The spine of KEHC will consist of linked Office for National Statistics (ONS) birth [28] and death registration [29] and National Health Service (NHS) birth notification data [30] on all children born in England since 1^st January 2006 until December 2023 at first build. This spine will then be linked to administrative health and education data from multiple data providers as well as small-area level (property and postcode) environmental data (Table 2). Children will be linked to their mother’s hospital and address records up to a year before their birth to allow maternal environmental exposures and health during pregnancy to be taken into account; children will also be linked to their siblings in the cohort via their mother’s records. Finally, for a subset of children born between 2010 and 2012, and between 2020 and 2022, their records will additionally be linked to their mothers’ 2011 or 2021 Census records, respectively.

With approximately 570,000 live births occurring in England annually, the total cohort will include data for around 10.5 million children born between 2006 and 2023 (the first build of the database). In England, regardless of where a birth occurs, parents are legally required to register their child’s birth within 42 days. Therefore, the cohort will include births occurring at home, at sea, on military bases or in a hospital. The cohort will also include stillbirths which occur at an rate of 3.9 per 1,000 births [14]. Children whose mothers are not resident in England (i.e., visitors or residents in other UK countries) will be included in the cohort but are unlikely to have any linked data available. Additionally, children whose parents have opted out of sharing their NHS data beyond NHS England through the National Opt-Out Programme will not have linked health or education data available, as this data is provided by NHS England. Similarly, women who have opted out through the National Data Opt-Out Programme will not have their records included. Furthermore, NHS England will exclude children with child protection flags in the Personal Demographics Service (PDS). Children will be followed longitudinally via linked administrative datasets from birth until death, emigration from England (as indicated in the child’s PDS record), or the latest health or education data available.

Figure 1 shows the expected follow-up from birth available for cohort children in a Lexis diagram. We anticipate that KEHC will be updated annually to capture new births each year and to incorporate the most recent health and education data for children already in the cohort. Ethical approval and a legal basis have been secured for conducting these annual updates. However, some time lags are expected due to the release schedules of data providers and logistical considerations in updating the dataset.

A diagram of data flows is shown in the Appendix Figure 1. Please note the linkage processes described below is subject to agreement by each data provider.

ONS will extract identifiers for mothers and children from the live and stillbirth registration and NHS birth notification records (Table 2, Dataset 3), including mother and child NHS numbers, child’s sex and date of birth, the residential postcode of the mother/child at delivery/birth, child’s name and mother’s names (forename, surname, and maiden name). ONS will link the birth registration data (Table 2, Dataset 1) to mortality data for liveborn children (Table 2, Dataset 2), using the NHS number, child’s name, date of birth and sex. The de-identified birth notification-registration attribute data and linked mortality records will be transferred into a secure ONS Trusted Research Environment (TRE) with individual cohort ID numbers to delineate each mother and child (live or stillborn) in the cohort. The linkage between birth data and mortality data is done to capture all instances of death, including infant deaths under 1 year of age, as well as deaths occurring later in childhood. This ensures a comprehensive record of mortality for the entire cohort, allowing us to study patterns and factors associated with early and later childhood mortality.

Individual mother IDs will be derived by ONS based on the mother’s NHS number, date of birth, and postcode at delivery to allow mothers, children, and siblings to be linked in KEHC. Identifiers for mothers and liveborn children (NHS numbers, postcode at birth/delivery and child sex and date of birth) together with cohort IDs will be securely transferred by ONS to NHS England.

NHS England will link the KECH identifiers to children’s health data, including data from HES Admitted Patient Care and A&E/Emergency Care (see Dataset 5 in Table 2), Mental Health Services (Table 2, Dataset 7), Maternity Services (see Dataset 8 in Table 2), Community Dispensing (Table 2, Dataset 6), and the PDS (Table 2, Dataset 4) using an established, deterministic linkage algorithm (the Master Person Service - MPS) [46]. NHS England will also link children in KEHC to their mother’s HES Admitted Patient Care data up to one year before birth.

