Obsessive-Compulsive Disorder (OCD) is a disabling anxiety disorder characterized by upsetting, unwanted cognitions (obsessions) and intense and time consuming recurrent compulsions (American Psychiatric Association, 2000). The idiosyncratic nature of the symptoms of OCD (Whittal et al., 2010) represents a challenge to completing accurate and comprehensive assessments, which if not achieved, can have a deleterious effect on the provision of effective treatment for the disorder (Kim et al., 1989; Taylor, 1995; Steketee and Barlow, 2002; Deacon and Abramowitz, 2005).

Accurately assessing the full range of symptoms of OCD requires reliable and psychometrically sound diagnostic instruments and measures (Taylor, 1995, 1998; Rees, 2009) alongside the standard clinical interview. Although the most commonly used psychometric instrument for assessing OCD, the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) (Goodman et al., 1989a,b), has acceptable reliability and convergent validity, it has been criticized by Taylor (1995) for weak discriminant validity. Taylor also highlighted that it remains susceptible to administration variance, relies on client memory recall, and is time consuming to administer. As with all measures completed retrospectively, selective memory biases affect the type of information reported by clients about their symptoms (Clark, 1988; Stone and Shiffman, 2002; Stone et al., 2004). Glass and Arnkoff (1997, p. 912) have summarized several disadvantages of structured inventories; first, they contain prototypical statements which may fail to capture the idiosyncratic nature of the client's actual thoughts; second, they can be affected by post-hoc reappraisals of what clients feel, as the data is subject to memory recall biases; and finally they may fail to adequately capture the client's internal dialog due to the limitations of the best fit question structure.

Discrepancies have been reported between data collected in the client's natural environment (in situ) and those based on the client's later recall (de Beurs et al., 1992; Marks and Hemsley, 1999; Stone et al., 2004). Such discrepancies may be further affected by factors such as the complexity and diversity of obsessions and compulsions, not to mention the ego-dystonic nature of many OCD clients' obsessional thoughts. It seems likely that clients with distressing ego-dystonic obsessions, for example, those involving sexual, aggressive, and/or religious themes may experience a heightened level of discomfort in reporting their obsessions in a face to face assessment with a clinician, thus reducing their willingness to accurately report (Taylor, 1995; Newth and Rachman, 2001; Grant et al., 2006; Rees, 2009). This may contribute to an underreporting of these obsessions, and hence an inaccurate understanding and a restriction of the clinician's ability to adequately treat the client (Grant et al., 2006; Rachman, 2007).

Exposure and response prevention, cognitive therapy, and pharmacological interventions have been shown to be effective in the treatment of OCD (Abramowitz, 1997, 2001; Foa and Franklin, 2001; Steketee and Barlow, 2002; Fisher and Wells, 2008; Chosak et al., 2009). Self-monitoring is a useful therapeutic technique that provides essential information to assist in the development of exposure hierarchies and behavioral experiments used in cognitive therapy (Tolin, 2009). Clients typically observe and record their experiences of target behaviors, including triggers, environmental events surrounding those experiences, and their response to those experiences (Cormier and Nurius, 2003). Such self-monitoring can be used to both assist assessment and/or as an intervention. Cormier and Nurius (2003) explained that the mere act of observing and monitoring one's own behavior and experiences can produce change. As people observe themselves and collect data about what they observe, their behavior may be influenced.

A form of self-monitoring and alternative to the typical clinic-based assessment of OCD is the use of sampling from the client's real-world experiences, a procedure known as Ecological Momentary Assessment (EMA) (Schwartz and Stone, 1998; Stone and Shiffman, 1994, 2002). EMA does not rely on measurements using memory recall within the clinical setting, but rather allows for collection of information about the client's experiences in their natural setting, potentially improving the assessment's ecological validity (Stone and Shiffman, 2002). In situ sampling techniques have been successfully used in psychology, psychiatry, and occupational therapy (for a more detailed account see research by Morgan et al., 1990; de Beurs et al., 1992; Kamarack et al., 1998; Litt et al., 1998; Kimhy et al., 2006; Gloster et al., 2008; Putnam and McSweeney, 2008; Trull et al., 2008). Generally it is agreed that EMA offers broader assessment within the client's natural environment, as it includes random time sampling of the client's experience, recording of events associated with the client's experience, and self-reports regarding the client's behaviors and physiological experiences (Stone and Shiffman, 2002). Because this assessment method accesses information about the client's situation, the difficulties of memory distortions like recall bias are reduced (Schwartz and Stone, 1998; Stone and Shiffman, 2002).

Given that accurate assessment of obsessions and compulsions is a critical aspect of treatment planning and that reliance on self-report and clinician interview has some known limitations, the purpose of this study was to investigate the utility of EMA as a potential adjunct to the conventional assessment of OCD. Specifically, we sought to compare the amount and type of information regarding obsessions and compulsions collected via EMA vs. standard assessment using the gold-standard symptom interview for OCD. As this is a pilot clinical case study, we offer the following tentative hypothesis: (1) EMA will yield additional types of obsessions and compulsions not captured by the Y-BOCS.

