Why
There is a lot of vaccine scepticism, which means even if government published vaccine effects are shared, there will be a cohort of people who will still have doubts. If we could use publicly available data that would be hard for a government to have doctored to generate some visualisation we might be able to see if the vaccine is working.
How
Data Sources
We can obtain the rates of new infection by age from the UK Government website.
#
Comparator
The early vaccine campaign in the UK concentrated on patients over 80 years of age, and front-line healthcare staff. Therefore we should be able to compare the over 80 population to the rest of the population. As the roll-out matures, we will need to add Over 70’s and then younger groups.
#
Outcome
Defining what outcome we are looking for with the vaccine campaign is an interesting debate which could be the subject of a very long blog post in its self. The clinical trials reduced hospital admissions and serious infection rates, however, only the AstraZeneca vaccine checked participants for mild / asymptomatic infection. It is even harder to explicitly check for transmission. We could use hospital admissions, we could use deaths. However, both have a lag from infection, and reducing infection is likely to reduce the rate of hospital admission and death.
Analysis
All the analysis is within the R code described below, using the initial steps described above.
I started planning this process in late December 2020, before the UK’s third lock-down was announced. Infection falling does not guarantee that the fall is the result of vaccination, it could be the effect of lock-down. We will therefore need to make some comparisons with previous lock-downs. The first UK lock-down was in late March 2020 however, testing capacity was significantly constrained at the time so the data on infection rates is unreliable. The second lock-down was not as intensive, or long as the first or third, however we should include that data for comparison.
Results
The aim is to visualise the results, so we should produce a graph.
#
Conclusions
The rate of infection in November Lockdown for over 80s fell slower than the rate of infection for the under 70-79s and under 70s. Both the 70-79s and under 70s age groups fell at very similar rates. The number of infections in the 70-79 age group and over 80s is roughly similar and therefore this is unlikely to simply be a numbers issue.
In the January Lockdown