Strategies Used With Intrusive Thoughts: A Comparison of OCD Patients With Anxious and Community Controls

Strategies Used With Intrusive Thoughts: A Comparison of OCD Patients With Anxious and Community Controls
[Articles]
Ladouceur, Robert1,6; Freeston, Mark H.2; Rhéaume, Josée3; Dugas, Michel J.4; Gagnon, Fabien5; Thibodeau, Nicole5; Fournier, Sarah2
1Psychology Department, Université Laval, Laval, Québec, Canada
2Centre de Recherche Fernand-Seguin, Montreal, Québec, Canada
3Psychology Department, Hôpital Hôtel Dieu de Lévis, Lévis, Québec, Canada
4Psychology Department, Concordia University, Montreal, Québec, Canada
5Psychiatry Department, Centre Hospitalier Universitaire de Québec, Québec, Canada.
6Correspondence concerning this article should be addressed to Robert Ladouceur, Psychology Department, Université Laval, Laval, Quebec G1K 7P4, Canada. Electronic mail may be sent to robert.ladouceur@psy.ulaval.ca.
This study was supported by grants from the Medical Research Council of Canada and the Fonds de Recherche en Santë du Quëbec. We express our appreciation to Martin Provencher and France Blais, who were instrumental in organizing data collection and conducting the interviews. Some of these data were presented at the annual meeting of the Association for the Advancement of Behavior Therapy, Miami, Florida, November 1997.
Received Date: September 25, 1997; Revised Date: May 19, 1999; Accepted Date: June 10, 1999
Abstract
Models of intrusive thoughts attribute a key role to strategies used by people to cope with their unwanted cognitions. In an extension of previous work, the authors conducted structured interviews with 38 people with obsessive–compulsive disorder, 38 people with another anxiety disorder, and 19 healthy volunteers. The interview identified the repertoire of strategies used with the participant's most troubling thought. The 2 clinical groups reported significantly more strategies than the nonclinical group. The clinical groups also reported significantly greater intensity of the thought and their emotional response and lower efficacy for the strategies. People with OCD reported a significantly higher proportion of strategies that were specifically linked to the thought content (as distinct from nonspecific strategies that were only linked sequentially in time). The results identify both common and differential characteristics of intrusive cognition in anxiety disorders.
________________________________________


Unwanted, intrusive thoughts are a widespread phenomenon in both clinical and nonclinical populations. Intrusive thoughts are an important feature of many psychological disorders. However, they are particularly prevalent in the anxiety disorders, and several disorders are partially defined by their presence. Both obsessive–compulsive disorder (OCD) and generalized anxiety disorder (GAD) are characterized by persistent or repetitive thinking, although the specific characteristics of the thoughts may differ (Diagnostic and Statistical Manual of Mental Disorders, 4th edition [DSM–IV]; American Psychiatric Association, 1994). Furthermore, intrusive experiences, including thoughts, are typical of posttraumatic stress disorder (Ehlers & Steil, 1995).
Current models of intrusive cognition, particularly in OCD, attribute an important role to what people do with their intrusive thoughts and how they interpret their experience in controlling these thoughts (e.g., Clark & Purdon, 1993; Freeston & Ladouceur, 1993, 1997; Rachman, 1997; Salkovskis, 1985; Salkovskis, Richards, & Forrester, 1995). We have studied what nonclinical volunteers and patients with obsessions do with their intrusive thoughts and obsessions by using both questionnaires and structured interviews (see later). The goal of the present study is to extend these findings by using structured interviews to compare strategies reported by people with OCD with a clinical group of people with anxiety disorders other than OCD and with a group of community volunteers. We first review the existing literature.
Several studies have been conducted on strategies used with intrusive thoughts reported by nonclinical volunteers by using questionnaires (Edwards & Dickerson, 1987; Freeston & Ladouceur, 1993; Freeston, Ladouceur, Thibodeau, & Gagnon, 1991, 1992; Langlois, Freeston, & Ladouceur, 2000a, 2000b; Purdon & Clark, 1994) and structured interviews (Freeston, Ladouceur, Provencher, & Blais, 1995). Three questionnaire studies have been conducted on groups of people with mixed OCD symptoms compared with a group of nonclinical controls (Amir, Cashman, & Foa, 1997; Calamari & Janeck, 1997; Freeston, Ladouceur, Gagnon, & Thibodeau, 1991). Two studies have been conducted on groups of people with OCD without overt compulsions. In the first study, a structured interview was used but no comparison group (Freeston & Ladouceur, 1997) and in the second study, a group was compared with matched nonclinical controls on a questionnaire measure (Freeston, Ladouceur, Gagnon, & Thibodeau, 1992). Only one previous study (Rachman & de Silva, 1978) has used an interview to compare strategies reported by people with OCD with those reported by nonclinical controls, but the clinical group was small (n = 8). Thus, there have been no studies to date that have compared people with OCD with people with other anxiety disorders or that used structured interviews with a reasonable number of people with OCD with overt compulsions. The overall findings from studies conducted to date may be summarized as follows:
1. Most individuals use a variety of different responses to intrusive thoughts. These strategies include doing nothing, self-reassurance or rational self-talk, analyzing or thinking the thought through, seeking reassurance, replacing the thought with another, performing a mental or concrete action to remove the thought, using a distracting activity, distracting oneself with surroundings, and thought stopping (Edwards & Dickerson, 1987; Freeston, Ladouceur, Gagnon, & Thibodeau, 1991, 1992; Freeston, Ladouceur, Provencher, & Blais, 1995; Freeston, Ladouceur, Thibodeau, & Gagnon, 1991, 1992; Langlois, Freeston, & Ladouceur, 2000a; Purdon & Clark, 1993; Rachman & de Silva, 1978). They may use different strategies with different thoughts (e.g., Freeston, Ladouceur, Gagnon, & Thibodeau, 1991) and different strategies to cope with a given thought (e.g., Freeston & Ladouceur, 1997; Freeston et al., 1995).
