Some of the team have just returned from this year’s excellent International Medical Geography Symposium , and it was a good opportunity to think about this project’s geographical approach to estimating people’s ‘exposure’ to natural environments. In trying to figure out relationships between natural environments and health and wellbeing,we’re taking a pretty typical epidemiological approach – we consider natural environments to be an ‘exposure’ and various aspects of health and wellbeing as ‘outcomes’ and try to assess relationships between the two. It’s therefore pretty important to consider how that exposure happens – e.g. people can spend time in natural environments (actively or passively); they can just look at them; they can simply be aware that they’re out there somewhere.

We’ve been thinking a lot about this, about how we try to estimate exposure using the kind of data we’re using in this project, and what assumptions we’re making along the way. In common with many similar studies (e.g. this one on radon and skin cancer), we’re estimating exposure based on the home residence location of people that are included in the various datasets we’re using (e.g. the census). So, we’re making the assumption that the environments in the vicinity of where someone lives are a good indicator of the environments that they typically spend time in, look at, or otherwise experience.

Clearly that assumption is subject to error. For example, I spend a lot of my waking hours time sitting at my desk in the office , looking at some big trees:

That’s an everyday visual exposure, which may well be significant (e.g. see Ulrich, 1984) – but is not at all reflected by measuring what the environment is like around where I live [slight tangent – Kat Gilchrist is researching workplace greenspace…]. I do spend a lot of time in the area around where I live, so it’s not a bad estimate – it’s just not a complete picture of the environments I’m exposed to.

There are ways and means to be a bit more sophisticated than simply using home location. We could consider other significant locations in people’s daily lives (work, school – even second homes) and combine multiple measures of local environments – but often we don’t have the data to do that. We can get even more complex and ask people to carry a GPS datalogger with them so we can track exactly where they go and what environments they spend time in – e.g. see the PEACH project. However, that’s pretty intensive data to collect, and usually we can only get it for a short period of time for a relatively small number of individuals. Also, it doesn’t capture the potential importance of ‘unvisited’ local environments.

So, linking geographical measures on natural environments to home residence locations might not be a perfect way to estimate exposure to natural environments, but it does allow us to make use of some very large and powerful datasets in this research. We just have to be conscious and clear that it’s only one of many ways of thinking about how nature, health and wellbeing might be related.