Evaluating the Family Nurse Partnership in England

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: − HES:Civil Registration (Deaths) bridge

Objective for processing

[32 paragraphs unchanged] To achieve this linkage, UCL are requesting a re-purposing of a separate [17 words unchanged] a current HES extract and sent by the research team to NHS Digital. England. [3 paragraphs unchanged] - UCL will send a list of HESIDS from initial FNP HES linkage to NHS Digital. England. - NHD Digital England will share identifiers with DfE. [1 paragraph unchanged] This was to allow UCL to prepare a list of HESIDs for the study cohort, from the initial FNP-HES linkage with the request being that NHS Digital England will share the identifiers with DfE to allow onward linkage to the NPD. This builds on previous activity, whereby NHS Digital England extracted the identifiers for the cohort of FNP participants, and performed a [53 words unchanged] control mothers) and the identifiers for this cohort were shared by NHS Digital England with DfE for onward linkage with the NPD. [1 paragraph unchanged] UCL have requested to use data for all mothers aged between 13-24 who delivered a baby between 01 April 2010 and 31 March 2017. 2019. UCL previously requested that FNP data for approximately 25,000 mothers who participated [41 words unchanged] now proposing to focus on educational attainment and social care histories, too). [2 paragraphs unchanged] For this purpose, the research team are requesting pseudonymised HES APC and A&E data for the year's corresponding to the data in the FNP cohort (births during 2010-2017, 2010-2019, with data on mothers’ previous health going back to April 2005 and [52 words unchanged] the research questions, and only request data for mothers in their cohort. [2 paragraphs unchanged] 1. FNP programme data for approximately 33,000 mothers and 34,000 children with the FNP study ID but no identifiers have been transferred from NHS Digital England to the UCL Data Safe Haven. [2 paragraphs unchanged] 1. A list of HESIDs for the wider HES-FNP cohort (approximately 1,000,000 mothers and 800,000 children) will be transferred to NHS Digital. England. This includes identifiers for the 33,000 FNP mothers and 34,000 FNP children; the remainder of the cohort are controls. 2. NHS Digital England will extract historical identifiers for the HES-FNP cohort list of HESIDs from PDS and share these identifiers with DfE for linkage with NPD. [12 paragraphs unchanged] 2022 Amendment: [1 paragraph unchanged] UCL request access to be given to 1 individual who is under honorary contract. Access is only permitted once the honorary contract has been signed by all required parties and meets NHS Digital’s standard on honorary contracts. The individual on the Honorary Contract will, in due course, be acting on the instructions of the applicant and not carrying out data controllership activities for the Health Foundation.

Processing activities

[1 paragraph unchanged] First, the FNP programme data will be linked with HES, by NHS Digital, England, which took place in May/June 2020. A link key (HES-FNP) was created [33 words unchanged] key will be created (HES-FNP-NPD) and transferred ONS Secure Research Service (SRS). [1 paragraph unchanged] - De-identified FNP programme data transferred from NHS Digital England with a study ID [1 paragraph unchanged] - The HES-FNP link key, transferred from NHS Digital England to the UCL Data Safe Haven. [6 paragraphs unchanged] 1. FNP programme data with the FNP study ID but no identifiers were transferred from NHS Digital England to the UCL Data Safe Haven. 2. Identifiers for FNP mothers and their first child are currently held by NHS Digital England (Open Exeter). These identifiers were checked against PDS to obtain identifier histories (e.g. previous and up to date addresses) within NHS Digital. England. 3. The FNP study ID, plus identifiers from NHS Digital England (NHS number, GP code, name, sex, date of birth, postcode), were made available for linkage within NHS Digital England (FNP programme data is held separately from the identifiers). [2 paragraphs unchanged] 2. NHS Digital England linked the FNP cohort with HES based on sex, date of birth, NHS number, name, postcode, and GP code. 3. A link-key was transferred from NHS Digital England to the UCL Data Safe Haven to allow merging of the FNP programme data with the existing HES data. 4. UCL prepared a list of HESIDs for the FNP-HES cohort and transferred these back to NHS Digital. England. 5. In the 2021 amendment, NHS Digital England extracted historical identifiers for the list of HESIDs (including name where available for FNP participants) and shared the identifiers with DfE for linkage with NPD [1 paragraph unchanged] 1. NHS Digital England will transfer identifiers for the FNP-HES cohort to DfE for linkage with NPD (no FNP programme data or HES data will be transferred). [6 paragraphs unchanged] All UCL students are expected to undertake annual training on handling highly confidential information. All Trainees and students register for and complete NHS Digital’s England’s Data Security Awareness (NHSD) course provided by e-Learning for Health. The course [7 words unchanged] to data security, breaches and incidents, and the General Data Protection Regulation. All students working on this study are students from UCL. All students working on the study will undertake the NHS Digitals England's Data Security Awareness course provided by e-Learning for Health. UCL has a [38 words unchanged] face potential sanctions in the event of a breach of the policy. [1 paragraph unchanged]