2. The choice of response appears to be associated with a variety of factors, such as the intensity of the thought, its appraisal, the context when the thought occurs, other strategies that have been used previously, and mood state (Freeston & Ladouceur, 1997; Freeston et al., 1995).
3. There are broad similarities between strategies used by nonclinical people and those used by obsessive patients, but there are also some differences (Amir et al., 1997; Calamari & Janeck, 1997; Freeston, Ladouceur, Gagnon, & Thibodeau, 1991, 1992).
4. No strategy is consistently more effective than another, although specific strategies may be relatively effective for some people. Many people report great variability in the efficacy of a given strategy (Freeston & Ladouceur, 1997; Freeston et al., 1995).
5. There are specific links between appraisal of intrusions and some types of responses (Freeston & Ladouceur, 1993; Freeston, Ladouceur, Thibodeau, & Gagnon, 1991). For example, thoughts appraised as more egodystonic were associated with greater use of escape or avoidance strategies (Langlois et al., 2000b).
6. Traditional descriptions of cognitive rituals emphasize the stereotyped nature of the mental activity and the necessity to repeat the action according to strict rules. Although activities meeting such descriptions of cognitive rituals were reported by some people with OCD, all people used a wide range of strategies to deal with their thoughts, many of which were not stereotyped nor repeated according to strict rules (Freeston & Ladouceur, 1997; Rachman & de Silva, 1978).
The studies to date indicate that there are at least superficial similarities between strategies reported by people with OCD and those reported by community volunteers. It is therefore likely that other clinically anxious people also use similar strategies to control their thoughts. To understand the possible role of these strategies in the maintenance of intrusive cognition, it is important to study the repertoires of people with other anxiety disorders. It is also important to identify whether repertoires that may be topographically similar (i.e., similar behavior) are used in different ways in different groups. Two new variables are considered: The first takes into account the extent to which a strategy is used on any given occasion (i.e., perseverance), and the second identifies the degree of connection or linkage between the thought's content and the behavior that is performed. Thus, this study is an extension of previous work in that it uses structured interviews to examine new groups and new variables.
In the current study we used a structured interview developed in our earlier work to examine the repertoire of strategies used by a new group of people with OCD (including both participants who reported overt rituals and those who did not). The repertoires reported by the OCD group were compared with those reported by a group of participants with various anxiety disorders and with a group of community volunteers. This design can examine whether the differences observed between the target group and nonclinical controls are due to the fact that the participants are anxious and are consulting for their anxiety or whether there are features that may be specific to OCD. The repertoires were examined as a function of the size, the content, other features such as the efficacy of the strategy, and certain characteristics of the thought when a particular strategy is used. In an extension of previous work, perseverance as measured by the duration or repetition of the strategy was also considered. We expected that intrusive thoughts would be more troubling in both of the clinical groups. Thus, we hypothesized that, compared with nonclinical participants, the OCD and anxious groups would report (a) more strategies than nonclinical volunteers, (b) greater perseverance in using the strategies, (c) greater intensity of the thought and associated emotions, and (d) lower efficacy. We expected that the OCD group would show signs of greater difficulty in controlling thoughts, a stronger link between the thought content and the strategy used, and greater stereotypy in the deployment of strategies. We thus hypothesized that, compared with the anxious group, the OCD group would report (a) greater perseverance in using the strategies, (b) lower efficacy of strategies, (c) greater linkage between strategies and thoughts than the anxious group, and (d) a higher proportion of strategies used in specific sequences and contexts. Our additional objective was to describe and compare the profile of strategies used by participants with an anxiety disorder (OCD or other) and by nonclinical participants, and by participants with OCD compared with those with other anxiety disorders.
Method
Participants and Procedure
All participants were taking part in a series of studies on processes involved in anxiety disorders. The series involved a diagnostic interview, three semistructured interviews, and a comprehensive questionnaire, each addressing separate research questions. The data presented here (apart from the sociodemographic and descriptive clinical data) are from the first semistructured interview on strategies used with intrusive thoughts and are not reported elsewhere.
Participants with anxiety disorders were referred by clinicians for assessment. If the participants wished, their referring professional (family practitioner, psychologist, or psychiatrist) received a written report on the differential diagnosis and symptom profile. Some participants were later offered treatment. Participants in the nonclinical control group were recruited from the community and were paid a small fee for their participation. The community volunteers were selected so that sex, age, and years of schooling were matched at the group level with the clinical groups.
