- Email: [email protected]
Virtual Reality Exposure Therapy for PTSD Symptoms After a Road Accident: An Uncontrolled Case Series
Behavior Therapy 38 (2007) 39 – 48
www.elsevier.com/locate/bt
Virtual Reality Exposure Therapy for PTSD Symptoms After a Road Accident: An Uncontrolled Case Series J. Gayle Beck, Sarah A. Palyo, SUNY Buffalo Eliot H. Winer, Brad E. Schwagler, Eu Jin Ang, NYSCEDII and SUNY Buffalo
This report examined whether Virtual Reality Exposure Therapy (VRET) could be used in the treatment of posttraumatic stress disorder (PTSD) symptoms in the aftermath of a serious motor vehicle accident. Six individuals reporting either full or severe subsyndromal PTSD completed 10 sessions of VRET, which was conducted using software designed to create real-time driving scenarios. Results indicated significant reductions in posttrauma symptoms involving reexperiencing, avoidance, and emotional numbing, with effect sizes ranging from d = .79 to d = 1.49. Indices of clinically significant and reliable change suggested that the magnitude of these changes was meaningful. Additionally, high levels of perceived reality (“presence”) within the virtual driving situation were reported, and patients reported satisfaction with treatment. Results are discussed in light of the possibility for VRET to be useful in guiding exposure in the treatment of PTSD following road accidents.
C O N S I D E R A B L E AT T E N T I O N H A S F O C U S E D recently on the treatment of posttraumatic stress disorder (PTSD). At present, strong empirical evidence exists to support the efficacy of cognitive Eliot Winer is now at Iowa State University. Brad Schwagler is now at the Georgia Institute of Technology. This project was supported in part by a grant from the University at Buffalo (IRCAF 24623). Additional support from NYSCEDII and the Department of Psychology is acknowledged as well. We are especially appreciative of help received from Dr. Ken English in completion of the project. We appreciate the assistance offered by Sumeet Parashar, Gautam Agrawal, Sherry Farrow, Luana Miller, and Berglind Gudmundsdottir in the completion of this project. Dr. Barbara Rothbaum provided helpful input on an earlier draft of this manuscript. Address correspondence to J. Gayle Beck, Department of Psychology, Park Hall, University at Buffalo, Buffalo, New York, 14260, USA; e-mail: [email protected]. 0005-7894/06/039–048/$1.00/0 © 2006 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.
behavioral treatments for PTSD, particularly those treatments that emphasize exposure to traumarelated stimuli and cues (Rothbaum, Meadows, Resick, & Foy, 2000). Exposure therapy (ET) has been used with a wide range of trauma populations (e.g., combat veterans, female survivors of sexual and physical assault), and the controlled studies that support its use meet the “gold standard” for clinical research (Foa & Meadows, 1997). Successful ET involves providing the patient with PTSD with prolonged contact with trauma-related cues in order to facilitate habituation and the development of neutral memory structures that “override” anxiety-provoking memories related to the trauma (Foa & Kozak, 1986). Although ET clearly is an efficacious approach in the treatment of PTSD, there are circumstances in which in vivo forms of ET become difficult to implement. The treatment of PTSD following a motor vehicle accident (MVA) represents one such circumstance, given the nature of the trauma-related cues and the fact that many individuals with MVA-related PTSD avoid driving completely (Blanchard & Hickling, 2004). In this article, preliminary data are presented on the use of virtual reality–assisted ET (VRET) in the treatment of acute-phase PTSD following a serious MVA in a beginning effort to examine the efficacy and acceptability of this treatment modality. The use of VRET is relatively recent, although studies to date have shown positive results of this treatment for a panoply of anxiety disorders, including claustrophobia (e.g., Botella et al., 1998), acrophobia (e.g., Emmelkamp et al., 2002), and fear of flying (e.g., Maltby, Kirsch, Mayers, & Allen, 2002). In these studies, virtual reality techniques are used to create environments that are immersive and that provoke anxiety. The patient is encouraged to stay in the virtual environment for a prolonged time interval, until fear reduction occurs. Thus, VRET
40
beck et al.
represents a step between imaginal exposure and in vivo exposure. As discussed by Krijn, Emmelkamp, Olafsson, and Biemond (2004), patients need to feel present in the virtual environment in order to experience it in the first person. Within this literature, “presence” is conceptualized as the subjective perception that the virtual world is somewhat real. Stated differently, “at some level and to some degree, . . . objects, events, entities, and environments are perceived as if the technology was not involved in the experience” (International Society for Presence Research, n.d., 2005). Additionally, the virtual environment needs to be able to elicit affect. Lastly, the virtual world needs to resemble the natural environment in order to ensure generalization of fear reduction. Thus, simulation sickness (which is akin to motion sickness in the virtual world) must be kept to a minimum. These qualities have been highlighted as necessary in order for VRET to be effective, following from the guidelines presented by Foa and Kozak (1986). Preliminary data collected on Vietnam veterans suggest that VRET can be used effectively for treating PTSD. Rothbaum and colleagues have presented a clinical trial (Rothbaum, Hodges, Ready, Graap, & Alarcon, 2001) in which VRET was used as one component of a multicomponent intervention for combat-related PTSD. In this report, VRET was included along with breathing relaxation (for general stress management) and imaginal exposure. In the open clinical trial, 9 veterans completed 8 to 16 treatment sessions. Statistically significant and clinically meaningful reductions in PTSD symptomatology were noted, suggesting that VRET (when combined with other interventions) holds promise in the treatment of chronic combat-related PTSD. Given the substantial number of individuals who are involved in an MVA each year and the estimate that 800,000 new cases of PTSD result from these crashes (Blanchard & Hickling, 2004), the need for a safe and cost-efficient treatment is clear. VRET represents a viable option for the treatment of MVArelated PTSD for a number of reasons. First, in contrast to in vivo exposure, VRET permits closer control over stimulus presentation. For example, if the patient wanted to work on exposure to driving near trucks in a highway environment, this would be easily accomplished within a VRET situation. In real life, setting up an exposure scenario to address this fear might prove challenging because it is difficult to predict when trucks would appear on a given highway. Second, VRET is especially well suited to exposure situations that involve time-limited stimuli (e.g., merging onto a freeway), as it is possible to
conduct repeated exposure easily. In real-life exposure, this is considerably more difficult. Thus, VRET offers greater convenience for therapist and patient alike. Third, VRET may reduce safety risks that may be associated with in vivo exposure, particularly given that individuals with MVA-related PTSD may report overly defensive driving styles that heighten risk for another MVA (e.g., J.G. Beck & Coffey, 2005). Fourth, VRET may offer a more “tangible” exposure experience, relative to imaginal exposure, possibly facilitating fear reduction more effectively. Lastly, some individuals with PTSD report a complete inability to drive, based on intense fear and anxiety. VRET may be a less threatening (and thus, more acceptable) form of treatment compared to in-vivo exposure for individuals who are extremely avoidant. Although there have been no published studies of VRET for MVA-related PTSD, there has been one effort to treat individuals with a specific phobia of driving using this approach (Wald, 2004). In this report, a commercially available driving simulator was used (driVR, Imago Systems, 1996). Six standardized scenes were used, in which the individual proceeded down a road in a predetermined path. Each scene was 3 to 5 minutes in duration, which is quite brief in light of the usual requirements for ET. A variety of options were available with respect to weather and time of day, although the specific driving route was repeated on subsequent exposure trials. Five women were treated in a multiple-baseline design, each receiving 8 weekly sessions. Three individuals showed reductions in phobia symptoms based on a clinician-administered interview, although the other two did not. VRET did not result in an increase in actual driving time or driving frequency for any patient in this report. Thus, even with short exposure stimuli and limited flexibility, VRET was somewhat effective for 3 of 5 (60%) of the cases of driving phobia. Clearly, this technology may hold promise for the treatment of driving-related fear. Although driving phobia and MVA-related PTSD share the common characteristic of avoidance of driving cues, many writers believe that the psychopathology and treatment of these two disorders are somewhat different (Kuch, Swinson, & Kirby, 1985; Taylor & Koch, 1995). In particular, the range of feared cues and degree of anxiety appears greater among patients with MVA-related PTSD. Thus, the goal of this report was to examine whether VRET could be used, in conjunction with relaxation training, to facilitate the treatment of MVA-related PTSD. Individuals in the acute phase of the disorder were selected for treatment and provided with 10 sessions of treatment. This report utilized an
virtual reality exposure therapy for ptsd uncontrolled case design, in light of the absence of previous studies on this issue. We hypothesized that VRET would reduce PTSD symptoms (intrusive reminders of the trauma, avoidance of traumarelated cues, emotional numbing, and physiological hyperarousal). Additional measures of anxiety and depression were included to examine the generalizability of treatment effects.
Method participants Participants were recruited for this trial as part of a larger study on the treatment of MVA-related PTSD. Recruitment efforts included newspaper ads as well as referrals from physicians, psychiatrists, counselors, chiropractors, physical therapists, and massage therapists. To qualify for participation, individuals must have experienced an MVA that involved actual or threatened death or serious injury and evoked an emotional response that included intense fear, helplessness, horror, or the perception that they would die (American Psychiatric Association, 2000), as assessed using the Motor Vehicle Accident Interview (Blanchard & Hickling, 2004). Only individuals who had experienced an MVA in the past 6 months were included. Exclusion criteria included cognitive impairment (indexed by a score below 24 on the Mini Mental Status Exam; Folstein, Folstein, & McHugh, 1975), current substance abuse or dependence, or suicidality. Eleven individuals were offered entry into the study. Three patients declined participation (two did not desire VRET and the third reported acute distress which warranted a more broad-band treatment to address severe depression as well as PTSD).1 Participants met criteria for full (n = 6) or subsyndromal (n = 2) PTSD. Full PTSD was defined as meeting DSM-IV criteria for the disorder. Subsyndromal PTSD was defined as meeting diagnostic criteria for the reexperiencing cluster and for either the avoidance and numbing or hyperarousal cluster (but not both), in accordance with Blanchard, Hickling, Taylor, Loos, and Gerardi (1994).2 Participants were required to suspend other psychological treatments during their participation. Individuals taking psychotropic
1 All three individuals who declined VRET met full criteria for PTSD. 2 Our decision to include patients with subsyndromal PTSD was based on reasons similar to those discussed by Blanchard and Hickling (2004), specifically that these individuals report clinically significant distress and impairment resulting from PTSD symptoms.
