15 July 2021
Lies, Damned Lies and Coronavirus
Jabs work
by David Chilvers
Last week we highlighted how the current rise in positive COVID test results was being driven by school children and university students. We looked at data from the smallest areas that the Government releases (MSOAs – Middle Level Super Output Areas) and saw that MSOA’s with University in the name had higher case rates than other areas, replicating what we had seen when university students returned to campus last autumn. Recently, data on vaccine uptake has also been released at MSOA level and this made me wonder whether age was still a contributory factor to determining case positivity or is it all down to the vaccine rollout which has been focused at older age groups and as such may be the reason behind the apparent age differences in case positivity. This analysis is not straightforward, as many factors are conflated at a local level:
- The presence of a university affects the age distribution in a local area
- Deprivation, ethnicity, economic activity levels and type of work undertaken are all inter-related
So, a multivariate approach to analysis is necessary, to try and tease out the specific impact of factors as opposed to those that appear to occur just because they are related to another impactful variable.
We used this approach in September last year in relation to death rates from COVID-19 and concluded that “Ethnicity, for example, shows much lower death rates for those of White ethnicity, slightly lower for people of mixed race and above average for all other ethnic groups. Males have higher death rates than females and there is a substantial gradient by age. Those unemployed and with pre-existing health conditions have higher deaths rates, having already taken account of ethnicity, gender, age etc.” COVID-19 in its initial waves had the greatest effect among the elderly, those with pre-existing medical conditions, men and black and Asian ethnic groups, whether the metric was case positivity, hospitalisation or death.
Since then, the vaccine programme has been initially targeted at the older age group and those with pre-existing health conditions and so it is hardly surprising that infection levels are greater among younger (and thus less well vaccinated) groups. That vaccination is the reason for the changing age profile of those catching COVID-19 has been taken as a given in the media but there has been little direct evidence of the link between vaccination and infections, hospitalisations and deaths and whether this removes any differential by age.
To try and understand this link a little more, we have taken data at MSOA level for:
- The % of the population in the MSOA that had a positive coronavirus test in the last seven days (as at 6th July, the latest date for which this data is available)
- The cumulative level of vaccines in the MSOA, both first dose and second dose (as at 10th July, again the latest date for which this data is available)
- A wide range of demographic variables which are available at MSOA level from the 2011 Census; although these are out of date, the general demographic and economic make-up of many local areas will still be broadly right
This mirrors the analysis of mortality undertaken in September last year and allows us to determine:
- The simple correlation across 6,728 MSOAs in England between the % of coronavirus positive tests and all the other variables
- A more complicated regression analysis which adds variables one at a time to a model, which at the end of the process contains all the factors that individually are significant predictors of current coronavirus positive test levels and collectively are the “best” set that can be derived
So, looking first at the simple correlations, these are the ten factors with the highest positive correlation with the % of positive coronavirus cases in the MSOA:
| Variable | correlation with covid % |
| Economic Activity: Economically inactive: Long-term sick or disabled | 0.266 |
| Cars: None | 0.260 |
| Health: Bad | 0.249 |
| Age 16-24 | 0.231 |
| Method of Travel to Work: Bus, minibus or coach | 0.231 |
| Health: Very bad | 0.229 |
| Method of Travel to Work: Taxi | 0.211 |
| Occupation: Sales and customer service occupations | 0.209 |
| Size: Single person | 0.209 |
| Deprivation: Level 3 | 0.190 |
At the top of the list are the long-term sick or disabled. MSOA’s with high percentages claiming their health is bad or very bad are among the top ten correlates of high current positive test rates. Those in one person households feature on this list as do those with no cars and those who travel to work by taxi or bus, where there will be a degree of social mixing. These metrics are summed up by the presence of three of the four deprivation measures used by ONS which comprise:
- Employment: Where any member of a household, who is not a full-time student, is either unemployed or long-term sick.
- Education: No person in the household has at least Level 2 education (see highest level of qualification), and no person aged 16 to 18 is a full-time student.
