Archives

Natural Language Processing: Interrogating free text in mental healthcare records to capture experiences of violence

    Violence can be categorised in a variety of ways for example physical, sexual, emotional, and domestic but all cause significant physical and mental morbidity within general populations. Individuals with a severe mental illness have been found to be significantly more likely to experience domestic, physical, and sexual violence compared to the general population. For these individuals, experiences of violence are important risk factors however, this is not routinely collected by mental health services.

    In general data on all forms of violence has been inadequately available from healthcare records. This is partly due to the lack of routine enquiry by professionals at points of clinical contact, and partly because instances of violence are difficult to identify in healthcare data in the absence of specific coding systems.

    A general challenge for using health records data for research is that the most valuable and granular information is frequently contained in text fields (e.g., routine case notes, clinical correspondence) rather than in pre-structured fields; this includes mentions of violence whether experienced as a victim or perpetrated. Capturing violence experiences across mental healthcare settings can be challenging because most instances are likely to be recorded as unstructured text data. Therefore, natural language processing (NLP), is increasingly in use to extract information automatically from unstructured text in electronic health records, particularly in mental healthcare, on clinical entities.

    Dr Ava Mason from Kings College London and VISION researchers Professor Robert Stewart, Dr Angus Roberts, Dr Lifang Li, and Dr Vishal Bhavsar worked with colleagues to apply NLP across different clinical samples to investigate mentions of violence. They ascertained recorded violence victimisation from the records of 60,021 patients receiving care from a large south London NHS mental healthcare provider during 2019. Descriptive and regression analyses were conducted to investigate variation by age, sex, ethnic group, and diagnostic category.

    Results showed that patients with a mood disorder, personality disorder, schizophrenia spectrum disorder or PTSD had a significantly increased likelihood of victimisation compared to those with other mental health diagnoses. Additionally, patients from minority ethnic groups for Black and Asian had significantly higher likelihood of recorded violence victimisation compared to White groups. Males were significantly less likely to have reported recorded violence victimisation than females.

    The researchers demonstrated the successful deployment of machine learning based NLP algorithms to ascertain important entities for outcome prediction in mental healthcare. The observed distributions highlight which sex, ethnicity and diagnostic groups had more records of violence victimisation. Further development of these algorithms could usefully capture broader experiences, such as differentiating more efficiently between witnessed, perpetrated and experienced violence and broader violence experiences like emotional abuse.

    To download the paper: Frontiers | Applying neural network algorithms to ascertain reported experiences of violence in routine mental healthcare records and distributions of reports by diagnosis

    To cite: Mason AJC, Bhavsar V, Botelle R, Chandran D, Li L, Mascio A, Sanyal J, Kadra-Scalzo G, Roberts A, Williams M, Stewart R. Applying neural network algorithms to ascertain reported experiences of violence in routine mental healthcare records and distributions of reports by diagnosis. Frontiers in Psychiatry 2024 Sep 10. doi:103389/fpsyt.2024.1181739

    Illustration from Adobe Photo Stock subscription

    Centring otherness with migrant women affected by domestic abuse

      Victims-survivors with insecure immigration status in the UK are subject to complexities that limit their access to safety, support, and justice. While campaigners have been advocating for more equitable pathways for provision and support over the years, migrant women continue to navigate hostile environments characterised by dehumanising language and anti-migrant bureaucratic systems.

      This chapter, written by VISION researcher Dr Olumide Adisa for the book, Otherness in Communication Research: Perspectives in Media, Interpersonal, and Intercultural Communication, reports on how a feminist dialogic approach (characterised by open, inclusive dialogue and a foundational understanding of social, economic, and political equality for women) was used to centre the often ‘silent voices’ of migrant women affected by domestic abuse.

      Feminist dialogical approach acknowledges the complexities that characterise the migrant victim’s journey through the system—the relationship between the self-other, in a peculiar hostile environment which views the other as a ‘threat’. Migrant women continue to endure this othering within agencies as they seek safety and support. For example, some professionals conflating ‘foreignness’ with ‘insecure immigration statuses’, when confronted with difference. This theorisation of self and other lends itself to a social justice-oriented practice.

