Linking anonymised GP records to other datasets (for example, on housing or collected by schools) can reveal important insight into how circumstances interact and influence our health. At the International Population Data Linkage Network conference, between 7-9 September, CEG researchers presented three projects that use innovative data linkage methods to address important research questions that have broad implications for policy.
Image: Freepik
Levels of household overcrowding are estimated using the Government’s English Housing Survey, but this does not give insight into the characteristics of the people who live there. Using CEG’s ASSIGN algorithm, Marta Wilk’s study uses pseudonymised UPRNs (Unique Property Reference Numbers) to link demographic information from GP records to property information from the Energy Performance Certificate (EPC) dataset. Together, the data reports the characteristics of the occupants and number of rooms for each property, providing a new estimate of household level overcrowding and revealing who is most likely to be affected – without disclosing addresses or identities to the researchers.
Marta successfully linked more than one million records from across north east London, finding that people who are most likely to live in overcrowded households are: Living in three-generational households (41% overcrowded), of South Asian ethnic background (29% overcrowded); Living in neighbourhoods that are in the top two quintiles for deprivation (26% overcrowded).View the abstract
Childhood obesity is a major public health concern in England, where more than a quarter of children start primary school with a BMI considered overweight or obese. Nicola Firman’s study uses pseudonymised NHS numbers to link data from the National Child Measurement Programme, conducted by local authorities, with the health records of 60,000 11-year-olds who took part in the programme. The aim was to ascertain whether obesity is a risk factor for knee pain in childhood. Nicola’s analysis found that, in general, boys were more likely to experience knee pain than girls. At age 11, obesity was a risk factor for knee pain among girls, but not boys. This could be explained by the relationship between sex and participation in physical activity - previous studies have shown that boys are more likely to participate in vigorous physical activity and are more likely to experience injuries in comparison to girls.
Routine linkage of National Child Measurement Programme data to GP health records could inform GPs of their patient’s recent weight status without the need for further measurements during a consultation.View the abstract
In London, coverage of the Measles, Mumps and Rubella (MMR) vaccine at 24 months of age is around 82%, and it is following a downward trend. The World Health Organization recommends 95% coverage to prevent outbreaks of disease. Understanding which children are missing out on this important protection is key to tackling the problem.
Using CEG’s ASSIGN algorithm, Milena’s study uses pseudonymised Unique Property Reference Numbers (UPRNs) to link the health records of children who share a household – without addresses or identities being disclosed to the researchers. Milena’s analysis found that children who are late in receiving the MMR are significantly more likely to live with other children whose MMR was delayed. She also found that children are less likely to receive the MMR on time if they live in a three-generational or single-adult household, or in households with four or more children.
The findings will enable more targeted interventions by the NHS and local authorities, and could make the case for home visits from health visitors to ensure each child has a healthy start in life. View the abstract