Abstract(resumo):

Objective: To investigate the individual and combined associations of obesity and physical inactivity with costs of primary heathcare in mid-age adults from Brazil. Methods: A representative cohort of 620 mid-age (median 63.9y) users of healthcare services in Brazil was followed from 2010 to 2014. Body mass index (BMI) and physical activity were assessed in 2010, and costs of all primary care services from 2010 to 2014 were retrieved from medical records. Quantile regression was used to estimate associations between obesity and inactivity in 2010 and health costs from 2010 to 2014. Results: In 2010 median annual costs were 19.4% higher in women ($29.5) than men ($24.7; p=0.02). The costs increased over 4 years by around 80% and it was higher in participants with BMI?30 (84%) (than in those with BMI<30; 67%) and markedly higher in those who were inactive (77%) (than in those who were active; 32%). In adjusted analyses, compared with the obese/inactive group, costs were lower in the obese/active group (?= -40.83[-110.63 a 40.18]), and lower in the non-obese/active group (?= -66.03 [-122.85 a -9.21]) over this 4 years period. Conclusion: These real-world data show that increases in health care costs among people in their sixties could be attenuated by both obesity prevention and physical activity promotion.

Keywords(palavra-chave):

health care costs, exercise, obesity.

Content(conteúdo):

Introduction
Obesity and physical inactivity are major global concerns, which are associated with numerous health and social problems worldwide1,2. Estimates from the World Health Organization suggest that globally, in 2016, 13% of adults were obese, and 27.5% were physically inactive3,4. Health concerns attributable to obesity and physical inactivity are particularly concerning in Brazil, which has the sixth largest population in the world (projected to reach 250 million people in 2050) and which faces substantial social challenges due to increases in population ageing5.
The prevalence of obesity in Brazilian adults (?20 years-old) increased from 11.9% to 17.5% between 2006 and 20136 and in 2018 was estimed ito be 19.8%7. In recent years there has been some improvement in inactivity levels in Brazilian adults, from 69.7% in 2009 to 62.3% in 20168. However, among adults aged 50-59 years, the prevalence of insufficient physical activity in 2016 was 84.6%, making older Brazilians 55% more likely to be inactive than their younger counterparts (20-29 years-old)9.
The increasing prevalence of obesity is creating a significant economic burden in terms of health care costs in both developed and developing countries across the globe. For example, recent estimates suggest that annual healthcare expenditure is US$1,029 higher per person with obesity type 1 than in a normal weight person10. Similarly, using standardized methods and data from 142 countries, Ding et al.11 have shown that physical inactivity cost health care systems nearly $54 billion worldwide in 2013. In Brazil, it is estimated that that costs attributable to obesity were US$1.1 billion per year from 2008 to 2010 and that the costs of inactivity were of the order of US$797 million in 201311.
Few researchers have examined the combined effects of obesity and inactivity on health care costs. Our earlier cross-sectional research with Australian women found that inactivity and high BMI are both associated with higher health care costs, but costs were lower for overweight active women than for healthy-weight sedentary women 12. However, a lack of prospective studies limits our understanding of the real impact of obesity and physical inactivity on health costs, especially in rapidly developing nations like Brazil. Our previous research in one industrial region of Brazil has shown that local health costs increased significantly from 2010 to 2014. This is not surprising, as the numbers of older people in the community was also rapidly increasing at that time. However, costs appeared to be mitigated by leisure time sports participation13.
Given the paucity of data on the effects of obesity and inactivity on health costs, the aim of this study was to investigate the individual and combined associations of obesity and physical inactivity with costs of primary heath care over four years in a cohort of mid-age adults from a regional industrial city in Brazil.

Material and methods
This is a cohort study of mid-age adults living in the city of Bauru, a medium-size city (~344,000 inhabitants in 2010), located in the most industrialized Brazilian region (State of Sao Paulo). In terms of services offered to the population, the Brazilian National Health Service [Sistema Unico de Saude (SUS)] is divided into primary, secondary and tertiary levels14. This research focused on the costs of primary care services, which are delivered to the population in small-to-medium medical facilities called Basic Health care Units (BHU).
In terms of geography, each BHU covers all the people living in the neighborhood around it, offering non-complex medical services, such as medical consultations, vaccinations and medicine prescriptions. Bauru has 17 BHU. The criteria used to selected the BHUs was the biggest BHUs (higher number of registered patients) in each geographical region [north, south, east, west and central]).