      Using different art forms (co-produced with migrant women) and purposeful conversations, attendees were able to encounter migrant women as not a distant ‘other’ whom ‘we’ observe and theorise but as equal partners in the creating and reshaping on knowledge systems on safety, support, and justice.

      This chapter draws on quotes from survivors to funnel through a hopeful lingering over otherness that positions migrant women as deserving of consideration and care, and considers empowering aspects about the other that may often be dismissed in professional circles, but nonetheless are important as a protective element of a safety net.

      To download the chapter: Centring Otherness with Migrant Women Affected by Domestic Abuse | SpringerLink

      To cite: Adisa, O. (2025). Centring Otherness with Migrant Women Affected by Domestic Abuse. In: Magalhaes, L. (eds) Otherness in Communication Research. Palgrave Studies in Otherness and Communication. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-031-73788-6_16

      For further information, please contact Olumide at olumide.adisa@city.ac.uk

      Illustration from Adobe Photo Stock subscription

      Addressing abuse in teenage relationships

        This VISION Policy Brief highlights emerging findings and policy recommendations from ongoing research and stakeholder engagement into abuse in teenage relationships carried out by the UKPRP VISION consortium.

        Abuse—whether physical, emotional, or sexual—within young people’s relationships is often overlooked in both research and policy. The Crime Survey for England and Wales (CSEW) finds that young women aged 16 to 19 are more likely to experience domestic abuse than any other age group. Despite this high prevalence, this age group is less likely to be referred to support services. Furthermore, the CSEW does not cover individuals under the age of 16, leaving a major gap in understanding of prevalence.

        Key findings:

        • Lack of consistent terminology and recognition – various terms are used to describe abuse in teenage relationships, including ‘teen dating violence’, ‘adolescent domestic abuse’, ‘teenage relationship abuse’ and ‘youth intimate partner violence’. Both the workshop with young people and the roundtables identified that young people generally do not associate the behaviours they experience with any of these terms and are more likely to use language like ‘toxic relationships’.
        • Very limited UK research on risk and protective factors for under 16s – our rapid review found that in the last 10 years there was only one UK academic study that looked into risk and protective factors for abuse in teenage relationships for those aged under 16.
        • Importance of schools and communities – unlike adult domestic abuse, which is largely experienced in private, abuse experienced in teenage relationships is more likely to occur outside of the home, especially within schools.
        • Very difficult to measure extent of issue – due to the current Home Office definition of domestic abuse there is very limited and consistently recorded administrative data collected on those under 16 who are experiencing abuse.
        • Need to take a more radical review of systems – our discussion highlighted the difficulty of addressing abuse in teenage relationships within the current systems.

        Recommendations for change:

        • Develop a national strategy – prevention and early intervention
        • Explore support for young people – victims and those carrying out harmful behaviours
        • Commission research into under 16s – including those with lived experience and taking a whole systems approach
        • Improve measurement in under 16s
        • Agree terminology and produce an associated education programme

        To download the policy briefing: VISION Policy Brief: Addressing Abuse in Teenage Relationships

        To cite: Weir, Ruth; Barrow-Grint, Katy (2025). VISION Policy Brief: Addressing Abuse in Teenage Relationships. City, University of London. Report. https://doi.org/10.25383/city.26539906.v1

        For further information, please contact: Ruth at ruth.weir@city.ac.uk

        Intimate partner violence impacts affected by relationship status and offence type

          Intimate partner violence and abuse has a detrimental impact on victim-survivors’ health and wellbeing. However, intimate partners include a range of different relationship types, which are rarely differentiated or contrasted in research. In this paper, VISION researchers, Dr Niels Blom and Dr Polina Obolenskaya, investigate with Dr Jessica Phoenix and Merili Pullerits, whether different types of intimate partners commit different types of violence/abuse and whether the injury and wellbeing impact on victim-survivors varies by intimate partner relationship type.

