BEHAVIOUR RESEARCH AND THERAPY
PERGAMON
Behaviour Research and Therapy 36 (1998) 17±35
A clinical study of spider phobia: prediction of outcome after self-help and therapist-directed treatments Lars-GoÈran OÈst a,*, Britt-Marie Stridh b, Martina Wolf b a
Department of Psychology, Stockholm University, S-106 91 Stockholm, Sweden b Department of Psychology, University of Uppsala, Uppsala, Sweden Received 25 August 1997
Abstract The present clinical study was aimed at investigating predictors of treatment success, attrition and the extent of treatment needed to achieve clinically signi®cant improvement in spider phobic patients. A total of 103 patients were included in the study after a detailed screening interview. There were four treatment conditions; self-help manual, video, group, and individual treatment, which the patients received in a hierarchical order providing they were not clinically signi®cantly improved after the previous treatment. Pre and post each treatment the patients went through a behavioral approach test and ®lled in a number of self-report questionnaires. The results showed that 38 patients dropped out during the manual treatment, and 59 ful®lled the treatments to become clinically improved. The patients achieving clinical improvement after the two self-help treatments were signi®cantly predicted, as was the extent of treatment needed. The signi®cant predictors were credibility of the manual treatment and motivation for psychotherapy in general. # 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction Speci®c phobias are the most common of the anxiety disorders with an estimated lifetime prevalence of 10±15% in the American population (DSM-IV; APA, 1994; Chapman, 1997). Among the speci®c phobias it seems that phobia of `bugs, mice, snakes, and bats', a category which includes fear of spiders, is the most common in the population with a life-time prevalence of 6.1% (Bourdon et al., 1988; Chapman, 1997). The therapy proven to be most successful for phobias is exposure (e.g. Chambless, 1990; Marks, 1987). Most exposure treatments * Author for correspondence. 0005-7967/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 5 - 7 9 6 7 ( 9 7 ) 1 0 0 1 8 - 3
18
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
have been therapist-directed where the therapist usually has weekly sessions with the patient until the problem has been remedied. Recent research has shown that short intensive treatment during a single session produces just as good results as more spaced programs do and could be considered the treatment of choice for speci®c phobias (HellstroÈm and OÈst, 1995; HellstroÈm et al., 1996; OÈst, 1996; OÈst et al., 1998; OÈst et al., 1998; OÈst et al., 1997; OÈst et al., 1992; OÈst et al., 1991). This treatment has also been replicated in spider phobia (Arntz and Lavy, 1993; GoÈtestam, 1994), and there is even earlier research showing that brief treatment of animal phobias is eective (e.g. Bandura et al., 1969). The ecacy of patient-directed exposure treatment has also been investigated, in search of other methods by which to increase the eectiveness and/or reduce the cost of treatment (AlKubaisy et al., 1992; Marks, 1987; OÈst et al., 1991). The contact with the therapist has varied from one visit per week to plan new exposure tasks (Mathews et al., 1981), to no contact whatsoever during the treatment period (Ghosh and Marks, 1987; HellstroÈm and OÈst, 1995; OÈst et al., 1991). The results vary in that some studies found self-directed exposure to be as successful as standard therapist-directed treatment (Ghosh and Marks, 1987), while others found it to do signi®cantly worse (HellstroÈm and OÈst, 1995; OÈst et al., 1991). However, in the HellstroÈm and OÈst (1995) study of spider phobia, the group of patients who used the speci®c manual at the clinic showed somewhat better eects, especially at the 1-yr follow-up when 63% were clinically improved. Across our four studies on spider phobia, 89% of the individually treated, 74% of those obtaining group treatment, and 31% of those receiving the self-help manual treatment were clinically signi®cantly improved (according to stringent criteria) at post-treatment. Since a smaller proportion of spider phobics do quite well with a self-help treatment manual it would be of great clinical interest to investigate if it is possible to predict which patients will improve with this treatment, and which will need therapist-directed treatment in order to achieve clinical improvement. HellstroÈm and OÈst (1996) reviewed prediction studies in various anxiety disorders and concluded that it is doubtful if any consistent predictors have been found. Most studies in this area had included too many predictor variables in relation to the number of patients and no cross validation had been done. Thus, the results can very well be considered as chance ®ndings, which rarely are replicated. This was also the outcome of the HellstroÈm and OÈst (1996) prediction study (with cross validation) on a large sample of speci®c phobias. While diastolic blood pressure was a signi®cant predictor in the ®rst sample, it was not found in the second sample. In that sample credibility of the treatments and expectancy to be improved were signi®cant predictors. Despite this gloomy picture, the aim of the present study was to investigate if there are variables that can predict treatment success and attrition from self-help treatment, as well as the extent of treatment needed to achieve clinically signi®cant improvement. Besides the variables which have been used in previous prediction studies (see HellstroÈm and OÈst, 1996, for a review) the present study included motivation for psychotherapy in general. Keijsers (1994) has developed a brief questionnaire which he used in a study of 53 anxiety disorder patients. The results showed that the degree of motivation explained 33% of the treatment outcome, and was the strongest individual predictor.
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
19
In order to address this question, a clinical design was used in which a large group of spider phobics all received the self-help manual treatment to begin with. Those who were not suciently improved were then oered gradually more therapist-involved treatments until they reached the criteria of improvement decided beforehand.
