Pain during midlife: a cross-national analysis of cohort differences in reports of pain in the United States, Europe, South Korea, and Mexico

Abstract

Middle-aged Americans today are reporting poorer mental, cognitive, and physical health compared to previous cohorts, but this trend has not been consistently observed in other nations. It is an open question whether pain shows similar cohort differences among US middle-aged adults compared to other nations. We used harmonized data on pain from nationally representative longitudinal panel surveys from the United States, 13 European nations (England, Continental, Mediterranean, and Nordic regions), South Korea, and Mexico to directly quantify cohort similarities and differences in midlife pain. Results from multilevel models demonstrated that midlife pain is higher among later-born cohorts in the United States than among earlier-born cohorts. The increased odds for later-born cohorts emerged in the early years of midlife, when people are in their early 50s. A similar pattern of increased odds of reporting pain for later-born cohorts was observed in England and Mexico. In contrast, decreased odds of reporting pain for later-born cohorts were observed in Continental, Mediterranean, and Nordic Europe as well as South Korea. Results for pain severity revealed a highly similar pattern. Our discussion focuses on potential explanations, including population-level discrepancies in use and quality of healthcare services and how pain is conceptualized across nations.

This article is part of a Special Collection on Cross-National Gerontology.

Introduction

Lifespan developmental psychology has long been interested in examining the role that cohort differences and cultural contexts play in shaping mental, cognitive, and physical health across the adult lifespan.^1,2 For example, older adults (ages 65 and older) in the United States and peer nations today are performing better cognitively and reporting better health and well-being than older adults several decades ago.³ In contrast, empirical research over the past decade has shown that later-born cohorts of middle-aged adults in the United States often report poorer mental, cognitive, and physical health and elevated daily stress, compared to same-aged peers several decades ago.^4-8 For the United States, research has also shown rising rates of chronic illness^9,10 and has documented that life expectancy is stagnating/declining, which is largely due to increasing mortality rates among working-age adults.¹¹ Our objective in the current study is to apply the principles of lifespan developmental psychology that focus on the role of cohort differences and cultural context in shaping development by investigating birth year cohort differences in midlife pain in the United States and how this pattern compares with those observed among middle-aged adults in Europe, South Korea, and Mexico.

Midlife

Midlife is a developmental period between the ages of 40 and 65 that is less well understood relative to other phases of life.¹² Middle-aged adults are vital to societal success because they comprise the majority of the workforce and constitute family leadership through bridging younger and older generations.^13,14 Several studies have documented that better midlife health foreshadows better health and more financial assets in old age.^15,16 Midlife is a period of the lifespan where key domains show stability, growth, or decline. Empirical evidence suggests that crystallized cognitive abilities are well preserved, if not increasing, in midlife, whereas fluid cognitive abilities start to show noticeable signs of decline.^17,18 Well-being and emotional experience show relative stability and, in some instances, improve among middle-aged adults.¹⁹ Midlife is typically the age of onset for chronic health issues. High blood pressure, cancer, and arthritis are usually detected for the first time in midlife.¹² However, recent decades have seen profound changes in culture and lifestyle characteristics that have wide-ranging implications for middle-aged adults. Following the Great Recession, midlife stress, chronic illness, and mental health problems rose,^4,6,7 and this is coupled with stagnating/declining life expectancy.¹¹ With the widespread declines in mental, cognitive, and physical health noted among later-born cohorts of middle-aged adults in the United States, troublesome implications for broader society arise because middle-aged adults are the backbone of modern society, given that they comprise most of the workforce, raise the next generation of citizens, and have caregiving responsibilities with older adults.^13-15

