Jonathan Marsh
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Jonathan Marsh has a total of 52 co-authors. Here are the top co-authors by number of shared publications:
- Jordan E. DeVylder
Silver School of Social Work New York University New York New York USA
6 shared publications - Christophe Lemey
URCI Mental Health Department, Brest Medical University Hospital, Brest, France
3 shared publications - Deok-Hee Kim-Dufor
Limics, Sorbonne Université, Université Sorbonne Paris-Nord, INSERM, Paris, France
3 shared publications - Michel Walter
Fondation FondaMental, 94000 Créteil, France
3 shared publications - Sofian Berrouiguet
Mental Health Department, Brest Medical University Hospital, Brest, France
3 shared publications - Aziliz Le Glaz
URCI Mental Health Department, Brest Medical University Hospital, Brest, France
2 shared publications
Latest Publications
- 29 July 2026
- Published by Wiley in Child & Family Social Work
How Credible Messenger Peer Mentors Support Young People Entering and Aging Out of Out‐of‐Home Care
Research shows that young people who experience out‐of‐home care (OOHC) are at heightened risk for poor socio‐economic and health outcomes as they age out of care in early adulthood. EnlightenMENT is a credible messenger peer mentoring programme that was proposed by a council of young adults with lived experiences in the foster care system as a way to connect young people in OOHC with trained peer mentors that have similar lived experiences. This study aimed to understand the strategies and approaches of these peer mentors in engaging and supporting young people who participated in EnlightenMENT. Mentors ( N = 12) who were employed in EnlightenMENT were recruited to participate in individual, semi‐structured interviews. Thematic analysis was used to identify and extract themes. The findings showed that credible messenger peer mentors can connect with young people in OOHC in unique and profound ways that distinguish their support from that provided by clinical professionals, case managers and other professionals who interact with young people in care. We describe the strategies and approaches that mentors use to promote young people's healthy transitions into adulthood. Implications for practice surrounding credible messenger peer mentoring for young people in OOHC are discussed.
- 29 July 2026
- Published by Wiley in Child & Family Social Work
How Credible Messenger Peer Mentors Support Young People Entering and Aging Out of Out‐of‐Home Care
Research shows that young people who experience out‐of‐home care (OOHC) are at heightened risk for poor socio‐economic and health outcomes as they age out of care in early adulthood. EnlightenMENT is a credible messenger peer mentoring programme that was proposed by a council of young adults with lived experiences in the foster care system as a way to connect young people in OOHC with trained peer mentors that have similar lived experiences. This study aimed to understand the strategies and approaches of these peer mentors in engaging and supporting young people who participated in EnlightenMENT. Mentors ( N = 12) who were employed in EnlightenMENT were recruited to participate in individual, semi‐structured interviews. Thematic analysis was used to identify and extract themes. The findings showed that credible messenger peer mentors can connect with young people in OOHC in unique and profound ways that distinguish their support from that provided by clinical professionals, case managers and other professionals who interact with young people in care. We describe the strategies and approaches that mentors use to promote young people's healthy transitions into adulthood. Implications for practice surrounding credible messenger peer mentoring for young people in OOHC are discussed.
- 25 July 2024
- Published by SAGE Publications in International Social Work
- Vol. 67 (6) , 1411-1428
A narrative review of the international literature on social work within the military
There is a dearth of understanding about the international literature on military social work with active service members. This narrative review spans the literature between 1944 and 2019 in an effort to grow the international military social work knowledge base. Of the 92 countries searched, 24 produced at least one publication about military social work with active military service members. Major themes identified among the articles reviewed included: the general scope of practice, practice in mental health and trauma, practice with special populations, practice with families, interdisciplinary practice and collaboration, ethical issues and dilemmas, and education and training.
- 25 July 2024
- Published by SAGE Publications in International Social Work
- Vol. 67 (6) , 1411-1428
A narrative review of the international literature on social work within the military
There is a dearth of understanding about the international literature on military social work with active service members. This narrative review spans the literature between 1944 and 2019 in an effort to grow the international military social work knowledge base. Of the 92 countries searched, 24 produced at least one publication about military social work with active military service members. Major themes identified among the articles reviewed included: the general scope of practice, practice in mental health and trauma, practice with special populations, practice with families, interdisciplinary practice and collaboration, ethical issues and dilemmas, and education and training.