2. Method 2.1. Subjects The Ss for the study were recruited through advertisements in local newspapers, or were referred by their physicians in the Stockholm and Uppsala counties. There were 149 Ss, all women, who went through a screening interview consisting of a modi®ed version of the Anxiety Disorders Interview ScheduleÐIV (Brown et al., 1994) and a behavioral approach test (see below). In order to participate in the study, the following criteria had to be satis®ed: (1) ful®lling the DSM-IV criteria of speci®c phobia, animal (spider) type (APA, 1994); (2) be between the ages of 18 and 60 yr; (3) be afraid and exhibit avoidance of a number of situations where confrontation with spiders occurred, this being the primary problem for which the patient had sought treatment; (4) a minimum of 1 yr duration of the phobia; (5) be willing to participate in the study for a certain period; (6) be incapable of inserting their hands into a plastic container with a spider (during the behavioral test); (7) have no other psychiatric problems requiring immediate treatment; (8) have no disease of the heart or lungs. Forty-six Ss did not ful®ll the inclusion criteria, 26 of whom performed too much on the behavioral test. This left 103 patients who started the study. Their average age was 29.4 yr (SD = 8.1; range 19±55), and the average age at which the phobia began was 7.9 yr (SD = 4.4; range 4±27). Forty-®ve (43.7%) of the patients were married or living together with a steady partner, 56 (54.4%) were single and two (1.9%) divorced. There were 83 (80.6%) who worked or studied full-time, 16 (15.5%) part-time, while four (3.9%) were unemployed. All of the patients were handicapped by their phobia of spiders in their daily lives or work. 2.2. Design The study had a clinical design with four consecutive treatments, that were gradually more demanding on the therapist's time. However, in order for the patients to do their very best in the self-help treatment (manual) they were not informed about the consecutive design. Instead, they were told that the study was a comparison between three conditionsÐself-help manual, video observation, and therapist-directed group treatmentÐand that allocation to the conditions was random. In fact, all patients got the manual treatment, and those who were not clinically signi®cantly improved (see below) after this treatment were oered the next one, video treatment. Again, those who were not clinically improved after the video treatment were oered the next one, group treatment. Finally, the few who were not clinically improved after the group treatment were oered the last one, individual treatment. When a patient ful®lled the criteria for clinical improvement, irrespective after which treatment, her participation in the study was over, and she was oered to take part in the voluntary
20
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
maintenance program for 6 months. At the same time, she was fully informed about the actual design of the study, and the reasons for not giving this information initially. The research assistants were also prepared to deal with any negative reactions on the part of the patients due to this, but in no case were any displayed. 2.3. Clinically signi®cant improvement In order to decide when a patient had achieved a clinically signi®cant improvement, the criteria set forth by Jacobson et al. (1984) were used. The change from pre- to post-treatment must be statistically reliable, and the post-treatment score must either be within the range of a normal sample or outside the range of the patient group, de®ned as M2 2 SD in the direction of functionality. The cut-o scores applied in the present study were calculated on a total of 174 spider phobics, participants in our four previous studies on one-session treatment of spider phobia (HellstroÈm and OÈst, 1995; OÈst, 1996; OÈst et al., 1991, 1997a). The following criteria were used. 2.3.1. Behavioral approach test (0±12) The change must be 2 points and the cut-o score 10, i.e. touching the spider. 2.3.2. Assessor rating of phobic severity (0±8) The change must be 2 points and the cut-o score 4. 2.3.3. Self-rating of anxiety (0±100) The change must be 10 and the cut-o score 42. However, at pre-treatment assessment, a number of patients stopped the behavioral test at such a low step that they did not experience very much anxiety. When they approached the spider to a much greater extent at post-assessment, their anxiety did not decrease from an already fairly low level. For these patients, the Spider Phobia Questionnaire was used. The change must be 3 points and the cut-o score 16. In order for a patient to be considered clinically improved all three criteria had to be ful®lled at the same assessment. 2.4. Assessment The patients were assessed before and after the treatments they participated in. All the screening-interviews and the behavioral tests were performed by two graduate students, the junior authors, and the therapist was in no way involved with the assessments. 2.4.1. Self-report measures In order to measure the degree of spider phobia, the Spider Phobia Questionnaire (SPQ; Klorman et al., 1974) and the Spider Questionnaire by Watts and Sharrock (1984) were used. The patients also self-assessed on 0±8 scales the degree of fear and avoidance in relation to each of a small, medium and large spider. There was also a 0±8 rating of overall degree of handicap that the patient experienced from her spider phobia. The Fear Survey ScheduleÐIII (FSS; Wolpe and Lang, 1964) was used to measure general phobic tendencies, the State±Trait
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
21
Anxiety Inventory (STAI; Spielberger et al., 1970) was used to measure general anxiety, while the Beck Depression Inventory (BDI; Beck, 1967) and the Beck Anxiety Inventory (BAI; Beck et al., 1988) were used to measure depression and anxiety, respectively. The Nijmegen Motivation List (NML; Keijsers, 1994) was used to assess motivation for psychotherapy. It consists of 12 statements that the patients rate on a 0±3 scale regarding the degree to which they agree with the statement. 2.4.2. Credibility To assess the credibility of the treatments, before the initiation of any therapy, the patients were asked to answer a questionnaire comprising four questions, on a scale of 1±10 (Nau et al., 1974, originally from Borkovec and Nau, 1972). The patients were also asked to estimate, on a scale of 1±10, how successful they expected the treatment to be. It was explained that this was routine procedure in the project and that the treatment itself was being evaluated and not the therapist. 2.4.3. Assessor rating At the close of the screening interview, the assessor rated the severity of the patient's spider phobia on a scale of 0±8; where 0 = free from symptoms, and 8 = extremely severe and disabling, all aspects of normal life aected by the phobia. 2.4.4. Behavioral test A behavioral test was used to assess the patient's avoidance of spiders and anxiety experienced. This test is described in detail in OÈst et al. (1991). The behavioral test was scaled from 0±12; where 0 = refusal to enter the room, and 12 = holding the spider for at least 20 sec. Immediately after having received the instructions for the behavioral test, the patient's self-ecacy was measured by using a questionnaire describing the 13 steps of the test. The patient rated her con®dence regarding being able to complete each step on a 0±100 scale. The level of self-ecacy was calculated as the proportion of tasks rated 20 or higher, and the strength of self-ecacy as the mean rating across the 13 steps. Patients were also asked to rate (0±100) the anxiety they anticipated, and that they actually experienced at the point of interruption. Blood pressure (BP) and heart rate (HR) were measured with a portable digital blood pressure apparatus (model UA-751). 2.5. Procedure 2.5.1. Pre-treatment The Ss who volunteered for the study were contacted by phone to schedule a date for the screening interview, and they were mailed the questionnaires described above. The interview took 60 min and its purpose was to establish whether or not the patient ful®lled the criteria necessary to be included in the study, and to conduct the behavioral test. After the interview, blood pressure (BP) and heart rate (HR) were measured, followed by a 5-min rest period (adaptation), after which they were measured again (baseline). The patient then received instructions regarding the behavioral test where upon the assessor and the patient left the interview room and walked 10 m to conduct the behavioral test in another room. BP and HR were
22
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
