International Journal of Psychiatry Research
Open AccessAn AI-Enabled, Trauma-Informed Rehabilitation Model for Ethiopian Women with Complex Trauma: Programmatic Implementation, Early Outcomes, and Implications for Scalable Care in Africa
Authors: Yared Alemu; Patrick Ohiomoba, Yoel Kahsay, Fikirte Weldearegay, Hermon Amare, Selam Negussie, Delessa Neger, Tsion Fekadu, Octavian Balatel, Abadi Degefu, Selam Abdi, Randa Ibrahim, Kaleb Amare.
Abstract
The Lenegewa mental health treatment model is a Trauma-Informed, AI-enabled digital health model that integrates culturally relevant clinical mental health services and AI-enabled digital tools to address complex trauma among marginalized Ethiopian women. This study evaluates outcomes from the second cohort (N = 413) of the Lenegewa Women's Rehabilitation and Skill Development Center program, comprising women with histories of severe childhood adversity, multiple traumas, including sexual violence, and substance use disorders. The program achieved a high retention rate of 92%, significantly higher than in comparable programs. The Lenegewa model comprises three core components: (1) AI-enabled digital screening and risk stratification through ClarityConnect. This end-to-end AI-enabled digital platform integrates standardized self-report measures—including the PHQ-9, GAD-7, CAGE-AID, SDS, and the ACE questionnaire— alongside an innovative 30-second voice-based emotionrecognition algorithm designed to quantify emotional distress objectively; (2) a task-shared physician–coach care delivery system, in which trained Physician-Coaches provide frontline clinical services under psychiatric supervision; and (3) an integrated trauma and substance use treatment framework that combines motivational interviewing, trauma-focused cognitive behavioral therapy (TF-CBT), and evidence-based substance use disorder interventions. Quantitative findings show that 17% of participants required urgent psychiatric or SUD intervention, and 24% had comorbid SUD diagnoses with mixed treatment outcomes. The cohort experienced a 2.5% suicide attempt rate, prompting enhanced crisis protocols. Culturally responsive adaptations significantly improved treatment engagement, including collaboration with local religious entities and psychoeducation addressing local spiritual beliefs (e.g., evil-eye attributions). Implementation challenges included treatment resistance (110 missed therapy sessions), somatic symptom preoccupation, and infrastructure limitations. The study demonstrates the feasibility of this comprehensive model in low-resource settings. It provides critical insights for scaling traumainformed care through a multidisciplinary team-based approach, AI-enabled digital tool integration, and culturally adapted mental health interventions. These findings have important implications for global mental health programs serving trauma-affected populations in resource-limited contexts.
Editor-in-Chief
View full editorial board →