Guido et al. The Journal of Headache and Pain (2020) 21:52 https://doi.org/10.1186/s10194-020-01108-3 The Journal of Headache and Pain RESEARCH ARTICLE Open Access Pain rates in general population for the period 1991–2015 and 10-years prediction: results from a multi-continent age-period- cohort analysis Davide Guido1, Matilde Leonardi1*, Blanca Mellor-Marsá2, Maria V. Moneta2, Albert Sanchez-Niubo2,3, Stefanos Tyrovolas2,3,4, Iago Giné-Vázquez2,3, Josep M. Haro2,3, Somnath Chatterji5, Martin Bobak6, Jose L. Ayuso-Mateos3,7,8, Holger Arndt9, Ilona Koupil10,11, Jerome Bickenbach12,13, Seppo Koskinen14, Beata Tobiasz-Adamczyk15, Demosthenes Panagiotakos4 and Alberto Raggi1 Abstract Background: Pain is a common symptom, often associated with neurological and musculoskeletal conditions, and experienced especially by females and by older people. The aims of this study are to evaluate the temporal variations of pain rates among general populations for the period 1991–2015 and to project 10-year pain rates. Methods: We used the harmonized dataset of ATHLOS project, which included 660,028 valid observations in the period 1990–2015 and we applied Bayesian age–period–cohort modeling to perform projections up to 2025. The harmonized Pain variable covers the content “self-reported pain experienced at the time of the interview”, with a dichotomous (yes or no) modality. Results: Pain rates were higher among females, older subjects, in recent periods, and among observations referred to cohorts of subjects born between the 20s and the 60s. The 10-year projections indicate a noteworthy increase in pain rates in both genders and particularly among subjects aged 66 or over, for whom a 10–20% increase in pain rate is foreseen; among females only, a 10–15% increase in pain rates is foreseen for those aged 36–50. Conclusions: Projected increase in pain rates will require specific interventions by health and welfare systems, as pain is responsible for limited quality of subjective well-being, reduced employment rates and hampered work performance. Worksite and lifestyle interventions will therefore be needed to limit the impact of projected higher pain rates. Keywords: Pain, Projection, Bayesian age period cohort model, Headache disorders, Musculoskeletal disorders, Employment * Correspondence: matilde.leonardi@istituto-besta.it 1 Neurology, Public Health and Disability Unit, Fondazione IRCCS Istituto Neurologico Carlo Besta, Via Celoria 11, 20133, Milan, Italy Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 2 of 11 Introduction to 2025. These variations were analyzed by gender con- Pain is defined as “a distressing experience associated sidering age, periods (year of survey) and cohorts (birth with actual or potential tissue damage with sensory, year), as derived by the ATHLOS project (Ageing Tra- emotional, cognitive, and social components” [1] and is jectories of Health – Longitudinal Opportunities and one the most common symptoms. It can be associated Synergies) harmonized dataset, which includes data col- with many conditions, including some highly prevalent lected in different studies carried out in the five and disabling ones, such as headaches, musculoskeletal continents. diseases or injuries, but also to sequelae of other disor- ders, such as diabetes [2]. These conditions often have Materials and methods long-lasting duration: they begin in adulthood, endure ATHLOS project and the harmonized dataset for several years and often worsen, and are associated ATHLOS project was funded by the European Union’s with disability [3–5]. Pain rates are expected to rise due Horizon 2020 Research and Innovation Program, and it to the increased prevalence of musculoskeletal and aims to achieve a better understanding of the impact of neurological diseases, increased life expectancy and in- ageing on health. To achieve this result, a new cohort creased proportion of elderly in societies [6]. has been composed from harmonized datasets of exist- The results of epidemiological studies show that the ing international longitudinal cohorts related to health prevalence of pain-associated conditions is between 10% and ageing (see: https://github.com/athlosproject/athlos- and 40% approximately [7–17]: the only exception to project.github.io). The harmonized dataset includes re- this is for the one-year prevalence of tension-type head- cords of participants from 17 different studies and is ache, which peaked up to 60% [13–15]. Literature also fully described elsewhere [34]: most of these studies show that pain is more frequently experienced by fe- were run between 2000 and 2010 and have at least two males and by older people, which is consistent with the waves, but there are both older and more recent ones, as epidemiological presentation of the main drivers of pain, well as studies with one wave only. The harmonized i.e. headache and musculoskeletal disorders. Based on ATHLOS mega-dataset comprises approximately 411, these figures, it could be hypothesized that most of the 000 respondents. Most of the studies whose data are in- general population would report pain as a daily or near- cluded in ATHLOS dataset were from high-income daily problem. However, reliable information on the countries and upper-middle-income countries: the only presence and time trend of pain in the general popula- exceptions are India and Ghana. tion, irrespective of underlying health conditions, is lim- ATHLOS harmonized dataset is composed of a wide ited to few studies. A recent one, which analyzed 18-year range of variables covering a variety of health conditions, trends in the overall rates of noncancer pain prevalence sociodemographic variables, personal functioning and in the U.S., showed that the proportion of adults report- contextual factors, which are usually assessed in popula- ing pain due to painful health conditions increased from tion studies. The ATHLOS dataset variables were classi- 32.9% to 41.0% between 1997 and 98 and 2013–14 [18]. fied in the following domains: sociodemographic and Understanding such a trend, as well as the projections economic characteristics, lifestyle and health behaviors, towards future periods, is of relevance for the manage- health status and functional limitations, diseases, living ment of pain as a symptom, particularly for the risk of status, physical measures, psychological measures, la- overuse of medications such as opioids [18–20]. In boratory measures, social environment and life events, addition to this, issues connected to worse health out- and administrative information [34]. come, e.g. higher disability, reduced employment rates Pain is included within health status and functional and loss of productivity are also of relevance [21–28]. limitations and was measured in 14 out of the 17 studies, This is, in turn, an important determinant of economic with different approaches (please refer to supplementary burden, which is high because of both per-case cost and materials for information on the distribution of pain general prevalence of pain-related conditions [29–33]. variable across studies and countries). Some studies, e.g. However, while higher pain rates among females and the Collaborative Research on Ageing in Europe [35] among older subjects, and occasionally differences by and the China Health and Retirement Longitudinal period have been shown, with higher pain rates in recent Study [36], addressed it in terms of pain severity (i.e. periods [6, 18], information on pain rates by cohort is None, Mild, Moderate, Severe, Extreme or other similar basically lacking. The integration of gender, age, period formats). Other studies, e.g. the Australian Longitudinal