IntroductionSynthetic data is an emerging tool for health and social care researchers which can be used to expedite research processes and enhance the safety of personal data handling. Although data owners across the UK and globally currently make synthetic datasets available, public awareness and perceptions of its use for research remain underexplored.ObjectivesWe aimed to agree a set of recommendations for data owners on producing and releasing synthetic data and communicating with the UK public about this topic.MethodsForty-four public members were invited to four deliberative workshops to co-create a set of recommendations for data owners. Notes and transcripts from workshops 1-3, along with public member survey feedback, were reviewed to identify key topics.ResultsThirty-seven individuals contributed. We developed ten recommendations spanning five areas: introducing synthetic data; explaining its purpose; creating datasets; managing access, use, and misuse; and clear communication. These were iteratively refined with public input during the final workshop.ConclusionsThis study offers a unique insight into public attitudes towards synthetic data use in health and administrative research. The recommendations developed are intended to support UK data owners in releasing synthetic data responsibly, aligning with public expectations and fostering trust in its use as a research innovation tool. Future research should continue to assess evolving public attitudes to both low and high-fidelity data as its use expands.

IntroductionMonitoring and addressing health inequalities is important. However, socioeconomic variables are usually unavailable within health datasets. Area deprivation measures provide access to open-source reliable socioeconomic data within low/middle-income countries and can contribute to the monitoring of the Sustainable Development Goals and assessing the growing burden of health inequalities.ObjectiveTo create a small-area deprivation measure for the whole of Brazil - the Brazilian Deprivation Index (Índice Brasileiro de Privação - IBP),MethodsUsing Census Sector data (mean population size=615) from the most recently available Brazilian Demographic Census (2010), variables measuring literacy, household income and housing conditions were standardised using z-scores and summed into a single measure. The IBP was validated using regional small-area measures of vulnerability: Belo Horizonte’s Health Vulnerability Index (IVS) and São Paulo’s Social Vulnerability Index (IPVS). Mortality data from Minas Gerais were used to estimate age-standardised mortality rates (ASMR) by ill-defined causes across IBP deprivation quintiles.ResultsThe IBP was created for 303,218 (97.8%) census sectors (99.7% population). Substantial regional variation in deprivation was found using the IBP measure, with higher deprivation in rural than urban areas. The IBP was correlated with the other indicators used for validation: the IVS (r = 0.96) and the IPVS (r = 0.68). We found gradients across the ill-defined causes ASMR, in Minas Gerais mortality was 2.6 higher in the most deprived quintile of IBP, compared with the least deprived. Main challenges in creating a deprivation measure for LMICs and possible solutions are demonstrated.ConclusionA small area deprivation index was created for Brazil, a large and highly diverse middle-income country. The IBP improves our understanding and monitoring of inequalities, serving as a valuable tool for informing targeted public policies. Although the index is based on Brazil’s specific context, the challenges faced, and the strategies implemented to tackle them are relevant for other low- and middle-income countries aiming to develop similar tools.

