MR623 - NATIONAL MOTHER AND CHILD COHORT

This Data Sharing Agreement permits the retention of the data provided under previous iterations of this Agreement for an interim period. This is a pragmatic approach to provide an active Agreement whilst enabling University College London to complete the necessary actions to enable a subsequent application to extend the Agreement meeting all applicable data sharing standards as published in NHS Digital’s website (see: https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance).

The following provides background information on the purpose of the original study:

University College London requires data for the purpose of a project which originally aimed to establish the feasibility and methods for long-term passive follow-up of children born to women living with HIV in England and Wales using national routine (registration) data to monitor deaths and cancer diagnoses. The National Mother and Child Cohort is an extension of the National Surveillance of HIV in Pregnancy and Childhood (NSHPC). The NSHPC conduct population level surveillance in the UK, collecting obstetric and paediatric data on the pregnancies of and infants born to women living with HIV (WLHIV).

The vast majority of these children were not only exposed to HIV in utero but also to antiretroviral drugs, used to prevent vertical (mother-to-child) transmission and to treat maternal HIV disease. The project provides the infrastructure for monitoring cancer incidence in and survival of this population and providing comprehensive data on cancer morbidity and all-cause mortality. Antiretroviral drugs have both known and unknown safety concerns, highlighting the importance of pharmacovigilance in this uniquely exposed population. At the time the project was set up, it was recognised that children HIV-exposed but uninfected (CHEU) with exposure to zidovudine, or other antiretroviral drugs might be at risk of mutagenic and carcinogenic effects at older ages. There were also some reports of mitochondrial dysfunction in a small number of uninfected infants. There is now some limited evidence of increased cancer risk associated with a first generation antiretroviral drug called didanosine (no longer used), but understanding of potential cancer risk and other serious adverse outcomes in CHEU remains poor.

The NSHPC has been undertaking comprehensive active surveillance of pregnant women living with HIV and their children for 30 years, and is now (since 2018), embedded within the Integrated Screening Outcomes Surveillance Service (ISOSS) of the Public Health England Infectious Diseases in Pregnancy Screening (IDPS) programme. The ISOSS is run from the University College London Great Ormond Street Institute of Child Health, which has operated the NSHPC since its inception. The IDPS programme is concerned with safe-guarding the health of children born to women with HIV, and a primary activity of the service is to ensure that the health of uninfected children born to infected women can be monitored. The NSHPC has had ongoing approval for collection of patient data without consent since its initiation.

The first children exposed to antiretroviral drugs in utero and neonatally in the UK were born in the mid-1990s, after a clinical trial showed zidovudine to be an effective intervention to reduce vertical transmission risk. Since then, changes in use of antiretroviral drugs for treatment and prevention has meant that an increasing proportion of children born to women living with HIV are exposed from conception. Recognising that the medium- to long-term outcomes in children exposed to antiretroviral drugs were unknown, the NSHPC pursued a number of research and surveillance strategies to address this evidence gap.

A flagging study was established in 1995 to monitor the cancer and death registrations of infants born to mothers living with HIV in England and Wales based on manual matching due to limited availability of NHS number at the time, with around 400/590 children born between 1996 and 1999 flagged.

Research funding was then secured to conduct the Children exposed to AntiRetroviral Therapy (CHART) Study from 2002-2005, which evaluated the feasibility of annual clinical follow-up of CHEU. Significant barriers to such follow-up including lack of capacity and family mobility were identified, although 94% of surveyed parents (131/139) agreed that it was important to follow-up uninfected children to see if there are any side effects from antiretroviral drugs. Follow-up via flagging and linkage with death and cancer registrations was therefore identified as a more feasible method for monitoring longer-term severe adverse outcomes in CHEU.

The ensuing development of the current project capitalised on the NHS “Numbers for Babies” initiative, with a new protocol established for flagging with the “National Mother and Child Cohort” in 2005. This was initially via the NHS Central Register, administered by the Office of National Statistics, but following national restructuring, was then conducted via the Health and Social Care Information Centre (HSCIC).

Article 6(1)(e) - 'processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller;' and Article 9(2)(j) - '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...' provide the legal basis for processing. There are no viable alternatives to processing to achieve the aim of this project, where processing is necessary in the interest of public health to identify any excess risk associated with exposure to antiretroviral drugs in-utero and ensure a high standard of medicinal products (antiretroviral drugs in pregnancy). Surveillance of this exposed population of children is ongoing and long-term to account for the lead-time between exposure and outcome; for statistical purposes a significant follow-up time and number of events (events are expected to be rare) are required in order to reasonably evaluate any associations between exposures (antiretroviral drugs) and outcomes (cancer and death events).

