To address problems facing the health and care system, the ESHCRU team undertakes quantitative research that uses sophisticated statistical techniques. We rarely use qualitative methods such as interviewing people, nor do we often collect data directly from patients or carers. Instead, our research generally relies on analyses of ‘big data’, that is large datasets containing many thousands or even millions of cases.

The data we use come from administrative sources, such as hospital or GP records, or from major surveys or cohort studies undertaken by other organisations.  Cohort studies contain information on a large sample of individuals, with data collected at regular intervals over a number of years.  Though expensive to conduct, cohort studies can help answer many different problems with health and care – so they can represent good value for money.

All of this means we have no contact with the individual people whose data we are using.  Indeed, one of the conditions of access to these datasets is that we are not permitted to even try to identify individuals!

The data we use relate to people’s lives but the public may be unaware or feel disconnected from how their data are used or what benefits their use in research can bring in terms of improving the health and care system.  However, they have a right to know this and their support and buy-in are essential.

Involving patients, carers and members of the public in quantitative research brings both opportunities and challenges. Involvement is a good way of helping people appreciate the benefits that come from data-intensive research.  But there are some challenges of involving the public with our research.

• Recruiting the right people: ESHCRU is a programme of work spanning multiple projects, so who should we involve? It is quite straightforward to identify the right sort of people to involve when a study focuses on a particular health condition or age group.  But in ESHCRU, our work more often covers broad populations with many different health and/or social care needs, and individual projects study different populations.
• The ‘dismal science’: our research is about economics, so projects may examine issues that the public are unaware of or even uninterested in: for example, how health and care organisations function, how contracts are designed, or how providers are paid. Indirectly this sort of work can make a big difference to how care is experienced by very large numbers of people and also to the value for money delivered to the tax payers whose money funds such services.  But engaging with the public on these topics may be more challenging than for clinical trials, for example, because the impact on society is less direct.
• Communicating: the methods we use are often very technical and so explaining clearly what we did and what we found calls for excellent communication skills. It’s easy to lapse into jargon, and difficult to provide the right amount of detail. But if we don’t communicate well, we risk making people feel disengaged from the work and our research won’t benefit from their insights and experience.

The papers below all describe PPI in data-intensive research.  They consider who to involve and why and how – and also, more controversially perhaps, whether.  Several studies also discuss the skills researchers need if involvement and engagement is to be done well.

Why involve?

• A 2019 consensus statement on involvement and engagement in data-heavy research urges researchers to see the public as part of the solution to ensuring that the research benefits society. Benefits include raising awareness of the value of such research, empowering the public to make decisions about how their data are used and shared, as well as bridging the gap between the research and the people whose data are being used.
• In the context of clinical trials, involving patients can help them understand how statistics are used and why. Patients can also reality-check the relevance and meaningfulness of proposed outcome measures, based on their personal experiences.
• Johnson and colleagues (2021) found that involving people in end-of-life research improved their study, which involved analysing existing datasets. PPI helped the research team to prioritise their research questions and work out what could and could not be answered, and provide real-life examples to illustrate findings.

Who to involve?

• Hannigan’s 2018 paper considers the potential for and challenges of PPI in statistical research. Hannigan notes the difficulties of recruiting sufficient numbers of people, as well as representative samples of the population of interest.  In clinical trials, important outcomes can be missed if good involvement is not in place at an early stage.  Hannigan also underscores the need for research teams to have appropriate skills to facilitate PPI and build partnerships.
• A ‘data safe haven’ is a secure place where data can be stored and accessed only by individuals with special training in how to handle data safely. Involving the public in the work of a data safe haven in Wales is described by Jones and colleagues. The organisation considers involvement of the public as a duty and has made it a strategic priority. A consumer panel composed of 16 members of the general public has been set up to advise on a wide range of topics including how to help GPs engage with data sharing, and the use of mobile phone data for research.

How to involve?

• Kandiyali and colleagues (2018) discuss PPI in health economics research on children. Setting out case studies to show how patients and members of the public can be involved in cost-effectiveness analyses, the authors then draw out a number of potential benefits and recommend ways forward for future practice.
• Researchers at Imperial College describe how they involved the public in a mathematical modelling study of ‘flu vaccination. Participants were offered a short training course on disease modelling and shared an online guide on how to read a scientific paper. At a one-day workshop, public contributors critiqued a draft journal article and commented on draft questionnaire surveys.

Whether to involve?

• Oliver and colleagues (2019) sound a note of warning about “the dark side of coproduction”. Coproduction is a collaboration that involves stakeholders in the research process, though the term is used loosely and can be used to describe many approaches. Its aim is to make science more inclusive and democratic.  Oliver et al argue that coproduction has potential benefits, but also potential risks and costs –that the latter are rarely considered.  The authors argue for a more reflective approach to ensure coproduction is used where it can be most useful.

Skills required for researchers

• A consensus statement outlines 8 principles for involving the public with data intensive health research.  These boil down to being inclusive and accessible, having clarity of purpose and approach, ensuring communication is two-way, and doing it well: good governance, regular evaluation and keeping an eye on the end goal – producing an impact.
• Communicating statistics to the public is tricky at the best of times, but is especially challenging when those numbers are about survival rates in children’s heart surgery. The NIHR’s Five-step guide to involving the public in communicating research emerged from work by Christina Pagel to provide parents with access to reliable information on surgical outcomes. Her aim was to develop a website explaining how survival rates are calculated, and what they mean and what they don’t mean. The guide explains how people were involved at each step and the differences their contributions made to the website. For example, parents wanted information they considered most important to be placed up front, and they helped choose the best words to describe statistical concepts.
• Halle Johnson and colleagues ran a workshop to identify opportunities for meaningful involvement with palliative care research using big datasets. Explaining terminology was the essential first step, with a longer-term aim being to involve the public in supporting new data collections.
• Several important steps are needed for meaningful public involvement in the work of a data safe haven. People need to feel welcome; be encouraged to ask questions; are provided with information in a digestible format; and given enough time to think through issues. Engagement, which a two-way process, can best be achieved by fostering respectful relationships. This means being inclusive of different types of people whilst taking account of each individual’s level of interest. It also means treating people as equals, and communicating clearly and openly.