- Research article
- Open Access
- Open Peer Review
Body mass index and overweight in relation to residence distance and population density: experience from the Northern Finland birth cohort 1966
© Näyhä et al.; licensee BioMed Central Ltd. 2013
- Received: 6 April 2013
- Accepted: 17 September 2013
- Published: 8 October 2013
The effect of urban sprawl on body weight in Finland is not well known. To provide more information, we examined whether body mass index (BMI) and the prevalence of overweight are associated with an individual’s distance to the local community centre and population density in his/her resident area.
The sample consisted of 5363 men and women, members of the Northern Finland Birth Cohort 1966 (NFBC), who filled in a postal questionnaire and attended a medical checkup in 1997, at the age of 31 years. Body mass index (BMI; kg/m2) and the prevalence of overweight (BMI ≥ 25.0 kg/m2) were regressed on each subject’s road distance to the resident commune’s centre and on population density in the 1 km2 geographical grid in which he/she resided, using a generalized additive model. Adjustments were made for sex, marital status, occupational class, education, leisure-time and occupational physical activity, alcohol consumption and smoking.
The mean BMI among the subjects was 24.7 kg/m2, but it increased by increasing road distance (by 1.3 kg/m2 from 5–10 to 20–184 km) and by decreasing population density (by 1.7 kg/m2 from 1000–19,192 to 1–5 inhabitants/km2). The respective increases in overweight (overall prevalence 41%) were 13 per cent units for distance and 14 per cent units for population density. Adjusted regressions based on continuous explanatory variables showed an inverse L-shaped pattern with a mean BMI of 24.6 kg/m2 at distances shorter than 5 km and a rise of 2.6 kg/m2 at longer distances, and an increase of 2.5 kg/m2 from highest to lowest population density. The associations with road distance were stronger for women than men, while the sex difference in association with population density remained indeterminate.
We conclude that young adults in Northern Finland who live far away from local centres or in the most sparsely populated areas are fatter than those who live close to local centres or in densely populated areas. The likely explanations include variations in everyday physical activity in different residential environments, although causality of the associations remains to be confirmed.
- Body mass index
- Medical geography
- Population density
It is frequently observed that body weight is associated with the physical living environment. This is mostly described in conjunction with urban sprawl characterized by inadequate walkability of streets and roads, dependence on private cars and poor accessibility to well-equipped food stores [1–3], which would lead to a positive energy balance and increased body weight. Some authors report that the effects of urban sprawl on body mass index (BMI) in county-level comparisons can be as great as 1 kg/m2, which would markedly increase the prevalence of overweight and the risk of diabetes and cardiovascular diseases. The main underlying factor is considered to be everyday physical activity, which shows wide variations across countries. In the Netherlands, for example, 48% of trips are made by foot or bicycle, but only 10% in the USA . From the public health point of view, the amount of physical activity undertaken in daily routines is more important than that used in sports and leisure-time physical exercise .
In Finland, studies on this topic are few, but they have revealed variations in body weight and obesity among administrative areas [5–7], community types [8, 9] and 10-square-kilometre geographical grids . The variations are attributed mostly to socioeconomic factors . However, the physical community structures have greatly changed in Finland in recent decades due to in-migration to towns , and this may have affected the amount of physical activity needed in normal daily activities, with consequent changes in body weight. Therefore, any associations found between the physical community structures and body weight would be useful, especially as such associations are modifiable by community planning  and could provide tools for preventing obesity. This is particularly relevant in Finland, where the population is sparse and the distances are long.
The present study, based on a population-based cohort from northern Finland, tests the assumption that body weight and the prevalence of overweight depend on the physical structure of individuals’ residential environment defined by (1) road distance to the local community centre and (2) population density in the 1-square-kilometre geographical grid in which he or she resides. The former depicts the distance which the person would travel while doing his or her daily errands, and the latter not only population density itself but also the type of residence in the urban–rural scale . The underlying rationale is that due to better connectivity, people walk and cycle more in densely populated than in sparsely populated areas .