41
medication were asked to keep the dosage constant, in consultation with their prescribing physician. Of the 8 participants (7 females and 1 male), 2 participants were African American, and 6 participants were Caucasian. The mean age for the sample was 49.50 (SD = 7.03). Participants were initially assessed an average of 3.50 months (SD = 2.07) after their MVA. Two patients dropped out of VRET. One participant dropped after Session 3 due to an increase in family stress. The other participant, who had been on disability prior to VRET, dropped after Session 4 because she had to return to work and no longer had time to attend treatment sessions.3 The sample of treatment completers included 5 females and 1 male. Four patients met criteria for full PTSD, while 2 met criteria for subsyndromal PTSD. The final sample had an average age of 49.17 (SD = 8.28). Two participants were African American, and the remainder was Caucasian. Four participants were currently on disability, while the other 2 participants were employed. Half of the participants reported that their annual income was less than $50,000. These participants were initially assessed an average of 3.5 months (SD = 1.87) after their MVA (3 patients <4 months post-MVA).
outcome measures PTSD measures. Measures of PTSD included a clinician-administered interview and two selfreport measures. The Clinician-Administered PTSD Scale (CAPS; Blake et al., 1990) is a structured interview designed to assess the frequency and intensity of the 17 symptoms of PTSD during the last month rated on a 5-point Likert-type scale, where 0 indicates that the symptom has not occurred or is not distressing and 4 indicates that the symptom occurs nearly every day or is severely distressing. The CAPS was scored by summing the frequency and intensity ratings for both the entire measure (total CAPS score) and each symptom cluster. The CAPS has excellent support for its reliability, with alpha coefficients generally ranging from .73 to .98 and 2- to 3-day test-retest reliability ranging from .78 to .87 (Weathers, Keane, & Davidson, 2001). Alpha coefficients for the CAPS derived from a similar sample (n = 112; J.G. Beck et al., 2004) were .80 (reexperiencing symptom cluster), .71 (avoidance-numbing symptom cluster), .69 (physiological symptom cluster), and .87 (total score). In order for a symptom to count toward a diagnosis of full or subsyndromal PTSD, a participant
3
Nonsignificant differences were noted between treatment completers and noncompleters on pretreatment scores on each of the outcome variables.
42
beck et al.
must score a 1 or higher on frequency and a 2 or higher on intensity (Blanchard et al., 1996). The CAPS was administered by extensively trained doctoral students in clinical psychology. Interviews from a larger sample (N = 409) that included the 6 participants in this study were videotaped, and 29% (n = 119) were randomly selected and reviewed by an independent clinician to establish diagnostic reliability. Interrater agreement in PTSD diagnosis, reflected by the kappa statistic, was strong for PTSD (k = 0.88). Additionally, the Posttraumatic Stress Scale–Self Report (PSS-SR; Foa, Riggs, Dancu, & Rothbaum, 1993) and the Impact of Event Scale–Revised (IES-R; Weiss & Marmar, 1997) were administered. The PSS-SR is a 17-item self-report scale that measures the frequency of DSM-IV symptoms of PTSD in the past month. Each item is rated on a 0 to 3 scale, and ratings are summed to form a total score. The PSSSR has been shown to have high internal consistency (α = .91) and good 1-month test-retest reliability (r = .74) in a sample of 46 female rape victims and 72 female non-sexual-assault victims (Foa et al., 1993). In a similar sample as the present report, the PSS-SR was shown to have good internal consistency (α = .93, N = 112; J.G. Beck et al., 2004). The IES-R is a 22-item self-report measure designed to assess for the experiences of intrusion, avoidance, and hyperarousal following a distressing event. Items are rated on a 0 to 4 scale and summed to form a total score, as well as subscales reflecting intrusion, avoidance, and hyperarousal symptoms. Psychometric data reported by Weiss and Marmar (1997) suggest that the IES-R has high internal consistency (alphas ranging from .79 to .92 on each of the subscales) and good test-retest reliability (correlations ranging from .51 to .94 on each of the subscales). Data from a larger sample derived from the same site as the current report indicate alphas ranging from .85 to .90 on each of the subscales (N = 199; J.G. Beck et al., 2006). Measures of anxiety and depression. Self-reported anxiety was measured using the Beck Anxiety Inventory (BAI; A.T. Beck, Epstein, Brown, & Steer, 1988). The BAI is a 21-item self-report inventory of common symptoms of anxiety (e.g., difficulty breathing, feelings of choking). Participants rate the extent to which they were bothered by each item during the past week on a 0-3 scale ranging from “not at all” to “severely.” The inventory has been shown to have high internal consistency (a = .92) and test-retest reliability (r = .75) in a sample of psychiatric outpatients (A. T. Beck et al., 1988). In a similar sample as the present report, the BAI also was internally consistent (α = .93, N = 199; J.G. Beck et al., 2006).
The Beck Depression Inventory-II (BDI-II; A.T. Beck, Steer, & Brown, 1996) was used to measure depression. The BDI-II is a 21-item measure used to evaluate depressive symptoms during the past two weeks. In a study of 137 students receiving treatment at a university counseling center, the BDI-II yielded reliability coefficients of .91 and .93 (Sprinkle et al., 2002). A one-week test-retest correlation of .93 was reported for a sample of 26 outpatients (A.T. Beck et al., 1996). In a similar sample as the present report, the BDI-II also was internally consistent (α = .93, N = 199; J.G. Beck et al., 2006). Treatment-related measures. Two self-report measures were included to assess simulation sickness and the sense of presence that was created by the virtual environment. Additionally, participants completed the Client Satisfaction Questionnaire (CSQ; Larsen, Attkisson, Hargreaves, & Nguyen, 1979) at the end of treatment. The Simulator Sickness Questionnaire (SSQ; Kennedy, Lane, Berhaum, & Lilienthal, 1993) lists 26 symptoms that are associated with physical discomfort experienced in VR simulators. Each symptom is rated as either none, slight, moderate, or severe. Symptoms are based on three subscales that describe specific dimensions of simulator sickness (i.e., nausea, oculomotor disturbances, and disorientation). The SSQ has been used extensively in studies of simulator sickness and has been shown to be a valid measure of this construct (Kennedy et al., 1993). The established norms for the SSQ were derived from 1,119 exposures to simulators at 10 different simulator sites (Kennedy et al., 1993). In the current study, the SSQ was administered to participants immediately after they had been in the simulator and averaged across exposure sessions to compare with established norms. Given the small sample size, coefficient alpha could not reliably be computed for the current report. The Presence Questionnaire (PQ; Witmer & Singer, 1998) is designed to measure the degree to which individuals experience presence in a virtual environment. This 19-item measure asks the participant to indicate on a 7-point Likert-type scale how immersed they felt in the virtual environment, with higher scores indicating a greater sense of presence in the virtual environment. The PQ has three subscales: Realism, Interface Quality, and Involvement/Control. The Realism subscale refers to the extent to which the interactions between the participant and the virtual environment felt consistent with reality. The Interface Quality subscale measures whether control devices distract from task performance. The Involvement/Control subscale addresses perceived control of events in the virtual
virtual reality exposure therapy for ptsd environment, responsiveness of the virtual environment to user-initiated actions, and how involved in the experience the participant became. Established norms were determined using a sample of 152 college students, and within this sample, the PQ was shown to have good internal consistency (a = .88; Witmer & Singer, 1998). In the current study, the PQ was administered to participants immediately after exiting the simulator and averaged across exposure sessions to compare with established norms. Given the small sample size, coefficient alpha could not reliably be computed for the current report. The CSQ is a measure of participant's satisfaction with treatment. This 8-item measure has been shown to be reliable (a = .95) in a sample of 49 outpatients (Larsen et al., 1979). Participants rate their satisfaction with treatment on a 4-point Likert-type scale. Scores on the CSQ can range from 8 to 32, with higher scores indicating greater satisfaction. The CSQ was administered at the end of the last VRET session. Given the small sample size, coefficient alpha could not reliably be computed for the current report.
vret VRET was administered in 10 individual sessions. The first 2 sessions included psychoeducation about the symptoms of PTSD and the purpose of VRET, relaxation training (Bernstein & Borkovec, 1973), construction of a hierarchy of anxiety-producing driving situations, and familiarization with the VR simulator using neutral (non-MVA-related) scenarios and were 90 minutes long. In the remaining 8 sessions, participants were instructed on how to do exposure and engaged in virtual reality exposure. In this protocol, relaxation training was included to address defensive behaviors that have been observed in this population (J.G. Beck & Coffey, 2005), including closing ones eyes while experiencing anxiety in the car or rigidly clutching the steering wheel in a fixed path, owing to extreme anxiety. This intervention was designed to help patients to focus on the exposure situation by giving them a coping strategy. Therapists were instructed not to encourage the use of relaxation as a distraction from exposure. Depending on the specific exposure situation, the participant would either be the driver or the passenger, with the therapist taking the other role. The therapist guided the participant through the exposure process, encouraging them to use relaxation skills, and to provide a rating of anxiety using the subjective units of distress (SUDS) scale (range: 0–100). In keeping with other exposure-based treatments (e.g., Foa & Rothbaum, 1998), participants were encouraged to
43
remain in the specific VR driving situation until their anxiety had reduced by 50%. Patients initially selected items on their hierarchies that caused mild anxiety and, as habituation occurred, gradually moved on to more anxiety-provoking situations. Sessions 3 through 10 were 60 minutes in length, and participants were assigned imaginal and in vivo exposure homework between sessions.4 The treatment was administered by two doctoral students, with supervision by a licensed psychologist. All sessions were audiotaped and 52% (n = 31) were selected for treatment integrity monitoring. A checklist of techniques deemed both appropriate (e.g., exposure) and inappropriate (e.g., cognitive therapy) for VRET was used to monitor treatment integrity. Examination of these data revealed no violations in the treatment protocol.
apparatus VRET was conducted using an SGI Origin 3400 and a Moog six degree-of-freedom motion base for the computational and graphics system. The motion base was outfitted with a steering wheel, gas pedal, and a brake pedal. Through PVM software (Parallel Virtual Machine), inputs from the motion base were sent to the graphics supercomputer to be integrated with the immersive visual representations being generated. In this manner, correct viewpoints and perspectives were computed and displayed for a user. The visualization system projected the simulated scene on a 10 foot by 8 foot screen that was located 5 feet away from the motion base. During VRET, both the patient and the therapist wore stereoscopic glasses to view the simulation in true 3D. virtual environments The visual simulation consisted of several adaptable and customizable driving scenarios. Four general types of scenes were available: highway, urban, suburban, and rural. Within each scene type, the therapist was able to predetermine the amount and type of traffic, time of day, weather conditions (i.e., snow, rain), and specific driving events (e.g., a car rapidly comes up from behind the driver, tailgating). This gave direct control over the level of anxiety produced for the patient as well as the patient’s overall comfort level with the VRET. None of the driving events were designed to re-create an MVA but were intended to simulate a real-world driving experience. During each VRET experience, the driver was able to determine his or her own 4
In the homework, imaginal exposure to tasks on the patient’s graded hierarchy was assigned, not exposure to memories of the traumatic MVA.
