Introduction

COVID-19, the disease caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has spread to become a global pandemic causing significant morbidity and mortality in many countries. As of February 2021, total worldwide COVID-19 cases are estimated to be over 100 million, and deaths related to COVID-19 number over 2.1 million [1]. In recent months, thanks to an unprecedented global research effort, a number of vaccines against SARS-CoV-2 have been developed and approved [2,3], and it is hoped that mass vaccination programmes will aid in slowing transmission of the virus as well as reducing hospitalisation and death from COVID-19.

As the pandemic has progressed, it has become clearer that certain factors may increase the risk of acquiring SARS-CoV-2 infection, including age, obesity and the presence of particular comorbidities (e.g., diabetes and cardiovascular disease), occupation, and household size [4–6]. Amongst these ‘high-risk’ groups are healthcare workers (HCWs) [6,7], in whom an increased risk of hospitalisation with COVID-19 has also been demonstrated [8]. Within a HCW population, it has been shown that the risk of SARS-CoV-2 infection differs by occupational role and is highest in ‘front-door’ and patient-facing specialities [8,9], implying that at least some of the increased risk faced by HCWs is mediated through occupational exposure to SARS-CoV-2. In recognition of this risk, the UK Joint Committee on Vaccination and Immunisation (JCVI) listed frontline HCWs as a priority group for receiving vaccination against SARS-CoV-2 [10].

The COVID-19 pandemic has also disproportionately affected those from ethnic minority groups, with previous work demonstrating an increased risk of SARS-CoV-2 infection and adverse outcomes relative to White individuals [5,11,12]. Furthermore, HCWs of minority ethnicity have been shown to be at higher risk of infection than their White colleagues [9,13].

In light of the increased risk of COVID-19 infection and adverse outcomes faced by ethnic minority HCWs, concerns have been raised regarding uptake of the SARS-CoV-2 vaccine in this group, both in the UK and in the US [14,15]. These concerns are founded upon previous work conducted in the general population, which has demonstrated reduced vaccine uptake by ethnic minority individuals [14], as well as recent survey studies investigating intentions to receive vaccination against COVID-19, which have demonstrated an increased likelihood of SARS-CoV-2 vaccine hesitancy in ethnic minority groups, including amongst HCWs [15–17].

SARS-CoV-2 vaccination in a HCW cohort is important not only for protection of the individual but, given that a significant proportion of COVID-19 inpatients acquire their infection in hospital and that this has been attributed to HCW-to-patient transmission, may also prove to be important for reducing nosocomial transmission of COVID-19 [18,19]. Despite HCWs being important targets for vaccination, there are few studies examining SARS-CoV-2 vaccine uptake (as opposed to vaccine intention) amongst HCWs. A small study of staff working in a specialist orthopaedic hospital in London, UK, determined vaccine uptake to be 62% [20]. A larger nationwide cohort study in UK HCWs determined overall vaccine uptake to be 89% as of 5 February 2021, although, as the authors acknowledge, this estimate may not be generalisable due to the potential for self-selection bias in a consented cohort study [21]. Both studies found lower vaccine uptake amongst those from minority ethnic groups (compared to those from White ethnic groups), females, and those in nursing and portering/estates roles. Additionally, it has been suggested that UK HCWs with a history of COVID-19 may be less likely to take up vaccination than those without evidence of previous SARS-CoV-2 infection [21]. Data from outside the UK are limited; an Israeli study reported SARS-CoV-2 vaccine uptake in HCWs to be 90% but did not attempt to determine factors associated with uptake [22].

To add to the limited evidence base on this important public health issue, particularly in light of the potential requirement for a SARS-CoV-2 booster vaccination for UK HCWs, we sought to determine the effects of demographic factors (including ethnicity), occupational factors, and previous COVID-19 on SARS-CoV-2 vaccine uptake in a large multi-ethnic UK healthcare workforce.

Methods

This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist).

Results

Demographic and occupational characteristics of the cohort

In total, 19,044 HCWs were included in the final analysis (see Table 1). In total, 47.7% were under 40 years of age and 75.6% were female; 60.3% were White, 25.5% were South Asian, and 7.1% were Black. Data were missing for 110 values for IMD and 1 value for date of vaccination. The number of vaccinations per week peaked in the week 11–17 January 2021 and has been in decline since (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Number and percentage of staff vaccinated over time by ethnic group.

