With the recent shift in health provision away from a medical model to a focus on health promotion, increasing attention is being paid to the role of the built environment, particularly within residential areas [1–4]. According to Jackson, 'we now realise that how we design the built environment may hold tremendous potential for addressing many of the nation's greatest public health concerns' . Much of the research in this area has concentrated on the links between neighbourhood characteristics and physical activity or obesity [6, 7], though there has also been interest in respiratory health [8, 9] and mental health . Another key area of research is the impact of low carbon design on winter deaths .
There are good reasons for investigating the role of the built environment in healthy ageing. As in clinical medicine it is recognised that people need to receive different medical treatments at different stages of life, so it can be argued that non-medical determinants of health will differ in their impacts over the life-course . Longevity is a pressing issue for public health, not least because of the greater likelihood of frailty and disability in older age . Consideration of the built environment is particularly pertinent for older people: as they age, they are likely to spend more time in their home and community environments, and declining health and functional status can make them more susceptible to barriers in them . Escalating care costs and evidence that the majority of older people prefer to remain living in their own homes has led to widespread adoption of 'ageing in place' policies .
For ageing in place to work well, housing and neighbourhood environments need to facilitate older people's independence and wellbeing. According to Liu et al , recent research suggests that wellbeing in later life is closely related to the physical environment, which is an important mediator of ageing experiences and opportunities. The physical character of the neighbourhood in particular seems to have a significant impact on the mobility, independence and quality of life of older people living in the local community . The idea of age-friendly communities was developed by the World Health Organisation through their Global Age Friendly Cities Project launched in 33 cities in 2005. There are alternative names for such communities: for example, they tend to be called 'liveable communities' in the US and 'lifetime neighbourhoods' in the UK . The features of each are similar though they differ in terms of the emphasis given to physical or social elements.
There is a common view, now apparent in UK planning policies and elsewhere, that older people need higher-density homes (usually apartments) in urban locations. This is based on several assumptions: older households prefer less space and freedom from the burden of looking after a large house and garden ; being in urban locations allows older people easy access to public transport, shops, health facilities and other amenities at a time when they may no longer be able to drive or afford to own a car ; and living in higher-density, urban locations provides older people with greater opportunities for social interaction, involvement in the local community and stimulation/interest . It is important to investigate whether policies intended to address the needs of older people are in fact delivering benefits to them.
What is lacking at the moment, in terms of both research and practice, is an integration of the health and built environment areas of expertise [22, 23]. Recognition of the relationship between the built environment and health opens up new avenues for health-promoting interventions, but little progress can be made without evidence of the role of different aspects of the built environment . There are many methodological challenges in conducting research on the built environment and health and wellbeing, and these are fairly well documented: e.g. impracticalities in carrying out trials or controlled experiments, the need to account for residential self-selection or drift, and difficulties in controlling for other influences on wellbeing [25, 26].
From a built environment perspective, perhaps the most significant shortcoming in research to date is that it is often difficult to translate the findings into practice - i.e. to know how the built environment should be changed or designed differently to optimise wellbeing . This is largely because of how the built environment has been measured:
Often it is treated as a 'black box', despite comprising a huge number of individual elements: for example, a neighbourhood environment is unique in terms of its street layout, street design, amount of greenery, presence of street furniture, and the type and variety of buildings within it (to name but a few different elements). In order to know how best to intervene, built environment practitioners (designers, developers etc.) need to know not just whether the built environment makes a difference but also which elements or characteristics are important .
Measures are often subjective - i.e. ratings of quality by residents or researchers [e.g. . Subjective measures are prone to recall error, and same-source bias (e.g. unhappy people are likely to rate their environment more negatively) . Further, they have the same limitations as those described above - i.e. they are difficult to translate into practice (what is a 'better quality' environment?). Lin and Vernez Moudon  found that objective measures of the built environment had stronger associations with amount of walking than subjective measures of the same attributes. Several tools for measuring built environments take the form of an audit or evaluation rather than a simple description - 'walkability' tools are examples of these .
Measures are often a combination of physical and social characteristics, including condition of buildings and streets and prevalence of vandalism, litter and graffiti (e.g. the REAT instrument ) [34, 35]. Researchers are right to believe all these characteristics play a role in people's wellbeing, but from the perspective of a built environment practitioner, they need guidance on those elements over which they have some control.
Measures are collected for a spatial unit, such as a 'lot' or block face, rather than for an individual  - as Clarke and Nieuwenhuijsen  argue, this may not represent the experiential environment of the respondent from whom health or wellbeing data is being collected. Wood et al  measured the residential environment of individuals, but only on four attributes.
The ultimate goal in terms of investigating the impact of the residential environment on healthy ageing might be an experiment or trial involving the evaluation of an intervention, or a prospective controlled study investigating outcomes for older people moving into different settings. However, while there is some evidence that the built environment makes a difference in healthy ageing, there is little understanding of which individual design characteristics are important. The use of objective measures of the built environment has been growing in recent years, mainly through research on obesity . These measures usually include characteristics such as street connectivity, block size, land use mix or presence of 'destinations', development density, presence and width of footpaths, topography, and sprawl [40–43]. Less commonly, studies address street surveillance, street lighting, presence of trees, urban-rural location type, traffic speed, presence of traffic control devices and public transportation infrastructure provision [44, 45].
Studies of the built environment in relation to mental health, stress and social capital use some of the measures above but additional ones related to the more immediate home environment, for example dwelling form/type, storey height of housing, floor level of dwelling, characteristics of the street/home interface (e.g. setback of home, presence of porches/stoops) [46–50]. Purciel et al  have attempted to measure more subtle urban design constructs, including imageability, enclosure, human scale, transparency and complexity, using objective data from GIS. This inventive work is potentially of great relevance to urban designers, but the difficulty is ensuring the objective measures are a true representation of these constructs, and so far they have been developed for spatial units (block faces) rather than individuals. Brown et al  at the University of Miami have devised the Built Environment Coding System, which has many strengths. It measures objectively a wide range of urban design attributes. It is, however, limited in use because, again, it obtains measures for the spatial unit of 'lots' rather than individuals' residential environments.
The purpose of the research reported here was to develop a tool which measures a wide range of urban design characteristics within an individual's residential environment, and to test this tool in a preliminary study, in order to identify potential predictors of older people's wellbeing that are worthy of investigation in future research.