- Open Access
Psychosis Caused by a Somatic Condition: How to Make the Diagnosis? A Systematic Literature Review
Background: First episode of psychosis (FEP) is a clinical condition that usually occurs during adolescence or early adulthood and is often a sign of a future psychiatric disease. However, these symptoms are not specific, and psychosis can be caused by a physical disease in at least 5% of cases. Timely detection of these diseases, the first signs of which may appear in childhood, is of particular importance, as a curable treatment exists in most cases. However, there is no consensus in academic societies to offer recommendations for a comprehensive medical assessment to eliminate somatic causes. Methods: We conducted a systematic literature search using a two-fold research strategy to: (1) identify physical diseases that can be differentially diagnosed for psychosis; and (2) determine the paraclinical exams allowing us to exclude these pathologies. Results: We identified 85 articles describing the autoimmune, metabolic, neurologic, infectious, and genetic differential diagnoses of psychosis. Clinical presentations are described, and a complete list of laboratory and imaging features required to identify and confirm these diseases is provided. Conclusion: This systematic review shows that most differential diagnoses of psychosis should be considered in the case of a FEP and could be identified by providing a systematic checkup with a laboratory test that includes ammonemia, antinuclear and anti-NMDA antibodies, and HIV testing; brain magnetic resonance imaging and lumbar puncture should be considered according to the clinical presentation. Genetic research could be of interest to patients presenting with physical or developmental symptoms associated with psychiatric manifestations.
Open AccessPsychosis Caused by a Somatic Condition: How to Make the Diagnosis? A Systematic Literature Review
Background: First episode of psychosis (FEP) is a clinical condition that usually occurs during adolescence or early adulthood and is often a sign of a future psychiatric disease. However, these symptoms are not specific, and psychosis can be caused by a physical disease in at least 5% of cases. Timely detection of these diseases, the first signs of which may appear in childhood, is of particular importance, as a curable treatment exists in most cases. However, there is no consensus in academic societies to offer recommendations for a comprehensive medical assessment to eliminate somatic causes. Methods: We conducted a systematic literature search using a two-fold research strategy to: (1) identify physical diseases that can be differentially diagnosed for psychosis; and (2) determine the paraclinical exams allowing us to exclude these pathologies. Results: We identified 85 articles describing the autoimmune, metabolic, neurologic, infectious, and genetic differential diagnoses of psychosis. Clinical presentations are described, and a complete list of laboratory and imaging features required to identify and confirm these diseases is provided. Conclusion: This systematic review shows that most differential diagnoses of psychosis should be considered in the case of a FEP and could be identified by providing a systematic checkup with a laboratory test that includes ammonemia, antinuclear and anti-NMDA antibodies, and HIV testing; brain magnetic resonance imaging and lumbar puncture should be considered according to the clinical presentation. Genetic research could be of interest to patients presenting with physical or developmental symptoms associated with psychiatric manifestations.
- 18 November 2022
- Published by Oxford University Press (OUP) in Schizophrenia Bulletin
- Vol. 49 (2) , 385-396
Ethnoracial Variation in Risk for Psychotic Experiences
Background & Hypothesis: Psychotic disorders are inequitably distributed by race in the United States, although it is not known whether this is due to assessment biases or inequitable distributions of risk factors. Psychotic experiences are subclinical hallucinations and delusions used to study the etiology of psychosis, which are based on self-report and therefore not subject to potential clinician biases. In this study, we test whether the prevalence of psychotic experiences (PE) varies by race and if this variance is explained by socioenvironmental risk factors.Study Design: Data on demographics, PE, and socioenvironmental risk factors were collected through the National Survey of Poly-victimization and Mental Health, a national probability sample of US young adults. Logistic regression analyses were used to determine whether PE prevalence varied by race/ethnicity and, if so, whether this was attenuated with inclusion of indicators of income, education, urban/rural living, discrimination, and trauma exposure.Study Results: Black and Hispanic respondents reported PE at significantly greater rates than White or “other” ethnoracial groups, with hallucinations more commonly reported by Hispanic respondents. PE were significantly associated with police violence exposure, discrimination, adverse childhood experiences, and educational attainment. These factors statistically explained ethnoracial differences in the likelihood of overall PE occurrence and of nearly all PE subtypes.Conclusions: Previously observed racial differences in psychosis extend beyond clinical schizophrenia, and therefore, are unlikely to be explained entirely by clinician biases. Instead, racial disparities in PE appear to be driven by features of structural racism, trauma, and discrimination.