A research assistant contacted the participants. The goals of the study were explained and an interview was scheduled. Informed consent forms were completed. A clinician experienced with the Anxiety Disorders Interview Schedule for DSM–III–R (ADIS–R; Di Nardo & Barlow, 1988) conducted the initial diagnostic interview. We established interrater reliability by having a second clinician rate recorded interviews for all participants. Kappa varied from .66 to 1.00 for all anxiety disorder diagnoses on the clinical groups (M = .84). Kappa for a secondary diagnosis of depression was .75. In cases of disagreement, a case conference with team members was used to assign the primary and secondary diagnoses. Once the diagnosis (or, for the community volunteers, absence of diagnosis) had been established, participants completed three blocks of questionnaires and three structured interviews, each conducted on separate days. The Structured Interview on Neutralization was the first of the three structured interviews.
OCD.
The OCD group was made up of 38 people (53% were women) with a primary diagnosis of OCD. The duration of the disorder since first symptoms was 14.4 years (SD = 10.3). Fifty-eight percent were currently taking psychoactive medication but all were strongly symptomatic with mean severity ratings of 6.5 (SD = 1.1) on the 0–8 scale from the ADIS–R, corresponding to severe symptoms. Secondary diagnoses (anxiety or mood disorders) were identified in 71%, and 18% received a secondary depressive diagnosis. The mean scores for the Beck Anxiety Inventory and Beck Depression Inventory were 18.6 and 18.3, respectively. The mean age was 32.8 years (SD = 10.7), and participants had received an average of 13.9 years (SD = 3.0) of education.
Other anxiety disorders group (ANX).
The anxiety disorders comparison group was originally made up of 49 people with a primary diagnosis of another anxiety disorder, determined using the ADIS–R. Patients with a secondary diagnosis of OCD (n = 9) and one patient each with a primary diagnosis of posttraumatic stress disorder (n = 1) and anxiety disorder not otherwise specified were excluded from all analyses (n = 1). Thus, the final group consisted of 38 people (66% were women). Primary diagnoses were GAD (n = 19), panic disorder (with or without agoraphobia; n = 12), social phobia (n = 6), and simple phobia (n = 1). The duration of the disorder since first symptoms was 17.8 years (SD = 11.7). Seventy-two percent were currently taking psychoactive medication, but all were strongly symptomatic with mean severity ratings of 6.1 (SD = 1.1), corresponding to severe symptoms. The proportions of additional diagnoses were similar to the OCD group: 74% received at least one other Axis I diagnosis (anxiety or depression), and 18% had a secondary depressive diagnosis. The mean scores for the Beck Anxiety Inventory and Beck Depression Inventory were 19.9 and 19.1, respectively. The mean age was 39.7 years (SD = 11.2), and participants had received an average of 14.9 years (SD = 3.1) of education.
Nonclinical control group (NC).
There were 19 people (63% were women) in the nonclinical control group. None received a diagnosis with the ADIS–R, and none were currently in treatment or taking psychoactive medication. The mean age was 34.4 years (SD = 12.1), and participants had received an average of 14.4 years (SD = 2.1) of education.
Instrument
We developed the Structured Interview on Neutralization to study the repertoire of strategies used by people to deal with intrusive thinking (Freeston et al., 1995; Freeston & Ladouceur, 1997). In the interview we asked questions about strategies used with the most frequent intrusive thought. As in our previous work, the most frequent thought was used rather than the most troubling thought because the most troubling thought may be quite infrequent and recall may be more difficult. The most frequent thought was identified with the Cognitive Intrusions Questionnaire (CIQ; Freeston, Ladouceur, Thibodeau, & Gagnon, 1991; Freeston & Ladouceur, 1993). The CIQ was completed before the interview in the presence of the interviewer to (a) identify the target thought, (b) describe in detail the target thought, and (c) encourage the participant to think about the thought while answering a series of questions.
Clear written directives for the interview were available to the interviewer, including general attitudes, how to present the interview, the nature of intrusive thoughts, specific probe questions, and examples of subquestioning. Participants were asked to form the most frequent thought clearly at the start of the interview. The interviewer then used 10 probe questions to elicit examples of strategies used with the target thought. Subquestioning continued until operational descriptions of the strategies were obtained. Further examples were sought until no more were reported. Answers were recorded on a structured interview form, two strategies to a page.
Once the repertoire of strategies had been established, the participant was then questioned about each strategy in turn. For each strategy, the participant was asked the following: (a) whether the thought was used in a specific context, (b) whether the thought was used in a specific sequence, (c) the probability that the thought would come true in real life, (d) the intensity of the intrusion, (e) the mood state when the thought occurred, (f) the intensity of the mood, (g) the immediate efficiency of the strategy in removing the thought or decreasing discomfort, and (h) the number of times that the strategy was repeated or the length of time that the strategy was used. If the participant reported that the thought was used in specific sequences and could describe the sequence, the sequence was recorded as present. Likewise, if the participant reported that the thought was used in specific contexts and could describe the context, the context was recorded as present. The questions on probability, thought intensity, mood intensity, and efficiency (questions 3, 4, 6, and 7) were rated on a 5-point Likert scale from 0 (not at all) to 4 (extremely).
There were two modifications compared with the previous version (Freeston et al., 1995). First, the duration or number of repetitions was asked for each strategy. We initially attempted to assess both duration and number of repetitions. However, pilot testing suggested that participants became confused because strategies may be either short and repeated several times or used once but for a certain period of time. Thus, participants provided estimates of either the duration (minutes) or the number of times the strategy was repeated for each strategy. Second, after all of the strategies had been investigated in detail, participants were asked to name the three most frequently used strategies.
