Patterns and Determinants of Physical Inactivity in Rural and Urban Areas in Peru: a Population-Based Study

Patterns and Determinants of Physical Inactivity in Rural and Urban Areas in Peru: A Population-Based Study

J. Jaime Miranda, Rodrigo M. Carrillo-Larco, Robert H Gilman, Jose L. Avilez, Liam Smeeth, William Checkley, Antonio Bernabe-Ortiz, and the CRONICAS Cohort Study Group

Background: Physical inactivity and sedentary behaviors have been linked with impaired health outcomes. Establishing the physical inactivity profiles of a given population is needed to establish program targets and to contribute to international monitoring efforts. We report the prevalence of, and explore sociodemographical and built environment factors associated with physical inactivity in 4 resource-limited settings in Peru: rural Puno, urban Puno, Pampas de San Juan de Miraflores (urban), and Tumbes (semiurban). Methods: Cross-sectional analysis of the CRONICAS Cohort Study’s baseline assessment. Outcomes of interest were physical inactivity of leisure time (<600 MET-min/week) and transport-related physical activity (not reporting walking or cycling trips) domains of the IPAQ, as well as watching TV, as a proxy of sedentarism (≥2 hours per day). Exposures included demographic factors and perceptions about neighborhood’s safety. Associations were explored using Poisson regression models with robust standard errors. Prevalence ratios (PR) and 95% confidence intervals (95% CI) are presented. Results: Data from 3593 individuals were included: 48.5% males, mean age 55.1 (SD: 12.7) years. Physical inactivity was present at rates of 93.7% (95% CI 93.0%–94.5%) and 9.3% (95% CI 8.3%–10.2%) within the leisure time and transport domains, respectively. In addition, 41.7% (95% CI 40.1%–43.3%) of participants reported watching TV for more than 2 hours per day. Rates varied according to study settings (P < .001). In multivariable analysis, being from rural settings was associated with 3% higher prevalence of leisure time physical inactivity relative to highly urban Lima. The pattern was different for transport-related physical inactivity: both Puno sites had around 75% to 50% lower prevalence of physical inactivity. Too much traffic was associated with higher levels of transport-related physical inactivity (PR = 1.24; 95% CI 1.01–1.54). Conclusion: Our study showed high levels of inactivity and marked contrasting patterns by rural/urban sites. These findings highlight the need to generate synergies to expand nationwide physical activity surveillance systems.

Keywords: physical activity, sedentary lifestyle, television, prevalence, Peru

There is no doubt about the role of physical activity in accruing studies have reported that low levels of physical activity are common health gains.1 The benefits of physical activity have been widely in urban cities in Peru, for example 39% in Lima14 and 58% in documented in young2–4 and adult populations.5–7 Conversely, Arequipa.15 The opposite, higher levels of physical activity, has been sedentary lifestyle and physical inactivity have been established as documented in rural settings.14,16 These differences between rural risk factors for certain types of cancer such as breast or colorectal and urban areas could be due to urbanization, and understanding cancer.8,9 In addition, behavioral changes, such as slight reductions the role the built environment plays in this difference is important to in time spent sitting down, do have positive health effects.10 intervene. However, most of these were small studies conducted in A 10% reduction in the prevalence of insufficient physical specific sites, in isolation, which highlights the lack of nationwide activity by 2025 is 1 of the 9 global voluntary targets in the global data in Peru. action plan for the prevention and control of noncommunicable Peru is a diverse country with varied geographical scenarios diseases for the period 2013–2020.11 These country-level indicators spanning sea level, Andean mountains and Amazonian environ- should be monitored over time, yet Peru does not have a national ments, and with combinations of urban/rural and low/high-altitude survey to comply with such need.12 settings in each of them. The diversity of contexts, which will likely Physical activity is widely recognized as one of the key driv- impact on the profiles associated with physical activity, calls for ers of health changes related to urbanization.13 For example, some a better characterization of physical activity profiles in Peru and similar settings, especially since there is a lack of information regarding physical activity in high altitude settings with clear rural/ urban differences. These efforts could garner sufficient data and ini- Miranda ([email protected]), Carrillo-Larco, Gilman, Avilez, tiate momentum to step toward collection of nationwide prevalence Smeeth, Checkley, and Bernabe-Ortiz are with the CRONICAS Center of estimates to contribute to international monitoring mechanisms.11,12 Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, This study had 2 objectives: to report prevalence estimates Lima, Peru. Miranda is also with the Dept of Medicine, School of Medicine, of physical inactivity levels and TV watching, as a proxy of sed- Universidad Peruana Cayetano Heredia, Lima, Peru. Gilman and Checkley entarism, and to explore sociodemographical and neighborhood, are also with the Program in Global Disease Epidemiology and Control, as a proxy of built environment, factors associated with physical Department of International Health, Bloomberg School of Public Health, inactivity in 4 low-socioeconomic areas—Pampas de San Juan de Johns Hopkins University, Baltimore, MD; and the Asociación Benéfica Miraflores (highly urbanized), urban Puno, rural Puno and Tumbes PRISMA, Lima, Peru. Smeeth and Bernabe-Ortiz are with the Faculty (semiurban)—in Peru with a combination of rural/urban and low/ of Epidemiology and Population Health, London School of Hygiene and high altitude settings. We hypothesized that people in urban settings Tropical Medicine, London, United Kingdom. would be more physically inactive and sedentary than in rural areas.

