Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 1 | 14 ORIGINAL RESEARCH Childhood adversity and leisure time physical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study Phoebe W. Hwang1, Cristiano dos Santos Gomes2, Mohammad Auais3, Kathryn L. Braun1, Catherine M. Pirkle1 1 Office of Public Health Studies, University of Hawaii at Mānoa, Hawaii, USA; 2 Federal University of Rio Grande do Norte, Natal, Brazil; 3 School of Rehabilitation Therapy, Queen’s University, Ontario, Canada. Corresponding author: Phoebe W. Hwang Address: 1960 East West Road, Biomedical Sciences Bldg #D104T, Honolulu, Hawaii 96822; Telephone: (808) 232-3223; E-mail: pwnhwang@hawaii.edu Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 2 | 14 Abstract Aim: The purpose is to examine the relationship between childhood adversity and leisure time physical activity (LTPA) among community-dwelling older adults from high and middle-income sites. Methods: Cross-sectional analysis of 2012 data from older adult ages 64-75 years old from King- ston, Canada; St. Hyacinthe, Canada; Tirana, Albania; Manizales, Colombia; and Natal, Brazil. Principal exposure variables were childhood social and economic adversity. Covariates included participant age, sex, income, and educational attainment. Outcome variables were LTPA and lei- sure time sports activity (LTSA). Results: High-income sites had higher LTPA prevalence than middle-income sites. Females were less likely to engage in LTPA compared to males in Tirana (OR:0.53, 95%CI:0.30-0.94), but were more likely to engage in LTPA in Manizales (OR:2.54, 95%CI:1.54-4.18). Low education was less likely than high education to engage in LTPA in Kingston (OR:0.38, 95%CI:0.19-0.73) and Natal (OR: 0.52, 95%CI:0.28-0.97). Low income was less likely than high income to engage in LTPA in St. Hyacinthe (OR: 0.42, 95%CI:0.20-0.89) and Manizales (OR:0.33, 95%CI:0.16-0.55). In Tirana, low income was more likely than high income to engage in LTPA (OR:5.27, 95%CI:2.06-13.51). Conclusions: Childhood economic and social adversity were not significantly associated with LTPA. Sex, income, and education were associated with older adult PA engagement, however the direction of the association varied by site location. This suggests that the paradigms surrounding PA behavior may vary from city to city. Understanding the site-specific risk factors to PA engage- ment may better inform clinical recommendations and public health approaches to increase PA engagement among older adults across the globe. Keywords: childhood adversity, gerontology, global health, physical activity. Conflicts of interest: None declared. Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 3 | 14 Introduction Physical activity (PA) is protective against chronic diseases and delays the onset of age- related health complications (1). Leisure time physical activity (LTPA) in particular is more effective in improving overall health than transportation, occupational, and sport-re- lated PA among older adults. Unfortunately, the amount of LTPA decreases as age in- creases (2). A large portion of PA literature explores in- dividual level theories of PA behavior change, such as self-efficacy theory and the transtheoretical model, or individual-micro- environment level theories, such as the social cognitive model (3). Consequently, there is no shortage of behavior-based interventions directed to increase older adult PA. Whether home, group, or educational-based, evalua- tions of these interventions have come to the same conclusion: individual behavior rein- forcement strategies alone are not effective in maintaining older adult PA behavior (4). Eti- ological studies applying a life course per- spective may be informative for interventions aimed in improving PA (2). Among the many exposures life course re- searchers have examined, early-life expo- sures appear the most cogently popular. Stud- ies have shown that early life exposures and socio-demographic characteristics affect an individual’s health behaviors and outcomes. Gender, social and material adversity, and living in a disadvantaged neighborhood are all documented to influence overall health during adulthood (5,6). These findings sug- gest that early childhood events may have long-term consequences on health behaviors and that PA behaviors may have roots situ- ated in early life circumstances. This study focused on community dwelling older adults of diverse socioeconomic status and global settings, recruited as part of the In- ternational Mobility in Aging Study (IMIAS). The objective is to examine the re- lationship between childhood adversity, and self-reported PA behaviors. Since early-life adversity negatively impacts many later life health behaviors, the authors hypothesize that childhood adversity is associated with lower levels of older adult PA