Expected output

Outputs updated as per latest annual confirmation report submitted 15/01/24 [4 paragraphs unchanged] Update under v3: Expected outputs: - The main results paper is expected to be published in early 2024, a secondary paper is expected to be published in mid 2024, and further papers are expected to be published in early 2025. - Findings have been submitted as a peer review publication to BMJ Public Health. Secondary papers will be submitted to Archives of Disease in Childhood. - The study report to the funders has been prepared. A paper on the main study results has been submitted to BMJ Public Health. Published outputs: Peer reviewed publications: - Cavallaro et al. (2020). "Evaluating the real-world implementation of the Family Nurse Partnership in England: protocol for a data linkage study." BMJ Open 10(5): e038530. - Cavallaro et al. (2022) “Characteristics of enrolment in an intensive home-visiting programme among eligible first-time adolescent mothers in England: a linked administrative data cohort study”. J Epi Com Health; Online First: 05 October 2022. doi: 10.1136/jech-2021-217986 - Cavallaro et al. “Lessons learned from using linked administrative data to evaluate the Family Nurse Partnership in England and Scotland.” Int J Pop Data Sci; 2023; in press Online articles: - Cavallaro et al. Reducing barriers to data access for research in the public interest—lessons from covid-19. The BMJ Opinion 2020. https://blogs.bmj.com/bmj/2020/07/06/reducing-barriers-to-data-access-for-research-in-thepublic-interest-lessons-from-covid-19/ Blogs: - Harron. “It did mean a lot”. What public engagement with teenage mothers taught us about out research. https://blogs.ucl.ac.uk/public-engagement/2019/05/07/engaging-with-teenage-mothers/ Conference abstracts: - Cavallaro et al. Using data linkage to assess the identification of vulnerabilities in administrative hospital data among a cohort of young mothers. Data for Policy 2021, Virtual. - Cavallaro et al. Are the most vulnerable mothers in England being targeted for additional support during pregnancy? An analysis using linked administrative data. Health Service Research UK 2021, Virtual. - Cavallaro et al. Are the most vulnerable mothers in England being targeted for additional support? Characteristics of enrolment in the Family Nurse Partnership. Public Health Research and Science Conference 2021, Virtual. - Cavallaro et al. Are the most vulnerable mothers in England being targeted for additional support? An analysis of characteristics of enrolment in the Family Nurse Partnership using linked administrative data. Society for Social Medicine and Population Health 2021, Virtual. - Cavallaro et al. Challenges and lessons learned from two countries using linked administrative data to evaluate the Family Nurse Partnership. International Population Data Linkage Network 2022, Edinburgh. - Cavallaro et al. Evaluation of the real-world implementation of the Family Nurse Partnership in England: an observational cohort study using linked data from health, education and children’s social care. International Population Data Linkage Network 2022, Edinburgh. - Cavallaro et al. Evaluation of the real-world implementation of the Family Nurse Partnership in England: an observational cohort study using linked data from health, education and children’s social care. Public Health Data Science 2022, Glasgow. - Cavallaro et al. Challenges and lessons learned from two countries using linked administrative data to evaluate the Family Nurse Partnership. Public Health Data Science 2022, Glasgow. Webinars: - Cavallaro & Harron. Using linkage of administrative data to evaluate the Family Nurse Partnership in England. MatCHNet webinar, 2022. - Harron. Using propensity score analysis to evaluate interventions. MatCHNet Methods Workshop, 2022.

Benefits reported

[1 paragraph unchanged] Update under v3: The benefits from the study outputs (listed above) are that these have provided commissioners of the Family Nurse Partnership with up to date and new evidence on the effectiveness of the intervention, including for different groups of mothers, on a range of child and maternal outcomes. This has direct relevance to the 4% of babies who are born each year to teenage mothers, and their families. This research complements and extends existing evidence from a randomised controlled trial of the FNP which included 1600 families enrolled between 2009-2010. This study adds to this evidence by describing outcomes for the 25,000 families who have been enrolled in the FNP between 2010 and 2019. This will give commissioners and policy makers more information about how to deliver targeted interventions to families that need them in their local areas. Findings have also been shared with the FNP NU, who will use the information to inform ongoing adaptations of the FNP in England.

Unchanged: Expected measurable benefits.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