Pain

Chronic pain has become a public health issue in the United States, with its prevalence escalating at rates that coincide with those of opioid abuse and overdose.²⁰ Recent epidemiologic data have demonstrated rising rates of reporting pain among adults in midlife (ie, age 45²¹). Other studies have also noted rising prevalence rates of mild/moderate non-limiting pain across adulthood and old age in the United States and in some nations in Europe.^22,23 In the 2010s, over 100 million Americans reported impairment due to pain,^24,25 and multiple studies have shown that pain is a risk factor for poorer health outcomes. For example, more severe pain is associated with lower quality of life, steeper functional health declines, and mortality.^24-26 Case et al. found that the prevalence of self-reported problematic pain has been increasing beginning in young adulthood among successive birth cohorts born since the 1940s and could be a contributing factor to the rise in rates of deaths of despair.^27,28 One notable limitation of these studies was reliance on cross-sectional data instead of examining historical shifts in within-person changes using longitudinal data. Such an approach offers the advantage of thoroughly testing whether health outcomes are stable, increasing, or declining within persons over time.²⁹ An analysis of midlife pain in which the same individuals are followed during midlife has not yet been conducted, even though middle-aged individuals are the ones reporting more severe and frequent pain than any other age group.²⁷ Moreover, across the American adult population, it has become clear that recent generations are experiencing more maladaptive and debilitating pain than those who came before them, despite advancements in health technology, improved treatment, and ostensibly better access to resources.¹³

Pain in midlife: a uniquely American issue?

Although midlife can be challenging regardless of the country in which people live, the historical decline in midlife mental, cognitive, and physical health has been most pronounced in the United States^5,6,8 In other peer nations, the responsibilities of midlife (eg, work, child-rearing, caretaking) are buttressed by government support and cultural norms prioritizing health, family, and community.³⁰ Paid family leave and universal health care are mandated in most high-income nations, and the latter has repeatedly been shown to be directly linked with better mental, cognitive, and physical health outcomes and longer life expectancy.^6,31,32 In the United States, however, ongoing economic difficulties force younger adults to move back in with their parents and remain dependent on them for financial support well into their 20s. At the same time, older adults in the United States are restricted from independent retirement and adequate care, making the caregiving responsibilities for most Americans during midlife burdensome. Correspondingly, Feinberg reported that today’s middle-aged adults in the United States are more financially unstable than ever before.³³ Cross-national studies by Infurna et al. found that later-born cohorts of middle-aged adults in the United States reported more depressive symptoms, feelings of loneliness, and exhibited poorer memory compared to earlier-born cohorts.^5,6 Conversely, later-born cohorts of middle-aged adults in several European nations, South Korea, and Mexico reported better outcomes than their earlier-born peers, indicating that there is something uniquely American about increases in midlife mental and physical health problems.

The present study

The objective of the present study is to investigate cohort differences in pain among middle-aged adults from 16 nations: the United States, England, Continental Europe (France, Austria, Belgium, Germany, Switzerland, and Israel), Mediterranean Europe (Italy, Spain, and Greece), Nordic Europe (Denmark, The Netherlands, and Sweden), South Korea, and Mexico. The current study contributes to the literature by using harmonized longitudinal panel survey data to examine cohort differences in within-person trajectories in reports of pain in midlife. We hypothesized that (1) later-born cohorts in the United States would report increases in pain in midlife and (2) the United States would be the only nation in which later-born cohorts report more pain during midlife than earlier-born cohorts.

Methods

Participants

We used data from nationally representative samples obtained in 16 countries across 5 different datasets: the US Health and Retirement Study (HRS), the English Longitudinal Study of Ageing (ELSA), the Survey of Health, Ageing, and Retirement in Europe (SHARE), the Korean Longitudinal Study of Aging (KLOSA), and the Mexico Health and Aging Study (MHAS). These datasets have been harmonized by the Gateway to Global Aging team and can be accessed via their website (https://g2aging.org). The development of the harmonized data was funded by the National Institute on Aging (R01 AG030153, RC2 AG036619, R03AG043052). We only included data between the ages of 45 and 65 in our analysis because this is widely considered the typical range of midlife.¹² Below and in Appendix S1 are the sampling and data collection procedures for each of the datasets. Table 1 provides descriptive information for each dataset, including the data collection years, sample sizes, and gender distributions. Gender, education (harmonized measure consisting of 3 categories: less than upper secondary education, upper secondary education and vocational education, and tertiary education), and the number of health conditions (sum index of high blood pressure, diabetes, cancer or malignant tumor, lung disease, heart condition, stroke, psychiatric problems, and arthritis) were included as covariates into the model because of their association with pain.³⁴  Figure S1 shows the number of observations provided for each nation/region by birth year. This study was preregistered, and further information is provided through the Open Science Framework (https://osf.io/34j7b/?view_only=09d06a8a52044cb0aa98dd5e32f99d9b).

Table 1.

Descriptive statistics for the variables under study.

