Rare, Not Invisible: Inside the UK's Registries for Low-Prevalence Disease

A rare disease may be seen only once or twice in a career by a general practitioner. That is the defining problem of low-prevalence conditions: the individual patient is not rare to themselves, but the evidence base around them is thin. Registries exist to change that. By collecting standardised information across many centres and over many years, they turn scattered cases into something that can be studied. In the UK, that ambition is well established, but it is not straightforward.

Why registries matter when numbers are small

For common conditions, a single large hospital or a national survey can generate useful evidence. For rare diseases, no single site has enough patients. A registry creates a shared numerator: patients with the same condition, recorded in a consistent way, often across a whole country. That allows clinicians to describe natural history, researchers to identify cohorts for trials, and patients to see that their condition is being counted.

The UK has a long history of condition-specific registries, often developed with patient organisations and specialist centres. Some are national and long-running; others are regional, disease-specific or tied to a particular treatment. The National Congenital Anomaly and Rare Disease Registration Service in England, alongside equivalent arrangements in the devolved nations, illustrates the public-health ambition. The wider policy context—the UK's rare disease frameworks and action plans—has repeatedly emphasised better data as a route to better diagnosis, care and research.

The infrastructure behind the count

A registry is not simply a database. It is a set of agreements, definitions and technical choices. Patients must be identified, usually through specialist clinics, patient organisations or hospital records. Consent may be explicit, or data may be collected under statutory public-health provisions with an opt-out. Each model has consequences for coverage and trust.

The data must then be standardised. Rare disease presents a particular challenge because classification is uneven. Clinicians may use different terms for the same phenotype. Coding systems such as SNOMED CT, ICD-10 and ICD-11, Orphacodes and the Human Phenotype Ontology all have a role, but they do not map neatly onto one another. A registry that accepts free-text diagnoses without curation will accumulate noise. One that imposes too rigid a schema may exclude patients who do not fit.

Linkage is the next layer. Registries become more valuable when they can be connected to hospital activity, primary care, genomics and mortality data. In the UK, the NHS number is a powerful identifier, but its use is governed by strict information-governance rules. Pseudonymisation, trusted research environments and data-access committees are now standard parts of the architecture. They slow work down, but they also make it possible.

Small numbers, large statistical problems

Rare disease registries face a paradox. Their purpose is to aggregate patients, but aggregation does not remove the problem of small cells. A table showing five patients in a particular age band and region may be statistically unstable and potentially disclosive. Confidence intervals around estimates can be wide enough to make a difference between regions look meaningless. Trend analysis over time can be dominated by a handful of new diagnoses.

Researchers respond with careful methods: pooling across years, using Bayesian models, suppressing small counts, or reporting ranges rather than exact figures. These are not cosmetic adjustments. They shape what can be said. A registry may be able to describe the overall cohort well while saying very little about any single subgroup. That limitation matters for equity, because the patients most likely to be missing from a rare disease registry—those without a specialist diagnosis, those in underserved areas, those with atypical presentations—are often the ones about whom the health system most needs to learn.

Consent, privacy and the limits of openness

Rare disease data are inherently identifying. A combination of diagnosis, age, region and treatment can point to a very small number of people. That makes privacy a first-order concern rather than a compliance afterthought. Patients who contribute to registries often do so with a strong sense of public benefit, but that goodwill is not unconditional. Commercial access, insurance implications and re-identification risks all affect willingness to participate.

The UK's model is varied. Some registries rely on explicit consent; others use statutory arrangements with transparent opt-outs. Both can work, but both require clear communication. A registry that cannot explain who will see the data, for what purpose and with what safeguards will struggle to maintain public confidence. The ethical question is not only whether data are lawful, but whether their use is legitimate in the eyes of the people who supplied them.

Interoperability is a clinical issue

It is tempting to treat interoperability as a technical matter for IT departments. In rare disease, it is clinical. If a specialist centre records a diagnosis in one coding system and a neighbouring trust uses another, the registry may miss cases. If primary care records are not linked, the registry may capture only the most severe end of the spectrum. If genomic data sit in a separate silo, the registry cannot connect phenotype to genotype.

Progress depends on shared definitions, reliable identity management and governance that allows proportionate access. It also depends on clinical time. Registries are often maintained by clinicians and specialist nurses who already carry heavy workloads. Without funded data-entry and curation, even the best-designed system will be incomplete.

Reading registries critically

For patients, clinicians and journalists, the critical questions are consistent. What is the coverage of the registry? Which centres contribute, and which patients are missing? How are diagnoses validated? What linkage is permitted, and under what governance? How are small numbers protected? Who can access the data, and on what terms?

A registry is not a neutral mirror of disease. It is an infrastructure that makes some things visible and leaves others out. That does not diminish its value. It simply means the data should be read with the same care that went into collecting them. For conditions that rarely appear in a single GP surgery, that care is the difference between a numbers exercise and a genuine account of lives affected by rare disease.