- 18 November 2022
- Published by Oxford University Press (OUP) in Schizophrenia Bulletin
- Vol. 49 (2) , 385-396
Ethnoracial Variation in Risk for Psychotic Experiences
Background & Hypothesis: Psychotic disorders are inequitably distributed by race in the United States, although it is not known whether this is due to assessment biases or inequitable distributions of risk factors. Psychotic experiences are subclinical hallucinations and delusions used to study the etiology of psychosis, which are based on self-report and therefore not subject to potential clinician biases. In this study, we test whether the prevalence of psychotic experiences (PE) varies by race and if this variance is explained by socioenvironmental risk factors.Study Design: Data on demographics, PE, and socioenvironmental risk factors were collected through the National Survey of Poly-victimization and Mental Health, a national probability sample of US young adults. Logistic regression analyses were used to determine whether PE prevalence varied by race/ethnicity and, if so, whether this was attenuated with inclusion of indicators of income, education, urban/rural living, discrimination, and trauma exposure.Study Results: Black and Hispanic respondents reported PE at significantly greater rates than White or “other” ethnoracial groups, with hallucinations more commonly reported by Hispanic respondents. PE were significantly associated with police violence exposure, discrimination, adverse childhood experiences, and educational attainment. These factors statistically explained ethnoracial differences in the likelihood of overall PE occurrence and of nearly all PE subtypes.Conclusions: Previously observed racial differences in psychosis extend beyond clinical schizophrenia, and therefore, are unlikely to be explained entirely by clinician biases. Instead, racial disparities in PE appear to be driven by features of structural racism, trauma, and discrimination.
- 12 June 2022
- Published by Taylor & Francis in Journal of American College Health
- Vol. 72 (5) , 1350-1354
COVID-19 factors and self-injurious behaviors among US college students: findings from the healthy minds study 2020
The COVID-19 pandemic has dramatically altered the way of life in the United States, which may be linked to self-injurious behaviors.We conducted a secondary analysis of data from the Fall 2020 Coh...
- 12 June 2022
- Published by Taylor & Francis in Journal of American College Health
- Vol. 72 (5) , 1350-1354
COVID-19 factors and self-injurious behaviors among US college students: findings from the healthy minds study 2020
The COVID-19 pandemic has dramatically altered the way of life in the United States, which may be linked to self-injurious behaviors.We conducted a secondary analysis of data from the Fall 2020 Coh...