behavior. Previous early life adversity studies that utilize a life course model were unable to examine cross- societal influences on behaviors due to sam- ple homogeneity. Cross-societal investiga- tions may provide insights on the contribu- tion of broad social structures to PA behav- iors, which in turn, may improve interven- tions geared at individual behavior change. Methods Site Location Descriptions IMIAS is a longitudinal study focused on older adult health. Baseline data were ob- tained in 2012, with follow-up collections in 2014 and 2016 (7). Data were community samples collected from five distinct study sites: Kingston and St. Hyacinthe, Canada; Tirana, Albania; Natal, Brazil; and Maniza- les, Colombia. The entire sample size is 2002 (roughly 200 men and 200 women from each site), which is large enough to examine how childhood adversity influences later life physical activity behavior. Population socio- economic, cultural, and religious de- mographics within each study site are rela- tively homogenous, whereas between sites there is substantial heterogeneity in socio-de- mographic characteristics. This give us a broad spectrum of different life exposures and health outcomes, thus providing a comprehensive picture of life course exposures and later life health outcomes across the globe. For a detailed description on Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 4 | 14 site locations and study, please refer to Gomez et al (7). Population and Data Collection Participants of this study are male and female community-dwelling older adults age 65-74 years. At Canadian sites, university ethics committees did not allow researchers to re- cruit potential participants directly. Family physicians sent letters of invitation to poten- tial participants that invited them to contact a field coordinator for further information re- garding the study. Participants were recruited from health center registries in Tirana, Natal, and Manizales. A random sample of potential participants was drawn from health center registries, and these individuals were re- cruited directly by interviewers. Interviewers were trained with a standardized protocol. Comparisons of recruited participants to cen- sus data suggest samples are representative of the towns/cities from which they were re- cruited (7). Individuals who had four or more errors on Leganes Cognitive Test orientation scale (8) were excluded from the study. Low scores indicated inability to complete study procedures. Recruitment continued until about 400 responses were obtained in each locale. Exposure Childhood adversity was measured using a series of retrospective questions on events that occurred within the first 15 years of the participants’ life. IMIAS survey questions re- garding childhood adversity were from the Survey on Health and Well-being Elders (SABE study) (9,10), and the Canadian Com- munity Health Survey (CCHS) (11). The events were: death of parent, parental sub- stance abuse, parental divorce, witnessing physical violence in the family, low eco- nomic status, having been hungry, having been physically abused, and parental unem- ployment. Members of the IMIAS team pre- viously performed an exploratory factor anal- ysis on these indicators to yield two catego- ries: economic adversity (low economic sta- tus, hunger, and parental unemployment), and social adversity (parental substance abuse, witnessing family physical violence, having been physically abused) (12). Adver- sity summary scores of economic and social adversity were recoded into two variables with binary responses—having experienced adversity (having experienced >0 of the indicators listed above) in childhood and no adversity experiences in childhood (having experienced none of the indicators listed above). Covariates Education, income, age, and sex were chosen as covariates based on research into the social determinants of health (13). Education was previously trichotomized into three catego- ries: illiterate/primary school only, secondary schooling, and post-secondary schooling. Analyses indicated insufficient variability within sites for comparison across sites. To allow for comparisons across sites, total years of education was split categorically into ter- tiles of high, medium, and low education by site to obtain a variable called “relative edu- cation”. Thus, it is possible for a participant to have high educational attainment relative to his/her community, but medium or low at- tainment compared to another site in IMIAS. Sex is an interviewer reported categorical variable (male/female). Age is a self-reported continuous variable re-coded into a binary categorical variable (64-69/70-75 years). In- come is a self-reported continuous variable of annual income recoded into an ordinal varia- ble (poor/middle/high) based on