Improving prognostic prediction in depression: The utility of pre-treatment symptom variability and trajectory

Background: Although baseline depressive severity is widely used to predict prognosis in major depressive disorder (MDD), recent evidence questions its reliability. Depression is dynamic and person-specific, raising the possibility that pre-treatment symptom trajectories and variability provide more informative prognostic signals.Methods: Data were drawn from two randomized controlled trials of short-term psychotherapy for depression. Using a digital depression-assessment app with high convergent validity with the Hamilton Depression Rating Scale (HRSD), daily depressive symptoms were tracked during the week preceding treatment initiation. Two within-patient parameters were derived: (a) symptom trajectory (improving vs. worsening), and (b) variability (low vs. high, based on sample median). These parameters yielded four pre-treatment dynamic types. Linear mixed-effects models predicted HRSD outcomes across 16 weeks of treatment, controlling for baseline HRSD.Results: All groups showed significant symptom improvement across treatment (p < .001). Patients with worsening symptoms and high variability prior to treatment demonstrated the steepest improvement trajectory (β = –0.22, 95% CI: –0.33 to –0.10, p < .001), significantly outperforming other groups. No significant differences emerged among the remaining three groups.Conclusions: Pre-treatment dynamics of depressive symptoms, particularly the combination of worsening trajectory and high variability, predicted enhanced treatment response beyond baseline severity. These findings suggest that short-term monitoring of symptom dynamics immediately before treatment may provide clinicians with valuable prognostic information. Incorporating such measures could improve decision-making and refine guidelines that currently rely heavily on static severity scores.