Ongoing ethics (London Rec ref: MREC/04/009) and Section 251 approval (CAG ref: 15/CAG/0190 [previously 15/CAG/0153]) were in place for all surveillance and research activities including the National Mother and Child Cohort (MR623) until this was superseded by Regulation 3 approval by the PHE Caldicott guardian in September 2019. The rationale for the regulation 3 approval is as follows: (i) the activity addresses a public health hazard; (ii) the data processing is appropriate and consistent with PHE policy; (iii) there is clear justification for not seeking consent (ascertainment bias; barrier to participation); (iv) opt out was not possible because it is a risk to the public’s health.

UCL currently holds record-level, identifiable data for its cohort members. UCL has received mortality, demographics and cancer registrations data for the cohort from 1999 until March 2017.

UCL wishes to retain this data with the intention of using it in the future, under an amended version of this Data Sharing Agreement, for analyses of events and validation of matching and flagging processes in the renewal application that is planned next.

The project involves “flagging” children born to women with HIV in England and Wales and reported to the National Surveillance (previously “Study”) on HIV in Pregnancy and Childhood (NSHPC) on the NHS Central Register, with subsequent event notifications for deaths and cancers to the NSHPC. The identified aim was to establish the feasibility of the flagging protocol – and the successful flagging and subsequent notification of events and linkage of these deaths and cancer outcomes to NSHPC data on exposures (including exposure to antiretroviral drugs in early life) has demonstrated the viability of this surveillance approach.

The longer-term aim (and main rationale for the flagging protocol) is: (i) to monitor for specific adverse outcomes (cancer and deaths) in children with in utero and/or early life exposure to antiretroviral drugs; (ii) to characterise these outcomes and to investigate potential associations between outcomes and drug exposures (i.e. relating to type of drug and timing/duration of exposure). Monitoring of adverse outcomes needs to be long-term, considering the long latency periods between exposures to carcinogens and clinical onset of cancers, and the expectation that cancer and deaths will be rare events. Following up all children born to women living with HIV in the UK over time is necessary to minimise selection bias, and to evaluate the longitudinal effects of different exposures.

Pregnant women with HIV have been recommended to receive antiretroviral drugs to prevent vertical transmission of HIV and/or for their own health since 1994. The purpose of the original project was to establish and implement a method for flagging HIV-exposed children (most of whom will also have antiretroviral exposure) born in England and Wales with death and cancer registries in order to conduct long-term monitoring of mortality and cancer diagnoses in this population, and to assess potential associations between these outcomes and antiretroviral drug exposures. Data subjects in this project are children born in England and Wales to women who were diagnosed with HIV infection before or during the pregnancy. These pregnant women and their children were identified as a result of the surveillance activities of the NSHPC which has population level coverage of HIV in pregnancy and childhood. All children meeting these criteria born between 1995 and 2008, with any anti-retroviral therapy (ART) exposure and regardless of their own HIV infection status, were flagged in the National Mother and Child Cohort. The number of individuals included within the cohort is 8877.

Under this Agreement, the data may be securely stored but not otherwise processed. No new data will be provided by NHS Digital under this Agreement.

The study data, including data provided by NHS Digital under previous agreements, are currently held by the University College London.

The following provides background on the processing activities undertaken prior to this Agreement:

Stage 1: Extraction

The dataset of children to be flagged was extracted from the NHSPC database (input dataset) including the following variables: unique study ID, child date of birth, multiple/singleton birth, sex, NHS number, hospital of birth, birthweight, mother’s partial postcode (district of residence at delivery), mother’s date of birth and mother’s country of birth. Encrypted datasets were then emailed to the ONS Newport Mortality Team Leader via email. ONS Newport then sent the ONS input dataset (containing the variables detailed above in the input dataset) to ONS Titchfield where matching procedure on the births/deaths registration database (BDRD) was carried out.

Stage 2: Matching algorithm

The matching algorithm structure was a hierarchical set of six match types based on variables both in NSHPC database and the BDRD: child NHS number, date of birth and sex, mother’s date of birth and partial postcode.

Matching Algorithm

• Match type 1) Child’s date of birth, sex and NHS number

• Match type 2) Child’s date of birth, sex and mother’s date of birth

• Match type 3) Child’s date of birth and mother’s date of birth

• Match type 4) Child’s date of birth, sex and mother’s partial postcode

• Match type 5) Mother’s date of birth, sex and mother’s partial postcode

• Match type 6) NHS number

Child birthweight, mother’s country of birth and variables used in matching algorithm, as well as the match type and number of matches found for the child were provided in the output dataset. The output dataset was emailed in a password protected excel file via ONS London to the researcher at NSHPC.

Stage 3: Confirmation of matches

To assess whether correct matches were made the NSHPC team ran a programme to confirm all matches using probabilistic methods. This was necessary because NHS number was not available for all children reported to the NSHPC. A password protected file including only those cases to be flagged (confirmed matches) were then emailed back to ONS Newport.

Stage 4: Flagging Subjects on NHS Central Register (NHSCR)

A dataset with children to be flagged was then sent by ONS Newport to the NHS Central Register where the for tracing. Their records were traced on the NHSCR and flagged with the National Mother and Child Cohort identifier. The decision was taken to use a generic flag name (i.e. “National Mother and Child Cohort Study”), with no mention of HIV.

