Group Interpersonal Psychotherapy for Depression in Perinatal Adolescents in Kenya: A Pilot Randomized Clinical Trial
Lay health workers delivering group interpersonal psychotherapy sharply reduced depressive symptoms in pregnant adolescents, with the longer program showing more sustained benefit.
*Pilot randomized clinical trial; Level 1b (OCEBM).
*Pilot randomized clinical trial; Level 1b (OCEBM).
Citation
Kumar M, Tele A, Nyongesa V, et al. Group interpersonal psychotherapy for depression in perinatal adolescents in Kenya: a pilot randomized clinical trial. JAMA Network Open. 2026;9(6):e2618255. doi:10.1001/jamanetworkopen.2026.18255.
Background
Pregnant and parenting adolescents in low-resource settings often have depression but limited access to mental health specialists. This study tested whether a culturally adapted group psychotherapy could be delivered by non-specialists in primary care, and whether a shorter version works.
Patients
122 pregnant adolescents (ages 13-18) in the first to second trimester in Nairobi, Kenya, with significant depressive symptoms. Excluded: recent suicide plan/attempt, active substance misuse, severe cognitive/physical impairment, or severe mental illness.
Intervention
Group interpersonal psychotherapy delivered by community health promoters: either 8 weekly 90-minute sessions (full) or 4 weekly 90-minute sessions (mini), plus one individual pre-group session.
Control
Usual care: information sheets and routine follow-up calls.
Outcome
Depression symptoms (9-item Patient Health Questionnaire; range 0-27) and functional impairment (range 0-9). Secondary outcomes included distress and post-trauma symptoms.
Follow-up Period
Within 1 week after program completion; and 6 months.
Results
| Significant outcomes | Full program vs usual care | Mini program vs usual care |
|---|---|---|
| Depression score, 1 week after completion (primary) | Mean difference −5.79 (95% CI, −7.67 to −3.91) | Mean difference −3.97 (95% CI, −5.83 to −2.10) |
| Depression score, 6 months (primary) | Mean difference −2.22 (95% CI, −4.25 to −0.18) | NS |
| Overall distress, 1 week after completion | Mean difference −3.31 (95% CI, −6.17 to −0.45) | NS |
| Post-trauma symptoms, 1 week after completion | NS | Mean difference 5.04 (95% CI, 0.95 to 9.13) |
| Well-being, 6 months | NS | Mean difference 12.48 (95% CI, 0.79 to 24.18) |
NS = not statistically significant.
Analyses were intention-to-treat. Functional impairment did not differ between groups.
Limitations
Pilot size limits certainty and generalizability. End-of-treatment assessments occurred at different times (4 vs 8 weeks), complicating direct comparison. Attrition at 6 months may bias results. Outcomes were self-reported.
Funding
Fogarty International Center, US National Institutes of Health (K43TW010716); no funder role.
Clinical Application
In similar clinics, consider brief or full group interpersonal psychotherapy delivered by trained lay workers; use the full program when sustained symptom improvement is a priority.
Journal Club
Discussion questions and an EBM quiz for this paper. How to run a journal club →
Discussion question (1)
In this pilot 3-arm randomized clinical trial, lay health worker–delivered full IPT-G (8 sessions) lowered PHQ-9 more than TAU at 1 week (β −5.79; 95% CI −7.67 to −3.91), while mini IPT-G (4 sessions) was smaller (β −3.97). Given differential attrition and different end-of-treatment time points, would you change practice here?
EBM quiz (1 question)
1. The authors report an intention-to-treat analysis despite substantial and unequal loss to follow-up across arms (eg, 6-month follow-up: full IPT-G 60% vs mini 86.8% vs TAU 77.3%). Why is intention-to-treat generally preferred in randomized trials?
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.