and cohort data is of core importance to produce projec- Study of Aging [37] and the Survey of Health, Ageing tions on pain rates in the general population. and Retirement in Europe [38], dichotomously addressed The aims of this study are to evaluate the temporal the presence of pain (i.e. yes or no), sometimes address- variations of pain rates among the general populations ing the idea that pain is “often experienced”, such as in for the period 1991–2015 and to predict future rates up the Irish Longitudinal Study of Ageing [39]. The Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 3 of 11  harmonization procedure is aimed to generate inferen- Log λijk ¼ μijk ¼ μ þ αi þ β j þ γk tially equivalent content across studies so to make the content of variables collected in different studies uni- Here μ represented the general level (intercept), and form. In the case of pain variable, the content was “self- αi, βj, γk denoted age, period, and cohort random effects reported pain experienced at the time of the interview”, (to be estimated), respectively [45]. and the variable modality was dichotomous. A fully Bayesian approach based on Integrated Nested Laplace Approximations (INLA) was considered for Age period cohort analysis model fitting and inference [46]. A Bayesian APC model Age period cohort (APC) models are commonly used to provides a more robust methodology compared to a log- analyze and project rates [40, 41]. APC models account linear model, particularly for the prediction of future oc- for these processes on three time scales: age, year of sur- currence [47–49]. Indeed, in our case, Bayesian age– vey (period) and year of birth (cohort). The period and period–cohort modeling was also used to perform pro- cohort effects are both surrogates for exposure to exter- jections, and relative credible intervals (CI), of the pain nal factors. Period effects include environmental and symptom rates in the time interval 2015–2025, extrapo- diagnostic factors. For example, the introduction of a lating the trend of rate from 1990 to 2015 and consider- new diagnostic procedure may lead to a jump in disease ing the subjects at risk relative to the last period 2011– incidence across all age groups. Cohort effects represent 2015. For each projection, different degrees of CI were risk factors that change over time and may have a de- reported, the closer to the predicted mean being 10% CI layed effect on disease outcomes. For example, lifestyle and the largest being 95% CI. factors, such as alcohol and tobacco consumption, can The model had some a priori assumptions: (i) log-RR manifest themselves as cohort effects [41]. for each effect summed to zero over the observed inter- Data were organized by five-year periods from 1991 to val; (ii) the expected effects were hypothesized to be 2015, five-year age-groups from the group 31–35 to the constant, so that both the large and the small deviations 96–100 one and stratified by gender. In a preliminary from a constant rate are detected. The Bayesian APC analysis, age-specific trend rates by gender were com- model also considered hyper-parameters of random walk puted. Trend rates vary between 0 and 1, with higher type of first (RW1) and second order (RW2) [50] on values indicating higher proportion of pain reporting in which log-gamma prior distributions were elicited. We specific subgroups and observations, i.e. among males intentionally used highly non-informative log-gamma and females, by age, cohort and periods. Later, APC prior distributions (with parameters equal to 1 and models were fitted to analyze the joint effects of the age, 0.00005) in order to endorse and make more credibility period, and cohort, expressed in Rate Ratio (RR) terms. to harmonized ATHLOS dataset and avoid the impos- RR is a relative measure that allows to identify potential ition of assumptions for which no a priori knowledge protective (if RR < 1) and risk factors (if RR > 1), in this was available. A prior distribution on an overdispersion case related to particular ages, periods or cohorts. Pain parameter was also elicited by an independent log- counts yijk in age group i at time point j in kth cohort gamma with parameters equal to 1 and 0.005. can be assumed to be Poisson distributed, i.e., Parameter estimates in terms of mean and median  were obtained. Finally, to select the best model among yijk ¼ Po nijk λijk ; where i ¼ 1; …; I; j ¼ 1; …; J; k ¼ 1; …; K; the different proposals, the Deviance Information Criter- ion (DIC, lower was better) [51] were computed. In par- with mean nijkλijk, where nijk denotes the corresponding ticular, different specification as age period and age subjects at risk. In our application, the age index i run cohort models, and different combinations of priors of from 1 to I = 14 (both in males and females), while the RW1 and RW2 [50] were probed by considering the ran- period index j run from 1 to J = 5. Concerning the co- dom effects of age, period and cohort. In case RW1 and hort index k, it depended on the age group and period RW2 priors provided very similar DICs, we decided to index, but also on their intervals width [42, 43] and it select the complete model with RW2 priors because is a was defined as M × (I − i) + j, where M indicated the ratio standard choice as natural target for smoothing [52, 53]. between the width in years of the age group and the The technique was implemented in the software R v period intervals [44]. In our application M was equal to [5 3.5.2 [54] through the packages R-INLA (www.r-inla. years] / [5 years] = 1, and the cohort index k, following the org) [55–57] and Bayesian APC [50] for the projections. previous formula where the age groups were fourteen and the periods are five, generated K = 1 × (14–1) + 5 = 18 co- Results horts, i.e., the cohort index run from 1 to K = 18, but only In total 660,028 valid observations were used, of whom 9 not overlapping (i.e., 1891–1900, 1901–1910,…,1971– 293,484 reported some degree of pain, the rate being 1980). Given that, the model was specified as 44.5% (see Table 1): the rate of pain was higher among Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 4 of 11 Table 1 Presence of pain in the general population included in Age-specific rates by gender and per period are ATHLOS harmonized dataset by age class shown in Fig. 2. Rates were higher among females Age class Males Females Total and, in both genders, were higher among older sub- 31–50 Observations 18,287 37,510 55,797 jects. A trend related to the period was observed, in Cases with pain 6085 14,886 20,971 both genders and across all age groups, with partici- pants recruited in more recent periods experiencing % cases with pain 33.3% 39.7% 37.6% pain with a higher frequency. An exception to this is 51–70 Observations 182,094 223,289 405,383 observed for participants aged below 50 for whom, in Cases with pain 70,425 106,939 177,364 the period 2010–2015, a decline in trend was ob- % cases with pain 38.7% 47.9% 43.8% served in both genders. > 70 Observations 83,389 115,459 198,848 Figure 3 shows the results of the APC models. The ef- Cases with pain 34,331 60,818 95,149 fect of age was different among males and females: in fact, while a consistent increasing trend was observed in % cases with pain 41.2% 52.7% 47.8% females, with age becoming a risk factor for pain in the Total Observations 283,770 376,258 660,028 seventh decade of life, the age effect in males decreased Cases with pain 110,841 182,643 293,484 from the first age group up to the 50–55 group and then % cases with pain 39.1% 48.5% 44.5% it increased, becoming a risk factor in the eighth decade of life. The period effect