Research That Matters (January 17 - 20, 2008)

Regency Ballroom Wings (Omni Shoreham)

Assessing and Advancing the Validity of Adolescent Suicidality Questionnaires

Michael E. Woolley, PhD, University of Michigan-Ann Arbor.

Purpose: Completed suicides tripled among youth between 1952 and 1996. In 1999 the Surgeon General called for expanded research on youth suicide prevention and the specific need to advance the validity of youth suicidality assessment tools. A call for research that matters; the current study was a direct response to that call. Assessment questionnaire development is a quantitative endeavor; however, there is growing interest in the systematic application of mixed methods to questionnaire development and validity assessment. Such mixed methods in questionnaire development, or cognitive methods, are used to collect qualitative data directly from potential respondents about how they understand, think about, and choose responses to assessment questionnaires.

Method: Two such mixed methods were used to collect data from adolescents (13-17) who had been psychiatrically hospitalized for suicidality (ideation to attempt). Cognitive pretesting (CP) interviews involve asking a respondent, while they read and respond to an assessment question, to explain what/how he/she: (a) interprets what a question is asking, (b) thinks about while forming a response, such as experiences, feelings, beliefs, etc., and (c) chooses an answer option. CPed was the Suicidal Ideation Questionnaire-Junior (SIQ-JR: Reynolds & Mazza, 1999), the most widely utilized such questionnaire. Construct probe interviews (CPI) apply open-ended questions to gather qualitative data about how adolescents experience the context and content of a targeted construct, suicidality.

Results: To date, 34 adolescents have been interviewed, with the full sample to be collected by the end of June 2007. Preliminary analyses indicate that while many of the SIQ-JR items perform well, several items have validity problems. For example, the item I thought about writing a will…, which is infrequently endorsed but seen as a high risk sign; while some adolescents knew what a will was, most did not interpret this question as intended. Nearly half thought wills were what “old people write”, some felt wills are for people with money and property, and they have neither, finally several just did not know what a will was. The question I thought about telling people I plan to kill myself…, several serious suicide attempters interpreted this wording as attention seeking behavior, and not something they would endorse. For those respondents it does not assess suicide risk, even the opposite. For the SSWR poster, full analyses will be available for the cognitive pretesting, including describing a codebook to quantitatively code the validity of responses, and the inter-rater reliability of multiple coders of the data. Finally, the CPI data are revealing a wide range of circumstances, dynamics, and responses to stressors and struggles that lead adolescents to think about or attempt suicide, and make them want to live or help when they feel suicidal.

Implications: The implications of this research matters. We need valid teen suicide assessment risk questionnaires that can both be used to screen a population in a school or applied in a clinical setting. Current findings will be used to develop such a questionnaire. Finally, the application and ongoing evolutions of such mixed methods in questionnaire development will be discussed.