site-specific Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 5 | 14 poverty thresholds (7). Site location is based on the location of data collection. Outcomes The outcomes for this study were LTPA and LTSA. LTPA was defined as leisure time ac- tivity that involved bodily movement pro- duced by large skeletal muscles that require energy expenditure (14). LTSA was defined as any reported leisure time activity that is considered an official event in the Olympics (15). LTSA is a subset of LTPA. Participants were asked to report any leisure time activi- ties and to specify those activities. Responses were categorized into yes or no LTPA or LTSA based on the definitions above. Statistical Analysis Bivariate analyses were performed using Pearson’s chi-squared test for categorical data in order to assess potential differences in proportions between different groups. As- sumptions were met for all comparisons. The exposures and covariates listed above were tested as correlates to LTPA and LTSA be- havior using logistic regression. Preliminary analysis demonstrated a strong site-specific interaction with the outcome variables childhood social and economic adversities. This was expected given the substantial economic and societal differences between the sites. Therefore, this study focuses on the effect modification per site and analyses were stratified by site to highlight the different relationships. Please refer to the IMIAS cohort profile for additional information regarding study sites (7). All regression mod- els statistically adjusted for age, educational attainment, current income, sex, and site lo- cation. STATA/SE 14.0 was used to conduct the analyses. Results The prevalence of LTPA and LTSA engage- ment by site is displayed in Table 1. Kingston (68.1%) and St. Hyacinthe (51.4%) had higher prevalence of LTPA compared to Ti- rana (17.5%), Manizales (27.3%), and Natal (22.6%). Similar patterns were also observed in LTSA. Of all the participants, 36.7% in Kingston, 31.7% in St. Hyacinthe, 4.1% in Tirana, 5.7% in Manizales, and 5.5% in Natal engaged in LTSA. Table 1. Proportion of participants reporting leisure time physical and sports activity en- gagement by site Kingston (N=398) St. Hyacinthe (N=401) Tirana (N=394) Manizales (N=407) Natal (N=402) LTPA engagement, n (%)* 68.1% 51.4% 17.5% 27.3% 22.6% LTSA engagement, n (%)† 36.7% 31.7% 4.1% 5.7% 5.5% Missing data: Kingston=24; St. Hyacinthe= 46; Tirana= 7; Manizales= 10. *LTPA = activity done for leisure that results in energy expenditure by major skeletal muscles. †LTSA = activity done for leisure that requires physical exertion and skill for competition. Table 2 summarizes socio-demographic characteristics and adversity. In Manizales, compared to men, women were significantly more likely to report LTPA engagement (33.8% versus 21.9%). At both Canadian sites, those with higher levels of education were significantly more likely to report LTPA compared to those with medium and low site-specific education levels. In King- ston for example, 81.8% of highly educated participants report LTPA compared to 63.5% of those with low education. It should be Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 6 | 14 noted, however, that even low educated par- ticipants from Kingston and St. Hyacinthe re- ported more LTPA than any educational cat- egory at the middle-income sites. Income was significantly associated with LTPA engage- ment in St. Hyacinthe, Tirana, and Maniza- les. However, the nature of these associations varied by site. In both St. Hyacinthe and Manizales, high income participants were more likely to report LTPA engagement (67.4% and 41.7%, respectively), compared to poor income participants (47.2% and 23.2%, respectively). The opposite was true in Tirana. Poor income participants were more likely to report LTPA engagement (32.6%) compared to high income (10.6%). In Tirana, 21.3% of participants who experi- enced childhood economic adversity engaged in LTPA compared to 13.3% of those who didn’t experience childhood economic adver- sity. Table 2. Summary of leisure time physical activity engagement (LTPA)‡ by participant so- cio-demographic characteristics and childhood adversity, according to site LTPA Engagement Kingston (N=398) St. Hyacinthe (N=401) Tirana (N=394) Manizales (N=407) Natal (N=402) Sex (%) Male 76.6% 59.9% 21.7% 21.9%† 24.0% Female 68.7% 56.4% 14.3% 33.8% 21.4% Age in years (%) 64 to 69 72.4% 59.5% 17.1% 26.9% 24.2% 70 to 74 72.6% 55.5% 18.7% 29.2% 20.9% Education (%)¶ Low 63.5%† 50.0%† 17.9% 26.1% 16.7%† Medium 76.5% 61.3% 18.3% 23.2% 24.5% High 81.8% 66.1% 17.0% 35.9% 19.0% Income (%) Poor 64.4% 47.2%* 32.6%* 23.2%† 12.1% Middle 72.9% 66.4% 19.0% 28.07% 21.0% High 74.5% 67.4% 10.6% 41.7% 27.6% Childhood Economic Adver- sity (%)§ Yes 71.7% 55.0% 21.3%† 28.9% 22.0% No 72.7% 59.7% 13.3% 27.4% 22.0% Childhood Social