The area studied
Finland is a subarctic country located between 60 and 70° N latitude and 20 and 31° E longitude. The population density in 1997 was 17 inhabitants per square kilometre, and the total land area is 338,145 km2, of which only 30% is inhabited. The country has experienced marked depopulation of the countryside in recent decades, with a consequent increase of the urban population and its density and a thinning of the rural population. This has caused sprawling of urban areas, with a simultaneous decrease of population density and an expansion of suburbs and build-up areas. The thinning of the rural population is expected to continue in the future .
The Northern Finland Birth Cohort (NFBC) data
The subjects were members of NFBC 1966, which consists of all individuals in the two northernmost provinces of Finland (Oulu and Lapland) whose mothers’ expected time of delivery was in 1966. The total number of births was 12,231 (12,058 live births), which covers 96.3% of all births in the area in 1966. The cohort have been followed up on since birth; the present analysis is based on a survey conducted in 1997, when the subjects were 31 years of age. A postal questionnaire was sent to all those 11,541 subjects who were still alive and whose addresses were known, and 8767 (76%) returned it. The respondents who lived in the provinces of Lapland and Oulu or in the Helsinki metropolitan area, altogether 6759 subjects, were invited to participate in a medical examination. Of those, 5713 (85%) attended, and 5690 filled in a separate questionnaire inquiring about working conditions, including a question on physical activity at work. The study was approved by the Ethical Committee of the Northern Ostrobothnia Hospital District.
The road network data
To calculate each subject’s road distance to the local community centre, we used the Finnish road network data available in the Digiroad database, which is a national road and street database developed by the Finnish transport agency (see http://www.digiroad.fi/en_GB/). The database includes accurate geometry and length of road segments, which enables a calculation of actual distances between any two points along the road network. We used the earliest available version (2003) of the database. The road network in 2003 was essentially similar to that in 1997, as only a minimal number of new roads were built during that period; there were only renovations of the old road network [14, 15].
Coordinates of place of residence
Coordinates of the participants’ home address on 1 January 1997 were obtained from the Finnish Population Register Centre. Based on these coordinates, each cohort member living in the provinces of Oulu or Lapland or in the Helsinki metropolitan area and on whom the relevant health and work information was available, was attached to the 1-square-kilometre grid cell in which he/she resided using ArcGIS. A detailed description of how the grids were attached to the map is found elsewhere (http://www.stat.fi/meta/kas/yhtenaiskoordin_en.html). The record linkage failed for 327 subjects due to unavailability of the coordinates, errors in the datasets (incorrect coordinates in either data, coordinates that failed to match any inhabited grid or were on the boundary of two inhabited grids and therefore were ambiguous) and discrepant information on the resident commune in the cohort and grid data. This left 5363 subjects in the final study population.
In the medical examination, body height and weight (to an accuracy of 0.1 cm and 0.1 kg, respectively) were measured and converted to body mass index (BMI; kg/m2). For a small proportion of the subjects (3.5%), body height and weight were based on self-reports in the questionnaire. Overweight was defined as BMI ≥ 25.0 kg/m2. BMI (a continuous variable) and overweight (yes/no) were used as outcomes.
The shortest road distance between the subjects’ home and the midpoint of the densest populated 1-square-kilometre grid in his/her resident commune was used as the first explanatory variable and was calculated from the road network database. This was done using ArcGIS, by adding together the lengths of the individual road segments. The roads included regional and local main streets, collector streets, feeder streets and private streets, and in the costal archipelago, ferry connections between these streets. In the data analysis, distance was used as a continuous variable (in kilometres) but was also classified to arbitrary intervals (0–1.9 km, 2.0–4.9 km, 5.0–9.9 km, 10–19.9 km and 20–184 km) for descriptive purposes.
Population density in each 1-square-kilometre grid, obtained from Statistics Finland, was attached to each cohort member residing in this grid and was used as the second explanatory variable. Population density was analyzed as a continuous variable but was also classified to form residential area types: scattered settlements; rural areas proper; transitional zones; built-up areas & suburbs and high-rise centres having population densities of 1–5, 6–20, 21–100, 101–1000 and more than 1000 inhabitants per square kilometre, respectively .
