The impact of poor insight on the two year natural course of Obsessive-Compulsive Disorder

Background: Some patients with Obsessive Compulsive Disorder (OCD) think or are convinced that their obsessive-compulsive beliefs are true. These patients have OCD with poor or absent insight, a clinical profile that poses a challenge to the clinician. The purpose of this study is to determine the range of insight, characterize the clinical profile of poor insight OCD and study the impact of poor insight on the natural two year course of OCD. Methods: Data were analyzed of 253 adult patients with OCD, participating in the prospective naturalistic Netherlands Obsessive Compulsive Disorder Association (NOCDA) Study. Insight was measured using a standardized instrument, the Overvalued Ideas Scale. Results: Good, fair, poor and absent insight occurred at every severity level of OCD. A small but significant correlation between changes in OCD severity and changes in insight was found. Poor insight was associated with higher OCD symptom severity, more chronicity, more comorbidity and predicted poor outcome at two-year followup, independently of severity of OCDand depressive symptoms, age of onset, comorbidity and chronicity of OCD. Conclusions: These findings show that appropriate changes are introduced in DSM-5 diagnostic criteria for OCD; insight in OCD ranges from excellent to absent. Poor insight in OCD appears to be an independent phenomenon which is critically important to the prognosis of OCD. Therefore measuring insight and specifically targeting improvement of insight might be recommendable in OCD treatment. Future work should focus on validating treatment that specifically targets insight. INTRODUCTION A substantial proportion of the patients with Obsessive Compulsive Disorder (OCD) has poor insight1-11. These patients think or are convinced that their obsessive compulsive beliefs are true12. Patients with OCD and poor insight have a worse clinical condition than patients with good insight in terms of more severe symptoms2-11,13, lower quality of life10, more chronicity13,14 and suicidal ideation15. Furthermore among these patients a worse response to psychotropic medication3,7,16 and cognitive behaviour therapy (CBT)1,10,17-19 was found. However, the clinical profile of patients with OCD and poor insight has not been established firmly and knowledge of the impact of insight on the natural course of OCD is marginal. Furthermore, examinations so far leave unanswered whether poor insight is merely a proxy for severe OCD, or a distinct phenomenon. More knowledge about insight in OCD might help to comprehend the relevance of insight for OCD and its prognosis, and the possible necessity to specifically target insight in OCD treatment. The goals of this study were to (i) establish the range of insight in a clinical sample of patients with OCD, (ii) characterize the clinical profile of OCD with poor insight, (iii) study the natural course of insight and its correlation with the natural course of OCD severity and (iv) study the impact of insight on the two-year course of OCD. MATERIALS AND METHODS Design and Participants Data were drawn from the Netherlands Obsessive Compulsive Disorder Association (NOCDA) study. The NOCDA study is an ongoing multi-center 6-year longitudinal naturalistic cohort study which examines the course of OCD. The participants were patients with a life-time diagnosis of OCD, aged 18 years and over and referred to one of the participating second and third line mental health care centers. At baseline a total of 419 participants were included in the NOCDA study. No formal exclusion criteria were applied except for an inadequate understanding of the Dutch language. The study was approved by the local ethical committee, and all participants gave written informed consent. Detailed sample characteristics and methodology of NOCDA are described elsewhere20. An insight measure was administered at first at wave 3 (two years after baseline). In the present study we included all patients with OCD complaints at wave 3 of NOCDA Chapter 4 62 The impact of poor insight on the natural two year course of OCD 63 (N=253) (whereby insight in OCD could be measured) and analyzed their data at wave 3 and wave 4 (two years later). Data were collected between 2007and 2012. Assessments Assessments at wave 3 Insight into OCD symptoms was measured with the Overvalued Ideas Scale (OVIS). The OVIS21 is a 10 item clinician administered scale that assesses the severity of OCDrelated over-valued ideation. The concept of overvalued ideas can be considered equivalent to poor or absent insight as described in DSM-521. The scale comprises 10 items that investigate features of the main OCD related belief that the patient has had in the last week. The 10 items measure: strength of belief; reasonableness of belief; the extent to which others share the same belief; effectiveness of compulsions; attribution of different views by others; strength of the resistance; the extent to which the patients OCD has caused the belief; fluctuation and duration of belief. The score of each item ranges from 0 to 10 and the OVIS total score is the mean score of the 10 items, where a high score represents poor insight. Internal consistency (α=0.88-0.95) and inter-rater reliability (r=0.86) are adequate. Patients with an OVIS score <4 are considered to have good insight, an OVIS score >=4 and <6 reflects fair insight, an OVIS score >= 6 and < 7.5 OR OVIS score >= 6 and item 5 (how accurate is the belief)<9 reflects poor insight and an OVIS score >=7.5 and item 5>=9 reflects absent insight. OCD symptom severity was assessed