The relevance of analogue studies for understanding obsessions and compulsions
Introduction
Obsessive compulsive disorder (OCD) is a complex clinical condition affecting 2–3% of the population (Kessler et al., 2005). It is characterized by (a) intrusive and unwanted thoughts or images or urges (obsessions) and/or (b) repetitive, intentional rituals to neutralize obsessional distress (compulsions) (American Psychiatric Association [APA], 2013). The disorder is associated with considerable suffering, functional impairment, and economic burden to both the individual and the health-care system (Markarian et al., 2010). Obsessive–compulsive (OC) symptoms are phenomenologically heterogeneous (i.e., they take many different forms) and etiologically complex (Taylor, 2011).
Investigators have used many different types of research designs to study OC phenomena. A common method is to study clinical samples of convenience, namely individuals with OCD who are attending a clinic in which research is being conducted. This type of design can be useful, especially for treatment studies, but has limitations. For example, studies of clinic patients are based on the untested assumption that the results generalize to OCD in the population at large. Those seeking treatment for OCD represent a minority of the OCD population (Grabe et al., 2000), and likely differ from non-help seekers on social, economic, attitudinal, and personality factors. Confounding factors such as treatment types and treatment effects, duration of prior treatment, and comorbidity also pose challenges for studies of OC phenomena in clinical populations. This is one reason researchers have pursued various forms of analogue research in order to study OC phenomena. The two most commonly used analogue designs include (a) studies of animals, typically rodents, in which particular behaviors (e.g., excessive grooming or the burying of objects) are used as analogues of compulsions, and (b) studies of human non-clinical samples, such as college students, in which subclinical OC phenomena are regarded as analogues of OC symptoms observed in people diagnosed with OCD.
Both animal and human analogue designs have their strengths and limitations. In recent years, however, it has become increasingly difficult to publish human analogue research, particularly studies based on correlational, factor analytic, or structural equation designs. Indeed, an increasing number of journals indicate in their aims and scope that studies using analogue and non-clinical samples will be given low priority (or in some instances not considered for publication). Perhaps this is because of the often unquestioned assumption that such studies are less relevant than studies of clinical patients to understanding the psychopathology or treatment of OCD (indeed, reviewers of journal submissions often ask authors to provide a justification for using analogue samples, as well as to cite their use as a limitation of the study). In comparison, animal analogue research (often involving rodents) has been rarely criticized in the literature despite the significant limitations of this type of design. We have discussed the limitations of such animal studies of OC phenomena elsewhere (Abramowitz, Taylor, McKay, & Deacon, 2011). To summarize, the major problems are: (a) it is difficult to determine whether a repetitive behavior in animals such as rodents is a bona fide compulsion (as defined in DSM) or some other form of repetitive behavior1; (b) compulsions in humans often arise as a consequence of obsessions—it is unclear whether rodents experience intrusive obsessional thoughts of any kind; and (c) there is no evidence that rodents possess the cognitive capacity (or the frontal lobe development, which is an integral part of neuroanatomical models of OCD) to experience common obsessions, such as those pertaining to taboo acts concerning aggression, sex, or morality.
The motivation for the present article arose from our interest in clarifying the utility of human analogue research in OCD. Since OCD occurs in only 2–3% of the population, it can be time intensive and costly to recruit clinical samples of an adequate size. OC symptoms, however, occur in the general population (Adam et al., 2012, de Bruijn et al., 2010, Grabe et al., 2000, Rachman and de Silva, 1978), allowing researchers to recruit larger samples with relative convenience. It is likely that using human analogue samples thus allows for more research to be conducted and may make some projects feasible that would otherwise be impractical. For example, analogue samples provide an opportunity to examine subgroups of obsessions and compulsions (e.g., scrupulosity, checking)—which has become an emphasis in recent years with the conceptualization of OCD as a dimensional condition (e.g., Abramowitz, McKay, & Taylor, 2008). Two decades ago, Gibbs (1996) argued for the relevance of human analogue research for understanding OC phenomena. Since that time there have been many important research developments that further support the value of human analogue research. The purpose of this article is to review the evidence regarding the value of human analogue samples. We also consider the use of non-clinical samples in research on the treatment of OCD. As a shorthand, in the following text we will refer to studies of students or community samples as “analogue samples,” maintaining a focus on human rather than animal analogues.