Introduction & BackgroundSupermarket shopping data is an emerging data source for researchers, which offers valuable insights into consumer behaviours, as well as health and lifestyle choices. Despite growing interest in its use for health research, much remains to be explored. A key challenge in using shopping transactions in research is obtaining informed consent and little is currently known about the sociodemographic characteristics or attitudes which influence an individual’s willingness donate their data. Understanding these is crucial for improving recruitment, retention and ensuring more representative datasets. Objectives & ApproachThe core objectives of this research are to establish the likelihood of public participation in shopping data research and identify which sociodemographic factors influence willingness to donate shopping data for research purposes, as well as explore potential motivations and concerns related to this. This study uses a mixed-methods approach and harnesses data collected from the Avon Longitudinal Study of Patents and Children (ALSPAC). Alongside demographic variables, our analysis is based on two primary sources: a 2018 survey addressing hypothetical willingness to consent and a 2023 survey in which participants were asked to donate their shopping data. Questions about loyalty card use, perceptions about acceptability of data-sharing and perspectives around privacy concerns were also analysed. Our analysis considers factors including sex, ethnicity, educational attainment and employment status as potential influencers of consent. We deploy a range of statistical methods including regression modelling and correlation tests. Relevance to Digital FootprintsThis project contributes to advancing our understanding of applying data linkage between digital footprints data and cohort studies inclusively and accessibly, providing an essential foundation for future research on important and wide-ranging health and societal challenges. Results3768 participants responded to the 2018 survey. In 2023, 511 participants provided consent to collect their shopping data (8.3% of those invited). Of these, over two-thirds (67.1%) were female, 38.9% indicated that their highest level of educational qualification was at Level 4 or above and 76.3% reported that they were in full or part-time work. 68.6% of participants indicated ownership of at least one loyalty card - Tesco (47.5%) and Boots (41.6%) were the most commonly owned. In 2018, most participants (60.7%) expressed willingness to share their supermarket data. There was a significant association between gender for both willingness to consent (X2 = 18.3, p<0.001) and actual consent (X2 = 13.2, p<0.001). There was no significant association with educational attainment, employment status or ethnicity. Willingness to consent may predict actual consent, although there are no significant demographic predictors. Conclusions & ImplicationsOur findings from will help to inform inclusive consent practices in digital footprints research, as well as identify potential barriers and motivators to consent, and contribute towards current discourse on the ethical applications of shopping data.

Introduction & BackgroundThe study of loneliness and social networks has largely involved social media analysis for both social circle insights and loneliness predictions from language. While geospatial data have also been used to investigate loneliness, the approach has predominantly involved a traditional investigation of geolocations per se. Few studies have focused on the factors contributing to the success of human relationships and the associated loneliness reduction – which was the objective of this project. Objectives & ApproachThe data were obtained from B:friend – a wellbeing charity dedicated to tackling loneliness through befriending schemes for Older Adults – and joined with the Index of Multiple Deprivation (2019) for feature engineering. The objective was to identify features which contributed to a successful friendship. The dataset included neodemographic information: geospatial and demographic data for both sides of each pair (the Older Adult (Older Neighbour) and the paired Volunteer); as well as preference features for a potential pairing. Machine learning classification models (logistic regression, decision trees, random forest, support vector machines and xgboost) were used to determine which features contributed to successful short- and long-term befriending schemes. A total of 805 pairings were analysed. Relevance to Digital FootprintsThe project used geospatial information paired with demographic records collected by B:friend digitally to identify the features which contributed to a successful friendship. ResultsThe preliminary results offer insight into the features which contribute to short-and long-term friendships, with variables such as house suitability, third COVID-19 lockdown and income decile difference driving the predictions. On average, a third of the friendships appear to succeed. Additionally, we use a series of variable importance measures to rank order variables which contribute to successful befriending of the pairings. Conclusions & ImplicationsCombining demographic and geospatial information for machine learning helps identify important factors which contribute to short- and long-term relationships, and, by extension, to tackling loneliness. This novel approach can benefit the public, as well as professionals, by offering tailored suggestions for loneliness interventions, utilising factors contributing to the specific requirements of various individuals, as well as their preferences, thus further expanding on the pre-established findings such as homophily.

Introduction & BackgroundOne of the main challenges in the field of digital footprints is bridging the communication gap between researchers and the public, whose participation in sharing transaction data is essential. To address this, we have created an engaging data visualisation tool that demonstrates the value of loyalty card data, with the aim for this to be developed into a public facing application, sparking awareness and discussions around research using digital footprint data for public good. Objectives & ApproachThis visualization project emerged from a long-standing collaboration with We The Curious science centre in Bristol, UK, which hosted an interactive shopping experience within the Open City Lab. The exhibition sparked public discussions about data donation for the public good. Feedback from the exhibition revealed that many people were unaware they could access their personal data and did not understand its potential value for public health research. We developed the first prototype tool using Streamlit, a python-based package and included four main themes of aggregated data based on these discussions with the public (Spending, Location, Product Categories and Price Changes). Relevance to Digital FootprintsThis application showcases the power of digital footprint data in an interactive and interpretable way for members of the public. This is vital for future digital footprint research as the publics data is essential for ongoing research into public health. ResultsFollowing the creation of the initial prototype, a workshop was hosted with 10 participants. During the two-hour session, participants provided feedback on the application, and a 'Data Hazards Workshop’ was conducted. The main feedback was collected through a structured questionnaire. The questionnaire results indicated that the overall user experience was very high, though improvements could be made by improving the overall navigation of individual graphs. All feedback gained throughout this workshop and the conference will be implemented into the future version of the application. Conclusions & ImplicationsOur future aim is to continue to develop the utility of this tool and make it more widely accessible, to enhance public understanding, build trust, and foster ongoing participation in sharing shopping data for health research.