          They estimate models for victim-survivors’ emotional impact and injuries using the Crime Survey for England and Wales (2001–2020). Intimate partner relationships are differentiated into four groups (current versus former partner, and spouses/partners versus boy/girlfriends). Violence and abuse are grouped into physical violence/abuse, sexual violence/abuse, threats, and economic crimes.

          The team found that for both men and women, offences committed by current partners are more likely to involve physical violence/abuse than offences by former partners. Ordered logit models indicate that female victim-survivors of physical violence/abuse or economic crimes experience more severe emotional impacts when the perpetrator is their current or former spouse/partner compared to a current or former boy/girlfriend. Women’s risk of injuries from physical violence and economic offences are higher when committed by current compared to former partners. Few differences are identified for men’s emotional impact and injuries.

          The type of intimate partner relationship is associated with type of violence/abuse experienced, and for women, with the resulting emotional impact and injury. Future research and policies aimed at reducing harms from intimate partner violence and abuse and supporting victim-survivors should therefore consider distinctions in relationships to deliver more targeted interventions.

          To download the paper: Physical and Emotional Impacts of Intimate Partner Violence and Abuse: Distinctions by Relationship Status and Offence Type | Journal of Family Violence

          To cite: Blom, N., Obolenskaya, P., Phoenix, J. and M. Pullerits. Physical and Emotional Impacts of Intimate Partner Violence and Abuse: Distinctions by Relationship Status and Offence Type. J Fam Viol (2024). https://doi.org/10.1007/s10896-024-00786-w

          For further information, please contact Niels at niels.blom@manchester.ac.uk

          Photograph from Adobe Photo Stock subscription

          Deputy Chief Constable awarded Practitioner in Residence at Violence and Society Centre

            Katy Barrow-Grint, Deputy Chief Constable, Gloucestershire

            City St George’s, UoL, offers a Practitioner in Residence programme at the School for Policy and Global Affairs. It is for mid-level and senior policy practitioners within the UK and provides a platform to grow and explore their practice in partnership with the school.

            Katy Barrow-Grint, Deputy Chief Constable in Gloucestershire and an executive leader in national policing, became aware of the opportunity via her work with VISION Senior Research Fellow, Dr Ruth Weir,  on the VISION adolescent domestic abuse (ADA) research programme. Having recently written a book entitled ‘Policing Domestic Abuse’ with Ruth and others, the research identified a national gap academically and in policing with how ADA is understood.

            Katy’s focus will be on how police constabularies document ADA and developing a better understanding of the impact of the statutory age limitations on the practical work police officers do on the front line.

            Forces do not routinely record ADA as the statutory guidance states that domestic abuse occurs in relationships where both parties are aged 16 or over. As a result, whilst crimes against young people will be recorded and investigated, they are not necessarily classified as domestic abuse, and it may be that child protection, domestic abuse or front-line response teams deal with the case.

            Her project work will seek to understand how forces are recording such incidents, and what type of officer and role is investigating. Katy will work with policing nationally through the National Police Chief‘s Council (NPCC) domestic abuse and child protection portfolios and collate an up-to-date picture across all forces in England and Wales to understand how they are recording and who is investigating ADA.

            Katy is also undertaking specific localised work in Oxfordshire, Gloucestershire and Northumbria, hosting roundtables with Dr Ruth Weir and  practitioners from all relevant agencies to gain a qualitative understanding of the problems staff encounter when dealing with ADA.

            Photograph from Adobe Photo Stock subscription

            VISION-funded research: Surviving Economic Abuse survey initial findings released

              Tackling economic abuse should be part of the solution to meet the new government’s ambitious target to halve violence against women and girls in a decade. It is important that the government’s measurement approach can understand the range of ways that economic restriction, exploitation and sabotage that victim-survivors experience at scale across the UK.