[1 paragraph unchanged] The Family Nurse Partnership (FNP) is an intensive early home visiting programme [34 words unchanged] implementation of FNP in England. To do this, the research team at UCL University College London (UCL) will use electronic records that are routinely collected to compare outcomes for FNP participants with similar families who did not take part in FNP. [1 paragraph unchanged] The purpose of this application falls under Article 6 (1) (e) of [37 words unchanged] humanity’. The application also falls under Article 9 (2) (j), as scientific research. research: ‘processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject.’ [5 paragraphs unchanged] The data requested will achieve the aim of this study by providing [123 words unchanged] whether the effect of the programme differs according to whether the mothers has have long-term health conditions or has frequently visited the A&E department (from HES) or whether mothers have special educational needs (from NPD) The National Pupil Database (NPD) is controlled by the Department for Education, [36 words unchanged] needs, disability, and indicators for free school meals and a child in care, care. Personal details are linked to pupils' attainment and exam results over a [5 words unchanged] is this dataset that captures the details needed to explore educational attainment. [10 paragraphs unchanged] i) Propensity scores scores* To derive propensity scores, we will regress FNP participation on all available maternal characteristics, e.g. pre-pregnancy chronic conditions and educational attainment. Matched groups will be formed based on the propensity of participation. Effects will be estimated as the difference in outcomes between matched groups. Statistical models will allow for clustering of families within LAs, and multiple imputation will be used to account for missing data. * Propensity score matching is used when a group of subjects receive a treatment and their outcomes are compared with the outcomes of a control group. To derive propensity scores, UCL will regress FNP participation on all available maternal characteristics, e.g. pre-pregnancy chronic conditions and educational attainment. This means that UCL will create a model to predict how likely it is that a mother will be enrolled in the FNP, based on pre-pregnancy information in HES and NPD (for example, educational attainment at age 16, or emergency admissions to hospital prior to pregnancy). UCL will then create matched groups, based on the propensity (or likelihood) of participation in FNP. Effects will be estimated as the difference in outcomes between matched groups. Statistical models will allow for clustering of families within LAs, and multiple imputation will be used to account for missing data. [3 paragraphs unchanged] Sensitivity analyses will determine the strength of unmeasured confounding required to invalidate [15 words unchanged] any potential differences in results due to the use of real-world data, we UCL will use our cohort to replicate findings observed in BB. For each analysis strategy, we UCL will derive trial outcomes for a group of families in the administrative data cohort with the same aggregate baseline characteristics as trial participants. [3 paragraphs unchanged] UCL are proposing to link an existing cohort of mothers and babies [35 words unchanged] are currently provided via Open Exeter, on behalf of the Department of Health Health. [1 paragraph unchanged] A large control group of 975,000 rather mothers not enrolled in FNP (rather than a significantly smaller stratified sample sample) has been selected to allow for inconsistencies in the comparison of usual [31 words unchanged] of change in child development in the FNP population demonstrates that effects are concentrated in mothers with low education, but highlights the need for further research to examine other behaviours such as drug/alcohol abuse and domestic violence. behaviours such as drug/alcohol abuse and domestic violence. In 2021, the following additional requirements were included in this agreement: Amendment Request: The amendment is for the following: [3 paragraphs unchanged] This amendment is was to allow UCL to prepare a list of HESIDs for the study [14 words unchanged] share the identifiers with DfE to allow onward linkage to the NPD. This amendment is in addition to previous requests received under this agreement, the last iteration being DARS-NIC-136916-B7D5C-v0.24. This builds on previous activity, whereby NHS Digital extracted the identifiers for the cohort of FNP participants, and performed a list-clean to retrieve up-to-date identifiers for this cohort. FNP participants were then linked with HES, and a link-key was created so that the researchers could merge together data from their existing HES extract with FNP programme data. Following the initial FNP-HES linkage, UCL prepared a list of HESIDs for the study cohort (including control mothers) and the identifiers for this cohort were shared by NHS Digital with DfE for onward linkage with the NPD. The amendment to this application builds on previous activity, whereby NHS Digital extracted the identifiers for the cohort of FNP participants, and performed a list-clean to retrieve up-to-date identifiers for this cohort. FNP participants were then linked with HES, and a link-key was created so that the researchers could merge together data from their existing HES extract with FNP programme data. Following the initial FNP-HES linkage, UCL will prepare a list of HESIDs for the study cohort (including control mothers) and are now proposing and requesting that NHS Digital share the identifiers for this cohort with DfE for onward linkage with the NPD. [3 paragraphs unchanged] The research team are requesting that health data captured in HES and [7 words unchanged] are linked to an existing cohort of FNP participants and control mothers. This will include information from before the mother becomes pregnant as well as information on post-pregnancy outcomes (depending on the age of the mother at delivery). The purpose of this request is to answer a set of research questions, aiming to generate evidence on the real-world implementation of the FNP in England. [17 paragraphs unchanged] Data processor: processors: UCL, Department for Education Education, and Office for National Statistics [1 paragraph unchanged] The amendment is for the following: Data processing will only be carried out by substantive employees of the data processors or UCL PhD students who have been appropriately trained in data protection and confidentiality. - UCL will send a list of HESIDS from initial FNP HES linkage to NHS Digital. 2022 Amendment: - NHD Digital will share identifiers with DfE. UCL request access to be given to 2 PhD students. - DfE will provide onward linkage to NPD. UCL request access to be given to 1 individual who is under honorary contract. Access is only permitted once the honorary contract has been signed by all required parties and meets NHS Digital’s standard on honorary contracts. The individual on the Honorary Contract will, in due course, be acting on the instructions of the applicant and not carrying out data controllership activities for the Health Foundation. This amendment is in addition to previous requests received under this agreement, the last iteration being DARS-NIC-136916-B7D5C-v0.24 and is explained in further detail below.

Processing activities

[1 paragraph unchanged] First, the FNP programme data will be linked with HES, by NHS [9 words unchanged] key (HES-FNP) was created and transferred to the UCL Data Safe Haven. This The 2021 amendment focuses focused on the second linkage, where the FNP-HES cohort identifiers will be transferred [9 words unchanged] key will be created (HES-FNP-NPD) and transferred ONS Secure Research Service (SRS). [12 paragraphs unchanged] 3. The FNP study ID, plus identifiers from NHS Digital (NHS number, GP code, name, sex, date of birth, postcode), were made available for linkage within NHS Digital (FNP programme data is held separately from the identifiers). [4 paragraphs unchanged] 4. UCL will prepare prepared a list of HESIDs for the FNP-HES cohort and transfer transferred these back to NHS Digital. 5. In this amendment the 2021 amendment, NHS Digital will extract extracted historical identifiers for the list of HESIDs (including name where available for FNP participants) and share shared the identifiers with DfE for linkage with NPD. NPD Further to the amendment listed in point 5 above: [6 paragraphs unchanged] There will be no attempts to identify individuals. Risk of re-identification will [16 words unchanged] shared or published. Data processing will only be carried out by substantive or honorary employees of UCL who have been appropriately trained in data protection and confidentiality. Once in the data safe haven, researchers based at the Institute of Child Health (the researchers are all substantive employees of UCL or PhD students) will be able to access the data in the safe haven. All those involved in the processing of the data are substantive employees of UCL, or students on UCL MSc and doctorate courses under the supervision of UCL substantive employees. The work undertaken by the students is only for the purpose stated in this Purpose section. All UCL students are expected to undertake annual training on handling highly confidential information. All Trainees and students register for and complete NHS Digital’s Data Security Awareness (NHSD) course provided by e-Learning for Health. The course covers data security awareness, the law, threats to data security, breaches and incidents, and the General Data Protection Regulation. All students working on this study are students from UCL. All students working on the study will undertake the NHS Digitals Data Security Awareness course provided by e-Learning for Health. UCL has a specific data protection and information security policy, which applies to all staff and students when processing personal data on behalf of UCL. All UCL students working on the study are bound by this policy, and that they will face potential sanctions in the event of a breach of the policy. All students sign up to the UCL's Academic Manual. The Student Academic Misconduct section of the 2021-2022 manual Section 9.1, item 3 states "All instances of Research Misconduct whether by taught students, research students or members of staff will be investigated under UCL’s Procedure for Investigating and Resolving Allegations of Misconduct in Academic Research".