	United States (HRS)	England (ELSA)	Continental Europe (SHARE)	Mediterranean Europe (SHARE)	Nordic Europe (SHARE)	South Korea (KLoSA)	Mexico (MHAS)
Data collection on  years	1992 to 2018/2019 (biennially)	2004 to 2018/2019 (biennially)	2012/2013 to 2019/2020 (biennially)	2012/2013 to 2019/2020 (biennially)	2012/2013 to 2019/2020 (biennially)	2006 to 2018 (biennially)	2001, 2003, 2012, 2015, and 2018
Sample size	n = 30 421	n = 12 863	n = 15 024	n = 6704	n = 8931	n = 7289	n = 20 934
Age at baseline	M = 53.95 (SD = 4.22) range: 45-65	M = 55.12 (SD = 4.75) range: 45-65	M = 57.29 (SD = 4.71) range: 45-65	M = 57.77 (SD = 4.80) range: 45-65	M = 57.63 (SD = 4.80) range: 45-65	M = 54.24 (SD = 5.89) range: 45-65	M = 54.86 (SD = 4.77) range: 45-65
Gender (% women)	55%	55%	57%	56%	58%	55%	57%
Year of birth	M = 1948 (SD = 11.05) range: 1927-1974	M = 1951 (SD = 8.06) range: 1937-1973	M = 1955 (SD = 4.88) range: 1948-1974	M = 1955 (SD = 4.86) range: 1948-1974	M = 1956 (SD = 4.96) range: 1948-1974	M = 1953 (SD = 6.75) range: 1941-1963	M = 1953 (SD = 9.54) range: 1936-1973

Abbreviations: ELSA, English Longitudinal Study of Ageing; HRS, Health and Retirement Study; KLoSA, Korean Longitudinal Study of Aging; MHAS, Mexico Health and Aging Study; SHARE, Survey of Health, Ageing, and Retirement in Europe.

Continental Europe consists of France, Austria, Belgium, Germany, Switzerland, and Israel. Mediterranean Europe consists of Italy, Spain, and Greece. Nordic Europe consists of Denmark, The Netherlands, and Sweden.

US Health and Retirement Study

The HRS³⁵ began in 1992 and is a nationally representative sample of households in the contiguous United States. Individuals in the sample are aged 50 and older. The HRS collected data using telephone and in-person interviews. Participants reported their self-rated physical and psychological health via survey responses every other year. Every 6 years, a new cohort was recruited. The analyses for the present study included 30 421 individuals who provided at least 1 assessment of pain during midlife (55% women; 20% less than upper secondary education, 58% upper secondary education and vocational education, and 22% tertiary education; health conditions: M = 2.36, SD = 1.64, range 0-8). We use data from 1992 to 2018/2019.

English Longitudinal Study of Ageing

The ELSA³⁶ began in 2002 and is a nationally representative sample of adults aged 50 years and older in England. Participants reported their self-rated physical and psychological health via survey responses every other year. The analyses for the present study included 12 863 individuals who provided at least one assessment of pain during midlife (55% women; 26% less than upper secondary education, 54% upper secondary education and vocational education, and 20% tertiary education; health conditions: M = 1.40, SD = 1.29, range 0-7); we used biennial assessments from 2002 to 2019.

Survey of Health, Ageing, and Retirement in Europe

The SHARE³⁷ began in 2004 and is a cross-national longitudinal panel survey on the mental and physical health, and social and family networks of individuals in Europe and Israel. In our analyses, we only included nations that began in Wave 1 because these provide enough observations over a long enough time frame that allow for the examination of historical shifts in within-person longitudinal change. These nations were Austria, Belgium, Switzerland, Germany, Denmark, Spain, France, Greece, Italy, The Netherlands, Sweden, and Israel. We followed Wallace et al.³⁸, who grouped these nations based on the nation’s social model into Continental Europe (France, Austria, Belgium, Germany, Switzerland, and Israel; 57% women; 22% less than upper secondary education, 47% upper secondary education and vocational education, and 31% tertiary education; health conditions: M = 1.46, SD = 1.29, range 0-8), Mediterranean Europe (Italy, Spain, and Greece; 58% women; 55% less than upper secondary education, 28% upper secondary education and vocational education, and 17% tertiary education; health conditions: M = 1.21, SD = 1.23, range 0-8), and Nordic Europe (Denmark, The Netherlands, and Sweden; 56% women; 23% less than upper secondary education, 36% upper secondary education and vocational education, and 41% tertiary education; health conditions: M = 1.21, SD = 1.29, range 0-8). The measure of pain was assessed beginning in Wave 5 and up to 4 assessments were included for each participant. In total, 30 659 participants who provided data when they were aged 45-65 were included in the analyses.

