Identifying domestic abuse victims earlier could help to reduce future emergency medical admissions. Our new research shows how this could be done before a victim even involves the police by connecting information gathered by the police and hospitals.
Using mathematical models, we then identified the factors that increased the risk of negative outcomes, such as hospital and A&E admissions or death within 12 months of the PPN.
Our study showed that many victims of domestic abuse often visit accident and emergency departments before the police get involved. This means that healthcare professionals can play a crucial role in identifying and helping people who are experiencing abuse.
We combined data from the police with data from GPs and accident and emergency hospital admissions. We focused on residents in the South Wales Police catchment area who had experienced domestic abuse between August 2015 and March 2020, and who were given a public protection notification (PPN). This is a document that records safeguarding concerns about adults or children.
Connecting this data with health information gives a wide view of how domestic abuse affects people’s health. Health records are kept in a secure database called the Secure Anonymised Information Linkage (SAIL) Databank.
This provides access to different kinds of information, such as records from doctors’ visits, hospital stays, accident and emergency visits, and death records. All data in the databank is anonymous, ensuring that individuals cannot be identified.
Using mathematical models, we then identified the factors that increased the risk of negative outcomes, such as hospital and A&E admissions or death within 12 months of the PPN.
What we found
Of the 8,709 people who experienced domestic abuse, 71.8% were women. Within a year of experiencing abuse, 3,544 of the victims had negative outcomes, such as an A&E admission, while there were 48 deaths.
Victims who had conflicts related to child contact had a lower risk of experiencing negative outcomes, because the perpetrator might not be living with the victim.
We also found that certain factors increased the likelihood of negative outcomes. These included being younger, having multiple incidents of abuse, getting injured during the abuse, being assessed as high-risk, being referred to other agencies, having a history of violence, experiencing attempted strangulation, or being pregnant.
Pregnant victims, in particular, faced more risks, which affected their own health and the health of their babies. Certain factors like smoking, obstetric issues and taking specific medications (like antidepressants and antibiotics) increased the risk of having a negative outcome after experiencing domestic abuse.
By studying different patterns, we could predict how severe the cases of domestic abuse were in terms of risk. For example, victims who had frequent interactions with the police were at higher risk.
However, victims who had conflicts related to child contact had a lower risk of experiencing negative outcomes. This is because the perpetrator might not be living with the victim.
What are the implications?
Our findings show the importance of considering a victim’s health history in identifying domestic abuse. Identifying certain patterns could lead to earlier interventions.
It is crucial for different organisations to work together and share information to identify and help vulnerable individuals effectively. Identifying specific risk factors, like being younger or having a history of violence, could help identify victims more effectively.
This would include investigating previous visits to the hospital, conducting thorough assessments for pregnant victims who are at high risk and connecting different pieces of information.
These measures could help prevent further victimisation and ensure that people receive the right support and resources.
Our research highlights the importance of healthcare settings, especially emergency departments, in identifying and addressing domestic abuse. Training programmes could help emergency department staff identify potential cases of domestic abuse, even if the victim does not explicitly disclose the abuse.
By connecting different sources of information and identifying people at high risk, health professionals could take necessary actions and refer victims to support services.
By connecting different sources of information and identifying people at high risk, health professionals could take necessary actions and refer victims to support services.
Our study looked at situations where abuse was officially reported, so victims who did not report it were not included.
We did not include cases where women went to the emergency room for obstetric reasons either. This means that the impact of domestic abuse during pregnancy may be underestimated in our findings.
In future, further research should be undertaken to validate the findings of this study in different settings and populations. It would also be helpful to look at information from other sources, such as social services and housing records, to get a better picture of the factors that contribute to domestic abuse and its consequences.
Protecting privacy
While linking data from different organisations can be helpful for research, it is also important to protect people’s privacy. If we want to link data at a national level for purposes other than research, we would need a public consultation on what data is shared and to discuss how people’s privacy would be protected.
This is important because if people were afraid that their data would be shared with the police, they might not seek help from emergency services. When victims can be encouraged to talk, however, this study underlines the importance of training A&E staff to recognise and address potential cases of abuse.
This article first appeared on The Conversation, and is republished under a Creative Commons Licence; the original can be read here.
Dr Natasha Kennedy is a Senior Research Officer and Data Scientist for the National Centre for Population Health and Wellbeing Research at Swansea University. Her research interests lie in utilising statistics to analyse and model life course data such as social service, education, and health data.
Amrita Bandyopadhyay is a Research Officer and Data Scientist at the National Centre for Population Health and Wellbeing Research, Swansea University. Her research focuses on data-linkage, harmonisation and analysis of large-scale routine electronic healthcare and administrative data. Amrita works on statistical analysis and developing data-driven models using advanced machine-learning algorithms; she is an early-career researcher with several top quartile publications in epidemiology and public health, and leads investigations in public health projects.