Expected output

The main output will consist of a report on the effectiveness of [9 words unchanged] according to different health and education characteristics identified within HES and NPD. We UCL will disseminate these outputs by providing briefings of these results that will [95 words unchanged] e.g. fact sheets about the impact of FNP from a parent perspective. [3 paragraphs unchanged]

Expected measurable benefits

The research will hopefully benefit the provision of health care and the promotion of health, by [25 words unchanged] on the likely benefits to maternal and child health and development. This will may have direct relevance to the 4% of babies each year (22,465 babies [51 words unchanged] risk of inadequate prenatal care and unhealthy behaviours during pregnancy, these factors can may lead to greater healthcare and educational needs for their children. Approximately 80% [87 words unchanged] for evidence on the best ways to support young and vulnerable mothers. This research will hopefully complement existing evidence from the Building Blocks trial of FNP in approximately [15 words unchanged] for the 25,000 families enrolled in FNP since 2010. The study will hopefully directly benefit the Health and Social Care sector by providing NHS managers, [16 words unchanged] roll out and targeting of early interventions designed to support young mothers. Outputs from this study will may provide commissioners with information on variation in health outcomes and healthcare use according to different maternal characteristics (from health and education) and differing engagement with FNP. This will may help inform decision makers on those most likely to benefit from increased [15 words unchanged] use of hospital services, as well as improving child development. Our findings will may also be used by the FNP NU to inform ongoing adaptations of [44 words unchanged] of families participating in FNP, and those eligible but not participating, will hopefully be particularly useful for LAs wishing to target the most disadvantaged families. [1 paragraph unchanged] The study team will work with commissioners on the study steering group, [29 words unchanged] care for young mothers and their families in England. This study will hopefully therefore directly benefit the Health and Social Care sector by providing healthcare [19 words unchanged] education and social care outcomes for young vulnerable mothers and their families.

Unchanged: Benefits reported.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