In the city of Bauru, around 60% of all 344,000 citizens exclusively use the SUS (~206,400 users)15. The five selected BHUs had 114,386 registered patients, representing approximately 55% of all users in this city. A simple size calculation was performad and estimated, based on the percentage of Brazilian population that is attended exclusively by the Brazilian public health care system (60 %)15 and using as parameters a 3.8 % error (arbitrary because there are no other similar studies), 5 % significance (z = 1.96) and design effect of 50 % (using cluster sampling including BHU), a final sample size of at least 958 participants (minimum of 192 in each BHU). After fieldwork, carried out during 2010, the final sample was composed of 970 adults of both genders.
These participants were contacted by phone to invite their participation and to check if they met the following inclusion criteria: i) aged ? 50 years; ii) were registered for at least one year with BHU; iii) had at least one medical consultation in the past six months.
In terms of sampling, 4,209 patients attended one of the 5 BHUs seletcted for the study, in the last 6 months, 970 were randomly selected (194 in each BHU) using a statistical software (Statistical Pachage for Social Sciences- SPSS – version 13.0 for Windows).
Interviews were conducted with participants between August and December in 2010 (face-to-face), 2012 (telephone) and 2014 (telephone). From 2010 to 2014 there were 59 deaths and it was not possible to contact 291 participants after three attempts (Figure 1). The final sample comprised 620 participants with no missing data. All participants gave informed consent during the three interview. The study was approved by the Ethics Committee of Sao Paulo Stata University (number: 06834912.3.0000.5423). More details about the recruitment process have been described elsewhere16-18.
*** Figure 1 ***
Health care costs of primary care services are paid by the SUS through the local authority (Municipal Department of Health), which is responsible for the management of health expenses. Medical services delivered at BHUs and information about number and type of medical appointments, tests and medication prescribed in 2010, 2011, 2012, 2013 and 2014 were retrieved from participants’ medical record. The local authority provided the cost of all health care services (medical consultations, medicines, and exams). Prices were expressed in US dollars (US$) and adjusted using inflation rates. Calculations followed standard methods, as described in previous studies16-19.
Participants were invited to visit their BHU for the conduct of anthropometric measures and interviews in 2010. Body mass (kg) and height (m) were measured following a standardized protocol. An electronic scale (maximum weight 150kg) and a wall-mounted stadiometer were used to assess body mass and height, respectively. Body mass index (BMI) was estimated as body mass divided by height squared (kg/m2). Obesity was defined as BMI ?30.0 kg/m2.
Physical activity was measured in 2010 using the Brazilian version of the Baecke questionnaire20. The questionnaire comprises 16 questions, grouped in 3 domains (occupational, leisure-time and transportation). For this study, we only considered information about physical activity in leisure-time. Three self-reported constructs of leisure-time physical activity were analyzed: i) intensity (low, moderate, vigorous), ii) weekly time of practice, (< 1h/week; 1 - 2h/week; 2 - 3h/week; 3 - 4h/week; > 4h/week) and iii) activity history (< 1 month; 1 - 3 months; 4 - 6 months; 7 - 9 months; > 9 months). Participants were categorized as physically active if they reported either a minimum of 180 minutes per week (3-4 h/week) of moderate and vigorous activitiesor a minimum of 60 minutes per week (1-2h/week) of vigorous activities over the last four months (4-6 months)21,22.
Median and interquartile values were used to describe sample characteristics, which were compared for men and women, using Mann-Whitney tests. Categorical characteristics were described using frequencies and proportions, and comparedusing chi-square tests. Comparisons of costs in each of five years (2010, 2011, 2012, 2013 and 2014) were performed using Friedman’s test. Comparisons of costs between gruops, according to sex, BMI and leisure time PA were performed using Mann-Whitney tests. A combined variable was created to categorize participants into one of four BMI/physical activity groups as follows: Non-Obese/Active (n=66), Obese/Active (n=33), Non-Obese/Inactive (n=291) or Obese/Inactive (n 230). Comparisons of the median costs in each of the BMI and activity groups were examined using Mann-Whitney and Kruskal-Wallis tests. Quantile regression was performed to compare costs in the combined obesity and physical activity groups, after adjusting for sex and age. Statistical significance (p-value) was set at <0.05 and the statistical software Stata (version 16) was used to conduct all statistical tests.
