The Global Burden of Headache: A Consistent but More Detailed Picture – and a Wakeup Call for the Field, Too

Methodological refinements in the Global Burden of Disease Study (GBD) 2023 allow for a more precise estimate of the gender gap in headache, as well as a better understanding of the contribution of medication overuse headache to headache-attributed health loss and how headache burden changes with age. But GBD researchers are also alerting the field to a troubling trend that they don’t want to see continue.

The Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) initiative has become a mainstay for epidemiologists and others seeking to understand the impact of disease on health outcomes. To date, GBD efforts have examined hundreds of outcomes and risk factors, including headache.

Now, a systematic analysis of headache disorders for GBD 2023, the latest iteration of the GBD effort, adds important new details for a better understanding of the burden of headache disorders. But those involved in the effort are also pointing to what they say are fatally flawed studies that followed up on previous GBD updates – a warning call to the field that these secondary analyses be met with skepticism. They urge that the mistakes of the past not be repeated by those who try to publish their own papers based on the new GBD 2023 analysis.

Researchers including Andreas Husøy, Lars Stovner, and Timothy Steiner of the Norwegian University of Science and Technology in Trondheim, Norway, along with hundreds of GBD collaborators, report in the December 2025 issue of Lancet Neurology that the prevalence of headache disorders has remained stable from previous iterations of GBD.

However, while there were no new data sources included for GBD 2023, methodological refinements did allow for a new estimate of the gender gap in headache, with health loss from headache disorders being twice as high in females compared to males. The refinements also made possible an estimate of the contribution of medication overuse to headache burden, which accounted for a stunning one-fifth of years lived with disability (YLDs).

“As headache disorders remain a persistent driver of non-fatal health loss, the methodological rigour of GBD 2023 provides a roadmap for turning data into impactful evidence. Only by addressing both epidemiological and clinical gaps will the global burden of headache be alleviated,” wrote Ye Ran, Xiaopeng Wang, and Zhao Dong, Chinese PLA General Hospital, Beijing, China, in an accompanying editorial, also published in the December 2025 issue of Lancet Neurology. (Husøy and Steiner also penned an editorial on their findings in the Journal of Headache and Pain.)

But in another editorial published in Cephalalgia, Husøy, Stovner, and Steiner also draw attention to a disturbing trend that perhaps some in the headache field are unaware of. Specifically, they flagged more than two dozen papers whose authors, they say, did shoddy work when they published their own papers based on data for headache from GBD 2017 and onward. Here, Husøy and colleagues pointed to lax peer review, paper mills, and the use of artificial intelligence (AI) as key culprits for the shortcomings of those secondary analyses.

They also expressed concern that history might repeat itself now that GBD 2023 has been published – and the GBD initiative has already taken steps to prevent that from happening.

Females spend a lot more time with headache than males do
Husøy told Migraine Science Collaborative in an interview that he and his GBD 2023 collaborators were especially interested in accounting for how much time people spend with headache – something that previous GBD iterations did not take into consideration – and how that measure varied according to sex and age. To do so, they relied on a meta-analysis of individual participant data from more than 41,000 people from the general population of 18 countries across the world.

That information “is part of the Global Campaign against Headache, which is led by Lifting the Burden, an organization that I’m part of and some of my colleagues are as well,” according to Husøy. “During this campaign, we have conducted population-based studies in many different countries across the entire world, and we have data not only on prevalence, but also on how often people experience headache and how long their headaches [last]. Based on this data, we were able to calculate for each individual the amount of time they actually spend having headache.”

The researchers found that females spent more time with headache, including definite migraine, probable migraine, definite tension-type headache (TTH), or probable TTH, and did so across all age groups.

For example, for definite migraine, the average percentage of time in a year in which a female experienced headache was 9.3% for those under 35 years of age, 10.13% for those aged between 35 and 49 years, and 12.77% among those 50 years or older, compared to 6.34%, 7.86%, and 8.67%, respectively, for males. These data also reveal that, for both females and males, time spent with migraine increased with age.

“If you look only at individuals with migraine, and you look at the proportion of time they spend having headaches, then that increases more or less linearly with age,” Husøy said.

He suspects the reason for this is that “as you get older, if you have severe migraine, in terms of having many days per month with migraine, then you are less likely to lose your migraine. The other explanation would obviously be that some individuals experience a worsening of their migraine as they get older – more frequent and perhaps also longer attacks – but that seems to me less likely; the majority actually lose their migraine when they get older.”

Prevalence and years lived with disability (YLDs)
For overall prevalence, approximately 2.9 billion people had a headache disorder in 2023. That equates to an overall prevalence of 34.6%, with females having a higher prevalence (37.5%) than males (31.8%). YLDs reached 45.5 million globally, with a rate per 100,000 population that was twice as high among females than among males.

Breaking these statistics down specifically for migraine, there were 40.9 million YLDs in 2023, equating to a rate of 487.5 YLDs per 100,000 population. While that rate placed migraine eighth among all conditions considered by GBD 2023, compared to fifth place in 1990, the rate was similar across time. So the researchers stress that the migraine statistics have remained stable over the past few decades; changes in the rankings may simply reflect that the burden of other disorders has increased.

Overall migraine prevalence was 14.1%, being significantly higher among females (17.6%) than males (10.5%). And, although TTH was twice as prevalent as migraine, migraine accounted for 90% of YLDs in 2023.

Medication overuse headache (MOH): A startling contribution
One of the most striking findings from GBD 2023 is the eye-popping contribution of MOH.