University College London (UCL) are proposing to link an existing cohort of mothers and babies (held by the research team under a separate DSA with UCL; NIC-393510-D6H1D) with programme information from the Family Nurse Partnership (FNP) Information System. NHS Digital hold the FNP data and the data linkage will be completed by NHS Digital. Aim The data under NIC-393510-D6H1D is disseminated for a programme of research within the healthcare provision theme of the Policy Research Unit for Children, Young People and Families (CPRU), within University College London (UCL) which is funded by the Department of Health (DoH). The Family Nurse Partnership (FNP) is an intensive early home visiting programme for first time teenage mothers, delivered by trained nurses aiming to improve maternal and child outcomes by providing support throughout pregnancy and until the child’s second birthday. This study aims to evaluate the real-world implementation of FNP in England. To do this, the research team at UCL will use electronic records that are routinely collected to compare outcomes for FNP participants with similar families who did not take part in FNP. The HES data under NIC-393510-D6H1D is limited to those under the age of 56. This data will be further minimised for the purpose of this agreement to those mothers aged 13-25 giving birth between 2010 and 2019 in England. The Civil Registration and the bridging file data released under NIC-393510-D6H1D is limited to deaths registered in England between 1st January 1998 until as late as possible, for all persons who died aged 0-55. The data will be further minimised for the purpose of this agreement to those mothers aged 13-25 giving birth between 2010 and 2019 in England. Legal basis This request (and the purpose of this application) is for a longitudinal research study funded by NIHR. The Family Nurse Partnership (FNP) is an intensive early home visiting programme for first time teenage mothers, delivered by trained nurses aiming to improve maternal and child outcomes by providing support throughout pregnancy and until the child’s second birthday. This study aims to evaluate the real-world implementation of FNP in England. To do this, the research team at UCL will use electronic health records (HES Outpatient, Accident and Emergency and Inpatient data, along with Civil Registration Data) that are routinely collected to compare outcomes for FNP participants with similar families who did not take part in FNP. There will be two cohorts: ~ mothers aged between 13-24 and their babies born between 01 April 2010 and 31 March 2019. The cohort size is approximately 975,000 but this will be further reduced through detailed matching of mothers with FNP participants in terms of age, parity, and other demographic and health characteristics (e.g. deprivation, chronic conditions). and ~ those within the FNP. Cohort size approximately 25,000. The total sample size will be approximately 1,000,000 mother-baby pairs; around 25,000 of these will be FNP participants. In practice, a smaller control group will be created for the comparison (see Processing Activities – Propensity Score Matching). Creating this matched control group will involve detailed analysis of the characteristics of mothers prior to pregnancy, in order to achieve closely-matched groups that are required to account for differences between mothers who participated in FNP and those who did not. Characteristics used to create the matched comparator group will include maternal age, ethnicity, local authority, deprivation, number of A&E visits, admissions for mental health conditions, and admissions for injuries and adversity-related diagnoses (in the 5 years prior to pregnancy). Within some local authorities, there may be a small number of women that can be matched to a FNP mother. The propensity matching process will be iterative and thus the whole population of mothers is initially required in order to generate the most comparable groups that will allow for a robust analysis. The purpose of this request is to answer a set of research questions, aiming to generate evidence on the real-world implementation of the FNP in England. [1 paragraph unchanged] There is clear public interest for this application, as it will research an important group of the population (teenage mothers). More than 20,000 babies are born to teenage mothers in England each year (~4% of births). Teenage mothers face a number of challenges in pregnancy (including lower levels of education, less stable careers and lower income than older mothers), which can lead to worse outcomes for mothers and their children. Whilst teenage motherhood can be a positive experience for some, early pregnancy is also related to domestic violence, less engagement with education and employment, and rapid repeat pregnancies. Unhealthy behaviours during pregnancy and inadequate prenatal care can lead to adverse neonatal outcomes, and children born to teenage mothers are at greater risk of maltreatment and associated long-term consequences, including adverse physical, social, emotional and cognitive outcomes, and depression, anxiety and suicidal behaviour. Findings from the study will help policymakers decide whether FNP should be offered to families in their local setting. Evidence generated by this study will support commissioners in providing improved services for mothers and children who could benefit most, and lead to increased efficiency through more effective targeting of resources. This information will inform targeting of appropriate services for families who are most in need, and most likely to gain from, additional support during pregnancy and early childhood. The results of this study will be used to inform professionals about the best ways to offer the FNP to those who could benefit from the service. Potential moral or ethical issues, or risk of potential harm to the public by the dissemination, relate to the risk of re-identification. These risks have been minimized through requesting the minimal data required for analysis, and through security arrangements for data processing and storage. The researchers will receive pseudonymised data only. How the data requested will achieve the aim The data requested will achieve the aim of this study by providing information on characteristics of mothers and their children eligible for the FNP in England. For example, the UCL team will look at outcomes for children (e.g., emergency hospital admissions for possible neglect or abuse captured in HES, and early childhood development captured in NPD) and mothers (e.g., subsequent pregnancies or hospital admissions due to violence or injuries captured in HES, and return to education captured in NPD). Exploring whether FNP works better for some families (e.g. the youngest teenagers) than others will help improve targeting of resources and highlight groups in need of alternative support. Information from HES and NPD will also allow the researchers to determine for which groups the programme is most effective. For example, they will be able to determine whether the effect of the programme differs according to whether the mothers has long-term health conditions or has frequently visited the A&E department (from HES) or whether mothers have special educational needs (from NPD) The National Pupil Database (NPD) is controlled by the Department for Education, based on multiple data collections from individuals aged 2-21 in state funded education and Higher Education in England. Data are matched using pupil names, dates of birth and other personal and school characteristics, including special educational needs, disability, and indicators for free school meals and a child in care, Personal details are linked to pupils' attainment and exam results over a lifetime school attendance, and it is this dataset that captures the details needed to explore educational attainment. [1 paragraph unchanged] - Describe variation in the delivery of FNP and usual care across Local Authorities (LAs) - Describe variation in health and educational characteristics of families participating in FNP over time and by LA; compare characteristics of FNP participants with families who met eligibility criteria but did not enrol [1 paragraph unchanged] - Evaluate the effectiveness of FNP on a broad range of health health, education and social care outcomes for both children and mothers - Determine which families stand to benefit from FNP using detailed information on maternal health and education trajectories prior to pregnancy (e.g. chronic conditions) conditions or school attainment) [2 paragraphs unchanged] - Determine how the effect of FNP differs between LAs. LAs The co-investigators: UCL will compare outcomes for mothers ever enrolled in FNP versus those who were never enrolled. Two analysis strategies will be used to take account of measured confounders related to both participation in FNP and outcomes: i) propensity score matching; ii) adjusted analyses. ~ London School of Hygiene and Tropical Medicine i) Propensity scores ~ Tavistock and Portman NHS Foundation Trust To derive propensity scores, we will regress FNP participation on all available maternal characteristics, e.g. pre-pregnancy chronic conditions and educational attainment. Matched groups will be formed based on the propensity of participation. Effects will be estimated as the difference in outcomes between matched groups. Statistical models will allow for clustering of families within LAs, and multiple imputation will be used to account for missing data. ~ University of Cambridge The main analysis will restrict matching within the same LA and within the time periods in which FNP was offered within that LA. Secondary analyses aiming to achieve more closely