Every year, thousands of people across the UK attend emergency departments because of headaches. Some fear the worst — a brain haemorrhage or meningitis. Others arrive worn down by migraines or chronic headaches that no longer respond to treatment. But new research from Lancaster University suggests that when these emergency visits are recorded in NHS healthcare data, most of the clinical detail vanishes, and with it the insight needed to understand what these patients truly need.
The study, carried out by researchers at Lancaster Medical School in collaboration with Lancashire Teaching Hospitals NHS Foundation Trust, analysed 3,819 emergency department attendances for headache between 2021 and 2023. What the team found reveals a systemic issue in diagnostic coding — one with consequences that stretch far beyond administrative paperwork.
Have questions about this research or about how diagnostic coding shapes patient care? Write to us — we would be glad to hear from you and point you toward further resources.
A Common Complaint, an Unexpected Data Gap
Headache ranks among the most common reasons people seek emergency care. The presentations vary enormously. A sudden, thunderclap headache can signal bleeding around the brain. Fever and confusion alongside head pain may point to meningitis. Others attend because recurrent migraines or tension-type headaches have become unbearable, disrupting work, sleep and daily life.
Given this diversity, researchers expected routine hospital records to reflect a similarly wide range of diagnoses. Instead, they found the opposite.
“We expected to find a wide range of headache diagnoses. Instead, almost all patients were assigned one of just four diagnostic labels. When we looked at how these visits were recorded in hospital data, we found that most of the diagnostic detail had effectively vanished,” said Professor Hedley Emsley, Professor of Clinical Neuroscience and Consultant Neurologist at Lancaster University.
What the Numbers Reveal About Diagnostic Coding
Four Labels for Thousands of Patients
More than half of the 3,819 emergency attendances were coded simply as “headache”. Just over 40% were recorded as “migraine”. Alongside trigeminal neuralgia and tension-type headache, these four labels accounted for nearly 97% of all cases in the dataset.
That compression is striking when you consider that 34 different diagnostic terms were technically available to clinicians. The overwhelming majority of emergency headache patients were funnelled into a handful of broad categories, regardless of the underlying disorder.
Defining “Diagnostic Compression”
The researchers describe this pattern as “diagnostic compression” — the process by which clinically distinct conditions collapse into a few generic codes in routine records. It is not a coding error in the traditional sense. Rather, it reflects how emergency departments operate, how coding systems are structured, and what clinicians can realistically record during a fast-paced consultation.
The result is healthcare data that is systematically less detailed than the clinical reality it is meant to represent.
Why Diagnostic Coding Matters for UK Healthcare
Healthcare systems depend on coded data. These standardised labels make it possible to analyse information across thousands — or even millions — of patient encounters. Commissioners use them to estimate demand. Service planners use them to allocate resources. Researchers use them to identify patient groups who might benefit from targeted interventions.
But these analyses are only as good as the underlying data. When diagnostic coding flattens many conditions into a few categories, the true picture becomes distorted.
Headache medicine illustrates the problem clearly. Clinicians recognise many distinct headache disorders, each with different causes, treatments and long-term outcomes. The International Classification of Headache Disorders distinguishes migraine from cluster headache, medication-overuse headache, chronic tension-type headache and many other conditions.
Yet much of that complexity never appears in routinely collected emergency department data in the UK. A dataset that says only “headache” cannot tell a commissioner how many patients are struggling with painkiller overuse, how many have cluster headache, or how many return repeatedly because an underlying disorder is going unaddressed.
The Missing Diagnosis: Medication-Overuse Headache
One finding stood out to the research team. Medication-overuse headache — a condition that develops when people take acute pain-relieving medicines too frequently — was not recorded once in the entire dataset of nearly 4,000 attendances.
“This is unlikely to mean the condition was absent,” Professor Emsley explained. “Previous research suggests medication-overuse headache affects around 1% to 2% of the population and can lead to substantial healthcare use, including urgent and emergency care attendances.”
In other words, a common and treatable condition appears to be effectively invisible in routine emergency care records. When conditions like this disappear from healthcare data, the consequences ripple outward. Commissioners may underestimate demand for specialist services. Researchers may struggle to identify the patients who need help most. Resources may be directed elsewhere because the data never signalled a problem.
Have you or someone you know experienced repeated emergency visits for headaches? Share your experiences in the comments below — patient perspectives add invaluable context to research like this.
Not a Criticism of Emergency Clinicians
The researchers are careful to point out that diagnostic compression is not the fault of frontline staff.
“Importantly, this is not a criticism of emergency clinicians. Emergency departments are designed to identify serious illness and make safe, timely decisions,” said Professor Emsley.
A clinician seeing a patient with headache is primarily focused on excluding life-threatening causes — bleeding, infection or raised pressure inside the skull. Recording a highly specific headache diagnosis is rarely the immediate priority in that setting. The role of the emergency department is to rule out danger quickly, not to subtype a chronic headache disorder.
There is also a practical constraint. Clinicians typically select diagnostic codes from predefined lists during the consultation itself. Those options may not match the finer distinctions that headache specialists make in clinic, leaving little room for detailed recording even when clinicians recognise it.
Making Healthcare Data More Useful: Practical Solutions
The research points to several realistic ways healthcare systems could capture more clinically meaningful information without adding significant burden to emergency staff.
Prompt for the Details That Matter
Electronic record systems could prompt clinicians to record a small number of high-value details: whether a headache is new or longstanding, whether painkillers are being used frequently, or whether recognised warning signs are present. These few data points would allow analysts to distinguish between a first-ever severe headache requiring urgent investigation and a recurrent condition already managed in the community.
Align Clinical Classifications with Coding Systems
Better alignment between clinical classifications — such as the International Classification of Headache Disorders — and the coding systems used in routine datasets would help ensure common headache disorders are represented more accurately. When the codes available mirror the conditions clinicians actually diagnose, detail is less likely to be lost at the point of recording.
These changes may sound technical, but their effects could be significant. As healthcare increasingly depends on large-scale data analysis, conditions that are poorly coded risk being overlooked when services are designed or resources allocated.
What This Means for Patients, Planners and Researchers
For emergency headache patients, better diagnostic coding could mean services that recognise patterns invisible today — for example, rising numbers of people attending A&E because painkiller overuse has turned their migraines into a daily burden.
For commissioners and service planners across the UK, richer healthcare data would support more accurate forecasting of demand for neurology services, specialist headache clinics and community-based interventions.
For researchers, it would open the door to studies that are currently not possible — tracking outcomes for specific headache disorders, evaluating new interventions, and identifying inequalities in access to care.
Interested in contributing to research that improves patient care? Explore postgraduate study and research opportunities in health and medicine at Lancaster University, where work like this is conducted.
The Road Ahead for Emergency Headache Data
“More detailed coding will not, by itself, cure headaches,” Professor Emsley concluded. “But if common headache disorders remain largely invisible in routine data, we may miss opportunities to improve care for the many people who live with them.”
The message from Lancaster University is clear: healthcare data is only as valuable as the detail it preserves. Improving diagnostic coding will not change clinical practice overnight, but it is a necessary step toward ensuring that emergency departments, commissioners and researchers all work from the same accurate picture.
For the thousands of people who attend A&E with headaches each year in the UK, that accuracy matters. It determines whether their conditions are counted, understood and planned for — or whether they, like the medication-overuse headache patients in this study, simply disappear into the data.
Found this analysis useful? Share this article with colleagues and anyone interested in how healthcare data shapes the future of NHS care.