654 Physical Inactivity in Peru 655

Methods Other Study Variables Study Design and Setting Demographic (sex, age); socioeconomic variables, based on number of years of education (6 years or less, 7–11 years, 12+ years) and The CRONICAS Cohort Study was designed to address the geo- socioeconomic status (measured using a wealth index based on graphical variation in the progression toward some noncommuni- assets and household facilities separately for each study site20 and cable diseases in Peru, and its methodology has been described in then combined into a single variable and presented in tertiles); study detail elsewhere.17 As a result, 4 Peruvian settings that differed by site (Lima, urban capital, sea level; urban Puno and rural Puno; and, level of urbanization and altitude, were included. Urban areas are Tumbes, semiurban, sea level) were recorded. defined as those sections which have at least 100 households grouped Behavioral risk factors included daily smoking (≥1 cigarette/ together. Rural areas have either dispersed households or less than day, self-report), heavy alcohol drinking (2 or more nights of 100 households grouped together. The sites were: Pampas de San alcohol intake in the past month and having ever drunk 6 or more Juan de Miraflores in Lima, a highly-urbanized low altitude setting drinks at a time, self-report), and fruits and vegetables intake (<5 (sea level), which has experienced significant but unplanned popula- portions per day, 5+ portions per day, self-report); body mass index tion growth; Puno, a high altitude area in the Peruvian Andes (3825 (normal weight, BMI = 18.5–24.9; overweight, BMI = 25–29.9; and m above sea level), divided into rural Puno and urban Puno, due obesity, BMI ≥30). Hypertension status (systolic blood pressure to the many small villages which surround the urban sections; and ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or self-report Tumbes, a semiurban low altitude (sea level), coastal setting where of physician diagnosis and current use of antihypertensive drugs)21 rural areas, comprising vast traditional landscapes, have become and type-2 diabetes (self-report of physician diagnosis and currently intermixed with rapidly-growing urban areas. For this analysis, receiving antihyperglycemic medication; or, fasting glucose ≥126 information of the baseline assessment, conducted between 2010 mg/dL)22 were also measured. and 2012, was used. Perceptions about neighborhood safety as a proxy of built environment characteristics (traffic, crosswalks, street illumina- Study Participants tion, walking safety during the day or night) were measured using questions of the Neighborhood Environmental Walkability Scale.23 Potential participants were randomly identified from the settings The aforementioned variables were treated as exposure vari- and stratified by age and sex. Subjects aged≥ 35 years with full-time ables and potential confounders in the regression models: when one residence in the area were invited to participate in the study. A sex- was the exposure of interest, the others were included as potential and age-stratified (35–44, 45–54, 55–64, and 65+ years) single-stage confounders in the regression analysis. random sampling procedure was performed using information of the most updated census in each site. Only 1 participant per household Statistical Methods was enrolled into the study.17 Statistical analysis was conducted in STATA 13 (Stata Corp, Col- lege Station, TX, USA). Initially, overall description of the study Ethical Procedures population was performed and the prevalence of physical inactivity All participants provided oral consent due to high illiteracy rates, and watching TV ≥2 hours with their respective 95% confidence especially in rural areas. The study was approved by the Institutional intervals (95% CI) were reported. Review Boards at Universidad Peruana Cayetano Heredia and A.B. Then, physical inactivity levels, based on leisure time and PRISMA, in Lima, Peru, and the Johns Hopkins University, in transport-related physical activity domains as well as watching TV Baltimore, MD, US. ≥2 hours, were tabulated and compared with the characteristics of the study population. Comparisons were performed by using Chi- squared test. In addition, built environment variables were also Outcomes tabulated according to our physical activity indicators.

Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 In the baseline assessment of the CRONICAS Cohort Study we Crude and adjusted Poisson regression models with robust stan- administered the leisure time and transport-related domains of the dard errors24 were created to determine the strength of association International Physical Activity Questionnaire (IPAQ). Although this between demographic and behavioral variables with physical inac- version has not been tested for reliability or validity in Peru, our tivity levels. In the multivariable model, all the variables were added decision to gather this data were based on guidance derived from to look for factors independently associated with the outcomes of research conducted in the Latin American region. A study by Hallal interest. In the case of built environment variables, Poisson regres- et al contrasted self-reported information collected using the IPAQ sion models were also created controlling for sex, age, education with objective measurements, and recommends focusing on leisure level, socioeconomic status, and study site. For all the regression time and transport-related physical activity.18 We defined physical models, prevalence ratios (PR) and 95% CI were calculated. inactivity for each of these 2 domains: leisure time inactivity was defined as doing none or very little physical activity (ie, <600 MET- min/week) during leisure time, whereas transport-related inactivity Results was defined as not reporting walking or cycling trips (ie, a single Participant Characteristics walk or cycle trip for 10 minutes or more was considered to be classified as physical active).19 Overall response rate after enrollment was 62.9% (4325/6872), In addition, the time spent watching TV, as a proxy of seden- and of these 83.3% (3601/4325) responded the physical activity tarism, was also assessed as part of the evaluation of physical inac- questionnaires. Eight physical activity records were incomplete; as a tivity. This variable was defined as reporting≥ 2 hours of watching result, data from 3593 individuals, 48.5% male, mean age 55.8 years TV per day during weekdays. (SD: 12.7) with all complete data were included in the analyses. Of

