8  Linkage

8.1 Linkage between health and education data in ECHILD

The ECHILD linkage spine is generated by NHS England. NHS England receives identifiers from the DfE (name, date of birth, sex, postcode) and the aPMR. The DfE identifiers are linked with NHS identifiers, and are held separately from any health or education information (NHS England, 2023h). For each matched pair of identifiers from education and health, NHS England attaches a pseudonymised ID called a ‘Token Person ID’ (TPI) (NHS England, 2024d). The TPI is created specifically for ECHILD and cannot be used to identify individuals. Further detail on the linkage processes can be found in Appendix A Appendix A.

A bridging file, containing the linked aPMR and TPI, can be used to link data for individuals across health and education datasets within ECHILD.

Image showing overview of preocesses involved in the creation of the ECHILD database, including matching process. Figure 5: Overview of processes involved in the creation of the ECHILD database.

8.2 Mother-baby linkage in ECHILD

The ECHILD team have created a ‘mother-baby’ link (Feng et al., 2024a), whereby HES delivery and birth records for both mother and baby are linked together using a probabilistic linkage algorithm (Harron et al., 2016). In summary, the probabilistic linkage approach uses ‘indirect identifiers’ (such as gestational age and birth weight) to link together mother-baby dyads. Previous work has demonstrated high linkage rates using this approach (Feng et al., 2024b). The mother-baby linkage means that it is possible within ECHILD to look at how maternal characteristics (including those captured in health, education and social care datasets) are related to child outcomes. It is also possible to identify sibling clusters.