recorded once more at the point the patient interrupted the behavioral test. Then followed a brief description of the study and the patient signed an informed consent form. Finally, the patients were told that they had been randomized to the self-exposure manual treatment and this was presented. They were also informed that the research assistant would phone them during the ®rst week to ®nd out how they were progressing and schedule a time for the posttreatment assessment, and that they could contact the assistant should they encounter unexpected problems. 2.5.2. Second assessment The patients performed the behavioral test and answered the questionnaires again. There was also a brief interview concerning how much time they had spent on the self-exposure and their experiences from this treatment. The study had a minimum requirement of 2 hr of self-exposure (not disclosed to the patients) in order to say that the treatment had been given a fair chance, and any patient having done less than this was not oered further treatment. A patient who ful®lled the criteria for clinically signi®cant improvement was presented with the voluntary maintenance program (OÈst, 1989b) and instructed how to continue her own self-exposure. A patient who did not ful®ll the improvement criteria was oered the video treatment and paid 100 SEK (approximately $15) deposit for the video tape. 2.5.3. Third, fourth and ®fth assessment The patients handed in the ®lled out self-report questionnaires, went through a brief interview about the previous treatment, and then the behavioral test was performed. Those who were clinically improved were presented with the maintenance program, and those who were not improved were oered group treatment and individual treatment, after assessment three and four, respectively. All patients who participated in the ®fth assessment were clinically improved, and were presented with the maintenance program. 2.6. Treatments 2.6.1. Manualized self-exposure The patients received a 30-page manual developed by OÈst and Salkovskis speci®cally for spider phobia and used in two studies (HellstroÈm and OÈst, 1995; OÈst et al., 1991). The manual contains a brief description of phobias and treatment of phobias, as well as what is required from the patient to use the manual. The main part of the manual consists of gradual exposure and the patient follows instructions on what to do in the exposure and how to rate anxiety experienced during the various procedures. The patients were instructed to secure the help of a relative or a friend to ®nd at least three spiders in dierent sizes to expose themselves to in their own homes. They were also strongly recommended to do at least two 2-hr exposure sessions during a 14 day period. 2.6.2. Video treatment The patients received a video tape (120 min) of an authentic individual treatment of a spider phobic woman, with the ®rst author as therapist. They were instructed to watch the video tape at least once in order to get new ideas on how to carry out their own self-exposure treatment.
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
23
2.6.3. Group treatment The therapist met the groups of patients for 30 min immediately before the treatment started. The purposes of this meeting were ®rstly, that the patients should share some of their experiences concerning spiders, and secondly, that the therapist should describe how the treatment was going to be carried out. Based on the description given by OÈst (1989a), the 1-session therapist-directed exposure includes a combination of gradual exposure and modelling, and is described in detail in OÈst et al. (1991) and OÈst (1997). In the present study, the treatment was carried out in groups of eight patients during a 3-hr session. The patients were seated in a circle, and in front of each there was a small table with a plastic container on it. The therapist sat on an oce chair on wheels to enable him to move around easily and help the patient most in need of his assistance at any time. For each patient there was a set of three spiders of gradually increasing sizes. The following pre-treatment instructions were given. The patients are told that the treatment is carried out as team-work, and both the therapist and the patients have equal responsibility for achieving a good result. Moreover, the therapist will never do anything without ®rst describing it to the patient, then demonstrating it, and ®nally getting the patient's permission to do it. The therapist also informs the patients that even if the treatment includes exposing themselves to a much greater extent than they ever have done in natural situations, this will not `beat their personal record' of anxiety in the phobic situation. Finally, the advantages of group treatment, especially the vicarious fear reduction, are emphasized. The therapist usually demonstrates the dierent activities with one patient, while the others observe this. Then all the patients are instructed to perform the same activity and the therapist helps, by using physical guidance, those who need assistance. Usually, all the patients need the therapist's help, but to varying degrees. In some groups it happens that one of the patients spontaneously acts as a `co-therapist' by helping the one sitting next to her when the therapist is occupied with one of the other patients. When the patients can handle the smallest spider with a clearly reduced anxiety it is time to proceed to the next, and so on until the patients have been exposed to all three spiders. The timing is geared by the progress of the average patient in the group, and it sometimes happens that one patient cannot proceed to the next spider, but has to continue working with the smaller one. The goal is that all patients should be able to have two spiders (2 and 3 cm in size) walking on their hands. 2.6.4. Individual treatment The individual treatment was tailored to the needs of each patient and depending on how far she had progressed during the group treatment. It had the same components as the group treatment but took only 20±60 min, since all the patients had had 3 hr of group treatment 2±3 weeks previously. 2.7. Therapist The therapist conducting the group and individual treatments was the ®rst author of this paper. At the time of the study, LGOÈ had 27 yr of clinical experience with behavior therapy, and had treated eight small and six large groups before the start of this study.
24
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
3. Results 3.1. Distribution of patients in the ¯ow-chart Figure 1 presents the number of patients who entered, dropped out, were clinically improved, and not improved, respectively, at the consecutive phases of the study. A total of 103 started with the self-exposure manual; 65 of whom completed the treatment and 38 dropped out. 3.2. Patients excluded due to their behavioral test performance Twenty-six patients (20.2%) of those who ful®lled the general inclusion criteria were excluded from participation because they performed too much of the behavioral test, i.e. they managed to put their hands into the plastic bowl where the spider was kept. This group was compared with the 103 participants and had signi®cantly (P < 0.001) lower scores on the spider phobia questionnaires, the self-rated anxiety during the behavioral test, and the assessor rating of severity, and signi®cantly (P < 0.001) higher behavioral scores and self-ecacy. On the screening interview they also had signi®cantly lower distress and impairment scores. However, there were no dierences on the physiological measures at the behavioral test and on the psychopathology self-report measures. 3.3. Attrition Thirty-eight patients dropped out during the manual treatment phase, one during the video treatment phase, and ®ve never showed up for the group treatment, making a total of 44 drop outs. Attempts were made to contact these patients by phone and 34 (77.3%) were interviewed this way. When asked about the primary reason for not completing the treatment, the following answers were given: nine (26.5%) said they needed a therapist during the treatment, eight (23.5%) had encountered practical diculties like not being able to ®nd spiders, six (17.6%) had private reasons, four (11.8%) said that the treatment was too time-consuming, four (11.8%) that working with the spiders was too frightening, two (5.9%) felt that it was not that important to get rid of their phobia, and one (2.9%) was satis®ed with her result and did not want to have any further treatment. The patients were also asked if their fears of spiders had changed in any way. None of the patients reported any deterioration as a consequence of participating in the study. Eight (23.5%) patients reported no change, 17 (50%) reported a small, ®ve (14.7%) a moderate, and four (11.8%) a major positive change. The latter group could handle spiders of various sizes with few, or no anxiety reactions at all. 3.4. Outcome after the respective treatments To test whether there are signi®cant treatment eects after each of the consecutive treatments, data from the ®ve assessments were used, and only for the 59 patients who remained in treatment. The patients were grouped according to which treatment led them to ful®ll the criteria for clinically signi®cant improvement. Thus, there were 16 (27.1%) in the manual con-
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
25
Fig. 1. Flow-chart describing the dierent phases of the design. The ®gures indicate the number of patients at each step (A = assessment, CSI = clincally signi®cant improvement).