matched groups (with potentially smaller numbers) will match i) within the same LA but in different time periods, comparing outcomes for eligible families before vs. after FNP was offered; and ii) within the same time period but in different LAs, comparing outcomes for eligible families in LAs that did and did not offer FNP. ~ University of Kent ii) Adjusted analyses ~ University of Oxford This analysis will be an unmatched comparison, adjusting for maternal variables (e.g. pregnancy complications, ethnicity, Index of Multiple Deprivation, educational attainment) and neonatal variables (e.g. gestational age, birthweight, length of postnatal hospital stay, season of birth, congenital anomalies, admission to NICU). and the collaborators: Sensitivity analyses will determine the strength of unmeasured confounding required to invalidate results. To further assess the robustness of findings to the analysis approach and to evaluate any potential differences in results due to the use of real-world data, we will use our cohort to replicate findings observed in BB. For each analysis strategy, we will derive trial outcomes for a group of families in the administrative data cohort with the same aggregate baseline characteristics as trial participants. ~ FNP National Unit Relevant background to the request ~ University of Cardiff This is a longitudinal research study funded by The National Institute of Health Research. named in the protocol are not listed as joint data controllers. Although they were involved in the initial set-up of the study, they do not have any control or influence over the overall purpose of how the data will be used. UCL are the sole party who have control of the purpose and processing of the data for this study. Relationship between proposed project and associated work UCL are proposing to link an existing cohort of mothers and babies (held by the research team under a separate Data Sharing Agreement; NIC-393510-D6H1D) with programme information from the FNP Information System and education and social care data from the National Pupil Database (NPD). FNP cohort identifiers are currently provided via Open Exeter, on behalf of the Department of Health To achieve this linkage, UCL are requesting a re-purposing of a separate DSA (NIC-393510-D6H1D) to allow a list of the patient identifiers in HES (HESIDs) to be extracted from a current HES extract and sent by the research team to NHS Digital. A large control group of 975,000 rather than a significantly smaller stratified sample has been selected to allow for inconsistencies in the comparison of usual care, due to differences in the organisation of local services. Service providers also report variation between local authorities and over time on how FNP is delivered and targeted. Research examining drivers of change in child development in the FNP population demonstrates that effects are concentrated in mothers with low education, but highlights the need for further research to examine other behaviours such as drug/alcohol abuse and domestic violence. Amendment Request: The amendment is for the following: - UCL will send a list of HESIDS from initial FNP HES linkage to NHS Digital. - NHD Digital will share identifiers with DfE. - DfE will provide onward linkage to NPD. This amendment is to allow UCL to prepare a list of HESIDs for the study cohort, from the initial FNP-HES linkage with the request being that NHS Digital will share the identifiers with DfE to allow onward linkage to the NPD. This amendment is in addition to previous requests received under this agreement, the last iteration being DARS-NIC-136916-B7D5C-v0.24. The amendment to this application builds on previous activity, whereby NHS Digital extracted the identifiers for the cohort of FNP participants, and performed a list-clean to retrieve up-to-date identifiers for this cohort. FNP participants were then linked with HES, and a link-key was created so that the researchers could merge together data from their existing HES extract with FNP programme data. Following the initial FNP-HES linkage, UCL will prepare a list of HESIDs for the study cohort (including control mothers) and are now proposing and requesting that NHS Digital share the identifiers for this cohort with DfE for onward linkage with the NPD. Data subjects UCL have requested to use data for all mothers aged between 13-24 who delivered a baby between 01 April 2010 and 31 March 2017. UCL previously requested that FNP data for approximately 25,000 mothers who participated in the programme will be linked with HES, in order to define the comparison group. Comparator groups will include mothers who are closely matched with FNP participants in terms of age, parity, and characteristics (e.g. deprivation, chronic conditions, and UCL are now proposing to focus on educational attainment and social care histories, too). The purpose of the request The research team are requesting that health data captured in HES and education and social care data in NPD are linked to an existing cohort of FNP participants and control mothers. The purpose of this request is to answer a set of research questions, aiming to generate evidence on the real-world implementation of the FNP in England. For this purpose, the research team are requesting pseudonymised HES APC and A&E data for the year's corresponding to the data in the FNP cohort (births during 2010-2017, with data on mothers’ previous health going back to April 2005 and with outcome data going forward to March 2019). National data are required in order to capture all the FNP cohort, and to evaluate variation between Local Authorities. There are no alternative ways of achieving the purpose of this application. The research team will use the minimum data required in order to answer the research questions, and only request data for mothers in their cohort. Information about the cohort at each stage of processing: FNP data 1. FNP programme data for approximately 33,000 mothers and 34,000 children with the FNP study ID but no identifiers have been transferred from NHS Digital to the UCL Data Safe Haven. 2. Identifiers for the 33,000 FNP mothers and the 34,000 FNP children have been linked with HES and the link key has been transferred to the UCL Data Safe Haven. HES data 1. A list of HESIDs for the wider HES-FNP cohort (approximately 1,000,000 mothers and 800,000 children) will be transferred to NHS Digital. This includes identifiers for the 33,000 FNP mothers and 34,000 FNP children; the remainder of the cohort are controls. 2. NHS Digital will extract historical identifiers for the HES-FNP cohort list of HESIDs from PDS and share these identifiers with DfE for linkage with NPD. NPD data 1. DfE will use all available identifiers (sex, date of birth, name, postcode) to link the HES-FNP cohort (approximately 1,000,000 mothers and 800,000 children) with NPD. 2. A link-key will be transferred from DfE to the ONS SRS, allowing merging of the HES-FNP cohort with NPD. 3. De-identified NPD data for the HES-FNP cohort will be transferred to the ONS SRS with a study ID. 4. FNP data for the 33,000 FNP mothers and 34,000 FNP children will be transferred from the UCL Data Safe Haven to the ONS SRS and merged with the NPD using the link-key. Public and Patient Involvement (PPI) The Study Steering Committee (SSC) currently includes one lay member (a teenage mother), who has joined both of the SSC meetings to date. She played an active part in the discussions of the planning of future PPI within the study period, and in plans for the project itself. UCL will shortly be sharing results of the first analyses with the lay member, and working with her to develop a public facing article for the Conversation (a free, online resource aimed at the public). UCL have been working with the FNP National Unit with the aim of identifying a further current or former FNP participant as a second lay member to the SSC. Throughout the study, UCL will continue to work closely with the lay SSC member(s) so that they are able to contribute to the interpretation of results and, as UCL proceed with the analyses, to co-produce public facing material to summarise the results of the study. Towards the end of the study, UCL will hold a stakeholder engagement workshop, which will include FNP participants and graduates (recruited through the FNP national unit), as well as family nurses and commissioners. One of the aims of the workshop will be to feedback on results and to identify relevant streams of communication, e.g. social media and the Conversation. UCL will work with parents to co-produce a range of outputs suitable for FNP families, e.g. fact sheets about the impact of FNP from a parent perspective. UCL have also been conducting wider engagement, including a training day for parents interested in learning more about administrative data research at Great Ormond Street Hospital. Organisations involved Data controller: UCL Data processor: UCL, Department for Education and Office for National Statistics The study will involve collaborators from FNP National Unit, the Institute of Health Visiting, London School of Hygiene and Tropical Medicine and the University of Cambridge. The collaborating institutions will not process the data but will contribute to interpreting findings and generating outputs. There are no funders or commissioner directly involved in the project. The amendment is for the following: - UCL will send a list of HESIDS from initial FNP HES linkage to NHS Digital. - NHD Digital will share identifiers with DfE. - DfE will provide onward linkage to NPD. This amendment is in addition to previous requests received under this agreement, the last iteration being DARS-NIC-136916-B7D5C-v0.24 and is explained in further detail below.

