65 (0.84e3.22) 1.00

1.58 (0.74e3.38) 1.00

0.242

1.04 (1.02e1.07) 1.02 (1.00e1.04) 1.02 (1.00e1.04)

1.05 (1.01e1.09) 0.99 (0.97e1.02) 0.99 (0.96e1.02)

0.011 0.668 0.461

1.00 3.08 (1.29e7.33) 0.57 (0.37e0.87) 1.81 (0.92e3.57)

1.00 2.67 (1.04e6.89) 0.62 (0.39e0.99) 1.89 (0.89e4.02)

0.042 0.045 0.099

WHOQOLa Psychological domain Social domain Physical domain Insighta Without With CGIa Repetition compulsionsa

The outcome reference category is No Response. Variables with significance between 5% and 25% were retained in the model to control for confounding. The area under the ROC curve was 0.759; 95% CI: 0.677e0.841. Hosmer and Lemeshow goodness-of-fit in the final regression model was c2 ¼ 5.96; p ¼ 0.651. a At baseline.

the global severity score of the CGI at the beginning of treatment, the chance of response to treatment decreased 38%. Therefore, the lower the score in this scale, the greater the chances that a patient responds to treatment, which confirms that global disease severity is associated with GCBT results. We found no effect of OCD severity according to Y-BOCS on the results of GCBT ( p ¼ 0.695). The QOL of the patients before treatment was an important predictive factor of response to GCBT. Responders had mean QOL scores 13% greater than non-responders, which indicates a better QOL in all domains. Moreover, patients with complete remission of symptoms had mean WHOQOL-BREF scores about 25% greater than non-responders. However, results of the final models revealed that only the psychological domain remained a predictor of response to GCBT. The concomitant use of antiobsessional medication (SSRI) during GCBT did not affect treatment. Paradoxically, mean YBOCS scores before treatment of patients taking medication were practically the same as those of patients who did not take medication, and response to therapy was also similar in the two groups. Psychiatric comorbidities are frequent in OCD. The presence of comorbid social phobia in OCD was not associated with GCBT results in the final regression analysis, which suggests that it may constitute a confounding factor. A trend towards better response ( p ¼ 0.063) was found for patients who had no comorbidities. Patients with dysthymia tended to achieve partial remission ( p ¼ 0.031) or not to respond to GCBT ( p ¼ 0.072): only four (11.4%) achieved complete remission of symptoms, whereas 17 (48.6%) had partial remission. These associations were found only in the bivariate analysis and did not retain significance in the final logistic regression. The presence of comorbidities was not significantly associated with

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

GCBT results; we believe that such condition is due to intrinsic OCD factors. We also investigated whether the type of OCS was associated with treatment results. Only repetition compulsions tended towards an association with the GCBT result. People with OCD often feel compelled to perform ritualized actions a repeated number of times, such as going over a paragraph or book again and again, putting on and taking off the same piece of clothing repeatedly, urge to wash parts of the body, tying and untying shoelaces, switching lights on and off, standing up and sitting down, praying excessively before going to bed. The person with OCD believes that if the ritual is not performed, according to rules that must be applied rigidly, something dreadful will happen. Only by performing these rituals, OCD sufferers feel they can prevent these imagined dreaded events. However, this variable was not statistically significant in the results of the logistic regression model. 4. Discussion This study investigated whether sociodemographic factors, intrinsic OCD characteristics, cognitive factors, and other factors were associated with response to GCBT. The results related to gender confirm the results reported by Basoglu et al. [7], who found that male sex (n ¼ 27) was a predictive factor for poor prognosis in a randomized clinical trial with 49 patients treated with individual ERP. Castle et al. [10] analyzed predictors of response to BT in 178 patients with OCD and found differences between sexes: male sex (n ¼ 57) and living alone or having no partner were predictors of poor response. Conversely, McLean et al. [37] in one of the few studies that investigated group treatment of OCD with BT or cognitive therapy, did not find any association between sex and response to treatment in a group of 63 patients. Studies suggest that sex may have an indirect role in response to treatment. Female sex was associated with greater severity of depressive symptoms [23], whereas men had a lower age at OCS onset [36], greater baseline severity and greater frequency of repetition rituals [23], as well as a more frequent association with Tourette syndrome or tic disorder [35]. These findings may indicate an effect of biological factors that would make men more vulnerable or more often subject to forms of OCD that are more resistant to psychological treatment. Another possible explanation may be found in possible psychological and cultural differences, which would make men more resistant to the idea of searching assistance and only doing it when symptoms are very severe. Our findings do not support the specific reasons why men have a poorer response to GCBT, reason why it is still an open question. The results about severity of OCS confirm findings reported by Hollander et al. [27], who evaluated 274 patients from eight international centers and found that the severity of symptoms according to Y-BOCS, CGI, or both before the beginning of treatment (BT, SRI) was significantly greater among refractory patients. However, other studies found that patients with more severe symptoms also obtain benefits from therapies although their responses are delayed and frequently in the form of

