The epidemiology of peripartum cardiomyopathy in the United Kingdom.

The University of Glasgow is the data controller for a study for which the aim is to collect data on a rare pregnancy-related type of heart failure, Peripartum cardiomyopathy (PPCM) and describe the epidemiology of the condition in the UK. At present, it is understood that there are no studies of this condition from the UK and only two from Europe and there remain many unanswered questions about the condition.

The data requested will allow researchers to answer many important questions about the pattern of Peripartum cardiomyopathy (PPCM) in the UK. At present, there is no UK-based information to give to women with the condition or to guide decision-making with regards to their management and future pregnancy risk. Extrapolating information from studies from other countries can be difficult as the population demographics often differ markedly. For example, many studies are from South Africa, but in predominantly black women whose outcomes differ from women of other ethnicities. As the population in the UK is predominantly Caucasian, applying evidence from South Africa to women in the UK is difficult.

The study will form part of a programme of research into the epidemiology of PPCM in the UK. It will take the form of a retrospective cohort study with a nested case-control study. The same study is also taking place in Scotland using similar routinely collected NHS hospitalisation data collated and linked by NHS National Services Scotland. It will collectively form the basis of PhD project for an individual at the University of Glasgow.

The data subjects are all individuals who had an obstetric episode between 1997 to 2017 plus one of seven specific ICD diagnostic codes which indicate PPCM, heart failure (HF), hypertensive heart disease (HHD) or cardiomyopathy (CM), within 9 months prior and up to 2 years after the obstetric episode.

The data requested comprises hospitalisations for a limited number of specified health conditions/procedures related to the heart and to other comorbidities of interest to the study (e.g. diabetes, cancer, respiratory diseases, anaemia, etc.), maternity records and linkage with death records. This will allow an individual to be tracked to the end of life in order to report repeat hospitalisations and mortality.

The maternity records contain some information about the infants who are additional data subjects. This information will be used to assess the viability of analysing infant/child mortality and maternal outcomes and to undertake analyses if feasible.

For the comorbidities of interest, the data requested is limited to:

- Date of admission

- Date of discharge

- Length of stay

- Age at time of admission

- Diagnostic and procedural codes

- Deprivation marker

- Ethnicity

Ethnicity as a variable is relevant because it may be associated with both developing the condition and with outcomes after diagnosis. For example, Black race has been shown to increase the risk of developing PPCM by 16-fold. The study would investigate whether ethnicity is a risk factor in the UK, if the data contains sufficient numbers to be able to do so.

The data are pseudonymised to protect patient confidentiality. It will not be possible for the researchers to identify individuals from the data and no attempt will be made to identify individuals.

A period of 20 years is being requested; given that the condition is rare (the incidence in the only population level study from Europe was 1 in 10,500 live births) researchers believe that by expanding the study period to this extensive period will increase the study sample and ensure the statistical techniques employed are as robust as possible. The data controller request data for the whole of England to ensure this is a population-level study. This is important for two reasons:

1) currently only one population-level study in Europe exists but this is small (n=61 women)

2) this will allow researchers to report on an unselected cohort of women and reduce the bias associated with selecting certain geographical areas only. For example, some areas may have a specialist heart hospital or have a clinician with a specialist interest in the condition which could therefore artificially inflate the incidence of the condition.

Only data required to achieve the desired outcomes are being requested.

Only aggregate data (with small numbers suppressed) will be published.

In a subsequent Agreement, the researchers will request data for a control group. The control group will be required for a sub-set of the cohort provided under this Agreement i.e. women with ICD10 codes of PPCM, HF and CM (plus maternity). Researchers plan to request matched controls for this subset, which will be matched for age, health board/trust, and year of delivery. These matched controls will be requested at a later date and the University of Glasgow will submit a formal request to amend this Agreement when ready to request the control data.

Routinely collected HES data will be used, and therefore there is no direct relationship with patients. The way in which researchers process the personal data will be fully compliant with GDPR; the processing will be carried out using the following lawful basis:

Article 6(1)(e) Public Task: “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”

As Special categories of personal data is being processed, the following lawful basis will also apply:

Article 9(2)(i)Public Interest: “processing is necessary for reasons of public interest in the area of public health'.

The data controller believes that the outcome of this research will be of great public interest both nationally and internationally.

The organisations involved in this project and their roles are as follows:

The British Heart Foundation (BHF) has funded a BHF Clinical Research Training Fellowship for an individual studying for a PhD with the University of Glasgow. This project will form the basis of that PhD. The BHF has not been involved in the design of the study, makes no decisions in relation to the processing of any data under this Data Sharing Agreement and will not access any of the data under this Agreement.

