System Maintenance Notice

Please note that maintenance work is scheduled between 4 AM and 5 AM - CET. During this time, you may experience reduced performance. Thanks for your patience.

Publications

Journal Article Ahead of print 1 citation

PubMed: 42364659 DOI: 10.1016/j.jad.2026.122171

Measurement stability and differential domain-specific trajectories of the PHQ-9 in late-life depression: Evidence from a collaborative care trial.

J Affect Disord

Authors: Rebecca Ndukwe, Marcia Scazufca, Pepijn Van de Ven

Affiliations

  • Research Ireland Centre for Research Training in Foundations of Data Science, University of Limerick, Limerick, Ireland; Data-Driven Computer Engineering Research Centre (D(2)ICE), University of Limerick, Limerick, Ireland; Health Research Institute (HRI), University of Limerick, Limerick, Ireland.
  • Department of Psychiatry, Faculty of Medicine, Universidade de São Paulo (FMUSP); Laboratory of Psychopathology and Psychiatric Therapeutics (LIM-23), Institute of Psychiatry, Hospital das Clínicas (IPq-HCFMUSP), São Paulo, Brazil.
  • Data-Driven Computer Engineering Research Centre (D(2)ICE), University of Limerick, Limerick, Ireland; Health Research Institute (HRI), University of Limerick, Limerick, Ireland.

Abstract

BACKGROUND

The Patient Health Questionnaire-9 (PHQ-9) is routinely used to monitor depressive symptoms. However, whether its empirically derived two-factor structure remains stable across repeated assessments in clinically depressed populations is not well established.

METHODS

Using item-level data from a secondary analysis of the PROACTIVE cluster-randomised controlled trial (N=558; PHQ-9 ≥10 at baseline; aged ≥60 years), we evaluated longitudinal measurement invariance of an empirically derived two-factor PHQ-9 structure across three assessment points using WLSMV structural equation modelling with multiple imputation. We then modelled domain-specific latent growth trajectories by trial arm and examined within-person temporal dynamics using a random intercept cross-lagged panel model (RI-CLPM).

RESULTS

Metric invariance was supported and partial scalar invariance achieved across all three time points, after freeing thresholds for three interpretable items (sleep disturbance, fatigue, and suicidal ideation). The intervention was associated with a steeper improvement in cognitive-affective symptoms but showed no significant effect on somatic symptom trajectories. Within-person somatic deviations at baseline were temporally associated with subsequent cognitive-affective deviations, whereas the reverse was not.

CONCLUSIONS

The two-factor structure of the PHQ-9 is metrically stable through clinical recovery in older adults with moderate-to-severe depression. Domain-specific subscores capture differential treatment response and temporal symptom dynamics that total scores conceal, and can be calculated from existing PHQ-9 items at no additional burden. These findings support the routine use of domain-level monitoring in collaborative care programmes serving older adults in low- and middle-income country (LMIC) settings.

MeSH

Depression, Patient Health Questionnaire, Research Design, Signs And Symptoms