44
beck et al.
Table 1 Examination of pretreatment and posttreatment outcome measures (n = 6) Pre M (SD)
Post M (SD)
t (df = 5)
57.83 (17.02) 21.00 (7.77) 20.67 (2.26) 16.17 (6.49) 29.00 (11.98) 2.13 (0.91) 2.13 (1.11) 2.02 (0.94) 2.28 (1.08)
34.50 (20.81) 12.50 (8.31) 8.5 (8.04) 13.50 (9.89) 14.67 (8.17) .98 (.39) .73 (.29) 1.10 (.68) 1.17 (.70)
3.76* 3.26* 3.55* 1.02 3.56* 3.30* 3.59* 2.97* 2.07
6 6 5 6
Anxiety/Depression Measures BAI 23.00 (12.28) BDI-II 19.00 (9.51)
14.00 (10.02) 12.17 (7.68)
1.84 1.63
3 (50%) 3 (50%)
PTSD Measures CAPS Total CAPS Reexperiencing CAPS Avoidance/Numbing CAPS Hyperarousal PSS-SR Total IES-R Total IES-R Intrusion IES-R Avoidance IES-R Hyperarousal
CSC n (%)
(100%) (100%) (83%) (100%)
RC n (%)
Effect size (d)
Effect size for Intent-to-treat (n = 8)
4 (67%) 5 (83%) 3 (50%) 3 (50%)
0.91 0.79 1.49 0.24 1.04 1.21 1.25 0.83 0.90
0.74 0.67 0.93 0.21 0.85 0.97 0.97 0.71 0.73
1 (17%) 2 (33%)
0.60 0.60
0.50 0.49
Note. CAPS = Clinician Administered PTSD Scale; PSS-SR = Posttraumatic Stress Scale–Self-Report; IES-R = Impact of Event Scale, Revised; STAI-S = State-Trait Anxiety Inventory, State subscale; STAI-T = State-Trait Anxiety Inventory, Trait subscale; BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory–II; CSC = Clinically Significant Change; RC = Reliable Change. *p < .05., **p < .001.
course, including choosing which direction to turn, whether to enter a side street, etc.
procedure All participants received a pretreatment assessment, VRET, and a posttreatment assessment 1 month following the last VRET session. Posttreatment CAPS were conducted by interviewers who were unaware of the treatment status of the patient. Participants were paid following completion of their posttreatment assessment.
Results outcome of vret Measures of PTSD, anxiety, and depression were examined using dependent-samples t-tests, comparing scores pre- and posttreatment for all treatment completers, given that directional hypotheses were tested. As shown in Table 1, significant reductions from pretreatment were noted on the CAPS total (p < .05), as well as the CAPS Reexperiencing (p < .05) and Avoidance/Numbing (p < .05) cluster scores. No significant difference was found on the CAPS Hyperarousal cluster score. An identical pattern of results was noted on the IES-R, with significant reductions in scores from pre- to posttreatment noted on the IES-R total score (p < .05), Intrusion subscale (p < .05), and Avoidance subscale (p < .05). On the PSS-SR, a significant difference was noted (p < .05) as well. Examination of scores on the anxiety and depression measures failed to reveal any significant differences from preto posttreatment.
Because 2 patients dropped out of VRET, it was important to examine outcome using an intent-totreat analysis, in order to bracket effect sizes of VRET on each dependent measure. Thus, analyses were completed for the entire sample of 8, using pretreatment data for the 2 individuals who prematurely discontinued VRET. The same pattern of significant effects was found as noted with the treatment completers, although effect sizes are smaller as would be expected (see Table 1). In order to place the results in a larger context, participants’ individual scores were examined to determine if they represented clinically significant and/or reliable change from pre- to posttreatment. Clinically significant change (CSC) occurs when an individual's posttreatment score is closer to the functional community mean than to the pretreatment mean of the sample. A participant's posttreatment score is considered to meet this criteria when it exceeds a cutoff score (Jacobson & Truax, 1991). Using the method described by Jacobson and Truax, cutoff scores were calculated using means and standard deviations from the current sample at pretreatment and from functional community samples for the IES-R (Creamer, Bell, & Failla, 2003; cutoff scores as follows: IES-R total 1.99, IES-R Intrusion 1.94, IES-R Avoidance 1.81, IES-R Hyperarousal 2.25), BAI (Creamer, Foran, & Bell, 1995; cutoff score = 17.44), and BDI-II (Whisman, Perez, & Ramel, 2000; cutoff score = 12.93).5 As 5
The CAPS and PSS-SR were not included in these analyses because normative data from community samples were not available for these measures.
virtual reality exposure therapy for ptsd noted in Table 1, the entire sample met criteria for CSC on the IES-R Total, Intrusion, and Hyperarousal scales at posttreatment, while 5 of 6 participants (83%) met criteria on the Avoidance scale. Three participants (50%) were classified as showing CSC on the BAI and BDI-II. In contrast to CSC, reliable change (RC) represents a means of quantifying how much change has occurred over the course of treatment, and as such, is a more stringent index of change. As discussed by Jacobson and Truax (1991), when functional and pretreatment distributions are overlapping (as was the case in this study), it is possible for a patient’s posttreatment score to exceed the cutoff but not represent statistically reliable change. Using Jacobson and TruaxTs (1991) method for calculating RC, the posttreatment scores for the sample were examined in order to determine if they met criteria for RC. As shown in Table 1, 4 (67%) participants met criteria for RC on the IES-R Total, 5 (83%) met criteria on the Intrusion subscale, 3 (50%) on the Avoidance subscale, and 3 (50%) on the Hyperarousal subscale of the IES-R. One individual (16%) met criteria on the BAI. Two participants (33%) showed RC on the BDI-II.
reactions to vret In order to examine patients’ reactions to treatment, scores on the SSQ and PQ were calculated and compared to published norms (Kennedy et al., 1993; Witmer & Singer, 1998), using one-sample t-tests. As can be seen in Table 2, significantly higher levels of disorientation were reported by
Table 2 Means and standard deviations for simulator sickness and presence questionnaires1 (n = 6) Current Sample (Aggregated across sessions)
Established Norms
M
SD
M
SD
SSQ 2 Nausea Oculomotor disturbances Disorientation
12.72 20.55 17.01*
6.75 11.83 7.81
7.70 10.60 6.4
15.00 15.00 15.00
PQ Realism Interface Quality Involvement/Control
17.52** 17.46* 67.43**
2.22 2.29 2.46
12.36 14.65 57.39
3.44 3.40 8.96
1
Statistically significant differences involve comparison with established norms, *p < .05, **p < .001. 2 Data are derived from n = 5, as one participant reported a vestibular disorder, which rendered her responses on this measure unusable. SSQ = Simulator Sickness Questionnaire; PQ = Presence Questionnaire.
45
this sample, t(4) = 3.04, p < .05, relative to established norms, although no differences were noted on the nausea and oculomotor disturbance subscales, ts(4) = 1.66 and 1.88, respectively. In contrast, this sample reported significantly higher levels of presence relative to established norms on each of the three subscales of the PQ: realism, t(5) = 5.68, p < .005; interface quality, t(5) = 3.01, p < .05; and involvement/control, t(5) = 52.55, p < .0001. Scores on the CSQ were examined to determine participants’ overall satisfaction with the treatment. All participants scored 30 or greater (M = 31, SD = 0.89), suggesting high levels of satisfaction with VRET.
adherence to exposure within vret sessions As mentioned, participants were encouraged to stay in the selected VR simulation until their anxiety decreased by 50%. In total, 80 exposure scenarios were attempted (across all patients) and 90% (n = 72) were responded to with a SUDS rating of 20 or higher, based on written progress notes following each session. Of the exposure scenarios that were attempted, reductions of 50% (or more) SUDS units were achieved in 67% (n = 48). With 33% of scenarios not meeting this criterion, it appears that a notable number of scenarios were attempted but not endured until the patient’s SUDS reduced by 50%. Examination of the association between the percentage of scenarios in which a 50% reduction in SUDS occurred and the difference between the patient’s CAPS total score at pretreatment minus posttreatment revealed an r = .33, ns. This suggests that outcomes might have been improved had the therapists been more consistent with achieving 50% reduction in anxiety within VRET.
Discussion The current report examined the efficacy of VRET for the treatment of MVA-related PTSD in an uncontrolled case series. Results from 6 patients indicated significant reductions in posttrauma symptoms involving reexperiencing, avoidance, and emotional numbing, with effect sizes ranging from d = .79 to d = 1.49. Indices of clinically significant and reliable change (Jacobson & Truax, 1991) suggested that the magnitude of these changes was meaningful and reliable. Measures of anxiety and depression did not show significant reductions before and after treatment, although moderate effect sizes were noted for the BAI and the BDI-II. High levels of presence were achieved by the virtual display, and participants reported satisfaction with treatment.
46
beck et al.