Adversity (%)II Yes 74.7% 55.0% 21.7% 27.7% 26.2% No 71.6% 59.1% 17.0% 28.2% 21.4% Pearson’s Chi-square analysis was used to test for association of categories within sites *p<0.001 †p<0.05 ‡Leisure time physical activity is defined as activity done for leisure that results in energy expenditure by major skeletal muscles. §Childhood economic adversity is defined as having experienced poor economic status, hunger, or parental unemployment. IIChildhood social adversity is defined as having experienced parental substance abuse, family physical violence, or physical abuse. ¶Education calculated from total years of education categorized by site-specific tertiles Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 7 | 14 Table 3 summarizes socio-demographic characteristics and adversity by LTSA en- gagement. Men were significantly more likely to report LTSA engagement in King- ston (49.3%), St. Hyacinthe (43.1%), and Ti- rana (7.1%) compared to women (29.8%, 29.3%, and 1.5%, respectively). The younger age group (43.8%) was significantly more likely to engage in LTSA compared to the older age group (32.9%) only in Kingston. In Manizales, high education and high income were significantly associated with LTSA en- gagement. In Tirana, presence of childhood economic adversity was significantly associ- ated with LTSA engagement. In Natal, pres- ence of childhood social adversity was signif- icantly associated with LTSA. Table 3. Summary of leisure time sport activity engagement (LTSA)‡ by participant socio- demographic characteristics and childhood adversity, according to site LTSA Engagement Kingston (N=398) St. Hyacinthe (N=401) Tirana (N=394) Manizales (N=407) Natal (N=402) Sex (%) Male 49.3%* 43.1%† 7.1%† 6.1% 6.8% Female 29.8% 29.3% 1.5% 5.5% 4.3% Age in years (%) 64 to 69 43.8%† 35.3% 3.6% 7.1% 6.5% 70 to 74 32.93% 36.7% 4.7% 4.3% 4.3% Education (%)¶ Low 33.9% 34.5% 5.2% 4.9%† 5.1% Medium 43.1% 33.1% 3.9% 2.9% 5.8% High 39.4% 39.1% 3.0% 10.3% 5.8% Income (%) Poor 34.0% 32.0% 5.2% 4.9%† 5.1% Middle 45.9% 36.6% 3.9% 2.9% 5.8% High 39.4% 40.2% 3.0% 10.3% 5.6% Childhood Economic Ad- versity (%)§ Yes 44.2% 34.9% 6.3%† 5.4% 4.6% No 36.8% 36.3% 1.2% 6.09% 7.1% Childhood Social Adversity (%)II Yes 35.8% 34.1% 5.8% 2.1% 9.4%† No 40.3% 36.4% 3.8% 7.0% 4.1% Pearson’s Chi-square analysis was used to test for association of categories within sites. * p<0.001 † p<0.05 ‡ Leisure time sport activity is defined as activity done for leisure that requires physical exertion and skill for competition. § Childhood economic adversity is defined as having experienced poor economic status, hunger, or parental unemployment. II Childhood social adversity is defined as having experienced parental substance abuse, family physical violence, or physical abused ¶ Education calculated from total years of education categorized by site-specific tertiles Tables 4 and 5 summarize the results of the multivariate models. Childhood social and economic adversities were not significantly associated with LTPA engagement in all sites. In Kingston, participants with lower ed- ucation were less likely to engage in LTPA (OR:0.38, 95%CI:0.19-0.73) compared to high education. In St. Hyacinthe, poor in- come participants were less likely to engage Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 8 | 14 in LTPA (OR:0.42, 95%CI:0.20-0.89) com- pared to high income. The opposite relation- ship was observed in Tirana. Poor (OR:5.27, 95%CI:2.06-13.51) and middle income (OR:2.44, 95%CI:1.20-4.99) participants were more likely to engage in LTPA com- pared to high income. In Manizales, women were more likely to engage in LTPA com- pared to men (OR:2.54, 95%CI:1.54-4.18). Also, poor (OR:0.33, 95%CI:0.16-0.65) and medium income participants (OR:0.46, 95%CI:0.23-0.92) were less likely to engage in LTPA compared to high income partici- pants from this site. In Natal, participants with low education were also less likely to engage in LTPA compared to high education (OR:0.52, 95%CI0.28-0.97). For LTSA in Natal, participants who experienced child- hood social adversity were more likely to en- gage in LTSA compared to those who did not (OR:3.31, 95%CI:1.31-8.41). Females were less likely to engage in LTSA compared to males in Kingston (OR:0.40, 95%CI:0.25- 0.65) and Tirana (OR:0.17, 95%CI:0.04- 0.64). In Manizales, participants with medium level education were less likely to engage in LTSA compared to high level (OR:0.25, 95%CI:0.10-0.82). In Natal, mid- dle income participants were less likely to en- gage in LTSA compared to high income (OR:0.29, 95% CI:0.10-0.82). Table 4. Association of participant socio-demographic characteristics and childhood adver- sity measures with self-reported LTPA†,‡ Kingston (N=398) St. Hyacinthe (N=401) Tirana (N=394) Manizales (N=407) Natal (N=402) OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI Sex Male 1.00 1.00 1.00 1.00 1.00 Female 0.74 0.44-1.23 1.24 0.74-2.05 0.53* 0.30-0.94 2.54* 1.54-4.18 0.99 0.60-1.65 Age (years) 64 to 69 1.00 1.00 1.00 1.00 1.00 70 to 74 1.18 0.72-1.95 0.89 0.56-1.43 0.91 0.53-1.58 1.08 0.68-1.70 0.81 0.50-1.31 Education¶ Low 0.38* 0.19-0.73 0.78 0.80 0.38-1.71 0.83 0.44-1.54 0.52* 0.28-0.97 Medium 0.88 0.43-1.78 1.05 0.55-2.00 0.83 0.41-1.69 0.65 0.36-1.19 0.85 0.47-1.52 High 1.00 1.00 1.00 1.00 1.00 Income Poor 0.82 0.40-1.67 0.42* 0.20-0.89 5.27* 2.06-13.51 0.33* 0.16-0.65 0.35 0.11-1.10 Middle 1.31 0.72-2.39 0.95 0.49-1.83 2.44* 1.20-4.99 0.46* 0.23-0.92 0.76 0.44-1.31 High 1.00 1.00 1.00 1.00 1.00 Childhood Econo- mic Adversity§ Yes 0.96 0.56-1.64 0.86 0.53-1.39 1.65 0.92-2.93 1.29 0.79-2.09 1.19 0.71-1.99 No 1.00 1.00 1.00 1.00 1.00 Childhood Social Ad- versityII Yes 1.42 0.78-2.56 1.08 0.63-1.85 0.93 0.47-1.87 0.90 0.52-1.56 1.43 0.84-2.45 No 1.00 1.00 1.00 1.00 1.00 *p<0.05 †Leisure time physical activity is defined as activity done for leisure that results in energy expenditure by major skeletal muscles. ‡Logistic regression models have been adjusted for age, sex, education, and income. §Childhood economic adversity is defined as having experienced poor economic status, hunger, or parental unemployment. IIChildhood social adversity is defined as having experienced parental substance abuse, family physical violence, or physical abuse ¶Education calculated from total years of education categorized by site-specific tertiles Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 9 | 14 Table 5. Association of participant socio-demographic characteristics and childhood adver- sity measures with self-reported LTSA†,‡ Kingston (N=398) St. Hyacinthe (N=401) Tirana (N=394) Manizales (N=407) Natal (N=402) OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI Sex Male 1.00 1.00 1.00 1.00 1.00 Female 0.40* 0.25-0.65 0.65 0.39-1.08 0.17* 0.04-0.64 1.09 0.44-2.67 0.70 0.27-1.85 Age (years) 64 to 69 1.00 1.00 1.00 1.00 1.00 70 to 74 0.64 0.40-1.02 1.14 0.71-1.83 1.04 0.36-2.97 0.59 0.24-1.46 0.64 0.25-1.60 Education¶ Low 0.89 0.49-1.61 0.95 0.51-1.75 2.05 0.46-9.10 0.54 0.17-1.69 1.56 0.48-5.10 Medium 1.62 0.90-2.93 1.17 0.62-2.19 1.28 0.29-5.65 0.25* 0.07-0.87 1.44 0.46-4.54 High 1.00 1.00 1.00 1.00 1.00 Income Poor 1.01 0.50-2.05 0.59 0.28-1.24 3.88 0.79-10.07 0.65 0.18-2.36 0.53 0.10-2.69 Middle 1.21 0.69-2.10 0.84 0.44-1.59 1.37 0.37-5.03 1.42 0.44-4.58 0.29* 0.10-0.82 High 1.00 1.00 1.00 1.00 1.00 Childhood Econo- mic Adversity§ Yes 1.62 0.98-2.66 0.88 0.54-1.43 4.35 0.94-20.13 1.31 0.50-3.39 0.60 0.24-1.51 No 1.00 1.00 1.00 1.00 1.00 Childhood Social AdversityII Yes 0.78 0.46-1.34 1.10 0.64-1.92 0.76 0.22-2.69 0.24 0.5-1.10 3.31* 1.31-8.41 No 1.00 1.00 1.00 1.00 1.00 *p<0.05 †Leisure time physical activity is defined as activity done for leisure that results in energy expenditure by major skeletal muscles. ‡Logistic regression models have been adjusted for age, sex, education, and income. §Childhood economic adversity is defined as having experienced poor economic status, hunger, or parental unemployment. IIChildhood social adversity is defined as having experienced parental substance abuse, family physical violence, or physical abuse. ¶Education calculated from total years of education categorized by site-specific tertiles. Discussion This study examined the relationship be- tween childhood adversity, occurring before 15 years of age, and self-reported later life PA behaviors among community-dwelling older adults from diverse global settings. This study hypothesized that since previous IMIAS studies demonstrated a strong associ- ation between childhood adversity and older adult physical performance, there must also be a relationship between childhood adver- sity and physical activity behavior. However, findings from these studies demonstrated that childhood social adversity was associated with self-reported LTPA only in Tirana and Natal, and childhood economic adversity was not associated with PA engagement at all. As expected, sex, income and education were as- sociated with older adult PA engagement, however the direction of the association var- ied by site location. This suggests that the paradigms surrounding PA behavior may vary, possibly depending on geographical, cultural, social, and/or historical influences. Thus, the risk factors associated with low PA engagement differ from city to city. Under- standing the site-specific risk factors to PA Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 10 | 14 engagement may better inform clinical rec- ommendations and public health approaches to increase PA engagement among older adults across the globe. Childhood adversity and physical activity behavior In a previous IMIAS study, the presence of social and economic childhood adversity was associated with poor physical