A single interviewer conducted all of the interviews. She had previously conducted over 30 interviews of this type and was unaware of any hypotheses. She was also unaware of group membership. However, for some participants, the likely group membership may have become apparent as the interview progressed. The reliability of all recorded answers (comparing the interview reporting form with a tape of each interview) had previously been established at 97%. Spot checks were conducted at intervals throughout the study.
We modified the scoring grid used in earlier studies (Freeston & Ladouceur, 1997; Freeston et al., 1995) in two ways for this sample. First, we sought to increase sensitivity for overt compulsions, mental checking, avoidance, information seeking, and reassurance seeking. We retained the seven main categories identified in previous studies, although we modified the definitions for four categories. Specifically, physical action was split into distracting behavior and overt superstitious behavior; the latter was ultimately dropped because of very low frequency. Talk about the thought was split into reassurance seeking and distracting conversation. Trying to convince oneself that the thought was unimportant and evaluating the thought as unimportant were combined because of difficulty in obtaining reliable distinctions. Second, we operationalized a new variable in this study, namely, a functional link between the thought and the strategy used. A functional link was recorded as present only when the action taken by the person referred specifically to the thought or to the situation mentioned in the thought. For example, for a thought referring to fear of losing control and attacking a loved one, asking one's spouse whether one is dangerous is specifically linked to the content. Trying to think about what one did on vacation is not specifically linked. However, if the thoughts about the vacation are conjured up as examples of not being dangerous and are used for self-reassurance, the strategy would be coded as linked.
Independent raters who had not conducted the interview categorized the written responses according to strategy type and functional linkage. They were unaware of group membership. There were a total of 1,064 strategies for the entire sample including the 11 participants who were removed from the analyses (see earlier). Interrater reliability on the categorization of five different samples of strategies all exceeded 85%. Kappa for interrater reliability for functional linkage on five different samples of strategies varied from .65 to .72 (interrater reliability = 85–87%). The scoring grid provided the following variables: (a) number of strategies reported, (b) number of different categories (i.e., classes of strategy), (c) number of strategies reported within each category, and (d) presence or absence of a functional linkage for each strategy.
Analyses
The plan of analysis was based on three considerations. First, the study design enabled examination of characteristics that were specific to one clinical group and not to another. Second, two types of hypothesis were postulated: differences between clinical and nonclinical groups, and differences between the anxious and OCD groups. Third, compared with post hoc comparisons, a smaller number of planned contrasts will decrease the chance of Type II error while still controlling for Type I error. Thus, the overall plan was to use planned contrasts rather then omnibus tests. When variables could be meaningfully grouped together, multivariate analyses were conducted first.
Parametric analyses were used for variables based on counts, duration, scale ratings, and proportions of repertoire. In all cases, planned orthogonal contrasts are reported rather than the main effects. The first compared the clinical groups to the nonclinical group, and the second compared the OCD and ANX groups. Each multivariate planned contrast was tested at p < .025. A nonparametric, repeated measures analysis of variance (ANOVA) was conducted for the proportions of participants using each strategy. If the interaction term was significant, chi-square tests were conducted as post hoc pairwise comparisons with Bonferroni correction. All analyses were conducted twice, first for the entire repertoire and then for the three most frequently used (i.e., main) strategies. Only the analyses for the entire repertoire are reported unless there were important differences in the observed results.
Results
Intercorrelations
Examination of the intercorrelation matrix on the interview variables (see Table 1) revealed only two strong and five medium correlations out of a total of 55 correlations. As may be expected, the total number of strategies and the number of different types of strategies (i.e., categories) were strongly correlated, as the former is the upper bound of the latter. Likewise, there was a strong correlation between the intensity of the thought and the intensity of the emotion when the strategy was used, indicating that stronger emotions were associated with more intense thoughts. Medium negative correlations were observed between the efficacy of the strategy and the intensities of the thought and emotion, indicating lower efficacy when thoughts and emotions were more intense. The other medium correlations were between the probability of the thought coming and the number of categories, the intensity of the thought, and the intensity of the emotion.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 1 Intercorrelations Between Interview Variables for Entire Sample
________________________________________
Repertoire Characteristics
Number of strategies and categories.
Two separate ANOVAs were conducted, given that one variable was derived from the other. The clinical groups reported significantly more strategies than the nonclinical group, F(1, 92) = 7.00, p < .025, but there was no difference between the clinical groups. Following categorization, the same pattern of results was observed, with the clinical groups reporting more types of strategy than the nonclinical group, F(1, 92) = 5.82, p < .025. Means are presented in Table 2. Thus, as expected, the clinical groups reported a broader repertoire both in terms of total number of strategies and in number of types of strategies.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 2 Means and Standard Deviations for Number of Strategies and Categories and for Duration and Repetition in the Obsessive–Compulsive Disorder (OCD), Other Anxiety Disorders (ANX), and Nonclinical (NC) Groups
________________________________________
Duration and repetition.