In interpreting these data, it seems important to place these results alongside other VRETs for PTSD and driving-related fears. In drawing these comparisons, factors such as the severity and chronicity of the study samples and the use of clinician versus self-report measures are salient. Rothbaum et al. (2001) reported an average 15% reduction in CAPS total score following 8 to 16 sessions of VRET for 9 veterans with chronic combat-related PTSD. Wald (2004) obtained an average 60% reduction in selfrated global phobia severity, following 8 sessions of VRET with 5 individuals diagnosed with driving phobia. In the current report, an average 40% reduction was noted on the CAPS total score following treatment, with greater percentage reductions noted for self-report measures (IES-R: 54% average reduction; PSS-SR: 49% average reduction). Thus, the 10-session VRET used in this trial obtained results that were comparable or better than results published in other uncontrolled trials of VRET for related problems. One feature of these results which deserves note is the lack of treatment effects for hyperarousal symptoms. Although relaxation training was included in the first two sessions of treatment, it would appear that patients may not have mastered this skill or may not have generalized this skill to nondriving situations. Specifically, therapists prompted the use of relaxation during VRET to facilitate habituation. Because the hyperarousal symptom cluster includes a number of nonspecific symptoms (e.g., sleep disturbance, irritability), the lack of significant pre-post treatment differences on both clinician and self-report measures of hyperarousal suggest that relaxation skills did not generalize or were not used by patients outside of driving situations. From a theoretical standpoint, one would expect that the three symptom clusters that comprise PTSD would show equal responsiveness to exposure therapy, whether conducted using virtual reality technology or not (e.g., Foa & Rothbaum, 1998). Although the process of symptom change has been studied with prolonged exposure (Nishith, Resick, & Griffin, 2002), it has not been examined with more gradual approaches to exposure, which theoretically could be expected to produce less synchrony across symptom clusters. From a conceptual standpoint, further study of this issue could deepen our understanding of the mechanisms through which exposure therapy works. One facet of this report that deserves additional comment is the visual simulation system. This software was developed specifically for this project. It was constructed for maximum flexibility, in order to permit the creation of virtual environments that matched individual feared situations (Ang, Schwag-
ler, English, Beck, & Winer, 2006). The software was designed to mimic a natural driving situation where a driver can determine her own route. As such, all aspects of the simulation (i.e., roads, traffic, signage, and buildings) are rendered in real time, based on the route that the driver takes. As noted, patients reported significantly higher levels of disorientation, relative to published norms, possibly owing to the use of stereoscopic glasses. Further refinement of the simulation system should consider adaptation to a head-mounted display, as simulation sickness appears less likely with this type of application. Efforts are underway to adapt the VRET software to a desktop computing system, to facilitate general use within clinical and research settings. In considering these findings, it is relevant to discuss the limitations of this report. A clear limit to the generalizability of these results rests with the fact that it involved an open clinical trial research design. Future studies of VRET for MVA-related PTSD should utilize a controlled design with randomization in order to account for numerous possible confounds. Second, because the patient sample in the current report involved individuals who had experienced a traumatic MVA less than 7 months prior to entry into the trial, it is possible that natural symptom remission could have influenced these results. While possible, it is unlikely that remission fully accounted for these effects as half of the patients in this report had experienced symptoms for at least 4 months. As reviewed by Blanchard and Hickling (2004), the rate of natural remission appears to stabilize approximately 4 months after a traumatic MVA. Ideally, future research could examine the influence of chronicity of PTSD on the efficacy of VRET and the role of natural remission in this process. Third, in a notable percentage of exposure trials, the trial was ended before the patient’s SUDS reduced by 50%. Given the association between the number of exposure trials where 50% anxiety reduction occurred and CAPS scores at posttreatment, this clearly influenced the efficacy of VRET in this report. This suggests that participants who exhibit greater habituation of fear within each session will realize better treatment outcomes. To ensure that this occurs, the use of therapists who have more experience with exposure-based therapy may be warranted. This observation also highlights the key role that virtual reality technology can play in parametric studies designed to examine the salient parameters of exposure during ET. Finally, this report is limited by the lack of follow-up data. Given the positive outcomes that were noted at posttreatment, follow-up data would be an
virtual reality exposure therapy for ptsd important methodological addition to examine the durability of change from VRET. In sum, preliminary data from this uncontrolled case series suggest that VRET can be used for the successful treatment of PTSD symptoms among survivors of traumatic MVAs. Observed effect sizes were comparable to trials involving the virtual treatment of related problems, with high levels of patient satisfaction. Future controlled studies are needed to provide stronger experimental support for VRET for this patient population. References American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders, (4th ed., text revision). Washington, DC: Author. Ang, E. J., Schwagler, B., English, K., Beck, J. G., & Winer, E. (2006). Development of an automated tile-based scene generator for driving simulation in virtual reality exposure therapy. Manuscript submitted for publication. 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. Beck, A. T., Steer, R., & Brown, G. (1996). Beck Depression Inventory manual, (2nd ed.). San Antonio, TX: Psychological Corporation. Beck, J. G., & Coffey, S. F. (2005). Group cognitive behavioral treatment of PTSD: Treatment of motor vehicle accident survivors. Cognitive and Behavioral Practice, 12, 267–277. Beck, J. G., Coffey, S. F., Palyo, S. A., Gudmundsdottir, B., Miller, L. M., & Colder, C. R. (2004). Psychometric properties of the Posttraumatic Cognitions Inventory (PTCI): A replication with motor vehicle accident survivors. Psychological Assessment, 16, 289–298. Beck, J. G., Grant, D., Read, J. P., Clapp, J., Coffey, S., Miller, L., & Palyo, S. A. (2006). The Impact of Event Scale–Revised: Psychometric properties in a sample of motor vehicle accident survivors. Manuscript in preparation. Bernstein, D. A., & Borkovec, T. D. (1973). Progressive relaxation training: A manual for helping professions. Champaign, IL: Research Press. Blake, D., Weathers, F., Nagy, L., Kaloupek, . D., Klauminzer, G., Charney, D., & Keane, T (1990). Clinician-administered PTSD scale (CAPS). Boston: National Center for PostTraumatic Stress Disorder, Behavioral Science Division. Blanchard, E. B., & Hickling, E. J. (2004). After the crash: Assessment and treatment of motor vehicle accident survivors, (2nd ed.). Washington, DC: American Psychological Association. Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., Forneris, C. A., & Jaccard, J. (1996). Who develops PTSD from motor vehicle accidents? Behaviour Research and Therapy, 34, 1–10. Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., & Gerardi, R. J. (1994). Psychological morbidity associated with motor vehicle accidents. Behaviour Research and Therapy, 32, 283–290. Botella, C., Banos, R., Perpina, C., Villa, H., Alcaniz, M., & Rey, A. (1998). Virtual reality treatment of claustrophobia: A case report. Behaviour Research and Therapy, 36, 239–246. Creamer, M., Bell, R., & Failla, S. (2003). Psychometric
47
properties of the Impact of Event Scale–Revised. Behaviour Research and Therapy, 41, 1489–1496. Creamer, M., Foran, J., & Bell, R. (1995). The Beck Anxiety Inventory in a non-clinical sample. Behaviour Research and Therapy, 33, 477–485. Emmelkamp, P., Krijn, M., Hulsbosch, A., deVries, S., Schuemie, M., & van der Mast, C. (2002). Virtual reality treatment versus exposure in vivo: A comparative evaluation in acrophobia. Behaviour Research and Therapy, 40, 509–516. Foa, E. B., & Kozak, M. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Foa, E. B., & Meadows, E. (1997). Psychosocial treatments for post-traumatic stress disorder: A critical review. In J. Spence, J. Darley, & D. Foss (Eds.), Annual review of psychology, (vol. 48). pp. 449–480. Palo Alto: Annual Reviews. Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. O. (1993). Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress, 6, 459–473. Foa, E. B., & Rothbaum, B. O. (1998). Treating the trauma of rape: Cognitive behavioral therapy for PTSD. New York: The Guilford Press. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Minimental state”: A practical method of grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. Imago Systems, Inc. (1996). DriVR user guide. Vancouver, BC: Author. International Society of Presence Research. (n.d.). The concept of presence: Explication statement. Retrieved August 3, 2005, from http://ispr.info/ Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19. Kennedy, R. S., Lane, N. E., Berhaum, K. S., & Lilienthal, M. G. (1993). A simulator sickness questionnaire (SSQ): A new method for quantifying simulator sickness. International Journal of Aviation Psychology, 3, 203–220. Krijn, M., Emmelkamp, P. M. G., Olafsson, R. P., & Biemond, R. (2004). Virtual reality exposure therapy for anxiety disorders: A review. Clinical Psychology Review, 24, 259–281. Kuch, K., Swinson, R. P., & Kirby, M. (1985). Post-traumatic stress disorder after car accidents. Canadian Journal of Psychiatry, 30, 426–427. Larsen, D. L., Attkisson, C., Hargreaves, W. A., & Nguyen, T. D. (1979). Assessment of client/patient satisfaction: Development of a general scale. Evaluation and Program Planning, 2, 197–207. Maltby, N., Kirsch, I., Mayers, M., & Allen, G. (2002). Virtual reality exposure therapy for the treatment of fear of flying: A controlled investigation. Journal of Consulting and Clinical Psychology, 70, 1112–1118. Nishith, P., Resick, P. A., & Griffin, M. G. (2002). Pattern of change in prolonged exposure and cognitive-processing therapy for female rape victims with posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 70, 880–886. Rothbaum, B., Hodges, L., Ready, D., Graap, K., & Alarcon, R. (2001). Virtual reality exposure therapy for Vietnam veterans with posttraumatic stress disorder. Journal of Clinical Psychiatry, 62, 617–622. Rothbaum, B., Meadows, E., Resick, P., & Foy, D. (2000). Cognitive-behavioral therapy. In E. Foa, T. Keane, & M. Friedman (Eds.), Effective treatments for PTSD (pp. 60–83). New York: The Guilford Press.
48
beck et al.
Sprinkle, S. D., Lurie, D., Insko, S. L., Atkinson, G., Jones, G. L., Logan, A. R., & Bissada, N. N. (2002). Criterion validity, severity cut scores, and test-retest reliability of the Beck Depression Inventory-II in a university counseling center sample. Journal of Counseling Psychology, 49, 381–385. Taylor, S., & Koch, W. J. (1995). Anxiety disorders due to motor vehicle accidents: Nature and treatment. Clinical Psychology Review, 15, 721–738. Wald, J. (2004). Efficacy of virtual reality exposure therapy for driving phobia: A multiple baseline across-subjects design. Behavior Therapy, 35, 621–635. Weathers, F. W., Keane, T., & Davidson, J. (2001). ClinicianAdministered PTSD Scale: A review of the first ten years of research. Depression and Anxiety, 13, 132–156. Weiss, D. S., & Marmar, C. R. (1997). The Impact of Event Scale-Revised. In J. P. Wilson, & T. M. Keane (Eds.),
Assessing psychological trauma and PTSD (pp. 399–411). New York: The Guilford Press. Whisman, M., Perez, J. E., & Ramel, W. (2000). Factor structure of the Beck Depression Inventory-Second Edition (BDI-II) in a student sample. Journal of Clinical Psychology, 56, 545–551. Witmer, B. G., & Singer, M. J. (1998). Measuring presence in virtual environments: A presence questionnaire. Presence, 7, 225–240.
