Different neighbourhoods had very different Covid vaccination rates: why?

What is it about?

After the first year of the Covid pandemic in Britain, vaccines were freely available from the National Health Service. They saved lives. In some areas of Liverpool, where I live, most people took the vaccines but in other areas they did not. Was this just a matter of chance? Was it about the types of people who live in different parts of Liverpool? Or was it about where they live within the city? We can ask the same questions elsewhere in England but does it matter which town or city we look in? I made many statistical models and compared them to find out which factors were most able to explain or predict the vaccine uptake in Middle Super Output Areas all across England. MSOAs are small geographical areas used in the Census, with around 8,000 people in each. At the same time as the 1st Primary vaccines were being rolled out, the 2021 Census recorded up-to-date anonymous information in each MSOA on people’s age, gender, ethnicity, education, employment, housing, whether they had recently moved within the UK or from another country, if they lived in a prison or a military base, if they declined to answer questions about their religion or sexuality... Other public data showed deprivation levels, the proportion of people in various areas receiving flu vaccinations, the proportion voting to Leave the European Union. The models used this public data. They also allowed for unexplained differences between local authorities (like cities, boroughs, county councils) and between different regions (like the North West or South East). I chose the model which did the best job of predicting what happened in each MSOA when it was tuned up using data from other MSOAs. In any particular MSOA it identified the factors which had the most impact on the uptake there. It also identified the key factors for predicting results across all MSOA in England. The proportion of people who described themselves as “Other White” (including Eastern European ethnicities like Polish or Romanian) turned out to be more important than most other factors, including age and deprivation, when it came to predicting the uptake across England. But deprived areas with rapid population turnover also had much lower uptake. Looking at the unexplained differences between different local authorities, it seems that talking with people face-to-face in their own languages helped to improve uptake.

Featured Image

Photo by CDC on Unsplash

Photo by CDC on Unsplash

Why is it important?

An infectious disease pandemic affects everyone, so no-one is safe unless everyone is safe. With a future pandemic there may be other vaccines. But the factors which predicted the uptake of Covid vaccines may still be relevant with different vaccines and a different disease. Finding a model which predicted Covid vaccination uptake in each local area from the data in other areas, is a step towards predicting the uptake in different circumstances. Understanding the barriers which people faced with Covid vaccination could help to overcome them before the next pandemic strikes. You can read the paper without following all the maths. Key takeaways: • Bayesian models can predict Covid vaccination area uptake at MSOA level in England, with very few outliers. • Problems occurred at the outset of the vaccination programme. Area uptake of the 1st Primary was decisive in predicting area uptake of the Booster / 3rd Injection. • Key factors include: local authority and regional “random effects”, the ratio of Vaccine Registry to eligible population, the area proportion of “Other White” ethnicity, deprivation, housing in military bases and prisons, various occupations, aged 55 and over, international migration, indicators of distrust in government, “African and Other Black” ethnicity... • Qualitative evidence from the highest and lowest local authority effects and from one outlier, suggests that speaking with people in their own language is essential.