As the unit of measure was different, the duration and repetition data were both standardized across the sample to a mean of zero and a standard deviation of one. A mixed-model, repeated measures ANOVA showed a significant main effect for group, F(2, 90) = 5.53, p < .01, and an interaction between group and duration/repetition, F(2, 90) = 6.49, p < .01. The interaction was broken down with two contrasts within each simple main effect, controlling Type I error for the interaction at [alpha] = .15/4. The effects showed that both clinical groups reported significantly longer duration, F(1, 90) = 5.81, p < .0375, and more repetitions than the nonclinical group, F(1, 90) = 8.18, p < .0375. Furthermore, examination of raw and standardized data in Table 2 shows significantly greater duration in the ANX group than in the OCD group, F(1, 90) = 4.95, p < .0375, and significantly fewer repetitions in the ANX group than in the OCD group, F(1, 90) = 4.48, p < .0375. Thus, as may be expected, both clinical groups showed more perseverant strategy use. However, the strategies were more perseverant in terms of duration in the ANX group, and more perseverant in terms of repetition in the OCD group.
Probability, thought intensity, emotion intensity, and strategy efficiency.
These four variables, rated on the 0 to 4 scale, were each averaged across all strategies and entered in a multivariate analysis of variance (MANOVA; see Table 3). The multivariate clinical versus nonclinical contrast was significant (Wilks's [LAMBDA] = .634), F(4, 89) = 12.85, p < .025. The multivariate OCD versus ANX contrast was not significant (Wilks's [LAMBDA] = .489, F < 1). The clinical groups reported significantly lower efficacy than the nonclinical group, F(1, 92) = 16.03, p < .025. The clinical groups reported significantly greater intensity of the thought, F(1, 92) = 39.75, p < .025, and of the emotion, F(1, 92) = 42.01, p < .025, than the nonclinical group. The clinical versus nonclinical contrast on the probability that the thought's content would come true was nonsignificant (p < .10). An identical pattern of results was obtained for all strategies and for the main strategies used, but the value of the means changed. Thus, participants in the two clinical groups reported that the thoughts were more intense, were associated with more intense emotions, and that the strategies used were less effective.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 3 Means and Standard Deviations for Strategy Ratings and Proportions in the Obsessive–Compulsive Disorder (OCD), Other Anxiety Disorders (ANX), and Nonclinical (NC) Groups
________________________________________
Sequence, context, and linkage.
The proportions of strategies for which sequence, context, and linkage were present were calculated for each participant and entered into a MANOVA. The multivariate OCD versus ANX contrast was significant, Wilks's lambda = .897, F(3, 89) = 3.45, p < .025, but the clinical versus nonclinical contrast was not. The univariate OCD versus ANX contrasts for the proportion of strategies used in a particular sequence or in a particular context were not significant. The great majority of strategies were used in a specific context (.75 to .79) and almost all strategies were used in a particular sequence (.90 to .98). The OCD versus ANX contrast for the proportion of strategies with functional linkage was significant, F(1, 92) = 8.40, p < .025. In fact, the mean in the OCD group was .65 compared with .51 in the ANX group. The mean in the nonclinical group fell between the others at .57. Similar results were observed for the three main strategies only, but the differences observed were more pronounced than for all strategies. The mean in the OCD group was .73 compared with .43 in the ANX group; the nonclinical group fell between them at .67. Thus, linkage was significantly higher in the OCD group than in the GAD group.
Repertoire Content
As in previous studies, the strategies were rank ordered in terms of frequency for the combined sample. This was examined in three ways (see Table 4): First, the percentage of all participants using the strategy (N = 95); second, the percentage of the total pool of strategies (N = 965); and third, the percentage of the pool of main strategies (N = 281). The seven most frequent strategies were reported by at least 40% of participants and represented at least 4% of the total pool. As in previous studies, the seven major categories made up the great majority of both the total pool (73.4%) and of the pool of main strategies (70.8%). The scoring grid would appear superior to the previous version of the grid because only 14% of participants had unclassifiable strategies, and these made up only 1.2% of the total pool.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 4 Repertoire of Strategies Used by the Combined Sample
________________________________________
To compare strategy use across groups, each type of strategy was coded as present or absent. The twelve most frequent strategies were submitted to a nonparametric Group × Category repeated measures analysis (PROC CATMOD; SAS Institute, 1989). The group effect was nonsignificant, [chi]2(2) = 4.06, but the category, [chi]2(12) = 514.67, p < .001, and the interaction term were both significant, [chi]2(24) = 158.83, p < .001. Three 2 × 2 chi-square tests were then conducted that were analogous to unplanned pairwise contrasts. A significance level of .017 was set for each test (.05/3). Percentages are reported in Table 5.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 5 Percentages of Participants Reporting Strategies in the Obsessive–Compulsive Disorder (OCD), Other Anxiety Disorders (ANX), and Nonclinical (NC) Groups
________________________________________
Compared with anxious controls, participants in the OCD group were more likely to engage in overt compulsions, [chi]2(1) = 20.72, p < .017, and mental checking, [chi]2(1) = 5.68, p < .017, and less likely to use distracting activities, [chi]2(1) = 6.79, p < .017, and replacement with a positive thought, [chi]2(1) = 6.79, p < .017. Compared with nonclinical controls, participants in the OCD group were more likely to engage in saying stop, [chi]2(1) = 69.33, p < .017, self-questioning, [chi]2(1) = 6.51, and overt compulsions, [chi]2(1) = 12.21, p < .017, and less likely to do nothing, [chi]2(1) = 6.00, p < .017. Anxious controls also reported more thought stopping than the nonclinical controls groups, [chi]2(1) = 5.98, p < .017. Finally, trying to convince oneself and reassurance seeking were highly endorsed in all three groups, whereas relaxation and distracting conversations were endorsed less frequently in all three groups.