R E C E I V E D : October 13, 2005 A C C E P T E D : April 19, 2006 Available online 22 September 2006
www.elsevier.com/locate/bt
Virtual Reality Exposure Therapy for PTSD Symptoms After a Road Accident: An Uncontrolled Case Series J. Gayle Beck, Sarah A. Palyo, SUNY Buffalo Eliot H. Winer, Brad E. Schwagler, Eu Jin Ang, NYSCEDII and SUNY Buffalo
This report examined whether Virtual Reality Exposure Therapy (VRET) could be used in the treatment of posttraumatic stress disorder (PTSD) symptoms in the aftermath of a serious motor vehicle accident. Six individuals reporting either full or severe subsyndromal PTSD completed 10 sessions of VRET, which was conducted using software designed to create real-time driving scenarios. Results indicated significant reductions in posttrauma symptoms involving reexperiencing, avoidance, and emotional numbing, with effect sizes ranging from d = .79 to d = 1.49. Indices of clinically significant and reliable change suggested that the magnitude of these changes was meaningful. Additionally, high levels of perceived reality (“presence”) within the virtual driving situation were reported, and patients reported satisfaction with treatment. Results are discussed in light of the possibility for VRET to be useful in guiding exposure in the treatment of PTSD following road accidents.
C O N S I D E R A B L E AT T E N T I O N H A S F O C U S E D recently on the treatment of posttraumatic stress disorder (PTSD). At present, strong empirical evidence exists to support the efficacy of cognitive Eliot Winer is now at Iowa State University. Brad Schwagler is now at the Georgia Institute of Technology. This project was supported in part by a grant from the University at Buffalo (IRCAF 24623). Additional support from NYSCEDII and the Department of Psychology is acknowledged as well. We are especially appreciative of help received from Dr. Ken English in completion of the project. We appreciate the assistance offered by Sumeet Parashar, Gautam Agrawal, Sherry Farrow, Luana Miller, and Berglind Gudmundsdottir in the completion of this project. Dr. Barbara Rothbaum provided helpful input on an earlier draft of this manuscript. Address correspondence to J. Gayle Beck, Department of Psychology, Park Hall, University at Buffalo, Buffalo, New York, 14260, USA; e-mail: [email protected]. 0005-7894/06/039–048/$1.00/0 © 2006 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.
behavioral treatments for PTSD, particularly those treatments that emphasize exposure to traumarelated stimuli and cues (Rothbaum, Meadows, Resick, & Foy, 2000). Exposure therapy (ET) has been used with a wide range of trauma populations (e.g., combat veterans, female survivors of sexual and physical assault), and the controlled studies that support its use meet the “gold standard” for clinical research (Foa & Meadows, 1997). Successful ET involves providing the patient with PTSD with prolonged contact with trauma-related cues in order to facilitate habituation and the development of neutral memory structures that “override” anxiety-provoking memories related to the trauma (Foa & Kozak, 1986). Although ET clearly is an efficacious approach in the treatment of PTSD, there are circumstances in which in vivo forms of ET become difficult to implement. The treatment of PTSD following a motor vehicle accident (MVA) represents one such circumstance, given the nature of the trauma-related cues and the fact that many individuals with MVA-related PTSD avoid driving completely (Blanchard & Hickling, 2004). In this article, preliminary data are presented on the use of virtual reality–assisted ET (VRET) in the treatment of acute-phase PTSD following a serious MVA in a beginning effort to examine the efficacy and acceptability of this treatment modality. The use of VRET is relatively recent, although studies to date have shown positive results of this treatment for a panoply of anxiety disorders, including claustrophobia (e.g., Botella et al., 1998), acrophobia (e.g., Emmelkamp et al., 2002), and fear of flying (e.g., Maltby, Kirsch, Mayers, & Allen, 2002). In these studies, virtual reality techniques are used to create environments that are immersive and that provoke anxiety. The patient is encouraged to stay in the virtual environment for a prolonged time interval, until fear reduction occurs. Thus, VRET
40
beck et al.
represents a step between imaginal exposure and in vivo exposure. As discussed by Krijn, Emmelkamp, Olafsson, and Biemond (2004), patients need to feel present in the virtual environment in order to experience it in the first person. Within this literature, “presence” is conceptualized as the subjective perception that the virtual world is somewhat real. Stated differently, “at some level and to some degree, . . . objects, events, entities, and environments are perceived as if the technology was not involved in the experience” (International Society for Presence Research, n.d., 2005). Additionally, the virtual environment needs to be able to elicit affect. Lastly, the virtual world needs to resemble the natural environment in order to ensure generalization of fear reduction. Thus, simulation sickness (which is akin to motion sickness in the virtual world) must be kept to a minimum. These qualities have been highlighted as necessary in order for VRET to be effective, following from the guidelines presented by Foa and Kozak (1986). Preliminary data collected on Vietnam veterans suggest that VRET can be used effectively for treating PTSD. Rothbaum and colleagues have presented a clinical trial (Rothbaum, Hodges, Ready, Graap, & Alarcon, 2001) in which VRET was used as one component of a multicomponent intervention for combat-related PTSD. In this report, VRET was included along with breathing relaxation (for general stress management) and imaginal exposure. In the open clinical trial, 9 veterans completed 8 to 16 treatment sessions. Statistically significant and clinically meaningful reductions in PTSD symptomatology were noted, suggesting that VRET (when combined with other interventions) holds promise in the treatment of chronic combat-related PTSD. Given the substantial number of individuals who are involved in an MVA each year and the estimate that 800,000 new cases of PTSD result from these crashes (Blanchard & Hickling, 2004), the need for a safe and cost-efficient treatment is clear. VRET represents a viable option for the treatment of MVArelated PTSD for a number of reasons. First, in contrast to in vivo exposure, VRET permits closer control over stimulus presentation. For example, if the patient wanted to work on exposure to driving near trucks in a highway environment, this would be easily accomplished within a VRET situation. In real life, setting up an exposure scenario to address this fear might prove challenging because it is difficult to predict when trucks would appear on a given highway. Second, VRET is especially well suited to exposure situations that involve time-limited stimuli (e.g., merging onto a freeway), as it is possible to
conduct repeated exposure easily. In real-life exposure, this is considerably more difficult. Thus, VRET offers greater convenience for therapist and patient alike. Third, VRET may reduce safety risks that may be associated with in vivo exposure, particularly given that individuals with MVA-related PTSD may report overly defensive driving styles that heighten risk for another MVA (e.g., J.G. Beck & Coffey, 2005). Fourth, VRET may offer a more “tangible” exposure experience, relative to imaginal exposure, possibly facilitating fear reduction more effectively. Lastly, some individuals with PTSD report a complete inability to drive, based on intense fear and anxiety. VRET may be a less threatening (and thus, more acceptable) form of treatment compared to in-vivo exposure for individuals who are extremely avoidant. Although there have been no published studies of VRET for MVA-related PTSD, there has been one effort to treat individuals with a specific phobia of driving using this approach (Wald, 2004). In this report, a commercially available driving simulator was used (driVR, Imago Systems, 1996). Six standardized scenes were used, in which the individual proceeded down a road in a predetermined path. Each scene was 3 to 5 minutes in duration, which is quite brief in light of the usual requirements for ET. A variety of options were available with respect to weather and time of day, although the specific driving route was repeated on subsequent exposure trials. Five women were treated in a multiple-baseline design, each receiving 8 weekly sessions. Three individuals showed reductions in phobia symptoms based on a clinician-administered interview, although the other two did not. VRET did not result in an increase in actual driving time or driving frequency for any patient in this report. Thus, even with short exposure stimuli and limited flexibility, VRET was somewhat effective for 3 of 5 (60%) of the cases of driving phobia. Clearly, this technology may hold promise for the treatment of driving-related fear. Although driving phobia and MVA-related PTSD share the common characteristic of avoidance of driving cues, many writers believe that the psychopathology and treatment of these two disorders are somewhat different (Kuch, Swinson, & Kirby, 1985; Taylor & Koch, 1995). In particular, the range of feared cues and degree of anxiety appears greater among patients with MVA-related PTSD. Thus, the goal of this report was to examine whether VRET could be used, in conjunction with relaxation training, to facilitate the treatment of MVA-related PTSD. Individuals in the acute phase of the disorder were selected for treatment and provided with 10 sessions of treatment. This report utilized an
virtual reality exposure therapy for ptsd uncontrolled case design, in light of the absence of previous studies on this issue. We hypothesized that VRET would reduce PTSD symptoms (intrusive reminders of the trauma, avoidance of traumarelated cues, emotional numbing, and physiological hyperarousal). Additional measures of anxiety and depression were included to examine the generalizability of treatment effects.
Method participants Participants were recruited for this trial as part of a larger study on the treatment of MVA-related PTSD. Recruitment efforts included newspaper ads as well as referrals from physicians, psychiatrists, counselors, chiropractors, physical therapists, and massage therapists. To qualify for participation, individuals must have experienced an MVA that involved actual or threatened death or serious injury and evoked an emotional response that included intense fear, helplessness, horror, or the perception that they would die (American Psychiatric Association, 2000), as assessed using the Motor Vehicle Accident Interview (Blanchard & Hickling, 2004). Only individuals who had experienced an MVA in the past 6 months were included. Exclusion criteria included cognitive impairment (indexed by a score below 24 on the Mini Mental Status Exam; Folstein, Folstein, & McHugh, 1975), current substance abuse or dependence, or suicidality. Eleven individuals were offered entry into the study. Three patients declined participation (two did not desire VRET and the third reported acute distress which warranted a more broad-band treatment to address severe depression as well as PTSD).1 Participants met criteria for full (n = 6) or subsyndromal (n = 2) PTSD. Full PTSD was defined as meeting DSM-IV criteria for the disorder. Subsyndromal PTSD was defined as meeting diagnostic criteria for the reexperiencing cluster and for either the avoidance and numbing or hyperarousal cluster (but not both), in accordance with Blanchard, Hickling, Taylor, Loos, and Gerardi (1994).2 Participants were required to suspend other psychological treatments during their participation. Individuals taking psychotropic
1 All three individuals who declined VRET met full criteria for PTSD. 2 Our decision to include patients with subsyndromal PTSD was based on reasons similar to those discussed by Blanchard and Hickling (2004), specifically that these individuals report clinically significant distress and impairment resulting from PTSD symptoms.