performance. However, the mechanisms of this relation- ship were unexplored. Since physical activity is commonly associated with good physical performance (16), we hypothesized that low physical activity engagement may partially explain the association observed by Sousa et al. Contrary to our hypotheses, self-reported childhood adversity experiences did not cor- relate strongly with LTPA/LTSA engage- ment among older adults. Moreover, the na- ture of the association differed from what we hypothesized. In Tirana, self-reported child- hood economic adversity was marginally as- sociated with both LTPA/LTSA engagement. While not statistically significant, partici- pants in Tirana who reported childhood eco- nomic adversity had 4.35 times the odds of reporting LTSA engagement. In Natal, re- porting childhood social adversity was also associated with LTSA. There is no doubt that early life adversity is associated to poor health behaviors and health outcomes in later life. Therefore, it was puzzling to find that early life adversity did not correlate strongly with LTPA/LTSA. Unfortunately, there is currently no literature that examines the relationship between child- hood adversity and later life physical activity behaviors to which we can compare this study. Our current results suggest that physi- cal activity behavior may not explain the re- lationship between early life adversity and physical performance. One possible explanation for our contrary findings may be selective survival, since data were collected only among older adults aged 65-74 (17), and the average life expectancies at birth between the sampled sites varied greatly. For example, in 1960, the life expec- tancy at birth in Brazil was 54.7 years, whereas Canada’s average life expectancy was 71.13 years old (18). Therefore, those in Brazil who survived until study recruitment reflect the survivors of their birth cohort. Se- lective survival has been observed in previ- ous studies where the differences in health and mortality between groups of high and low socioeconomic statuses decline as age in- creases (19). In fact, a study conducted in Is- rael found that older adults who survived past 61 years old have higher community resili- ence scores compared to the younger popula- tion, indicating that healthy older adults have a better ability to alleviate the detrimental ef- fects of adverse events (20). This may ex- plain why childhood adversity was associated with physical activity engagement in the mid- dle income sites. Those who managed to overcome childhood adversity and live past the average life expectancy of their cohort may have distinctively different behaviors from those who did not survive. Site-specific influences on physical activity behavior Overall, childhood adversity did not correlate as strongly to LTPA/LTSA as compared to the other socio-demographic factors that were observed in this study. LTPA/LTSA en- gagement was notably greater in high income (Kingston, St. Hyacinthe) compared to mid- dle-income sites (Tirana, Manizales, Natal). These results were consistent with a study that analyzed physical activity trends using data from the World Health Organization. Among adults aged 15 years and over, Brazil, Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 11 | 14 Colombia, and Albania’s physical inactivity rates were higher than Canada’s (21). Addi- tionally, the authors found that LTPA in- creased as occupational PA decreased over time in high-income countries. The same analysis could not be done with low and mid- dle-income countries because these data were not available (21). Our study is one of the first to estimate LTPA prevalence in community dwelling older adults from middle-income settings. The observed associations between socio-de- mographic factors and reported PA behavior varied notably by study site as well. The re- lationships between LTPA/LTSA engage- ment and socio-demographic variables may be dependent on site-specific norms. For ex- ample, income was significantly associated with LTPA engagement in St. Hyacinthe, Ti- rana, and Manizales, but not in a consistent direction. In Tirana, poor income participants were five times more likely to engage in LTPA compared to high income participants, whereas in St. Hyacinthe, poor income par- ticipants were less likely to engage in LTPA compared to high income. Our study further justifies that social norms may influence PA behaviors. Similar results can be found within the United States (22), and high-in- come East Asian countries (23). However, to the authors’ knowledge, no studies have iden- tified cross-societal differences of factors as- sociated to LTPA engagement across study sites of varying income categorization. LTPA versus LTSA This paper examined PA behavior by type— LTPA and LTSA. LTSA is a subcategory of LTPA. It can be