dition, ®ve (8.5%) in the video condition, 26 (44.1%) in the group treatment condition, and 12 (20.3%) in the individual treatment condition. Two-way ANOVAs with group (4) as between-factor and assessment (2) as within-factor (only the ®rst and the last assessment point for each group) were performed. A signi®cant time factor was followed-up with dependent t-tests within each group. Finally, one-way ANOVAs between the groups at each assessment point were carried out. 3.4.1. Behavioral test measures Table 1 presents the outcome on the behavioral test measures at the ®ve assessment points. The two-way ANOVAs yielded signi®cant time eects on all the measures, and the subsequent t-tests showed that all four groups displayed signi®cant changes on all measures, except for the
26
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
Table 1 Means for the respective conditions and assessment points on the behavioral test Variable
Condition
A1
A2
A3
A4
A5
Behavioral score
Manual Video Group Individual F-valueb
5.7 6.0 4.5 5.3 1.7
11.0 8.6 6.9 6.8 21.3$
Ð 10.0 7.7 7.8
Ð Ð 10.4 9.0
Ð Ð Ð 10.5
Self-rated anxiety
Manual Video Group Individual F-valueb
70.6 85.0 80.6 71.2 1.4
33.4 32.0 56.9 59.2 4.4**
Ð 56.0 56.3 66.5 0.6
Ð Ð 29.4 33.3 0.3
Ð Ð
Self-ecacy level
Manual Video Group Individual F-valueb
62.6 67.8 63.7 65.4 0.2
93.8 90.8 66.7 66.7 22.6$
Ð 95.4 72.5 70.6 13.2$
Self-ecacy strength
Manual Video Group Individual F-valueb
49.3 51.0 47.2 54.4 0.5
86.6 80.8 56.3 58.4 19.2$
Assessor rating of severity Manual Video Group Individual F-value2
5.5 5.0 5.9 5.4 1.1
1.0 3.4 4.7 4.6 45.9$
t-value 7.9$ 4.8** 13.8$ 8.1$ 3.9*
F-valuea G: 2.6 T: 325.7$ I: 1.1 35.1$
22.9
4.8*** 1.4 9.6$ 6.1***
G: 3.0* T:130.6$ I: 1.2
Ð Ð 96.0 88.7 5.8*
Ð Ð Ð 98.8
7.4$ 3.5* 7.3*** 9.0$
G: 0.5 T: 162.6$ I: 0.4
Ð 88.6 65.2 63.8 7.7***
Ð Ð 85.2 78.3 5.2*
Ð Ð Ð 92.5
7.5$ 4.3* 8.5$ 10.0$
G: 2.3 T: 218.7$ I: 6.4***
Ð 1.4 4.2 4.0 24.0$
Ð Ð 1.2 3.1 55.8$
Ð Ð Ð 1.2
8.3$ 4.4* 18.4$ 9.5$
G: 0.9 T: 425.9d I: 0.7
a
G = group, T = Time, I = Interaction eect in the two-way ANOVA. One-way ANOVA. * P < 0.05, **P < 0.01, ***P < 0.001, $P < 0.0001. b
self-rated anxiety. On this measure the video group did not achieve a signi®cant improvement between the ®rst and the third assessment. However, this group's mean at the second assessment (32.0) was signi®cantly lower than at the ®rst assessment. The one-way ANOVAs at the dierent assessment points (A1±A4) showed that the groups did not dier at pre-treatment on any of the measures. At A2 the manual group diered from the other groups on the behavioral score and the assessor rating of severity, and both the manual and the video groups diered from the therapist-treated groups on the self-rating of anxiety and the two self-ecacy measures. At A3 the video group diered from the therapist-treated groups on the behavioral score, the self-ecacy measures and the assessor rating of severity, while there was no dierence on self-rated anxiety. Finally, at A4 the group treatment condition diered from the individual treatment on all measures but self-rated anxiety.