Korean Longitudinal Study of Aging

The KLoSA³⁹ began in 2006 and is a nationally representative sample of South Korean residents, excluding those from Jeju Island. Individuals in the sample are aged 45 and older. Participants reported their self-rated physical and psychological health via survey responses every other year. In total, 7289 participants who provided data when they were aged 45-65 were included in the analyses (55% women; 45% less than upper secondary education, 41% upper secondary education and vocational education, and 14% tertiary education; health conditions: M = 0.92, SD = 1.08, range 0-7).

Mexico Health and Aging Study

The MHAS⁴⁰ is a nationally representative sample of both rural and urban Mexican residents and their spouses. Individuals in the sample were aged 50 and older. The data in the MHAS are longitudinal (assessments were in 2001, 2003, 2012, 2015, and 2018); participants reported their self-rated physical and psychological health via survey responses over the course of several years. In total, 20 934 participants who provided data when they were aged 45-65 were included in the analyses (57% women; 82% less than upper secondary education, 3% upper secondary education and vocational education, and 15% tertiary education; health conditions: M = 1.28, SD = 1.17, range 0-6).

Outcome

Reports of pain were assessed with a single-item dichotomous variable referring to whether the respondent reported being often troubled with pain (1 = yes, 0 = no). Pain severity was assessed with a single item referring to the respondent’s self-reported usual degree of pain (0 = no pain, 1 = mild pain, 2 = moderate pain, 3 = severe pain). Previous research documents the reliability and validity of these 1-item measures (see^41,42). The sample size for the analyses pertaining to reports of pain contained 101 592 individuals (301 971 observations), and the sample size for the analyses pertaining to pain severity contained 100 381 individuals (298 281 observations).

Data analysis

Time-in-study, age, and cohort

Following Gerstorf et al., we examined intraindividual change as time-in-study, a time-varying variable quantified for each assessment as the number of years since baseline (T1) and centered at the middle of each individual’s repeated measures time series.⁴³ Age-related differences (age gradients) were examined as individuals’ chronological age (at their middle assessment) and centered at age 50. Cohort-related differences were examined as individuals’ birth year and centered at 1960.

Intraindividual changes, age-related, and birth year cohort differences were examined using multilevel models.⁴⁴ We analyzed our data using multilevel models because our objective was to track how pain frequency developed as people moved through midlife (ie, within-person change) and the extent to which this differed across cohorts (ie, between-person differences). Appendix S1 contains the equations for our models. Models were fit using Mplus⁴⁵ and the models accounted for the dichotomous (ie, pain frequency) and Poisson distribution (ie, pain severity) nature of the outcome variables, which takes into consideration the application of the appropriate non-normal error distribution (for discussion, see⁴⁶). The findings for pain frequency are interpreted as odds ratios for each 1-unit increase in the predictor, and the findings for pain severity are interpreted as expected log counts for each 1-unit increase in the predictor. Following usual practice (eg,⁴⁴), we applied Full Information Maximum Likelihood procedures to accommodate incomplete data under usual missing at random assumptions.⁴⁷ Given the large sample size for the analyses and to guard against false positives (for discussion, see⁴⁸), we use the P < .0001 cut-off for statistical significance for intercept-related parameters and P < .001 for time-related parameters. We have chosen these different cut-offs because the statistical power to detect intercept effects is greater than the power to detect slope effects (see⁴⁹).