Stage 5: Event Notifications

Once confirmation of flagging was received, a ‘members and posting (M&P) listing’ was requested from the Medical Research Information Service (MRIS). The MRIS team sent a named team member in the NSHPC event notifications annually. For death registration, date and cause of death was provided and for cancer registration, year of diagnosis, site and type of cancer was provided with a pseudonymised identifier.

Stage 6: Updating NSHPC Database

A flagging table on the NSHPC database contained pseudonymised data on all children flagged in the National Mother and Child Cohort. Event notifications were also stored on a table within the database in a pseudonymised format.

The NSHPC database is currently stored on the AIMES secure ISO27001 environment. All NHS Digital data files from processing are stored in UCL’s secure ISO27001 Data Safe Haven. Data is not linked to any third parties, mortality data received to date was used to update the NSHPC (data processors and controllers) database with confirmation of deaths on the paediatric and maternity surveillance interfaces. Data on cancers were not stored on the NSHPC database as there were few events; this data was stored in UCL’s secure ISO27001 Data Safe Haven.

Data processing is conducted solely by employees of University College London, specifically members of the ISOSS/NSHPC team and accessed in the secure ISOSS/NSHPC office at the UCL GOS Institute of Child Health. Mandatory annual information governance and GDPR training is required by all UCL staff. The ISOSS/NSHPC team members also complete Level 1 NHS Digital IG training and Data Security Awareness on an annual basis.

No new outputs will be produced under this Data Sharing Agreement.

There will be no dissemination of results under this extension. The nature of this project is to facilitate long-term surveillance with respect to the incidence of cancers and of survival in CHEU through the establishment of the National Mother and Child Cohort. Given that these are rare events and with knowledge of the substantial latency period between exposures to potential carcinogens and cancer onset, it is not appropriate to conduct analyses at this time. However, there have been communications activities with respect to the project methods with key stakeholders, including Public Health England (who are now commissioning this work), the European Medicines Agency, the Medical and Healthcare Products Regulatory Agency, researchers and the wider HIV community.

UCL have achieved the aim of the initial project, which was to establish feasibility and methods for the use of national cancer and death registration data to monitor adverse health outcomes in CHEU in England and Wales through the establishment of the National Mother and Child Cohort, via linkage with the NSHPC database. Methods for matching appear to be sufficiently robust to continue this longitudinal surveillance activity. There have been 100 death and cancer events reported for the cohort to date. The first cancer event was reported in 1996 and cancer event reporting has demonstrated no concerns (e.g. around missing data).

It would not be appropriate or valid to share project results at this point in this long-term surveillance activity. The planned later dissemination of findings following access to data under a future version of this Agreement will be in the public interest for the following reasons:

1. The vast majority of children born to women living with HIV were not only exposed to HIV in utero but also to antiretroviral drugs, used to prevent vertical (mother-to-child) transmission and to treat maternal HIV disease. All people with HIV require lifelong treatment with antiretroviral drugs (ARVs) and it is recommended that treatment starts as soon as a new diagnosis is made; early initiation of ARVs with good adherence to medication should result in achievement of life spans similar to those seen in uninfected people. Specifically, for pregnant women, alongside the benefits for their own health, widespread and early use of ARVs has resulted in the risk of vertical transmission decreasing from around 18-25% without treatment to 0.2-0.3% in the current treatment era. The enormous benefits of ARVs within and outside pregnancy are therefore undisputable. However, ARVs have both known and unknown safety concerns when used in pregnancy, highlighting the importance of pharmacovigilance in this uniquely exposed population. Antiretroviral drugs have had reported mutagenic and carcinogenic effects, in addition to haematological and mitochondrial toxicities. The recent safety signal of increased risk of neural tube defects with periconception use of the integrase inhibitor Dolutegravir highlights the evidence gaps around use of newer ARVs in pregnancy. Understanding the effects of HIV and ART exposure in foetal and perinatal life will inform treatment guidelines and contribute to risk-benefit analyses of the use of different combinations of ARVs.

2. There is a growing body of research that suggests children who are HIV-exposed and uninfected (CHEU) have poorer morbidity and mortality outcomes than children HIV-unexposed and uninfected (CHUU). This project will provide an evidence-base for evaluating these outcomes for CHEU in the UK and enable key stakeholders such as PHE and the NHS to design and implement measures that could alleviate health inequities in this growing population.

3. Given that the in utero exposures to ARVs have already taken place in thousands of individuals born to mothers living with HIV, should there be a signal of concern from the data collected then this will require an appropriate, ethical and proportionate response in terms of communication of results requiring expert input from key stakeholders such as the MHRA and PHE.

This project can be considered a first phase in the long-term surveillance of CHEU in England and Wales.

Flagging methods have been established and 100 death and cancer event notifications received for 8877 children reported in the NSHPC with a flagged status until 2017.