Recruitment Is Not Representation: Why Trial Evidence Still Drifts From the NHS

The UK's clinical research system has a recruitment problem, but it also has a representation problem. The two are related, and they are not the same. As research bodies and NHS partners push recruitment in 2026, there is a risk that success is measured by the number of participants enrolled rather than by how well those participants reflect the populations who will eventually use the evidence. For a health service built on universal access, that distinction matters. Trial evidence that does not travel to the NHS's everyday caseload is evidence with limited practical value.

The recruitment target trap

Recruitment targets are necessary. Trials need sufficient numbers to answer their questions, and under-recruitment wastes time, money and participant goodwill. But a target is an arithmetic device, not an equity strategy. When recruitment is treated primarily as a numbers game, research teams are pulled towards the patients easiest to reach: those already attending large teaching hospitals, those with the confidence to volunteer, those with the time and transport to attend repeated visits, and those who have previously taken part in research. These are not the only patients in the NHS, but they are often the easiest to recruit. A 2023 review of the commercial clinical trials landscape drew attention to UK research capacity and delivery, and subsequent policy attention has rightly focused on speed and access. The harder question is whether the people being recruited are the people whose care the trial is meant to inform.

Where a trial is done shapes who is asked

Site selection is an under-discussed determinant of diversity. A trial run only from a specialist centre in a large city will draw from that centre's referral population and from people willing to travel. It may miss patients in coastal towns, rural areas, deprived neighbourhoods and communities with historically lower engagement with research. Moving trials into primary care, community pharmacies and local NHS settings can widen access, but only if those sites are resourced to deliver research safely and if community teams are involved early. Decentralised and remote designs can help, but they are not automatically inclusive. Apps, video calls and electronic consent assume stable connectivity, digital confidence, suitable devices and private space. For older patients, people with sensory impairments, people with caring responsibilities or households without reliable broadband, a 'remote' trial may be less accessible, not more.

The burden behind the consent form

Representation is also a matter of practical burden. Parking costs, repeated blood tests, time off work, childcare, language support and the physical layout of a clinic all shape who can say yes and stay in a study. A trial that requires six weekday visits may be scientifically sound and socially narrow. Researchers often discuss willingness to participate, but less often discuss the design choices that make participation disproportionately costly for some groups. Community engagement is not a last-minute recruitment tactic. It is a design input: it should influence the question being asked, the outcome measures chosen, the schedule of visits and the way results are communicated back. Without that, 'hard-to-reach' communities become a label for people whom research has not tried hard enough to reach.

Trust is built before recruitment begins

Mistrust of research is sometimes framed as a barrier located in particular communities. It is better understood as a response to history and experience, including unequal treatment, poor communication and research that has extracted data without returning benefit. A recruitment campaign cannot fix that in a poster. It requires sustained relationships with community organisations, general practices and faith groups, and it requires researchers to explain what happens to data, who owns findings and how participants will hear the results. The NHS's own data and digital infrastructure offer opportunities, but they also raise legitimate questions about consent, commercial access and surveillance. Public health surveillance and clinical research are not identical, but they share a need for public confidence. That confidence is earned through transparency, not through branding.

What real-world care needs from evidence

The NHS treats multimorbidity, deprivation-related illness, frailty, mental health problems and diverse ethnic populations. If trials repeatedly recruit younger, healthier, more affluent and less diverse participants, the resulting evidence may be internally valid but externally thin. Clinicians are then left to extrapolate: to ask whether a treatment effect observed in a narrow group will hold in the patient in front of them. That is not a failure of individual clinicians. It is a limitation of the evidence base. Translating research into everyday care means designing trials that can answer questions for the population the NHS actually serves, not only for the population that finds it easiest to enrol.

Practical questions for 2026

Clinicians can ask who is missing from a study before applying its findings: which age groups, ethnicities, deprivation levels and care settings were included, and where were they recruited? Researchers can ask whether site selection, visit schedules and digital tools were tested for accessibility, and whether community partners had a role in design rather than only in recruitment. Funders can ask for representation plans with the same seriousness as recruitment plans. Patients and the public can ask what the trial is trying to learn, what participation involves, what support is available, how data will be used and what will be shared back. These questions are not obstacles to research. They are quality controls.

Representation by design, not by afterthought

A recruitment push that does not change who is asked, where they are asked and what they are asked to do will produce more participants but not necessarily better evidence. The 2026 agenda should be judged not only by enrolment figures but by whether trial populations look more like the NHS. That requires deliberate site selection, realistic support for participation, accessible digital design and community involvement from the start. It also requires honesty about what a trial can and cannot show. Representation is not a diversity metric bolted onto a study; it is part of whether the study is fit to inform care. On that test, recruitment is only the beginning.