Thousands of people attend A&E with headaches every year but research suggests the way these visits are recorded may be hiding important clues about the needs of patients.
Headache is one of the most common reasons people attend emergency departments. Some arrive worried about a possible brain haemorrhage or meningitis. Others are living with recurrent migraines or chronic headaches that have become unbearable.
Researchers from Lancaster Medical School and Lancashire Teaching Hospitals NHS Foundation Trust conducted an evaluation of routinely collected NHS data at the Trust, analysing 3,819 emergency department attendances for headache between 2021 and 2023.
Professor of Clinical Neuroscience and Consultant Neurologist Hedley Emsley said: "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."
More than half were simply coded as "headache", while just over 40% were coded as "migraine". Together with trigeminal neuralgia and tension-type headache, these four labels accounted for almost 97% of all cases. Although 34 different diagnostic terms were technically available, the overwhelming majority of patients were funnelled into a handful of broad categories.
This phenomenon, which the researchers describe as "diagnostic compression", has implications far beyond administrative paperwork.
Why coding matters
Healthcare systems rely on coded data - standardised labels that allow information to be analysed across thousands or even millions of patient encounters, but these analyses are only as good as the data they are based on.
Headache medicine provides a good example of the problem. Clinicians recognise many distinct headache disorders with different causes, treatments and long-term outcomes. The International Classification of Headache Disorders contains detailed diagnostic categories that distinguish between migraine, cluster headache, medication-overuse headache, chronic tension-type headache and many others.
Yet much of that complexity never appears in routinely collected emergency department data.
The missing patients
Professor Emsley said: "One finding stood out - medication-overuse headache, a condition that develops when people use acute pain-relieving medicines too frequently, was not recorded once in our dataset.
"This is unlikely to mean the condition was absent. 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.
If conditions such as these are invisible in healthcare data, it becomes much harder to understand their true impact. Commissioners and service planners may underestimate demand, while researchers may struggle to identify patient groups who could benefit from targeted interventions.
"Importantly, this is not a criticism of emergency clinicians. Emergency departments are designed to identify serious illness and make safe, timely decisions. A clinician seeing a patient with headache may be primarily focused on excluding life-threatening causes such as bleeding, infection or raised intracranial pressure. Recording a highly specific headache diagnosis is often not the immediate priority.
"In addition, clinicians usually select diagnostic codes from predefined lists during the consultation itself. Those options may not match the distinctions headache specialists routinely make in clinical practice."
Making health data more useful
Researchers suggest that healthcare systems could do a better job of capturing a small number of clinically meaningful details. For example, electronic records might prompt clinicians to record factors such as whether a headache is new or longstanding, whether painkillers are being used frequently, or whether recognised warning signs are present.
Better alignment between clinical classifications and coding systems could also help ensure that common headache disorders are represented more accurately in routine datasets. 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 becoming overlooked when services are designed or resources allocated.
Professor Emsley said: "More detailed coding will not, by itself, cure headaches. 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."