was constantly growing in both observations referred to females than to males. As males and females and it became a risk factor for pain shown in Table 2, the average age of participants at the from the period 2006–2010 onwards. The cohort effect time of observation was comprised between 64 and 65 had a parabolic shape: being born between the 20s and years across all periods, with the exception of the 1991– the 60s was a risk factor for pain in both genders. The 1995 one, where the average age was approximately 5 selected models for both males (DIC = 650.6) and fe- years below that of the other periods. Supplementary males (DIC = 723.8) involved the complete age period Tables S1-S4 report distribution of all observations and cohort models and the choice of RW2 priors (see sup- of observations with pain by period and by 5-years age plementary Table S9). It is worth to point out that the groups; supplementary Tables S5-S8 report the same in- estimated decreasing trend for the youngest age groups formation not aggregated by 5-years periods. in males could be due to the small number of periods in Figure 1 reports the observed pain rates by age, period relation to the age groups. In order to account for that, and cohort for males and females. Among males rates by we have also performed a sensitivity analysis where the age varied approximately between 0.3 and 0.4, increasing periods were non-aggregated (data are included in sup- with age and remaining basically stable after 80 years of plemental materials, Tables S5-S8). The results of this age; among females, rates by age varied approximately last analysis were very similar to that of those carried between 0.4 and 0.54, peaking at the age class 81–85 out with aggregated data: only for the period effects dif- and declining thereafter. With regard to rates by periods, ferences on the trends that presented a more irregular a consistent increase was observed in both genders: increase were detected (see Fig. S1). Moreover, because among males it varied between 0.2 and 0.4, among fe- of data sparsity, sensitivity analysis could not produce males between 0.3 and 0.5. Regarding rates by cohort, it projection as estimates intervals would have been too has parabolic shapes in both genders: among males the wide and unstable. peak was around 0.4 for the cohort born in 1941–1950, Figures 4 and 5 show observed and predicted rates for among females it was around 0.5 for the cohorts born all age groups and for males and females, respectively. between 1921 and 1950. For both males and females, the projection is indica- tive of an increase in pain rates between 2015 and 2025, and such an increase was wider among older Table 2 Age by period subjects. Among males, the projected increase over Period Males Females Total the decade was higher than 10% in all age groups 1991–1995 60.6 ± 8.7 58.5 ± 10.1 59.4 ± 9.6 from the 66–70 one, and get close to 20% for the age 1996–2000 64.8 ± 11.0 64.8 ± 12.5 64.8 ± 11.9 groups 91–95 and 96–100 (Fig. 4m-n). The same ap- 2001–2005 64.6 ± 10.2 64.1 ± 11.5 64.3 ± 10.9 plies to females, for whom however the projection over the decade was higher than 20% in the oldest 2006–2010 65.1 ± 10.9 65.0 ± 11.9 65.1 ± 11.5 group (Fig. 5n). Moreover, an increase comprised be- 2011–2015 64.9 ± 10. 7 64.8 ± 11.5 64.8 ± 11.2 tween 10% and 15% over the decade was also ob- Total 64. 7 ± 10.6 64.4 ± 11.7 64.5 ± 11.3 served among females for the age groups comprised Note. Data are expressed in years as mean ± standard deviations between 36 and 50 years (Fig. 5b-d). Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 5 of 11 Fig. 1 Observed rate of pain in males and females by age, period and cohort. a: pain rates by age. b: pain rates by period. c: pain rate by cohort. Note. The dark central line indicates rate = 0.5 Discussion with age and through the years as an effect of population The results of this study confirm part of the finding of ageing [58, 59]. In addition to the known effect of age, previous literature, namely higher rates of pain among some reports have also shown that people enrolled in re- females and among older subjects, and reinforce the few cent years in population studies tend to report higher available data on period analysis by showing higher pain pain rates, i.e. an effect of age [6, 18] which may underlie rates in more recent periods, with specific reference to differences in pain reporting, that might be due to cul- the observations referred to older subjects. In addition tural or biological issues. Our data do not enable to to this, our results show that pain rates were higher make hypotheses on what is causing such a period effect, among the observations referred to cohorts of subjects but it has to be acknowledged that the prevalence of born between the 20s and the 60s (i.e. between 1921 and some conditions is rising worldwide. Examples of this in- 1970). Finally, the most innovative aspect of our work clude diabetic neuropathy, which affects up to 50% of lies in the 10-year projection for pain rates, which indi- patients with diabetes [60] and is expected to increase in cate a noteworthy increase in pain rates across all age consideration of the rising prevalence of diabetes world- groups and in both genders. Among males aged 66 and wide [61, 62], and musculoskeletal conditions, which are over, the increase is projected to be by 10–20% over the among the main pain drivers, and represent approxi- decade, and the same is for females: in addition to this, mately 19% of non-communicable diseases in terms of however, also females aged 36–50 will likely face a 10– prevalence and 20% in terms of disability [2]. 15% increase in pain rates. In consideration of the higher impact of such condi- Taken as a whole, our results tell the story of an ex- tions among older subjects, and in consideration of glo- pansion of pain-related morbidity, irrespective of its aeti- bal population ageing, actions aimed to prevent and ology. Chronic pain prevalence has been occasionally control pain are and will be more and more needed. It is addressed, but available reports show that it is increasing interesting to notice the increase in pain rates which is Fig. 2 Observed pain rate in males and females in the ATHLOS dataset. a: age-specific pain rates by age for males. b: age-specific pain rates by age for females Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 6 of 11 Fig. 3 Estimated effects of age, period and cohort in rate ratio terms by gender. a: estimated effect of age. b: estimated effect of period. c: estimated effect of cohort. Note. The dark central line indicates RR = 1 specific for females aged 36–50. This age group is at 2020–2030 [69]. Therefore, it can be hypothesized that higher risk of having migraine or tension-type headache headache disorders will contribute more and more to [3, 63–66]. It has to be noted that, as shown by GBD the overall burden of pain in general populations. In data [3], the decrease in prevalence of headache disor- addition to this, headache disorders and back pain, ders begins after the age of 30, but the decline is particu- which are among the most prevalent and disabling con- larly pronounced after the age of 50, particularly among ditions [2], show considerable rates of comorbidity, with women; this may be the reason for the differential in- odds of association comprised between 1.55 and 8.00 crease by gender in the 36–50 age group. In addition to [70]. this, there are biological differences that make women to It is reasonable to hypothesize that in the older age be exposed to specific pain, such as cyclic menstrual groups, the difference in the increase of pain might be pain or menstrual migraine, which affects 22% to 62% of associated to the higher impact of some skeletal condi- female migraineurs [67, 68]. Such an increase in pain tions, such as osteoporosis, and to the associated higher rates will also be accompanied by an increase in the fracture rates. In fact, osteoporosis is approximately amount of the portion of population in that age group three-fold more prevalent in women than in men [71], (third to fifth decade of life) in high and upper-middle and women older than 60 years have a 44% lifetime risk income countries. In fact, in high-income countries, the of fractures, compared to 25% among men of the same median age between 1990 and 2015, moved from 33.4 to age [72]. 