said that all LTSA is consid- ered LTPA, but not all LTPA is considered LTSA. LTSA have a set of rules and goals to train and excel in specific athletic skills. Moreover, LTSA in general, has a more com- petitive edge (24). In this study, sex was a significant correlate to LTSA engagement for all sites except Natal. Males were more likely to engage in LTSA compared to females. Yet, sex was not significantly associated to LTPA. Results from this study were congruent to other studies that examined sex differences in PA behaviors. In the United States, females are less likely to engage in vigorous PA from adolescence to adulthood (2). Among college attending young adults, females were less likely to engage in sports compared to males (25). Historical and anthropological studies suggest that males experience an evolution- ary history of physical competition for court- ship and warfare more often than females (26). Further, men are more likely to engage in extreme physical competitive aggression compared to women (27). Understanding how sex is correlated with physical activity type preference may give us insight on the so- cial norms of PA, and guide sex-specific PA intervention design. Limitations Although the large gap between middle and high-income sites clearly shows a difference in PA engagement prevalence, bivariate site- specific analyses that examine the correlates to PA engagement may have been underpow- ered as very few participants from middle in- come sites reported LTSA engagement, and relatively few reported LTPA engagement. A second limitation to this study is that the LTPA/LTSA measure used has not been pre- viously validated. However, widely used LTPA measurement tools such as Godin Lei- sure Time Questionnaire, International Phys- ical Activity Questionnaire, and Sedentary Behavior Questionnaire have been only vali- dated with populations aged 18 to 69 years old, just missing the older adult population. Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 12 | 14 Further, pilot studies were previously con- ducted to validate the IPAQ in Santa Cruz, Brazil, using accelerometers. Results showed that IPAQ had poor validity (28); therefore, it was not considered for this study. Lastly, since this is a secondary data analysis, sample size could not be determined a priori. Thus, the sample size may not be powered for this particular analysis. However, given the rich- ness of the data, it allows us to deeply exam- ine the multiple factors involved in the life course. Conclusions Since the 1990’s, there has been a progress in research that examines environment-level factors correlates and causes of PA. Unfortu- nately, many studies focused only on high-in- come countries (29). As the world ages, and the global burden of non-communicable dis- eases increase, health behaviors such as PA are becoming more relevant in lower-income settings. Several studies have shown a link between childhood adversity at adult PA be- havior, but have not followed the participants into older adulthood (30). This study is one of the few that makes cross-societal infer- ences on the effects of childhood adversity on older adult PA behavior and it highlights the powerful influences of social norms on LTPA/LTSA engagement. References 1. Hackney ME, Hall CD, Echt KV, Wolf SL. Dancing for Balance: Fea- sibility and Efficacy in Oldest-Old Adults With Visual Impairment. Nurs Res 2013;62:138-43. 2. Caspersen CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc 2000;32:1601-9. 3. King AC, Stokols D, Talen E, Brass- ington GS, Killingsworth R. Theoret- ical approaches to the promotion of physical activity: forging a transdis- ciplinary paradigm. Am J Prev Med 2002;23:15-25. 4. Ashworth NL, Chad KE, Harrison EL, Reeder BA, Marshall SC. Home versus center based physical activity programs in older adults. Cochrane Database of Systematic Reviews; 2005. Available from: http://onlineli- brary.wiley.com/doi/10.1002/146518 58.CD004017.pub2/full (accessed: July 26, 2020). 5. Gustafsson PE, San Sebastian M. When does hardship matter for health? Neighborhood and individual disadvantages and functional somatic symptoms from adolescence to mid- life in the Northern Swedish Cohort. PloS One 2014;9:e99558. 6. Kajeepeta S, Gelaye B, Jackson CL, Williams MA. Adverse childhood experiences are associated with adult sleep disorders: a systematic review. Sleep Med 2015;16:320-30. 7. Gomez F, Zunzunegui MV, Al- varado B, Curcio CL, Pirkle CM, Guerra R, et al. Cohort Profile: The International Mobility In Aging Study (IMIAS). Int J Epidemiol 2018;47:1393. 8. Yébenes MJG, Otero A, Zunzunegui MV, Rodríguez-Laso A, Sánchez- Sánchez F, Del Ser T. Validation of a short cognitive tool for the screen- ing of dementia in elderly people with low educational level. Int J Ger- iatr Psychiatry 2003;18:925-36. 