- 1 May 2022
- Published by Elsevier in Schizophrenia Research
- Vol. 243, 449-450
Population density is unrelated to prevalence of psychotic experience across 4 U.S. cities
- 1 May 2022
- Published by Elsevier in Schizophrenia Research
- Vol. 243, 449-450
Population density is unrelated to prevalence of psychotic experience across 4 U.S. cities
- Open Access
- 28 May 2021
- Published by Elsevier in Biological Psychiatry Global Open Science
- Vol. 1 (4) , 310-316
COVID-19 Infection and Psychotic Experiences: Findings From the Healthy Minds Study 2020
- 28 May 2021
- Published by Elsevier in Biological Psychiatry Global Open Science
- Vol. 1 (4) , 310-316
Open AccessCOVID-19 Infection and Psychotic Experiences: Findings From the Healthy Minds Study 2020
- Top Cited PapersOpen Access
- 4 May 2021
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 23 (5) , e15708
Machine Learning and Natural Language Processing in Mental Health: Systematic Review
Background: Machine learning systems are part of the field of artificial intelligence that automatically learn models from data to make better decisions. Natural language processing (NLP), by using corpora and learning approaches, provides good performance in statistical tasks, such as text classification or sentiment mining. Objective: The primary aim of this systematic review was to summarize and characterize, in methodological and technical terms, studies that used machine learning and NLP techniques for mental health. The secondary aim was to consider the potential use of these methods in mental health clinical practice Methods: This systematic review follows the PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis) guidelines and is registered with PROSPERO (Prospective Register of Systematic Reviews; number CRD42019107376). The search was conducted using 4 medical databases (PubMed, Scopus, ScienceDirect, and PsycINFO) with the following keywords: machine learning, data mining, psychiatry, mental health, and mental disorder. The exclusion criteria were as follows: languages other than English, anonymization process, case studies, conference papers, and reviews. No limitations on publication dates were imposed. Results: A total of 327 articles were identified, of which 269 (82.3%) were excluded and 58 (17.7%) were included in the review. The results were organized through a qualitative perspective. Although studies had heterogeneous topics and methods, some themes emerged. Population studies could be grouped into 3 categories: patients included in medical databases, patients who came to the emergency room, and social media users. The main objectives were to extract symptoms, classify severity of illness, compare therapy effectiveness, provide psychopathological clues, and challenge the current nosography. Medical records and social media were the 2 major data sources. With regard to the methods used, preprocessing used the standard methods of NLP and unique identifier extraction dedicated to medical texts. Efficient classifiers were preferred rather than transparent functioning classifiers. Python was the most frequently used platform. Conclusions: Machine learning and NLP models have been highly topical issues in medicine in recent years and may be considered a new paradigm in medical research. However, these processes tend to confirm clinical hypotheses rather than developing entirely new information, and only one major category of the population (ie, social media users) is an imprecise cohort. Moreover, some language-specific features can improve the performance of NLP methods, and their extension to other languages should be more closely investigated. However, machine learning and NLP techniques provide useful information from unexplored data (ie, patients’ daily habits that are usually inaccessible to care providers). Before considering It as an additional tool of mental health care, ethical issues remain and should be discussed in a timely manner. Machine learning and NLP methods may offer multiple perspectives in mental health research but should also be considered as tools to support clinical practice.
- 4 May 2021
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 23 (5) , e15708
Top Cited PapersOpen AccessMachine Learning and Natural Language Processing in Mental Health: Systematic Review
Background: Machine learning systems are part of the field of artificial intelligence that automatically learn models from data to make better decisions. Natural language processing (NLP), by using corpora and learning approaches, provides good performance in statistical tasks, such as text classification or sentiment mining. Objective: The primary aim of this systematic review was to summarize and characterize, in methodological and technical terms, studies that used machine learning and NLP techniques for mental health. The secondary aim was to consider the potential use of these methods in mental health clinical practice Methods: This systematic review follows the PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis) guidelines and is registered with PROSPERO (Prospective Register of Systematic Reviews; number CRD42019107376). The search was conducted using 4 medical databases (PubMed, Scopus, ScienceDirect, and PsycINFO) with the following keywords: machine learning, data mining, psychiatry, mental health, and mental disorder. The exclusion criteria were as follows: languages other than English, anonymization process, case studies, conference papers, and reviews. No limitations on publication dates were imposed. Results: A total of 327 articles were identified, of which 269 (82.3%) were excluded and 58 (17.7%) were included in the review. The results were organized through a qualitative perspective. Although studies had heterogeneous topics and methods, some themes emerged. Population studies could be grouped into 3 categories: patients included in medical databases, patients who came to the emergency room, and social media users. The main objectives were to extract symptoms, classify severity of illness, compare therapy effectiveness, provide psychopathological clues, and challenge the current nosography. Medical records and social media were the 2 major data sources. With regard to the methods used, preprocessing used the standard methods of NLP and unique identifier extraction dedicated to medical texts. Efficient classifiers were preferred rather than transparent functioning classifiers. Python was the most frequently used platform. Conclusions: Machine learning and NLP models have been highly topical issues in medicine in recent years and may be considered a new paradigm in medical research. However, these processes tend to confirm clinical hypotheses rather than developing entirely new information, and only one major category of the population (ie, social media users) is an imprecise cohort. Moreover, some language-specific features can improve the performance of NLP methods, and their extension to other languages should be more closely investigated. However, machine learning and NLP techniques provide useful information from unexplored data (ie, patients’ daily habits that are usually inaccessible to care providers). Before considering It as an additional tool of mental health care, ethical issues remain and should be discussed in a timely manner. Machine learning and NLP methods may offer multiple perspectives in mental health research but should also be considered as tools to support clinical practice.
