A Validity Study of a Brief Suicidal Thinking Measure with Older Adults

Purpose: Suicide is a significant problem among older adults. While the prevalence of suicide in the general population is 11 per 100,000, amongst older adults it increases to 14.3 per 100,000. The prevalence of suicidal ideation among older adults is between 1% and 12%. Research has suggested the probability of a person who is contemplating suicide developing a suicide plan is about .34, and the probability that he or she then makes an attempt is .72. Consequently, the identification of suicidal thinking in older adult clients is a critically important clinical task. A number of measures of suicidal ideation exist, but few have been validated with older adults. This is a critical gap in the literature on measures of suicidal thinking (Brown, 2002). This validity study of a four item suicidal thinking screener addresses this gap

Method: Older adults entering services in two agencies that offer health and mental health services were informed of this validity study of the Brief Suicidal Thinking Screener (BSTS). Two-hundred clients signed informed consent documents and completed the BSTS, the Geriatric Depression Scale (GDS), and other measures at the end of a psychosocial assessment conducted by agency social workers. During the psychosocial assessment, and prior to their completion of the BSTS and other measures, clients were assessed for suicidal ideation and risk. These ratings of suicidal thinking and suicide risk, along with the scores from the BSTS, the GDS, and scores from other measures were analyzed using multiple methods, including logistic regression, receiver operator characteristic curve (ROC) analysis, confirmatory factor analysis, and methods from generalizability theory and item response theory (IRT).

Results: The results from the multiple analytic methods converged to provide various forms of evidence supporting the reliability and validity of the scores from the BSTS. For example, the coefficient alpha estimate of the reliability of the BSTS scores was .90. The correlation between scores on the BSTS and the independent clinician ratings of suicidal thinking was .63; and the partial correlation between scores on the BSTS and the independent clinician ratings, controlling for demographics and scores on the GDS was .56. Patterns of correlations between scores on the BSTS, the scores from the other measures, and demographic variables were consistent with convergent and divergent validity. A clinical cutting score was identified that had a sensitivity of .85, a specificity of .80, and a dependability index of .90. Analyses based upon IRT suggested no gender bias in the BSTS scores.

Implications: The results provided strong evidence for the validity of the scores from the BSTS with older adults. In contrast to most other measures of suicidal ideation, which take about 10 minutes to complete (Brown, 2002), the BSTS can be completed in five minutes. The results of this study suggest the BSTS can be a valuable assessment tool for social workers who work with older adults. Reference: Brown, G. (2002). A review of suicide assessment measures for intervention research with adults and older adults. Bethesda, MD: National Institute of Mental Health. Online: http://www.nimh.nih.gov/suicideresearch/adultsuicide.pdf