Processing activities

The data linkage will be performed in two stages. First, the FNP programme data will be linked with HES, by NHS Digital, which took place in May/June 2020. A link key (HES-FNP) was created and transferred to the UCL Data Safe Haven. This amendment focuses on the second linkage, where the FNP-HES cohort identifiers will be transferred to DfE for linkage with NPD. A second link key will be created (HES-FNP-NPD) and transferred ONS Secure Research Service (SRS). Only de-identified data will be accessed by UCL. The first set of de-identified, linked data (HES-FNP) will be stored on the UCL Data Safe Haven, and was created by joining: - De-identified FNP programme data transferred from NHS Digital with a study ID - De-identified HES data extracted from an existing HES cohort held by UCL (DSA NIC-393510-D6H1D). - The HES-FNP link key, transferred from NHS Digital to the UCL Data Safe Haven. The second set of de-identified, linked data (HES-FNP-NPD) will be stored on the ONS Secure Research Service (SRS). This will be created by joining: - The de-identified, linked HES-FNP dataset, transferred from the UCL Data Safe Haven - De-identified NPD data, transferred from DfE with a study ID - The HES-FNP-NPD link key, transferred from DfE to the ONS SRS. [1 paragraph unchanged] A. FNP cohort (as some identifiers might have changed since enrolment (e.g. mother’s name, postcode), identifiers will first be updated using the Personal Demographic Service within NHS Digital, so that the most relevant set of identifiers can be used for linkage). A. FNP data 1. Identifiers for FNP mothers and their first child are currently held by NHS Digital via Open Exeter. These identifiers (NHS number, GP code, name, sex, date of birth, postcode) will be transferred to the NHS Digital HES Production team and linked to records held in PDS as some identifiers might have changed since enrolment. 1. FNP programme data with the FNP study ID but no identifiers were transferred from NHS Digital to the UCL Data Safe Haven. 2. FNP cohort with the FNP study ID but no identifiers will be transferred separately to the secure data safe haven at UCL. 2. Identifiers for FNP mothers and their first child are currently held by NHS Digital (Open Exeter). These identifiers were checked against PDS to obtain identifier histories (e.g. previous and up to date addresses) within NHS Digital. 3. The FNP study ID, plus identifiers (NHS number, GP code, name, sex, date of birth, postcode), were made available for linkage within NHS Digital (FNP programme data is held separately from the identifiers). [1 paragraph unchanged] 1. A HES cohort of mothers and babies will be prepared by the researchers at UCL An extract based on an existing dataset (NIC-393510-D6H1D) de-identified HES cohort held by the research team at UCL. Encrypted HESIDs for these records will be UCL (DSA NIC-393510-D6H1D), containing HESID and a number of analysis variables, was transferred to NHS Digital. a new server within the UCL Data Safe Haven. 2. A limited version of the HES cohort, containing encrypted HESID and a number of analysis variables, will be extracted from an existing HES cohort held at UCL (DSA NIC-393510-D6H1D) to a new server within the secure setting. 2. NHS Digital linked the FNP cohort with HES based on sex, date of birth, NHS number, name, postcode, and GP code. 3. NHS Digital will extract identifiers for the list of encrypted HESIDs (sex, date of birth, NHS number) and updated identifiers from PDS (name, postcode, GP code) and use these for linkage with the FNP cohort (using the key from the abstract for NIC-393510-D6H1D) . 3. A link-key was transferred from NHS Digital to the UCL Data Safe Haven to allow merging of the FNP programme data with the existing HES data. 4. A pseudonymised link-key will be transferred from NHS Digital to the UCL data safe haven. 4. UCL will prepare a list of HESIDs for the FNP-HES cohort and transfer these back to NHS Digital. C. Secure setting 5. In this amendment NHS Digital will extract historical identifiers for the list of HESIDs (including name where available for FNP participants) and share the identifiers with DfE for linkage with NPD. 1. The link-key will be used to merge the de-identified FNP programme data (cohort) with the HES analysis variables within the secure setting of the data safe haven. Identifiers will not be held in the secure setting. The data will remain pseudonymised as the data is encrypted. Further to the amendment listed in point 5 above: C. NPD data 1. NHS Digital will transfer identifiers for the FNP-HES cohort to DfE for linkage with NPD (no FNP programme data or HES data will be transferred). 2. DfE will use all available identifiers (sex, date of birth, name, postcode) to link the FNP-HES cohort with NPD. 3. A link-key will be transferred from DfE to the ONS SRS, allowing merging of the HES-FNP cohort with NPD. 4. De-identified NPD data will be transferred to the ONS SRS with a study ID. 5. The HES-FNP cohort will be transferred from the UCL Data Safe Haven to the ONS SRS and merged with the NPD using the link-key. [1 paragraph unchanged] The researcher will compare outcomes for mothers ever enrolled in FNP versus those who were never enrolled. Two analysis strategies will be used to take account of measured confounders related to both participation in FNP and outcomes: i) propensity score matching; ii) adjusted analyses. Propensity Score Matching (statistical matching technique used to estimate the effect of a treatment) To derive propensity scores, the researcher will regress FNP participation on all available maternal characteristics, e.g. pre-pregnancy chronic conditions. Matched groups will be formed based on the propensity of participation. Effects will be estimated as the difference in outcomes between matched groups. Statistical models will allow for clustering of families within LAs, and multiple imputation will be used to account for missing data. The main analysis will restrict matching within the same LA and within the time periods in which FNP was offered within that LA. Secondary analyses aiming to achieve more closely matched groups (with potentially smaller numbers) will match i) within the same LA but in different time periods, comparing outcomes for eligible families before vs. after FNP was offered; and ii) within the same time period but in different LAs, comparing outcomes for eligible families in LAs that did and did not offer FNP. Adjusted analyses This analysis will be an unmatched comparison, adjusting for maternal variables (e.g. pregnancy complications, ethnicity, Index of Multiple Deprivation) and neonatal variables (e.g. gestational age, birthweight, length of postnatal hospital stay, season of birth, congenital anomalies, admission to the Neonatal Intensive Care Unit (NICU)). Sensitivity analyses will determine the strength of unmeasured confounding required to invalidate results. To further assess the robustness of findings to the analysis approach and to evaluate any potential differences in results due to the use of real-world data, the researcher will use the cohort to replicate findings observed in the Building Blocks Trial (a randomised control trial to evaluate the Family Nurse Partnership in England which was conducted by Cardiff University and ended in March 2016). For each analysis strategy, the researcher will derive trial outcomes for a group of families in the administrative data cohort with the same aggregate baseline characteristics as trial participants.