303

partial OCS remission [7,8,14]. Differently from our findings using the CGI, we found no effect of OCD severity according to Y-BOCS on the results of GCBT. An explanation may be assigned to the fact that Y-BOCS evaluates OCS severity, whereas severity score of the CGI investigates global aspects of severity; therefore, other symptoms, even those that are consequences of OCD, may affect results. Another explanation for that could be found in some conclusions in the following studies: Abramowitz [1], in a quantitative review of the controlled treatment outcome literature for OCD, found differences between the effect size calculated from the clinical evaluations and from the patients’ evaluations. Kobak et al. [34], in a quantitative analysis of the relative efficacy of all five currently available SSRIs and ERP for OCD, showed that effect sizes were larger for studies without a control group or random assignment, for self-reported outcome measures, and varied significantly by method of effect size calculation. Therefore, methodological differences can significantly impact the effect size and results. Studies show that QOL is probably lower among patients with OCD because their symptoms compromise performance at work and in social and academic life [28,49]. Patients with OCD date less, have fewer stable marriages or do not get married, and are more often unemployed than the general population [49]. Patients with a good insight achieved better results in GCBT. Hollander et al. [27] also reported a similar association: poor insight was found in only 10% of their patients who responded to treatment, but in 20% of non-responders. Neziroglu et al. [39] evaluated 20 patients with OCD in a study that investigated overvalued ideas and response to BT or CBT. Patients (n ¼ 10) with higher levels of overvalued ideas (Overvalued Ideas Scale e OVIS  6) showed no changes in the levels of anxiety or depression with treatment, a result that is a factor of poor prognosis for the treatment of OCD with CBT. Patients with a poor insight do not consider their symptoms meaningful or excessive; they may accommodate to symptoms and seek assistance later than other patients. Consequently, poor insight may be associated with longer disease duration or longer treatment time. De Arau´jo et al. [13] found that the best responders to BT were patients who performed ERP exercises at home as early as on the first week of treatment. They also found that a better insight predicted early compliance. Very intense distorted personal beliefs may reflect more severe forms of OCD and, consequently, greater discomfort during the performance of ERP exercises, which may decrease the number of exercises performed at home. Response to medication is low in general [40,48], and most patients achieve only partial remission of symptoms [48]. Studies show that, although ERP, CBT and antiobsessional medications (either combined with psychotherapy or not) are effective in the treatment of OCD, CBT alone is better in reducing OCS than medication [22,40,48]; also, there is no difference in response to CBT or ERP among patients who started antiobsessional medication before treatment and continued taking stable doses during CBT [22,40]. It could be

