A new global picture of pain
Pain is one of the most common reasons people seek medical care, yet comparing its prevalence across countries has remained surprisingly difficult. A major international study has now assembled self-reported pain data from more than six million people to map how different kinds of pain change throughout life.
Published in Nature Medicine on 5 October 2026, the analysis drew on 6,125,459 participants from 902 population-based data sources in 118 countries and territories. It provides a reference framework for comparing pain by age, sex, body region and national development.
One of its most striking findings challenges a common assumption about global inequality. Older adults in countries with lower human development reported substantially more bodily pain than those in countries with higher development.
The findings suggest that conventional estimates based primarily on diagnosed medical conditions may miss an important part of the pain burden in places with limited access to healthcare.
How researchers combined 902 data sources
Matt Fillingim and colleagues brought together individual-level information from 138 population-based study programmes conducted between 1990 and 2025.
The participants ranged in age from five to more than 100 years, and approximately 55% were female. Studies represented all inhabited continents, although coverage varied markedly between regions.
The researchers harmonised different survey questions and body maps into eleven anatomical pain categories. These included the head, face, neck or shoulder, back, stomach or abdomen, hip, knee and several limb and chest regions.
They used generalised linear mixed-effects models to estimate age-related patterns while accounting for the structure of the contributing cohorts and regions.
Flexible statistical curves allowed pain prevalence to rise or fall at different ages rather than forcing a straight-line relationship.
Estimates were standardised to a common past-month recall period and a general definition of pain. This step improved comparability but could not eliminate every difference between original surveys.
The team also developed an open-access benchmarking tool through which other researchers can compare datasets with the resulting reference curves.
Back pain was far more common than facial pain
Across the eleven anatomical sites, estimated pain prevalence ranged from approximately 2% for facial pain to 40% for back pain.
These figures describe population-level estimates across harmonised data, not the probability that a particular individual will develop a medical condition.
The distinction matters because back pain can arise from many causes, including muscle strain, injury and chronic disease. A self-report of back pain is not itself a diagnosis.
Different anatomical regions also followed different age patterns. Back, hip and knee pain generally increased through adulthood, with particularly rapid growth between ages 20 and 55.
Several other forms of pain reached their highest prevalence earlier. Headache peaked around age 34 and abdominal pain around age 23 in the modelled trajectories.
Neck or shoulder pain peaked near age 62, while elbow pain peaked near age 59.
Overall, seven of the eleven anatomical categories followed a curved pattern in which prevalence eventually declined rather than increasing indefinitely.
These differences challenge the idea that every form of pain becomes steadily more common with each additional year of life.
Women reported more pain at every body site
Women had a higher prevalence of pain than men across all eleven anatomical sites examined.
The reported female-to-male risk ratios ranged from 1.08 to 1.83, indicating that the magnitude of the difference varied considerably by location.
The largest difference was observed for facial pain, with a risk ratio of 1.83. Headache followed at 1.74, while abdominal pain had a ratio of 1.60.
A risk ratio of 1.83 means that the estimated prevalence was 83% higher in one group than the other, not that 83% of women experienced facial pain.
The study also found higher rates among women for high-intensity pain and generalised pain.
These patterns are descriptive associations. They do not isolate the roles of biological differences, working conditions, social expectations or reporting practices.
Future research will need to investigate how these factors interact in different countries and at different ages.
Pain does not simply rise throughout old age
The probability of reporting pain anywhere in the body increased across the lifespan, from approximately 30% at age five to 70% at age 100 or older.
However, the more severe outcomes did not follow exactly the same pattern.
High-intensity pain, defined as an average rating of at least seven on a ten-point scale, peaked at approximately age 50.
Generalised pain, measured across multiple body regions, peaked later, around age 70.
Among people who reported pain, average intensity peaked near age 48, at about 5.0 for women and 4.6 for men, before declining at older ages.
This distinction is important for prevention. Waiting until people reach advanced old age could miss periods when several types of pain are increasing most quickly.
The findings also demonstrate why measuring whether pain exists is different from measuring its intensity or distribution across the body.
The largest inequality appeared late in life
Researchers compared countries using the Human Development Index, which summarises dimensions of health, education and living standards.
From approximately age 40 onward, bodily pain increased more steeply in lower-development settings.
By age 80 or older, the estimated prevalence of any bodily pain was 30.6 percentage points higher in low-HDI settings than in high-HDI settings.
The study reported a 95% confidence interval from 26.9 to 34.3 percentage points for that difference.
Joint pain showed a corresponding gap of 21.0 percentage points, while back pain differed by 20.2 percentage points.