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
27
3.4.2. Correlations The correlations between the behavioral score on one hand and experienced anxiety and strength of self-ecacy on the other are presented in Table 2. At the ®rst two assessments both anxiety and self-ecacy correlate signi®cantly with behavioral scores, but from A3 only selfecacy does. However, when partial correlations are computed and the eects of self-ecacy are controlled for, the correlation between anxiety and behavioral score is signi®cant only at A1. On the other hand, when anxiety is controlled for, the correlation between self-ecacy and behavioral score is signi®cant at all assessments. 3.4.3. Physiological measures The results on the physiological measures are presented in Table 3. The two-way ANOVAs indicated signi®cant time eects on systolic and diastolic blood pressure, but not on heart rate. The subsequent t-tests indicated that on SBP only the manual and group treatment conditions showed signi®cant improvements, while on DBP the manual and individual treatment conditions did so. 3.4.4. Spider phobia questionnaires Table 4 presents the results on the spider phobia questionnaires. The two-way ANOVAs yielded signi®cant time eects for all the measures and the subsequent t-test also showed that all groups improved signi®cantly from pre to the last assessment. The one-way ANOVAs showed no dierences between the groups at pre-treatment assessment. At A2 there were signi®cant F-values on all measures. The manual group had signi®cantly lower means than the therapist-treated groups on all measures. At A3 the video group was signi®cantly lower than the therapist-treated groups on SPQ and self-assessed avoidance, while at A4 no dierences emerged. 3.4.5. Psychopathology questionnaires Table 5 displays the results on the psychopathology questionnaires. The two-way ANOVAs yielded signi®cant time eects on all measures, and the subsequent t-tests showed signi®cant pre±post changes for all groups on BDI, STAI-Trait, and FSS-III. On STAI-State all but the video group improved, and on BAI only the therapist-directed treatment groups changed sigTable 2 Product-moment and partial correlations between behavioral scores, anxiety, and strength of self-ecacy measures at the dierent assement points Product-moment correlations Variable df Anxiety Self-ecacy
A1 99
A2 61
ÿ0.39$ 0.59$
ÿ0.45$ 0.84$
A3 44
Partial correlations A4 34
ÿ0.20 ÿ0.20 0.80*** 0.52***
* P < 0.05, **P < 0.01, ***P < 0.001, $P < 0.0001.
A5 8 0.02 0.72**
A1 99 ÿ0.35$ 0.38$
A2 61 ÿ0.05 0.38***
A3 44 0.01 0.53$
A4 34 ÿ0.1 0.34*
A5 8 0.56 0.89***
28
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
Table 3 Means for the respective conditions and assessment points on the physiological measures Variable
Condition
A1
A2
A3
A4
A5
t-value
F-valuea
Systolic blood pressure
Manual Video Group Individual F-valueb
136.0 126.8 132.1 131.3 0.6
125.8 124.8 119.2 125.9 1.3
Ð 127.4 118.9 123.8 1.6
Ð Ð 119.7 120.4 0.0
Ð Ð Ð 120.8
2.3* 0.1 4.5*** 1.9
G: 0.7 T: 24.4$ I: 0.9
Diastolic blood pressure
Manual Video Group Individual F-valueb
91.3 88.0 91.9 94.9 0.4
86.1 88.8 80.5 90.4 3.2*
Ð 89.4 85.2 87.4 0.6
Ð Ð 85.7 91.0 1.1
Ð Ð Ð 81.9
2.3* 0.3 1.9 2.7*
G: 0.0 T: 12.2*** I: 1.3
Heartrate
Manual Video Group Individual F-valueb
86.4 83.8 81.1 75.4 1.0
78.1 68.8 75.9 81.5 1.6
Ð 79.8 75.5 78.1 0.5
Ð Ð 78.7 78.7 0.0
Ð Ð Ð 89.3
1.3 0.5 0.7 2.2
G: 0.2 T: 0.1 I: 2.5
a
G = group, T = Time, I = Interaction eect in the two-way ANOVA. One-way ANOVA. * P < 0.05, **P < 0.01, ***P < 0.001, $P < 0.0001. b
ni®cantly. The one-way ANOVAs indicated no signi®cant dierences between the groups, either at pre-treatment, or at any later assessment. 3.4.6. Credibility and motivation The results regarding the credibility of the treatments, the patients' expectancy for improvement, and their motivation for therapy are presented in Table 6. Credibility and expectancy were assessed prior to the manual, video, and group treatment, respectively, while motivation was assessed ®rst before the screening-interview and secondly immediately before the group treatment. The one-way ANOVA on the ®rst assessment scores showed that the patients who were to become clinically improved after manual and video treatment rated the manual treatment as signi®cantly more credible than did the patients who were going to require therapist-directed treatment. There were no dierences at A2, but at A3 those patients who were to become clinically improved after group treatment rated this as more credible than those who required individual treatment. On the expectancy item and the motivation scale there were no dierences between the groups on any assessment. 3.5. Predictors of treatment eects after self-help treatments In order to identify possible predictors of treatment eects a standard multiple regression analysis was performed. Based on previous prediction studies in other anxiety disorders some pre-treatment variables were selected together with other variables of theoretical interest. This resulted in the following 14 variables: age at onset, duration of the phobia, family prevalence of spider phobia, self-assessed fear, avoidance, and handicap from spiders, credibility, expect-
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
29
Table 4 Means for the respective conditions and assessment points on the spider phobia questionnaires Condition
A1
A2
A3
A4
A5
SPQ
Manual Video Group Individual F-valueb
23.4 22.8 24.1 23.0 0.3
10.3 15.4 20.0 21.9 18.3$
Ð 10.6 19.5 20.1 6.2*
Ð Ð 10.3 12.8 2.1
Ð Ð Ð 11.7
9.3$ 3.7* 12.1$ 5.5$
G: 0.1 T: 259.8$ I: 0.5
SPQ-WS
Manual Video Group Individual F-valueb
25.6 26.0 27.3 24.8 1.1
14.3 18.6 24.6 24.8 14.7$
Ð 17.8 22.7 23.8 2.0
Ð Ð 15.1 16.8 0.9
Ð Ð Ð 15.3
10.8$ 3.2* 13.2$ 6.3$
G: 0.6 T: 319.5$ I: 1.6
SA-Fear
Manual Video Group Individual F-valueb
21.9 21.6 21.9 21.2 0.3
7.8 15.8 18.2 20.1 15.3$
Ð 13.2 16.6 18.8 2.1
Ð Ð 7.7 9.7 1.5
Ð Ð Ð 8.5
8.1$ 3.0* 14.5$ 9.2$
G: 1.1 T: 311.1$ I: 2.0
SA-Avoidance
Manual Video Group Individual F-valueb
19.5 20.4 19.1 17.5 1.4
7.1 13.4 15.7 18.1 13.8$
Ð 10.6 14.0 16.2 3.3*
Ð Ð 7.0 8.4 1.2
Ð Ð Ð 6.8
8.6$ 7.2** 12.7$ 12.1
G: 1.7 T: 362.1$ I: 0.6
SA-Handicap
Manual Video Group Individual F-valueb
5.1 6.4 5.1 4.3 1.5
Ð 2.0 3.5 3.0 1.4
Ð Ð 1.6 1.9 0.7
Ð Ð Ð 1.3
7.1$ 5.4** 8.7$ 5.1$
G: 1.8 T: 175.4$ I: 0.8
a b
1.3 3.8 3.7 3.5 7.0***
t-value
F-valuea
Variable
G = group, T = Time, I = Interaction eect in the two-way ANOVA. One-way ANOVA.