304

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

speculated that GCBT might bring additional benefits to patients already under psychopharmacologic treatment, and may be prescribed at the beginning or at any stage of treatment with medication. However, because of the uncontrolled study design, this issue cannot be firmly concluded. Patients with OCD and associated psychiatric comorbidities are more frequently economically inactive and report more interference in work and usual activities than patients without comorbidities [52]. Psychiatric comorbidities associated with OCD may also negatively affect the patient’s general conditions. Our data support the hypothesis that patients with major depression may find difficulties in doing the ERP exercises and confirm findings of other studies [10,32]. The patients with dysthymia, similarly to patients with major depression, may find difficulties in doing the ERP exercises, as reported by Abramowitz et al. [2]. Patients with social phobia often avoid situations in which they have to interact with others or perform tasks in the presence of others. Thus these patients were expected to show a trend not to respond to GCBT. However, the fact that they did not avoid the group therapy sessions may suggest that their comorbidity was not severe enough as to influence treatment. Several authors found an association between hoarding [44,46] and sexual/religious obsessions and poorer response to CBT [3,18]. However, our data did not confirm their findings. A possible explanation for these differences in results may lie in the fact that the intensity and severity of symptoms were not taken into consideration in the analyses and that the samples in the different studies were not comparable. One of the limitations of this study is a possible memory bias, as some information was collected in interviews with patients. Another limitation is the fact that the severity of specific OCS was not controlled, which complicated the identification of subgroups with possible common characteristics of response to GCBT. Although this study aimed at evaluating possible predictors of response to GBCT, some studies [8,13,14,41] indicate that these treatments also have long-term influence. Since we have partial follow-up data for our sample, we have not assessed these possible predictors. Some studies [13,41] indicate that performing the home tasks assigned has an important impact on the overall performance in the treatment. We have not measured home task performance specifically; we only performed a subjective analysis, evaluating whether the patient performed or not the assigned task. We used item 11 of the YBOCS to evaluate patient’s insight, but there are other scales that may measure this aspect more accurately (OVIS, BABS). Finally, other factors that affect response to treatment may have been left out in this study. Clinical implications of this study indicate that special attention should be paid to patients with poor insight. In such cases, a more direct treatment may be planned to increase the cognitive aspects of the therapy in an attempt to modify the rigidity and fixity of patients’ beliefs. Psychological education may also be an important instrument and may contribute to increasing the insight of patients and their motivation to comply with ERP exercises. Another important implication derives from the fact that patients with more severe global symptoms (CGI) are poorer responders to GCBT, which

indicates that not only OCS should be evaluated because other symptoms, often resulting from OCD, such as depression and anxiety, may affect the treatment. Special attention should also be paid to male patients, since they showed lower chance of response to GCBT when compared to women. The presence of better QOL before treatment was associated with better response to GCBT, although only the psychological domain remained a significant predictor after the regression analysis. The use of antiobsessional medication during GCBT did not affect response to treatment, indicating that medication prescription should be better evaluated in some cases, mainly for patients who are not tolerant to the side effects. Finally, a higher number of comorbidities associated with OCD showed a trend to a worse response to treatment, indicating that an attempt to reduce symptoms prior to the treatment of OCD, in some cases (dysthymia and major depression, for example), should be considered as an option. 5. Conclusion GCBT to treat OCD is a relatively new and little studied form of treatment. This study provided some important evidence which may offer an additional help in predicting treatment outcomes in group CBT. This is an important contribution to improve the efficacy of public health strategies, since this modality aims at treating a greater number of patients at a lower cost. The same results were found in other studies, however therapeutics were not similar since these studies focused mainly on behavioral therapy alone. The identification of factors associated with improvement of symptoms remains open. The reasons may be associated with the heterogeneity of OCD and of the patient samples used in different studies. Another relevant aspect is the lack of standardization of psychotherapeutic techniques which means an inability to compare results between the different studies. In this case, the size of our sample was also a relevant aspect, as we included 181 participants that followed the same treatment protocol. To sum up, non-specific factors associated with the person of the therapist, the quality of the therapeutic relationship, or the patient’s personal characteristics, such as motivation and ability to withstand failure, may play important roles that remain to be investigated. Acknowledgments The authors wish to acknowledge the Brazilian Coordinating Agency for Advanced Training of Graduate Personnel (Coordenac¸~ao de Aperfeic¸oamento de Pessoal de Nı´vel Superior, CAPES) and the Research Incentive Fund (Fundo de Incentivo a` Pesquisa e Eventos; FIPE) of Hospital de Clı´nicas de Porto Alegre, Brazil, for the financial support granted to the first author. References [1] Abramowitz JS. Effectiveness of psychological and pharmacological treatments for obsessive-compulsive disorder: a quantitative review. J Consult Clin Psychol 1997;65(1):44e52.

