The question of whether the method of transport you use to get to work affects your mental wellbeing is bound to grab attention[1] and this has certainly been the case with the study by Adam Martin, Yevgeniy Goryakin and Marc Suhrcke[2] from the Health Economics Group at the University of East Anglia.

An ONS study[3] published in February 2014 found that those who walked, cycled or took the bus to work had lower personal wellbeing than car commuters. A rather different result is reported by Martin et al. Their analysis used 18 years of data from the British Household Panel Survey (BHPS), containing repeated observations for individuals of commuting mode and travel time and psychological symptoms. The longitudinal data gave them a stronger basis for causal inference than previous studies, as they could not only control for measured differences between individuals that could confound the relationship between commute mode and wellbeing (these were also taken into account by ONS), but also unmeasured differences (e.g. preferences for travel) and they could assess how transitions between commute modes affect wellbeing. Hence their work can be considered the most robust evidence to date on the relationship between commuting and wellbeing.

Martin et al found that actively commuting to work (walking or cycling) and using public transport are associated with higher psychological wellbeing than driving to work. Not only that but they show a switch from car driving to active travel is associated with a simultaneous increase in wellbeing.

These were the headline findings reported in the media when the paper was published. When you look closely at the results there are nuances that merit mention. In one of the analyses which accounted for specific travel modes, it is only walking and bus commuters that are associated with higher wellbeing (however an acknowledged limitation of the dataset was that relatively few people ever cycled or took the train). The longer the duration of the walk to work the larger the increase in wellbeing (with the opposite effect for car commute duration). When looking at mode switching, it is only those commuters that switch from car driving to walking that improve their wellbeing (and not from car driving to cycling or bus or rail). Also, switching from walking to work to car driving is not examined.

Having got these results, it seems important to try to explain them. Why would it be expected that how you get to work might influence mental wellbeing? What is the causal mechanism? We could suggest that walking (like other forms of exercise) involves physical movement that relieves stress and provides time for reflection that helps in dealing with issues of daily life. Other modes do not offer this quite as well – for example, cycling requires concentration to attend to risks on the road. In fact, other modes could introduce stress due to uncertainty about journey times.

The main results of Martin et al focus on an overall score for psychological wellbeing constructed from 12 different psychological symptoms. When looking separately at the 12 symptoms, it is found that active commuting (results are not separated for walkers and cyclists) is associated with lower scores for being ‘constantly under strain’ and ‘unable to concentrate’ but is not statistically significant at 95% level with different scores for other symptoms. The association with these two symptoms does however support the hypothesis that walking relieves stress and assists in dealing with daily life.

So, why did we find this work of great interest? What was common in the Martin et al study and our Life Transitions and Travel Behaviour study is that they both analysed changes to commute mode recorded in panel surveys. Martin et al looked at the effect of commute mode changes on wellbeing. We were looking in the opposite direction at what factors influence a change in commute mode to occur in the first place. We used the first two waves of Understanding Society, the successor to BHPS. Put the two studies together (as suggested by Martin et al in their paper) and you can consider in what circumstances people are likely to switch from car to active commuting and gain the wellbeing benefits that they discovered. This could have important implications for the design of behaviour change interventions.

In the Life Transitions and Travel Behaviour project, we found that switches from non-active commuting to active commuting were strongly predicted by job or home location changes that shortened distances to work (to within 3 miles) and brought people into more supportive environments for active travel (i.e. living in mixed land use neighbourhood). They were also more likely among those with pro-environmental attitudes.

Incidentally, this raises the question of whether the increased wellbeing that Martin et al found with walking to work could be picking up the benefits of wider lifestyle changes that occur at the same time as switching to walk to work. The researchers accounted for changes in residence, job and neighbourhood and job satisfaction and hence provides strong evidence that walking to work has an effect in its own right. Nevertheless, it would be worthwhile to conduct further research to examine if life changes that occur at the same time as a change in commute mode have a separate, direct effect on wellbeing, as well an indirect effect through a change in commuting mode. Although the Martin et al study used data from an 18 year period, only changes that occurred between any two consecutive waves were considered in the analysis. A development would be to track commuting behaviour over a longer period and look at whether commuting dissatisfaction is a prior prompt to a job or home move and whether improved wellbeing arising from a commuting mode change is maintained or grows over the longer term. Building on the work of the Life Transitions and Travel Behaviour project, which linked data from the panel survey with secondary data on built environment characteristics, future research could also consider whether the effect on wellbeing of switching to more active travel modes varies in neighbourhoods with different characteristics – it would be expected that walking to work would be more relaxing in parkland than along main roads. This would strengthen the case for investment in environmental improvements.

So, encouraging workers to live closer to their jobs and in walkable neighbourhoods (especially those workers with pro-environmental attitudes) could facilitate increased walking to work with the attendant benefits to wellbeing. While commuter cycling did not come out as improving psychological wellbeing in Martin et al study, there is cross-sectional evidence that it increases physical health and from that perspective it is also worth encouraging workers to live within distances that are able to be cycled, even if not walked. Other schemes, such as that recently proposed by Lord Darzi[4], whereby incentives are used to encourage public transport commuters to walk or cycle the first mile to/from home or work, could also be used to support active commuting amongst those for whom longer journeys are unavoidable. It might also be hoped that current day improvements to cycling environments could enable the wellbeing impacts of walking to be appreciated also by cyclists.

—-

Kiron Chatterjee and Ben Clark (20/10/14)

[1] Walking or cycling to work ‘improves well-being’ Available at: https://www.bbc.co.uk/news/health-29175088

[2] Martin, A., Goryakin, Y., Suhrcke, M. (2014) Does active commuting improve psychological wellbeing? Longitudinal evidence from eighteen waves of the British Household Panel Survey, Preventive Medicine. Available at: https://dx.doi.org/10.1016/j.ypmed.2014.08.023

[3] Office for National Statistics (2014) Commuting and personal wellbeing. Available at: https://www.ons.gov.uk/ons/rel/wellbeing/measuring-national-well-being/commuting-and-personal-well-being-2014/art-commuting-and-personal-well-being.html

[4] London Health Commission (2014). Better Health for London. Available at https://www.londonhealthcommission.org.uk/better-health-for-london/