Functional Linkage of Strategies to Thoughts
As noted earlier, a significantly greater proportion of strategies were functionally linked to the thoughts in the OCD group than in the other anxiety disorders group. To examine whether linkage is related to particular characteristics of the thoughts, we reanalyzed repetition and duration, probability, thought and emotion intensity, efficacy, sequence, and context as a function of linkage in a MANOVA (Group × Linkage). The effects of interest are the linkage main effect and the interaction; the group effect simply replicates the original analyses and is not discussed further. The multivariate main effect for linkage was significant, Wilks's lambda = 0.745, F(8, 85) = 3.59, p < .05, but the Group × Linkage interaction was not significant (F < 1).
Univariate ANOVAs were conducted for each variable. Significant linkage effects were found for the following: the number of times an action was repeated, F(1, 92) = 16.83, p < .05; probability, F(1, 92) = 7.57, p < .05; thought intensity, F(1, 92) = 13.89, p < .05; and emotion intensity, F(1, 92) = 8.62, p < .05. There were no significant linkage effects for duration, efficacy, context, or sequence. Thus, linked strategies were repeated more (M = 2.6, SD = 2.1 vs. M = 1.3, SD = 1.9) and were used when the thought was more probable (M = 2.1, SD = 1.0 vs. M = 1.9, SD = 1.0), more intense (M = 2.7, SD = 0.9 vs. M = 2.3, SD = 1.1), and associated with a stronger emotion (M = 2.6, SD = 0.8 vs. M = 2.3, SD = 1.1).
Discussion
As expected, people in the clinical groups reported more strategies and more different types of strategy than nonclinical volunteers. They also deployed the strategies with greater perseverance and rated them as less efficient. The clinical groups reported higher mean thought and emotion intensity than did the nonclinical group but did not rate the possibility of the thought content coming true as more probable. Thus, hypotheses about the clinical versus nonclinical distinction were confirmed, suggesting that intrusive thoughts are more troubling among people with anxiety disorders. People with anxiety disorders develop more complex and more persistent use of strategies to cope with the thoughts, but the strategies are ultimately less efficient. The finding that there was no difference in the probability of the thought coming true may appear counterintuitive because cognitive models generally postulate that people with anxiety disorders overestimate the probability of negative events. One possibility is that the negative events referred to in the thoughts of nonclinical participants may be objectively more probable. In addition, many people with OCD have fluctuating levels of insight. They may have less accurate perceptions when they are in situations that trigger the thought or when the thought is present. However, the interview may invite people with OCD to take a more detached perspective toward the thoughts, consider them more objectively, and thus rate the thoughts as less likely to come true.
The hypothesis that the OCD group would show greater perseverance was partially confirmed, but for repetition of strategies only. The differential results obtained for the repetition or duration of strategies provides some insight into the nature of interference caused by anxious thinking in anxiety disorders. In the OCD group, strategies were repeated more. Any action that must be repeated, even when doing it once is unnecessary, will not only waste time but will also lead to frustration from being compelled to engage in what many people with OCD will admit is ultimately an unnecessary activity. In the anxiety disorder group, in which many people were diagnosed with primary or secondary GAD, people reported that the strategies took significantly longer to execute. Analyzing the thought while doing something else or engaging in an activity to distract oneself may cause less direct interference in the person's life than being compelled to repeat a meaningless act. However, this does not mean that the strategies do not cause interference. First, the mental load of executing simultaneous tasks may lead to fatigue. Second, the obligation to engage in an activity to distract oneself leads to loss of liberty. Third, only partial success in distracting oneself from one's unwanted thoughts may lead to loss of enjoyment of the activity itself. Although interference caused by intrusive thinking in OCD may be more focalized and may necessitate engaging in specific activities such as overt compulsions or mental rituals, attempts to control intrusive thinking there may cause a similar degree of interference in other anxiety disorders but it may be more diffuse and subtle.
Contrary to expectations, the OCD group did not differ from the anxious group on strategy efficacy. Remember that half of the anxious control group were people with a primary diagnosis of GAD, a disorder that is also characterized by difficulty in controlling one's thoughts. Thus, the relative inefficiency of thought control in OCD is not unique, even though it may have different characteristics and origins than inefficiency of thought control in other disorders. The problem in controlling thoughts may well go beyond GAD; it may be a feature of anxiety disorders in general. To the extent that worry may be conceptualized as the central cognitive feature of anxiety (see Mathews, 1990), this finding is, perhaps in retrospect, not so unexpected. The mean efficacy score for the nonclinical group (2.5 vs. 2.40) is similar to that in our previous study (Freeston et al., 1995) and falls somewhere between moderately and efficient. The efficacy in the OCD group in the current study appears somewhat higher than our previous study (2.0 vs. 1.65; Freeston & Ladouceur, 1997). In the previous study, all of the participants had few or no overt compulsions, whereas in the current study, over half of the group reported dominant overt compulsions that may, at least in the short term, appear to provide some control over the intrusive thoughts.