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306 [2] Abramowitz JS, Franklin ME, Street GP, Kozak MJ, Foa EB. Effects of comorbid depression on response to treatment for obsessive-compulsive disorder. Behav Ther 2000;31:517e28. [3] Alonso P, Menchon JM, Pifarre J, Mataix-Cols D, Torres L, Salgado P, et al. Long-term follow-up and predictors of clinical outcome in obsessive-compulsive patients treated with serotonin reuptake inhibitors and behavioral therapy. J Clin Psychiatry 2001;62:535e40. [4] American Psychiatric Association [APA]. Manual diagno´stico e estatı´stico de transtornos mentais: DSM-IV-TR. Porto Alegre: Artmed; 2002. [5] Amorim P. Mini-International Neuropsychiatric Interview: validation of a short structured diagnostic psychiatric interview. Rev Bras Psiquiatr 2000;22:106e15. [6] Anderson RA, Rees CS. Group versus individual cognitive-behavioural treatment for obsessive-compulsive disorder: a controlled trial. Behav Res Ther 2007;45(1):123e37. [7] Basoglu M, Lax T, Kasviskis Y, Marks IM. Predictors of improvement in obsessive-compulsive disorder. J Anxiety Disord 1988;2:299e317. [8] Braga DT, Cordioli AV, Niederauer K, Manfro GG. Cognitive-behavioral group therapy for obsessive-compulsive disorder: a 1-year follow-up. Acta Psychiatr Scand 2005;112:180e6. [9] Buchanan AW, Meng KS, Marks IM. What predicts improvement and compliance during the behavioral treatment of obsessive-compulsive disorder? Anxiety 1996;2(1):22e7. [10] Castle DJ, Deale A, Marks IM, Cutts F, Chadhoury Y, Stewart A. Obsessive-compulsive disorder: prediction of outcome from behavioural psychotherapy. Acta Psychiatr Scand 1994;89:393e8. [11] Volpato Cordioli A, Heldt E, Braga Bochi D, Margis R, Basso de Souza M, Fonseca Tonello J, et al. Cognitive-behavioral group therapy in obsessivecompulsive disorder: a clinical trial. Rev Bras Psiquiatr 2002;24:113e20. [12] Volpato Cordioli A, Heldt E, Braga Bochi D, Margis R, Basso de Sousa M, Fonseca Tonello J, et al. Cognitive-behavioral group therapy in obsessive-compulsive disorder: a randomized clinical trial. Psychother Psychosom 2003;72:211e6. [13] De Arau´jo LA, Ito LM, Marks IM. Early compliance and other factors predicting outcome of exposure for obsessive-compulsive disorder. Br J Psychiatry 1996;169:747e52. [14] De Haan E, van Oppen P, van Balkom AJ, Spinhoven P, Hoogduin KA, Van Dyck R. Prediction of outcome and early vs. late improvement in OCD patients treated with cognitive behaviour therapy and pharmacotherapy. Acta Psychiatr Scand 1997;96:354e61. [15] Eddy KT, Dutra L, Bradley R, Westen D. A multidimensional metaanalysis of psychotherapy and pharmacotherapy for obsessive-compulsive disorder. Clin Psychol Rev 2004;24(8):1011e30. [16] Emmelkamp PMG, Bouman TK, Blaauw E. Individualized versus standardized therapy: a comparative evaluation with obsessive-compulsive patients. Clin Psychol Psychother 1994;1:95e100. [17] Falls-Stewart W, Marks AP, Schafer J. A comparison of behavioral group therapy and individual behavior therapy in treating obsessive-compulsive disorder. J Nerv Ment Dis 1993;181:189e93. [18] Ferr~ao YA, Shavitt RG, Bedin NR, Mathis ME, Carlos Lopes A, Fontenelle LF, et al. Clinical features associated to refractory obsessive-compulsive disorder. J Affect Disord 2006;94:199e209. [19] Flament MF, Bisserbe JC. Pharmacologic treatment of obsessivecompulsive disorder: comparative studies. J Clin Psychiatry 1997;58: 18e22. [20] Fleck MP, Louzada S, Xavier M, Chachamovich E, Vieira G, Santos L, et al. Application of the Portuguese version of the abbreviated instrument of quality life WHOQOL-bref. Rev Saude Publica 2000;34:178e83. [21] Foa EB, Grayson JB, Steketee GS, Doppelt HG, Turner RM, Latimer PR. Success and failure in the behavioral treatment of obsessive-compulsives. J Consult Clin Psychol 1983;51(2):287e97. [22] Fleck MP, Louzada S, Xavier M, Chachamovich E, Vieira G, Santos L, et al. Randomized, placebo-controlled trial of exposure and ritual prevention, clomipramine, and their combination in the treatment of obsessive-compulsive disorder. Am J Psychiatry 2005;162:151e61. [23] Fontenelle LF, Marques C, Versiani M. The effect of gender on the clinical features and therapeutic response in obsessive-compulsive disorder. Rev Bras Psiquiatr 2002;24:7e11.