- Health and disability: Any person in the household has general health that is ‘bad’ or ‘very bad’ or has a long-term health problem.
- Housing: The household’s accommodation is either overcrowded, with an occupancy rating -1 or less, or is in a shared dwelling, or has no central heating.
The final two factors positively correlated with case positivity are people working in sales and customer service occupations, which will include most of the retail sector and finally those aged 16-24. As outlined last week, this current wave is having its greatest effect on younger people and this analysis confirms that and the impact on those either with health conditions or those in the least advantaged segments of our society.
At the other extreme, the ten factors with the highest negative associations with coronavirus case rates are:
| Health: Good | -0.332 |
| Economic Activity: Economically active: Self-employed | -0.331 |
| Cars: 1 or more | -0.260 |
| Method of Travel to Work: Work mainly at or from home | -0.253 |
| Industry: Agriculture, forestry and fishing | -0.224 |
| Method of Travel to Work: Motorcycle, scooter or moped | -0.216 |
| Occupation: 1. Managers, directors and senior officials | -0.185 |
| Occupation: 5. Skilled trades occupations | -0.184 |
| Size: Family of 2 or more | -0.184 |
| Age 65-74 | -0.169 |
Some of these are the polar opposites of the factors generating high positive test rates: those in good health, working from home, in managerial or skilled positions, the self-employed, those with one or more cars and in larger households and those aged 65-74.
Having one or more doses of the vaccine also correlate highly and negatively with current coronavirus case rates: for the first dose the correlation is -0.107 and for the second dose -0.143. Clearly, vaccination rates are correlated with lower infection rates when looking across this group of 6,728 MSOA’s. But these two metrics are correlated with one another – it is hard to have a high second dose % without having a high first dose % and so the question arises of which of these is the most significant (on the face of it the second dose as this has the higher correlation).
More generally, this simple correlation analysis may be obscuring the real impact of factors; areas with large numbers of older people will generally have seen more vaccination, so is it the age range or the vaccination that is impacting on test positivity and if vaccination is a significant driver of infection levels, is that instead of age or as well as age? The regression approach helps to answer this and the results are quite illuminating.
| Factor | Direction |
| Have had second dose | Negative |
| Method of Travel to Work: Motorcycle/scooter/moped | Negative |
| Cars: None | Positive |
| First dose only | Positive |
| Qualification: Apprenticeship | Positive |
| Industry: Manufacturing | Negative |
| Method of Travel to Work: Driving a car or van | Positive |
| Health: Good | Negative |
| White (GB) | Positive |
| Occupation: Process plant and machine operatives | Positive |
The single most important factor negatively associated with case positivity is having a second dose – the MSOAs with high levels of second dose will tend to have lower case positivity. The other factors with a negative association – helping to reduce case positivity – are travelling to work by motorcycle/scooter/moped (a small group), working in manufacturing and being in good health. The top factors with a positive association – helping to increase case positivity – are having no car, having a first dose but not second, being an apprentice, travelling to work by car, being of White:GB ethnicity and working as a machine operative. Age does not feature in the top ten factors, the first age group featured is those aged 16-24 with a positive association (more likely to have a higher case positivity) at 17th place in the list of 50 significant predictors.
So, this analysis tells us that whilst infection in the current wave is skewed towards the young, this appears to be mainly due to a lack of two vaccinations rather than age itself (remember that pre vaccination, age was the single most important determinant of catching COVID-19 and being seriously ill and needing hospitalisation). Interestingly, having only one dose is positively associated case rates. From this analysis, we can see that age is much less relevant and it is vaccination status that is driving case positivity (and thus subsequently hospitalisation). Get jabbed and get double jabbed – a message the Government has been repeating frequently – does indeed seem to be the single most important factor associated with positive case rates. Finally, deprivation continues to be an important driver of infection, as it has been through the pandemic. This is clearly not levelling up.
This article is one of a series, the previous article on the kids are alright is here.