Results

The results shown in Table 2 convey the main effects and interactions between country, birth year, and age in predicting the likelihood of reporting pain. Figure 1 graphically illustrates the findings pertaining to how reports of pain differ across birth years and the nations/regions. This conjoint model includes participants from all 16 nations, with participants from the United States serving as the reference group. We observed that relative to the United States, each nation (except Mexico) had a greater propensity to report pain at the centering point of a 50-year-old born in 1960. Compared to the United States, middle-aged adults in England, Continental, Nordic, and Mediterranean Europe, and South Korea had 1.82, 3.59, 2.21, 2.29, and 4.83 higher odds of reporting pain, respectively. Findings revealed that, for middle-aged Americans (blue line in Figure 1), later-born cohorts had increased odds of reporting pain. For each birth year after 1960 (where the data were centered), middle-aged Americans, on average, had 10% increased odds of reporting pain. Although all other countries’ initial reports of pain superseded those seen among earlier-born cohorts in the United States, most saw decreasing trends in reported pain across birth years (eg, Mediterranean Europe by birth year = 0.83); later-born US respondents ultimately had the highest odds of reporting pain, starting around those born in 1970 (see blue line in Figure 1). A similar pattern of increasing odds of reporting pain among later-born cohorts was observed in England and Mexico. In contrast, later-born cohorts of middle-aged adults in Continental Europe (gray line in Figure 1), Mediterranean Europe (yellow line in Figure 1), Nordic Europe (light-blue line in Figure 1), and South Korea (green line in Figure 1) had decreased odds of reporting pain. For each year of birth after 1960 (where the data were centered), middle-aged adults in Continental Europe, Mediterranean Europe, Nordic Europe, and South Korea had a 7%, 17%, 2%, and 7% decreased odds of reporting pain, respectively.

Table 2.

Results from multilevel model examining historical changes in reports of pain within and between nations

	Odds ratio	95% CI
Parameters
Birth year	1.103*	[1.092, 1.114]
Birth year squared	1.003*	[1.002, 1.003]
Age	1.115*	[1.091, 1.139]
Birth year*age	1.009*	[1.007, 1.011]
England	1.817	[1.568, 2.104]
Continental Europe	3.586*	[2.845, 4.519]
Nordic Europe	2.208*	[1.604, 3.039]
Mediterranean Europe	2.289	[1.689, 3.101]
South Korea	4.831*	[4.154, 5.618]
Mexico	1.155*	[1.015, 1.314]
England*birth year	1.023*	[1.007, 1.039]
Continental Europe*birth year	0.930	[0.889, 0.972]
Nordic Europe*birth year	0.979	[0.921, 1.041]
Mediterranean Europe*birth year	0.827*	[0.778, 0.879]
South Korea*birth year	0.927*	[0.902, 0.953]
Mexico*birth year	1.067*	[1.053, 1.082]
England*age	1.082	[1.053, 1.112]
Continental Europe*age	0.946	[0.899, 0.995]
Mediterranean Europe*age	0.949*	[0.893, 1.009]
South Korea*age	1.028	[0.993, 1.065]
Mexico*age	1.083*	[1.061, 1.107]
Mediterranean Europe*birth year*age	1.021	[1.009, 1.033]
Mediterranean Europe*birth year squared	1.013	[1.003, 1.023]
Mexico*birth year squared	1.007*	[1.006, 1.007]
Women	1.806*	[1.736, 1.878]
Education	0.598*	[0.581, 0.614]
Health conditions	2.250*	[2.215, 2.286]
Fixed Effects	Estimate	SE
Intercept (threshold)	1.35*	0.05
Time	0.06*	0.01
Random effects
Intercept	4.63*	0.07
Time	0.03*	0.001

n = 101 592 individuals and 301 971 observations. Reports of Pain are scored 0 (no) and 1 (yes). The model was centered at birth year 1960 and age 50; the United States is the reference group.

^*P < .0001.

Figure 1.

Probability of reporting pain according to birth year by country. In the United States, England, and Mexico, the odds of reporting pain increased the more recently people were born (ie, with increases in year of birth). In contrast, in Continental, Mediterranean, and Nordic Europe, as well as in South Korea, the odds of reporting pain decreased the more recently people were born.

Focusing on the findings pertaining to age and age-by-country interactions, middle-aged Americans had increasing odds of reporting pain for each birth year after age 50. A similar pattern was observed for Mexico. In comparison to the United States, age-related increases in reports of pain were reduced for middle-aged adults in Mediterranean Europe, and no age-related differences were observed in Continental and Nordic Europe and South Korea.

Results thus indicated that age-related increases in the odds of reporting pain were more pronounced with later birth years for the United States and Mexico. Figure 2 graphically illustrates the birth year × age interaction for middle-aged adults in the United States. For each age trajectory, the reported pain levels were higher in later birth years. For example, for those aged 60 (gray line), the probability of reporting pain is higher for those born in 1955 than for those born in 1940. Likewise, for those aged 50 (blue line), the probability of reporting pain is higher for those born in 1965 than for those born in 1950.