ObjectivesTo investigate the link between permanent school exclusion and perpetration of (i) serious violence (ii) homicide or near-miss homicide in young people. MethodsWe used linked national education and justice data to emulate a randomised controlled trial. Individuals aged 11 to 16 years enrolled in a state secondary school in England who were permanently excluded were matched 1:1 to an unexcluded control (meeting the same eligibility criteria). Follow up started on the date of exclusion for the matched pair and ended at the outcome of interest or after 12 months, whichever occurred first. Cox proportional hazards regression was used to compare outcomes in the two groups. We used a fixed bias analysis as a sensitivity analysis to unmeasured confounding. ResultsAmong the approximately 650,000 individuals meeting eligibility criteria there were just over 21,000 exclusions with complete data on covariates. 98% of these were matched to an eligible control. Of those excluded, 4.8% had a conviction or caution for serious violence within a year, compared to 2.4% among those not excluded (adjusted hazard ratio (HR)=2.05, 95% confidence interval (CI) 1.83 to 2.29). Homicide or near-miss homicide was much rarer (<0.1% of those excluded), with adjusted HR=2.36, 95% CI 1.04 to 5.36. The fixed bias analysis indicated that an unmeasured confounder would have to be strongly related to both the exposure and outcome in order to fully explain these associations. ConclusionPermanent exclusion is associated with approximately a doubling of the risk of being convicted or cautioned for serious violence and for homicide or near-miss homicide in the following 12-months. Since we cannot determine whether these associations are causal, targeting the upstream common causes of exclusion and violence should be prioritised.

ObjectivesTo identify the factors influencing nurses’ geographic relocation within their first year after registration as they seek employment. MethodIn England and Wales, 630,000 nurses are listed on the professional register maintained by the Nursing and Midwifery Council. This register has been linked to the 2021 census for England and Wales, creating a unique dataset that integrates geographic, social, and professional data. The register indicates the length of time since initial registration, while migration is determined based on census responses detailing current area of residence (as of 21 March 2021) and location 12 months prior. The census also provides data on occupation and employment status, including full-time and part-time work. ResultsPatterns of nurse migration within the first year of practice in England and Wales will be analyzed and compared with those registered for 2–5, 6–10, and more than 10 years. Geographic variations in in-migration and out-migration rates will be mapped at the Government Office Region level. Logistic regression models will identify factors predicting relocation for employment among newly registered nurses, compared to those who remain in place. Employment patterns (full-time vs. part-time), caregiving responsibilities, housing tenure, as well as other socio-demographic and professional factors will be examined to assess influences on mobility, including life-course stage, within the nursing workforce. ConclusionNurse relocation may contribute to localized workforce shortages. By understanding the factors that encourage stability or drive mobility, we can enhance workforce planning, funding allocations, and nursing education strategies. Policy recommendations will be provided to support recruitment and retention efforts, which are critical to delivering the NHS Long-term Plan.

