Symptom Profiles Among Individuals at-Risk for Major Depressive Disorder and Their Correlates Among a Nationally Representative Sample

Background and Purpose: Despite overall improvements in treatment, major depressive disorder (MDD) remains the most prevalent mental health problem in the U.S (Hassin et al., 2005). Current interventions to improve outcomes of individuals with MDD emphasize a prevention-based approach, with intention to reduce the likelihood of future MDD onset. Yet within social work research, MDD onset is usually examined as a binary event: either an individual experiences it or they do not. While there is some evidence that symptoms associated MDD are multidimensional (Sullivan, 1998; Kendler et al. 1996), these claims have not been examined within a prevention-based framework among individuals at risk for MDD, but who are yet to experience onset. In the current study, we used nationally representative data from participants with prodromal indicators of MDD to examine whether distinct symptom profiles for MDD exist, and whether sociodemographic and clinical variables are correlates of these profiles.

Methods: We analyzed the Collaborative Psychiatric Epidemiology Survey (CPES), a nationally representative survey of non-institutionalized U.S. adults (2001-2003). Participants completed the World Mental Health Composite International Diagnostic Interview (WMH-CIDI), which is used to establish DSM-IV consistent psychiatric disorder criteria, including (MDD). Our sample included all respondents who met criteria for a past-year major depressive episode (n= 1,543). Nine MDD criteria, consistent with DSM-IV guidelines, were our latent class indicators. Social demographics (age race/ethnicity, gender, marital status, education and annual household income) and clinical variables (the presence of a lifetime DSM-IV alcohol or drug use disorder, general anxiety disorder, major depressive disorder and post-traumatic stress disorder (PTSD) were examined in a follow-up analysis to establish correlates of class membership. Mplus was used to establish MDD symptom profiles, where model fit statistics and substantive theory was used to identify the appropriate number of latent classes. Next, multinomial regression analysis examined social and clinical correlates of the latent classes. CPES survey weights were used in all analyses to obtain accurate standard error estimates for the population.

Results: Four MDD symptom profiles emerged; High Cognitive Functioning (3.8% of the sample), Absolute Depression (63.7% of the sample), Diminished Cognitive Functioning/Low Suicidal Ideation (24.7% of the sample) and Diminished Cognitive Functioning/High Suicidal Ideation (7.7% of the sample). The multinomial regression model showed that individuals in the High Cognitive Functioning class and the Absolute Depression class were more likely to be racial/ethnic minority. Individuals in the Absolute Depression class and the Diminished Cognitive Functioning/Low Suicidal Ideation were more likely to experience PTSD.

Conclusions and Implications: Symptom profiles clearly exist among individuals at risk for MDD onset. Race/ethnicity and PTSD were associated with these profiles, suggesting that culturally appropriate assessments are essential for the effective detection of depression profiles, and that integrated treatment approaches focusing on the intersection of trauma and depression are in high need. Finally, future research associated with MDD needs to move beyond MDD onset as a binary event. Through identifying symptom profiles of MDD, social workers may be able to implement targeted interventions that match client experiences to necessary services.