JPAH Vol. 13, No. 6, 2016 656 Miranda et al

them, 22.2% had ≥12 years of education and 90.5% had a family status, and among heavy drinkers. Conversely, it was less common income of USD <550 per month, indicating the predominance of among those aged 45+ years. Both Puno sites had lower prevalence low socioeconomic status among participants across sites. of this profile when compared with Lima. In terms of environmental factors (Table 4), only 2 parameters Levels of Physical Inactivity by Study Settings yielded an association with the outcomes of interest. Reporting feeling unsafe to walk at night, compared with those who do not, was Patterns of physical inactivity are depicted in Table 1. Of note, associated with a 2% lower prevalence of physical inactivity in the 93.7% (95% CI 93.0%–94.5%) of all population were physically leisure time physical activity domain. Using the transport-related inactive during leisure time. Rates varied across studies sites (P < domain, too much traffic in the living area was associated with 15% .001), ranging from 91% in urban Puno to 96% in rural Puno, and greater prevalence of physical inactivity. they were highest in both rural sites. In contrast, 9.3% (95% CI 8.3%–10.2%) of participants reported physical inactivity levels using the transport-related domain Discussion of the IPAQ. Rates also varied according to study site (P < .001), without a clear rural/urban pattern (eg, semiurban Tumbes in the Main Findings upper range of inactivity contrasted with rural Puno on the lower Our results show marked differences in the patterns of inactivity range). using the leisure time and transport-related domains of the IPAQ. A total of 1501 individuals (41.7%; 95% CI 40.1%–43.3%) These prevalence estimates varied widely across settings, and while reported watching TV ≥2 hours per day and varied depending on leisure time physical inactivity rates in the rural and semiurban sites the study site (P < .001): ranging from 14% in rural Puno to 51% were higher than in both urban sites, such clear differential rural/ in semiurban Tumbes. urban pattern was not observed in transport-related physical inactivity or in duration of watching TV, our surrogate for sedentarism. Profiles of Leisure Time and Transport-Related Similar findings have been observed in India25 and United States.26 Physical Inactivity Therefore, we could not find a definitive answer to our hypothesis that people in urban settings would be more physically inactive and The highest rates of physical inactivity according to the leisure sedentary than in rural areas. time domain of the IPAQ were observed in rural Puno (96%), in In multivariable analysis, being from rural sites was associated females (96%), and in those with the lowest education and the with 3% higher prevalence of leisure time physical inactivity relative lowest socioeconomic status (96% and 97%). On the contrary, in to highly urban Lima. The pattern is different for transport-related terms of the transport-related domain, the highest rates of physical physical inactivity in that both urban and rural sites in Puno had inactivity were observed among those living in Tumbes (21%), around 75% to 50% lower prevalence whereas semiurban Tumbes those aged ≥65 years (15%), and those with diabetes (19%) and had 186% higher prevalence of physical inactivity. Watching TV ≥2 hypertension (18%). hours per day was less frequent in both urban and rural Puno than Watching TV ≥2 hours per day was very common among heavy in Lima; and Tumbes showed no difference. These observations alcohol drinkers and among daily smokers (60%); whereas it was highlight a stark difference in patterns of transport-related physical lower in rural Puno (14%), those with <6 years of education (33%), activity and TV watching between Tumbes, a semiurban area, and and those in the lowest tertile of socioeconomic status (33%). the other settings. This constitutes a finding that would have gone When assessed by environmental factors, between 92% and unnoticed when country estimates are aggregated into traditional 95% of subjects were physically inactive during leisure time, national averages. Given that our definition of transport-related transport-related physical inactivity ranged between 6% and 12%, physical inactivity was based on as not reporting walking or cycling and watching TV ≥2 hours per day varied between 39% and 74% trips, these findings signal toward potential areas of high priority (Table 2). for the promotion of active commuting. Male sex and higher socioeconomic status were associated with Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 Factors Associated With Physical Inactivity lower physical inactivity levels when leisure time and transport- related domains of the IPAQ were assessed in multivariable models, In multivariable analysis (Table 3), being from rural Puno was but greater intake of fruits and vegetables was only associated with associated with 3% higher prevalence of leisure time physical lower leisure time physical inactivity levels. inactivity, when compared with highly urban Lima. The opposite With regard to environmental associated factors, insecurity to pattern, in the range of 5% to 6% lower prevalence, was observed walk at night was associated with a 2% lower prevalence of physical among males, higher number of years of education, and higher inactivity based on the leisure time domain of the IPAQ. This finding, socioeconomic status. albeit borderline significant, was in the unexpected direction and When using the transport-related domain of the IPAQ, older may reflect the role of chance, and therefore merits confirmation age, higher socioeconomic status, and having heavy alcohol drinking in other studies. In addition to our opposite-to-expected results, a were associated with greater prevalence, up to 60% to 78% higher, previous study found no association between perceived insecurity of physical inactivity (Table 3). The opposite pattern was observed and leisure time physical activity.27 Too much traffic was indeed among males, those with 7 to 11 years of education, in the order of a predictor of up to a quarter of higher levels of transport-related 22% to 26% lower. In terms of study sites, and compared with highly physical inactivity. urban Lima, participants living at both Puno sites had around 75% Given that every country should act well beyond its national to 50% lower prevalence of physical inactivity, whereas semiurban aggregates, our findings suggest that efforts should prioritize Tumbes had 186% higher prevalence of physical inactivity. improving physical activity initiatives in those areas with poorer Watching TV ≥2 hours per day was more frequent among those activity profiles. Taken together, these findings signal toward a with more years of education, those with better socioeconomic need for a better profiling of physical inactivity in Peru, including

