Modifiable predictors of suicidal ideation during psychotherapy for late-life major depression. A machine learning approach

Abstract

This study aimed to identify subgroups of depressed older adults with distinct trajectories of suicidal ideation during brief psychotherapy and to detect modifiable predictors of membership to the trajectories of suicidal ideation. Latent growth mixed models were used to identify trajectories of the presence of suicidal ideation in participants to a randomized controlled trial comparing Problem Solving Therapy with "Engage" therapy in older adults with major depression over 9 weeks. Predictors of membership to trajectories of suicidal ideation were identified by the convergence of four machine learning models, i.e., least absolute shrinkage and selection operator logistic regression, random forest, gradient boosting machine, and classification tree. The course of suicidal ideation was best captured by two trajectories, a favorable and an unfavorable trajectory comprising 173 and 76 participants respectively. Members of the favorable trajectory had no suicidal ideation by week 8. In contrast, members of the unfavorable trajectory had a 60% probability of suicidal ideation by treatment end. Convergent findings of the four machine learning models identified hopelessness, neuroticism, and low general self-efficacy as the strongest predictors of membership to the unfavorable trajectory of suicidal ideation during psychotherapy. Assessment of suicide risk should include hopelessness, neuroticism, and general self-efficacy as they are predictors of an unfavorable course of suicidal ideation in depressed older adults receiving psychotherapy. Psychotherapeutic interventions exist for hopelessness, emotional reactivity related to neuroticism, and low self-efficacy, and if used during therapy, may improve the course of suicidal ideation.

Fig. 1. Latent growth mixture model (LGMM) estimated growth curves of the presence of suicidal ideation in 249 older adults with major depression randomly assigned to “engage” or problem-solving therapy.

The figure presents two LGMM trajectories with 95% CI of the probability of the presence of suicidal ideation over 9 weeks. Red color represents an unfavorable (31% of participants) and blue color represents a favorable trajectory of suicidal ideation (69% of participants). The presence of suicidal ideation during the course of the 9-week treatment trial is defined as a score of 1, 2, or 3 in the Suicide item of the 24-item Hamilton Depression Rating Scale.

Fig. 2. Variable importance in predicting membership to suicidal ideation trajectories in older adults with major depression (N = 249) during 9 weeks of treatment with “Engage” or problem-solving therapy estimated by gradient-boosted machine.

Predictors are ordered from top to bottom in order of importance in the Gradient-Boosted Machine. The horizontal axis represents the variable importance measure in the Gradient-Boosted Machine that quantifies the improvement in prediction accuracy due to a predictor.

Fig. 3. Latent growth mixture model (LGMM) estimated growth curves of the presence of suicidal ideation in 99 older adults with major depression and suicidal ideation at baseline randomly assigned to “Engage” or Problem-solving therapy.

The figure presents two LGMM trajectories with 95% CI of the probability of the presence of suicidal ideation over 9 weeks. Red color represents an unfavorable (42% of participants) and blue color represents a favorable trajectory of suicidal ideation (58% of participants). The presence of suicidal ideation at baseline and during the treatment trial is defined as a score of 1, 2, or 3 in the Suicide item of the 24-item Hamilton Depression Rating Scale.

Fig. 4. A classification tree predicting membership to the two suicidal ideation trajectories using the three strongest predictors of suicidal ideation in older adults with major depression randomly assigned to “engage” or problem-solving therapy.

The classification tree offers a clinical decision rule for identifying membership to either the favorable or the unfavorable trajectory of suicidal ideation during psychotherapy. At each box, participants are classified into either the left or right succeeding box based on the decision rule displayed. Green boxes signify higher probability to have a favorable trajectory of suicidal ideation and blue boxes signify a higher probability to have an unfavorable trajectory of suicidal ideation. The final prediction of trajectory membership is determined by the final box a participant falls into. Each box presents the proportion of total participants who fall into that box (bottom proportion), as well as the proportion of who belongs to the favorable and the unfavorable trajectories (left and right proportions in center of box, respectively), irrespective of the box prediction. For example, the first decision rule split indicates that 58% of participants had hopelessness scores <19; 83% of these participants had favorable trajectories, while 17% of these participants had unfavorable trajectories.