Potential confounding factors
Several suspected confounders were controlled for in the analysis. Firstly, we asked about the frequency of leisure-time physical activity (response options: once a month or less often; 2–3 times a month; once a week; 2–3 times a week; 4–6 times a week; daily) and the duration of physical activity at a time (not at all; less than 20 minutes; 20–39 minutes; 40–59 minutes; 1–1.5 hours; more than 1.5 hours), separately regarding light physical activity (no breathlessness or sweating ) and brisk physical activity (at least some breathlessness and sweating). The responses were converted to metabolic equivalents (MET) and MET hours per week. In the calculations, an intensity value of 3 METs was used for light physical activity and 5 METs for brisk physical activity . Secondly, occupational physical activity was elicited by a question classifying the subjects as having light sedentary work, other sedentary work, light standing or moving work, medium heavy moving work, heavy manual work and very heavy manual work. In the analysis, the two lowest and two highest groups were merged (light sedentary/other sedentary work and heavy/very heavy manual work, respectively). The detailed questions have been described elsewhere [17, 18].
Thirdly, the subjects were requested to indicate the frequency of their habitual consumption of rye or crisp bread, fresh vegetables, roots or salads, fruit or berries and sausages during the past six months (response options: less frequently than once a month or not at all; 1–2 times a month; once a week; a couple of times per week; almost every day; once a day or more often). An unhealthy diet was defined as one containing sausages daily or almost every day, and consumption of the rest of the food items a couple of times a week or less frequently .
Smoking was classified according to those smoking 5–7 days a week, 4 days a week or fewer and not smoking at all. The questionnaire also asked about the frequency of consumption of beer, wine and spirits during the last year, and the usual amount consumed per drinking occasion. The daily amount of alcohol consumed was calculated using the following alcohol contents (vol %): beer: 4.8; light wine: 5.0; table wine: 14.5; and spirits: 37.0. The subjects were classified into quartiles of alcohol consumption (grams per day). The method has been validated against 7-day food records .
In addition to sex and marital status, demographic factors allowed for in the analysis were based on self-reports and included socioeconomic group (entrepreneurs; higher administrative employees; lower administrative employees; blue collar workers; other) , occupational class (office work; industry; agriculture/forestry; health & social work; business)  and education (university; college or polytechnic; vocational school or course; no vocational school).
We first drew the histograms of residence distance and population density which we smoothed by Gaussian kernel density function using smoothing windows of 0.5 km and 150 inhabitants/km2, respectively. Then BMI and the prevalence of overweight were regressed on logged distance and population density using a generalized additive model in which smoothing was achieved by cubic splines with 4 degrees of freedom. We used the Gaussian error distribution for BMI and binomial for overweight, and the identity link function in both cases. The results were expressed as smoothed predictions of BMI (kg/m2) and the prevalence of overweight together with their 95% confidence bands. The method has the advantage of retaining the continuity of the explanatory variables without assuming any regular shape of the relationship. Confounders were added to the models depending on whether they caused any marked change in the smoothed predictions. The intra-class correlations of BMI and overweight within the resident communes were close to zero (< 0.01). The calculations were performed using the R software, release 2.15.0 .
Distribution of the subjects according to major demographic, work-related and lifestyle factors, road distance from resident commune’s centre and population density of the subject’s resident grid
Demographic, work-related and lifestyle factors
No. (%) of subjects
Road distance (km)
Population density (inhabitants/km2)
Region of residence
City of Oulu
Rest of northern Finland
Helsinki metropolitan area
Blue collar workers
No vocational school
Occupational physical activity
Light standing or moving work
Medium heavy moving work
Heavy/very heavy manual work
Leisure-time physical activity e (quintiles, MET-hours per week)
Alcohol consumption (quartiles, grams per day)
On 5–7 days/week
On 4 days/week at mostg
Not at all
The median distance from home to the resident commune’s centre was 4.1 km, but it varied among subgroups (Table 1). Thus the subjects living in Oulu or elsewhere in northern Finland resided closer (median distance 4.9 km and 3.3 km, respectively) than those living in the Helsinki metropolitan area (7.3 km). The longest distances in Oulu and Helsinki were only 25 km and 26 km, respectively, but could reach 184 km in the countryside. People engaged in agriculture lived farther (median distance 11.2 km) than those engaged in other occupations (3.7–4.1 km), and entrepreneurs lived farther (7.4 km) than people belonging to other socioeconomic groups (3.7–4.2 km). Table 1 also shows variations in residence distance depending on marital status, education and lifestyle factors, but these were minor or inconsistent.