The University of Glasgow is the sole data controller. The University of Glasgow employs the Chief Investigator/Study Statistician and the two co-investigators – one of which is the PhD student referred to above. These individuals designed the study and make all decisions regarding how and why the data under this Agreement are processed.

NHS Greater Glasgow and Clyde was named as the study Sponsor in the study Protocol v2.0. The decision to name this organisation as a Sponsor was taken because, for the element of the study population based in Scotland, the University of Glasgow chose to undertake a sub-study which involved validating the cases and collecting more data directly from patient records at NHS sites. This would be done through the NHS and these data would be held on an NHS platform. For this reason, NHS GGC was added as the sponsor for this aspect of the study. NHS GGC has no involvement with the element of the study which involves the collection and processing of NHS Digital data under this Agreement. For clarity, NHS GGC has had no input into the decisions how or why to process this data and will not access this data.

The study Protocol v2.0 also named West Glasgow Ambulatory Care as the location of the Sponsor Representative. Again, this is only relevant to the data being collected in Scotland and has no relevance to the data under this Agreement.

The University of Glasgow will be the sole processor of the data. The processing will not involve any other data processors as all of the processing will be undertaken by the University’s Research Team in-house. The Research Team carrying out the processing has the required extensive experience to undertake the processing activities required for the research task(s). The Research Team will monitor the processing of the data in line with the relevant data protection laws and regulations as advised by the University’s Data Protection & FOI Office. The data controller's DPIA will also continue to be reviewed by both the Research Team and the Data Protection & FOI Office to ensure that it is updated if and when necessary to include changes in legislation and/or GDPR. The Research Team have also consulted with the University’s Information Technology team’s ‘Information Security Policies’ to ensure that data is processed in line with their recommendations.

Patients with an incident discharge diagnosis of PPCM will be identified using the relevant ICD codes. Cases from 1997-2017 will be identified. This time frame will allow researchers to provide a minimum of 10-year follow-up for women earlier in the cohort (and reporting long-term outcomes is one of the research aims) as well as increasing the number of cases in the study. Given the anticipated rarity of the condition (there are no published data on the incidence in the UK but, from an unpublished pilot study in Scotland conducted by the research team, the incidence was reported as 1 in 11,000 live births), using all the available data will allow researchers to most accurately report the epidemiology of this uncommon condition.

Given that PPCM presents with signs and symptoms of heart failure (HF) or cardiomyopathy (CM), some women with PPCM may have been given a diagnosis of HF or CM rather than the more specific diagnosis of PPCM. Therefore, data for women who have an ICD code for HF and CM in the same time period will also be included in the dataset; those with a new diagnosis of HF or CM (and with no evidence of an alternative cause and with no prior history of cardiac disease) around the time of delivery, will be identified as possible cases of PPCM. Prior cardiac diagnoses will be applied as exclusion criteria to further describe possible cases and ensure these are as close to ‘true’ PPCM as possible. In addition, researchers will identify possible cases up to 1 year either side of the birth and sensitivity analyses of these time restrictions will be performed. This novel approach of examining possible cases, and expanding the diagnostic window to a year either side of the birth, aims to identify a cohort of women that may have had PPCM, particularly given the changes in the diagnostic criteria over in the past decade.

Researchers therefore request the following cohorts of women:

1-those with a diagnostic code for PPCM

2-women with a diagnostic code for HF or CM in the peripartum period and no prior cardiac history (possible PPCM)

All data will be collected and linked by NHS Digital. Data will be minimised appropriately before dissemination and only relevant ICD10 codes will be provided. Routinely collected hospitalisation data (HES) linked with death records will be used to carry out analysis. The linked data will be presented in a pseudonymised format to the Research Team at University of Glasgow.

Data will be processed in accordance with regulations stipulated by NHS Digital. Only the Data Controller's Clinical Research Fellow (PhD student) will have access to the data. The data will not be shared with a third party. All data is pseudonymised and there will be no need nor any attempt to re-identify individuals.

Data will be accessed via a secure password-protected environment (in line with NHS Digital policy for data storage) at the University of Glasgow British Heart Foundation Glasgow Cardiovascular Research Centre. The following statistical methods will be employed:

- Descriptive analyses of numbers of cases, baseline characteristics and obstetric characteristics; these will be compared across groups of women (definitive PPCM cases and possible PPCM cases). Parametric and nonparametric tests will be used as appropriate.

- The association of ethnicity with the development of PPCM and with outcomes will be examined if there are sufficient numbers to allow this analysis.

