Two hundred and seventy eight OCD patients (n = 278: 148 male; 130 female), and 54 TTM patients (n = 54; 5 male; 49 female), ranging in age between 8 and 75 years, took part in the study (Table 1). These patients were referred to our research unit from a wide range of sources (including the OCD Association of South Africa, community based primary care practitioners, and psychiatrists). Either a clinical psychologist or a psychiatrist with expertise in the field interviewed participants. Participants met the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria  for either a primary diagnosis of OCD or TTM on the Structured Clinical Interview for Axis I Disorders (SCID-I) . Patients were included irrespective of whether they were at baseline (i.e. not receiving any form of treatment for their primary psychiatric disorder), or were receiving treatment for OCD / TTM, but those with comorbid OCD and TTM (N = 25) were excluded from subsequent analysis. A history of psychosis was also an exclusion criterion. Referring clinicians were contacted to establish, where possible, a longitudinal expert evaluation of the diagnostic status of the patient. All subjects gave informed written consent to participate after confidentiality was guaranteed and risks and benefits had been fully explained. The study was approved by the Institutional Review Board of the University of Stellenbosch.
Specific demographic data, including age when interviewed, age of onset of OCD/TTM, highest level of education, current employment status, and population group were obtained from all participants. In addition to the SCID-I, and selected parts of the SCID-II (obsessive-compulsive, avoidant, schizotypal, borderline personality disorders) for adult patients (aged 18 or older) , the interview also included the Structured Clinical Interview for Obsessive-Compulsive Spectrum Disorders (SCID-OCSD) to determine the presence of other obsessive-compulsive related conditions .
The Yale-Brown Obsessive-Compulsive Severity Scale (Y-BOCS)  was implemented to assess the severity of OCD symptoms. Severity of hair-pulling symptoms was assessed with the Massachusetts General Hospital Hair-pulling Scale . The Trichotillomania Behaviour Profile (TBP, available from the first author on request) was administered to TTM patients to assess hair-pulling phenomenology.
Patients' level of insight into the senselessness or excessiveness of their symptoms was assessed on the relevant YBOCS item. When an adequate trial of pharmacotherapy with an SRI (i.e. for both OCD and TTM groups, at least 10 weeks on the medication with a minimum of 6 weeks on mid-range dose) had been undertaken, response to pharmacotherapy was assessed using the global improvement item of the Clinical Global Impression (CGI) scale; subjects with CGI scores of 1 ('very much improved') or 2 ('much improved') were defined as responders . Similarly, when patients received an adequate trial of cognitive behavioural therapy (CBT) (i.e. for both OCD and TTM groups, 8 or more sessions with an expert CBT psychotherapist), response to treatment was rated using the CGI. The Disability Profile questionnaire (DP)  was included in the interview to assess current (i.e. past two weeks) and lifetime impairment in eight domains. The DP was initially developed for use in patients with social anxiety disorder; nevertheless, the scale has since been used to assess disability in patients with other anxiety disorders as well .
Questions addressing potential precipitating or exacerbating factors, including the impact of menstrual/reproductive cycle changes, brain trauma and history of autoimmune infections on OCD/TTM symptom fluctuations, were included in the interview.
Severity of comorbid depression was evaluated with the Beck Depression Inventory (BDI) . The Childhood Trauma Questionnaire (CTQ) , a scale proven to be a valid and reliable measure of past traumatic experiences , was used as a self-report questionnaire to assess the nature and severity of childhood trauma. Sub-scales of the CTQ include measures of emotional abuse, physical abuse, sexual abuse, emotional neglect and physical neglect.
The self-report Temperament and Character Inventory (TCI)  was also used to measure behaviours associated with seven personality dimensions, namely novelty seeking, harm avoidance, reward dependence, persistence, self-directedness, cooperativeness, and self-transcendence. In addition, participants completed the self-report Young Schema Questionnaire (YSQ)  to assess the current profile of fundamental maladaptive beliefs (cognitive schemas) in OCD and TTM. For each item of the 75-item "short form" of the YSQ (which includes 15 schemas), the answer is required to be placed on a 6-point Likert-type scale (1= 'completely untrue of me', 2 = 'mostly untrue of me', 3 = 'slightly more true than untrue', 4 = 'moderately true of me', 5 = 'mostly true of me', 6 = 'describes me perfectly').
As there were few males with TTM (Table 2), only clinical data from females with OCD and TTM were analyzed. Chi-square and t-tests were performed to investigate the differences in OCD/TTM phenomenology where appropriate. A one-way analysis of variance (ANOVA) was done to investigate the effects of the primary and comorbid disorders on disability. Subsequently, a two-way fixed effects ANOVA was used to assess the main interactions between primary diagnosis and comorbidity on disability and to test for the main (fixed) effects of the different diagnoses on disability. Residuals of ANOVA's of cognitive schema data in OCD and TTM groups suggested non-normality of the data. As a result, pair-wise comparison tests (Mann-Whitney U) were implemented to compare the two groups on cognitive schemas.