Depression subtypes and their response to cognitive behavioral therapy: A latent transition analysis

Depression is a heterogeneous condition, with multiple possible symptom-profiles leading to the same diagnosis. Descriptive depression subtypes based on observation and theory have so far proven to have limited clinical utility.

To identify depression subtypes and to examine their time-course and prognosis using data-driven methods.

Latent transition analysis was applied to a large (N = 8380) multi-service sample of depressed patients treated with cognitive behavioral therapy (CBT) in outpatient clinics. Patients were classed into initial latent states based on their responses to the Patient Health Questionnaire-9 of depression symptoms, and transition probabilities to other states during treatment were quantified. Qualitatively similar states were clustered into overarching depression subtypes and we statistically compared indices of treatment engagement and outcomes between subtypes using post hoc analyses.

Fourteen latent states were clustered into five depression subtypes: mild (2.7%), severe (9.8%), cognitive-affective (23.7%), somatic (21.4%), and typical (42.4%). These subtypes had high temporal stability, and the most common transitions during treatment were from severe toward milder states within the same subtype. Differential response to treatment was evident, with the highest improvement rate (63.6%) observed in the cognitive-affective subtype.

Replicated evidence indicates that depression subtypes are temporally stable and associated with differential response to CBT.