The figure shows the number of staff vaccinated (grey bars) and the cumulative percentage of the total number of staff of each ethnic group vaccinated (coloured lines) each week since the start of the University Hospitals of Leicester NHS Trust vaccination programme. It should be noted that the first and last time points do not represent complete weeks. There was 1 missing value for date of vaccination, and this observation is excluded from the figure.

https://doi.org/10.1371/journal.pmed.1003823.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Description of the cohort by vaccination status.

https://doi.org/10.1371/journal.pmed.1003823.t001

Factors associated with SARS-CoV-2 vaccine uptake

Table 2 shows univariable and multivariable (adjusted for age, sex, ethnicity, IMD, occupation, previous SARS-CoV-2 testing, and COVID-19-related work absences) logistic regression models for factors associated with vaccination against SARS-CoV-2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Univariable and multivariable analysis of factors associated with SARS-CoV-2 vaccine uptake.

https://doi.org/10.1371/journal.pmed.1003823.t002

After adjustment, factors associated with uptake of vaccination included older age (age group ≤30 years: aOR 0.48, 95% CI 0.44–0.53 [p < 0.001]; age group 51–60 years: aOR 1.19, 95% CI 1.07–1.31 [p = 0.001]; both compared to age 41–50 years) and male sex (aOR 1.24, 95% CI 1.15–1.35 [p < 0.001]).

HCWs from ethnic minority backgrounds were significantly less likely than their White colleagues to be vaccinated, an effect most marked in those of Black ethnicity (Black: aOR 0.30, 95% CI 0.26–0.34 [p < 0.001]; South Asian: aOR 0.67, 95% CI 0.62–0.72 [p < 0.001]). We found that vaccination uptake decreased with increasing deprivation (decrease in IMD quintile; test for trend p < 0.001).

In comparison to doctors, allied health professionals (aOR 1.40, 95% CI 1.20–1.62 [p < 0.001]), administrative and executive staff (aOR 1.48, 95% CI 1.32–1.66 [p < 0.001]), and healthcare scientists (aOR 1.69, 95% CI 1.41–2.02 [p < 0.001]) were all around 1.5 times more likely to be vaccinated. However, in a sensitivity analysis excluding those with locum/bank contracts (S6 Table), doctors were not less likely than other groups to be vaccinated and were, in fact, more likely than nurses/midwives/HCAs and estates/facilities staff to take up vaccination. Other significant findings remained unchanged.

Staff who had never undergone serology or PCR testing for SARS-CoV-2 were significantly less likely to have been vaccinated than those who had tested negative (serology: aOR 0.46, 95% CI 0.43–0.50 [p < 0.001]; PCR: aOR 0.70, 95% CI 0.64–0.77 [p < 0.001]). Staff with a history of a previous positive SARS-CoV-2 PCR result were significantly less likely to be vaccinated than those with only negative results (aOR 0.71, 95% CI 0.60–0.85 [p < 0.001]). To ensure that this effect was not simply due to these individuals not accessing vaccination due to recent COVID-19 infection, we conducted a sensitivity analysis excluding from the multivariable model individuals who tested positive for SARS-CoV-2 by PCR within the 28 days prior to the vaccine programme starting (n = 289). The significant findings remained unchanged.

Discussion

In this observational analysis in one of the largest and most ethnically diverse populations of HCWs in the UK, we found SARS-CoV-2 vaccine uptake to be significantly lower in those who were younger, female, or living in more deprived areas. We also found that ethnic minority HCWs were significantly less likely to take up vaccination than those of White ethnicity and that this difference was particularly marked for Black HCWs and certain South Asian HCW groups.

We provide real-life observational data from an entire hospital workforce demonstrating that SARS-CoV-2 vaccine uptake is lower in ethnic minority HCWs than those of White ethnicity. This finding aligns with the limited number of published studies conducted on vaccine uptake in HCWs to date [20,21] as well as recent survey studies on SARS-CoV-2 vaccine hesitancy in HCWs in the UK and US [17,31]. It also adds significant weight to emerging data in the general population which also suggest reduced uptake in ethnic minority groups. Our findings closely align with this population-level data as we also demonstrate that those of Black ethnicity were least likely to take up vaccination and that, amongst South Asian ethnic groups, those of Pakistani and Bangladeshi ethnicity were less likely to take up SARS-CoV-2 vaccination than those of Indian ethnicity [32].