Section snippets
Assumptions underlying human OC analogue research
Analogue designs are based on the notion that the sample or research design is relevant to a particular target disorder. If studies of student or community samples are to be relevant for understanding OCD, then the following assumptions would optimally be met:
- (a)
It should be demonstrated that OC symptoms in student or community samples are sufficiently prevalent so that correlational studies of OC symptoms in students would be relevant to correlational studies in patients diagnosed with OCD. If OC
Prevalence of OC symptoms in non-clinical samples
Studies of analogue samples (i.e., student and community participants) highlight the prevalence of subclinical OC symptoms. According to surveys, up to 90% of people report that they at least occasionally experience intrusive thoughts that are similar in form and content to clinical obsessions (Clark, 1992, Clark and de Silva, 1985, Freeston et al., 1991, Freeston et al., 1992, Purdon and Clark, 1993, Purdon and Clark, 1994, Rachman and de Silva, 1978, Salkovskis and Harrison, 1984). To further
Categories versus dimensions
The question of whether a given psychopathic phenomenon is dimensional or categorical has important implications for the way in which the phenomenon is studied. If OCD is categorical it would mean that people either do or do not have the disorder, with no intermediate. In such a case there is little point in studying non-clinical samples because most people will not have the disorder. But if OCD is dimensional, ranging from no symptoms through to subclinical OC symptoms and finally to symptoms
Presence of obsessions and compulsions
The vast majority of individuals diagnosed with OCD experience both obsessions and compulsions (Foa et al., 1995, Rasmussen and Tsuang, 1986). Two early studies, however, found that non-clinical participants tended to report either obsessions or compulsions (Karno et al., 1988, Weissman et al., 1994). One explanation for this discrepancy is that the functional relationship between obsessions (which trigger anxiety) and compulsions (which aim to reduce anxiety) may not exist in non-clinical
Etiology: behavioral and genetic factors
The use of non-clinical analogue samples in OC research is based on the assumption that non-clinical and clinical samples are comparable with regard to certain characteristics. One important characteristic is etiology. Twin studies can provide evidence of whether such an etiological overlap exists between clinical and non-clinical samples because twin studies use biometric structural equation modeling to determine, among other things, the proportion of variance in psychological symptoms (or
Development and maintenance processes
Also central to the question of whether the findings from non-clinical analogue samples are useful for understanding the nature of OC symptoms as they occur in clinical samples is the degree to which the psychological processes and mechanisms hypothesized to govern the development and maintenance of OC symptoms overlap across types of samples. Cognitive–behavioral models are the most well-articulated and well-researched psychological models of OC symptoms. These models posit that obsessions
Use of analogue samples in OCD treatment research
The field of cognitive–behavioral therapy (CBT) has a rich history of conceptualizing and developing treatment methods using analogue samples. In fact, the very procedures that would come to form the heart of CBT for OCD – exposure and response prevention – were initially studied with dogs that had been conditioned to fear the sudden illumination of their shuttle box which signaled the onset of electric shock (Solomon, Kamin, & Wynne, 1953). The dogs learned to avoid the shock by jumping
Guidelines for using analogue designs and samples
On the basis of the findings reviewed in this article, we provide recommendations and criteria for the use of analogue designs and samples that allow for understanding the psychological processes involved in OCD, as well as for generalization to clinically severe populations. The suitability of analogue designs and samples depends on the nature of the investigator's research question, and on the associated assumptions underlying the use of analogue samples.
Summary and conclusions
Researchers use analogue samples to study OC-related phenomena based on the assumption that the results derived from such samples are relevant and generalizable to understanding these phenomena in individuals with a diagnosis of OCD. The present review synthesizes research pertaining to five specific assumptions underlying the use of analogue samples: that OC symptoms (a) are prevalent in non-clinical populations, (b) are dimensional as opposed to categorical, (c) have similar etiologies in
References (166)
- et al.
Psychometric properties and construct validity of the Obsessive-Compulsive Inventory—Revised: Replication and extension with a clinical sample
Journal of Anxiety Disorders
(2006) - et al.
The role of cognitive factors in the pathogenesis of obsessive–compulsive symptoms: A prospective study
Behaviour Research and Therapy
(2006) - et al.
Obsessive–compulsive disorder
Lancet
(2009) - et al.
Animal models of obsessive–compulsive disorder
Biological Psychiatry
(2011) - et al.
Is thought-action fusion specific to obsessive–compulsive disorder: A mediating role of negative affect
Behaviour Research and Therapy
(2003) - et al.
Attenuation of attention bias in obsessive–compulsive disorder
Behaviour Research and Therapy
(2009) - et al.
The relationship between religion and thought–action fusion: Use of an in vivo paradigm
Behaviour Research and Therapy
(2010) - et al.
Deteriorating memory confidence, responsibility perceptions and repeated checking: Comparisons in OCD and control samples
Behaviour Research and Therapy
(2007) - et al.
Obsessions and worry beliefs in an inpatient OCD population
Journal of Anxiety Disorders
(2010) - et al.
What do low-dysfunctional beliefs obsessive–compulsive disorder subgroups believe?
Journal of Anxiety Disorders
(2010)