ancy, motivation, BAI, BDI, STAI-S, STAI-T, and FSS-III. The dependent variable in the analysis was the degree of clinically signi®cant improvement. Based on three criteria the patient could get 0±3 points on this measure. As a preparatory step, partial correlations between each predictor variable and the dependent variable were calculated, while controlling for all other predictors. This resulted in the inclusion of the following seven predictor variables in the multiple regression analysis: duration of the phobia, self-assessed avoidance, credibility, expectancy, motivation, BDI, and STAI-S. Table 7 displays the result of the regression analysis. There was a signi®cant R2 of 0.27, and credibility had the highest prediction value, i.e. a higher credibility for the manual treatment increased the probability of a positive treatment eect. 3.6. Predictors of attrition from self-exposure treatment In order to identify possible predictors of drop-out from the manual treatment a logistic regression analysis was performed. The same independent variables as in the multiple regression analysis above were used, and the patients were divided into two groups: (1) those who
30
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
Table 5 Means for the respective conditions and assessment points on the psychopathology questionnaires Variable
Condition
A1
A2
A3
A4
A5
t-value
F-valuea
BAI
Manual Video Group Individual F-valueb
8.8 9.0 8.2 10.2 0.2
5.8 6.4 8.0 8.9 0.8
Ð 4.0 6.6 7.1 0.3
Ð Ð 5.0 4.3 0.1
Ð Ð Ð 4.1
1.6 1.3 3.2** 3.0*
G: 0.1 T: 20.9d I: 0.7
BDI
Manual Video Group Individual F-valueb
6.0 5.0 5.3 5.7 0.1
3.3 4.0 3.7 5.5 0.7
Ð 1.6 3.8 3.6 0.2
Ð Ð 3.5 4.0 0.1
Ð Ð Ð 2.5
2.7* 3.6* 2.2 2.8*
G: 0.1 T: 22.2$ I: 0.4
STAI-Trait
Manual Video Group Individual F-valueb
39.8 44.0 36.1 40.5 1.2
36.3 37.4 33.6 36.2 0.5
Ð 34.2 33.3 35.8 0.4
Ð Ð 32.0 31.0 0.1
Ð Ð Ð 30.4
2.4* 5.1** 2.9** 4.2**
G: 1.1 T: 38.4$ I: 3.2*
STAI-State
Manual Video Group Individual F-valueb
40.4 39.0 36.2 38.5 0.6
32.1 37.6 34.0 36.6 0.9
Ð 29.8 31.8 36.6 1.2
Ð Ð 31.1 33.0 0.3
Ð Ð Ð 29.6
2.6* 2.7 2.9** 3.1*
G: 0.4 T: 28.3$ I: 0.6
FSS-III
Manual Video Group Individual F-valueb
144.3 168.4 150.2 157.4 1.1
130.4 142.6 140.1 144.7 0.5
Ð 126.8 133.0 135.8 0.2
Ð Ð 128.8 125.1 0.1
Ð Ð Ð 121.3
3.0** 6.6** 4.7*** 4.6***
G: 0.2 T: 67.9$ I: 3.4*
a
G = group, T = Time, I = Interaction eect in the two-way ANOVA. One-way ANOVA. * P < 0.05, **P < 0.01, ***P < 0.001, $P < 0.0001. b
dropped out (n = 38); and (2) those who completed (n = 59) the manual treatment. Overall, 65% of the patients were correctly classi®ed (32% of drop-outs and 86% of completers), but the w2(7) = 8.70, P < 0.28, was not signi®cant. 3.7. Predictors of the extent of treatment needed In order to identify possible predictors of the extent of treatment needed to obtain a clinically signi®cant improvement, a standard discriminant analysis was performed with the same independent variables as in the regression analyses. Excluding the dropouts, the remaining 59 patients were divided into two groups: (1) those who were improved after self-exposure treatment (n = 21); and (2) those who needed therapist-directed treatment of some form (n = 38). The result of this analysis is presented in Table 8. The canonical R (0.59) was signi®cant
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
31
Table 6 Means for the respective conditions and assessment points on the credibility and motivation scales Variable
Condition
A1
A2
A3
Credibility
Manual Video Group Individual F-valuea
33.4 35.6 28.1 27.1 3.3*
Ð 32.8 29.0 29.3 0.5
Ð Ð 35.6 30.5 14.7***
Expectancy
Manual Video Group Individual F-valuea
7.9 8.0 7.8 6.7 1.5
Ð 8.0 7.1 7.2 0.6
Ð Ð 7.8 7.6 0.3
Motivation
Manual Video Group Individual F-valuea
26.5 26.8 27.2 26.6 0.9
Ð Ð Ð Ð Ð
Ð Ð 22.8 21.5 1.8
a
One-way ANOVA. * P < 0.05, **P < 0.01, ***P < 0.001.
(w2(7) = 22.48, P < 0.005), and credibility and motivation were the two strongest predictors. A total of 71% of the patients were correctly classi®ed (67% of the self-exposure successes, and 74% of the therapist treated patients).