Evidence on the specific barriers to COVID-19 vaccination in ethnic minority groups is limited [14]. However, when vaccine uptake is considered more broadly, factors such as a lack of trust in the government or in healthcare systems (e.g., due to unethical and non-ethnically heterogenous research practices in vaccine studies or structural and institutional racism), a lower perception of the risk of COVID-19 or a higher perception of the risk of side effects from vaccination, and other sociodemographic factors interrelated with ethnicity (educational level, socioeconomic status, and religion) have all been suggested as barriers to vaccine uptake [14,33–35]. Additionally, a recent nationwide survey study of UK HCWs determined that trust in employers and belief in COVID-19 conspiracy theories are important factors in predicting SARS-CoV-2 vaccine hesitancy [17]. Ethnic disparities in vaccine uptake are not unique to COVID-19; this phenomenon was also observed during the 2009 H1N1 influenza pandemic [36,37]. Regardless of the underlying mechanisms, these findings give significant cause for concern, as ethnic minority groups (especially those working in healthcare) are at higher risk of infection with SARS-CoV-2 and adverse outcomes from COVID-19, yet are not taking up this critical preventative intervention [3,5,9,11]. Previous work, including by the UK Scientific Advisory Group for Emergencies (SAGE), suggests that interventions aimed at overcoming barriers to vaccination in ethnic minority groups might involve multilingual, non-stigmatising vaccine endorsements from trusted sources (including information aimed at overcoming religious concerns about the vaccine) and community engagement (utilising trusted sources, e.g., general practitioners, within local communities to respond to concerns about vaccine safety and effectiveness) [14,17,38]. Implementing such interventions urgently is arguably of higher priority in HCWs than in the general population as HCWs represent an important source of health information for ethnic minority communities [14].

Alongside ethnicity, we also found deprivation to be associated with SARS-CoV-2 vaccination uptake, with those living in the most deprived areas being most likely to be unvaccinated. Deprivation has previously been shown to be associated with lower vaccine uptake in the general UK population [39], and, more recently, increased deprivation has been shown to be associated with lower SARS-CoV-2 vaccine uptake in UK HCWs [21]; this may be mediated through many of the same mechanisms discussed in relation to ethnicity above.

In accordance with previous studies [17,31], younger healthcare workers were less likely to be vaccinated than their older colleagues. A likely explanation for this finding is a reduced perception of personal risk of adverse outcomes from COVID-19. However, alongside the obvious greater risk of transmitting infection to more vulnerable individuals, long-term sequelae of COVID-19 (termed ‘long-COVID’), which may cause significant morbidity, have been demonstrated to be prevalent even in a young ‘low-risk’ population [40], suggesting that this cohort may still derive significant personal benefit from SARS-CoV-2 vaccination. A further explanation for this finding is that the vaccination programme was initially targeted at those with risk factors for severe COVID-19 (including those advanced in age), and thus older staff may have had more time and opportunity to be vaccinated compared to their younger colleagues.

We found that doctors were significantly less likely to take up SARS-CoV-2 vaccination than other staff groups (including allied health professionals). However, these findings should be interpreted with caution, as exclusion of individuals with locum/bank contracts resulted in higher uptake amongst doctors. It is possible, therefore, that locum doctors are not taking up vaccination through UHL. This may be due to limited access to trust communications or due to taking up offers of vaccination elsewhere. It should also be noted that this finding is in contrast to the results of 2 previous studies of SARS-CoV-2 vaccine uptake in UK HCWs, which both showed medical staff to be amongst the most likely to take up vaccination [20,21]. Estates and facilities staff also had lower levels of vaccine uptake than many other groups; support staff have been found to have low levels of vaccine uptake previously [20,21,35], and possible explanations for this observation in our cohort include limited access to the email communications regarding vaccination, as well as factors interrelated with occupational role such as educational level, deprivation, and ethnicity.