40.4, and is projected to span between 41.5 and 43.5 in Increasing pain rates will most likely be associated 2020–2030; in upper-middle income countries, the me- with increased healthcare utilization and consumption of dian age between 1990 and 2015, moved from 24.5 to different kinds of analgesics [73–75], and in particularly 33.9, and is projected to span between 35.6 and 39.9 in opioids, which however may in turn produce negative Fig. 4 Pain rates and 10-years pain rate projection per age group in males. a: pain rate projection, males, age 31–35. b: pain rate projection, males, age 36–40. c: pain rate projection, males, age 41–45. d: pain rate projection, males, age 46–50. e: pain rate projection, males, age 51–55. f: pain rate projection, males, age 56–60. g: pain rate projection, males, age 61–65. h: pain rate projection, males, age 66–70. i: pain rate projection, males, age 71–75. j: pain rate projection, males, age 76–80. k: pain rate projection, males, age 81–85. l: pain rate projection, males, age 86–90. m: pain rate projection, males, age 91–95. n: pain rate projection, males, age 96–100. Note. The predictive mean is shown as solid line. The different shadings indicate pointwise credible intervals. The central interval represents 10% CI, and the largest interval 95% CI. Observed rates are shown as a filled circle. The vertical dashed line indicates where prediction started Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 7 of 11 Fig. 5 Pain rates and 10-years pain rate projection per age group in females. a: pain rate projection, females, age 31–35. b: pain rate projection, females, age 36–40. c: pain rate projection, females, age 41–45. d: pain rate projection, females, age 46–50. e: pain rate projection, females, age 51–55. f: pain rate projection, females, age 56–60. g: pain rate projection, females, age 61–65. h: pain rate projection, females, age 66–70. i: pain rate projection, females, age 71–75. j: pain rate projection, females, age 76–80. k: pain rate projection, females, age 81–85. l: pain rate projection, females, age 86–90. m: pain rate projection, females, age 91–95. n: pain rate projection, females, age 96–100. Note. The predictive mean is shown as solid line. The different shadings indicate pointwise credible intervals. The central interval represents 10% CI, and the largest interval 95% CI. Observed rates are shown as a filled circle. The vertical dashed line indicates where prediction started health outcomes and a further increase of healthcare use aims of this study, but some general considerations can be [76, 77]. In addition to this, pain is associated with re- made. Evidence exist that effective interventions act upon duced work productivity and sick leave [22, 26, 27, 78– different levels, i.e. both patients and the workplaces, and 81], early retirement and reduced employment rates [21, involve several stakeholders, including treating pain spe- 25, 28, 82, 83]. The old-age dependency ratio, i.e. the ra- cialists, occupational physicians and employers, and the tio of the population aged 65 years or over to the popu- most commonly reported strategies include appropriate lation aged 15–64 -- referred to as the number of pharmacotherapy, provision of ergonomic furniture, pa- dependants per 100 persons of working age -- has been tients’ education on pain management, relaxation/posture constantly increasing and is projected to further increase exercises. Lifestyle interventions are aimed to reduce fac- in both high-income and upper-middle-income coun- tors contributing to pain exacerbation and pain triggers, tries. In high-income countries the dependency ratio in- and most of them target diet and physical activity, as well creased from 18.3 to 25.7 between 1990 and 2015, and is as appropriate drug prescription and adherence to treat- projected to span between 25.8 and 36.3 between 2020 ment. Finally, preliminary research evidence is available and 2030; in upper-middle-income countries the in- on the effectiveness of behavioral treatments, alone or as crease was from 8.1 to 10.5 between 1990 and 2015, and add-on to medical ones, for chronic pain control [93–95]. the rate will span between 12.3 and 16.2 between 2020 More research is needed in this field, that seems however and 2030 [69]. Thus, our projections are not only indica- promising and could provide a substantial contribution to tive of an increased risk of future worse health status, control pain experience and reduce the consumption of but also of higher healthcare expenditure and overall fi- medications. nancial burden which the health and welfare systems of Some limitations have to be acknowledged. The age high and upper-middle income countries, i.e. countries period cohort models are commonly used to analyse and from which most of observations included in ATHLOS project mortality or morbidity rates from health registers dataset were drawn [34], likely will struggle to bear. to routinely collect demographic rates. However, the Interventions are therefore needed to prevent pain and data herein reported do not derive from registries but limit the impact of projected higher pain rates, which from cross-sectional and longitudinal surveys, harmo- should include worksite interventions [84–89], lifestyle in- nized in the ATHLOS dataset. This entails that the terventions and control over prescribed drugs [89–92]. population is not the same across periods, and therefore Worksite intervention should be aimed to limit work ces- the sample sizes were different. Moreover, our analysis sation as well as enhance work productivity, and should did not consider population dynamics either. In addition ideally to be tailored to the features of each individual. to this, it has to be remembered that our analysis is Practically, these interventions can be tailored to the fea- based on a harmonized dataset, whose core definition of tures of specific disorders as well, and may thus involve pain variable is “self-reported pain experienced at the specific actions. A synthesis of the most effective strategies time of the interview”, with a dichotomous output. This included among available literature [84–92] is beyond the creates an important limitation, namely the fact that this Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 8 of 11 variable does not account for two issues of relevance in A.M.P.). The authors gratefully acknowledge the work of the 10/66 Dementia pain experience. The first is the severity and impact of Research Group who provided data for this paper. - The Australian Longitudinal Study of Ageing (ALSA): The ALSA study was pain, which spans between mild and disabling; the sec- supported by grants from the South Australian Health Commission, the ond is the frequency with which pain is experienced, Australian Rotary Health Research Fund, the US National Institute on Aging which might be daily or near-daily, such as pain due to (Grant No. AG 08523–02), the