305

[24] Goodman WK, Price LH, Rasmussen SA, Mazure C, Fleischmann RL, Hill CL, et al. The YaleeBrown Obsessive Compulsive Scale. I. Development, use, and reliability. Arch Gen Psychiatry 1989a;46:1006e11. [25] Goodman WK, Price LH, Rasmussen SA, Mazure C, Delgado P, Heninger GR, et al. The YaleeBrown Obsessive Compulsive Scale. II. Validity. Arch Gen Psychiatry 1989b;46:1012e6. [26] Guy W. Clinical global impressions (CGI). In: Guy W, Bonato RR, editors. ECDEU assessment manual for psychopharmacologydRevised. Rockville: National Institute of Mental Health; 1976. p. 217e22. [27] Hollander E, Bienstock CA, Koran LM, Pallanti S, Marazziti D, Rasmussen SA, et al. Refractory obsessive-compulsive disorder: state-ofthe-art treatment. J Clin Psychiatry 2002;63:20e9. [28] Hollander E, Kwon JH, Stein DJ, Broatch J, Rowland CT, Himelein CA. Obsessive-compulsive and spectrum disorders: overview and quality of life issues. J Clin Psychiatry 1996;57:3e6. [29] Hosmer Jr DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons; 1989. [30] Jenike MA. Approaches to the patient with treatment-refractory obsessive compulsive disorder. J Clin Psychiatry 1990;51:15e21. [31] Keeley ML, Storch EA, Merlo LJ, Geffken GR. Clinical predictors of response to cognitive-behavioral therapy for obsessive-compulsive disorder. Clin Psychol Rev 2008;28:118e30. [32] Keijsers G, Hoogduin C, Schaap C. Predictors of treatment outcome in the behavioural treatment of obsessive-compulsive disorder. Br J Psychiatry 1994;165:781e6. [33] Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 2005;62:593e602. [34] Kobak KA, Greist JH, Jefferson JW, Katzelnick DJ, Henk HJ. Behavioral versus pharmacological treatments of obsessive-compulsive disorder: a meta-analysis. Psychopharmacology 1998;136(3):205e16. [35] Leckman JF, Grice DE, Barr LC, de Vries AL, Martin C, Cohen DJ, et al. Tic-related vs. non-tic-related obsessive-compulsive disorder. Anxiety 1994e1995;1:208e15. [36] Lensi P, Cassano GB, Correddu G, Ravagli S, Kunovac JL, Akiskal HS. Obsessive-compulsive disorder. Familial-developmental history, symptomatology, comorbidity and course with special reference to genderrelated differences. Br J Psychiatry 1996;169:107e17. [37] McLean PD, Whittal ML, Thordarson DS, Taylor S, So¨chting I, Koch WJ, et al. Cognitive versus behavior therapy in the group treatment of obsessive-compulsive disorder. J Consult Clin Psychol 2001;69:205e14. [38] Meier S, Fricke S, Moritz S, Hand I, Rufer M. Outpatient behavioural group therapy for obsessive-compulsive disorder e an effective treatment approach? Verhaltenstherapie 2006;16:173e82. [39] Neziroglu FA, Stevens KP, Yaryura-Tobias JA. Overvalued ideas and their impact on treatment outcome. Rev Bras Psiquiatr 1999;21:209e16. [40] O’Connor KP, Aardema F, Robillard S, Guay S, Pe´lissier MC, Todorov C, et al. Cognitive behaviour therapy and medication in the treatment of obsessive-compulsive disorder. Acta Psychiatr Scand 2006;113:408e19. ´ Sullivan G, Noshirvani H, Marks I, Monteiro W, Lelliott P. Six-year [41] O follow-up after exposure and clomipramine therapy for obsessivecompulsive disorder. J Clin Psychiatry 1991;52:150e5. [42] Piccinelli M, Pini S, Bellantuono C, Wilkinson G. Efficacy of drug treatment in obsessive-compulsive disorder. A meta-analytic review. Br J Psychiatry 1995;166:424e43. [43] Raffin AL, Ferr~ao YA, De Souza FP, Cordioli AV. Fatores preditores de resultados no tratamento do transtorno obsessivo-compulsivo com as terapias comportamental e cognitivo-comportamental: uma revis~ao sistema´tica. Rev psiquiatr Rio Gd Sul 2008;30, suppl [online]. ISSN 0101-8108. Available at . Acessed 05Jan2009. [44] Rufer M, Fricke S, Moritz S, Kloss M, Hand I. Symptom dimensions in obsessive-compulsive disorder: prediction of cognitive-behavior therapy outcome. Acta Psychiatr Scand 2006;113:440e6. [45] Salkovskis PM. Understanding and treating obsessive-compulsive disorder. Behav Res Ther 1999;37:29e52.