Though the numbers of pregnant HCWs in our analysis were small, we found that pregnancy was associated with lower odds of being vaccinated. Pregnant HCWs have previously been shown to be SARS-CoV-2 vaccine hesitant [17]. This finding may be a result of the advice given to pregnant women at the start of the vaccine rollout to delay vaccination until after delivery, on the basis that there was a lack of safety data in this cohort. However, there is now accumulating evidence that SARS-CoV-2 vaccination is safe during pregnancy and also that pregnancy increases the risk of severe COVID-19 [41,42]. Therefore, these findings have important implications for targeting catch-up vaccination programmes.

We also investigated the relationship of previous COVID-19 infection to SARS-CoV-2 vaccine uptake in the UHL population of HCWs. In accordance with a recent study [21], we found that those who were never tested for evidence of current/previous SARS-CoV-2 infection by swab or serology were more likely to be unvaccinated. This is unsurprising given that many of the barriers to vaccination (e.g., mistrust of the healthcare system, ‘needle phobia’, and low perception of personal risk from COVID-19) may also influence decisions about testing for evidence of current/previous SARS-CoV-2. Those with a history of SARS-CoV-2 PCR positivity were also less likely to take up vaccination. Some of this effect could be mediated by those who were isolating due to a positive swab having no access to vaccination, as well as advice from UHL that those with a positive swab in the last 28 days should avoid vaccination. However, exclusion of individuals testing positive within 28 days prior to the start of the vaccination programme did not change the result, implying the influence of other factors. It is possible that some of those who have had confirmed COVID-19 would be less likely to take up vaccination, believing themselves to have acquired sufficient immunological protection against SARS-CoV-2. This is likely to be true in the short term; however, risk of infection may increase with time since infection, given evidence concerning waning humoral immunity to SARS-CoV-2 and the short-lived immunity after infection with other coronaviruses [43,44]. Therefore, this group may represent an important group to target in subsequent SARS-CoV-2 vaccination drives.

This study has limitations. Although the population is large, data are from a single centre, affecting their generalisability. We only have vaccination data on those who were vaccinated through UHL. HCWs who obtained vaccination through primary care will be coded as unvaccinated in our analysis, although we expect these numbers to be small given that few other vaccination centres were in operation prior to the establishment of the vaccination hubs at UHL. We cannot predict whether HCWs who are currently unvaccinated will take up vaccination in the future; however, the numbers of staff taking up the vaccine over time are falling, implying that most who will accept vaccination have already done so. Furthermore, the predictors of vaccine uptake identified in our surveillance are similar to those in a recent UK-wide survey study of SARS-CoV-2 vaccine hesitancy in HCWs [17] and 2 UK-based studies on SARS-CoV-2 vaccine uptake [20,21], providing reassurance about the generalisability of our results. SARS-CoV-2 PCR testing has been available at other non-UHL centres, and PCR results from HCWs accessing testing via these centres were not available within UHL records; however, given the convenience and availability of PCR testing within UHL, it is likely that the vast majority of staff would have accessed testing via this route. There are other factors that may influence vaccine uptake (e.g., past medical history and educational level) on which we do not have data as we felt this was beyond the scope of an audit, and therefore we cannot adjust for these in our analysis. We are unable to determine which vaccine (Pfizer–BioNTech or Oxford–AstraZeneca) was offered to HCWs, only whether or not they were vaccinated. Therefore, we are unable to determine whether vaccine uptake is associated with the particular vaccine that is offered. This might form a focus of future studies. Despite these limitations, our work has many novel findings that will be of direct relevance to policymakers involved in designing SARS-CoV-2 vaccination programmes, particularly in light of the potential booster vaccination programme for UK HCWs [45].

In summary, we have found that in a population of UK HCWs, those from ethnic minority groups and from more deprived areas, as well as those who are younger, female, or from particular occupational groups, are less likely to take up SARS-CoV-2 vaccination. These findings have major implications for the effective ongoing delivery of SARS-CoV-2 vaccination programmes, including planned booster vaccinations, both in HCWs and in the wider population. Urgent actions should be taken to boost vaccine uptake in the identified groups and to prevent the disparities caused by the COVID-19 pandemic from being allowed to widen further.

Supporting information

Acknowledgments

Tony Doherty and Michael Dobson assisted with data collection. Matt Archer was involved in establishing the UHL vaccination programme.