The median population density in the subjects’ resident grid was 712 inhabitants/km2. The subjects living in Helsinki, those who were not married, those who were engaged in occupations other than agriculture, those who were higher administrative employees and those who had university education lived in relatively dense grids (Table 1). The median population density increased with increasing leisure-time physical activity and decreasing occupational physical activity. Population density also increased with growing alcohol consumption, and smokers and subjects who had a healthy diet lived in relatively dense grids.
Crude associations of BMI and overweight with residence distance, population density and potential confounders
Crude associations of body mass index (BMI) and overweight with road distance, residential area type and demographic and lifestyle factors
No. (%) overweight (BMI ≥ 25)
No. of subjects
Road distance to commune’s centre (km)
Residential area type (population density, inhabitants per km 2 )
Scattered settlements (1–5)
Rural areas proper (6–20)
Transition zones (21–100)
Built-up areas & suburbs (101–1000)
High-rise centres (1000–19192)
Region of residence
City of Oulu
Rest of northern Finland
Helsinki metropolitan area
Blue collar workers
No vocational school
Occupational physical activity
Light standing or moving work
Medium heavy moving work
Heavy/very heavy manual work
Leisure-time physical activity (quintiles, MET-hours per week)
Alcohol consumption (quartiles, grams per day)
On 5–7 days/week
On 4 days/week at most
Not at all
Table 2 also shows higher BMI and higher prevalence of overweight in men than women, in subjects living in northern Finland versus those living in Helsinki, in entrepreneurs compared with other socioeconomic groups and in people having lower education versus higher education. The subjects working in agriculture and industry were fattest, and those in the health sector were leanest. The BMI and the prevalence of overweight decreased with increasing leisure-time physical activity, while the subjects doing heavy physical work had the highest BMI and those doing light physical work the lowest. Unhealthy diet was associated with a slightly higher BMI and more overweight than healthy diet, and smokers were fatter than non-smokers. Subjects in the highest alcohol consumption quartile had higher BMI and more overweight compared with those in other quartiles.
Analyses allowing for confounders and curvilinear associations
A breakdown by sex in (Additional file 2: Figure S2) indicated greater distance-related increases in BMI beyond 5 km among women than men, in both unadjusted (women 3.5, men 1.9 kg/m2) and adjusted (women 3.0, men 2.1 kg/m2) analyses, and a corresponding sex difference in overweight (unadjusted increase 37 versus 20 per cent units in women and men, respectively; adjusted increase 40 versus 18 per cent units, respectively). However, the wider confidence intervals and less precise regression estimates left the associations more equivocal than in the unstratified analysis.
BMI and overweight also increased with decreasing population density, even though the relation was less curvilinear than that for residence distance (Figure 3). The unadjusted model-estimated increase over the whole population density scale was 2.8 kg/m2 for BMI and 28 per cent units for overweight, and these diminished to respectively 2.5 kg/m2 and 20 per cent units when allowance was made for the confounders mentioned above. The sex differences were less consistent than those regarding distance-related sex differences (Additional file 2: Figure S2).
We found that among young Finnish adults, BMI and the prevalence of overweight increased with increasing road distance to the local commune centre, but only at distances longer than approximately 5 kilometres, and they increased from densely to sparsely populated areas. This is the first attempt in Finland to explain the geographical variation in body weight [6–10] in terms of residence distance and grid-based population density. The variations of BMI and overweight were large, and in case the associations can be interpreted causally, the findings provide opportunities for prevention of overweight by health-based planning of the built environment.