Office for the Ageing (SA), Elderly Citizens Homes (SA), the National Health and Medical Research Council (NH&MRC musculoskeletal conditions, episodic with variable fre- 22922), the Premiers Science Research Fund (SA) and the Australian Research quency, such as in the case of headache disorders, or oc- Council (DP0879152; DP130100428). The authors gratefully acknowledge the casional. Finally, we interpreted the results of pain work of the project team at the Flinders Centre for Ageing Studies, Flinders University who provided data for this paper. trends in light of the trends that can be reasonably ex- - The China Health and Retirement Longitudinal Study (CHARLS): The pected for musculoskeletal and headache disorders in CHARLS study has received critical support from Peking University, the reason of their prevalence in the general population. National Natural Science Foundation of China, the Behavioral and Social Research Division of the National Institute on Aging and the World Bank. The However, none of the original pain-related variables in- authors gratefully acknowledge the work of the project team at the Peking cluded specification of pain in terms of aetiology or University who provided data for this paper. location. - Collaborative Research on Ageing (COURAGE) in Europe: The COURAGE study was supported by the European Community’s Seventh Framework Programme (FP7/2007–2013) under grant agreement number 223071 Conclusion (COURAGE in Europe). Data from Spain were also collected with support from the Instituto de Salud Carlos III-FIS research grants number PS09/00295, In conclusion, we reported data on the temporal varia- PS09/01845, PI12/01490, PI13/00059, PI16/00218 and PI16/01073; the Spanish tions of pain rates among the general population for the Ministry of Science and Innovation ACI-Promociona (ACI2009-1010); the Euro- period 1991–2015 and predicted 10-years future rates. pean Regional Development Fund (ERDF) ‘A Way to Build Europe’ grant numbers PI12/01490 and PI13/00059; and by the Centro de Investigación Our results are based on a very large international data- Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Salud Carlos III. set, which included data from populations from the five Data from Poland were collected with support from the Polish Ministry for continents. Results show that pain rates were higher Science and Higher Education grant for an international co-financed project (number 1277/7PR/UE/2009/7, 2009–2012) and Jagiellonian University Med- among females and among older subjects, and that rates ical College grant for project COURAGE-POLFUS (K/ZDS/005241). The authors were also higher among the respondents enrolled in gratefully acknowledge the work of COURAGE researchers who provided more recent periods. Finally we showed that the trends data for this paper. - The English Longitudinal Study of Ageing (ELSA): ELSA is supported by the of pain are increasing, in particular among females and U.S. National Institute of Aging, the National Centre for Social Research, the among older subjects, for whom a 10–20% increase is University College London (UCL) and the Institute for Fiscal Studies. The projected over the 10-year period. authors gratefully acknowledge the UK Data Service and UCL who provided data for this paper. Pain is strongly associated with reduced employment - The Seniors-ENRICA: The Seniors-ENRICA cohort was funded by an uncondi- rates and hampered work performance which, in consider- tional grant from Sanofi-Aventis, the Ministry of Health of Spain, FIS grant 12/ ation of population ageing and projected increase of de- 1166 (State Secretary for R + D and FEDER-FSE) and the Centro de Investiga- ción Biomédica en Red de Epidemiología y Salud Pública (CIBERESP), Instituto pendency ratio of older people on people of working age, de Salud Carlos III. The authors gratefully acknowledge the work of the pro- will require specific actions by health and welfare systems. ject team at the Universidad Autónoma de Madrid who provided data for Worksite and lifestyle interventions will be needed to limit this paper. - The Health, Alcohol and Psychosocial factors In Eastern Europe (HAPIEE) the impact of projected higher pain rates. study: The HAPIEE study was supported by the Wellcome Trust [grant numbers WT064947, WT081081], the US National Institute of Aging [grant number 1RO1AG23522] and the MacArthur Foundation Initiative on Social Supplementary information Upheaval and Health. The authors gratefully acknowledge the work of the Supplementary information accompanies this paper at https://doi.org/10. project teams at University College London, the National Institute of Public 1186/s10194-020-01108-3. Health in Prague, the Jagiellonian University Medical College in Krakow and the Kaunas University of Medicine who provided data for this paper. Additional file 1. Trends Pain – Supplementary Materials – February 10, - The Health 2000/2011 study: The authors gratefully acknowledge the 2020. supplementary results not included in the full text. Finnish Institute for Health and Welfare who provided data for this paper. - Health and Retirement Study (HRS): The HRS study is supported by the National Institute on Aging (grant number NIA U01AG009740) and the Social Abbreviations Security Administration, and is conducted by the University of Michigan. The AP: Age Period Cohort; ATHLOS: Ageing Trajectories of Health – Longitudinal authors gratefully acknowledge the University of Michigan who provided Opportunities and Synergies; CI: Credible Intervals; DIC: Deviance Information data for this paper. Criterion; INLA: Integrated Nested Laplace Approximations; RR: Rate Ratio; - The Korean Longitudinal Study of Ageing (KLOSA): The KLOSA study is RW1: Random Walk type of first order; RW2: Random Walk type of second funded by the Korea Employment Information Service (KEIS) and was order supported by the Korea Labor Institute’s KLOSA Team. The authors gratefully acknowledge the KEIS who provided data for this paper. - The Mexican Health and Aging Study (MHAS): The MHAS study is partly Acknowledgements sponsored by the National Institutes of Health/National Institute on Aging The ATHLOS project researchers are grateful for data contribution and (grant number NIH R01AG018016) and the INEGI in Mexico. The authors funding in the following studies: gratefully acknowledge the MHAS team who provided data for this paper - The 10/66 study (10/66): The 10/66 study is supported by the Wellcome retrieved from www.MHASweb.org Trust (GR066133/ GR080002), the European Research Council (340755), US - The Study on Global Ageing and Adult Health (SAGE): The SAGE study is Alzheimer’s Association, WHO, FONDACIT (Venezuela) and the Puerto Rico funded by the U.S. National Institute on Aging and has received financial State Government, and the Medical Research Council (MR/K021907/1 to Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 9 of 11 support through Interagency Agreements (OGHA 04034785; YA1323-08-CN- Health Sciences, Centre for Health Equity Studies, Stockholm University, 0020; Y1-AG-1005–01) and Grants (R01-AG034479; IR21-AG034263-0182). The Stockholm, Sweden. 11Department of Public Health Sciences, Karolinska authors gratefully acknowledge the World Health Organization who provided Institutet, Stockholm, Sweden. 12Department of Health Sciences and Health data for this paper. Policy, University of Lucerne, Lucerne, Switzerland. 13Swiss Paraplegic - The Survey of Health, Ageing and Retirement in Europe (SHARE): The Research, Nottwil, Switzerland. 14Finnish Institute for Health and Welfare SHARE study is funded by the European Commission through FP5 (QLK6-CT- (THL), Helsinki, Finland. 