Results
Of the 970 participants included in the study in 2010, 59 died, and of the remainder 68.1% (95%CI: 65.1 to 71.1) were retained in 2014. (See Figure 1). At baseline, participants who were lost to follow up (n= 350) were similar to those who were followed at three time points (n= 620) in terms of age (p-value= 0.727), health care costs (p-value= 0.132), physical activity (p-value=0.076), obesity (p-value= 0.126) and sex (p-value= 0.961).
Baseline demographic and BMI/activity data, for the whole sample, and for men and women, are shown in Table 1. The analysis sample of 620 included more women (73.2% [95%CI: 69.7 to 76.7) than men (26.8% [95%CI: 23.3 to 30.2]). Median age was 63.9 years and median BMI was 28.8 kg.m-2. (See Table 1). Both obesity and inactivity were more common in women than in men. When categorized into four groups, more women (42.5% [95%CI:37.9 to 47.1]) than men (22.9% [95%CI:16.5 to 29.3]) were obese/inactive, while more men (56% [95%CI: 48.5 to 63.6]) then women (44.5% [95%CI: 39.9 to 49.1]) were non-obese/inactive. Few participants of either sex were obese and active. The prevalence of obesity was 42.4% [95%CI: 38.5 to 46.3] in 2010 and remained stable to 2014 (41.8% [95%CI: 38.1 to 45.8]; p= 0.724). Similarly, the prevalence of inactivity was 84% [95%CI: 81.1 to 86.9] at baseline and remained largly unchanged in 2014 (82.4% [95%CI: 79.4 to 85.4]; p=0.447). (See Table 1)
*** Table 1 ***
Median health costs (US$ per person) in each year, for all participants, and by sex, BMI and physical activity category, are shown in Table 2. Total health care costs, in each year, were higher in women than men. Costs over the 4-year period, for all participants were US$ 212.14 (121.44-369.72); these were higher in women (US$ 220.45 (132.99-372.04)) (73% of the sample) than in men (US$ 165.94 (102.49-359.25)) (p=0.028).
In 2010, costs were higher in participants with obesity than their non-obese counterparts, but similar in inactive and active participants (See Table 2). Over 4 years, health care costs increased significantly in both the obese and non-obese groups (by 84% and 67% respectively) Similarly, health care costs increased by 77% from 2010 to 2014 in the inactive participants, but this increase was much smaller in those who were active (32%).
*** Table 2 ***
When these costs were considered for the four groups of combined obesity and physical activity levels, there was very little difference in costs in 2010. (See Figure 2). Costs then increased in all four groups, but the rate of increase was much lower in the non-obese, active group and higher in obese-inactive group.By the end of the follow-up (2014) costs were highest in the obese-inactive group, similar in the obese-active and non-obese-inactive groups, and lowest in those who were non-obese-active.
*** Figure 2 ***
Median health care costs from 2010 to 2014 are compared according to physical inactivity and obesity in Table 3. Costs were 26.8% higher in obese than in non-obese participants (p= 0.002), and 33% higher in the inactive than in the active participants (p=0.019) Quantile regression models (adjusted for age and sex) showed that costs were US$48 lower in the non-obese (than in those with obesity), and $35 lower in the active than the inactive. The latter difference was not statistically significant, because of the low numbers and wider variation in the active group.
The combined costs of obesity and physical inactivity are shown in the lower half of Table 3. In the combined models, costs were lowest in the non-obese/active group and highest in obese/inactive group. Compared with per person costs in that group (obese/inactive), after adjustment for age and sex, median costs were $40 lower in participants who were non-obese/inactive and $66 lower in the non-obese/active participants.
*** Table 3 ***

Discussion
The aim was to investigate the individual and combined associations of obesity and physical inactivity with primary health care costs from 2010 to 2014 in a population based cohort of mid-age Brazilian adults. As expected, the median annual costs of health services were lower for mid-age adults who reported being physically active and those with BMI<30 than in those who were inactive or obese. Both obesity and physical inactivity had a significant impact on health care costs. Our study is the first to present data on the combined impact of obesity and physical inactivity on health costs in mid-age Brazilian adults.