The strength of our study is the large population-based cohort in well-defined geographical areas of Finland, where the distances and population density vary widely. The geographical grids on which the information on population density was based were smaller than those used previously  and were small enough to reflect influences of individuals’ close environment. The response rate was satisfactory, and any marked selection at different phases of the study is unlikely. We were able to use road distance as an explanatory factor, which is likely to serve as a reasonable estimate for the length of actual distances people go for their daily trips—although we did no individual measurements nor did we measure the duration of the trips.
One limitation of the study is that we had no information on the actual length of work travels or on the mode of travelling to work. Therefore, we cannot differentiate between the effects of leisure time physical activity and those related to commuting. Some confounding may have been caused by higher unemployment rates and consequently less work-related physical activity in sparsely populated areas, which was not specifically controlled for. We did control for socioeconomic group, in which unemployed persons are included in the class “other”, but this did not markedly change the parameter estimates and this factor was not included in the final model.
The questions on leisure-time and occupational physical activity have been previously used and have adequate face validity [17, 18]. We similarly assume that the composite variable used to measure the healthiness of diet is valid because it predicts abdominal obesity , even though calorie intake was not measured. We cannot rule out a possibility of residual confounding, but we believe that our findings serve as reasonable estimates for how much residence distance and population density may affect body weight.
Geographical studies of BMI and obesity conducted in Finland have found variations between localities  and larger areas  and a lower BMI in the capital area of Helsinki than in other areas . The sole grid-based analysis of body weight conducted in this country observed a lower prevalence of obesity in cities compared with other areas . Investigations performed elsewhere have either found  or have not found  more obesity in rural than in urban areas. The lower BMI in urban areas has been linked with higher socioeconomic status  and higher population density [2, 25–27], and is interpreted in terms of proximity of daily destinations accessible by walking or bicycle and the consequently greater physical activity. We used road distance as an operational measure for actual distances which people travel in everyday life, and population density to indicate the subject’s community type in urban–rural scale. By a reasonable assumption, both measures are related to the intensity of physical activity and the accessibility of services and amenities, the former on account of travel lengths and mode of moving, the latter through better connectivity with daily destinations.
We observed higher body weight among people living in remote and sparsely populated areas. The association was attributable partly to social, demographic and lifestyle factors including leisure-time and occupational physical activity. As the associations persisted after adjustments, true effects of distance and population density are a possibility. The cross-sectional study design and the possibility of unidentified confounders prevent us from drawing causal conclusions. We also acknowledge a possibility of migration-related bias. A Finnish study concluded that heavier adolescents who reside in the countryside may be less likely to move to towns than normal weight adolescents , but we do not know whether this is true for adults.
Substantive evidence from elsewhere favours the view that individuals’ physical environment modifies the amount of physical activity in daily life. The relevant issues include how far away the daily destinations are located and whether the distances are travelled by foot, bicycle, public transport or private car. More overweight has been reported among people who live far from large supermarkets  or recreational facilities  or who have poor access to them [30, 31], those who live far from bike paths  or have no shops or paths within walking distance  and those who have few non-residential destinations . Proximity of employment establishments, grocery stores and business centres is believed to be linked to lower body weight .
Less overweight has been observed in individuals who have an easy access to healthy food stores  or supermarkets  or have healthy grocery or multiple food options in their neighbourhood , while people living close to fast-food restaurants are fatter . Factors associated with low BMI and less obesity also include easy access to green areas [37, 38] or living in neighbourhoods with much vegetation . Increased walkability of streets would decrease the risk of overweight , each kilometer walked per day decreasing the likelihood of obesity by 5% and each hour spent in a car per day increasing the likelihood of obesity by 6% . People are leaner in areas in which a higher percentage of the population walks to work , while those using a car for work travels or trips to grocery stores are fatter [41, 42].
Even in sparsely populated Finland, urban centres have sprawled in recent decades as a consequence of migration of the country population to towns , but the public health consequences of this with respect to body weight are not well recognized nor taken into account in the national health policy. However, the high-level health and population records in Finland provide excellent opportunities to foresee and control the adverse effects of urban sprawl and rural depopulation. Our results suggest that approximately 40% of young adults live in areas that may adversely affect their body weight because their place of residence is located too far away from local centres or is too sparsely inhabited. We assume that variations in everyday physical activity in different areas explain the findings, but individual-based measurements of daily activity are needed to strengthen this causal argument. The prevalence of overweight attributable to geographical location represents a significant and removable fraction of the public health burden and calls for more studies on the effects of community planning on health . Even though causality needs to be confirmed for the associations reported here, the findings emphasize that community planning is a health issue that should be closely examined not only by city planners, but also by public health scientists.