              Recent survey results from Surviving Economic Abuse (SEA) tell a powerful story that highlights experiences of economic abuse across the UK. The full report will be launched by SEA in March 2025, but their early release of key findings include:

              • Economic abuse is often understood to only be about creating dependency through restriction, but it can take many forms e.g., having a partner or ex-partner steal money, refuse to pay bills, or scare their partner into taking out credit. Early analysis suggests that a wider range of behaviours may continue post-separation than previously thought.
              • The data shines a light on the dangerous situation for young women- an area that SEA and VISION are seeking funding to explore further. 18–24-year-olds experienced more economic abuse than any other age group, for example 12% of this sample had been prevented from having log-in information (e.g. passwords, usernames) to key accounts such as online banking, utilities accounts, emails by a partner or ex-partner compared to 4% of all women.
              • Black, Asian and racially minoritised women in the UK may be more than twice as likely to experience economic abuse from a partner or ex-partner than White women, with women with a Black/African/Caribbean or Black British ethnicity particularly at risk.
              • Disabled women in the UK may be nearly twice as likely to experience economic abuse from a partner or ex-partner as non-disabled women

              The VISION consortium was delighted to financially support SEA’s research, A rapid impact survey to monitor the nature and prevalence of economic abuse in the UK, through our Small Projects Fund in spring 2024. Their full report will be widely shared in 2025, including on the VISION website and through our networks.

              Photograph from Adobe Photo Stock subscription

              Migrants’ experiences of violence while in insecure migration status

                Violence is a major public health issue. Moreover, there is evidence that violence is significantly related to social inequality. Existing studies have found links between violence and gender, ethnicity, place of residence and socioeconomic status.

                Although economic globalization impacts trade, goods, and services, the movement of people has been increasingly restricted since the 1990s. The number of people globally who live with insecure migration status is difficult to estimate, but includes people worldwide undertaking irregular journeys and crossing international borders without authorization, people living without the correct immigration documentation, and people in temporary or dependent statuses in destination countries.

                The global movement of people in the context of strict immigration laws and policies places significant numbers of people in insecure migration status worldwide. Insecure status leaves people without recourse to legal, governmental or social protection from violence and abuse.

                This review synthesized qualitative studies that reported how migrants associated physical and physically enforced sexual violence they experienced with their insecure migration status. VISION researchers, Andri Innes, Annie Bunce, Hannah Manzur, and Natalia V. Lewis, generated robust qualitative evidence showing that women experienced sexual violence while in transit or without status in a host state, and that they associated that violence with their insecure migration status. This was the case across the various geographic routes and destination countries.

                They found evidence that women associated intimate partner violence with lacking (legal) access to support because of their insecure migration status. Women connected their unwillingness to leave violent circumstances, and therefore their prolonged or repeated exposure to violence, with a fear of immigration removal produced by their insecure migration status.

                To protect people in insecure migration status from experiencing violence that they associated with their migration status, it’s necessary to ensure that the reporting of violence does not lead to immigration enforcement consequences for the victim.

                To download the paper: Experiences of violence while in insecure migration status: a qualitative evidence synthesis | Globalization and Health | Full Text

                To cite: Innes, A., Bunce, A., Manzur, H. et al. Experiences of violence while in insecure migration status: a qualitative evidence synthesis. Global Health 20, 83 (2024). https://doi.org/10.1186/s12992-024-01085-1

                For further information, please contact Andri at alexandria.innes@city.ac.uk

                Photograph from Adobe Photo Stock subscription

                VISION/VASC Webinar Series: Into the Light Index

                  We are pleased to announce our next webinar for the VISION and Violence & Society Centre (VASC) Webinar Series on Tuesday, 21 January, 1100 – 1150.

                  Deborah Fry, Director of Data at Childlight – Global Child Safety Institute and Professor of International Child Protection Research at University of Edinburgh, will present on the Into the Light Index, published last year on the prevalence of technology-facilitated child sexual exploitation and abuse. She will also discuss some of the measurement challenges in this field and how they are documenting and exploring those challenges.