using the Yale-Brown Obsessive Compulsive Scale (Y-BOCS)-severity scale22. The YBOCS-severity scale is a 10-item measure of current severity of obsessions and compulsions with total scores ranging from 0 to 40. Higher scores indicate greater severity. This reliable and valid scale is widely accepted as the major outcome measure for OCD22. Chronicity of OCD was defined as “continuous presence of at least moderately severe OCD symptoms during at least two years”, as proposed by Visser et al.,23 and was assessed retrospectively using the Life-Chart Interview (LCI)24. To establish OCD and other DSM-IV-TR axis I disorders, the Structured Clinical Interview for DSM-IV-TR (SCID-I/ P25) was administered. Severity of depressive and anxiety symptoms were assessed with the Beck Depression Inventory26and the Beck Anxiety Inventory27. Severity of symptoms of Attention Deficit Hyperactivity Disorder (ADHD) was assessed using an 18item interview28. Severity of symptoms within the autism spectrum was rated using the 50 item Autism-Spectrum Quotient29. Tic symptoms and severity were measured with the Yale Global Tic Severity Scale30. Severity of psychotic symptoms was measured using the psychosis items of the Comprehensive Psychopathological Rating Scale (CPRS)31. The Padua Inventory Revised (PI-R)32 was used to determine presence and severity of subtypes. The Interpretation of Intrusion Inventory (Triple I)33 was administered to assess appraisals of obsessions. The age at which patients first fulfilled DSM-IV criteria for OCD was marked as the age of onset. Information on the presence of OCD among first degree relatives was obtained using the family tree method34. Quality of life was assessed with the EuroQol (EQ)35. The Daily Hassles questionnaire measures stress arising from daily circumstances such as work, arguments or financial problems36. As an indicator of loneliness the Loneliness Scale37 was administered. Demographic characteristics (age, gender, education level, living together with a partner, having a paid job) were recorded during the interview. Assessment at wave 4 (follow-up) At follow-up OCD symptom severity was measured with the YBOCS severity scale22 and insight in OCD with the OVIS21. Follow-up data were present in respectively 220 (87%) and 192 (76%) participants. Statistical analyses The range of insight and the proportions of patients with good, fair, poor and absent insight were studied using descriptive statistics. To examine (cross-sectional) whether insight differed per OCD severity category a one way ANOVA with post-hoc tests was conducted. One way ANOVAs with post-hoc tests and chi-square tests were performed to determine (cross-sectional) whether patients with good, fair, poor and absent insight were distinguished by different clinical and demographic characteristics. Although the main analysis was based on an intention to treat basis using linear mixed effects models (described below), we also performed some completer analyses where we tested whether severity of OCD and insight into OCD changed significantly over two years using a paired t-test, and where we described the relationship between changes in OCD symptom severity and changes in insight in terms of Pearson’s correlation coefficient. In order to assess the generalizability of the completer analysis to the full sample, we determined whether participants that i) were lost at wave 4 or ii) of whom the insight score was missing at wave 4 had different characteristics than participants that completed wave 4, using t-tests and chi-square tests. To examine the predictive value of insight for the two-year course of OCD symptom severity, given equal initial OCD severity, on an intention-to-treat basis a linear mixedeffects model was used. Therefore, the data were transformed into “long format”, so that participants either appear once (wave 3 only, in case of drop-out) or twice (both Chapter 4 64 The impact of poor insight on the natural two year course of OCD 65 wave 3 and wave 4) in the data set. The fixed-effects part of the model included a time indicator to estimate the change in OCD severity between wave 3 and wave 4 in the reference group (patients with poor or absent insight), and two time-by-insight groupinteraction terms: one for the good insight group and one for the fair insight group, to estimate to what extent the change in OCD severity differs between these groups and the reference group. The two insight group indicators themselves are left out of the model. In this way, the model corrects for the differences between the groups with respect to OCD severity at wave 3 (see Twisk, 2013)38. All variables were entered as fixed effects, with subject as the only random factor. Additionally an extended version of this model was estimated to determine whether insight predicts changes in OCD severity independently of factors that were previously found to predict the natural course of OCD, namely age of onset, comorbidity, chronicity of OCD and severity of depressive symptoms. The fixed effects part of the model was therefore extended by adding interaction terms of these variables (age at onset, comorbidity, etc.) with the time indicators, since in a long data structure covariates should be added as interaction terms. Statistical analyses were performed using the Statistical Package for Social Sciences version 18. All p values were two-tailed and statistical significance was set at p <.05. Bonferroni correction for multiple testing was applied to the 42 regression analyses testing the relationship between insight and other characteristics, the threshold for statistical significance was set at 0.05/42=0.0012.