4. Discussion The aim of this study was to investigate if there are variables that can predict treatment success and attrition from self-help treatment, as well as the extent of treatment needed to achieve clinically signi®cant improvement. A prerequisite in order to study predictors at all is that there are clinical eects of the treatment. Table 1, Table 3, Table 4 and Table 5 show that of the 18 outcome variables, all but one (heart rate) yielded highly signi®cant time eects (P < 0.001 or better) in the ANOVAs. Furthermore, it is important to compare the proportions of clinically signi®cant improvement in the present study and in our previous four studies. The self-help manual yielded 27% clinically improved, which is comparable to the 31% in the OÈst et al. (1991) and HellstroÈm and OÈst (1995) studies. The video treatment yielded 10%, which is lower than the 31% of the OÈst et al. (1997a) study. The group treatment gave 68%, which is the same as the 69% of the OÈst (1996) and OÈst et al. (1997a) studies. Finally, the individual treatment resulted in 100% clinically improved, compared to 89% in the OÈst et al. (1991) and HellstroÈm and OÈst (1995) studies. Thus, except for the video treatment, the eects of the present study were comparable to those of the authors' previous studies. However, it should be pointed out that there is a dierence between the group format of video treatment in the OÈst et al. (1997a) study, where eight
32
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
Table 7 Multiple regression analysis regarding predictors of clinically signi®cant improvement after self-exposure treatment (manual and video) Predictor variable
B
b
R
R2
Motivation STAI-S BDI Phobia duration Credibility Expectancy Phobic avoidance
0.088 0.020 0.027 0.273 ÿ0.454* 0.608 ÿ0.331
0.22 0.17 0.03 0.21 ÿ0.43 0.22 ÿ0.16
0.51
0.27
Adj. R2
F(7,57)
0.18
3.01*
* P < 0.01. Table 8 Discriminant analysis regarding clinically signi®cant improvement after self-exposure and therapist-directed exposure, respectively Predictor variable Credibility Motivation Phobic avoidance Phobic duration STAI-S BDI Expectancy Canonical R Eigenvalue
Correlation between predictor variable and discriminant function ÿ0.56 0.50 ÿ0.24 0.21 0.20 0.20 ÿ0.14 0.59 0.52
Univariate F (1,57) 9.39* 7.46* 1.76 1.35 1.21 1.18 0.54
* P < 0.01.
patients sit in a group and watch the video together, and the individual watching of the tape in each patient's home in the present study. Turning to the prediction analyses, the following picture emerged. When it comes to the issue of predicting treatment success after the self-help treatments, the multiple regression analysis showed that the tested variables explained 27% of the variance in the degree of clinical improvement. The credibility of manual treatment had the highest prediction value of the independent variables. This was also the case in one of the samples in the cross validation study of HellstroÈm and OÈst (1996), which included patients with spider phobia, blood phobia, and injection phobia. In that study expectancy was also a signi®cant predictor. Thus, it might be that the extent to which speci®c phobics experience the treatment as credible and expect to become improved are important predictors for treatment outcome. However, this has to be replicated in a new study using a cross validation design. When the authors attempted to predict attrition from the self-help treatments using a logistic regression analysis and the same independent variables the result was not signi®cant. While
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
33
86% of the completers were correctly classi®ed only 32% of the drop-outs were so classi®ed, for a total of 65%. The results from the telephone interviews of the drop-outs also indicated that there are other variables that may be of greater importance. The primary reasons for not completing the manual treatment were that they needed the support of a therapist during the treatment, and practical diculties like not being able to ®nd spiders to work with. However, from a clinical point of view it is reassuring that none of the patients had become worse as a consequence of their incomplete participation in the manual treatment. On the contrary, 76.5% reported various degrees of positive changes, and 12% of the drop-outs considered themselves improved to such an extent that they no longer needed any treatment. Finally, predictors of the extent of treatment needed to become clinically improved was tested with a discriminant analysis. Using the same independent variables led to a signi®cant result, and credibility and motivation for psychotherapy were the strongest predictors. Overall 71% of the patients were correctly classi®ed. Thus, the ®ndings of Keijsers (1994) who found that motivation was a strong predictor was replicated in this study. Future prediction studies on speci®c phobics might consider including variables that were important in the drop-out analysis, namely the need for active therapist support during treatment and the access to social support during the self-help treatment. Furthermore, a recent study by Thorpe and Salkovskis (1995) showed that spider phobics have catastrophic cognitions concerning the phobic situation, e.g. being physically harmed, going crazy, or losing control. The degree to which the patients hold these beliefs is also a predictor worth assessing in future studies. From the clinician's point of view this study has shown that two variables that we as therapists believe are important for the outcome, credibility and motivation, also turned out signi®cant in the multivariate analyses. Thus, the therapist is recommended to assess the patient's motivation for psychotherapy in general, and the credibility of the suggested treatment in particular before initiating treatment. If it turns out that these variables are low, the therapist is advised to spend some time trying to analyze why this is so, and to increase them using e.g. information and cognitive restructuring methods. From a methodological point of view the design used has its weaknesses. First, the many assessments necessary for the patients needing most treatment naturally means that this eect comes into play as a threat to internal validity. However, there does not seem to be any way around this problem. Second, the design implies that the patients did not unmask that there were other treatments besides the one they were randomized to, i.e. the manual treatment, available. If the patients assumed that in case they were not clinically improved after the selfexposure treatments they would be oered therapist-directed treatment, this would most probably lead them not to do their very best with the previous treatments. However, the information gathered at assessments 2, 3, and 4, has in no way indicated that the patients believed there to be another treatment `waiting'. On the contrary, the patients expressed surprise and delight at being oered further treatment. Thus, we believe that it is possible to conclude that the design worked as it was supposed to do. Regarding the assessment instruments, well developed and psychometrically sound instruments were used, with one exception, the Nijmegen Motivation List. This newly developed instrument has not yet gone through psychometric testing, which means that we do not know its reliability and validity. The study by Keijsers (1994) found that only six of the 12 items were used in the analysis, due to low item-total correlations in the others. Despite this fact, the
34
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
motivation scale turned out signi®cant in the discriminant analysis which means that it has to have some degree of reliability. The issue of motivation for treatment is very important and further research is warranted in this area to develop new and psychometrically sound instruments of this concept. Three of the four treatments used in the present study have been used in the authors' previous studies of spider phobia. The fourth, the video treatment, was new in the format applied. This also yielded the smallest proportion of clinically improved patients. The major advantage of this treatment is that it is inexpensive and the patients reported increased con®dence in exposure treatment in general after having watched the video. However, a combination of the manual and a professionally produced video could be a more eective alternative as self-help treatment. Such a video could on the one hand be directly tailored to the dierent steps of the manual, and on the other contain pedagogical instructions to the patient watching it. In future research on self-help treatment such a combined condition could be compared with a manual only condition, and a group treatment condition. In conclusion, the present study adds to the eight previous one-session studies from the authors' clinic (HellstroÈm and OÈst, 1995; HellstroÈm et al., 1996; OÈst, 1996; OÈst et al., 1991, 1992, 1997a,b,c) indicating that this is the treatment of choice for speci®c phobias. Moreover, it is the ®rst study in which a stepwise design of this type has been used, and it yielded some useful predictors for treatment outcome of the self-help form of exposure therapy. Acknowledgements This research was partly supported by Grant 05452 from the Swedish Medical Research Council.