Evidence on refugee integration outcomes in the UK is lacking, partly due to an absence of datasets which permit refugees to be identified. The RIO longitudinal cohort study designed in collaboration with the Home Office and MHCLG aims to address this by linking various administrative data longitudinally. RIO covers cohorts granted asylum and those who resettled in the UK via several resettlement schemes between 2015 and 2023. Administrative datasets linked to the RIO cohorts include the Personal Demographic Service (PDS), Home Office Borders and Immigration (HOBI) data, Births, Deaths and 2021 Census. Our latest iteration linked Pay as you earn (PAYE), Energy performance certificate (EPC) and Valuation Office Agency (VOA). Alongside improvements to our deterministic and associative linkage algorithms, we’ve now incorporated Ministry of Justice’s Splink coding package to enable us to link via probabilistic methods and further improve our linkage rates and quality. Experimental analysis has looked at social and economic outcomes for these refugee cohorts. Linkage to NHS data helps us understand access to health services and time taken to access these services once resettled. Our publication due for release in April 2025 provides us with a rich understanding and evidence of how linked administrative and Census data can bridge the gap in understanding of how refugees integrate into the labour market and housing systems in the UK up to 6 years after their arrival or grant of asylum. We also break these insights down into key characteristics to not only show how integration varies by asylum and refugee cohorts but also by age, sex and geographical region. RIO will ultimately help inform local authorities, government, charities and other organisations with resource allocation for these vulnerable populations. We are planning to make this dataset available to Accredited Researchers via the ONS Secure Research Service (SRS) and the Integrated Data Service (IDS).

ObjectivesPrevious research has established an association between diagnosis of a neurodevelopmental condition and poorer academic outcomes. Many education policies advocate for early diagnosis and special education needs (SEN) support, however little is known about the association between timing of SEN identification and educational outcomes. MethodsTo address this gap, we linked Welsh administrative education data to primary and secondary health care records to identify the age at which a child (i) receives a neurodevelopmental diagnosis and (ii) receives SEN support. Using data from over 90,000 children across 5 cohorts, we conducted logistic and linear regressions to examine the association between SEN identification and children’s attainment, exclusions, and attendance at age 16. In the most recent cohort, we show that timing of SEN identification significantly associated with attainment, exclusions, and attendance. ResultsChildren identified for SEN in their first year of school were eight times less likely (than those without SEN) to meet national attainment expectations at age 16, whereas those identified at age 16 were three times less likely, X2= 41.7, p<.001). The opposite pattern held true for exclusions and attendance: those identified in their first year were two times more likely to be excluded, compared to five times more likely for those identified in their final year X2= 7.3, p=.007). Those identified in first year were 1.6 times more likely to be persistently absent (missing >10% of sessions), while those identified in their final year were 2.5 times more likely X2= 8.9, p=.003). We show how these associations vary by type of need and timing of health diagnosis. ConclusionEarly identified SEN pupils are less likely to achieve national attainment expectations. Late identification is more strongly associated with absence and exclusion issues. Variation by SEN classification and timing of diagnosis are important considerations for the new Additional learning Needs framework in Wales, and wider health and education policy and practice.

ObjectivesPolicies designed to support the sustainability of farming are multiple, although developed with little knowledge about the intricacies of farming households. To inform policy and support strategies, more information is needed about the families involved, including their characteristics, economic circumstances and resilience. MethodsDe-identified data on farm households from the Administrative Data | Agricultural Research Collection (AD|ARC) is utilised within the SAIL Databank. AD|ARC contains characteristics of the farmer, farm household (all children and adults living in the household) and farm business, for all members of households in Wales living at an address for which a farm business received a subsidy payment in 2010. The population comprises approximately 18,000 farm businesses and linked farm households in Wales, and a comparison group of non-farming rural households. The demographic, household and self-reported health characteristics of individuals in the two groups was compared. ResultsDescriptive and bivariate statistics will be used to describe farm households in Wales in terms of the farmer (e.g. farmer’s age, gender, ethnicity, education), farm household (e.g. household members age, gender, ethnicity, education, employment, income, household composition, total income, deprivation), and farm business (e.g. number of farmers and non-farmers, employed workers, agricultural training, farm activity, area of land, total subsidy payment) characteristics. The profile of farming households in Wales will then be compared with those of the rural control group held in the AD|ARC data set. Through bivariate and multivariate statistics and generalised linear models, the analysis will highlight characteristics specific to the farming community or reflective of rural communities in general. ConclusionThis research investigates the characteristics, economic circumstances and resilience of farming households in Wales. The findings have important implications for agricultural policy, and for health and other services, and will aid policy makers and farming stakeholders in anticipating the diverse impacts of policies across the farming community.