Expected output

The main output will consist of a report on the effectiveness of FNP for different groups of families. The researcher families, including those categorised according to different health and education characteristics identified within HES and NPD. We will disseminate these outputs by providing briefings of these results that will be prepared for policy makers and disseminated using the FNP Neonatal Unit’s NU’s existing networks. Findings will be used by the FNP NU to inform [37 words unchanged] Public Health, BMJ, JAMA Paediatrics, submitting within 3 years of data access). The researchers will also work with parent representatives to co-produce a range of outputs suitable for communicating results to families participating in FNP, e.g. fact sheets about the impact of FNP from a parent perspective. The researchers will also work with parent representatives to co-produce a range of outputs suitable for communicating results to families participating in FNP, e.g. fact sheets about the impact of FNP from a parent perspective. The researchers have already had input on their study from a number of teenage mothers, and these mothers will continue to be involved in dissemination of results, e.g. by co-producing outputs and ensuring that public-facing materials are age-appropriate. The FNP National Unit are already very experienced in producing material that is appropriate for the ages of their participants, and they will have input to the outputs from this study. Two mothers sit on the study steering committee and will advise on appropriate routes to disseminate outputs, e.g. social media and blogs on the FNP study website. [1 paragraph unchanged] Outputs from the study will help policy-makers decide whether FNP should be offered to families in their local setting. Outputs will also provide commissioners with information on variation in health outcomes and healthcare use according to different maternal characteristics and differing engagement with FNP. [1 paragraph unchanged] Outputs will contain only aggregate level data with small numbers suppressed (in line with the HES Analysis Guide). Data will not be used for sales or marketing purposes. Only aggregated data with small numbers supressed (in line with the HES Analysis Guide) will be used by the organisations mentioned in the protocol.

Expected measurable benefits

The research will benefit the provision of health care and the promotion of health, by informing policy on the effectiveness of the FNP for vulnerable families in England, England (including vulnerabilities defined by education and social care data), and providing evidence on the likely benefits to maternal and child health. health and development. This will have direct relevance to the 4% of babies each year [58 words unchanged] and unhealthy behaviours during pregnancy, these factors can lead to greater healthcare and educational needs for their children. Approximately 80% of infants born to first time [81 words unchanged] for evidence on the best ways to support young and vulnerable mothers. This research will complement existing evidence from the Building Blocks trial of FNP in approximately 1600 families recruited in England between 2009-2010, by evaluating health, education and social care outcomes for the 25,000 families enrolled in FNP since 2010. The study [29 words unchanged] roll out and targeting of early interventions designed to support young mothers. The research Outputs from this study will provide commissioners with information on variation in health outcomes and healthcare use according to different maternal characteristics (from health and education) and differing engagement with FNP. This will help inform decision makers on those most likely to benefit from [7 words unchanged] gains from reducing maltreatment, abuse and neglect, and emergency use of hospital services. Findings services, as well as improving child development. Our findings will also be used by the FNP NU to inform ongoing adaptations [57 words unchanged] be particularly useful for LAs wishing to target the most disadvantaged families. Targeted support for the most disadvantaged children and their families is recognised [27 words unchanged] to young mothers and their families could help improve maternal and childhood health and development outcomes. Research is outputs are therefore likely to have significant impact, given the ongoing roll-out of the FNP internationally. The study team will work with commissioners on the study steering group, [59 words unchanged] detailed evidence to inform policy and aid decision making in relation to health, education and social care outcomes for young vulnerable mothers and their families. Evidence generated by this study will support commissioners in providing improved services for mothers and children who could benefit most, and lead to increased efficiency through more effective targeting of resources. This information will inform targeting of appropriate services for families who are most in need, and most likely to gain from, additional support during pregnancy and early childhood. The results of this study will be used to inform professionals about the best ways to offer the FNP to those who could benefit from the service.

Benefits reported

Yielded Benefits is not a requirement for new applications. Due to a delay in data linkage and supply, there are no yielded benefits associated with this DSA to date.