15Department of Epidemiology and Population 2001–00360), FP6 (SHARE-I3: RII-CT-2006–062193, COMPARE: CIT5-CT-2005– Studies, Faculty of Health Sciences, Jagiellonian University Medical College, 028857, SHARELIFE: CIT4-CT-2006–028812) and FP7 (SHARE-PREP: N°211909, Krakow, Poland. SHARE-LEAP: N°227822, SHARE M4: N°261982). Additional funding from the German Ministry of Education and Research, the Max Planck Society for the Received: 12 February 2020 Accepted: 13 April 2020 Advancement of Science, the U.S. National Institute on Aging (U01_AG09740-13S2, P01_AG005842, P01_AG08291, P30_AG12815, R21_AG025169, Y1-AG-4553–01, IAG_BSR06-11, OGHA_04-064, HHSN271201300071C) and from various national funding sources is gratefully References acknowledged (see www.share-project.org). 1. Williams AC, Craig KD (2016) Updating the definition of pain. Pain. 157: - The Irish Longitudinal study on Ageing (TILDA): The authors gratefully 2420–2423 acknowledge the Trinity College Dublin and the Irish Social Science Data 2. GBD (2017) Disease and Injury Incidence and Prevalence Collaborators Archive (www.ucd.ie/issda) who provided data for this paper. (2018) Global, regional, and national incidence, prevalence, and years lived Alberto Raggi is supported by a grant from the Italian Ministry of Health with disability for 354 diseases and injuries for 195 countries and territories, (Ricerca Corrente, Fondazione Istituto Neurologico C. Besta, Linea 1990–2017: a systematic analysis for the Global Burden of Disease Study 4—Outcome Research: dagli Indicatori alle Raccomandazioni Cliniche). 2017. Lancet 392:1789–1858 Stefanos Tyrovolas was supported by the Foundation for Education and 3. GBD (2016) Headache Collaborators (2018) Global, regional, and national European Culture, the Miguel Servet programme (reference CP18/00006), burden of migraine and tension-type headache, 1990–2016: a systematic and the Fondos Europeos de Desarrollo Regional. analysis for the Global Burden of Disease Study 2016. Lancet Neurol 17:954– 976 4. GBD (2016) Traumatic Brain Injury and Spinal Cord Injury Collaborators Authors’ contributions (2019) Global, regional, and national burden of traumatic brain injury and DG, planned and ran the analyses; ML, planned the analyses and revised the spinal cord injury, 1990–2016: a systematic analysis for the Global Burden of manuscript for intellectual content; BMM, MVM, ASN, ST, IGV, JMH, SC, MB, Disease Study 2016. Lancet Neurol 18:56–87 JLAM, HA, IK, JB, SK, BTA and DP revised the manuscript for intellectual 5. Sebbag E, Felten R, Sagez F, Sibilia J, Devilliers H, Arnaud L (2019) The content; AR, planned the analyses, drafted the manuscript and revised it for world-wide burden of musculoskeletal diseases: a systematic analysis of the intellectual content. The author(s) read and approved the final manuscript. World Health Organization burden of diseases database. Ann Rheum Dis 78: 844–848 Funding 6. Ahacic K, Kareholt I (2010) Prevalence of musculoskeletal pain in the general The ATHLOS project (Ageing Trajectories of Health: Longitudinal Swedish population from 1968 to 2002: age, period, and cohort patterns. Opportunities and Synergies) has received funding from the European Pain. 151:206–214 Union’s Horizon 2020 research and innovation programme under grant 7. Failde I, Dueñas M, Ribera MV, Gálvez R, Mico JA, Salazar A et al (2018) agreement No 635316. The funding body had no role in the design of the Prevalence of central and peripheral neuropathic pain in patients attending study, analysis and interpretation of data, and in writing the manuscript. pain clinics in Spain: factors related to intensity of pain and quality of life. J Pain Res 11:1835–1847 Availability of data and materials 8. Chenaf C, Delorme J, Delage N, Ardid D, Eschalier A, Authier N (2018) The datasets generated and/or analysed during the current study are not Prevalence of chronic pain with or without neuropathic characteristics in publicly available due restrictions imposed by part of the owners (COURAGE France using the capture-recapture method: a population-based study. Pain. in Europe, HAPIEE, ALSA, the Health 2000 and 2011 Surveys- Finland, ENRICA 159:2394–2402 and the 10/66 study), but may be available from the corresponding author 9. French HP, Smart KM, Doyle F (2017) Prevalence of neuropathic pain in upon reasonable request and once consent form ATHLOS project intellectual knee or hip osteoarthritis: a systematic review and meta-analysis. Semin property and dissemination board is obtained. Arthritis Rheum 47:1–8 10. Mansfield KE, Sim J, Jordan JL, Jordan KP (2016) A systematic review and Ethics approval and consent to participate meta-analysis of the prevalence of chronic widespread pain in the general Not applicable for ATHLOS study: ethics approval and participants’ consent population. Pain. 157:55–64 to participate were obtained by each study at the time point of surveys’ field 11. Burch RC, Loder S, Loder E, Smitherman TA (2015) The prevalence and trials. burden of migraine and severe headache in the United States: updated statistics from government health surveillance studies. Headache. 55:21–34 Consent for publication 12. Mundal I, Gråwe RW, Bjørngaard JH, Linaker OM, Fors EA (2014) Prevalence Not applicable. and long-term predictors of persistent chronic widespread pain in the general population in an 11-year prospective study: the HUNT study. BMC Competing interests Musculoskelet Disord 15:213 The authors declare that they have no competing interests. 13. Straube A, Aicher B, Förderreuther S, Eggert T, Köppel J, Möller S et al (2013) Period prevalence of self-reported headache in the general population in Author details Germany from 1995-2005 and 2009: results from annual nationwide 1 Neurology, Public Health and Disability Unit, Fondazione IRCCS Istituto population-based cross-sectional surveys. J Headache Pain. 14:11 Neurologico Carlo Besta, Via Celoria 11, 20133, Milan, Italy. 2Parc Sanitari Sant 14. Ayzenberg I, Katsarava Z, Sborowski A, Chernysh M, Osipova V, Tabeeva G et Joan de Déu, Fundacion Sant Joan de Deu, Barcelona, Spain. 3Instituto de al (2012) The prevalence of primary headache disorders in Russia: a Salud Carlos III, Centro de Investigación Biomédica en Red de Salud Mental, countrywide survey. Cephalalgia. 32:373–381 CIBERSAM, Madrid, Spain. 4Department of Nutrition and Dietetics, School of 15. Stovner LJ, Andree C (2010) Prevalence of headache in Europe: a review for Health Science and Education, Harokopio University, Athens, Greece. the Eurolight project. J Headache Pain. 11:289–299 5 Information, Evidence and Research, World Health Organization, Geneva, 16. Linde M, Stovner LJ, Zwart JA, Hagen K (2011) Time trends in the Switzerland. 6Research Department of Epidemiology and Public Health, prevalence of headache disorders. The Nord-Trondelag health studies University College London, London, UK. 7Department of Psychiatry, (HUNT 2 and HUNT 3). Cephalalgia. 31:585–596 Universidad Autónoma de Madrid, Madrid, Spain. 8Hospital Universitario de 17. Lj S, Hagen K, Jensen R, Katsarava Z, Lipton R, Scher A et al (2007) The La Princesa, Instituto de Investigación Sanitaria Princesa, Madrid, Spain. global burden of headache: a documentation of headache prevalence and 9 SPRING TECHNO GMBH & Co. KG, Bremen, Germany. 10Department of Public disability worldwide. Cephalalgia. 