41
medication were asked to keep the dosage constant, in consultation with their prescribing physician. Of the 8 participants (7 females and 1 male), 2 participants were African American, and 6 participants were Caucasian. The mean age for the sample was 49.50 (SD = 7.03). Participants were initially assessed an average of 3.50 months (SD = 2.07) after their MVA. Two patients dropped out of VRET. One participant dropped after Session 3 due to an increase in family stress. The other participant, who had been on disability prior to VRET, dropped after Session 4 because she had to return to work and no longer had time to attend treatment sessions.3 The sample of treatment completers included 5 females and 1 male. Four patients met criteria for full PTSD, while 2 met criteria for subsyndromal PTSD. The final sample had an average age of 49.17 (SD = 8.28). Two participants were African American, and the remainder was Caucasian. Four participants were currently on disability, while the other 2 participants were employed. Half of the participants reported that their annual income was less than $50,000. These participants were initially assessed an average of 3.5 months (SD = 1.87) after their MVA (3 patients <4 months post-MVA).
outcome measures PTSD measures. Measures of PTSD included a clinician-administered interview and two selfreport measures. The Clinician-Administered PTSD Scale (CAPS; Blake et al., 1990) is a structured interview designed to assess the frequency and intensity of the 17 symptoms of PTSD during the last month rated on a 5-point Likert-type scale, where 0 indicates that the symptom has not occurred or is not distressing and 4 indicates that the symptom occurs nearly every day or is severely distressing. The CAPS was scored by summing the frequency and intensity ratings for both the entire measure (total CAPS score) and each symptom cluster. The CAPS has excellent support for its reliability, with alpha coefficients generally ranging from .73 to .98 and 2- to 3-day test-retest reliability ranging from .78 to .87 (Weathers, Keane, & Davidson, 2001). Alpha coefficients for the CAPS derived from a similar sample (n = 112; J.G. Beck et al., 2004) were .80 (reexperiencing symptom cluster), .71 (avoidance-numbing symptom cluster), .69 (physiological symptom cluster), and .87 (total score). In order for a symptom to count toward a diagnosis of full or subsyndromal PTSD, a participant
3
Nonsignificant differences were noted between treatment completers and noncompleters on pretreatment scores on each of the outcome variables.
42
beck et al.
must score a 1 or higher on frequency and a 2 or higher on intensity (Blanchard et al., 1996). The CAPS was administered by extensively trained doctoral students in clinical psychology. Interviews from a larger sample (N = 409) that included the 6 participants in this study were videotaped, and 29% (n = 119) were randomly selected and reviewed by an independent clinician to establish diagnostic reliability. Interrater agreement in PTSD diagnosis, reflected by the kappa statistic, was strong for PTSD (k = 0.88). Additionally, the Posttraumatic Stress Scale–Self Report (PSS-SR; Foa, Riggs, Dancu, & Rothbaum, 1993) and the Impact of Event Scale–Revised (IES-R; Weiss & Marmar, 1997) were administered. The PSS-SR is a 17-item self-report scale that measures the frequency of DSM-IV symptoms of PTSD in the past month. Each item is rated on a 0 to 3 scale, and ratings are summed to form a total score. The PSSSR has been shown to have high internal consistency (α = .91) and good 1-month test-retest reliability (r = .74) in a sample of 46 female rape victims and 72 female non-sexual-assault victims (Foa et al., 1993). In a similar sample as the present report, the PSS-SR was shown to have good internal consistency (α = .93, N = 112; J.G. Beck et al., 2004). The IES-R is a 22-item self-report measure designed to assess for the experiences of intrusion, avoidance, and hyperarousal following a distressing event. Items are rated on a 0 to 4 scale and summed to form a total score, as well as subscales reflecting intrusion, avoidance, and hyperarousal symptoms. Psychometric data reported by Weiss and Marmar (1997) suggest that the IES-R has high internal consistency (alphas ranging from .79 to .92 on each of the subscales) and good test-retest reliability (correlations ranging from .51 to .94 on each of the subscales). Data from a larger sample derived from the same site as the current report indicate alphas ranging from .85 to .90 on each of the subscales (N = 199; J.G. Beck et al., 2006). Measures of anxiety and depression. Self-reported anxiety was measured using the Beck Anxiety Inventory (BAI; A.T. Beck, Epstein, Brown, & Steer, 1988). The BAI is a 21-item self-report inventory of common symptoms of anxiety (e.g., difficulty breathing, feelings of choking). Participants rate the extent to which they were bothered by each item during the past week on a 0-3 scale ranging from “not at all” to “severely.” The inventory has been shown to have high internal consistency (a = .92) and test-retest reliability (r = .75) in a sample of psychiatric outpatients (A. T. Beck et al., 1988). In a similar sample as the present report, the BAI also was internally consistent (α = .93, N = 199; J.G. Beck et al., 2006).
The Beck Depression Inventory-II (BDI-II; A.T. Beck, Steer, & Brown, 1996) was used to measure depression. The BDI-II is a 21-item measure used to evaluate depressive symptoms during the past two weeks. In a study of 137 students receiving treatment at a university counseling center, the BDI-II yielded reliability coefficients of .91 and .93 (Sprinkle et al., 2002). A one-week test-retest correlation of .93 was reported for a sample of 26 outpatients (A.T. Beck et al., 1996). In a similar sample as the present report, the BDI-II also was internally consistent (α = .93, N = 199; J.G. Beck et al., 2006). Treatment-related measures. Two self-report measures were included to assess simulation sickness and the sense of presence that was created by the virtual environment. Additionally, participants completed the Client Satisfaction Questionnaire (CSQ; Larsen, Attkisson, Hargreaves, & Nguyen, 1979) at the end of treatment. The Simulator Sickness Questionnaire (SSQ; Kennedy, Lane, Berhaum, & Lilienthal, 1993) lists 26 symptoms that are associated with physical discomfort experienced in VR simulators. Each symptom is rated as either none, slight, moderate, or severe. Symptoms are based on three subscales that describe specific dimensions of simulator sickness (i.e., nausea, oculomotor disturbances, and disorientation). The SSQ has been used extensively in studies of simulator sickness and has been shown to be a valid measure of this construct (Kennedy et al., 1993). The established norms for the SSQ were derived from 1,119 exposures to simulators at 10 different simulator sites (Kennedy et al., 1993). In the current study, the SSQ was administered to participants immediately after they had been in the simulator and averaged across exposure sessions to compare with established norms. Given the small sample size, coefficient alpha could not reliably be computed for the current report. The Presence Questionnaire (PQ; Witmer & Singer, 1998) is designed to measure the degree to which individuals experience presence in a virtual environment. This 19-item measure asks the participant to indicate on a 7-point Likert-type scale how immersed they felt in the virtual environment, with higher scores indicating a greater sense of presence in the virtual environment. The PQ has three subscales: Realism, Interface Quality, and Involvement/Control. The Realism subscale refers to the extent to which the interactions between the participant and the virtual environment felt consistent with reality. The Interface Quality subscale measures whether control devices distract from task performance. The Involvement/Control subscale addresses perceived control of events in the virtual
virtual reality exposure therapy for ptsd environment, responsiveness of the virtual environment to user-initiated actions, and how involved in the experience the participant became. Established norms were determined using a sample of 152 college students, and within this sample, the PQ was shown to have good internal consistency (a = .88; Witmer & Singer, 1998). In the current study, the PQ was administered to participants immediately after exiting the simulator and averaged across exposure sessions to compare with established norms. Given the small sample size, coefficient alpha could not reliably be computed for the current report. The CSQ is a measure of participant's satisfaction with treatment. This 8-item measure has been shown to be reliable (a = .95) in a sample of 49 outpatients (Larsen et al., 1979). Participants rate their satisfaction with treatment on a 4-point Likert-type scale. Scores on the CSQ can range from 8 to 32, with higher scores indicating greater satisfaction. The CSQ was administered at the end of the last VRET session. Given the small sample size, coefficient alpha could not reliably be computed for the current report.
vret VRET was administered in 10 individual sessions. The first 2 sessions included psychoeducation about the symptoms of PTSD and the purpose of VRET, relaxation training (Bernstein & Borkovec, 1973), construction of a hierarchy of anxiety-producing driving situations, and familiarization with the VR simulator using neutral (non-MVA-related) scenarios and were 90 minutes long. In the remaining 8 sessions, participants were instructed on how to do exposure and engaged in virtual reality exposure. In this protocol, relaxation training was included to address defensive behaviors that have been observed in this population (J.G. Beck & Coffey, 2005), including closing ones eyes while experiencing anxiety in the car or rigidly clutching the steering wheel in a fixed path, owing to extreme anxiety. This intervention was designed to help patients to focus on the exposure situation by giving them a coping strategy. Therapists were instructed not to encourage the use of relaxation as a distraction from exposure. Depending on the specific exposure situation, the participant would either be the driver or the passenger, with the therapist taking the other role. The therapist guided the participant through the exposure process, encouraging them to use relaxation skills, and to provide a rating of anxiety using the subjective units of distress (SUDS) scale (range: 0–100). In keeping with other exposure-based treatments (e.g., Foa & Rothbaum, 1998), participants were encouraged to
43
remain in the specific VR driving situation until their anxiety had reduced by 50%. Patients initially selected items on their hierarchies that caused mild anxiety and, as habituation occurred, gradually moved on to more anxiety-provoking situations. Sessions 3 through 10 were 60 minutes in length, and participants were assigned imaginal and in vivo exposure homework between sessions.4 The treatment was administered by two doctoral students, with supervision by a licensed psychologist. All sessions were audiotaped and 52% (n = 31) were selected for treatment integrity monitoring. A checklist of techniques deemed both appropriate (e.g., exposure) and inappropriate (e.g., cognitive therapy) for VRET was used to monitor treatment integrity. Examination of these data revealed no violations in the treatment protocol.
apparatus VRET was conducted using an SGI Origin 3400 and a Moog six degree-of-freedom motion base for the computational and graphics system. The motion base was outfitted with a steering wheel, gas pedal, and a brake pedal. Through PVM software (Parallel Virtual Machine), inputs from the motion base were sent to the graphics supercomputer to be integrated with the immersive visual representations being generated. In this manner, correct viewpoints and perspectives were computed and displayed for a user. The visualization system projected the simulated scene on a 10 foot by 8 foot screen that was located 5 feet away from the motion base. During VRET, both the patient and the therapist wore stereoscopic glasses to view the simulation in true 3D. virtual environments The visual simulation consisted of several adaptable and customizable driving scenarios. Four general types of scenes were available: highway, urban, suburban, and rural. Within each scene type, the therapist was able to predetermine the amount and type of traffic, time of day, weather conditions (i.e., snow, rain), and specific driving events (e.g., a car rapidly comes up from behind the driver, tailgating). This gave direct control over the level of anxiety produced for the patient as well as the patient’s overall comfort level with the VRET. None of the driving events were designed to re-create an MVA but were intended to simulate a real-world driving experience. During each VRET experience, the driver was able to determine his or her own 4
In the homework, imaginal exposure to tasks on the patient’s graded hierarchy was assigned, not exposure to memories of the traumatic MVA.