JPAH Vol. 13, No. 6, 2016 Table 1 Physical Inactivity Levels According to Characteristics of the Study Population Physical inactivity rates Leisure timea Transport-relatedb Sedentarismc Variables n/N (%) P-value* n/N (%) P-value* n/N (%) P-value* Sex Female 1782/1849 (96.4) <0.001 184/1849 (10.0) 0.15 725/1849 (39.2) 0.001 Male 1582/1740 (90.9) 149/1740 (8.5) 775/1740 (44.5) Age 35–44 years 800/857 (93.4) 0.03 71/857 (8.3) <0.001 428/857 (49.9) <0.001 45–54 years 853/921 (92.6) 64/921 (7.0) 399/921 (43.3) 55–64 years 852/914 (93.2) 63/914 (6.9) 370/914 (40.5) 65+ years 859/897 (95.8) 134/897 (14.9) 302/897 (33.7) Years of education 6 years or less 1582/1640 (96.5) <0.001 199/1640 (12.1) <0.001 544/1640 (33.2) <0.001 7–11 years 1083/1152 (94.0) 80/1152 (6.9) 547/1152 (47.5) 12+ years 700/798 (87.7) 53/798 (6.6) 407/798 (51.0) Socioeconomic status (tertiles) Lowest 1192/1225 (97.3) <0.001 104/1225 (8.5) 0.07 407/1225 (33.2) <0.001 Middle 1099/1182 (93.0) 100/1182 (8.5) 509/1182 (43.1) Highest 1077/1186 (90.8) 129/1186 (10.9) 584/1186 (49.2) Study site Lima (highly urbanized) 1021/1103 (92.6) <0.001 78/1103 (7.1) <0.001 541/1103 (49.1) <0.001 Urban Puno (urbanized) 694/760 (91.3) 26/760 (3.4) 336/760 (44.2) Tumbes (semiurban) 6980/1032 (95.0) 217/1032 (21.0) 526/1032 (51.0) Rural Puno (rural) 673/698 (96.4) 12/698 (1.7) 97/698 (13.9) Daily smoking No 3262/3483 (93.7) 0.26 320/3483 (9.2) 0.33 2048/3483 (41.2) <0.001 Yes 105/109 (96.3) 13/109 (11.9) 65/109 (59.6) Heavy alcohol drinking No 3190/3396 (93.9) 0.04 309/3396 (9.1) 0.003 1382/3396 (40.7) <0.001 Yes 178/197 (90.4) 24/197 (12.2) 118/197 (59.9) Fruits and vegetables intake < 5 portions per day 3227/3434 (94.0) 0.01 323/3434 (9.4) 0.20 1428/3434 (41.6) 0.29 Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 5+ portions per day 139/157 (88.5) 10/157 (6.4) 72/157 (45.9) Body mass index Normal (< 25 kg/m2) 891/948 (94.0) 0.01 91/948 (9.6) 0.009 322/948 (34.0) <0.001 Overweight (25 to <30 kg/m2) 1292/1402 (92.2) 121/1402 (8.6) 626/1402 (44.7) Obese (≥30 kg/m2) 826/867 (95.3) 109/867 (12.6) 422/867 (48.7) Hypertension No 2417/2596 (93.1) 0.04 208/2596 (8.0) <0.001 1092/2596 (42.1) 0.14 Yes 608/638 (95.3) 113/638 (17.7) 289/638 (45.3) Type 2 diabetes No 2716/2900 (93.7) 0.70 277/2900 (9.6) <0.001 1221/2900 (42.1) 0.005 Yes 205/220 (93.2) 41/220 (18.6) 114/220 (51.8) * Comparisons were performed by using Chi squared test. a Doing none or very little physical activity (ie, <150 MET-min/week) during leisure time. b Not reporting walking or cycling trips in the transport domain. c Daily TV watching for over 2 hours a day.

JPAH Vol. 13, No. 6, 2016 657 658 Miranda et al

Table 2 Physical Inactivity Levels According to Built Environment Variables Physical inactivity rates Leisure time Transport-related Sedentarism Variables n/N (%) P-value* n/N (%) P-value* n/N (%) P-value* Too much traffic where you live Totally disagree/disagree 2570/2729 (94.2) 0.07 229/2729 (8.4) 0.001 1092/2729 (40.0) <0.001 Agree/totally agree 797/862 (92.5) 104/862 (12.1) 407/862 (47.2) Too much traffic close to where you live Totally disagree/disagree 2199/2331 (94.3) 0.05 204/2331 (8.8) 0.14 915/2331 (39.3) <0.001 Agree/totally agree 1168/1260 (92.7) 129/1260 (10.2) 584/1260 (46.4) Crosswalks help feel safe Totally disagree/disagree 2905/3089 (94.0) 0.11 303/3089 (9.8) 0.006 1275/3089 (41.3) 0.22 Agree/totally agree 461/500 (92.2) 30/500 (6.0) 221/500 (44.2) Streets well lit at night Totally disagree/disagree 1225/1288 (95.1) 0.01 84/1288 (6.5) 0.006 885/1288 (68.7) 0.006 Agree/totally agree 2142/2303 (93.0) 248/2303 (10.8) 1477/2303 (64.1) Unsafe walking during the day Totally disagree/disagree 2556/2713 (94.2) 0.05 274/2713 (10.1) 0.003 1713/2713 (63.1) <0.001 Agree/totally agree 809/876 (92.4) 59/876 (6.7) 646/876 (73.7) Unsafe walking at night Totally disagree/disagree 1764/1856 (95.0) 0.001 208/1856 (11.2) <0.001 1092/1856 (58.8) <0.001 Agree/totally agree 1600/1732 (92.4) 124/1732 (7.2) 1267/1732 (73.2) * Comparisons were performed by using Chi squared test. the expansion of our study to Amazonian jungle urban and rural 57% to 79% (they defined inactive if a subject never engages in areas. This characterization of patterns, across a variety of fac- moderate or vigorous leisure time physical activity for as long as tors and settings can serve to inform how best to design adequate 10 minutes at once).33,34 interventions to improve upon leisure time physical activity and to Our results of transport-related physical inactivity are within maximize the gains of low rates observed in the case of transport- the ranges observed in major Latin American cities. For example, related physical inactivity. Becerra et al reported that 4.3% of adults in Bogota cycle, while 40.5% walk as a mean of transportation. In Curitiba, subjects showed Comparison With Literature similar high rates of walking (55%) relative to cycling (8%).35 Our results were not conclusive regarding the effect of environ- Physical activity has been associated with various health benefits, mental characteristics and physical activity, with the exception of both for youngsters and adults.28 Conversely, sedentarism has traffic as a predictor of higher transport-related physical inactivity.

Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 been associated with higher mortality rates: all-cause mortality Our observation of insecurity to walk at night being associated with and mortality derived from noncommunicable diseases such as lower prevalence of leisure time physical inactivity requires further cardiovascular disease, cancer and diabetes.29 Despite this evidence, careful evaluation in other studies. Unlike our findings, other studies there is a low prevalence of leisure time physical activity in Peru. performed in Latin American countries have reported an associa- Although most adults (75.5%) claim to undertake some kind of tion between feeling safe during the day and leisure time physical leisure time physical activity, 72% do it for less than 150 minutes activity.36 Nevertheless, the association between crime-related safety per week,30 well below the recommendations of the World Health and physical activity is complex because different types of crime, Organization: healthy adults should do a minimum of 150 minutes timing and context, emotional responses, and coping strategies.37 of moderate-intensity aerobic physical activity during a given For example, if the place where participant lives is unsafe, subjects week, or 75 minutes of vigorous-intensity aerobic physical activity could explore other locations to perform physical activity. In addi- in a given week, or an equivalent combination of moderate- and tion, other confounders, such as residential density, have not been vigorous-intensity activity.13 considered in our models and might affect our results. Thus, further The scenario in Latin America is rather similar with low studies in the Peruvian context are needed to improve our under- levels of physical activity, presumably signaling higher levels of standing of the impact of environment on physical activity levels. physical inactivity. In Brazil, 13% of adults reported engaging in 31 30 minutes of leisure time physical activity at least once a week. Relevance to Public Health A wrist accelerometer study in Pelotas, Brazil, revealed that overall moderate/high physical activity, when defined as a >100 mg accel- Acting upon insufficient physical activity is 1 of the 9 global eration magnitude, was below 1 hour per day for all age groups.32 targets for the prevention and control of noncommunicable dis- In Colombia, leisure time levels of physical inactivity ranges from eases.11 Our findings signal to different patterns of leisure time

JPAH Vol. 13, No. 6, 2016 Physical Inactivity in Peru 659

Table 3 Variables Associated With Physical Inactivity: Crude and Adjusted Models Leisure time Transport-related Sedentarism Bivariable Multivariable* Bivariable Multivariable* Bivariable Multivariable* Variables PR (95%CI) PR (95%CI) PR (95%CI) PR (95%CI) PR (95%CI) PR (95%CI) Sex Female 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Male 0.85 (0.82–0.88) 0.95 (0.94–0.97) 0.86 (0.70–1.06) 0.78 (0.63–0.97) 1.14 (1.05–1.23) 1.02 (0.94–1.10) Age 35–44 years 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 45–54 years 1.00 (0.96–1.05) 0.99 (0.97–1.02) 0.84 (0.61–1.16) 0.82 (0.60–1.13) 0.87 (0.79–0.96) 0.88 (0.79–0.97) 55–64 years 1.03 (0.99–1.08) 0.99 (0.96–1.01) 0.83 (0.60–1.15) 0.80 (0.58–1.12) 0.81 (0.73–0.90) 0.86 (0.78–0.95) 65+ years 1.07 (1.03–1.12) 1.00 (0.98–1.03) 1.80 (1.37–2.37) 1.78 (1.32–2.39) 0.67 (0.60–0.76) 0.80 (0.71–0.90) Years of education 6 years or less 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 7–11 years 0.92 (0.89–0.95) 1.00 (0.98–1.01) 0.57 (0.45–0.73) 0.74 (0.57–0.96) 1.43 (1.31–1.57) 1.19 (1.08–1.31) 12+ years 0.81 (0.77–0.84) 0.94 (0.91–0.97) 0.55 (0.41–0.73) 0.83 (0.59–1.18) 1.54 (1.40–1.69) 1.16 (1.03–1.30) Socioeconomic status (tertiles) Lowest 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Middle 0.93 (0.90–0.96) 0.97 (0.95–0.99) 1.00 (0.77–1.30) 1.12 (0.87–1.44) 1.30 (1.17–1.44) 1.24 (1.12–1.37) Highest 0.86 (0.83–0.90) 0.96 (0.94–0.98) 1.28 (1.01–1.64) 1.60 (1.25–2.06) 1.48 (1.34–1.64) 1.36 (1.23–1.50) Study site Lima (highly urbanized) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Urban Puno (urbanized) 0.98 (0.94–1.03) 1.01 (0.98–1.04) 0.48 (0.31–0.75) 0.48 (0.31–0.78) 0.90 (0.82–0.99) 0.87 (0.78–0.97) Tumbes (semiurban) 1.07 (1.03–1.12) 1.02 (1.00–1.04) 2.97 (2.33–3.80) 2.86 (2.23–3.65) 1.04 (0.95–1.13) 1.05 (0.97–1.15) Rural Puno (rural) 1.16 (1.11–1.20) 1.03 (1.01–1.05) 0.24 (0.13–0.44) 0.24 (0.13–0.43) 0.28 (0.23–0.34) 0.30 (0.25–0.36) Current smoking No 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Yes 0.95 (0.86–1.04) 1.05 (1.01–1.10) 1.30 (0.77–2.19) 1.02 (0.61–1.71) 1.45 (1.23–1.70) 1.14 (0.97–1.34) Heavy alcohol drinking No 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Yes 0.93 (0.86–1.00) 0.99 (0.95–1.04) 1.34 (0.91–1.98) 1.74 (1.17–2.57) 1.47 (1.30–1.66) 1.28 (1.13–1.45) Fruits and vegetables

Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 intake < 5 portions per day 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 5+ portions per day 0.87 (0.79–0.96) 0.99 (0.96–1.01) 0.68 (0.37–1.24) 0.84 (0.43–1.65) 1.10 (0.93–1.32) 1.02 (0.85–1.22) * Adjusted for all the variables listed.