At the oldest ages, lower-back pain in low-development settings approached twice the prevalence estimated in high-development settings.
These comparisons concern country-level development and modelled pain prevalence. They do not prove that low income or poor services caused each individual’s pain.
Nevertheless, they raise questions about occupational exposure, access to treatment and the conditions in which people age.
Sub-Saharan Africa followed a distinctive trajectory
The analysis found that sub-Saharan Africa and several other regions tended to report comparatively low pain prevalence in early adulthood, followed by steeper increases later in life.
By older ages, prevalence in these regions often approached or exceeded the global reference curve.
This pattern differs from some high-development regions, where certain pain measures levelled off or declined in later adulthood.
Regional estimates should not be mistaken for national statistics for South Africa or any other individual country.
The study combined unevenly distributed sources, with an average of 2.1 data sources per country in sub-Saharan Africa compared with 19.4 in western Europe.
That imbalance underlines the need for stronger local surveys and more consistent long-term pain surveillance.
Health systems may need to consider pain prevention and management across working-age and older populations, rather than treating it only as a problem of advanced ageing.
Three risk factors explained only part of the burden
The researchers also examined smoking, obesity and low household income as potentially modifiable or addressable risk factors.
Across anatomical sites, their combined population-attributable fraction was estimated at 18.3% globally.
The estimate varied substantially, from 12.6% in sub-Saharan Africa to 27.1% in eastern Europe.
A population-attributable fraction estimates the share of burden associated with specified exposures under modelling assumptions. It is not a guarantee that eliminating those factors would prevent exactly that proportion of pain.
Obesity showed particularly strong associations with weight-bearing joints. The risk ratio was 1.91 for foot or ankle pain and 1.86 for knee pain compared with people of normal weight.
Low household income was associated with pain across body regions. For chest pain, the lowest-income group had more than twice the estimated risk of the highest-income group.
These patterns may help identify prevention priorities, but the comparatively low explained fraction in some regions suggests that other causes remain insufficiently understood.
Why self-reported pain matters
Pain has no single objective biological marker that captures everyone’s experience.
Health records can identify diagnosed conditions, but they may overlook people who experience pain without receiving medical assessment.
This problem can be particularly important where specialist care, imaging and diagnostic services are difficult to access.
Self-reported measures therefore offer a complementary perspective on population health.
However, answers can be affected by cultural interpretations, language, recall periods and the way questions are presented.
The study’s harmonisation process aimed to make these diverse reports comparable while acknowledging that perfect equivalence is impossible.
Combining self-reported pain with clinical and functional measures could provide a fuller picture of how pain affects everyday life.
Limitations and future directions
The analysis brought together observational surveys conducted over 35 years, rather than following every participant through their lifetime.
Age-related curves therefore describe estimated patterns across different people and cohorts. They do not demonstrate how any one person’s pain will change as they age.
Regional coverage was uneven, and some settings contributed relatively few datasets.
Questionnaires differed in their original definitions, time frames and body-location coding. Statistical harmonisation reduces but cannot remove this measurement uncertainty.
Associations with income, smoking and obesity cannot by themselves establish causation.
The study also focused on pain prevalence and related measures, not the effectiveness of specific treatments.
Future research could improve geographic coverage, investigate unexplained late-life inequalities and link pain trajectories with disability, employment and access to healthcare.
Longitudinal studies would help clarify whether apparent late-life declines in some pain categories reflect changes within individuals or differences between generations.
What the findings mean
The study provides a common reference point for understanding pain across ages and regions, using data from more than 6.1 million participants.
It shows that pain differs sharply by anatomical site, that women report higher prevalence across all eleven sites and that many pain trajectories rise most quickly before age 55.
Perhaps most importantly, it identifies substantial late-life inequalities between lower- and higher-development settings that deserve greater attention.
For researchers and policymakers, the findings support more inclusive surveillance and prevention strategies that consider the experiences of people who may never receive a formal diagnosis.
The broader lesson is that the burden of pain cannot be understood through age or diagnosed disease alone. Geography, socioeconomic circumstances and measurement choices also shape the global picture.
Source Information
Study: Global and regional reference curves for pain across the lifespan in 6.1 million individuals in 118 countries
Authors: Matt Fillingim, Christophe Tanguay-Sabourin, Lindsay Neuert and colleagues.
Journal: Nature Medicine
Published: 5 October 2026
Methods: Harmonised individual-level self-reported pain data from 6,125,459 participants across 902 population-based sources in 118 countries and territories, analysed using mixed-effects models.