Contrary to expectations, there was no evidence of greater stereotypy in the OCD group in terms of more rules about sequence and context. In fact, the proportions for both variables were high in all three groups, perhaps leading to ceiling effects. Although, for the majority of the analyses, the results for the three main strategies followed closely the results for all strategies, the differences between the groups in the proportion of linked strategies were more pronounced when only the main strategies were considered. In fact, a mean of 73% of the main strategies were linked to the thought's content for people with OCD, but only 43% of the strategies were linked for people with other anxiety disorders. Although people may have extensive repertoires of strategies, it may be important to consider both the entirety of the repertoire and the strategies that people rely on most.
The significantly higher mean proportion of linkage in the OCD group compared with the anxious group is consistent with current models of OCD. In such models, the thought's presence and content are subject to appraisal in terms of responsibility, overimportance of thoughts, and so on (see Freeston, Rhéaume, & Ladouceur, 1996; Rachman, 1997; Salkovskis, 1985; Salkovskis et al., 1995). Thus, seeing the thought as something that needs to be acted on would lead to the content-related actions that were coded as linkage. The lower proportion of linked strategies in the anxious group and the greater use of distraction and thought replacement suggest that the thought is being perceived as contributing to anxiety, but the thought itself does not represent danger. In other words, in the OCD group, the thoughts are the problem (overimportance of thoughts), or having the thought implies that something must be done to prevent harm from happening (responsibility).
The fairly high degree of linkage reported in the nonclinical group was somewhat surprising. Close to two thirds of the strategies reported were functionally linked to the thoughts. This may mean a number of things, but we offer two possibilities. First, the linked strategies may be more akin to instrumental problem-solving strategies. The grid may not have been able to differentiate one type of functional linkage from another, potentially more pathological form. Second, if the strategies are indeed truly linked to the thoughts, we may have some insight into how OCD can develop. That is, acting on intrusive thoughts in a linked manner would then, in fact, be a fairly normal response. In the absence of other contributing factors (such as preexisting beliefs or the activation of preexisting beliefs, disturbed mood, or critical incidents), there are no long-term consequences. However, under some circumstances, or combination of circumstances, such normal actions could be the first step toward the development of OCD.
Strategies that were functionally linked to the thought through their meaning were associated with more intense thoughts and emotions and a greater perceived probability of the thought coming true. Linked strategies were also more likely to be repeated than unlinked strategies. These results support the idea that functional linkage may again be an indicator that the thoughts themselves are the problem and that the person will attempt to neutralize them in a more focalized way.
There were a number of broad similarities in the types of strategies used, but a number of differences emerged. The OCD group differed from both the anxious and nonclinical control groups in terms of the strategies used. Thus, although people may draw on a common pool of strategies, the repertoires vary according to the type of intrusions and the degree of severity. Compared with anxious controls, OCD participants were more likely to engage in overt compulsions and mental compulsions and less likely to use distracting activities and replacement with a positive thought. The first two results are hardly surprising and indeed support the validity of the interview. The last two results are coherent with the notion of linkage. Both distraction and thought replacement are likely to be nonspecific strategies. Results indicate that OCD patients with compulsions also engage in a number of other strategies that are not as easily identifiable as pathological. In some cases, response prevention of overt compulsions may lead to switching to covert strategies, including seemingly harmless coping strategies. However, these strategies may prevent functional exposure or may interfere with challenging underlying beliefs (see Freeston & Ladouceur, 1997). Even in some cases of standard exposure and response prevention for classic compulsions, outcome may be improved by directly targeting covert strategies as well. Applying broader response prevention, targeting all effortful strategies aimed at either the thought (its presence and its content) or the associated anxiety, may be useful in the case of partial response to treatment.
This study provides strong support for the notion that people with OCD, those with other anxiety disorders, and community volunteers all draw on a similar pool of strategies to deal with intrusive thoughts when they occur. Although there are a number of important differences that emerge from this study, the similarities that were observed argue for a general model of intrusive thought that cuts across diagnostic boundaries and across the clinical versus nonclinical distinction. However, superimposed on a general model of intrusive thought are some critical differences that may provide the specificity found in current accounts of OCD and other anxiety disorders such as GAD. There are a number of important implications of a general conceptualization of intrusive thinking. At a clinical level, techniques that have proved successful in the treatment of OCD may be applied with appropriate modification to other groups. We are in no way arguing for an atheoretical approach to the treatment of anxiety disorders. The history of behavior and cognitive therapy is marked by clinical innovations that then provoke empirical study and theoretical debate. At a theoretical level, the existence of common or overlapping psychological processes in the maintenance of putatively distinct anxiety disorders presents a challenge to current nosology. Common processes may appear to threaten the basis of nosological systems that depend on the presence or absence of key features. Identifying common processes is not to argue, for example, that GAD and OCD are the same thing. Rather, it argues for an approach to understanding psychopathology in which the nosological category represents a clearly defined prototype. The implication of such a view is that, between clearly defined poles, various gradations and mixed phenomena can be found that may share some features of both prototypes.
