Maternity - Birth Data Study HES Extract

Royal College of Surgeons require an extension to the data sharing agreement as further time is required to prepare, finish and publish the remaining three papers (outputs) which will yield important benefits for Health and Social Care.

Royal College of Surgeons (RCS) request patient-level Hospital Episode Statistics (HES) linked with mortality data to support a programme of work on maternal and child health, specifically focusing on how maternal factors prior to and during pregnancy affect childhood service use and outcomes. RCS propose to link key characteristics from maternal admission records to hospitalisation patterns for their offspring, and to use this information to provide an integrated approach for understanding infant healthcare use and childhood outcomes.

Linked hospital data on mothers and their offspring has been used to support service evaluation, planning, and research in smaller populations (including within Scotland, Australia, the US, Canada, and the Netherlands) and prospective mother-child data will be made available as part of the Maternity and Children’s Data Set (http://content.digital.nhs.uk/maternityandchildren?tabid=2).

Bringing together information on mothers and their offspring using retrospective, national hospital data in England will allow RCS to evaluate maternal and household determinants of service use and outcomes in children, on a substantially larger and more detailed scale than previously possible.

Firstly, RCS will analyse HES-mortality data to investigate how maternal hospitalisation patterns are related to birth outcomes and childhood readmissions. Determinants of variation in healthcare use and outcomes will be described for a variety of patient groups, defined either by specific diagnoses and/or procedure (e.g. obstetric procedures, pregnancy-related conditions, chronic conditions, mental health or indicators of social risk/adversity).

Secondly, RCS will bring together information on outcomes of previous pregnancies to investigate how previous birth outcomes and inter-pregnancy interval influence childhood outcomes and hospitalisation patterns.

Finally, RCS will combine information on healthcare use in mothers and all her offspring, to determine changing influences throughout childhood. This increasingly complex information will help RCS to understand how determinants of healthcare use change through childhood – i.e. from clinical birth outcomes in infancy to maternal and household influences throughout childhood.

The data requested will only be accessed by staff within the RCS on a need to know basis and only for the purposes of this application, not for any other purposes.

Data will be stored on the secure server at the Royal College of Surgeons (RCS) and will be accessed by substantive employees of the Royal College of Surgeons and researchers who have an honorary contract with RCS.

Date of birth is required to calculate accurate post-natal length of stay and time to readmission.

To identify corresponding maternal and baby records (linking mother to baby), RCS will use non-disclosive, clinical and demographic variables common to both maternal and baby HES admission records. These would be episode dates, provider code, postcode district, GP practice and delivery information.

Linked HES-mortality data will be used to account for maternal and infant deaths within analyses; date of death will be transformed into days since birth/postnatal discharge and then destroyed. Taking into account timing of deaths is important when assessing trends in maternal and infant readmissions over time. Date of birth is required to calculate accurate post-natal length of stay and time to readmission.

Agreement between HES and national aggregate published data from the Office for National Statistics will be assessed, in relation to distributions of important data items (e.g. gestational age, birth weight, maternal age).

The data will not be linked to any other datasets.

RCS will not attempt, nor have any requirement, to re-identify any participants within the data.

RCS will not share the data with any 3rd parties, including the Royal College of Obstetricians and Gynaecologists or linked with the National Maternity Audit data.

RCS agree to the destruction of the date of birth and date of death data once the required calculations have been completed. Data destruction documents will be submitted to NHS Digital once completed.

All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data).

To date, we have published three papers for this study. A further three papers are being prepared for publication.

1. Harron K , Gilbert R , Cromwell D , et al . Linking data for mothers and babies in de-identified electronic health data. PLoS One 2016;11:e0164667.doi:10.1371/journal.pone.0164667

2. Harron K , Gilbert R , Cromwell D , et al . International comparison of emergency hospital use for infants: data linkage cohort study in Canada and England. BMJ Qual Saf 2018;27:31–9.doi:10.1136/bmjqs-2016-006253

3. Harron K, Gilbert R, Cromwell D, et al. (2017) Newborn length of stay and risk of readmission. Paediatr Perinat Epidemiol 31:221–32.

Results of the study have also been presented at the following conferences:

Public Health Science 2016

International Population Data Linkage 2016

Society for Social Medicine 2016

Society for Longitudinal and Lifecourse Studies 2016

Health Services Research UK 2016

Administrative Data Research Network 2016

Farr Institute Conference 2015

Public Health Science 2015

Further outputs of the analysis, in the form of aggregate data with small numbers suppressed ( in line with the HES Analysis Guide), will be submitted for publication in peer-review journals and presented at national and international conferences.