Annual health costs increased by more than 80% from 2010 to 2014. This increase is explained by the fact that health care costs are strongly affected by aging and the development of non-communicable chronic diseases, with concomitant higher use of medicines, which increases markedly from the age of 6523. Because people with obesity are more likely to have these health problems, it was not surprising that costs were higher in 2010 for those with obesity than in those with BMI<30. Differences in costs for the active and inactive participants were not significant in 2010. However, cost increases over the next 4 years were notably less marked in those with favourable obesity/activity profiles, and as marked in the inactive participants as in those who were obese. When obesity and physical activity were considered together, it was clear that costs increased in all the obesity/activity groups, but the increase was notably attenuated in the non-obese/active participants. Hence, by 2014, costs were highest in the obese/inactive group, similar in the obese/active and non-obese/inactive groups, and lowest in the non-obese/active group.
In our study the prevalence of obesity remained stable from 2010, but was much higher than observed in Brazilian adults aged 55-64 years-old , in whom there was an increase from 2006 [17.6%] to 2012 [24.4%])24. The differences in terms of prevalence might be explained, at least in part, by methodological differences. For example, the national surveys included randomly selected participants, and used self-reported data, while our study considered patients from community health services, whose BMI was directly measured Our data suggest that the costs of primary care services could be reduced by US$48 per participant over four years (~US$12 per person per year) if obesity could be prevented. Although these potential costs savings are much smaller than those reported for secondary and tertiary services25, with such a high prevalence of obesity, and a population of 344,000 in Bauru, the savings for the local health system would be substantial.
The finding that health care costs were lower in active (than in inactive) mid-age adults has been observed in previous studies26-28. Our estimate of a potential US$35 saving (16%) in active participants is similar to that reported in studies of American adults where the saving attributed to physical activity ranged from 15% to 22.8%26. Although these savings for active participants may reflect the lower occurrence of obesity, metabolic and cardiovascular diseases among people who are active [28], it should be acknowledged that this group may also incur greater health care costs from sports and recreation injuries, which might offset some of the benefits29. An important finding of our study was the potential saving in health costs which could hypothetically be achieved by elimination of both physical inactivity and obesity (US$ -66.03 [-122.85 to -9.21] per participant in 4 years). Adopting a population attributable fraction approach, Ding et al.11 have estimated that the direct health care costs (primary, secondary and tertiary levels) attributed to physical inactivity in Brazil in the calendar year of 2013 reached US$ 1.6 billion.
Brazil is home for 39,007,220 adults aged ?50 years old of whom ~22% are obese (8,581,588 adults)6 and 82% are insufficiently active in leisure time (32,063,935 adults)30. Using the data from our study, we estimate that the potential mitigation of primary care costs attributed to obesity (~US$ -12 per person per year) would be US$ ~103 million/year, while for physical inactivity the impact of elimination in the population (~US$ -8.80 per person per year) would be US$ ~282 million/year. The elimination of both obesity and physical inactivity would reduce costs by around US$ 385 million/year, for primary care services alone. Even in a median size city like Bauru, with a population of 79,645 people aged >50, we estimate that elimination of obesity and inactivity would reduce costs by around US$ 785,000/year, for primary care services alone.
Some limitations of this study should be considered. Physical activity was based on self-report, which might introduce recall bias. Moreover, only primary care services were considered, and it is plausible to assume that all economic forecasts generated in this study are underestimated, assuming that relevant health outcomes related to obesity and physical inactivity, such as hospitalizations and surgeries, were not considered. There may have been residual confounding due to unmeasured variables. Strengths of this study include the prospective design. This is important because most studies of relationships between physical inactivity/obesity and health care costs have been cross-sectional in design, and economic forecasts have been based on theoretical models. In our longitudinal survey, the costs were directly assessed, providing a real perspective about the actual costs incurred by these participants.
Conclusion
In summary, our findings illustrate the significant impact of obesity and physical inactivity on primary health care costs in Brazil, illustrating the relevant economic loss attributed to the combination of both. These findings highlight the importance of physical activity promotion and obesity prevention efforts for older adults.

References
1. le Roux CW, Hartvig NV, Haase CL, Nordsborg RB, Olsen AH, Satylganova A. Obesity, cardiovascular risk and healthcare resource utilization in the UK. Eur J Prev Cardiol. 2021 Sep 20;28(11):1235-1241. doi: 10.1177/2047487320925639.