Figure 2.

Probability of reporting pain according to year of birth by age in the United States. For each age trajectory, the level of pain reports is higher at later years of birth. For example, for those aged 60, the probability of reporting pain is higher for those born in 1955 compared with those born in 1940. Likewise, for those aged 50, the probability of reporting pain is higher for those born in 1965 compared with those born in 1950.

Table 3 reports results focusing on pain severity. On average, later-born cohorts of middle-aged adults in the United States reported greater severity of pain (estimate = 0.036, P < .0001). Reports of pain increased with age (estimate = 0.042, P < .0001), and the birth year × age interaction indicates that age differences in pain severity were more pronounced for later-born cohorts (estimate = 0.003, P < .0001). Like reports of pain, compared to the middle-aged adults in the United States, middle-aged adults in England, Continental, Nordic, and Mediterranean Europe, and South Korea, on average, reported greater pain severity at the model centering point of age 50 and birth year 1960. The country × birth year interactions indicate whether there are differences with the United States. In contrast to the United States, later-born cohorts of middle-aged adults in Continental, Mediterranean, and Nordic Europe, South Korea, and Mexico reported less severe pain than earlier-born cohorts.

Table 3.

Results from multilevel model examining historical changes in pain severity within and between nations.

	Estimate	SE
Parameters
Birth year	0.036*	0.002
Birth year squared	0.001*	0.000
Age	0.042*	0.004
Birth year*age	0.003*	0.000
England	0.227*	0.034
Continental Europe	0.529*	0.046
Nordic Europe	0.356*	0.069
Mediterranean Europe	0.261*	0.057
South Korea	0.231*	0.032
Mexico	0.030	0.027
England*birth year	−0.024*	0.005
Continental Europe*birth year	−0.065*	0.009
Nordic Europe*birth year	−0.058*	0.013
Mediterranean Europe*birth year	−0.071*	0.010
South Korea*birth year	−0.066*	0.005
Mexico*birth year	−0.014*	0.003
Continental Europe*age	−0.067*	0.010
England*age	−0.028*	0.006
Mediterranean Europe*age	−0.053*	0.011
Nordic Europe*age	−0.057*	0.014
South Korea*age	−0.040*	0.007
Mexico*age	−0.016	0.005
England*birth year squared	0.001*	0.000
Mexico*birth year squared	0.002*	0.000
Women	0.227*	0.008
Education	−0.244*	0.006
Health conditions	0.286*	0.003
Fixed effects
Intercept	−0.586*	0.022
Time	0.02*	0.004

N = 100 381 individuals and 298 281 observations. Pain severity was scored 0 = no pain, 1 = mild pain, 2 = moderate pain, and 3 = severe pain. The model was centered at birth year 1960 and age 50. The United States is the reference group.

^*P < .0001.

Each of the covariates was predictive of reporting pain and pain severity. Being a woman, attaining fewer years of education and reporting more health conditions were each associated with an increased odds of reporting pain and greater pain severity.

Discussion

The overarching objective of our study was to examine whether reports of pain have increased, decreased, or remained stable across birth years or cohorts among US middle-aged adults and how these changes compare to peer nations. Later-born cohorts of middle-aged adults in the United States had increased odds of reporting pain, and their pain was more severe. Later-born cohorts also tended to start reporting pain earlier in midlife than their earlier-born counterparts (see Figure 2). Contrary to predictions, the United States was not the only country in which this trend was found. Middle-aged adults in England and Mexico also saw increased pain among later-born cohorts, although, unlike in the United States, these reports of pain did not supersede those of earlier-born cohorts. Nations/regions that saw decreases in reports of pain across birth years or cohorts included Continental, Mediterranean, and Nordic Europe and South Korea. Our findings document that cohort differences in pain operate differently across nations for middle-aged adults, and we discuss potential reasons underlying these findings below.