References Al-Kubaisy, T., Marks, I. M., Logsdail, S., Marks, M. P., Lovell, K., Sungur, M., & Araya, R. (1992). Role of exposure homework in phobia reduction: A controlled study. Behavior Therapy, 23, 599±621. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders, 4th ed. Washington, DC: Author. Arntz, A., & Lavy, E. (1993). Does stimulus elaboration potentiate exposure in-vivo treatment? Two forms of one-session treatment of spider phobia. Behavioural Psychotherapy, 21, 1±12. Bandura, A., Blanchard, E. B., & Ritter, B. (1969). Relative ecacy of desensitisation and modelling approaches for inducing behavioral, aective, and attitudinal changes. Journal of Personality and Social Psychology, 13, 173±199. Beck, A. T. (1967). Depression: Causes and treatment. Philadelphia: University of Philadelphia Press. Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893±897. Borkovec, T. D., & Nau, S. D. (1972). Credibility of analogue therapy rationales. Journal of Behavior Therapy and Experimental Psychiatry, 3, 257±260. Bourdon, K. H., Boyd, J. H., Rae, D. S., Burns, B. J., Thompson, J. W., & Locke, B. Z. (1988). Gender dierences in phobiasÐ results of the ECA community survey. Journal of Anxiety Disorders, 2, 227±241. Brown, T. A., DiNardo, P., & Barlow, D. H. (1994). Anxiety Disorders Interview ScheduleÐIV (ADIS-IV). Albany, NY: Center for Stress and Anxiety Disorders. Chambless, D. L. (1990). Spacing of exposure sessions in treatment of agoraphobia and simple phobia. Behavior Therapy, 21, 217±229. Chapman, T. F. (1997). The epidemiology of fears and phobias. In Davey, G. C. L. (Ed.), Phobias. A handbook of theory, research and treatment (pp. 415±434). London: Wiley.
L. OÈst et al. / Behaviour Research and Therapy 36 (1998) 17±35
35
Ghosh, A., & Marks, I. M. (1987). Self-treatment of agoraphobia by exposure. Behavior Therapy, 18, 3±16. GoÈtestam, K. G. (1994, September). Three forms of one-session group treatments for spider phobia. Paper presented at the EABCT-congress in Corfu. HellstroÈm, K., Fellenius, J., & OÈst, L. -G. (1996). One versus ®ve sessions of applied tension in the treatment of blood phobia. Behaviour Research and Therapy, 34, 101±112. HellstroÈm, K., & OÈst, L. -G. (1995). One-session therapist directed exposure vs. two forms of manual directed self-exposure in the treatment of spider phobia. Behaviour Research and Therapy, 33, 959±965. HellstroÈm, K., & OÈst, L.-G. (1996). Prediction of outcome in the treatment of speci®c phobia. A cross-validation study. Behaviour Research and Therapy, 34, 403±411. Jacobson, N. S., Follette, W. C., & Revenstorf, D. (1984). Psychotherapy outcome research: Methods for reporting variability and evaluating signi®cance. Behavior Therapy, 15, 336±352. Keijsers, G. (1994). Prognostic factors in the treatment of anxiety and failure in behaviour therapy. Doctoral dissertation, University of Nijmegen, Holland. Klorman, R., Weerts, T. C., Hastings, J. C., Melamed, B. G., & Lang, P. (1974). Psychometric description of some speci®c-fear questionnaires. Behavior Therapy, 5, 401±409. Marks, I. M. (1987). Fears, phobias and rituals. Oxford: Oxford University Press. Mathews, A. M., Gelder. G., & Johnston, D. W. (1981). Agoraphobia. Nature and treatment. New York: Guilford. Nau, S. D., Caputo, J. A., & Borkovec, T. D. (1974). The relationship between credibility of therapy and simulated therapy eects. Journal of Behavior Therapy and Experimental Psychiatry, 5, 129±133. OÈst, L. -G. (1989a). One-session treatment for speci®c phobias. Behaviour Research and Therapy, 27, 1±7. OÈst, L. -G. (1989b). A maintenance program for behavioral treatment of anxiety disorders. Behaviour Research and Therapy, 27, 123± 130. OÈst, L.-G. (1996). One session group treatment of spider phobia. Behaviour Research and Therapy, 34, 707±715. OÈst, L.-G. (1997). Rapid treatment of speci®c phobias. In Davey, G. C. L. (Ed.), Phobias. A handbook of theory, research and treatment (pp. 227±246). London: Wiley. OÈst, L.-G., Alm. T., Brandberg. M., & Breitholtz, E. (1998). One vs. ®ve sessions of exposure and cognitive therapy in the treatment of claustrophobia. Manuscript submitted for publication. OÈst, L.-G., Brandberg. M., & Alm. T. (1997). One versus ®ve sessions of exposure in the treatment of ¯ying phobia. 35, 987±996. OÈst, L.-G., Ferebee, I., & Furmark, T. (1997). One-session group therapy of spider phobia: Direct vs. indirect treatments. Behaviour Research and Therapy, 35, 721±732. OÈst, L. -G., HellstroÈm, K., & KaÊver, A. (1992). One versus ®ve sessions of exposure in the treatment of injection phobia. Behavior Therapy, 23, 263±282. OÈst, L. -G., Salkovskis, P. M., & HellstroÈm, K. (1991). One-session therapist directed exposure vs. self-exposure in the treatment of spider phobia. Behavior Therapy, 22, 407±422. Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the STAI. Palo Alto: Consulting Psychologists Press. Thorpe, S. J., & Saslkovskis, P. M. (1995). Phobic beliefs: Do cognitive factors play a role in speci®c phobia?. Behaviour Research and Therapy, 33, 805±816. Watts, F. N., & Sharrock, R. (1984). Questionnaire dimensions of spider phobia. Behaviour Research and Therapy, 22, 575±580. Wolpe, J., & Lang, P. (1964). A fear schedule for use in behavior therapy. Behaviour Research and Therapy, 2, 27±30.