44
beck et al.
Table 1 Examination of pretreatment and posttreatment outcome measures (n = 6) Pre M (SD)
Post M (SD)
t (df = 5)
57.83 (17.02) 21.00 (7.77) 20.67 (2.26) 16.17 (6.49) 29.00 (11.98) 2.13 (0.91) 2.13 (1.11) 2.02 (0.94) 2.28 (1.08)
34.50 (20.81) 12.50 (8.31) 8.5 (8.04) 13.50 (9.89) 14.67 (8.17) .98 (.39) .73 (.29) 1.10 (.68) 1.17 (.70)
3.76* 3.26* 3.55* 1.02 3.56* 3.30* 3.59* 2.97* 2.07
6 6 5 6
Anxiety/Depression Measures BAI 23.00 (12.28) BDI-II 19.00 (9.51)
14.00 (10.02) 12.17 (7.68)
1.84 1.63
3 (50%) 3 (50%)
PTSD Measures CAPS Total CAPS Reexperiencing CAPS Avoidance/Numbing CAPS Hyperarousal PSS-SR Total IES-R Total IES-R Intrusion IES-R Avoidance IES-R Hyperarousal
CSC n (%)
(100%) (100%) (83%) (100%)
RC n (%)
Effect size (d)
Effect size for Intent-to-treat (n = 8)
4 (67%) 5 (83%) 3 (50%) 3 (50%)
0.91 0.79 1.49 0.24 1.04 1.21 1.25 0.83 0.90
0.74 0.67 0.93 0.21 0.85 0.97 0.97 0.71 0.73
1 (17%) 2 (33%)
0.60 0.60
0.50 0.49
Note. CAPS = Clinician Administered PTSD Scale; PSS-SR = Posttraumatic Stress Scale–Self-Report; IES-R = Impact of Event Scale, Revised; STAI-S = State-Trait Anxiety Inventory, State subscale; STAI-T = State-Trait Anxiety Inventory, Trait subscale; BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory–II; CSC = Clinically Significant Change; RC = Reliable Change. *p < .05., **p < .001.
course, including choosing which direction to turn, whether to enter a side street, etc.
procedure All participants received a pretreatment assessment, VRET, and a posttreatment assessment 1 month following the last VRET session. Posttreatment CAPS were conducted by interviewers who were unaware of the treatment status of the patient. Participants were paid following completion of their posttreatment assessment.
Results outcome of vret Measures of PTSD, anxiety, and depression were examined using dependent-samples t-tests, comparing scores pre- and posttreatment for all treatment completers, given that directional hypotheses were tested. As shown in Table 1, significant reductions from pretreatment were noted on the CAPS total (p < .05), as well as the CAPS Reexperiencing (p < .05) and Avoidance/Numbing (p < .05) cluster scores. No significant difference was found on the CAPS Hyperarousal cluster score. An identical pattern of results was noted on the IES-R, with significant reductions in scores from pre- to posttreatment noted on the IES-R total score (p < .05), Intrusion subscale (p < .05), and Avoidance subscale (p < .05). On the PSS-SR, a significant difference was noted (p < .05) as well. Examination of scores on the anxiety and depression measures failed to reveal any significant differences from preto posttreatment.
Because 2 patients dropped out of VRET, it was important to examine outcome using an intent-totreat analysis, in order to bracket effect sizes of VRET on each dependent measure. Thus, analyses were completed for the entire sample of 8, using pretreatment data for the 2 individuals who prematurely discontinued VRET. The same pattern of significant effects was found as noted with the treatment completers, although effect sizes are smaller as would be expected (see Table 1). In order to place the results in a larger context, participants’ individual scores were examined to determine if they represented clinically significant and/or reliable change from pre- to posttreatment. Clinically significant change (CSC) occurs when an individual's posttreatment score is closer to the functional community mean than to the pretreatment mean of the sample. A participant's posttreatment score is considered to meet this criteria when it exceeds a cutoff score (Jacobson & Truax, 1991). Using the method described by Jacobson and Truax, cutoff scores were calculated using means and standard deviations from the current sample at pretreatment and from functional community samples for the IES-R (Creamer, Bell, & Failla, 2003; cutoff scores as follows: IES-R total 1.99, IES-R Intrusion 1.94, IES-R Avoidance 1.81, IES-R Hyperarousal 2.25), BAI (Creamer, Foran, & Bell, 1995; cutoff score = 17.44), and BDI-II (Whisman, Perez, & Ramel, 2000; cutoff score = 12.93).5 As 5
The CAPS and PSS-SR were not included in these analyses because normative data from community samples were not available for these measures.
virtual reality exposure therapy for ptsd noted in Table 1, the entire sample met criteria for CSC on the IES-R Total, Intrusion, and Hyperarousal scales at posttreatment, while 5 of 6 participants (83%) met criteria on the Avoidance scale. Three participants (50%) were classified as showing CSC on the BAI and BDI-II. In contrast to CSC, reliable change (RC) represents a means of quantifying how much change has occurred over the course of treatment, and as such, is a more stringent index of change. As discussed by Jacobson and Truax (1991), when functional and pretreatment distributions are overlapping (as was the case in this study), it is possible for a patient’s posttreatment score to exceed the cutoff but not represent statistically reliable change. Using Jacobson and TruaxTs (1991) method for calculating RC, the posttreatment scores for the sample were examined in order to determine if they met criteria for RC. As shown in Table 1, 4 (67%) participants met criteria for RC on the IES-R Total, 5 (83%) met criteria on the Intrusion subscale, 3 (50%) on the Avoidance subscale, and 3 (50%) on the Hyperarousal subscale of the IES-R. One individual (16%) met criteria on the BAI. Two participants (33%) showed RC on the BDI-II.
reactions to vret In order to examine patients’ reactions to treatment, scores on the SSQ and PQ were calculated and compared to published norms (Kennedy et al., 1993; Witmer & Singer, 1998), using one-sample t-tests. As can be seen in Table 2, significantly higher levels of disorientation were reported by
Table 2 Means and standard deviations for simulator sickness and presence questionnaires1 (n = 6) Current Sample (Aggregated across sessions)
Established Norms
M
SD
M
SD
SSQ 2 Nausea Oculomotor disturbances Disorientation
12.72 20.55 17.01*
6.75 11.83 7.81
7.70 10.60 6.4
15.00 15.00 15.00
PQ Realism Interface Quality Involvement/Control
17.52** 17.46* 67.43**
2.22 2.29 2.46
12.36 14.65 57.39
3.44 3.40 8.96
1
Statistically significant differences involve comparison with established norms, *p < .05, **p < .001. 2 Data are derived from n = 5, as one participant reported a vestibular disorder, which rendered her responses on this measure unusable. SSQ = Simulator Sickness Questionnaire; PQ = Presence Questionnaire.
45
this sample, t(4) = 3.04, p < .05, relative to established norms, although no differences were noted on the nausea and oculomotor disturbance subscales, ts(4) = 1.66 and 1.88, respectively. In contrast, this sample reported significantly higher levels of presence relative to established norms on each of the three subscales of the PQ: realism, t(5) = 5.68, p < .005; interface quality, t(5) = 3.01, p < .05; and involvement/control, t(5) = 52.55, p < .0001. Scores on the CSQ were examined to determine participants’ overall satisfaction with the treatment. All participants scored 30 or greater (M = 31, SD = 0.89), suggesting high levels of satisfaction with VRET.
adherence to exposure within vret sessions As mentioned, participants were encouraged to stay in the selected VR simulation until their anxiety decreased by 50%. In total, 80 exposure scenarios were attempted (across all patients) and 90% (n = 72) were responded to with a SUDS rating of 20 or higher, based on written progress notes following each session. Of the exposure scenarios that were attempted, reductions of 50% (or more) SUDS units were achieved in 67% (n = 48). With 33% of scenarios not meeting this criterion, it appears that a notable number of scenarios were attempted but not endured until the patient’s SUDS reduced by 50%. Examination of the association between the percentage of scenarios in which a 50% reduction in SUDS occurred and the difference between the patient’s CAPS total score at pretreatment minus posttreatment revealed an r = .33, ns. This suggests that outcomes might have been improved had the therapists been more consistent with achieving 50% reduction in anxiety within VRET.
Discussion The current report examined the efficacy of VRET for the treatment of MVA-related PTSD in an uncontrolled case series. Results from 6 patients indicated significant reductions in posttrauma symptoms involving reexperiencing, avoidance, and emotional numbing, with effect sizes ranging from d = .79 to d = 1.49. Indices of clinically significant and reliable change (Jacobson & Truax, 1991) suggested that the magnitude of these changes was meaningful and reliable. Measures of anxiety and depression did not show significant reductions before and after treatment, although moderate effect sizes were noted for the BAI and the BDI-II. High levels of presence were achieved by the virtual display, and participants reported satisfaction with treatment.
46
beck et al.
