Active commuting associations with BMI and self-rated health: a cross-sectional analysis of the Healthy Ireland survey

The Healthy Ireland survey uses a representative sample of the population in the Republic of Ireland and is conducted annually to collect information on health and lifestyle behaviours of the Irish population. Multistage probability sampling was used in the Healthy Ireland survey to ensure a representative distribution of sampling points around the Republic of Ireland. In brief, 686 sampling points were created using electoral divisions stratified by region and demographic characteristics with 21 of the divisions containing more than one sampling point due to larger size (Department of Health 2020). The current cross-sectional study population was derived from Wave 3 of the Healthy Ireland survey 2017 which included 7487 in-person interviews (Department of Health 2017). The sample was restricted to participants who worked or attended education (n = 4038). Ethics approval for the parent Healthy Ireland survey was granted by the Research Board at the Royal College of Physicians in Ireland. Ethics exemption for the current study was granted by the Research Ethics Committee, School of Public Health, Physiotherapy and Sports Science, University College Dublin.

Specially trained professionals visited the selected households and conducted face-to-face interviews with study participants who completed a questionnaire using computer-assisted personal interviewing to gather information related to lifestyle and health behaviours. In the present study values were recoded as missing data and excluded in cases where participants responded “Do not know” or who refused to answer a question.

SRH was determined by asking participants to rate their health in general with the following question – “How is your health in general?”. Responses included – very good, good, fair, bad and very bad. SRH was classified as high (responses were very good or good) or low (responses were fair, bad or very bad). Participants (78%) completed a health assessment involving objective measurement of BMI and waist circumference. BMI was calculated as (weight (kg)/height (m)²). Standard cut-off points for BMI were used to define underweight (<18.5 kg/m²), normal weight (18.5–24.9 kg/m²), overweight (25–29.9 kg/m²), obesity (≥30 kg/m²) (WHO 2021). In total a response rate of 60.4% was achieved for the Healthy Ireland 2017 Wave 3 survey and data was weighted to minimise bias in the dataset generated by non-responses (Department of Health 2020). Regarding commuting mode participants were asked the following – “How do you usually travel from home to your regular place of work or college?”, selecting only one mode from the following: travel on foot, travel by bicycle, travel by public transport, travel by car or motorbike. Active travel was defined as travel by bicycle or on foot. Non-active travel included travel by car or motorbike or public transport.

Potential confounders and relevant covariates were identified from literature. These were: gender, age (15 to 24 years, 25 to 44 years, 45 to 64 years, 65 years and older), education (primary/post primary or third level), socioeconomic classification using the National Statistics Socio-Economic classification (NS-SEC) (Office of National Statistics 2021), distance from work (less than 3 km travelled or 3 km and longer), residential location (urban/rural), current smoking status (yes/no), presence of a chronic health condition limiting activities of daily living (yes/no), daily fruit consumption (yes/no), daily vegetable consumption (yes/no) and daily consumption of unhealthy food in the form of cakes, muffins, or biscuits (2 categories yes/no).

Comparative demographic and lifestyle characteristics of active and inactive commuters are presented in Table 1. Active commuters were more likely to be younger (68.6% were less than 45 years of age), have lower educational attainment and socioeconomic status, have a healthy BMI, live in an urban location and have a shorter commuting distance compared to the inactive commuters. Inactive commuters were more likely to be non-smokers and consume vegetables on a daily basis compared to active commuters. No differences were detected in self-rated health or in the percentage with a chronic health condition limiting activities of daily living between commuting groups.

Binary logistic regression analyses were performed to investigate the relationships between active commuting modes and overweight and obesity (Table 2). Those who used any form of active commuting had decreased odds of being obese compared to non-active commuters (AOR 0.69, 95% CI 0.52, 0.90). When active commuter modes were analysed separately, cycling was associated with further decreased odds of obesity relative to those who did not cycle (AOR 0.23, 95% CI 0.09, 0.56). Walking to work or place of education was not associated with obesity (AOR 0.85, 95% CI 0.64, 1.13) relative to those who did not commute on foot. There was no association identified between active commuting (overall or by individual modes) and odds of being overweight.

Binary logistic regression analyses were performed to investigate the relationship between active commuting modes and high SRH with adjustment for covariates (Table 3). There were no statistically significant associations between SRH and active commuting overall or when cycling and walking were analysed separately.

When commuters were stratified according to the distance they travelled to work or place of education (i.e. less than 3 km and 3 km or greater using active means) (Table 4) active transport with a journey of at least 3 km was associated with decreased odds of obesity (AOR 0.54, 95% CI 0.30, 0.98) relative to non-active commuters. No association with obesity was observed for a commute distance of less than 3 km nor was active travel associated with overweight or self-rated health at either journey distance.

In this cross-sectional analysis of a representative sample of the population in the Republic of Ireland we investigated associations between active commuting and weight status and self-rated health. The main findings suggest that active commuting is inversely associated with obesity, with further reduced obesity risk observed among cyclists and those who actively commuted ≥3 km. No associations were observed between active commuting (overall or individual modes) with overweight or SRH.

