National Pupil Database: linkage accuracy using bespoke cohorts
Linkage of multiple datasets to the same administrative dataset enables assessment of study characteristics associated with linkage error. This study compares characteristics of linked and unlinked children within three cohorts and the evaluation of these linked databases can help better assessing the linkage accuracy.
To evaluate potential biases in the linked cohorts which can be affected by the consent process, completeness of identifiers, and methods for capturing data
Datasets and name of the Government departments
More about the project
Linkage of a range of different cohorts within the ADRC-E offers the opportunity to evaluate potential biases in the linked cohorts as a function of the consent process, completeness of identifiers, and methods for capturing data. This study compares characteristics of linked and unlinked children within 3 cohorts: i) the Millennium cohort study (a British birth cohort); ii) the CRIS database (Clinical record interactive search), which involves unconsented linkage of all mental health service data for 4 local authorities (1.2 million residents) to National Pupil Database and Hospital Episode Statistics (NHS Digital); iii) the CRANE database (a national audit of children with cleft lip or palate in England, collected by specialist services).
ADRC-E Bloomsbury investigators have been involved in linkage of three databases to the National Pupil Database. This study, led by ADRC-E in collaboration with CLOSER and Millenium Cohort Study, CRANE and CRIS, involves an evaluation of the linked databases and assessment of linkage accuracy.
Dr Efrosini Setakis, ADRC-E and UCL Institute of Education
Dr Johnny Downs, King's College London
Dr Tarek Mostafa, UCL Institute of Education
Prof Jan van der Meulen, ADRC-E and London School of Hygiene and Tropical Medicine
Prof Lorraine Dearden, ADRC-E and UCL Institute of Education
Prof Ruth Gilbert, ADRC-E and Great Ormond Street Institute of Child Health, UCL