In interpreting these data, it seems important to place these results alongside other VRETs for PTSD and driving-related fears. In drawing these comparisons, factors such as the severity and chronicity of the study samples and the use of clinician versus self-report measures are salient. Rothbaum et al. (2001) reported an average 15% reduction in CAPS total score following 8 to 16 sessions of VRET for 9 veterans with chronic combat-related PTSD. Wald (2004) obtained an average 60% reduction in selfrated global phobia severity, following 8 sessions of VRET with 5 individuals diagnosed with driving phobia. In the current report, an average 40% reduction was noted on the CAPS total score following treatment, with greater percentage reductions noted for self-report measures (IES-R: 54% average reduction; PSS-SR: 49% average reduction). Thus, the 10-session VRET used in this trial obtained results that were comparable or better than results published in other uncontrolled trials of VRET for related problems. One feature of these results which deserves note is the lack of treatment effects for hyperarousal symptoms. Although relaxation training was included in the first two sessions of treatment, it would appear that patients may not have mastered this skill or may not have generalized this skill to nondriving situations. Specifically, therapists prompted the use of relaxation during VRET to facilitate habituation. Because the hyperarousal symptom cluster includes a number of nonspecific symptoms (e.g., sleep disturbance, irritability), the lack of significant pre-post treatment differences on both clinician and self-report measures of hyperarousal suggest that relaxation skills did not generalize or were not used by patients outside of driving situations. From a theoretical standpoint, one would expect that the three symptom clusters that comprise PTSD would show equal responsiveness to exposure therapy, whether conducted using virtual reality technology or not (e.g., Foa & Rothbaum, 1998). Although the process of symptom change has been studied with prolonged exposure (Nishith, Resick, & Griffin, 2002), it has not been examined with more gradual approaches to exposure, which theoretically could be expected to produce less synchrony across symptom clusters. From a conceptual standpoint, further study of this issue could deepen our understanding of the mechanisms through which exposure therapy works. One facet of this report that deserves additional comment is the visual simulation system. This software was developed specifically for this project. It was constructed for maximum flexibility, in order to permit the creation of virtual environments that matched individual feared situations (Ang, Schwag-
ler, English, Beck, & Winer, 2006). The software was designed to mimic a natural driving situation where a driver can determine her own route. As such, all aspects of the simulation (i.e., roads, traffic, signage, and buildings) are rendered in real time, based on the route that the driver takes. As noted, patients reported significantly higher levels of disorientation, relative to published norms, possibly owing to the use of stereoscopic glasses. Further refinement of the simulation system should consider adaptation to a head-mounted display, as simulation sickness appears less likely with this type of application. Efforts are underway to adapt the VRET software to a desktop computing system, to facilitate general use within clinical and research settings. In considering these findings, it is relevant to discuss the limitations of this report. A clear limit to the generalizability of these results rests with the fact that it involved an open clinical trial research design. Future studies of VRET for MVA-related PTSD should utilize a controlled design with randomization in order to account for numerous possible confounds. Second, because the patient sample in the current report involved individuals who had experienced a traumatic MVA less than 7 months prior to entry into the trial, it is possible that natural symptom remission could have influenced these results. While possible, it is unlikely that remission fully accounted for these effects as half of the patients in this report had experienced symptoms for at least 4 months. As reviewed by Blanchard and Hickling (2004), the rate of natural remission appears to stabilize approximately 4 months after a traumatic MVA. Ideally, future research could examine the influence of chronicity of PTSD on the efficacy of VRET and the role of natural remission in this process. Third, in a notable percentage of exposure trials, the trial was ended before the patient’s SUDS reduced by 50%. Given the association between the number of exposure trials where 50% anxiety reduction occurred and CAPS scores at posttreatment, this clearly influenced the efficacy of VRET in this report. This suggests that participants who exhibit greater habituation of fear within each session will realize better treatment outcomes. To ensure that this occurs, the use of therapists who have more experience with exposure-based therapy may be warranted. This observation also highlights the key role that virtual reality technology can play in parametric studies designed to examine the salient parameters of exposure during ET. Finally, this report is limited by the lack of follow-up data. Given the positive outcomes that were noted at posttreatment, follow-up data would be an
virtual reality exposure therapy for ptsd important methodological addition to examine the durability of change from VRET. In sum, preliminary data from this uncontrolled case series suggest that VRET can be used for the successful treatment of PTSD symptoms among survivors of traumatic MVAs. Observed effect sizes were comparable to trials involving the virtual treatment of related problems, with high levels of patient satisfaction. Future controlled studies are needed to provide stronger experimental support for VRET for this patient population. References American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders, (4th ed., text revision). Washington, DC: Author. Ang, E. J., Schwagler, B., English, K., Beck, J. G., & Winer, E. (2006). Development of an automated tile-based scene generator for driving simulation in virtual reality exposure therapy. Manuscript submitted for publication. 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. Beck, A. T., Steer, R., & Brown, G. (1996). Beck Depression Inventory manual, (2nd ed.). San Antonio, TX: Psychological Corporation. Beck, J. G., & Coffey, S. F. (2005). Group cognitive behavioral treatment of PTSD: Treatment of motor vehicle accident survivors. Cognitive and Behavioral Practice, 12, 267–277. Beck, J. G., Coffey, S. F., Palyo, S. A., Gudmundsdottir, B., Miller, L. M., & Colder, C. R. (2004). Psychometric properties of the Posttraumatic Cognitions Inventory (PTCI): A replication with motor vehicle accident survivors. Psychological Assessment, 16, 289–298. Beck, J. G., Grant, D., Read, J. P., Clapp, J., Coffey, S., Miller, L., & Palyo, S. A. (2006). The Impact of Event Scale–Revised: Psychometric properties in a sample of motor vehicle accident survivors. Manuscript in preparation. Bernstein, D. A., & Borkovec, T. D. (1973). Progressive relaxation training: A manual for helping professions. Champaign, IL: Research Press. Blake, D., Weathers, F., Nagy, L., Kaloupek, . D., Klauminzer, G., Charney, D., & Keane, T (1990). Clinician-administered PTSD scale (CAPS). Boston: National Center for PostTraumatic Stress Disorder, Behavioral Science Division. Blanchard, E. B., & Hickling, E. J. (2004). After the crash: Assessment and treatment of motor vehicle accident survivors, (2nd ed.). Washington, DC: American Psychological Association. Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., Forneris, C. A., & Jaccard, J. (1996). Who develops PTSD from motor vehicle accidents? Behaviour Research and Therapy, 34, 1–10. Blanchard, E. B., Hickling, E. J., Taylor, A. E., Loos, W. R., & Gerardi, R. J. (1994). Psychological morbidity associated with motor vehicle accidents. Behaviour Research and Therapy, 32, 283–290. Botella, C., Banos, R., Perpina, C., Villa, H., Alcaniz, M., & Rey, A. (1998). Virtual reality treatment of claustrophobia: A case report. Behaviour Research and Therapy, 36, 239–246. Creamer, M., Bell, R., & Failla, S. (2003). Psychometric
47
properties of the Impact of Event Scale–Revised. Behaviour Research and Therapy, 41, 1489–1496. Creamer, M., Foran, J., & Bell, R. (1995). The Beck Anxiety Inventory in a non-clinical sample. Behaviour Research and Therapy, 33, 477–485. Emmelkamp, P., Krijn, M., Hulsbosch, A., deVries, S., Schuemie, M., & van der Mast, C. (2002). Virtual reality treatment versus exposure in vivo: A comparative evaluation in acrophobia. Behaviour Research and Therapy, 40, 509–516. Foa, E. B., & Kozak, M. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Foa, E. B., & Meadows, E. (1997). Psychosocial treatments for post-traumatic stress disorder: A critical review. In J. Spence, J. Darley, & D. Foss (Eds.), Annual review of psychology, (vol. 48). pp. 449–480. Palo Alto: Annual Reviews. Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. O. (1993). Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress, 6, 459–473. Foa, E. B., & Rothbaum, B. O. (1998). Treating the trauma of rape: Cognitive behavioral therapy for PTSD. New York: The Guilford Press. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Minimental state”: A practical method of grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. Imago Systems, Inc. (1996). DriVR user guide. Vancouver, BC: Author. International Society of Presence Research. (n.d.). The concept of presence: Explication statement. Retrieved August 3, 2005, from http://ispr.info/ Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19. Kennedy, R. S., Lane, N. E., Berhaum, K. S., & Lilienthal, M. G. (1993). A simulator sickness questionnaire (SSQ): A new method for quantifying simulator sickness. International Journal of Aviation Psychology, 3, 203–220. Krijn, M., Emmelkamp, P. M. G., Olafsson, R. P., & Biemond, R. (2004). Virtual reality exposure therapy for anxiety disorders: A review. Clinical Psychology Review, 24, 259–281. Kuch, K., Swinson, R. P., & Kirby, M. (1985). Post-traumatic stress disorder after car accidents. Canadian Journal of Psychiatry, 30, 426–427. Larsen, D. L., Attkisson, C., Hargreaves, W. A., & Nguyen, T. D. (1979). Assessment of client/patient satisfaction: Development of a general scale. Evaluation and Program Planning, 2, 197–207. Maltby, N., Kirsch, I., Mayers, M., & Allen, G. (2002). Virtual reality exposure therapy for the treatment of fear of flying: A controlled investigation. Journal of Consulting and Clinical Psychology, 70, 1112–1118. Nishith, P., Resick, P. A., & Griffin, M. G. (2002). Pattern of change in prolonged exposure and cognitive-processing therapy for female rape victims with posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 70, 880–886. Rothbaum, B., Hodges, L., Ready, D., Graap, K., & Alarcon, R. (2001). Virtual reality exposure therapy for Vietnam veterans with posttraumatic stress disorder. Journal of Clinical Psychiatry, 62, 617–622. Rothbaum, B., Meadows, E., Resick, P., & Foy, D. (2000). Cognitive-behavioral therapy. In E. Foa, T. Keane, & M. Friedman (Eds.), Effective treatments for PTSD (pp. 60–83). New York: The Guilford Press.
48
beck et al.
Sprinkle, S. D., Lurie, D., Insko, S. L., Atkinson, G., Jones, G. L., Logan, A. R., & Bissada, N. N. (2002). Criterion validity, severity cut scores, and test-retest reliability of the Beck Depression Inventory-II in a university counseling center sample. Journal of Counseling Psychology, 49, 381–385. Taylor, S., & Koch, W. J. (1995). Anxiety disorders due to motor vehicle accidents: Nature and treatment. Clinical Psychology Review, 15, 721–738. Wald, J. (2004). Efficacy of virtual reality exposure therapy for driving phobia: A multiple baseline across-subjects design. Behavior Therapy, 35, 621–635. Weathers, F. W., Keane, T., & Davidson, J. (2001). ClinicianAdministered PTSD Scale: A review of the first ten years of research. Depression and Anxiety, 13, 132–156. Weiss, D. S., & Marmar, C. R. (1997). The Impact of Event Scale-Revised. In J. P. Wilson, & T. M. Keane (Eds.),
Assessing psychological trauma and PTSD (pp. 399–411). New York: The Guilford Press. Whisman, M., Perez, J. E., & Ramel, W. (2000). Factor structure of the Beck Depression Inventory-Second Edition (BDI-II) in a student sample. Journal of Clinical Psychology, 56, 545–551. Witmer, B. G., & Singer, M. J. (1998). Measuring presence in virtual environments: A presence questionnaire. Presence, 7, 225–240.
R E C E I V E D : October 13, 2005 A C C E P T E D : April 19, 2006 Available online 22 September 2006