Given the unprecedented scale and impact of the COVID-19 outbreak, it is perhaps unsurprising the heightened level of discussion, debate and analysis of every aspect of the pandemic. However, it is striking and heartening to see the data community at the forefront of this; individuals and organisations collating, compiling and presenting statistics in such a way as to help better understand what we are being confronted with.
There have been a number of articles and visualisations that have really caught my interest over the last few weeks, in particular this one examining the different experiences of two countries (Italy and South Korea) with similar levels of infection. What the data suggests is that Italy’s experience (in terms of mortality) has been exacerbated in part by (a) having an older age profile in general and (b) the infection taking hold and spreading within this cohort; the implication being that predicting the impact of the virus is very much dependent upon through which sections of society the virus spreads through, and this is why shielding those groups is a necessity.
With this in mind I decided to see what the picture for the UK looked like in terms of where are those ‘vulnerable’ populations? There are various open data sources available to help categorise this population, but in the end I chose Census 2011 data (for England, Wales, Scotland and Northern Ireland – Figure 1). Yes, there are issues with the timeliness, but the trade-off is that I was able to identify a suitably well-defined ‘vulnerable’ group; any resident over 65, reporting ‘Bad or very bad health’. Furthermore, the data was available to analyse this population by Local Authority and to specifically pick out social renters.
So what does the data tell us?
According to the census, overall 82% of people reported being in ‘very good or good health’, with only 5% of the population reporting ‘bad or very bad health’. However, if we narrow this down by adding age into the criteria we find that this figure increases markedly (unsurprisingly). When looking at those 65 years and older the proportion of those in ‘bad or very bad health’ rises to almost 15%.
But how does this compare when looking at those who are social renters? Across all social renters 12% report ‘bad or very bad health’, but this jumps to 26% for social renters who are over 65. So, one of the takeaway statistics is that 1 in 4 social renters over 65 are in this vulnerable grouping – equating to 423 thousand people across England, Wales, Scotland & Northern Ireland.
Specifically focussing on total numbers of this ‘vulnerable’ population (figure 2); the regional picture shows a considerable variation. Unsurprisingly, the regions with populous metropolitan areas tend to hold larger ‘vulnerable’ populations, although the fact that London and the South East (the most populous regions) are behind the North West, suggests either a slightly healthier and/or younger population in these two regions.
When isolating just social renters the picture shifts somewhat; London and Scotland and the North East become more pronounced, whilst regions such as the South East fall away a little. This is likely to reflect higher rates of social renting in these areas.
If we move away from total numbers and look at the rates of these ‘vulnerable’ populations amongst the community i.e. how many per 1,000 of the resident population (figure 3 which better communicates the density/severity in each area) then the variation changes considerably. What becomes apparent is that it is the less populous regions, such as Wales and the North East that proportionately have the highest numbers of ‘vulnerable’ people, with London dropping way down to the bottom.
The picture for those that are social renters broadly follows a similar profile, however, there are some significant exceptions with the North East, Scotland and London, in particular, bucking the trend (most likely a function of higher social renting rates). What these charts show is that whilst Wales is a real concern generally, in the social housing community it is areas such as the North East which could potentially be hit the hardest (again, due to the fact that the North East has a higher proportion of social renters than Wales – 23% and 16.5% respectively).
The last part of the puzzle was to look at the data at Local Authority level; are there patterns that haven’t been picked up at a broader, regional level? To help with this, I put the data onto a map, showing the ‘vulnerable’ rates per 1,000 resident population across all tenures, as well as for social renters. N.B. the data for Northern Ireland wasn’t available at this level so has been excluded.
What is clear from this map (figure 4) is that there are specific concentrations of these ‘vulnerable’ populations in South Wales, Greater Manchester/Liverpool and the North East, with a large degree of crossover with those who are also social renters. The most significant difference between these groups relates to the fact that social renting rates are significantly higher in the North East than in Wales. However, it is easy to characterise these areas, independent of tenure, as all having experienced post-industrial decline of one form or another; be it coal mining (Welsh Valleys, Bolsover, County Durham), shipbuilding (Glasgow, Sunderland) or manufacturing (Knowsley), etc.
My last point takes me back to the start of this article – in that the potential impact of Covid-19 in terms of mortality will very much be dependent upon which groups the virus spreads through. What has become apparent during the writing of this article is that Gwent is experiencing proportionately high incidents of infection, which is concerning due to the fact that there is a significant cohort of the population who are at risk (as shown by the data), which is why shielding these populations away is of utmost importance, given the experience of Northern Italy.
If you have any questions on this analysis or require further advice, please contact me at firstname.lastname@example.org.
No Related Items
The latest meeting of the Information Security Forum took place over two mornings via WebEx instead of a full day...
Friday 29 May 2020
HouseMark has today shared the results from month two of their sector-wide COVID-19 impact analysis