27:193–210 Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 10 of 11 18. Nahin RL, Sayer B, Stussman BJ, Feinberg TM (2019) Eighteen-year trends in 40. Riebler A, Held L, Rue H (2012) Estimation and extrapolation of time trends the prevalence of, and health care use for, noncancer pain in the United in registry data—borrowing strength from related populations. Ann Appl States: data from the medical expenditure panel survey. J Pain 20:796–809 Stat 6:304–333 19. Harrison JM, Lagisetty P, Sites BD, Guo C, Davis MA (2018) Trends in 41. Smith TR, Wakefield J (2016) A review and comparison of age–period– prescription pain medication use by race/ethnicity among US adults with cohort models for cancer incidence. Stat Sci 31:591–610 noncancer pain, 2000-2015. Am J Public Health 108(6):788–790 42. Holford TR (1983) The estimation of age, period and cohort effects for vital 20. Luo X, Pietrobon R, Hey L (2004) Patterns and trends in opioid use among rates. Biometrics 39:311–324 individuals with back pain in the United States. Spine (Phila Pa 1976) 29: 43. Holford T (2005) Age-period-cohort analysis. In: Armitage P, Colton T (eds) 884–890 Encyclopaedia of biostatistics, 2nd edn. Wiley, West Sussex, pp 105–123 21. Ropponen A, Narusyte J, Mittendorfer-Rutz E, Svedberg P (2019) Number of 44. Heuer C (1997) Modeling of time trends and interactions in vital rates using pain locations as predictor of cause-specific disability pension in Sweden- restricted regression splines. Biometrics 53:161–177 do common mental disorders play a role? J Occup Environ Med 61:646–652 45. Clayton D, Schifflers E (1987) Models for temporal variation in cancer rates. 22. Porter JK, Di Tanna GL, Lipton RB, Sapra S, Villa G (2019) Costs of acute II: age-period-cohort models. Stat Med 6:469–481 headache medication use and productivity losses among patients with 46. Rue H, Martino S, Chopin N (2009) Approximate Bayesian inference for migraine: insights from three randomized controlled trials. Pharmacoecon latent Gaussian models by using integrated nested Laplace approximations. Open 3:411–417 J R Stat Soc Ser B Stat Methodol 71:319–392 23. Raggi A, Corso B, De Torres L, Quintas R, Chatterji S, Sainio P et al (2018) 47. Eilstein D, Uhry Z, Lim TA, Bloch J (2008) Lung Cancer mortality in France. Determinants of mobility in populations of older adults: results from a Trend analysis and projection between 1975 and 2012, using a Bayesian cross-sectional study in Finland. Poland and Spain Maturitas 115:84–91 age-period-cohort model. Lung Cancer 59:282–290 24. Raggi A, Corso B, Minicuci N, Quintas R, Sattin D, De Torres L et al (2016) 48. Knorr-Held L, Rainer E (2001) Projections of lung cancer mortality in West Determinants of quality of life in ageing populations: results from a cross- Germany: a case study in Bayesian prediction. Biostatistics 2:109–129 sectional study in Finland. Poland and Spain PLoS One 11:e0159293 49. Girardi P, Bressan V, Merler E (2014) Past trends and future prediction of 25. de Sola H, Salazar A, Dueñas M, Ojeda B, Failde I (2016) Nationwide cross- mesothelioma incidence in an industrialized area of Italy, the Veneto region. sectional study of the impact of chronic pain on an individual's Cancer Epidemiol 38:496–503 employment: relationship with the family and the social support. BMJ Open 50. Riebler A, Held L (2017) Projecting the future burden of cancer: Bayesian 6:e012246 age-period-cohort analysis with integrated nested Laplace approximations. 26. Dorner TE, Alexanderson K, Svedberg P, Ropponen A, Stein KV, Biom J 59:531–549 Mittendorfer-Rutz E (2015) Sickness absence due to back pain or depressive 51. Spiegelhalter D, Best N, Carlin B, Van Der Linde A (2002) Bayesian measures episode and the risk of all-cause and diagnosis-specific disability pension: a of model complexity and fit. J R Stat Soc Ser B Stat Methodol 64:583–616 Swedish cohort study of 4,823,069 individuals. Eur J Pain 19:1308–1320 52. Besag J, Green P, Higdon D, Mengersen K (1995) Bayesian computation and 27. Selekler HM, Gökmen G, Alvur TM, Steiner TJ (2015) Productivity losses stochastic systems (with discussion). Stat Sci 10:3–66 attributable to headache, and their attempted recovery, in a heavy- 53. Rue H, Held L (2005) Gaussian Markov random fields, vol 104. Chapman & manufacturing workforce in Turkey: implications for employers and Hall/CRC Press, London politicians. J Headache Pain. 16:96 54. R Core Team (2018) R. A language and environment for statistical 28. Natvig B, Eriksen W, Bruusgaard D (2002) Low back pain as a predictor of computing. R Foundation for Statistical Computing, Vienna https://www.R- long-term work disability. Scand J Public Health 30:288–292 project.org/ (accessed 10 Feb 2020) 29. Raggi A, Leonardi M, Sansone E, Curone M, Grazzi L, D'Amico D (2020) The 55. Martins TG, Simpson D, Lindgren F, Rue H (2013) Bayesian computing with cost and the value of treatment of medication overuse headache in Italy: a INLA: new features. Comput Stat Data Analysis 67:68–83 longitudinal study based on patient-derived data. Eur J Neurol 27:62–e1 56. Lindgren F, Rue H, Lindstrom J (2011) An explicit link between gaussian 30. Messali A, Sanderson JC, Blumenfeld AM, Goadsby PJ, Buse DC, Varon SF et fields and gaussian markov random fields: the stochastic partial differential al (2016) Direct and indirect costs of chronic and episodic migraine in the equation. J R Stat Soc Ser B Stat Methodol 73:423–498 United States: a web-based survey. Headache 56:306–322 57. Rue H, Riebler A, Sorbye SH, Illian JB, Simpson DP, Lindgren FK (2017) 31. Gustavsson A, Bjorkman J, Ljungcrantz C, Rhodin A, Rivano-Fischer M, Bayesian computing with INLA: a review. Annu Rev Stat Its Appl 4:395–421 Sjolund KF et al (2012) Socio-economic burden of patients with a diagnosis 58. Jiménez-Sánchez S, Jiménez-García R, Hernández-Barrera V, Villanueva- related to chronic pain--register data of 840,000 Swedish patients. Eur J Pain Martínez M, Ríos-Luna A, Fernández-de-las-Peñas C (2010) Has the 16:289–299 prevalence of invalidating musculoskeletal pain changed over the last 15 32. Linde M, Gustavsson A, Stovner LJ, Steiner TJ, Barré J, Katsarava Z et al years (1993-2006)? A Spanish population-based survey. J Pain 11(7):612–620 (2012) The cost of headache disorders in Europe: the Eurolight project. Eur J 59. Celeste RK, Fritzell J (2018) Do socioeconomic inequalities in pain, Neurol 19:703–711 psychological distress and oral health increase or decrease over the life 33. Juniper M, Le TK, Mladsi D (2009) The epidemiology, economic burden, and course? Evidence from Sweden over 43 years of follow-up. J Epidemiol pharmacological treatment of chronic low back pain in France, Germany, Community Health 72:160–167 Italy, Spain and the UK: a literature-based review. Expert Opin Pharmacother 60. Tesfaye S (2011) Recent advances in the management of diabetic distal 10:2581–2592 symmetrical polyneuropathy. J Diabetes Investig 2:33–42 34. Sanchez-Niubo A, Egea-Cortés L, Olaya B, Caballero FF, Ayuso-Mateos JL, 61. Danaei G, Finucane MM, Lu Y, Singh GM, Cowan MJ, Paciorek CJ et al (2011) Prina M et al (2019) Cohort profile: The Ageing Trajectories of Health – National, regional, and global trends in fasting plasma glucose and diabetes Longitudinal Opportunities and Synergies (ATHLOS) project. Int J Epidemiol prevalence since 1980: systematic analysis of health examination surveys 48:1052–1053i and epidemiological studies with 370 country-years and 2·7 million 35. Leonardi M, Chatterji S, Koskinen S, Ayuso-Mateos JL, Haro JM, Frisoni G et participants. Lancet. 