The study has received funding from the Wellcome Trust, who will support the publication of results in open access journals and ensure the wide dissemination of this study’s results to healthcare professionals, NHS managers, commissioners and policy makers.

The study team is also closely linked with ongoing maternity service evaluation conducted by the Royal College of Obstetricians (RCOG). The RCOG set standards for clinical practice, publish clinical guidelines and patient information leaflets, contribute to policy development, and advise the government and other public bodies on healthcare matters.

RCS aim to make their results available in both clinical and methodological forums: RCS will aim to submit abstracts for presentation at the following conferences within 1 year of data access:

* International Population Data Linkage Network,

*Society for Longitudinal and Lifecourse Studies,

*Public Health Science, Society for Social Medicine.

RCS will aim to submit publications to the following journals within 2 years of data access:

*International Journal of Epidemiology,

*Journal of Clinical Epidemiology,

*PLoS One,

*Archives of Disease in Childhood, Paediatrics.

All papers will be published with open access.

Data will not be used for sales or marketing purposes.

Over 650,000 mothers give birth every year in England, and there has been growing recognition of the need for evidence on how maternal factors prior to and during pregnancy affect childhood service use and outcomes. For example, NHS Digital are developing a new Maternity and Children’s Data Set linking prospective mother and child data, in order to commission services in a way that improve health and reduce inequalities; HQIP have funded the RCOG to run the National Maternity and Perinatal Audit, which will evaluate the quality of care received by women and newborns in the UK.

Outputs from this study will add to and complement both of these initiatives by providing detailed information on trajectories of care for mothers and babies, and evaluating changes in service use over time using retrospective data analyses.

This study will directly benefit the Health and Social Care sector by providing healthcare professionals, NHS managers, commissioners and policy makers with detailed evidence to inform policy and aid decision making in relation to maternal and paediatric healthcare services.

Specifically, these outputs will provide information on variation in healthcare use in NHS organisations, which will be used to help to reduce inappropriate variation in practice and to improve outcomes of care for specific groups of mothers and babies; this research will provide detailed information on maternal and infant hospitalisation patterns and outcomes which will be used to identify which groups would benefit from increased support prior to birth, or intervention prior to delivery or following birth.

As well as presenting to paediatricians and other clinicians at national and international conferences, this study team will work with clinicians on the study Steering Group and members of the Royal College of Obstetricians (RCOG) and the Royal College of Surgeons (RCS) to ensure that the outputs of this study are used to support clinicians, policy makers and commissioners in their efforts to improve the quality of NHS care.

The study team is also closely linked with ongoing maternity service evaluation conducted by the Royal College of Obstetricians. Similar RCOG studies have had clear benefits for the Health and Social Care sector, for example by providing evidence on modifiable risk factors for stillbirth to inform better monitoring during pregnancy, and by informing NICE guidelines relating to providing patients with information explaining the long-term consequences of prematurity.

Results will also provide evidence to support the Children and Young People’s Health Outcomes Forum and the National Child and Maternal Health Intelligence Network. Therefore, this will directly influence policy, through adding to the body of evidence to allow policy-makers to improve maternity and paediatric services across the UK, and supporting them to meet the needs of women giving birth in England and their families.

Three studies based on these data to date have been published, with a further three publications in preparation. These papers have been influential in their fields.

Harron et al 2015 (Linking data for mothers and babies in de-identified electronic healthcare data) has led to a number of other studies that require linkage of maternal and baby records. For example, the recent paper in Lancet (Zylbersztejn et al 2018. Child mortality in England compared with Sweden: a birth cohort study), and PLoS Medicine (Knight et al 2018. Perinatal mortality associated with induction of labour versus expectant management in nulliparous women aged 35 years or over: An English national cohort study) were both conducted using linked data, a direct result of the current study. In addition, the methods used in this study have led to other studies in different contexts, including Brazil (Paixão et al 2017. Evaluation of record linkage of two large administrative databases in a middle income country: Stillbirths and notifications of dengue during pregnancy in Brazil).

Harron et al 2016 (Newborn Length of Stay and Risk of Readmission) reveals important information on the association between the length of time a newborn stays in hospital immediately following birth, and the risk of being admitted to hospital within the first week of life.

Harron et al 2016 (International comparison of emergency hospital use for infants: Data linkage cohort study in Canada and England) demonstrates how international comparisons can be conducted using linked healthcare data for mothers and babies. The study found that children attending ED in England were much more likely to be admitted than those in Ontario, and provides important information for policy-makers on how best to focus resources, including in-hospital, primary care and paediatric care in the community.