Handley TE, Hiles SA, Inder KJ, Kay-Lambkin FJ, Kelly BJ, Lewin TJ, McEvoy M, Peel R, Attia JR. Predictors of suicidal ideation in older people: a decision tree analysis.
Am J Geriatr Psychiatry 2014;
22:1325-35. [PMID:
24012228 DOI:
10.1016/j.jagp.2013.05.009]
[Citation(s) in RCA: 35] [Impact Index Per Article: 3.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 02/28/2013] [Revised: 05/24/2013] [Accepted: 05/28/2013] [Indexed: 10/26/2022]
Abstract
OBJECTIVES
Suicide among older adults is a major public health issue worldwide. Although studies have identified psychological, physical, and social contributors to suicidal thoughts in older adults, few have explored the specific interactions between these factors. This article used a novel statistical approach to explore predictors of suicidal ideation in a community-based sample of older adults.
DESIGN
Prospective cohort study.
PARTICIPANTS AND SETTING
Participants aged 55-85 years were randomly selected from the Hunter Region, a large regional center in New South Wales, Australia.
MEASUREMENTS
Baseline psychological, physical, and social factors, including psychological distress, physical functioning, and social support, were used to predict suicidal ideation at the 5-year follow-up. Classification and regression tree modeling was used to determine specific risk profiles for participants depending on their individual well-being in each of these key areas.
RESULTS
Psychological distress was the strongest predictor, with 25% of people with high distress reporting suicidal ideation. Within high psychological distress, lower physical functioning significantly increased the likelihood of suicidal ideation, with high distress and low functioning being associated with ideation in 50% of cases. A substantial subgroup reported suicidal ideation in the absence of psychological distress; dissatisfaction with social support was the most important predictor among this group. The performance of the model was high (area under the curve: 0.81).
CONCLUSIONS
Decision tree modeling enabled individualized "risk" profiles for suicidal ideation to be determined. Although psychological factors are important for predicting suicidal ideation, both physical and social factors significantly improved the predictive ability of the model. Assessing these factors may enhance identification of older people at risk of suicidal ideation.
Collapse