Historical change in midlife pain

Lifespan developmental psychology posits that various historical, developmental, and contextual factors combine to systematically influence individuals born and living at different historical points and geographic locations to produce unique developmental trajectories.¹ Thus, there are several potential contributors to cohort differences in midlife pain for middle-aged adults in the United States, as well as the comparative similarities and differences among peer nations. One possible explanation involves the disparities in SES and differential financial burdens that likely have direct effects on reported health and quality of life in the United States, whereas in many of the other nations, universal healthcare may reduce health disparities due to SES (eg, Continental, Nordic, and Mediterranean Europe³¹). Intermingled with SES differences (that we here in part accommodated by considering education), the vast discrepancies in health characteristics of countries as diverse as the United States and England may also play a role. In South Korea, where the population is more racially homogenous, the systematic gap in quality of life and life expectancy between different racial groups likely plays less of a role in measures of health like pain and illness.³²

Another plausible explanation for the increasing pain in the United States and not in parts of Europe is mass medicalization. Over time, people in the United States have utilized health services more, resulting in burdens previously constructed as normative being overdiagnosed and pathologized.⁵⁰ As a social phenomenon, medicalization may be attributed to intimate ties between the pharmaceutical industry and national investment; profits incurred by higher rates of “disease” are beneficial to economies highly dependent on the healthcare industry.⁵⁰ Populations from countries in which mass medicalization benefits the GDP tend to report more problems with health, including problems with pain. According to the OECD, the countries that spend the most on healthcare per capita (and thus have the greatest investment in the healthcare industry) include the United States, England, Switzerland, Germany, Austria, Sweden, The Netherlands, Belgium, and France.⁵¹ These countries, in our study, exhibited similar patterns in reports of pain insofar as the differences between them were not statistically significant. It would be valuable to explore the utilization of healthcare services and cultural norms surrounding health and disease when aiming to understand why the United States and some European nations have seen recent upticks in reports of midlife pain.

In the United States, the financial burdens of midlife are disproportionately high, and responsibilities characteristic of midlife (eg, childcare, elder care, and full-time occupation) are less likely to be supported by the social safety net.^31,33 Additionally, the unique structure of the American workforce produces the phenomenon of being “on call” 24/7, meaning that many Americans have little distinction between work life and personal life, cultivating the perception (and oftentimes reality) of being constantly busy.⁵² Research has shown a significant difference between on-call workers and not-on-call workers in “need for recovery” between shifts.⁵³ “Need for recovery” was defined as one’s self-reported perception of fatigue and exhaustion following work and the need for recuperation. Prior research has also found that chronic “need for recovery” relates to various health complaints.⁵⁴ Van de Ven et al.⁵³ also showed that workers who were on-call and called had the highest reported need for recovery, followed closely by workers who were on-call and not called. Both conditions were significantly higher than the not-on-call condition, suggesting that being required to work outside of working hours, regardless of whether one is actually “called in,” may disrupt the ability to recuperate and “recharge” between shifts. The implications of these findings suggest that working conditions resulting in a higher “need for recovery” may be one key contributing factor to subjective experiences of poorer health, which may affect reports and severity of pain. Demanding, laborer-unfriendly working conditions may help explain the U-shape cohort trend in reports of pain in Mexico as well, where high workplace stress and hostile working environments have been demonstrably linked to death and disease for decades.⁵⁵

Our findings revealed vast discrepancies in initial reports of pain. Each nation/region, except for Mexico, showed a higher likelihood of reporting pain and pain severity compared to the United States. Even with the comparatively higher pain among later-born cohorts, the United States sustains the lowest overall predicted odds ratio margins once the model is adjusted for gender, education, and health conditions because the pain reports of earlier-born cohorts in other countries are simply so much higher or conversely the pain reports of earlier born cohorts in the United States were so low (ie, born in the 1930s and 1940s; see Figure 1). Based on the estimates from Figure 1, whereas members of the earliest born-cohorts in the United States are just under 20% likely to report pain, the earliest-born participants in Continental and Mediterranean Europe and South Korea (1940s and 1950s cohorts) hovered around 80% probability of reporting pain. It is an open question as to which specific factors in these nations could lead to their citizens reporting such high levels of pain. Conceptually, there could be differences between nations in how pain is experienced and reported. For example, Sharma et al. observed differences in pain beliefs across countries; the authors suggest that pain treatments available in a country and health professionals’ pain beliefs play a role.⁵⁶ Methodologically, one potential explanation for later-born cohorts showing a decreased likelihood of reporting pain is that later-born cohorts are more positively selected, as survey participation rates have declined (for discussion, see ⁵⁷). However, follow-up analyses that targeted each nation/region separately revealed substantially similar findings to those reported in Tables 2 and 3. These findings warrant further exploration to determine the contributing mechanisms impacting these patterns of reports of pain within and between nations.