2. Katzmarzyk PT, Friedenreich C, Shiroma EJ, Lee IM. Physical inactivity and non-communicable disease burden in low-income, middle-income and high-income countries. Br J Sports Med. 2022 Jan;56(2):101-106. doi: 10.1136/bjsports-2020-103640.
3. World Health Organization. Obesity and overweight. WHO https://www.who.int/mediacentre/factsheets/ fs311/en/ (2016).
4. Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufficient physical activity from 2001 to 2016: a pooled analysis of 358 population-based surveys with 1·9 million participants. Lancet Glob Health. 2018 Oct;6(10):e1077-e1086. doi: 10.1016/S2214-109X(18)30357-7.
5. Lutz W, K C S. Dimensions of global population projections: what do we know about future population trends and structures? Philos Trans R Soc Lond B Biol Sci. 2010 Sep 27;365(1554):2779-91. doi: 10.1098/rstb.2010.0133.
6. Malta DC, Santos MA, Andrade SS, Oliveira TP, Stopa SR, de Oliveira MM, Jaime P. Time trend in adult obesity indicators in Brazilian state capitals, 2006-2013. Cien Saude Colet. 2016 Apr;21(4):1061-9. Portuguese. doi: 10.1590/1413-81232015214.12292015.
7. Brasil. Ministério da Saúde. Secretaria de Vigilância em Saúde. Departamento de Análise em Saúde e Vigilância de Doenças não Transmissíveis. Vigitel Brasil 2018: vigilância de fatores de risco e proteção para doenças crônicas por inquérito telefônico : estimativas sobre frequência e distribuição sociodemográfica de fatores de risco e proteção para doenças crônicas nas capitais dos 26 estados brasileiros e no Distrito Federal em 2018 / Ministério da Saúde, Secretaria de Vigilância em Saúde, Departamento de Análise em Saúde e Vigilância de Doenças não Transmissíveis. – Brasília: Ministério da Saúde, 2019.
8. Cruz MSD, Bernal RTI, Claro RM. Trends in leisure-time physical activity in Brazilian adults (2006-2016). Cad Saude Publica. 2018 Oct 22;34(10):e00114817. Portuguese. doi: 10.1590/0102-311X00114817.
9. Ide PH, Martins MSAS, Segri NJ. Trends in different domains of physical activity in Brazilian adults: data from the Vigitel survey, 2006-2016. Cad Saude Publica. 2020;36(8):e00142919. Portuguese. doi: 10.1590/0102-311x00142919.
10. Van den Broek-Altenburg E, Atherly A, Holladay E. Changes in healthcare spending attributable to obesity and overweight: payer- and service-specific estimates. BMC Public Health. 2022 May 13;22(1):962. doi: 10.1186/s12889-022-13176-y.
11. Ding D, Lawson KD, Kolbe-Alexander TL, Finkelstein EA, Katzmarzyk PT, van Mechelen W, Pratt M; Lancet Physical Activity Series 2 Executive Committee. The economic burden of physical inactivity: a global analysis of major non-communicable diseases. Lancet. 2016 Sep 24;388(10051):1311-24. doi: 10.1016/S0140-6736(16)30383-X.
12. Brown WJ, Hockey R, Dobson AJ. Physical activity, Body Mass Index and health care costs in mid-age Australian women. Aust N Z J Public Health. 2008 Apr;32(2):150-5. doi: 10.1111/j.1753-6405.2008.00192.x.
13. Codogno JS, Monteiro HL, Turi-Lynch BC, Fernandes RA, Pokhrel S, Anokye N. Sports Participation and Health Care Costs in Older Adults Aged 50 Years or Older. J Aging Phys Act. 2020 Feb 12:1-7. doi: 10.1123/japa.2019-0219.
14. Paim J, Travassos C, Almeida C, Bahia L, Macinko J. The Brazilian health system: history, advances, and challenges. Lancet. 2011 May 21;377(9779):1778-97. doi: 10.1016/S0140-6736(11)60054-8
15. Kilsztajn S, Silva DF, Camara MB, Ferreira VS. Grau de cobertura dos planos de saúde e distribuição regional do gasto público em saúde. Saude e Sociedade. 2001; 10:35-45.
16. Codogno JS, Turi BC, Kemper HC, Fernandes RA, Christofaro DG, Monteiro HL. Physical inactivity of adults and 1-year health care expenditures in Brazil. Int J Public Health. 2015 Mar;60(3):309-16. doi: 10.1007/s00038-015-0657-z.
