ABSTRACT
BACKGROUND: Although mental health challenges disproportionately affect people in humanitarian contexts, most refugee youth do not receive the mental health support needed. Uganda is the largest refugee-hosting nation in Africa, hosting over 1.58 million refugees in 2022, with more than 111,000 living in the city of Kampala. There is limited information about effective and feasible interventions to improve mental health outcomes and mental health literacy, and to reduce mental health stigma among urban refugee adolescents and youth in low- and middle-income countries (LMICs). Virtual reality (VR) is a promising approach to reduce stigma and improve mental health and coping, yet such interventions have not yet been tested in LMICs where most forcibly displaced people reside. Group Problem Management Plus (GPM+) is a scalable brief psychological transdiagnostic intervention for people experiencing a range of adversities, but has not been tested with adolescents and youth to date. Further, mobile health (mHealth) strategies have demonstrated promise in promoting mental health literacy. OBJECTIVE: The aim of this study is to evaluate the feasibility and effectiveness of two youth-tailored mental health interventions (VR alone and VR combined with GMP+) in comparison with the standard of care in improving mental health outcomes among refugee and displaced youth aged 16-24 years in Kampala, Uganda. METHODS: A three-arm cluster randomized controlled trial will be implemented across five informal settlements grouped into three sites, based on proximity, and randomized in a 1:1:1 design. Approximately 330 adolescents (110 per cluster) are enrolled and will be followed for approximately 16 weeks. Data will be collected at three time points: baseline enrollment, 8 weeks following enrollment, and 16 weeks after enrollment. Primary (depression) and secondary outcomes (mental health literacy, attitudes toward mental help-seeking, adaptive coping, mental health stigma, mental well-being, level of functioning) will be evaluated. RESULTS: The study will be conducted in accordance with CONSORT (Consolidated Standards of Reporting Trials) guidelines. The study has received ethical approval from the University of Toronto (#40965; May 12, 2021), Mildmay Uganda Research Ethics Committee (MUREC-2021-41; June 24, 2021), and Uganda National Council for Science & Technology (SS1021ES; January 1, 2022). A qualitative formative phase was conducted using focus groups and in-depth, semistructured key informant interviews to understand contextual factors influencing mental well-being among urban refugee and displaced youth. Qualitative findings will inform the VR intervention, SMS text check-in messages, and the adaptation of GPM+. Intervention development was conducted in collaboration with refugee youth peer navigators. The trial launched in June 2022 and the final follow-up survey will be conducted in November 2022. CONCLUSIONS: This study will contribute to the knowledge of youth-tailored mental health intervention strategies for urban refugee and displaced youth living in informal settlements in LMIC contexts. Findings will be shared in peer-reviewed publications, conference presentations, and with community dissemination. TRIAL REGISTRATION: ClinicalTrials.gov NCT05187689; https://clinicaltrials.gov/ct2/show/NCT05187689. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/42342.
ABSTRACT
COVID-19, due to its accelerated spread has brought in the need to use assistive tools for faster diagnosis in addition to typical lab swab testing. Chest X-Rays for COVID cases tend to show changes in the lungs such as ground glass opacities and peripheral consolidations which can be detected by deep neural networks. However, traditional convolutional networks use point estimate for predictions, lacking in capture of uncertainty, which makes them less reliable for adoption. There have been several works so far in predicting COVID positive cases with chest X-Rays. However, not much has been explored on quantifying the uncertainty of these predictions, interpreting uncertainty, and decomposing this to model or data uncertainty. To address these needs, we develop a visualization framework to address interpretability of uncertainty and its components, with uncertainty in predictions computed with a Bayesian Convolutional Neural Network. This framework aims to understand the contribution of individual features in the Chest-X-Ray images to predictive uncertainty. Providing this as an assistive tool can help the radiologist understand why the model came up with a prediction and whether the regions of interest captured by the model for the specific prediction are of significance in diagnosis. We demonstrate the usefulness of the tool in chest x-ray interpretation through several test cases from a benchmark dataset.