and transport-related physical inactivity by study site, which in Strengths and Limitations turn informs toward the need to establish a nationwide surveillance and monitoring system of physical inactivity across Peru’s Physical activity is usually classified under 4 domains: occupational, major geographical regions. Our findings also suggest that certain domestic, transportation and leisure time physical activity, and our areas are in greater need of, and may benefit from, more aggressive population-based study only addressed the latter 2 of them. Our implementation of physical activity initiatives. decision to focus on 2 domains, leisure time and transport-related Moreover, physical activity has also been shown to improve activity, was based on available evidence specific for our Latin global health-related quality of life indicators (eg, self-esteem, American context,18 as the occupational and domestic domains tend emotional well-being, sexuality, etc).38 Higher physical activity to be over-reported by the IPAQ. Despite this restriction, we were would translate into less disability-adjusted life years (DALYs), as able to identify clear patterns of differences between the leisure time 2010 physical inactivity and dietary risk factors were responsible and transport domains. It could be argued that in rural areas there for 1 in 10 DALYs worldwide.19 These gains can well extend into are lower levels of work-related physical inactivity, and we have not population-based improvements in quality of life if physical activity addressed this profile. However, only a small part of our population indicators were to be increased. was rural (rural Puno). The other study sites are highly urbanized or

JPAH Vol. 13, No. 6, 2016 PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Multivariable* 1.03 (0.95–1.12) 1.00 (0.92–1.08) 0.97 (0.86–1.08) 1.01 (0.92–1.11) 1.02 (0.93–1.12) 1.04 (0.95–1.13) Sedentarism Bivariable PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1.18 (1.08–1.28) 1.18 (1.09–1.28) 1.07 (0.96–1.19) 1.63 (1.48–1.79) 1.11 (1.01–1.20) 1.10 (1.02–1.19) PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Multivariable* 1.24 (1.01–1.54) 1.15 (0.93–1.42) 1.18 (0.80–1.74) 0.83 (0.65–1.05) 0.88 (0.65–1.18) 1.07 (0.83–1.37) Transport-related Bivariable PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1.44 (1.16–1.79) 1.17 (0.95–1.44) 0.61 (0.43–0.88) 1.65 (1.30–2.09) 0.67 (0.51–0.87) 0.64 (0.52–0.79) PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) Multivariable* 0.99 (0.97–1.02) 1.00 (0.98–1.02) 1.00 (0.98–1.03) 1.00 (0.97–1.02) 0.99 (0.96–1.01) 0.98 (0.96–1.00) Leisure time Leisure Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 Bivariable PR (95%CI) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 1 (Reference) 0.97 (0.93–1.00) 0.97 (0.94–1.00) 0.92 (0.88–0.97) 0.91 (0.89–0.94) 0.96 (0.92–0.99) 0.93 (0.91–0.96) Built Environment and Physical Inactivity Levels: Crude and Adjusted Models Adjusted and Crude Inactivity Levels: and Physical Built Environment Totally disagree/disagree Totally Totally disagree/disagree Totally Totally disagree/disagree Totally Totally disagree/disagree Totally Totally disagree/disagree Totally Totally disagree/disagree Totally Agree/totally agree Agree/totally agree Agree/totally agree Agree/totally agree Agree/totally agree Agree/totally agree Variables much traffic where you live Too much traffic close to where you live Too help feel safe Crosswalks Streets well lit at night during the day Unsafe walking at night Unsafe walking

Table 4 Table socioeconomic status, and study site. age, education level, * Models adjusted by sex,