References
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. [Context Link]
Amir, N., Cashman, L., & Foa, E. B. (1997). Strategies of thought control in obsessive–compulsive disorder. Behaviour Research and Therapy, 35, 775–777. [Context Link]
Calamari, J. E., & Janeck, A. S. (1997, March). Reactions to negative intrusive thoughts: Appraisal and response differences. Poster presented at the 17th National Conference of the Anxiety Disorders Association of America, New Orleans. [Context Link]
Clark, D. A., & Purdon, C. (1993). New perspectives for a cognitive theory of obsessions. Australian Psychologist, 28, 161–167. [Context Link]
Di Nardo, P. A., & Barlow, D. H. (1988). Anxiety Disorders Interview Schedule for DSM–III–R (ADIS–R). New York: Phobia and Anxiety Disorders Clinic, Center for Stress and Anxiety Disorders. [Context Link]
Edwards, S., & Dickerson, M. (1987). On the similarity of positive and negative intrusions. Behaviour Research and Therapy, 25, 297–211. [Context Link]
Ehlers, A., & Steil, R. (1995). Maintenance of intrusive memories in posttraumatic stress disorder: A cognitive approach. Behavioural and Cognitive Psychotherapy, 25, 217–249. [Context Link]
Freeston, M. H., & Ladouceur, R. (1993). Appraisal of cognitive intrusions and response style: Replication and extension. Behaviour Research and Therapy, 31, 181–191. [Context Link]
Freeston, M. H., & Ladouceur, R. (1997). What do patients do with their obsessive thoughts? Behaviour Research and Therapy, 35, 335–348. [Context Link]
Freeston, M. H., Ladouceur, R., Gagnon, F., & Thibodeau, N. (1991, November). Les intrusions cognitives: Implications pour le trouble obsessionnel–compulsif [Cognitive intrusions: Implications for obsessive–compulsive disorder]. In R. Ladouceur (Chair), Intrusions cognitives: Différentes perspectives. Paper presented at the XIVème Congrès Annuel de la Société Québécoise pour la Recherche en Psychologie, Trois-Rivières, Québec, Canada. [Context Link]
Freeston, M. H., Ladouceur, R., Gagnon, F., & Thibodeau, N. (1992, June). Intrusive thoughts, worry, and obsessions: Empirical and theoretical distinctions. In P. Salkovskis (Chair), Clinical and non-clinical intrusive thoughts. Paper presented at the World Congress of Cognitive Therapy, Toronto, Ontario, Canada. [Context Link]
Freeston, M. H., Ladouceur, R., Provencher, M., & Blais, F. (1995). Strategies used with intrusive thoughts: Context, appraisal, mood, and efficacy. Journal of Anxiety Disorders, 9, 201–215. Bibliographic Links [Context Link]
Freeston, M. H., Ladouceur, R., Thibodeau, N., & Gagnon, F. (1991). Cognitive intrusions in a non-clinical population: I. Response style, subjective experience, and appraisal. Behaviour Research and Therapy, 29, 585–597. [Context Link]
Freeston, M. H., Ladouceur, R., Thibodeau, N., & Gagnon, F. (1992). Cognitive intrusions in a non-clinical population: II. Associations with depressive, anxious, and compulsive symptoms. Behaviour Research and Therapy, 30, 263–271. [Context Link]
Freeston, M. H., Rhéaume, J., & Ladouceur, R. (1996). Correcting faulty appraisals of obsessional thoughts. Behaviour Research and Therapy, 34, 433–446. [Context Link]
Langlois, F., Freeston, M. H., & Ladouceur, R. (2000a). Differences and similarities between obsessive intrusive thoughts and worry in a non-clinical population: Study 1. Behaviour Research and Therapy, 38, 157–173. [Context Link]
Langlois, F., Freeston, M. H., & Ladouceur, R. (2000b). Differences and similarities between obsessive intrusive thoughts and worry in a non-clinical population: Study II. Behaviour Research and Therapy, 38, 175–189. [Context Link]
Mathews, A. (1990). Why worry? The cognitive function of anxiety. Behaviour Research and Therapy, 28, 455–468. [Context Link]
Purdon, C., & Clark, D. A. (1993). Obsessive intrusive thoughts in nonclinical subjects: I. Content and relation with depressive, anxious and obsessional symptoms. Behaviour Research and Therapy, 31, 713–720. [Context Link]
Purdon, C., & Clark, D. A. (1994). Obsessive intrusive thoughts in nonclinical subjects: II. Cognitive appraisal, emotional response and thought control strategies. Behaviour Research and Therapy, 32, 403–410. [Context Link]
Rachman, S. (1997). A cognitive theory of obsessions. Behaviour Research and Therapy, 35, 793–802. [Context Link]
Rachman, S. J., & de Silva, P. (1978). Normal and abnormal obsessions. Behaviour Research and Therapy, 16, 233–248. [Context Link]
Salkovskis, P. M. (1985). Obsessional–compulsive problems: A cognitive–behavioral analysis. Behaviour Research and Therapy, 23, 571–583. [Context Link]
Salkovskis, P. M., Richards, H. C., & Forrester, E. (1995). The relationship between obsessional problems and intrusive thoughts. Behavioural and Cognitive Psychotherapy, 25, 281–299. [Context Link]
SAS Institute. (1989). SAS/STAT user's guide, version 6 (Vol. 1, 4th ed.) [Computer software manual]. Cary, NC: Author. [Context Link]
________________________________________

Accession Number: 00004468-200005000-00001
Copyright (c) 2000-2007 Ovid Technologies, Inc.
Version: rel10.5.8, SourceID 1.13281.2.32.1.0.2.197.1.4.1.5