Future directions

We elaborate on areas for future research that may advance the burgeoning area of research focusing on historical change in midlife outcomes. The longitudinal nature of our study allowed us to provide strong empirical evidence that differential experiences within and between nations may account for discrepancies between birth-year cohorts and help explain the trajectories of increasing pain within birth-year cohorts. Future studies need to examine country-level variables (ie, healthcare quality and access to healthcare, as well as use of pain medications) and individual-level variables (ie, depressive symptoms, obesity, and sedentary behavior) that coincide with the pattern of increasing pain in the United States and other nations to determine potential links with public health phenomena (eg, opioid epidemic, mass medicalization), evolving work conditions, and cultural phenomena like interpretations and attitudes about pain. For example, Zimmer et al. observed that living in a nation with higher income inequality, greater population density, and more gender inequality was associated with an increased likelihood of reporting pain.⁵⁸ Most research in this regard has found that the worsening of midlife outcomes among later-born cohorts is most pronounced among low-SES individuals.²⁸ Future research should also use multiple measures to assess pain (eg, differentiating between chronic vs. acute pain and between different pain locations as well as pain sources/etiologies), evaluate the consistency and accuracy of the construct, and arrive at a common operationalization across contexts.

Limitations and conclusion

We acknowledge several limitations of our study. First, our measure did not specify the type, location, or source of pain. There may be systematic differences in conceptions of pain and pain thresholds between nations, and (un)systematic differences in conceptions of pain between individuals, which we were not able to assess. Also, the location of pain (eg, knee vs shoulder vs hand) could differentially impact reports of pain and we did include comorbidities as a covariate, presumably the variable that is most closely tied to sources of pain. As an inherently subjective experience,⁵⁹ pain is constructed socially by learned and internalized definitions of it, and those definitions depend largely on the culture and context people are living in.⁶⁰ The likelihood of reporting pain when being asked and the determination of what types of pain warrant reporting may differ across countries (ie, the phrase “troubled with pain often” can be interpreted differently). There is also the possibility of systematic attrition; participants reporting more pain may be less likely to provide more observations.

Our use of harmonized longitudinal panel survey data allowed us to investigate the extent to which there are historical changes over time within countries and among individuals. The pattern of findings could be due to cultural factors and cross-national differences, yet a close inspection of the mechanisms of changes in reports of pain is beyond the scope and possibilities of this study and requires future studies and mediation analyses.

We intended to bridge the gap in the current literature by moving beyond the mainly (time-lagged) cross-sectional data and examining cross-national differences in how pain evolves within persons as people move through their midlife years. Towards that end, we demonstrated that midlife pain is increasing in the United States across birth years or cohorts, with individuals born in later years already reporting more pain in the earlier years of midlife (a similar pattern was observed in England and Mexico). In Continental, Mediterranean, and Nordic Europe and South Korea, later-born cohorts reported less pain over the course of midlife compared to their earlier-born counterparts. The patterns of reported pain in the United States reflect similar declines in other measures of midlife health found in prior studies, including cognition, mental health, and functional limitations.^5,6,11 Overall, this trend in worsening health across cohorts suggests the existence of elements in modern American society and culture of health that may be exacerbating pain and other issues for middle-aged individuals. With the existence of these troubling pain trajectories established not only in America, but in several peer nations, an evaluation of proximal culture and context-driven factors related to self-reported pain (ie, financial burdens, use of health care resources, population-level differences, and the social construction of pain and health) is critical to understanding and intervening in this veritable health crisis (see⁵⁸). Midlife pain is an international issue with broad-reaching consequences and efforts must be made to implement nationwide solutions.

Supplementary material

Supplementary material is available at the American Journal of Epidemiology online.

Funding

This work was supported by the National Institute on Aging, National Institutes of Health (R01 AG079523) awarded to F.J.I., M.E.L., and D.G.

Conflict of interest

The authors declare no conflicts of interest.

Data availability

This study was preregistered, and further information is provided through the Open Science Framework (https://osf.io/34j7b/?view_only=09d06a8a52044cb0aa98dd5e32f99d9b).