Accessing Anonymised Data from Ireland’s Maternal and Newborn Clinical Management System for Research Purposes

Collaboration with clinical leads

A request to collaborate should be made to the clinical lead within each unit where data is being requested. If they agree, permission to add their name to the study protocol should be sought. It is advised that this request is made by the Principal Investigator or other senior member of the project in the first instance, as early as possible in the project timeline. Clinical leads may help to identify the appropriate person who would carry out the data extraction and anonymisation in their respective units. Additionally, it allows for an opportunity to provide further details of the project, if necessary. For example, a detailed description of the project could be presented to clinical leads/relevant persons to clarify any issues or answer any questions regarding the project, and to ensure the available data is suitable to address the proposed research question(s).

Study protocol and data collection sheet

Study protocol: A detailed study protocol outlining the purpose, aims, objectives, and methodology, as well as any pertinent information surrounding data protection, governance and insurance should be drafted. The protocol should also contain information on any inclusion and exclusion criteria, for example, specific years, singleton, or multiples births and from which maternity units’ data is being requested. A study protocol will allow for ethical considerations to be clarified early on and serve as a guide for members of the research team throughout the project timeline. Furthermore, this written record of the study plan allows for monitoring of project progress and prevents unnecessary deviations from the protocol¹¹. Institutional support may be available when drafting the protocol and should be utilised where available. While currently there is no standard study protocol template for requesting anonymised MN-CMS data, Table 1 outlines an example of the key components of a study protocol.

Data collection sheet: A data collection sheet outlining each data point/variable that is being requested for the purpose of the project must be created. This helps to ensure that the minimum amount of data is being collected for the study, and data is adequate, relevant, and limited to what is necessary for the project, reducing the risk of individuals being inadvertently identified^12,13. Furthermore, it allows data manager(s) within each unit to review the requested data points and assess their availability within the MN-CMS. While currently there is no standard data collection sheet template for requesting anonymised MN-CMS data, previous data collection sheets have been created using Excel and Word software.

National Information Governance Group

The National Information Governance Group is made up of health care professionals, scientists, informatics specialists, programme/project managers and administrators. This national group were instrumental in the initial development of the MN-CMS and continue to ensure the roll-out and on-going support of the national MN-CMS programme². Among other functions, the Information Governance group ensure compliance with relevant HSE policies and procedures as well as ensuring compliance with legislative and regulatory frameworks surrounding data protection.

If a data request is being made at a national level (i.e., from the four units that have the MN-CMS currently implemented), the protocol, data collection sheet and any other pertinent documents for the project must be reviewed and approved by the National Information Governance Group. All documents should be sent to MN-CMS Programme Manager for review at the subsequent governance group meetings, details of whom can be found here: https://www.ehealthireland.ie/ehealth-functions/acute-delivery/maternal-newborn-clinical-management-system-mn-cms/meet-the-team/meet-the-team/

Data Protection Impact Assessment

When a decision from the National Information Governance Group is made, a Data Protection Impact Assessment (DPIA) should be completed. A DPIA helps researchers to identify risks arising from the processing of personal data and to help minimise the risks early on in the project timeline. The DPIA template can be located here: https://hseresearch.ie/data-protection-and-research/ along with a companion guide to complete the DPIA.

Ethical Committee Review

Permission to access data from each maternity unit will need to be applied for separately from each unit’s local Ethics Committee and all local standard procedures regarding data-access must be followed. For example, ethical approval must be sought from the Clinical Research Ethics Committee of the Cork Teaching Hospitals (CREC) (for CUMH and UHK), the Research Ethics Committee of the National Maternity Hospital (for NMH) and Rotunda Hospital Research Ethics Committee (for Rotunda Hospital), prior to study commencement. Institutional support may be available when drafting ethics applications. Details on how to apply to each ethics committee are available here:

CUMH and UHK: https://www.ucc.ie/en/crec/

NMH: https://www.nmh.ie/home/research-ethics-committee.13650.html

Rotunda Hospital: https://rotunda.ie/knowledgebase/research-ethics/

Local Information Governance Group (LIGG)