- 5 April 2021
- Published by Taylor & Francis in Psychosis
- Vol. 14 (1) , 57-69
Violence exposure, psychotic experiences, and social disconnection in an urban community sample
Background: Psychotic experiences (PEs) occur in the general population, serving as indicators of vulnerability to psychotic disorders and other mental and physical concerns. Environmental stressors, including violence exposure and social disconnection are associated with PEs, although the exact nature of these relations is unknown.Methods: Data were drawn from an online community sample of 1,000 adults (476 male) residing in Baltimore or New York City. Binary logistic regression and mediation analyses were used to examine associations between PEs, violence exposure (adverse childhood experiences; intimate partner violence), and social disconnectedness (personal; neighborhood).Results: Personal social disconnectedness was significantly associated with delusional mood (p < .001). Neighborhood social disconnectedness was significantly associated with delusional mood (p < .01), delusions of reference and persecution (p < .001), delusions of control (p < .01), and hallucinations (p < .01). Both subtypes of violence exposure were found to have significant direct effects on PEs (p < .001). Personal social disconnectedness accounted for 6.72-7.87% of the total effect of violence exposure on PEs, whereas neighborhood social disconnectedness was found to account for 5.14-6.74% of the total effect.Conclusion: The results advance understandings of prominent environmental risks for psychosis, and offer new implications for public health consideration of psychosis.
- 5 April 2021
- Published by Taylor & Francis in Psychosis
- Vol. 14 (1) , 57-69
Violence exposure, psychotic experiences, and social disconnection in an urban community sample
Background: Psychotic experiences (PEs) occur in the general population, serving as indicators of vulnerability to psychotic disorders and other mental and physical concerns. Environmental stressors, including violence exposure and social disconnection are associated with PEs, although the exact nature of these relations is unknown.Methods: Data were drawn from an online community sample of 1,000 adults (476 male) residing in Baltimore or New York City. Binary logistic regression and mediation analyses were used to examine associations between PEs, violence exposure (adverse childhood experiences; intimate partner violence), and social disconnectedness (personal; neighborhood).Results: Personal social disconnectedness was significantly associated with delusional mood (p < .001). Neighborhood social disconnectedness was significantly associated with delusional mood (p < .01), delusions of reference and persecution (p < .001), delusions of control (p < .01), and hallucinations (p < .01). Both subtypes of violence exposure were found to have significant direct effects on PEs (p < .001). Personal social disconnectedness accounted for 6.72-7.87% of the total effect of violence exposure on PEs, whereas neighborhood social disconnectedness was found to account for 5.14-6.74% of the total effect.Conclusion: The results advance understandings of prominent environmental risks for psychosis, and offer new implications for public health consideration of psychosis.