378:31–40 al (2014) Determinants of health and disability in ageing population: the 62. NCD Risk Factor Collaboration (NCD-RisC) (2016) Worldwide trends in COURAGE in Europe project (collaborative research on ageing in Europe). diabetes since 1980: a pooled analysis of 751 population-based studies with Clin Psychol Psychother 21:193–198 4.4 million participants. Lancet 387:1513–1530 36. Zhao Y, Hu Y, Smith JP, Strauss J, Yang G (2014) Cohort profile: the China 63. Korolainen MA, Kurki S, Lassenius MI, Toppila I, Costa-Scharplatz M, health and retirement longitudinal study (CHARLS). Int J Epidemiol 43:61–68 Purmonen T et al (2019) Burden of migraine in Finland: health care resource 37. Luszcz MA, Giles LC, Anstey KJ, Browne-Yung KC, Walker RA, Windsor TD use, sick-leaves and comorbidities in occupational health care. J Headache (2016) Cohort profile: the Australian longitudinal study of ageing (ALSA). Int Pain. 20(1):13 J Epidemiol 45:1054–1063 64. Takeshima T, Wan Q, Zhang Y, Komori M, Stretton S, Rajan N et al (2019) 38. Börsch-Supan A, Brandt M, Hunkler C, Kneip T, Korbmacher J, Malter F et al Prevalence, burden, and clinical management of migraine in China, Japan, (2013) Data resource profile: the survey of health, ageing and retirement in and South Korea: a comprehensive review of the literature. J Headache Pain Europe (SHARE). Int J Epidemiol 42:992–1001 20(1):111 39. Whelan BJ, Savva GM (2013) Design and methodology of the Irish 65. Straube A, Andreou A (2019) Primary headaches during lifespan. J Headache longitudinal study on ageing. J Am Geriatr Soc 61(Suppl 2):S265–S268 Pain. 20(1):35 Guido et al. The Journal of Headache and Pain (2020) 21:52 Page 11 of 11 66. Woldeamanuel YW, Cowan RP (2017) Migraine affects 1 in 10 people 88. Rota E, Evangelista A, Ceccarelli M, Ferrero L, Milani C, Ugolini A et al (2016) worldwide featuring recent rise: a systematic review and meta-analysis of Efficacy of a workplace relaxation exercise program on muscle tenderness community-based studies involving 6 million participants. J Neurol Sci 372: in a working community with headache and neck pain: a longitudinal, 307–315 controlled study. Eur J Phys Rehabil Med 52:457–465 67. Vetvik KG, Macgregor EA, Lundqvist C, Russell MB (2014) Prevalence of 89. Parker C, Waltman N (2012) Reducing the frequency and severity of menstrual migraine: a population-based study. Cephalalgia. 34(4):280–288 migraine headaches in the workplace: implementing evidence-based 68. Vetvik KG, MacGregor EA, Lundqvist C, Russell MB (2018) Symptoms of interventions. Workplace Health Saf 60:12–18 premenstrual syndrome in female migraineurs with and without menstrual 90. Barton GR, Sach TH, Jenkinson C, Doherty M, Avery AJ, Muir KR (2009) migraine. J Headache Pain. 19(1):97 Lifestyle interventions for knee pain in overweight and obese adults aged > 69. United Nations (2017) Department of Economic and Social Affairs, or = 45: economic evaluation of randomised controlled trial. BMJ 339:b2273 population division. World population prospects: the 2017 revision, volume 91. Hagen K, Åsberg AN, Stovner L, Linde M, Zwart JA, Winsvold BS et al (2018) II: demographic profiles (ST/ESA/SER.A/400). United Nations, New York Lifestyle factors and risk of migraine and tension-type headache. Follow-up 70. Vivekanantham A, Edwin C, Pincus T, Matharu M, Parsons H, Underwood M data from the Nord-Trøndelag health surveys 1995-1997 and 2006-2008. (2019) The association between headache and low back pain: a systematic Cephalalgia. 38:1919–1926 review. J Headache Pain. 20(1):82 92. Verrotti A, Di Fonzo A, Penta L, Agostinelli S, Parisi P (2014) Obesity and 71. National Osteoporosis Foundation (2002) America’s Bone Health: The State headache/migraine: the importance of weight reduction through lifestyle of Osteoporosis and Low Bone Mass in Our Nation. National Osteoporosis modifications. Biomed Res Int 2014:420858 Foundation, Washington, DC 93. Majeed MH, Sudak DM (2017) Cognitive behavioral therapy for chronic 72. Nguyen ND, Pongchaiyakul C, Center JR, Eisman JA, Nguyen TV (2005) pain-one therapeutic approach for the opioid epidemic. J Psychiatr Pract 23: Identification of high-risk individuals for hip fracture: a 14-year prospective 409–414 study. J Bone Miner Res 20:1921–1928 94. Van Denburg AN, Vilardaga JP, Shelby RA, Keefe FJ (2018) Opioid therapy 73. Witt EA, Kenworthy J, Isherwood G, Dunlop WC (2016) Examining the and persistent pain: can cognitive behavioral therapy help? Pain. 159:411– association between pain severity and quality-of-life, work-productivity loss, 415 and healthcare resource use among European adults diagnosed with pain. J 95. Castelnuovo G, Giusti EM, Manzoni GM, Saviola D, Gatti A, Gabrielli S et al Med Econ 19:858–865 (2016) Psychological Treatments and Psychotherapies in the 74. Müller-Schwefe G, Freytag A, Höer A, Schiffhorst G, Becker A, Casser HR et al Neurorehabilitation of Pain: Evidences and Recommendations from the (2011) Healthcare utilization of back pain patients: results of a claims data Italian Consensus Conference on Pain in Neurorehabilitation. Front Psychol analysis. J Med Econ 14:816–823 7:115 75. Grazzi L, Grignani E, D’Amico D, Sansone E, Raggi A (2018) Is medication overuse drug specific or not? Data from a review of published literature and Publisher’s Note from an original study on Italian MOH patients. Curr Pain Headache Rep 22: Springer Nature remains neutral with regard to jurisdictional claims in 71 published maps and institutional affiliations. 76. Coutinho AD, Gandhi K, Fuldeore RM, Landsman-Blumberg PB, Gandhi S (2018) Long-term opioid users with chronic noncancer pain: assessment of opioid abuse risk and relationship with healthcare resource use. J Opioid Manag 14:131–141 77. Morasco BJ, Yarborough BJ, Smith NX, Dobscha SK, Deyo RA, Perrin NA et al (2017) Higher prescription opioid dose is associated with worse patient- reported pain outcomes and more health care utilization. J Pain 18:437–445 78. Agaliotis M, Mackey MG, Jan S, Fransen M (2014) Burden of reduced work productivity among people with chronic knee pain: a systematic review. Occup Environ Med 71:651–659 79. Kristoffersen ES, Stavem K, Lundqvist C, Russell MB (2019) Impact of chronic headache on workdays, unemployment and disutility in the general population. J Epidemiol Community Health 73:360–367 80. Leonardi M, Raggi A (2019) A narrative review on the burden of migraine: when the burden is the impact on people's life. J Headache Pain. 20:41 81. Vo P, Fang J, Bilitou A, Laflamme AK, Gupta S (2018) Patients' perspective on the burden of migraine in Europe: a cross-sectional analysis of survey data in France, Germany, Italy, Spain, and the United Kingdom. J Headache Pain. 19:82 82. Neupane S, Nygård CH, Prakash KC, von Bonsdorff MB, von Bonsdorff ME, Seitsamo J et al (2018) Multisite musculoskeletal pain trajectories from midlife to old age: a 28-year follow-up of municipal employees. Occup Environ Med 75:863–870 83. Laires PA, Canhão H, Rodrigues AM, Eusébio M, Gouveia M, Branco JC (2018) The impact of osteoarthritis on early exit from work: results from a population-based study. BMC Public Health 18:472 84. Keysor JJ, LaValley MP, Brown C, Felson DT, AlHeresh RA, Vaughan MW et al (2018) Efficacy of a Work Disability Prevention Program for People with Rheumatic and Musculoskeletal Conditions: A Single-Blind Parallel-Arm Randomized Controlled Trial. Arthritis Care Res (Hoboken) 70:1022–1029 85. Laires PA, Gouveia M, Canhão H (2017) Interventions aiming to reduce early retirement due to rheumatic diseases. Acta Reumatol Port 42:240–248 86. Oakman J, Kinsman N, Briggs AM (2017) Working with persistent pain: an exploration of strategies utilised to stay productive at work. J Occup Rehabil 27:4–14 87. Vicente-Herrero T, Burke TA, Laínez MJ (2004) The impact of a worksite migraine intervention program on work productivity, productivity costs, and non-workplace impairment among Spanish postal service employees from an employer perspective. Curr Med Res Opin 20:1805–1814