When ethical approval is granted by CREC to access data from CUMH and/or UHK, a research application form (which can be accessed from https://redcap.ucc.ie/surveys/?s=HCM7JPF8CMP3DWJH) must be submitted to LIGG. The LIGG governance group advise researchers around issues with their planned research and ethical considerations from a data governance perspective. The research application form must first be sent to the Secretary to Director of Midwifery at CUMH who will then forward the application to a LIGG representative. The LIGG meet monthly to consider all applications. Applications which are deemed appropriate for approval are then sent to the Ireland South Women and Infants Directorate, Executive Management Committee (EMC) who meet every two weeks. The Chair of the LIGG raises applications that require further discussion at the EMC meetings.

The administrative coordinator to the EMC notifies the Secretary to the Director of Midwifery of approved and/or declined applications who then informs the person named of the application form of the decision. Only requests from those affiliated with an education, healthcare or other research institution are considered.

Data anonymisation and data transfer

According to the Data Protection Commission, data can be considered anonymised if it “does not relate to an identified or identifiable natural person or where it has been rendered anonymous in such a manner that the data subject is not or no longer identifiable”. When data is fully anonymised, it is no longer considered personal data and is therefore not subject to the same restrictions placed on the processing of personal data under the General Data Protection Regulation (GDPR). Anonymisation techniques can vary, (examples of which have been published previously¹³) and data protection law does not prescribe any particular technique. Pseudonymisation is “the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information¹³.

Data from the MN-CMS can be extracted and anonymised by the data manager(s) at the respective units. Therefore, project members would not have any role in the data anonymisation process. Firstly, data is extracted by data manager(s) through generation of MN-CMS reports. For example, there are approximately 30 data reports that can be automatically generated within the MN-CMS (see Table 2 for list of MN-CMS reports). Data on mothers and babies are matched using a medical record number (MRN), a unique identifier number assigned to a patient within the MN-CMS. Secondly, the data should be fully anonymised for the purpose of the project, prior to transferring it securely to project personnel. This data anonymisation process should be irreversible (i.e., not pseudonymised) and therefore no longer considered ‘personal data’ from a data protection perspective¹³. This anonymisation process reduces the risk of data breaches, however it does not eliminate the risk and therefore precautions should be taken to avoid inadvertent data breaches. This includes precautions during the transfer of the anonymised data to project personnel.

Data transfer: An encrypted platform should be used when transferring the anonymised data from the maternity unit to pre-specified project personnel. One such option is FileSender, which is provided by HEAnet, Ireland's National Education and Research Network (https://filesender2.heanet.ie/). FileSender enables users to send large, encrypted files both within and outside of academic organisations.

If requesting data from one unit only (and the researcher is affiliated with that institution), institutional-based options may be used. For example, Network Attached Storage (NAS) and cloud storage facilities such as OneDrive for Business. NAS provides a shared folder, which is made available on the institutional network, to a specified group whose members are individually authenticated by their domain logon account, while OneDrive provides a cloud storage facility within the institution to store and share work files.

Timeline

The time between initial request to data access can vary according to the amount of data being requested (i.e., number of maternity units from which data is being requested and number of variables being requested from each unit). The timeline typically ranges from approximately 6–12 months and should be factored into the project timeline. Each stage of data access can be a lengthy process, in particular if further discussion is required to clarify any information to each of the relevant bodies involved in the data request. Therefore, it is very important to provide as much detail as possible in the project protocol as well as the ethics and LIGG applications to avoid the need to clarify details, as this can result in significant delays. The Gantt chart provided in Figure 1 provides an approximate estimate of the timeline involved in each stage of data access.

Figure 1. Approximate estimate of the current timeline involved between initial request and data access.