- Open Access
- 31 July 2019
- Published by JMIR Publications Inc. in JMIR Preprints
Machine Learning and Natural Language Processing in Mental Health: Systematic Review (Preprint)
BACKGROUND Machine learning systems are part of the field of artificial intelligence that automatically learn models from data to make better decisions. Natural language processing (NLP), by using corpora and learning approaches, provides good performance in statistical tasks, such as text classification or sentiment mining. OBJECTIVE The primary aim of this systematic review was to summarize and characterize, in methodological and technical terms, studies that used machine learning and NLP techniques for mental health. The secondary aim was to consider the potential use of these methods in mental health clinical practice METHODS This systematic review follows the PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis) guidelines and is registered with PROSPERO (Prospective Register of Systematic Reviews; number CRD42019107376). The search was conducted using 4 medical databases (PubMed, Scopus, ScienceDirect, and PsycINFO) with the following keywords: machine learning, data mining, psychiatry, mental health, and mental disorder. The exclusion criteria were as follows: languages other than English, anonymization process, case studies, conference papers, and reviews. No limitations on publication dates were imposed. RESULTS A total of 327 articles were identified, of which 269 (82.3%) were excluded and 58 (17.7%) were included in the review. The results were organized through a qualitative perspective. Although studies had heterogeneous topics and methods, some themes emerged. Population studies could be grouped into 3 categories: patients included in medical databases, patients who came to the emergency room, and social media users. The main objectives were to extract symptoms, classify severity of illness, compare therapy effectiveness, provide psychopathological clues, and challenge the current nosography. Medical records and social media were the 2 major data sources. With regard to the methods used, preprocessing used the standard methods of NLP and unique identifier extraction dedicated to medical texts. Efficient classifiers were preferred rather than transparent functioning classifiers. Python was the most frequently used platform. CONCLUSIONS Machine learning and NLP models have been highly topical issues in medicine in recent years and may be considered a new paradigm in medical research. However, these processes tend to confirm clinical hypotheses rather than developing entirely new information, and only one major category of the population (ie, social media users) is an imprecise cohort. Moreover, some language-specific features can improve the performance of NLP methods, and their extension to other languages should be more closely investigated. However, machine learning and NLP techniques provide useful information from unexplored data (ie, patients’ daily habits that are usually inaccessible to care providers). Before considering It as an additional tool of mental health care, ethical issues remain and should be discussed in a timely manner. Machine learning and NLP methods may offer multiple perspectives in mental health research but should also be considered as tools to support clinical practice.
- 31 July 2019
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessMachine Learning and Natural Language Processing in Mental Health: Systematic Review (Preprint)
BACKGROUND Machine learning systems are part of the field of artificial intelligence that automatically learn models from data to make better decisions. Natural language processing (NLP), by using corpora and learning approaches, provides good performance in statistical tasks, such as text classification or sentiment mining. OBJECTIVE The primary aim of this systematic review was to summarize and characterize, in methodological and technical terms, studies that used machine learning and NLP techniques for mental health. The secondary aim was to consider the potential use of these methods in mental health clinical practice METHODS This systematic review follows the PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis) guidelines and is registered with PROSPERO (Prospective Register of Systematic Reviews; number CRD42019107376). The search was conducted using 4 medical databases (PubMed, Scopus, ScienceDirect, and PsycINFO) with the following keywords: machine learning, data mining, psychiatry, mental health, and mental disorder. The exclusion criteria were as follows: languages other than English, anonymization process, case studies, conference papers, and reviews. No limitations on publication dates were imposed. RESULTS A total of 327 articles were identified, of which 269 (82.3%) were excluded and 58 (17.7%) were included in the review. The results were organized through a qualitative perspective. Although studies had heterogeneous topics and methods, some themes emerged. Population studies could be grouped into 3 categories: patients included in medical databases, patients who came to the emergency room, and social media users. The main objectives were to extract symptoms, classify severity of illness, compare therapy effectiveness, provide psychopathological clues, and challenge the current nosography. Medical records and social media were the 2 major data sources. With regard to the methods used, preprocessing used the standard methods of NLP and unique identifier extraction dedicated to medical texts. Efficient classifiers were preferred rather than transparent functioning classifiers. Python was the most frequently used platform. CONCLUSIONS Machine learning and NLP models have been highly topical issues in medicine in recent years and may be considered a new paradigm in medical research. However, these processes tend to confirm clinical hypotheses rather than developing entirely new information, and only one major category of the population (ie, social media users) is an imprecise cohort. Moreover, some language-specific features can improve the performance of NLP methods, and their extension to other languages should be more closely investigated. However, machine learning and NLP techniques provide useful information from unexplored data (ie, patients’ daily habits that are usually inaccessible to care providers). Before considering It as an additional tool of mental health care, ethical issues remain and should be discussed in a timely manner. Machine learning and NLP methods may offer multiple perspectives in mental health research but should also be considered as tools to support clinical practice.