9. Gomes MMF, Turra CM, Fígoli MGB, Duarte YA, Lebrão ML. Past and present: conditions of life during http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004017.pub2/full http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004017.pub2/full http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004017.pub2/full Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 13 | 14 childhood and mortality of older adults. Rev Saude Publica 2015;49:93. 10. Guerra RO, Alvarado BE, Zun- zunegui MV. Life course, gender and ethnic inequalities in functional disa- bility in a Brazilian urban elderly population. Aging Clin Exp Res 2008;20:53-61. 11. Fuller-Thomson E, Stefanyk M, Brennenstuhl S. The robust associa- tion between childhood physical abuse and osteoarthritis in adulthood: findings from a representative com- munity sample. Arthritis Care Res 2009;61:1554-62. 12. Albuquerque Sousa AC, Guerra RO, Tu MT, Phillips SP, Guralnik JM, Zunzunegui MV. Lifecourse adver- sity and physical performance across countries among men and women aged 65-74. PloS One 2014;9:e102299. 13. Marmot M. Social determinants of health inequalities. Lancet 2005;365:1099-104. 14. World Health Organization. Physical activity [Internet]. WHO; 2016. Available from: http://www.who.int/topics/physi- cal_activity/en/ (accessed: July 26, 2020). 15. International Olympic Committee. List of Summer and Winter Olympic Sports [Internet]. Available from: https://www.olympic.org/sports (ac- cessed: July 26, 2020). 16. Mazzeo RS, Cavanagh P, Evans WJ, Fiatarone M, Hagberg J, McAuley E, et al. ACSM Position Stand: Exer- cise and Physical Activity for Older Adults. Med Sci Sports Exerc 1998;30:992-1008. 17. Willson AE, Shuey KM, Elder Jr GH. Cumulative advantage processes as mechanisms of inequality in life course health1. Am J Sociol 2007;112:1886-924. 18. World Bank. Life expectancy at birth, total (years) [Internet]. Availa- ble from: http://data.worldbank.org/indica- tor/SP.DYN.LE00.IN (accessed: July 26, 2020). 19. Andersen O, Laursen L. Health and Factors causing disease-in a social perspective. Danmarks Statistics; 1998. 20. Cohen O, Geva D, Lahad M, Bolotin A, Leykin D, Goldberg A, et al. Community Resilience throughout the Lifespan–The Potential Contribu- tion of Healthy Elders. PLoS One 2016;11:e0148125. 21. Hallal PC, Andersen LB, Bull FC, Guthold R, Haskell W, Ekelund U, et al. Global physical activity levels: surveillance progress, pitfalls, and prospects. Lancet 2012;380:247-57. 22. Arredondo EM, Elder JP, Ayala GX, Campbell N, Baquero B, Duerksen S. Is parenting style related to chil- dren’s healthy eating and physical activity in Latino families? Health Educ Res 2006;21:862-71. 23. Chen DR, Lin YC. Social identity, perceived urban neighborhood qual- ity, and physical inactivity: A com- parison study of China, Taiwan, and South Korea. Health Place 2016;41:1-10. 24. Chick GE. The cross-cultural study of games. Exerc Sport Sci Rev 1984;12:307-37. 25. Deaner RO, Geary DC, Puts DA, Ham SA, Kruger J, Fles E, et al. A http://www.who.int/topics/physical_activity/en/ http://www.who.int/topics/physical_activity/en/ https://www.olympic.org/sports http://data.worldbank.org/indicator/SP.DYN.LE00.IN http://data.worldbank.org/indicator/SP.DYN.LE00.IN Hwang PW, Gomes CdS, Auais M, Braun KL, Pirkle CM. Childhood adversity and leisure time phys- ical and sports activity in older adults: A cross-sectional analysis from the International Mobility in Aging Study (Original research). SEEJPH 2020, posted: 26 October 2020. DOI 10.4119/seejph-389 P a g e 14 | 14 sex difference in the predisposition for physical competition: Males play sports much more than females even in the contemporary US. PLoS One 2012;7:e49168. 26. De Block A, Dewitte S. Darwinism and the cultural evolution of sports. Perspect Biol Med 2009;52:1-16. 27. Archer J. Does sexual selection ex- plain human sex differences in ag- gression? Behav Brain Sci 2009;32:249-66. 28. Forget MF. Étude sur la validité et la fiabilité d’un questionnaire sur l’ac- tivité physique de personnes âgées de 65 à 74 ans, du Québec et du Brésil. 2012 [in French]. Available from: https://papyrus.bib.umon- treal.ca/xmlui/handle/1866/7133 (ac- cessed: July 26, 2020). 29. Bauman AE, Reis RS, Sallis JF, Wells JC, Loos RJ, Martin BW, et al. Correlates of physical activity: why are some people physically active and others not? Lancet 2012;380:258-71. 30. Juneau CE, Benmarhnia T, Poulin AA, Côté S, Potvin L. Socioeco- nomic position during childhood and physical activity during adulthood: a systematic review. Int J Public Health 2015;60:799-813. __________________________________________________________________ © 2020 Hwang et al; This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. https://papyrus.bib.umontreal.ca/xmlui/handle/1866/7133 https://papyrus.bib.umontreal.ca/xmlui/handle/1866/7133