                  Professor Fry undertakes primary research to measure the magnitude, drivers and consequences of violence against children, barriers and enablers to appropriate prevention and response systems including in school settings and the effectiveness of existing interventions.

                  She leads the data division at Childlight – Global Child Safety Institute. The Data Institute, funded by the Human Dignity Foundation, aims to take a data driven, evidence-based approach to understanding the prevalence of child sexual exploitation and abuse across the globe and translating that data into sustainable action that safeguards children. The mission is to establish a world leading independent institute that gathers, translates and visualises the prevalence of child sexual exploitation and abuse across the world.  

                  To register for the event in order to receive the Teams invitation, please contact: VISION_Management_Team@city.ac.uk

                  The purpose of the VISION/VASC webinar series is to provide a platform for academia, government and the voluntary and community sector that work to reduce and prevent violence to present their work / research to a wider audience. This is a multidisciplinary platform and we welcome speakers from across a variety of fields such as health, crime, policing, ethnicity, migration, sociology, social work, primary care, front line services, etc. If interested in presenting at a future Series webinar, please contact: VISION_Management_Team@city.ac.uk

                  This webinar series is sponsored by the UK Prevention and Research Partnership consortium, Violence, Health and Society (VISION; MR-V049879) and the Violence and Society Centre at City St George’s, University of London.

                  Discovering the Potential of Large Language Models in Social Science Research: Takeaways from an Oxford Workshop

                    By Dr Maddy Janickyj, Research Fellow in Natural Language Processing (NLP) for the Violence, Health, and Society (VISION) Consortium, University College London

                    As a data-focused VISION researcher with a PhD specialising in Natural Language Processing (NLP; see our previous blog for more about this), I initially avoided ChatGPT and similar tools. ChatGPT, a type of Large Language Model (LLM) developed by OpenAI, offers capabilities like summarising information, translating text, and even coding.

                    While ChatGPT is potentially the most well-known example of a LLM, similar models are integrated into many everyday tools. For instance, LLMs are the underlying technology in many customer service chatbots, virtual assistants like Alexa, and writing tools such as Grammarly. These LLMs are trained on large sets of data with the intention of getting them to understand (and in some cases generate) language. The models draw on this training to complete various tasks and are finetuned to work for specific domains. Their breadth of abilities and the many open-source models that have been developed make them the perfect methodological tool for researchers in both computer science and the social sciences. For clarity, an open-source LLM is one whose code and architecture are publicly available.

                    To further understand how LLMs are being used by researchers and to consider how the tools would integrate with and support violence-related research, I – a mathematician turned computational social scientist – attended the Oxford LLMs workshop. The event, held at Oxford’s Nuffield College, aimed to bring early-career scholars up to speed with the technical foundations, real-world applications, and research potential of LLMs. Throughout the week, I met with PhD/Masters students and other Post-doctoral researchers interested in using LLMs to evaluate anything from economic, linguistic, and political issues, for example.

                    Understanding LLMs: Lectures and Industry Insights

                    The first few days provided foundational lectures and talks, showcasing the technical underpinnings and application of LLMs. One of the big draws was the calibre of speakers. We heard from industry experts working at well-known companies such as Meta, Ori , Qdrant, Wayfair, Intento, Arize AI, and Google.  

                    We then started our deep dive into LLMs, including how they are trained and evaluated. We heard about the numerous ways you can fine-tune LLMs, a step which occurs after general pre-training and tailors a model to meet domain/task-specific needs. Fine-tuning methods such as Continued Pre-training, Supervised Fine-tuning, and Preference Tuning were highlighted. Each technique offers different ways of adapting LLMs to specialised domains without needing to re-train them from scratch, saving computational resources.