17. Turi BC, Codogno JS, Fernandes RA, Sui X, Lavie CJ, Blair SN, Monteiro HL. Accumulation of Domain-Specific Physical Inactivity and Presence of Hypertension in Brazilian Public Healthcare System. J Phys Act Health. 2015 Nov;12(11):1508-12. doi: 10.1123/jpah.2014-0368.
18. Turi BC, Codogno JS, Fernandes RA, Sui X, Lavie CJ, Blair SN, Monteiro HL. Association of Different Physical Activity Domains on All-Cause Mortality in Adults Participating in Primary Care in the Brazilian National Health System: 4-Year Follow-up. J Phys Act Health. 2017 Jan;14(1):45-51. doi: 10.1123/jpah.2016-0067.
19. Codogno JS, Fernandes RA, Sarti FM, Freitas Júnior IF, Monteiro HL. The burden of physical activity on type 2 diabetes public healthcare expenditures among adults: a retrospective study. BMC Public Health. 2011 May 4;11:275. doi: 10.1186/1471-2458-11-275.
20. Codogno JS, Fernandes RA, Sarti FM, Freitas Júnior IF, Monteiro HL. The burden of physical activity on type 2 diabetes public healthcare expenditures among adults: a retrospective study. BMC Public Health. 2011 May 4;11:275. doi: 10.1186/1471-2458-11-275.
21. Fernandes RA, Christofaro DGD, Casonatto J, Codogno JS, Rodrigues EQ, Cardoso M, et al. Prevalência de dislipidemia em indivíduos fisicamente ativos durante a infância, adolescência e idade adulta. Arq Bras Cardiol 2011; 97: 317-2
22. Fernandes RA, Zanesco A. Early physical activity promotes lower prevalence of chronic diseases in adults. Hypertens Res 2010; 33: 926-31.
23. Narayan SW, Tordoff JM, Nishtala PS. Temporal trends in the utilisation of preventive medicines by older people: A 9-year population-based study. Arch Gerontol Geriatr. 2016 Jan-Feb;62:103-11. doi: 10.1016/j.archger.2015.10.007.
24. Malta DC, Andrade SC, Claro RM, Bernal RT, Monteiro CA. Trends in prevalence of overweight and obesity in adults in 26 Brazilian state capitals and the Federal District from 2006 to 2012. Rev Bras Epidemiol. 2014;17 Suppl 1:267-76. English, Portuguese. doi: 10.1590/1809-4503201400050021.
25. Krishna SG, Rawal V, Durkin C, Modi RM, Hinton A, Cruz-Monserrate Z, Conwell DL, Hussan H. Weight Loss Surgery Reduces Healthcare Resource Utilization and All-Cause Inpatient Mortality in Morbid Obesity: a Propensity-Matched Analysis. Obes Surg. 2018 Oct;28(10):3213-3220. doi: 10.1007/s11695-018-3345-2.
26. Coughlan D, Saint-Maurice PF, Carlson SA, Fulton J, Matthews CE. Leisure time physical activity throughout adulthood is associated with lower medicare costs: evidence from the linked NIH-AARP diet and health study cohort. BMJ Open Sport Exerc Med. 2021 Mar 5;7(1):e001038. doi: 10.1136/bmjsem-2021-001038.
27. Gomes GAO, Brown WJ, Codogno JS, Mielke GI. Twelve year trajectories of physical activity and health costs in mid-age Australian women. Int J Behav Nutr Phys Act. 2020 Aug 10;17(1):101. doi: 10.1186/s12966-020-01006-6.
28. Popkin BM, Kim S, Rusev ER, Du S, Zizza C. Measuring the full economic costs of diet, physical activity and obesity-related chronic diseases. Obes Rev. 2006 Aug;7(3):271-93. doi: 10.1111/j.1467-789X.2006.00230.x.
29. van Mechelen W. The severity of sports injuries. Sports Med. 1997 Sep;24(3):176-80. doi: 10.2165/00007256-199724030-00006.
30. da Silva ICM, Mielke GI, Bertoldi AD, Arrais PSD, Luiza VL, Mengue SS, Hallal PC. Overall and Leisure-Time Physical Activity Among Brazilian Adults: National Survey Based on the Global Physical Activity Questionnaire. J Phys Act Health. 2018 Mar 1;15(3):212-218. doi: 10.1123/jpah.2017-0262.