We wish to acknowledge financial aid from the Academy of Finland.
- Garden FL, Jalaludin BB: Impact of urban sprawl on overweight, obesity, and physical activity in Sydney, Australia. J Urban Health. 2008, 86: 19-30.View ArticlePubMedPubMed CentralGoogle Scholar
- Lovasi GS, Neckerman KM, Quinn JW, Weiss CC, Rundle A: Effect of individual or neighborhood disadvantage on the association between neighborhood walkability and body mass index. Am J Public Health. 2009, 99: 279-284. 10.2105/AJPH.2008.138230.View ArticlePubMedPubMed CentralGoogle Scholar
- McCann BA, Ewing R: Measuring the health effects of sprawl. A national analysis of physical activity, obesity and chronic disease. Smart Growth America. Surface Transportation Policy Project. 2003, Washington DC, http://smartgrowth.umd.edu/assets/ewingmccann_2003.pdf,Google Scholar
- Frumkin H: Urban sprawl and public health. Public Health Rep. 2002, 117: 201-217.View ArticlePubMedPubMed CentralGoogle Scholar
- Heliövaara M, Aromaa A: Height, weight and obesity of Finnish adults. 1980, Helsinki, Finland: Publications of the Social Insurance Institution, Finland, ML:19, Finnish, with English summaryGoogle Scholar
- Näyhä S, Hassi J: Life style, work and health of Finnish reindeer herders. Life style, work and health of Finnish reindeer herders. Helsinki: The Finnish Social Insurance Institution. Edited by: Näyhä S, Hassi J. 1993, Finland: Publications of the Social Insurance Institution, 217-237. ML:127 ]Google Scholar
- Pietinen P, Vartiainen E, Männistö S: Trends in body mass index and obesity among adults in Finland from 1972 to 1992. Int J Obes. 1996, 20: 114-120.Google Scholar
- Fogelholm M, Valve R, Absetz P, Heinonen H, Uutela A, Patja K, Karisto A, Konttinen R, Mäkelä T, Nissinen A, Jallinoja P, Nummela O, Talja M: Rural–urban differences in health and health behaviour: a baseline description of a community health-promotion programme for the elderly. Scand J Public Health. 2006, 34: 632-640. 10.1080/14034940600616039.View ArticlePubMedGoogle Scholar
- Jokela M, Kivimäki M, Elovainio M, Viikari J, Raitakari OT, Keltikangas-Järvinen L: Urban/rural differences in body weight. Evidence for social selection and causation hypothesis in Finland. Soc Sci Med. 2009, 68: 867-875. 10.1016/j.socscimed.2008.12.022.View ArticlePubMedGoogle Scholar
- Lahti-Koski M, Taskinen O, Similä M, Männistö S, Laatikainen T, Knekt P, Valsta LM: Mapping geographical variation in obesity in Finland. Eur J Public Health. 2008, 18: 637-643. 10.1093/eurpub/ckn089.View ArticlePubMedGoogle Scholar
- Rusanen J, Muilu T, Colpaert A, Naukkarinen A: Georeferenced data as a tool for monitoring the concentration of population in Finland in 1970–1998. Fennia, Int J Geography. 2003, 181: 129-144.Google Scholar
- Zick CD, Smith KR, Fan JX, Brown BB, Yamada I, Kowaleski-Jones L: Running to the store? The relationship between neighborhood environments and the risk of obesity. Soc Sci Med. 2009, 69: 1493-1500. 10.1016/j.socscimed.2009.08.032.View ArticlePubMedPubMed CentralGoogle Scholar
- Saelens BE, Sallis JF, Frank LD: Environmental correlates of walking and cycling: Findings from the transportation, urban design, and planning literatures. Ann Behav Med. 2003, 25: 80-91. 10.1207/S15324796ABM2502_03.View ArticlePubMedGoogle Scholar
