Stopping smoking: How to target smokers using geodemographics
October 6, 2016
Stoptober is back once again and since its inception in 2012 it has encouraged almost 1 million* people to quit. We first published this blog on 10.10.13 but it is still just as relevant now.
This month the Stoptober campaign gives smokers a great opportunity to quit by providing free support packs along with daily tips and advice throughout October.
This got us thinking about how we can use geodemographics to target smokers and direct anti-smoking campaigns to where they’d have maximum effect?
We decided to base this investigation in York and we therefore needed to check how York compares with other Local Authorities for premature deaths and number of smokers.
Recently Public Health England put York at the bottom of a league table for premature deaths:
The “similar areas” are chosen by what they call the “socioeconomic deprivation bracket”. One of the responses to the article from Council Leader, James Alexander pointed out that York was the odd one out in the list. Furthermore, he commented that the Government says York’s lung disease issues are related to both smoking and air pollution. Using the information that Beacon Dodsworth generates together with other surveys, we decided to investigate.
Our demographic classification, P² People & Places, classifies residential neighbourhoods using data gleaned from sources such as the UK census. At the highest level each postcode is given one of fourteen codes ranging from A to M (plus a code for unknowns U)**. Neighbourhoods in category A are the most affluent and those in category M are the least affluent. However, affluence is not the only determining characteristic; aspects like the degree of urbanisation also affect the classification of a neighbourhood. Here are the categories at this highest level, which we call the “tree level”:
You can view the categories for areas of York (or anywhere else in the UK) by opening our web profiler demonstration and entering a valid postcode. Here is an example map for the Clifton area of York:
There are two other levels of the P² People & places classification, the branch level, which has 41 categories and the leaf level which has 157 categories. For this analysis, I used the 41 categories of the branch level of P².
How does this relate to targeting smokers?
Public Health England chose the group of local authorities according to socioeconomic deprivation.
Their league table is based on the idea that the measure of social deprivation will predict the level of premature deaths. The higher the level of deprivation the greater the expected number of premature deaths.
To a large extent this works but it is a bit too simple, so I looked for other measures that might help to give a richer picture. One cost-effective way is to use a geodemographic classification like P² People & Places, in conjunction with other data.
The warnings on the cigarette packets tell us “Smoking Kills” so a good place to test for health-related similarities or differences is to look at the number of smokers in each local authority.
We could estimate the likely level of smoking within local authority areas using P² because it is coded against the Living Costs and Food Survey from the Office of National Statistics. For each of the 41 P² branch level categories this survey can be used to estimate their expenditure on tobacco.
Knowing the P² demographic breakdown of the local authorities enabled us to make predictions of the amount spent on tobacco in each of them. Here are the predictions for the amount spent on tobacco averaged over the population of each local authority:
These estimated figures give York a 17% higher spend on tobacco than Bromley because York has more people in the groups that are the heaviest smokers. Let us call the categories of households that nationally average more than £6 per week expenditure on tobacco as “the smoking classes”.
Here are the P² categories that make up the smoking classes:
My calculations show that 5.8% of Bromley’s population comes from these smoking classes compared to 16.7% of York’s population. This appears to show that the populations of Bromley and York are not so similar after all.
There are other ways of demographically ranking Local Authorities than using socio-economic deprivation. Geodemographic classifications, such as P² People & Places, could provide a more informed picture. In particular, Beacon Dodsworth can help with targeting anti-smoking campaigns in the neighbourhoods where the smoking classes live.
Find out more
*This information was taken from the Public Health England website
**Please note that we updated P2 People & Places in Jan 2016 to use the latest Census and lifestyle data and therefore accurately reflect the population.