2005 HSR&D National Meeting Abstract
1006 — Evaluating Computer Self-Assessments In Patients With Severe Mental Illness
Chinman MJ (VISN 22 MIRECC and RAND)
Young AS (VISN 22 MIRECC and UCLA)
Mintz J (VISN 22 MIRECC and UCLA)
Shih S (VISN 22 MIRECC)
Hassell J (VISN 22 MIRECC)
Usual care for severe mental illness (SMI) is often poor, yielding suboptimal outcomes. Improving care for this chronic illness has been difficult because routine data collection is rare and existing medical records contain little information regarding clinical needs in patients, such as symptoms or side-effects. Efficient collection of, and access to, this information is needed for evaluating and improving healthcare quality. Audio computer-assisted self-interviewing (ACASI) can facilitate data collection by presenting patients important clinical questions visually and aurally, typically in clinic waiting rooms. Collected information can then be used by providers to improve clinical decision-making and by researchers and policy makers to evaluate and improve care. Given that ACASI surveys have helped improve outcomes for a number of disorders, this study will shed light on their accuracy and reliability in people with SMI, a difficult to treat chronic illness.
In a large VAMC, ninety patients with schizophrenia or bipolar disorder (N=45 each) completed duplicate standardized symptom surveys 20 minutes apart, randomized by administration order. One was an in-person interview; the second an ACASI survey self-administered through an Internet browser using a touchscreen developed to meet the cognitive needs of people with SMI. We evaluated attitudes, understanding of the ACASI survey, reliability, concurrent validity, and a possible administration mode bias with scales from the BSI and revised BASIS-32 (Daily Living, Impulsivity, Interpersonal Competence, Depression and Anxiety, Psychosis, Substance Abuse). We evaluated reliability with Pearson correlations and bias with Mode (ACASI vs. face-to-face) x Order (ACASI first vs. ACASI second) repeated measures ANOVAs.
Across modes of administration, all ACASI and face-to-face interview scales were similar (at p<.05) and highly correlated (r=.81-.99). Patients completed the ACASI survey faster (13.6 vs.14.8 minutes, p<.05), and a large majority rated it as easier, more enjoyable, more preferable if monthly completion of a survey were required, more private, and 97% to 99% perfectly answered questions about how to use it.
ACASI-based data collection is reliable among people with bipolar disorder and schizophrenia.
ACASI systems could be a valuable component of systems to improve care in people with SMI.