Sustainedly high levels of COVID-19 vaccine acceptance in five Sub-Saharan African countries

In a previous blog post, we reported the results of a recent working paper and corresponding journal article finding high acceptance rates for an eventual COVID-19 vaccine in a sample of six Sub-Saharan African countries surveyed between September and December 2020. Since then, vaccine availability has become a reality, and vaccination campaigns are now underway in all countries covered. With evidence collected mostly in high-income countries suggesting that vaccine acceptance may vary substantially over time, we wondered: How stable have the high levels of COVID-19 vaccine acceptance in Sub-Saharan Africa been now that acceptance needs to translate into active demand?

To answer this question, we draw on a follow-up round of the vaccine module that is part of the LSMS-supported High-Frequency Phone Surveys (HFPS), collecting comparable cross-country data on, amongst others, vaccine attitudes in Burkina Faso, Ethiopia, Malawi, Nigeria, and Uganda. The data were collected between February and June 2021 and allow us to track vaccine acceptance overtime for a panel sample of individual survey respondents. As such, they now allow us to compare vaccine acceptance rates before the availability of vaccines to when the start of vaccination campaigns was imminent or had recently started. Sample selection and representativeness of the HFPS are discussed in more detail in the paper referenced above.

Vaccine acceptance remains high with variations over time and across countries

Our headline results indicate that acceptance rates in this sample remain generally high and above 80 percent in three countries (Ethiopia, Uganda, Nigeria), with a reduction in vaccine acceptance compared to the previous survey round in Malawi and an increase in Uganda. We find some decreases in point estimates of vaccine acceptance rates in Burkina Faso, Ethiopia, and Nigeria, but these are not significant at the conventional 95% level.Note: Estimated acceptance rates of a COVID-19 vaccine for individuals interviewed at wave 1 and wave 2 (i.e., panel individuals). In Burkina Faso and Malawi, the questionnaires in wave 2 included a response option for “not being aware” of the availability of a vaccine in lieu of asking for the willingness to be vaccinated. The share of respondents for which this is the case is negligible in Malawi (1.8% of the sample) but substantial in Burkina Faso (47% of the sample). We  exclude these respondents from the analysis since it would not be appropriate to infer anything form this response about vaccine acceptance or hesitancy. Full documentation and data access are available here and here.

Besides looking at average acceptance rates, the fact that the survey tracks respondents across waves offers us an opportunity to study individual respondents’ (as opposed to aggregate or household-level) attitudes over time. The alluvial charts below allow for several conclusions. First, vaccine attitudes seem to be mostly stable, with the majority of respondents willing to be vaccinated maintaining their stance. However, we also observe a number of individuals changing attitudes from willing to hesitant and vice versa. Specifically, there is little switching of attitudes in Ethiopia (7% of panel individuals), Nigeria (14%) and Uganda (15%), whereas a larger share of panel individuals change their stance in Burkina Faso (18% *) and Malawi (28%). In line with the aggregate changes in estimated vaccine acceptance rates, we observe that, among those changing attitudes, more people switch from willing to hesitant in Ethiopia (65%), Burkina Faso (61%), Malawi (58%), and Nigeria (55%) whereas the opposite is true in Uganda (36%). Overall, this suggests that trust in vaccines can be lost as well as gained. Therefore, there is value in and need for ongoing public information and communication campaigns that ensure continued support for national vaccination efforts.

_{* For Burkina Faso, we count respondents declaring to be unaware of the availability of a vaccine (instead of being asked about their attitude) in wave 2 as not switching vaccine attitudes. As such, the figure is a lower boundary and may be higher if some among them actually changed their attitude.}

_{Note: Burkina Faso and Malawi added response options for “not being aware” of the availability of COVID-19 vaccines in the country and did not ask thus who are “unaware” of their general attitude toward vaccines. This does not necessarily imply hesitancy or acceptance and as such we plot this as a distinct answer option.}

Some clustering of vaccine attitudes across demographics and space

Regarding the correlates of vaccine hesitancy, controlling for a number of individual and household characteristics in a series of multivariate logit regressions, the data indicate that vaccine hesitancy is higher in urban areas, among women, and in richer households. Furthermore, we find that internet access is associated with higher levels of vaccine hesitancy, a pattern that is stronger for Wave 2 of our data when the rollout of vaccination campaigns was imminent. This may indicate greater exposure to online misinformation about vaccines which a study by the Africa CDC found to be correlated with hesitancy, though we cannot explore this hypothesis with the available data. When it comes to why people are hesitant to get vaccinated, our respondents continue to report concerns around the safety and side effects of the vaccine as the primary reservations.

Motivated by the discrepancy in vaccine attitudes between urban and rural areas, we also explore the clustering of vaccine attitudes across geographical space. Early results from visual and regression analysis suggest a positive correlation of vaccine attitudes for people living in the same area, possibly pointing to network effects in the transmission of vaccine attitudes. This appears to be consistent with higher hesitancy among wealthier household which tend to clusters in networks. However, this is as yet only a hypothesis that upcoming work will have to explore further.

A sustained effort to turn high levels of vaccine acceptance to high levels of coverage is needed

Our findings offer reassurance that acceptance of COVID-19 vaccines in the five Sub-Saharan African countries we analyze data for remained high as vaccine rollout got underway. At the same time, they indicate that vaccine attitudes can change over time. Therefore, they caution that a sustained effort is required to ensure that vaccines' availability is met by active demand. As such, making sure that confidence in the safety and desirability of vaccination against COVID-19 stays high will remain important as supply side and infrastructural constraints are resolved – so that high levels of vaccine acceptance can turn into high levels of vaccine coverage. In upcoming survey rounds, we thus plan to put special emphasis on gauging active demand for vaccines, analyze barriers of access, and delve deeper into intra-household and community dynamics in vaccination uptake.

Resources

• RSVP and add to Outlook Calendar, Tuesday, December 7, 2021 | 11:00-12:30 PM EST
  Development Data Group Learning Series:
  Seminar on Attitudes toward COVID-19 vaccines in Africa and beyond: How can survey data support global vaccination efforts?

• Read the blog presenting the results of wave 1 of the survey and associated working paper

• Find out more about the LSMS-supported COVID-19 High-Frequency Phone Surveys