306

A.L. Raffin et al. / European Psychiatry 24 (2009) 297e306

[46] Saxena S, Maidment KM, Vapnik T, Golden G, Rishwain T, Rosen RM, et al. Obsessive-compulsive hoarding: symptom severity and response to multimodal treatment. J Clin Psychiatry 2002;63:21e7. [47] Skoog G, Skoog I. A 40-year follow-up of patients with obsessivecompulsive disorder. Arch Gen Psychiatry 1999;56:121e7. [48] Souza MB, Isolan LR, Oliveira RR, Manfro GG, Cordioli AV. A randomized clinical trial of cognitive-behavioral group therapy and sertraline in the treatment of obsessive-compulsive disorder. J Clin Psychiatry 2006;67:1133e9. [49] Steketee G. Disability and family burden in obsessive-compulsive disorder. Can J Psychiatry 1997;42:919e28.

[50] Steketee G. Social support and treatment outcome of obsessive-compulsive disorder at 9-month follow-up. Behav Psychother 1993;21:81e95. [51] Steketee G, Eisen J, Dyck I, Warshaw M, Rasmussen S. Predictors of course in obsessive-compulsive disorder. Psychiatry Res 1999;89:229e38. [52] Torres AR, Prince MJ, Bebbington PE, Bhugra D, Brugha TS, Farrell M, et al. Obsessive-compulsive disorder: prevalence, comorbidity, impact, and help-seeking in the British National Psychiatric Morbidity Survey of 2000. Am J Psychiatry 2006;163:1978e85. [53] Vogel PA, Stiles TC, Gotestam KG. Adding cognitive therapy elements to exposure therapy for obsessive-compulsive disorder: a controlled study. Behav Cogn Psychoth 2004;32:275e90.