660 JPAH Vol. 13, No. 6, 2016 Physical Inactivity in Peru 661

are indeed undergoing an active process of urbanization. Therefore, agreed to participate in the study. Special thanks to all field teams for their our results present useful and valid information for physical inac- commitment and hard work, especially to Lilia Cabrera, Rosa Salirrosas, tivity profiling and physical activity promotion for, at least, these Viterbo Aybar, Sergio Mimbela, and David Danz for their leadership in each urban areas. Physical activity presents also 4 dimensions including of the study sites, as well as Marco Varela for data coordination. We would mode, frequency, duration, and intensity.39 Given the constraints in also like to acknowledge different people who comment first versions of epidemiological surveys, most available data are generated through the manuscript, including Michael Pratt, Thiago H. de Sa, Luis F. Gomez, questionnaires, and newer objective measurements and protocols Diana C. Parra, Olga L. Sarmiento, and Pedro C Hallal. will become much more available in the future.18,40 Given the difficulties faced during the recruitment process, we cannot rule out the possibility of selection bias. Although our results are not References representative of the general population (ie, of all Peruvians), they are representative of subjects living in each of the resource-limited 1. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT. settings. Furthermore, data from the Amazon is lacking and should Effect of physical inactivity on major non-communicable diseases be collected in future studies. We also decided to focus on physi- worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838):219–229 doi:10.1016/S0140-6736(12)61031- cal inactivity as this is one of the major global indicators for the 11 9. PubMed prevention and control of noncommunicable disease and would 2. Liesea AD, Ma X, Maahs DM, Trilke JL. Physical activity, sedentary be most useful for policy makers and practitioners. behaviors, physical fitness, and their relation to health outcomes in youth with type 1 and type 2 diabetes: a review of the epidemiologic literature. JSHS. 2013;2(1):21–38. Conclusions 3. Quirk H, Blake H, Tennyson R, Randell TL, Glazebrook C. Physi- In summary, our study showed high levels of inactivity and marked cal activity interventions in children and young people with Type 1 contrasting patterns by study site. Subjects in the 4 settings studied diabetes mellitus: a systematic review with meta-analysis. Diabetic showed that they tend to engage more in transport-related than Medicine. 2014;31(10):1163–1173. doi:10.1111/dme.12531. 4. Janssen I, Leblanc AG. Systematic review of the health benefits of leisure time physical activity, yet patterns differ by rural/urban physical activity and fitness in school-aged children and youth. Int sites studies depending on the inactivity or sedentarism domain J Behav Nutr Phys Act. 2010;7:40 doi:10.1186/1479-5868-7-40. of interest. Certain groups based on socioeconomic status, disease PubMed conditions, and geographical areas showing much higher levels 5. Cornelissen VA, Smart NA. Exercise training for blood pres- leisure time and transport-related physical inactivity than others sure: a systematic review and meta-analysis. J Am Heart Assoc. could point toward potential avenues to deploy interventions for 2013;2(1):e004473 doi:10.1161/JAHA.112.004473. PubMed those who need them the most. This study also intends to gener- 6. Kahn EB, Ramsey LT, Brownson RC, et al. The effectiveness of ate synergies and momentum to expand these assessments into interventions to increase physical activity. A systematic review. Am J a Peruvian countrywide evaluation and monitoring of levels of Prev Med. 2002;22(4, Suppl):73–107. PubMed doi:10.1016/S0749- physical activity, and thus align with global monitoring initiatives. 3797(02)00434-8 7. Reiner M, Niermann C, Jekauc D, Woll A. Long-term health benefits Built environment variables could be gauged to construct tailor- of physical activity—a systematic review of longitudinal studies. BMC made interventions, specific to each setting. Stakeholders may use Public Health. 2013;13:813 doi:10.1186/1471-2458-13-813. PubMed countrywide evaluations to ensure evidence-based interventions 8. Kruk J, Aboul-Enein HY. Physical activity in the prevention of cancer. are applied across populations to mitigate health problems brought Asian Pac J Cancer Prev. 2006;7(1):11–21. PubMed about by physical inactivity and sedentarism. 9. Schmid D, Behrens G, Jochem C, Keimling M, Leitzmann M. Physical activity, diabetes, and risk of thyroid cancer: a systematic review and Acknowledgments meta-analysis. Eur J Epidemiol. 2013;28(12):945–958 doi:10.1007/ s10654-013-9865-0. PubMed The CRONICAS Cohort Study Group consists of Cardiovascular Disease: 10. Dunstan DW, Howard B, Healy GN, Owen N. Too much sitting—a Antonio Bernabé-Ortiz, Juan P. Casas, George Davey Smith, Shah Ebrahim, health hazard. Diabetes Res Clin Pract. 2012;97(3):368–376. PubMed

Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 Héctor H. García, Robert H. Gilman, Luis Huicho, Germán Málaga, J. Jaime doi:10.1016/j.diabres.2012.05.020 Miranda, Víctor M. Montori, Liam Smeeth; Chronic Obstructive Pulmo- 11. World Health Organization. Global action plan for the prevention nary Disease: William Checkley, Gregory B. Diette, Robert H. Gilman, and control of noncommunicable diseases 2013-2020. Geneva: World Health Organization; 2013. Luis Huicho, Fabiola León-Velarde, María Rivera, Robert A. Wise; and 12. World Health Organization. Noncommunicable diseases country Training and Capacity Building: William Checkley, Héctor H. García, profiles. Geneva: World Health Organization; 2014. Robert H. Gilman, J. Jaime Miranda, Katherine Sacksteder. This project has 13. World Health Organization. Global recommendations on physical been funded in whole with Federal funds from the United States National activity for health. Geneva: World Health Organization; 2010. Heart, Lung, and Blood Institute, National Institutes of Health, Department 14. Masterson Creber RM, Smeeth L, Gilman RH, Miranda JJ. Physical of Health and Human Services, under Contract No. HHSN268200900033C. activity and cardiovascular risk factors among rural and urban groups JJM is supported by Fogarty International Centre (R21TW009982), Grand and rural-to-urban migrants in Peru: a cross-sectional study. Rev Challenges Canada (0335-04), International Development Research Center Panam Salud Publica. 2010;28(1):1–8. Canada (106887-001), Inter-American Institute for Global Change Research 15. Medina-Lezama J, Morey-Vargas OL, Zea-Diaz H, et al. Prevalence (IAI CRN3036), Medical Research Council UK (M007405), National of lifestyle-related cardiovascular risk factors in Peru: the PREVEN- CION study. Rev Panam Salud Publica. 2008;24(3):169–179. PubMed Heart, Lung and Blood Institute (U01HL114180), National Institutes of doi:10.1590/S1020-49892008000900003 Mental Health (U19MH098780). WC was further supported by a Pathway 16. Fortunato L, Drusini AG. Socio-demographic, behavioral and func- to Independence Award (R00HL096955) from the National Heart, Lung and tional characteristics of groups of community and institutionalized Blood Institute. LS and ABO (103994/Z/14/Z) are both supported by Well- elderly Quechua Indians of Peru, and their association with nutritional come Trust. The funders had no role in decision to publish, or preparation status. J Cross Cult Gerontol. 2005;20(2):141–157 doi:10.1007/ of the manuscript. The authors are indebted to all participants who kindly s10823-005-9088-2. PubMed