All screening of participants, interviewing and assessment, as well as administration of the study, was conducted by the first author, who was a provisionally registered psychologist undergoing postgraduate training at the time of the research, and was supervised by the second author, an experienced OCD clinician and academic. Potential participants were recruited from the Curtin OCD clinic. They were screened via telephone to ascertain their suitability for the study. A face-to-face assessment session using the Structured Clinical Interview for DSM-IV (SCID-IV; First et al., 1997) was conducted to determine a primary OCD diagnosis, followed by the administration of the Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) interview and checklist (Goodman et al., 1989b) to identify the participants' obsessions and compulsions and their symptom severity. The Y-BOCS is the most widely used scale for OCD symptoms assessment and is considered by researchers to be the “gold standard” measure for symptom severity (Deacon and Abramowitz, 2005; Himle and Franklin, 2009). It consists of two parts; a checklist of prelisted types of obsessions, usually endorsed by the clinician based on disclosures made by the client; and the severity scale which requires the client to rate the severity of their experience by answering the questions based on their recall. Goodman et al. (1989) note that the Y-BOCS has shown adequate interrater agreement, internal consistency, and validity.

Suitable participants then attended a second session where they signed consent forms and were given instructions about the study procedure. During the data collection using the Ecological Momentary Assessment data (EMA-OCD), participants used an Olympus WS-110 digital voice recorder to record their experiences throughout a 12 h period. Participants used their existing mobile phones to receive prompts via the mobile phone Short Message Service (SMS) to record their responses to the research questions. All three participants were then provided with an envelope containing the Olympus WS-100 digital voice recorder, a spare battery, and the participant prompt questions (see Table 1).

Participants were asked to turn their mobile phones on during the data collection day by 10 am, ready to receive their SMS prompts. The researcher manually sent SMS prompts to the participants at random intervals; at least every 2 h (across 1 day, from 10 am to 10 pm), for a minimum of 10 data entries in keeping with research using EMA procedures (see, Stone and Shiffman, 2002); asking them to complete their responses to all four questions as details on the EMA-OCD Participant Questions Sheet. Participants were instructed not to respond to the SMS prompts if driving, and were asked to respond as quickly as possible to the prompts. Data was then downloaded from the voice recorder to the researcher's computer, and transcribed. During this process all identifying details were removed. During the debrief session open-ended questions were used to gather as much information as possible regarding the participant's experiences of the study, and suggestions for improvements. During the data collection day the researcher completed a journal to record his observations and reflections related to the use of the EMA. At the completion of the EMD-OCD data collection, each of the participants was provided with a debrief session (Mary by phone, and John and Paul, face to face). The debrief session focused on their experiences of the research and use of the digital voice recorder; and provided the opportunity for them to discuss anything else that arose they wished to tell the researcher. As stated above, the data was downloaded and transcribed by the first author. The Y-BOCS obsession and compulsion categories were used as a framework to compare the data generated from the EMA-OCD procedure. After the complete de-identified data set was tabled, it was provided to a second person who was an expert in OCD for verification of categories. In the case of any discrepancies agreement was reached via consensus.

This study investigated the utility of EMA as an adjunct assessment approach for OCD. Each of our study hypotheses was supported. As predicted the EMA procedure resulted in the identification of additional types of obsessions and compulsions not captured by the Y-BOCS interview. The finding that the EMA procedure identifies obsession and compulsion symptoms not captured by the Y-BOCS suggests that further studies in this area are warranted. As a pilot case study we cannot generalize from these initial findings but our results indicate that a larger study replicating the procedure used here, is justified. Importantly, the three participants in our study were representative of quite typical OCD clients in that they had severe levels of symptoms and had OCD for a number of years. The EMA procedure we used was found to be satisfactory to all three participants. Feedback from the participants at the de-briefing session included suggestions that this process would be helpful for therapy because it would provide the therapist and client with rich and current material regarding their symptom patterns. From a clinician's point of view, collecting the EMA data is not onerous because the entries are simply short answers collected on 12 occasions and thus is not a time-consuming exercise.

The EMA procedure as used in this study could provide clinicians with a new method by which to gain a current and accurate snap-shot of clients symptoms as they occur in real-time. This information could augment information gained from standard pencil and paper measures but also provide an “active” process which may help to engage clients in the therapeutic process. It seems likely that using a procedure like EMA with OCD clients will assist in understanding their OCD experiences, and thus assist in generating valuable information, supporting accurate assessment, client conceptualization, and ultimately treatment.

Despite these valuable findings, there are limitations of this study. As a pilot study and exploratory in nature, it is only possible to draw limited interpretations from the data provided. However, the preliminary findings of this study support the benefit of conducting further research into this procedure, where it may be possible to draw more empirically valid findings from a larger and more statistically powerful sample. Second, due to the lack of availability of date stamping, participants were asked to record the time they made each recording. Unfortunately this was not routinely provided by all participants, and hence creates an unanswerable question regarding the accuracy of the data recorded. As Stone and Shiffman (2002) discuss, a potential problem relates to participants recording their data based on their recall of what was occurring at the time of the SMS prompt, rather than immediately. Hence introducing possible memory bias, and undermining the premise of the study. Although this is certainly an unwanted variable, based on the EMA-OCD data provided it seems that except for Mary, both John and Paul responded promptly to the SMS messages, or recorded the time if they didn't. Mary on the other hand, reported during the debrief session that she was unable to record the time for the initial targets, but did so for subsequent SMS prompts. It was not possible to ascertain from her data the delay in time between the first SMS prompts and her recordings. In future applications of this procedure, it is recommended that the device used provides automatic date-stamping to address this limitation. Indeed, it may be possible to adapt the EMA methodology for use with smart phones via a dedicated OCD application.

The findings from this study of three patients with severe OCD suggest that the use of EMA provides important additional information regarding obsessions and compulsions and may thus be a useful adjunct to the clinical assessment of OCD.