Our finding, that active commuting is associated with lower likelihood of obesity, is consistent with findings from other studies (Gordon-Larsen et al. 2009; Lindström 2008). A systematic review by Brown and colleagues suggested small but important benefits of active travel for obesity management (Brown et al. 2017). The results from the current study support the proposed framework that active travel use results in higher energy expenditure than inactive travel modes which may result in healthy weight maintenance provided no significant increase in energy intake occurs in active travel users (Brown et al. 2017). When cycling and walking were investigated separately, cycling had a strong inverse association with obesity. This was in concordance with findings from other studies (Laverty et al. 2013; Millett et al. 2013; Mytton et al. 2016a; Nordengen et al. 2019a; Rissel et al. 2014; Wen and Rissel 2008). Results from another systematic review indicate that cyclists have lower BMI and decreased cardiovascular disease risk factors than non-cyclists (Nordengen et al. 2019b). Contrary to other studies which reported associations between active travel on foot and reduced obesity risk (Laverty et al. 2013; Millett et al. 2013; Rissel et al. 2014), we did not find any association between walking and obesity. Of note we used objective BMI measurements whereas some of the previous studies used self-reported BMI data which may affect the accuracy of results (Laverty et al. 2013; Rissel et al. 2014). However, our findings are plausible as cycling is often considered a more intense activity. Oja and colleagues reported that commuting by bicycle was associated with higher cardiorespiratory and metabolic improvements than walking in a 10-week RCT of sedentary commuters (Oja et al. 1991).

There was no association detected between active commute modes and overweight. This is inconsistent with findings from other cross-sectional studies (Barengo et al. 2006; Hu et al. 2002; Mendoza et al. 2011), all of which used objective BMI measurements. Eriksson and colleagues reported lower CVD risk for overweight participants who actively commuted (Eriksson et al. 2020). A possible explanation for the lack of an association between active commuting and overweight in the current study may be that active travel is generally of lower intensity than leisure time physical activity while regular higher intensity physical activity may be needed for prevention of overweight.

Higher SRH is hypothesised to be associated with active travel use, given the established links between physical activity and better health-related quality of life (Bize et al. 2007). No association was found between high SRH and active commute modes. This contrasts with findings from a study of 1196 commuters by Norgenen and colleagues (Nordengen et al. 2019a). It should be noted that in that study 40% of the study population cycled and 88% reported high SRH, whereas in the current study 3% of participants cycled and 92% reported high SRH, which may account for the lack of an association with active commuting. In a longitudinal study by Mytton and colleagues, higher physical well-being was associated with maintenance of cycling commuting in Cambridge, UK, using the Physical Component Summary (PCS-8), which includes an SRH component (Mytton et al. 2016b). It has been suggested that those who participate in healthy lifestyle behaviours can be more negative about their health (Layes et al. 2012). Therefore, it is important to appreciate the complex nature of SRH as a representation of overall health as it can be influenced by many underlying social and individual factors. Residual confounding factors such as job strain and sleep may also affect the association between SRH and active commuting in this study.

Interestingly we report a stronger inverse association with obesity in those who actively travelled at least 3 km which may suggest a dose-response relationship. This is consistent with previous studies which indicated similar exposure-response findings (Laverty et al. 2013; Martin et al. 2015). Those engaging in longer active journeys are more likely to achieve the 30-minute recommended physical activity threshold which is linked to a decreased disease risk (Department of Health and Children, Health Service Executive 2009).

Active travel specifically to work or education accounts for 27% of all trips in Ireland with the largest proportion of all trips taken during the week in Ireland (National Transport Authority 2017). In the present study almost 16% of study participants aged 15 years and older used active commuting. Although this represents an increase from 13% in 2016 (Central Statistics Office 2016) and is similar to that reported in the UK (Laverty et al. 2013; Flint et al. 2014) it is lower than in other European countries, for example in Sweden 38% report actively commuting (Eriksson et al. 2020). These differences may represent different ethnicities, cultural habits and infrastructure availability to the populations in question. Car use is more dominant in Ireland, England and the USA while government policies in other countries such as the Netherlands are more car restrictive in favour of active transport promotion (Pucher and Buehler 2010).

To our knowledge this is the first study in Ireland to investigate associations between active commuting and BMI and SRH. Additionally, our study which uses data from a nationally representative study of the Irish population, contributes to the evidence base regarding active commuting associations with SRH as a single item measure. Interviewers were professionally trained for face-to-face interviews reducing the risk of information bias and error. Objectively measured height and weight data was used for BMI measurement which is more reliable than self-reported data (Maukonen et al. 2018). Information on many potential confounders was collected.

In Ireland only 33% of adults and 12% of adolescents meet the WHO weekly physical activity recommendations (European Commission WHO 2018). Therefore, active travel use is an important lifestyle factor which has the potential to improve health by incorporating physical activity into daily commuting, re-enforcing the benefits of making small behavioural changes. This study, which demonstrated an inverse association between active commuting, particularly among cyclists and those actively commuting ≥3 km, represents a significant contribution to the evidence base regarding active commuting associations with obesity. The findings support the need for further incentives to actively travel. Further research should consider additional confounding factors in this association, in particular environmental factors such as traffic congestion, green space surroundings and public transport availability.

However, there are some limitations that should be noted. The cross-sectional study design, which precludes drawing conclusions regarding the temporal direction of relationships, limits inference with respect to causality. Reverse causality is possible; that is that participants who were obese may have been less likely to actively commute. Additionally, self-reported data for active commute use was collected during a face-to-face interview which may be subject to social desirability bias (Adams et al. 2005). Self-selection bias is possible in this study. Healthier people are more inclined to actively commute and have healthier lifestyle habits. Additional data on physical activity levels of participants would help to delineate if the inverse association between active commuting and BMI is related specifically to active travel and not to overall fitness in daily life. This was an Anonymised Microdata File version of the original 2017 Healthy Ireland dataset. Data analysis was limited as a result by the categorised nature of all data which may affect accuracy of results.