- Age-standardised rates of incident PPCM and possible cases will be calculated using census estimates.

- Actuarial life-tables and Kaplan-Meier survival curves will be used to describe maternal and infant/child mortality and risk of subsequent maternal events.

- The association between potential risk factors and the incidence of PPCM, infant/child and maternal outcomes will be analysed using multivariable logistic regression.

- Survival analyses will be performed using the Cox proportional hazards models, the Fine and Grey method for competing risks and joint frailty models for recurrent events.

- All analyses will be subjected to a number of sensitivity analyses.

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 i.e.: employees, agents and contractors of the Data Recipient who may have access to that data).

Results of the stated analyses will be disseminated through the following means:

- Submission to peer-reviewed medical journals for publication

- Presentation (oral and poster) at scientific conferences and seminars (obstetric and cardiology)

- Dissemination of information to patient groups (such as our focus group) and generation of patient-information leaflets to de distributed to e.g. cardiac/obstetric centres, specialist nurses, patient support groups

- Inform Position Statements from the European Society of Cardiology PPCM Working Group

- To form part of a PhD undertaken by Dr Jackson

Through these means the following groups of people will have access to the published study results:

Academics, clinicians, patients/public, policy-makers/managers.

The outputs will not be used for commercial purposes, not provided in record level form to any third party and not used for direct marketing.

All outputs will only contain results in highly aggregated format and as statistical summaries and measures of association. Small numbers will be suppressed in line with the HES Analysis Guide. Record level information will not be released to any third party. Where outputs include abortion episodes, all national figures will be suppressed if less than 8. For sensitive ages (<15) figures will be suppressed if less than 10.

Time frame (from receipt of data):

0-12 months: Cleaning of data and analysis

12-24 months: Generation of abstracts and manuscripts for submission to journals and conferences

24-36 months: Writing up of PhD

There are currently no data on the epidemiology of PPCM from the UK and only 2 studies (small case series of fewer than 10 patients also exist) of patients from Europe with the condition. Extrapolating data from other countries has been difficult for two main reasons. Firstly, there is wide geographical variation in the reported incidence, prevalence and survival and, secondly, there are few studies with longer follow-up or data on outcomes of children born to women with PPCM. The data will therefore be of direct relevance to the health and social care system in the UK and will add to the global picture of PPCM from which the UK is currently missing and European data are sparsely represented. The project will fill this knowledge gap. It will allow a range of health care professionals to provide real-world evidence to patients, to better inform women about their condition, to more effectively communicate risk (including risk surrounding a future pregnancy), and to answer questions about potential dangers to the new born babies. Most importantly, it will help to better identify those at risk of developing the condition and highlight those with a high likelihood of an adverse outcome; it is these women that can then be targeted with tailored interventions in order to try and reduce the occurrence of the condition and improve prognosis for those in whom it is established.

There may also be a benefit to service delivery in Scotland; through a better understanding of the risks associated with PPCM, services can be redesigned to account for this risk, such as tailoring the level of monitoring for women with prior PPCM depending on the absolute level of risk researchers find in the population. This may involve a re-organisation of services into regional or supra-regional centres or may lead to an increase in local services if the risks are low enough to be managed in this setting.

The findings will be of value to a wide range of clinical researchers, from cardiologists to obstetricians, paediatricians and specialist heart failure nurses. This may allow the development of new avenues of research into the disease by identifying associations with incidence and outcomes and may elucidate new treatment pathways for the condition. The findings will also be of interest to basic science researchers, who can use newly identified associations to guide basic research into the mechanisms behind PPCM. By producing accurate, contemporary numbers on the incidence and survival of patients and their children, the work will inform future power calculations for trials of therapies in PPCM. At present, the widely varying estimates in the literature and lack of European data have made this a difficult task for those undertaking clinical trials in this area.

The research team will devise a patient-leaflet explaining relevant findings in user-friendly terminology to provide information to patients and their families. The co-investigator is involved in work at the Global Jubilee National Hospital, a specialist heart hospital in Glasgow and also has a role in the charity Pumping Marvellous. Therefore the team will be able to provide information directly to patients affected by PPCM. Results will also be disseminated via national and international academic meetings, such as the British Cardiovascular Society and European Society of Cardiology (ESC) congresses. In addition, results will be disseminated within the network of specialists involved in the ESC Study Group on PPCM (via co-investigators who are official members) and to the academic community by preparing the work for publication in high-impact journals. It is therefore likely that the study will also inform future position statements and guideline-recommendations from the European Society of Cardiology PPCM Study Group.

Yielded Benefits is not a requirement for new applications.