- Finnish road statistics. http://www2.liikennevirasto.fi/julkaisut/pdf3/lti_2011-06_tietilasto_2010_web.pdf,
- Uimonen S: Measuring the highway capital in Finland 1900–2009. Tampere Economic Working Papers Net Series 81. 2010, Tampere, Finland, http://urn.fi/urn:isbn:978-951-44-8226-7,Google Scholar
- Kaakinen M, Läärä E, Pouta A, Hartikainen AL, Laitinen J, Tammelin TH, Herzig KH, Sovio U, Bennet AJ, Peltonen L, McCarthy MI, Elliot P, DeStavola B, Järvelin MR: Life course analysis of a fat mass and obesity-associated (FTO) gene Variant and body mass index in the Northern Finland birth cohort 1966 using structural equation modelling. Am J Epidemiol. 2010, 172: 653-665. 10.1093/aje/kwq178.View ArticlePubMedPubMed CentralGoogle Scholar
- Hu G, Qiao Q, Silventoinen K, Eriksson JG, Jousilahti P, Lindström J, Valle TT, Nissinen A, Tuomilehto J: Occupational, commuting, and leisure-time physical activity in relation to risk for type 2 diabetes in middle-aged Finnish men and women. Diabetologia. 2003, 46: 322-329.PubMedGoogle Scholar
- Tammelin T, Näyhä S, Rintamäki H, Zitting P: Occupational physical activity is related to physical fitness in young workers. Med Sci Sports Exer. 2002, 34: 158-165. 10.1097/00005768-200201000-00024.View ArticleGoogle Scholar
- Laitinen J, Pietiläinen K, Wadsworth M, Sovio U, Järvelin MR: Predictors of abdominal obesity among 31-y-old men and women born in Northern Finland in 1966. Eur J Clin Nutr. 2004, 58: 180-190. 10.1038/sj.ejcn.1601765.View ArticlePubMedGoogle Scholar
- Statistics Finland: Sosioekonomisen aseman luokitus 1989. 1989, Helsinki, Finland: Käsikirjoja 17, Classification of social status 1989: in FinnishGoogle Scholar
- Statistics Finland: Ammattiluokitus 1987. 1987, Helsinki, Finland: Käsikirjoja 14, Classification of occupations; in FinnishGoogle Scholar
- R Development Core Team: R: A language and environment for statistical computing. Release 2.15.0. 2012, Vienna, Austria: R Foundation for Statistical Computing, http://www.R-project.org, ISBN 3-900051-07-0,Google Scholar
- Reeder BA, Chen Y, Mcdonald SM, Angel A, Sweet L: Regional and urban–rural differences in obesity in Canada. Canadian Heart Health Survey Group. Can Med Assoc J. 1997, 57 (Suppl 1): S10-S16.Google Scholar
- Peytremann-Bridevaux I, Faeh D, Santos-Eggimann B: Prevalence of overweight and obesity in rural and urban settings in 10 European countries. Prev Med. 2007, 44: 442-446. 10.1016/j.ypmed.2006.11.011.View ArticlePubMedGoogle Scholar
- Lee IM, Ewing R, Sesso HD: The built environment and physical activity levels: the Harvard Alumni Health Study. Am J Prev Med. 2009, 37: 293-298. 10.1016/j.amepre.2009.06.007.View ArticlePubMedPubMed CentralGoogle Scholar
- Rundle A, Roux AV, Free LM, Miller D, Neckerman KM, Weiss CC: The urban built environment and obesity in New York City: a multilevel analysis. Am J Health Promot. 2007, 21 (4 Suppl): 326-334.View ArticlePubMedGoogle Scholar
- Zhao Z, Kaestner R: Effects of urban sprawl on obesity. J Health Econ. 2010, 29: 779-787. 10.1016/j.jhealeco.2010.07.006.View ArticlePubMedGoogle Scholar
- Liu GC, Wilson JS, Qi R, Ying J: Green neighborhoods, food retail and childhood overweight: differences by population density. Am J Health Promot. 2007, 21 (4 Suppl): 317-325.View ArticlePubMedGoogle Scholar
