Identifying youth at risk for psychosis using the Behavior Assessment System for Children, Second Edition

Abstract

Identification of youth at risk for or with early psychosis has become the focus of many research and clinical initiatives, as early intervention may be linked to better long-term outcomes. Efforts to facilitate identification have led to the development of several self-report instruments that intend to quickly assess “attenuated” psychosis, potentially screening people for further evaluation. The widely used Behavior Assessment System for Children, Second Edition (BASC-2) includes the atypicality scale, a scale that may be useful for risk screening as it is designed to recognize emerging symptoms of psychosis. The current study aimed to evaluate the utility of the BASC-2 for identifying youth at high clinical risk or with early psychosis within a sample of 70 help-seeking participants aged 12–22. Atypicality scores were compared to risk status (low-risk, high-risk or early psychosis) as determined by the clinician-administered Structured Interview for Psychosis-Risk Syndromes (SIPS). The relative accuracy of the atypicality scale was evaluated against three self-report screeners specifically designed to identify this population. Results indicate that the BASC-2 atypicality scale may be a useful tool for identifying youth in early stages of psychosis. Moreover, the atypicality scale is comparable if not superior to other specialized risk screening instruments in terms of predictive ability. Given the widespread use of the BASC-2 across educational and mental health settings, evidence for convergent validity between the BASC-2 atypicality scale and SIPS diagnoses has the potential to make screening available to a greater population and facilitate earlier detection and intervention.

Introduction

Psychotic disorders (e.g., schizophrenia) often have negative effects on emotional, social, and occupational functioning, and can impose a substantial economic burden on families and health care systems in general (Wu et al., 2005). A strong association has been established between the duration of untreated psychosis and a broad range of clinical and functional outcomes, including outcomes relating to symptom management, social and occupational functioning, and morbidity and mortality over time (Marshall et al., 2005). Evidence suggests that early identification and treatment can improve these negative sequelae (Stafford et al., 2013), however, many individuals with psychosis receive treatment only after symptoms have reached a level of severity that warrants urgent attention and may have already contributed to substantial psychosocial impairment (e.g., social isolation or drug and alcohol abuse; Bergner et al., 2008, Gerson et al., 2009, Tanskanen et al., 2011). Identification of clinical concerns at an earlier stage of symptom progression may lead to better outcomes for individuals on a trajectory toward psychotic illnesses. The goal of early identification that might lead to intervention, however, hinges on the improvement of strategies for identifying individuals early in the course of psychotic symptom progression.

The characterization of a high-risk or “prodromal” phase of psychosis marked by lower level or “attenuated” symptoms (psychosis-like in quality, but typically briefer and/or less intense; McGlashan et al., 2010) has expanded the potential opportunities for early intervention in this population (Fusar-Poli et al., 2012). Identifying individuals in early phases of psychosis would allow for targeted treatment of concomitant distress. Additionally, continuous psychosis symptom monitoring could facilitate the initiation of immediate specialized treatment should symptoms progress to a diagnosable psychotic disorder. Assessments such as the Structured Interview for Psychosis-Risk Syndromes (SIPS; Miller et al., 2003) and the Comprehensive Assessment for At-Risk Mental States (CAARMS; Yung et al., 2005) have made possible the reliable identification of patients at substantially increased risk for psychotic disorders; however, these tools require considerable time for clinician training and administration, and are unlikely to be widely adopted outside of specialty settings.

For these reasons, several brief self-report questionnaires have been developed with the goal of more efficiently screening people in the earliest stages of psychotic illness. Several screening tools targeting attenuated symptoms have been introduced (e.g., Miller et al., 2004, Ord et al., 2004, Loewy et al., 2011). Recent research has shown that these self-ratings of attenuated psychotic symptoms may be useful for identifying individuals with psychosis-risk or emergent psychotic symptoms among those seeking mental health services assessed within an early psychosis specialty clinic (Kline et al., 2012a, Kline et al., 2013). A few of these screening measures have been used for ‘real-world’ applications, for example in screening newly incarcerated men for mental health needs (Jarrett et al., 2012), and as a first-step assessment in a high-risk recruitment protocol (Ising et al., 2012, Rietdijk et al., 2012).

Despite promise, aspects of these screeners present challenges to clinical application. Concerns about the reliability of these measures exist, as screeners have been validated in different samples that may be distinct with regard to clinical concerns, cultural context, and other distinguishing features (Kline et al., 2012b). These differences in validation samples introduce uncertainty with regard to whether screener threshold scores are valid and stable across diverse populations. Additionally, screening specifically for symptoms signifying risk for psychosis could be perceived as highly stigmatizing (Corcoran et al., 2005, Reavley and Jorm, 2011). Further, despite findings that adolescence appears to represent a period of peak incidence for psychosis onset (Amminger et al., 2006, Kessler et al., 2007, Schimmelmann et al., 2007), screeners may pose questions that are developmentally inappropriate for younger respondents.

Broad-based mental health screening tools and checklists are commonly employed in behavioral and mental health screening for youth and have the potential to address some of the concerns raised by psychosis-specific screeners. Widely-used inventories such as the Behavior Assessment System for Children, Second Edition (BASC-2; Reynolds and Kamphaus, 2004) and the Child Behavior Checklist (CBCL; Achenbach, 1991) are composed of questions designed to assess many different areas of functioning, including numerous clinical domains such as depression, anxiety, conduct problems and other symptomatology related to common mental health diagnoses. As a result of their comprehensiveness, these measures offer clinically diverse information helping to inform a thorough case conceptualization. The effectiveness of these measures to better inform identification and treatment of early psychosis, however, is relatively unknown. One study found that the parent rating scale of the CBCL was not useful in discriminating at-risk participants who converted to psychosis from at-risk participants who did not in the year following initial assessment (Simeonova et al., 2011). Given other findings demonstrating poor agreement between parents and youth on the presence of psychotic-like symptoms (Kline et al., 2013, Nugent et al., 2013), self-report questionnaires may be more informative for the purpose of screening. Additionally, Simeonova and colleague's study attempted to distinguish future converters from within a sample at high-risk; it did not speak to whether or not parent CBCL reports could be used to identify individuals at risk or already with psychosis from individuals not at risk.

The BASC-2 is another widely used multi-informant behavior checklist that, in addition to more common behavioral health issues, includes an “atypicality” scale specifically designed to recognize the presence of symptoms thought to indicate heightened risk for psychosis. Items included in the atypicality scale assess symptoms similar to those targeted by psychosis-risk screeners, including odd behaviors, paranoia, and perceptual abnormalities. Although the BASC-2 has received extensive use in research and clinical practice, little is known about the atypicality scale and its utility in terms of psychosis and psychosis-risk screening (Nugent et al., 2013). Given its wide use in general mental health settings, its broad coverage of a variety of mental health concerns, and its developmentally sensitive structure and item content, the BASC-2 could be a useful tool to better facilitate early identification.

The aim of the current study is to investigate the ability of the BASC-2 atypicality scale to predict SIPS-defined psychosis or psychosis-risk status. This study also seeks to examine the predictive strength of the atypicality scale in relation to three leading psychosis-risk screeners with regard to SIPS status. We hypothesize that atypicality scores will be strongly associated with SIPS ratings as well as other screening measures designed specifically for assessing psychosis risk.