For migraine, MOH accounted for 6% of the prevalence for males and 5% of the prevalence for females. But it’s the YLD metric for migraine that really drives home MOH’s outsized role: MOH was responsible for 22.6% of the YLD estimates for males, and 14.1% for females. Overall, a whopping one-fifth of headache-attributed YLDs came from MOH.

“It’s not because medication overuse headache is very prevalent. It is not – its prevalence is maybe around 1% in most countries. But the key here obviously is that people with medication overuse headache spend a very great proportion of their time with headache, so that’s why it contributes so much to the overall estimates of health loss. And it’s also a disabling headache,” Husøy said.

Knowing that MOH plays such a large role is important, as it’s an actionable finding, considering that MOH is preventable.

A sequel you don’t want to see
In previous years, the publication of new GBD findings spurred efforts by other researchers looking to use the data as the basis for new papers of their own – anyone who follows the headache literature closely will likely have seen at least some of those papers. But in their editorial, Husøy, Stovner, and Steiner argue that these secondary analyses are often highly problematic.

“A little over two years ago, I think this [problem] came on our radar, because we were given so many of these papers to peer review,” Husøy said.

In particular, Husøy and colleagues identified “errors and misinterpretations” in 29 papers that relied on data from GBD 2017 and after, many of which “display limited or no understanding of GBD,” as they wrote in their editorial.

They pointed to a host of problems. For instance, some of the questionable papers cited the COVID-19 pandemic or the introduction of CGRP inhibitors to explain purported changes in migraine prevalence over time – but both the pandemic and the advent of anti-CGRP therapy occurred after the last entry of new headache data into GBD. In fact, there have been no new headache data included since GBD 2017; rather, there have only been methodological refinements.

In reality, despite claims in the questionable papers that headache prevalence has increased, the new GBD 2023 publication confirmed that the proportion of the population with headache has actually remained stable.

“These papers end up making interpretations of the data saying that the prevalence of migraine is increasing rapidly, whereas it is actually not,” Husøy said.

Another significant problem with the secondary analyses, the researchers highlighted, is that those analyses often cite the variable quality of certain types of data, such as hospital records and census data, as the reason why prevalence estimates may vary. But those data sources do not actually inform GBD headache estimates, which are in fact based only on population-based studies.

Making matters worse, they argue, the questionable papers often seem to have the imprimatur of GBD, but that is not the case – the authors have simply taken the original GBD data and tried to make it their own.

“[M]any of the papers copy, republish and effectively claim ownership of vast quantities of GBD data (sometimes in tables extending over more than 100 pages) that the authors did nothing to collect,” the authors wrote in their editorial.

They also wrote that most of the 29 dubious papers “bear the hallmarks of artificial intelligence (AI)-generation.”

“It’s relatively easy to produce a new paper once they have set up their AI algorithms – you just feed in another set of data. [For instance], instead of hip arthritis, you feed in migraine and then you have a new paper out there. One of my colleagues, when he peer reviewed one of these manuscripts on GBD and migraine, he actually saw that where the word “migraine” should have been, “hip arthritis” was erroneously used instead,” Husøy said.

In general, the advent of AI is increasingly a problem for headache journals, according to Fred Cohen, a headache specialist and researcher at the Icahn School of Medicine at Mount Sinai in New York City. Cohen, who studies migraine epidemiology, is also an associate editor of Headache, so he has first-hand experience with this issue.

“We have been seeing – many journals have – a lot of papers that are submitted very quickly that we have suspicion of being written by AI,” Cohen told Migraine Science Collaborative.

Regarding the questionable GBD secondary analyses, Husøy and colleagues also pointed to a failure of peer review, though they cut reviewers some slack here, considering that the shenanigans can be hard to recognize.

“This is a problem that we have actually written [about] to the editors of the main headache journals and warned them about,” Husøy said. “But I guess it’s also hard from the editor’s perspective to keep track of all of this as well. And to be fair, a lot of these papers look like they have an authoritative stamp from GBD on them, saying that these are GBD papers. You immediately think that they are approved centrally from GBD and that their quality is up to standard, and so you basically just take it at face value and publish it. So I think a lot of reviewers and editors are fooled in that way.”

Husøy also said that the only solution at this time seems to be to limit the availability of GBD data.

“I’m leaning towards that, unfortunately,” Husøy said

It’s an approach that the Institute of Health Metrics and Evaluation at the University of Washington in Seattle, which leads the GBD initiative, is already taking: Researchers are no longer able to download as much GBD data as they might like. Further, some of the data for headache disorders will not be made available online at all.

Husøy added that there will still be enough material available for the general public, governments, and other stakeholders, “because these are very important data, for allocation of health care resources, so it needs to be out there somehow. But for other researchers to download the entire catalog and just re-analyze the data – I don’t think that should be possible.”

Neil Andrews is a science journalist and executive editor of the Migraine Science Collaborative.

Image credit: fotoyou/123RF Stock Photo.

References
Global, regional, and national burden of headache disorders, 1990–2023: a systematic analysis for the Global Burden of Disease Study 2023.
GBD 2023 Headache Collaborators
Lancet Neurol. 2025 Dec;24(12):1005-15.

Global burden of headache disorders: innovations, disparities, and priorities for action.
Ran et al.
Lancet Neurol. 2025 Dec;24(12):988-89.

GBD 2023: over twice the health loss from headache disorders among females compared with males, and one fifth of the loss is attributable to medication overuse.
Husøy AK, Steiner TJ.
J Headache Pain. 2025 Oct 21;26(1):227. 

GBD, secondary analyses, paper mills and abuse of a valuable but vulnerable resource: A problem that must be confronted.
Husøy et al.
Cephalalgia. 2025 Nov;45(11):3331024251392884.