Strengths

There are several advantages to requesting MN-CMS data to conduct epidemiological research. First, use of secondary data for research purposes requires fewer resources than collecting primary data¹⁴. In addition to this, the MN-CMS electronic health record provides valuable maternity related data on a national level which can be used to answer important research questions with an aim to make pregnancy safer and improve outcomes for mothers and babies. This is evident in other countries that use electronic health record data for research. For example, the Swedish Medical Birth Register (MBR) is an electronic health record containing information on prenatal care, delivery, neonatal care, as well as maternal sociodemographic and lifestyle factors. It was founded in 1973, and now contains data on over five million children, their mothers, and their siblings. This MBR has formed the basis of more than 1,000 research publications since its inception, the majority of which were published since the early 2000s¹⁵.

Limitations and recommendations

There are also some limitations that should be noted as well as a number of recommendations to further facilitate epidemiological research using MN-CMS data. First, while MN-CMS data are real world data and reflects outcome in practice, a reliance on existing data can be a limitation in terms of data availability, unmeasured variables, and uncertainty around data quality. However, in order to minimise issues surrounding data quality, daily data quality checks are conducted within the MN-CMS by Data Quality personnel. Any issues highlighted from these checks are then communicated with relevant healthcare personnel^3,16.

Second, there is a need for an MN-CMS data dictionary to inform researchers on data availability and to support researcher analysis. This recommendation is in line with a recent report published by the Health Research Board (HRB), Ireland which highlights the need for national data dictionaries and common data formats to improve data quality, support aggregation of data and guide interpretation of findings⁹. Furthermore, this report suggests that the many valuable health related datasets that currently exist in Ireland should be accompanied by standardised metadata in order to drive use of data resources for public benefit⁹.

Third, the application process would likely benefit from a standard data collection sheet and protocol template. These standardised templates would need to be agreed by each unit. While we have provided details of the key components of a study protocol when requesting anonymised MN-CMS data for research purposes, this is not a comprehensive list and should be adapted to specific projects. Researchers should check if institutional supports are available when drafting the study protocol and ethics applications.

Future directions

As the MN-CMS generates more data year-on-year, the creation of a data warehouse to manage and safely store data will likely be necessary. This should be accompanied by a national standardised data access process with transparent expected timelines and will require an adequately resourced data access system in order to drive use of MN-CMS data for research. Furthermore, the formation of a data science team to assist data management and handling of data requests would enhance the approaches to providing data for secondary use. This data science function would require increased supports from funding agencies such as Department of Health, Health Research Board and Science Foundation Ireland.

The introduction of a nominal fee for data access may also be necessary to partly cover costs associated with data processing. This fee-based model for data processing and extraction is also used in other research settings. For example, the electronic Data Research and Innovation Service (eDRIS), as part of Public Health Scotland, are the central provider of Scottish health data for use outside its primary function. The eDRIS team work in collaboration with researchers and provide support throughout the data access process. A fee applies for this service, with the cost depending on the complexity of the request and whether the request is coming from the public sector, an academic organisation, charity, commercial or private industry¹⁷.

Similarly, ‘Born in Bradford’ is a birth cohort that tracks the health and wellbeing of over 13,500 children, and their parents born at Bradford Royal Infirmary between March 2007 and December 2010¹⁸. Born in Bradford does not receive funding to provide data extracts to individual projects and researchers. Therefore, a fee applies for accessing the data, and similar to the eDRIS service described above, they operate a tiered pricing structure based on the type of organisation researchers are applying from¹⁹.

Alternative models of data processing and extraction that could be considered by MN-CMS include transferring the management of data to other national agencies, as used in other research settings. For example, Hospital In-Patient Enquiry (HIPE) is the principal source of national data on discharges from acute public hospitals in Ireland. The HIPE database is managed by Healthcare Pricing Office (HPO) in the HSE, and among other tasks, are responsible for responding to requests for data from researchers²⁰. Similarly, data from Growing Up in Ireland (GUI), the national longitudinal study of children and young people, are managed by the Central Statistics Office (CSO) and researchers can apply to access GUI data through the CSO²¹. Finally, the introduction of a single National Ethics Review Board, particularly for secondary use of data, would provide a more streamlined data access approach and would likely reduce the timeline involved between initial request and data access.