JPAH Vol. 13, No. 6, 2016 662 Miranda et al

17. Miranda JJ, Bernabe-Ortiz A, Smeeth L, Gilman RH, Checkley W. 29. Biswas A, Oh PI, Faulkner GE, et al. Sedentary time and its associa- Addressing geographical variation in the progression of non-com- tion with risk for disease incidence, mortality, and hospitalization municable diseases in Peru: the CRONICAS cohort study protocol. in adults: a systematic review and meta-analysis. Ann Intern Med. BMJ Open. 2012;2(1):e000610 doi:10.1136/bmjopen-2011-000610. 2015;162(2):123–132 doi:10.7326/M14-1651. PubMed PubMed 30. Gelaye B, Revilla L, Lopez T, Sanchez S, Williams MA. Prevalence 18. Hallal PC, Gomez LF, Parra DC, et al. Lessons learned after 10 years of metabolic syndrome and its relationship with leisure time physical of IPAQ use in Brazil and Colombia. J Phys Act Health. 2010;7(Suppl activity among Peruvian adults. Eur J Clin Invest. 2009;39(10):891– 2):S259–S264. PubMed 898 doi:10.1111/j.1365-2362.2009.02191.x. PubMed 19. Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of 31. Monteiro CA, Conde WL, Matsudo SM, Matsudo VR, Bonsenor IM, burden of disease and injury attributable to 67 risk factors and risk Lotufo PA. A descriptive epidemiology of leisure-time physical activ- factor clusters in 21 regions, 1990–2010: a systematic analysis for the ity in Brazil, 1996-1997. Rev Panam Salud Publica. 2003;14(4):246– Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2224– 254. PubMed doi:10.1590/S1020-49892003000900005 2260 doi:10.1016/S0140-6736(12)61766-8. PubMed 32. da Silva IC, van Hees VT, Ramires VV, et al. Physical activity levels 20. Howe LD, Galobardes B, Matijasevich A, et al. Measuring socio- in three Brazilian birth cohorts as assessed with raw triaxial wrist economic position for epidemiological studies in low- and middle- accelerometry. Int J Epidemiol. 2014;43(6):1959–68. doi:10.1093/ income countries: a methods of measurement in epidemiology paper. ije/dyu203 PubMed Int J Epidemiol. 2012;41(3):871–886 doi:10.1093/ije/dys037. PubMed 33. Gamez R, Venegas SA, Yepes H, Ardilla AM, Rincón AT. Cono- 21. Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of cimiento, percepción y nivel de actividad física de los ciudadanos de the Joint National Committee on Prevention, Detection, Evaluation, Santa Fe de Bogotá, 1999. Rev Bras Cien e Mov. 2000;8(4):51–71. and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 34. Gomez LF, Mateus JC, Cabrera G. Leisure-time physical activity 2003;289(19):2560–2572 doi:10.1001/jama.289.19.2560. PubMed among women in a neighbourhood in Bogota, Colombia: prevalence 22. Diagnosis and classification of diabetes mellitus. Diabetes Care. and socio-demographic correlates. Cadernos de saude publica. 2014;37(Suppl 1):S81–S90. doi:10.2337/dc14-S081 PubMed 2004;20(4):1103–9. 23. Cerin E, Saelens BE, Sallis JF, Frank LD. Neighborhood Environ- 35. Becerra JM, Reis RS, Frank LD, et al. Transport and health: a look at ment Walkability Scale: validity and development of a short form. three Latin American cities. Cad Saude Publica. 2013;29(4):654–666. Med Sci Sports Exerc. 2006;38(9):1682–1691 doi:10.1249/01. PubMed doi:10.1590/S0102-311X2013000800004 mss.0000227639.83607.4d. PubMed 36. Arango CM, Paez DC, Reis RS, Brownson RC, Parra DC. Associa- 24. Coutinho LM, Scazufca M, Menezes PR. Methods for estimat- tion between the perceived environment and physical activity among ing prevalence ratios in cross-sectional studies. Rev Saude adults in Latin America: a systematic review. Int J Behav Nutr Phys Publica. 2008;42(6):992–998. PubMed doi:10.1590/S0034- Act. 2013;10:122 doi:10.1186/1479-5868-10-122. PubMed 89102008000600003 37. Ding D, Adams MA, Sallis JF, et al. Perceived neighborhood environ- 25. Komal W, Jaipanesh K, Seemal M. Association of leisure time physical ment and physical activity in 11 countries: do associations differ by activity, watching television, obesity & lipid profile among seden- country? Int J Behav Nutr Phys Act. 2013;10:57 doi:10.1186/1479- tary low-income south Indian population. East Afr J Public Health. 5868-10-57. PubMed 2010;7(3):225–228. PubMed 38. Mishra SI, Scherer RW, Geigle PM, et al. Exercise interventions on 26. Wilcox S, Castro C, King AC, Housemann R, Brownson RC. Deter- health-related quality of life for cancer survivors. Cochrane Database minants of leisure time physical activity in rural compared with urban Syst Rev. 2012;8:CD007566 10.1002/14651858.CD007566.pub2. older and ethnically diverse women in the United States. J Epidemiol PubMed Community Health. 2000;54(9):667–672. PubMed doi:10.1136/ 39. Strath SJ, Kaminsky LA, Ainsworth BE, et al. Guide to the assess- jech.54.9.667 ment of physical activity: Clinical and research applications: a sci- 27. Mendes MdA, da Silva ICM, Hallal PC, Tomasi E. Physical activ- entific statement from the American Heart Association.Circulation . ity and perceived insecurity from crime in adults: a population- 2013;128(20):2259–2279 doi:10.1161/01.cir.0000435708.67487.da. based study. PLoS One. 2014;9(9):e108136 doi:10.1371/journal. PubMed pone.0108136. PubMed 40. Hallal PC, Andersen LB, Bull FC, Guthold R, Haskell W, Ekelund 28. C3 Collaborating for Health. The benefits of physical activity for U. Global physical activity levels: surveillance progress, pitfalls, and health and well-being. 2011. http://www.c3health.org/wp-content/ prospects. Lancet. 2012;380(9838):247–257 doi:10.1016/S0140- uploads/2009/09/C3-review-of-physical-activity-and-health- 6736(12)60646-1. PubMed

Downloaded by Washington Univ In St Louis on 09/16/16, Volume 13, Article Number 6 v-1-20110603.pdf. Accessed June 17, 2015.

JPAH Vol. 13, No. 6, 2016