1 - (SYM 104) Latent Class Analysis of Depressive Symptoms and Associations with Suicidal Ideation, Plans, and Attempts Among a National Sample of Depressed Adults

To better understand depression phenomenology, studies have sought to develop symptom clusters that can aid clinicians in identifying and treating profiles of patients presenting with depressive disorders. Despite clinical overlap in depression and suicide risk, limited research to date has sought to investigate whether patients with particular clusters of depressive symptoms are more likely to report suicide outcomes.

We leveraged data from the National Survey on Drug Use and Health (NSDUH, years 2008-2014, n=41969). We use latent class analysis (LCA) of depressive symptoms to (1) identify classes of different DSM-IV depression symptom presentations among adults endorsing depressive symptoms and (2) explore differences among classes in suicidal thoughts, plans, and attempts. The depressive symptom items (except for thoughts of death or urges to self-harm given its overlap with the suicide-related outcomes of interests) were set as the indicator variables for the LCA. After identifying the best-fitting latent class model, chi-square analyses were used to examine whether particular classes were more likely to report suicide outcomes. Pairwise comparisons tested for differences between class for all significant omnibus chi-square analyses.

The four-class model had optimal fit. Class 1 (53% of sample) had high endorsement of all examined symptoms, except worthlessness and psychomotor disturbance. Class 2 (9% of sample) had high endorsement of worthlessness and low endorsement of psychomotor disturbance, with all other symptoms having moderate endorsement. Class 3 (31% of sample) had high endorsement of sleep issues, low endorsement of worthlessness and psychomotor retardation, with all other symptoms having moderate endorsement. Class 4 (7% of sample) had lower endorsement of all examined symptoms, except depressed mood. When examining differences in suicide-related outcomes, Class 1 had higher endorsement rates, followed by Class 2, with Classes 3 and 4 having significantly lower rates.

Results highlight the importance of examining both a composite depression score to understand the severity and quantity of symptoms endorsed, as well as the individual items that a given individual is experiencing. In addition to risk among the class with a higher endorsement of symptoms overall, Class 2 also had elevated rates of endorsement of suicidal plans and attempts, underscoring a need to pay close attention to suicide risk among individuals who endorse worthlessness.