- Boehmer TK, Lovegreen SL, Haire-Joshu D, Brownson RC: What constitutes an obesogenic environment in rural communities?. Am J Health Promot. 2006, 20: 411-421. 10.4278/0890-1171-20.6.411.View ArticlePubMedGoogle Scholar
- Giles-Corti B, Macintyre S, Clarkson JP, Pikora T, Donovan RJ: Environmental and lifestyle factors associated with overweight and obesity in Perth, Australia. Am J Health Promot. 2003, 18: 93-102. 10.4278/0890-1171-18.1.93.View ArticlePubMedGoogle Scholar
- Gordon-Larsen P, Nelson MC, Page P, Popkin BM: Inequality in the built environment underlies key health disparities in physical activity and obesity. Pediatrics. 2006, 117: 417-424. 10.1542/peds.2005-0058.View ArticlePubMedGoogle Scholar
- Petrella RJ, Kennedy E, Overend TJ: Geographic determinants of healthy lifestyle change in a community-based exercise prescription delivered in family practice. Environ Health Insights. 2008, 1: 51-62.PubMedPubMed CentralGoogle Scholar
- Yamada I, Brown BB, Smith KR, Zick CD, Kowaleski-Jones L, Fan JX: Mixed land use and obesity: an empirical comparison of alternative land use measures and geographic scales. Prof Geogr. 2012, 64: 157-177. 10.1080/00330124.2011.583592.View ArticlePubMedPubMed CentralGoogle Scholar
- Rundle A, Neckerman KM, Freeman L, Lovasi GS, Purciel M, Quinn J, Richards C, Sircar N, Weiss C: Neighborhood food environment and walkability predict obesity in New York City. Environ Health Perspect. 2009, 117: 442-447.View ArticlePubMedGoogle Scholar
- Larson N, Story MT, Nelsin MC: Neighborhood environments: disparities in access to healthy foods in the U.S. Am J Prev Med. 2009, 36: 74-81. 10.1016/j.amepre.2008.09.025.View ArticlePubMedGoogle Scholar
- Oreskovic NM, Winickoff JP, Kuhlthau KA, Romm D, Perrin JM: Obesity and the built environment among Massachusetts children. Clin Pediatr (Phila). 2009, 48: 904-912. 10.1177/0009922809336073.View ArticleGoogle Scholar
- Bell JF, Wilson JS, Liu GC: Neighborhood greenness and 2-year changes in body mass index of children and youth. Am J Prev Med. 2008, 35: 547-553. 10.1016/j.amepre.2008.07.006.View ArticlePubMedPubMed CentralGoogle Scholar
- Nielsen TS, Hansen KB: Do green areas affect health? Results from a Danish survey on the use of green areas and health indicators. Health Place. 2007, 13: 839-850. 10.1016/j.healthplace.2007.02.001.View ArticlePubMedGoogle Scholar
- Smith KR, Brown BB, Yamada I, Kowaleski-Jones L, Zick CD, Fan JX: Walkability and body mass index. Density, design, and new diversity measures. Am J Prev Med. 2008, 35: 237-244. 10.1016/j.amepre.2008.05.028.View ArticlePubMedGoogle Scholar
- Frank LD, Andresen MA, Schmid TL: Obesity relationships with community design, physical activity, and time spent in cars. Am J Prev Med. 2004, 27: 87-96. 10.1016/j.amepre.2004.04.011.View ArticlePubMedGoogle Scholar
- Lopez-Zetina J, Lee H, Friis R: The link between obesity and the built environment. Evidence from an ecological analysis of obesity and vehicle miles of travel in California. Health Place. 2006, 12: 656-664. 10.1016/j.healthplace.2005.09.001.View ArticlePubMedGoogle Scholar
- Pendola R, Gen S: BMI, auto use, and the urban environment in San Francisco. Health Place. 2007, 13: 551-556. 10.1016/j.healthplace.2006.02.004.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/13/938/prepub
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.