Studying area effects: How the old neighbourhood may shape our old age ShareThis


(…Or my approach to small talk at dinner parties…and possibly why I’m not invited to more)

Whenever I am at a dinner party, or meet someone new outside of academia, they always ask me what I do for a living. When I explain that I research how the places people live can affect their health, I usually get one of two reactions: (1) they’ve never heard the concept before but it makes sense intuitively or (2) they believe that they are master of their behaviours and this will override any environmental factors.

For the latter, I have a ready answer: why do we readily accept the idea that the air we breathe can affect our lungs, that polluted waters can spread disease, and contaminants in our soil can lead to cancer, birth defects, and a whole host of health problems? Yet we can’t accept that having more green space, recreational centres, and healthy food in our neighbourhoods can promote more physical activity and better diets? Or that having more individuals in a neighbourhood who smoke, drink alcohol or eat a certain way can also influence individual behaviour?

I usually end with a little add-on about how I have been attempting to study these ‘area effects’ across the whole of a person’s life. But this little ad-on is really the bulk of what my research has been focused on. Most existing studies of area effects on health rely on measures of areas assessed at a single point in time in adulthood. Factors that potentially determine health outcomes at older ages can reflect trajectories and processes that have occurred over decades. If over time people move to areas with varying degrees of these qualities (I myself have moved 15 times in the past 14 years), or the areas they live change, using a single measure of residence could misestimate the effects of area on the health outcome.

So it makes intuitive sense that if we want to examine area effects on a chronic disease, or biological ageing, we should measure an individual’s exposure to area conditions over the entire life course. Much easier said than done.

As a part of the Healthy Ageing across the Life Course (HALCyon) programme (, my colleagues and I linked prospectively collected residential addresses from the MRC National Survey of Health and Development (NSHD), the 1946 British birth cohort study, at ages 4, 26 and 53 years to census socioeconomic data in 1951, 1971, and 2001. Then we used these data to estimate when in the life course area deprivation was associated with physical and cognitive capability.

The main impetus of this research, that a ‘single point in time’ measurement of area is inaccurate of lifetime exposure, was fully born out: the number of individuals who experienced both absolute and relative area socioeconomic changes was considerable. For example, of cohort members who resided in the top 25% most overcrowded areas in 1950, a full 78.3% no longer resided in that most deprived quartile by 1999. Other available area measures of occupational social class, unemployment, household amenities and higher education displayed similar patterns.

This research project has had its share of challenges, discussed in detail elsewhere (Murray et al, 2012). One example is that for consistency across all time points we would have liked to have represented area deprivation with established deprivation indices, such as the Townsend or Carstairs. However, two variables included in these indices – car ownership and renting – were not collected before the 1971 census. It is also unknown whether changes in area socioeconomic measures were due to residential mobility or changes in areas, as municipal changes have made census boundaries not comparable across census years.

In general, for each of the separate area level measures, the more disadvantaged the area in which a cohort member resided at age of 53 years (1999), the worse their physical capability at the same age. However, this differed by the physical capability outcome, with a higher correlation for the ability to balance on one leg with eyes closed than to rise from a chair ten times without the use of the arms. In particular, the percent of persons in an area working in partly- or unskilled occupations had a stronger correlation with physical capability than all of the other area variables, including when all area measures were combined into one summary measure. It was therefore fortunate that the percentage of employed persons with low social class occupations in an area was collected in the census at all time points.

This methodological work led us to consider class of occupation as the potential mechanism by which ‘area’ over the life course is acting on physical capability. Our more recent work has incorporated this and other lessons to examine whether area deprivation during childhood and young adulthood contributes additional risk to poor physical and cognitive capability later in life, over and above the influence of individual social factors.

But that’s a story for another blog post…or another dinner party.

Dr Emily T Murray works on the Healthy Ageing across the Life Course (HALCyon) programme as a Post-Doctoral Fellow in Life Course Epidemiology at the National Institutes of Health and the Medical Research Council (MRC) Unit for Lifelong Health and Ageing. Special thanks to her colleagues on the project: Professor Rebecca Hardy, Professor Diana Kuh, Professor Yoav Ben-Shlomo, Professor Kate Tilling, Dr Humphrey Southall and Ms Paula Aucott.

For more information, see Murray ET, Southall H, Aucott P, Tilling K, Kuh D, Hardy R, Ben-Shlomo Y. (2012) Challenges in examining area effects across the life course on physical capability in mid-life: findings from the 1946 British Birth Cohort. Health Place, 18(2), pp. 366-74.