                    We also covered common challenges associated with finetuning models. One of these is “catastrophic forgetting,” where a model’s performance declines in one area when it’s fine-tuned on another. For example, if a model is adjusted to improve name recognition, it may inadvertently lose accuracy in identifying locations. This side effect is something I encountered when finetuning other NLP models during my PhD and illustrates the balance required when refining LLMs.

                    Applying LLMs: Collaborative Research Projects

                    In the latter half of the week, workshop attendees collaborated on research projects, exploring LLM applications across social science realms. This was a hands-on opportunity to test LLM methodologies discussed earlier and apply them to real-world social science challenges.

                    Leading up to the workshop we had the chance to review the proposed project briefs, gather literature showcasing how LLMs are used in our respective disciplines, and finally rank the four projects according to our own skillsets and research interests. One of the projects we decided to tackle as a group focused on developing an LLM purely for social science research. LLMs are considered to have some form of bias, for example against certain demographic groups, and with this ongoing project, we wanted to create a fair, unbiased, and open-source LLM suited to the social sciences.

                    In another project, we examined gender bias in academia. For this, we used Google’s Gemini to classify the gender of authors in academic syllabi. By experimenting with prompts, we measured how well the LLM could assess gender trends in syllabus authorship. Using tools like Google Colab, we collaboratively coded and refined our approach, leveraging Gemini’s capabilities to highlight gender disparities effectively. In some cases, we found the model to correctly classify 100% of the authors’ genders. This project underscored both the potential and the limitations of LLMs in accurately capturing nuanced social issues.

                    Appreciating the Potential: Be Cautious

                    Overall, the Oxford workshop demonstrated how LLMs can be powerful tools in social science research including violence-related research such as what we do at VISION, provided they are tailored to specific domain needs and applied with caution. Hearing directly from researchers and industry professionals offered invaluable guidance on both leveraging and responsibly implementing LLMs. Its also important to consider the data you are utilizing and the outputs you are expecting. In my current area (which focuses on technology-facilitated abuse), an increasing number of researchers are using sensitive data and the outcomes of such research can impact the lives of real individuals. Thus, for anyone in the social sciences looking to integrate cutting-edge NLP methods, understanding the complexities behind these models and their applications is essential. I encourage readers to look at the work being done currently by the workshop participants, and to keep an eye out for later outputs of the workshop!

                    For further information, please contact Maddy at m.janickyj@ucl.ac.uk

                    Domestic abuse in cancer care: Improving the identification and support

                      Although few studies have explored people’s experiences of domestic abuse and cancer, we know the two co-occur. The few studies we do have show that cancer can trigger an escalation of abuse. But there are no published domestic abuse interventions in the cancer setting.

                      In an attempt to plug that gap, Sandi Dheensa, University of Bristol researcher, and colleagues, including VISION Deputy Director Estela Capelas Barbosa, have conducted a service evaluation on a domestic abuse intervention for hospital-based cancer professionals. Their study, Identifying and responding to domestic abuse in cancer care: a mixed methods service evaluation of a training and support intervention – European Journal of Oncology Nursing, is the first to evaluate a DA training (and support) intervention for cancer professionals in England.

                      The key findings demonstrate that cancer and DA frequently co-occur, and that training and support intervention of hospital-based cancer staff increased the rate of DA identifications. There is an appetite for DA and cancer training amongst hospital-based cancer staff.

                      The evaluation contributes further evidence of the benefit of hospital-based domestic abuse coordinator roles and contributes new evidence for the feasibility of adapting the role for a specific context.

                      To read the article or download the paper free of charge:

                      Identifying and responding to domestic abuse in cancer care: a mixed methods service evaluation of a training and support intervention – European Journal of Oncology Nursing

                      To cite:

                      Identifying and responding to domestic abuse in cancer care: a mixed methods service evaluation of a training and support intervention. Dheensa, Sandi et al. European Journal of Oncology Nursing, Volume 0, Issue 0, 102724

                      Or for further information, please contact Sandi at sandi.dheensa@bristol.ac.uk

                      Photograph from Adobe Photo Stock subscription