Predicting incomplete COVID-19 vaccination schedules in Scotland

Plain English summary

It has been shown that uptake of the first dose of vaccine associated with the Scottish COVID-19 vaccination programme exceeded 90% yet, uptake of later doses reduced. There was concern that future vaccine uptake may further decline. Some individuals at higher risk of COVID-19 related hospitalisation or death including smokers, males, persons of Black ethnicity and pregnant women are among the most vaccine hesitant.

Why did we carry out this work?

Previous research has identified characteristics of not receiving any COVID-19 vaccines in adults in Scotland. However, this study is the first of its kind to characterise individuals who have previously received a COVID-19 vaccine but did not receive subsequent doses. In other words, we identified the population and medical predictors of individuals who did not complete the recommended COVID-19 vaccine schedule.

Read the summary of our previous work: Characterising unvaccinated adults in Scotland 

What data did we use?

We used the EAVE platform to access health records of 4.2 million residents in Scotland. Those who were registered with a GP in March 2020 and had received at least one COVID-19 vaccine before 31 May 2022 were included. Using the above data, two studies were designed to identify risk factors for those who did not complete their COVID-19 vaccine schedule, who were alive and who had sufficient time to be vaccinated with either the primary schedule or third dose/first booster.

These terms were defined as follows:

• Primary schedule: individuals who received two vaccines
• Third dose: individuals who have received three doses either via the primary schedule or as a booster (primary schedule + booster)

Individuals were included in the above studies if they were aged 18 and over with a GP registration in Scotland. The study period for the primary schedule was from the 08 December 2020. The third dose analysis started on 14 September 2021. Both study periods ended on 31 May 2022. Our main outcome was vaccination status on 31 May 2022 among those eligible for vaccination at the time.

We excluded individuals if they died during the study period, were no longer living in Scotland or did not have sufficient time between doses as per the vaccine schedule. This left the following individuals included in our study:

• Primary schedule analysis: 3,826,797 people
• Third dose: 3,711,756 people

The data sources are listed below and were linked via a pseudonymised Community Health Index (CHI) number.

• GP records: deprivation, household identifiers, smoking status, health board, urban/rural status, morbidity
• National Clinical Data Store (NCDS)/Vaccine Management Tool (VMT): vaccine status, age, sex, vaccine product
• ECOSS (Electronic Communication of Surveillance in Scotland): positive test for SARS-CoV-2 after previous vaccination,
• SMR01: Cause of hospitalisation due to  a potential AESI)
• COPS: Pregnancy status
• RAPID (Rapid Preliminary Inpatient Data), Unscheduled Care, CMS (Case Management System): ethnicity data

Using statistical analysis techniques, several models included characteristics that were computed to predict risk of incomplete vaccination. These characteristics were:

• Household vaccination status
• Household testing status
• Sex
• Age
• Ethnicity
• Deprivation
• Smoking status
• Positive SARS-CoV-2 test after previous vaccination
• Hospitalisation with a potential AESI
• Health board
• Urban-rural status

In addition, the impact of multimorbidity (people living with multiple health conditions), and certain health illnesses were explored in separate statistical models. Sub-analyses were also completed on incomplete vaccine schedule in pregnant women.

What did we find?

Of the 3,826,797 people who received one vaccine dose, 97.5% of these received a second dose and 86.5% received all available doses during the study period.

Individuals aged 18-29 were four times more likely not to receive a second dose, those who were hospitalised with a potential AESI were 3.5 times as likely, and those who resided in the most deprived area were three times as likely not to receive a second dose.

Those from Black, Caribbean, or African ethnic backgrounds were 37% more likely to have an incomplete primary schedule compared to those from white ethnicities. Also, individuals with the conditions below were also more likely to have an incomplete primary schedule compared to those who did not have these conditions:

• Thrombosis or pulmonary embolus
• Severe mental illness
• Cirrhosis of the liver
• Epilepsy
• History of pelvic or leg fracture
• History of stroke
• Chronic obstructive pulmonary disease (COPD)
• Learning disability.

Predictors of not receiving a third dose included being aged 18-29, living in the most deprived area, and having black, African or Caribbean ethnicity.

Other predictors of an incomplete vaccine schedule were pregnancy, having a history of smoking, individual and household COVID-19 positivity, having an unvaccinated adult in the household, or previous vaccination with a mRNA-1273 vaccine.

Why is this work important?

This is the first national study to investigate incomplete COVID-19 vaccine schedules in those with existing conditions. Our findings are from over 3.7 million people and have found common risk factors of people who did not receive the recommended number of COVID-19 vaccine doses. The predictors of incomplete COVID-19 vaccine schedule included:

• younger age
• higher deprivation
• ethnic minority background
• history of smoking
• Pregnancy
• Hospitalised with a potential AESI after a previous dose
• Household vaccination status and SARS-CoV-2 positivity
• SARS-CoV-2 infection history

In addition, having certain conditions such as, COPD, epilepsy or cirrhosis of the liver also were a risk factor for incomplete vaccine schedules. Although most people who were at higher risk of severe COVID-19 outcomes were more likely to be vaccinated, some high-risk groups had lower than expected uptake. These findings can help improve future vaccination programs by identifying specific groups for targeted health messages and actions.

This study has a few limitations. A minor limitation is inaccurate vaccination records for individuals who received their vaccine elsewhere within or outside of the UK.  Similarly, records for those who moved recently, had no recent attendance to their GP or were recently diagnosed with a condition may not have been identified. AESI may have been incorrectly associated with vaccination simply due to hospitalisations occurring closely following vaccination by coincidence. Some AESI caused by vaccination may have been excluded from the conditions of interest chosen and therefore missed. Immunocompromised individuals were unable to be identified. Finally, there was a shorter follow up period for the third dose compared to the completion of the primary schedule.

Note

This plain English summary was written by our Patient and Public Involvement Officer Anna Crawford, with the support and feedback of the EAVE II Public Advisory Group (PAG) and researchers. This particular summary was reviewed by PAG member David W.