Can delusions be self-assessed? Concordance between self- and observer-rated delusions in schizophrenia

Abstract

Several multi-dimensional self-report scales have been developed to assess delusional ideation in the general population. However, self-ratings of positive symptoms in patients with psychosis are often considered unreliable due to neuro-cognitive disturbance and lack of insight. This study tested associations of self- and observer-rated delusions as well as factors associated with discrepancies. Observer-rated delusions were assessed in 80 in- and outpatients with schizophrenia spectrum disorders by trained raters with the Positive and Negative Syndrome Scale. Self-rated delusions were assessed with the Peters et al. Delusions Inventory and the Paranoia Checklist. Correlations between self- and observer-rated overall delusions ranged from 0.49 to 0.57. Associations between specific delusions of persecution and grandiosity were moderate but unique. Good concordance of ratings was not restricted to outpatients or patients with fewer positive symptoms. Patients with lower self- than observer ratings of delusions were characterised by fewer years of education, lower functioning, more negative symptoms and less insight. The results indicate that patients can reliably provide information with regard to the presence and type of delusional beliefs. Thus, patient ratings are a valid additional source of diagnostic information.

Introduction

Observer-rated symptom scales are consistently used as main outcome in treatment studies of psychosis. However, despite many advantages, observer-based assessments are time-consuming, costly and prone to socially desirable answers and, in particular, observer-bias (Wykes et al., 2008).

Research on the continuum of psychosis in the population (Verdoux and Van Os, 2002, van Os et al., 2009) has brought along reliable and valid self-report single symptom measures, such as the Peters et al. Delusions Inventory (PDI, Peters et al., 1999) or the Paranoia Checklist (Freeman et al., 2005). Delusions are assessed by providing lists of delusional beliefs (e.g., “I believe people are observing me”) to be rated on several dimensions, such as frequency, distress and conviction. These scales are now widely used to investigate subclinical psychosis (Larøi and Van der Linden, 2005, Preti et al., 2007) and could potentially make a valuable contribution to clinical research.

The prevailing non-application of self-report scales in psychosis research is rooted in the concern that self-report of psychotic symptoms is unreliable. Foremost, it is argued that psychosis directly affects perceptions of reality which disrupt the ability of patients to adequately assess their mental status. This is supported by studies finding insight, consensually defined as awareness of having a mental disorder, its symptoms and its implications, to be at least partly lacking in approximately 50% of patients with schizophrenia (Lysaker et al., 2002, Sevy et al., 2004, Gharabawi et al., 2006). Furthermore, several studies support the idea that disordered thinking and neuro-cognitive impairment, such as executive functioning, attention and memory deficits are related to deficits in awareness of symptoms (for reviews see Cooke et al., 2005, Schad et al., 2007).

Nevertheless there has been little research on the concordance between self- and observer ratings. Studies employing global measures, such as the Brief Psychiatric Rating Scale (Overall and Gorham, 1962) or the Brief Symptom Inventory (Derogatis, 1993) have generally found satisfying associations between self- and observer-rated overall pathology (Morlan and Siang-Yang, 1998), negative symptoms (Bottlender et al., 2003) and positive symptoms (Dixon and King, 1995, Hamera et al., 1996, Preston and Harrison, 2003, Liraud et al., 2004). Two studies even found concordance between self- and observer ratings to be better for positive symptoms than for other symptom domains (Dixon and King, 1995, Hamera et al., 1996).

No study has investigated the concordance of delusion-specific self-rating scales and observer ratings. Delusions might be particularly difficult to self-assess as they result, per definition, from reality distortion. Thus, directly asking a deluded patient whether delusions are present is likely to be unproductive. Unvalued questions, however, as provided in the PDI or the Paranoia Checklist, that simply ask to indicate the presence or absence of specific beliefs, are more likely to be endorsed. Accordingly, validation studies of the PDI consistently find higher scores for patients, pointing to its ability to correctly identify delusional beliefs in patients (Verdoux et al., 1998, Peters et al., 1999, Peters et al., 2004, Lincoln, 2007).

Few studies have attempted to identify whether discrepancies between self- and observer ratings of delusions are predictable by patient characteristics. Although some researchers even use self-observer discrepancies as a measure of symptom unawareness (Dixon et al., 1998) lack of insight has not consistently been found to predict self-observer discrepancies (Selten et al., 2000, Liraud et al., 2004, Bell et al., 2007). However, Selten et al. (2000) found higher levels of depression to predict more concordance between self- and observer ratings of negative symptoms and ascribed this to depressive realism (Ackermann and DeRubeis, 1991).

The aim of this study is to investigate the concordance between patient-rated delusions and observer-rated delusions, hypothesising at least moderate associations. Furthermore, we aim to investigate whether discrepancies in self- and observer ratings are associated with clinical variables, such as symptom severity, duration of disorder, depression and lack of insight and neuro-cognitive variables, such as memory, intelligence and cognitive flexibility.