The de-identified attribute health data linked to the mother and child cohort IDs will be returned to ONS and placed in the ONS TRE. PDS addresses for mothers and children will be returned to the ONS. Linkage between PDS (Table 2, Dataset 4) and NPD (National Pupil Database) (Table 2, Dataset 10), developed for the ECHILD database [47] is based on a deterministic algorithm using name, date of birth, and chronological postcode, with a linkage rate of 95% [47]. NHS England will use the established PDS-NPD link to extract anonymised pupil reference numbers (aPRNs; the individual pupil ID in the NPD) for all children in the cohort. The aPRNs will be forwarded to ONS for onward linkage to NPD attribute data via the Department for Education (DfE).

ONS holds PDS records from 2016 onwards. Upon agreement to this process, ONS will link the KEHC spine to the ONS-held part of PDS and join these PDS records with those received from NHS England, generating longitudinal address records for the whole cohort of children and their mothers up to a year before the child’s birth. ONS will derive Unique Property Reference Numbers (UPRNs), a unique identifier which is assigned to every addressable location across Great Britain [48], and extract postcodes, all with timestamps, and store the longitudinal UPRNs and postcodes from PDS in the Data Access Platform (DAP), ONS’s internal secure data linkage platform, This will allow further linkages to environmental data. ONS will also link mothers’ address at delivery from birth registration to UPRNs, and store these in the DAP for researchers wanting to link environmental data to the birth dwelling only, and for validation of PDS address histories.

ONS will link the cohort to postcode-level data on air pollution (Table 2, Dataset 12) and temperature (Table 2, Dataset 15), and UPRN-level data on distance to roads (Table 2, Dataset 13), traffic flow (Table 2, Dataset 13), greenspace coverage and access (Table 2, Dataset 17), Energy Performance Certificates (EPCs) (Table 2, Dataset 14), and Valuation Office Agency data (Table 2, Dataset 16) using the longitudinal postcode and UPRN PDS histories, respectively. These UPRN-linked environmental data will be developed by the KEHC study team over the course of the project. ONS will also link the longitudinal PDS UPRNs to Index of Multiple Deprivation quintiles (a lower super output area index of deprivation) [49] and government office region indicator. All linked postcode- and dwelling -level environmental and deprivation datasets will be placed in the ONS TRE.

ONS will link the mothers’ identifiers (forenames, surnames, maiden names, dates of birth, and postcodes) from the birth records to the mothers’ 2011 and 2021 Census records (Table 2, Dataset 9) using deterministic and probabilistic linkage algorithms. Only registrations for births between March 2010 and March 2012, and March 2020 and March 2022 will be linked to the 2011 and 2021 Census, respectively. Once Census linkage is complete, the de-identified Census attribute data for the mothers in KEHC will be released to the ONS TRE. ONS will assess and provide a report on the quality of linkage required to construct the KEHC spine and link the spine to Census data.

ONS will forward the aPRNs received from NHS England to DfE who will extract the attribute NPD records for the children in KEHC who are old enough to have an aPRN. DfE will return the de-identified education data to the ONS TRE for inclusion into the KEHC. ONS will link the unique school IDs in NPD (Table 2, Dataset 10) to current and historical school address information published by DfE (Table 2, Dataset 11). The school addresses and school IDs will be held separately by ONS in the DAP. Researchers may request linkage of environmental data to KEHC via the school locations or school IDs to examine the impact of environments near schools on child health and education outcomes.

We will work with ONS to clean and set up the KEHC for access by external researchers. We will document all data cleaning steps, which will be described in the KEHC user guide, and any received or derived variables will be described in the KEHC data dictionary. All code used to clean the data will be made openly available on GitHub. Although the highly sensitive data in KEHC cannot be made openly available, we will strive to make the KEHC data resource adhere to Findable, Accessible Interoperable Reusability (FAIR) principles [50], supporting open science values.

The legal bases for KEHC data include the Digital Economy Act 2017 and Section 251 of the NHS Act 2006 (for processing of confidential patient information). Under the General Data Protection Regulations (GDPR), the legal bases are 6(1)e for public task and 9(2)j for archiving, research, and statistics, specifically for processing special category data. We have obtained approvals from London – City & East NHS Research Ethics Committee, National Statistician’s Data Ethics Committee and the Confidentiality Advisory Group (23/CAG/0060) to establish the KEHC as a research database. We are seeking approvals from Advisory Group for Data (AGD; NHS England), Data Sharing and Approval Panel (DSAP; DfE) and Research Accreditation Panel (UK Statistics Authority; UKSA).