Table 1 Publication characteristics of included studies
First Author (year) | Study Aim | Area of MH focus | Sample size and characteristics | Data analyzed |
|---|---|---|---|---|
Acion (2017) [36] | Predict substance abuse treatment success using 17 different machine learning models | Substance abuse | 99,013 Hispanic individuals | TEDS-D 2006–2011 |
Augsburger (2017) [31] | Assessed risk-taking behavior in refugees after exposure to trauma using a gamified BART | Trauma | 56 Refugees resettled in Germany | Surveys and data on BART |
Baird (2022) [35] | Used drawings by refugee children to estimate predictors of exposure to violence and mental wellbeing | Trauma | 2480 Syrian refugee children | USF 2016 dataset |
Castilla-Puentes (2021) [41] | To understand how Hispanic populations converse about depression by conducting big data analysis of digital conversations through machine learning | Depression | 441,000 unique conversations about depression; 43,000 (9.8%) conversations were by Hispanics | Conversations from open sources like topical MH websites, message boards, social networks, and blogs |
Choi (2020) [32] | Examined the predictive ability of discrimination-related variables, coping mechanisms, and sociodemographic factors on the psychological distress level of Korean immigrants in the U.S. during the pandemic | Psychological distress | 790 Korean immigrants, foreign and US-born | Surveys |
Drydakis (2021) [33] | Understanding associations between the number of mobile applications in use aiming to facilitate immigrants’ societal integration and increased level of integration, good overall health, and mental health | Depression | 287 immigrants in Greece | Surveys |
Erol (2022) [34] | Examine the PTSD and depression levels of Syrian refugee children and adolescents, the difficulties they experienced in access to food and education, and the changes in their family income, and evaluate the effects of these factors on symptom severities of depression and PTSD | Depression & PTSD | 631 Refugee children living in Turkey | Surveys |
Goldstein (2022) [37] | To examine the relationship between experiencing discrimination and suicidal ideation in Hispanic populations | Suicidal ideation | 22,968 Hispanic individuals | Holmusk and MindLinc EHR datasets, 52,703 patient-year observations from 2010 to 2020 |
Haroz (2020) [39] | Develop a model using ML methods to better identify those at highest risk for suicide in Native American communities | Suicidal ideation | 2,390 Native American individuals | Surveillance program data |
Huber (2020) [38] | Differentiated native Europeans and migrants as to their risk of having schizophrenia | Schizophrenia | 370 patients with diagnosed schizophrenia spectrum disorder | Hospital data from 1982 to 2016 |
Khatua (2021) [40] | Using social media data to identify the voices of migrants and refugees and analyze their MH concerns | General MH | 0.15 million tweets, 2% from self-identified refugees | 0.15 million tweets |
Liu (2021) [43] | Used ML algorithms to distinguish ADHD, depression, anxiety, autism, intellectual disabilities, speech/language disorder, delays in development, and oppositional defiant disorder in Blacks using the data from their genome. | ADHD, depression, anxiety, autism, intellectual disabilities, speech/language disorder, delays in development, oppositional defiant disorder | 4179 Black individuals | Genomic sequencing data |
Liu (2021) [42] | Used ML algorithms to distinguish ADHD in Blacks using the data from their genome. | ADHD | 524 Black individuals | Genomic sequencing data |