Understanding COVID-19 Vaccination Behaviors and Intentions in Ghana: A Behavioral Insights (BI) Study

"What drives COVID-19 vaccine intentions does not necessarily drive behaviors. The results of this study can be used to develop appropriate COVID-19 vaccine uptake strategies targeting the most important drivers of COVID-19 vaccine acceptance, using effective message frames."

Studies in Ghana show that a significant percentage of the adult population have been hesitant to take the COVID-19 vaccine. In response, Ghana Health Services (GHS) deployed multiple vaccine demand creation strategies, including community outreach through community leaders, social media, and mass media. They have also responded to circulating rumours and debunked misinformation. Behavioural insights (BI) can be used to address vaccine hesitancy by understanding the various factors that influence the decision to take or refuse a vaccine. This two-part study used BI to better understand factors that influence COVID-19 uptake in Ghana and tested the impact of differently framed nudges on the willingness of people to take the vaccine.

BI involves the study of human behaviour. It draws on empirical research in fields, including economics, psychology, and sociology, and uses these insights to design and make discrete changes in the environment to impact behaviours. For example, BI can uncover behavioural drivers by experiments and surveys using behavioural analytics to identify the relative influence and variety of factors that draw people to or away from vaccinations. BI can then create strategies to improve decision-making, called nudges, which alter how choices are presented, leading decision-makers to behave in predictable ways. In addition, BI can encompass strategies that, unlike nudges, require sustained effort. For example, implementation scientists have long recognised that social motivation, incentives, and rewards are crucial levers of behaviour change.

This study consisted of two components, which were both embedded in a survey administered to respondents from Ghana between December 2021 and January 2022. (COVID-19 vaccines, including the AstraZeneca/Oxford and Pfizer-BioNTech vaccines, were available starting on February 24 2021.) Working in partnership with Viamo, the researchers collected data from 1,494 participants, 1,089 (73%) of whom reported already being vaccinated and 405 of whom (27%) reported not being vaccinated yet. Viamo administered the survey via interactive voice response (IVR), an automated system that delivers pre-recorded voice messages to people over a mobile phone, in any of six generally spoken languages in Ghana.
 

• Part I of the survey was designed to identify the influence of various drivers of vaccine acceptance. Eight predictors were examined. These predictors measured various drivers of vaccination behaviour and intentions, wherein respondents rated their perceptions on each of these drivers.
• Part II of the survey included an experiment to test which of several behaviourally informed message frames had the greatest effect on vaccine acceptance. The BI message types were recorded by voice actors and were presented in the language chosen by the participant at the beginning of the survey. Respondents were randomly assigned to one of six possible messages: fear framing, altruism framing, social norms framing, GHS as the messenger, a doctor as the messenger, and religious leaders as messengers. Three dependent variables were also examined: (i) respondents' willingness to recommend the vaccine to family and friends; (ii) their willingness to share the benefits of vaccination with their family and friends; and (iii) their intention to get vaccinated (measured again after presenting the BI message).

Both components align with the Define, Diagnose, Design, and Test (DDDT) methodology used in BI and were developed in collaboration with UNICEF.

The findings indicated that vaccine uptake in Ghana is influenced more by social factors (what others think) than by practical factors such as ease of vaccination.

With regard to vaccination behaviour: To test the influence of the eight potential drivers (tested in the first part of the survey) on respondents' vaccination behaviour, a logistic regression analysis was conducted. The model explained 7.56% of the variance in respondents' vaccination behaviour. Of the eight drivers, family's attitudes and religious leaders' attitudes significantly predicted vaccination behaviour in a positive direction; when respondents perceived family and religious leaders to have favourable attitudes towards the vaccine, they were more likely to get vaccinated. However, perceptions of healthcare providers' attitudes had a significant negative relationship with respondents' vaccination behaviour. When respondents perceived healthcare providers to have positive attitudes towards the COVID-19 vaccine, they were less likely to get vaccinated. (One's own attitude, risk perception, effectiveness, ease of vaccination, and perceptions of the community's attitudes toward the COVID-19 vaccine did not significantly predict vaccination behaviours.)

With regard to vaccination intention: It was positively predicted by risk perception, ease of vaccination, and the degree to which respondents considered the vaccine effective. Perceptions of religious leaders' attitudes also significantly and positively predicted respondents' intention to get vaccinated. Although perceptions of religious leaders' views about the vaccine are an important driver of vaccine acceptance, results asking respondents to rank-order who influences them suggest that people may not be consciously aware - or do not want to admit - the degree to which they are affected by what religious leaders think. When it comes to forming intentions to get the COVID-19 vaccine in Ghana, perceptions of healthcare practitioners' attitudes are significantly less important than nearly every other variable in the model: risk perception, ease of vaccination, own attitudes, family's attitudes, community's attitudes, and religious leaders' attitudes.

Whereas a large proportion of the vaccinated respondents believed that the decision to take the vaccine was influenced by their own attitude, a good number of unvaccinated believed that their intention was influenced by the community and family around them. This finding suggests that social influence, which is the process by which perceptions of what other people think and do influence beliefs and behaviours, should be considered carefully in COVID-19 vaccine demand generation strategies in Ghana. As social norms define what is acceptable in the given context, it is important to gain the trust of family members and religious leaders as advocates for COVID-19 vaccines.

In addition, religious leaders can be enlisted in community mobilisation and engagement activities, as they interact with diverse communities of Ghana, which include more than 70 ethnic groups, and they are seen as informal liaisons between local communities and state institutions. The role of religious leaders extends to political, economic, educational, religious, and family life, which makes them promising messengers to family members. Health authorities should build the capacity of religious leaders to become strong advocates for COVID-19 vaccines as well as to establish a network for social listening purposes to be able to assist religious leaders when community members have information voids and concerns regarding the vaccine.

As discussed previously, in part II of the study, respondents were randomly assigned to one of the six BI message types. After listening to the assigned BI message, they were asked to report their willingness to recommend the COVID-19 vaccine to family members and friends. Between 85-93% expressed such a willingness, with slight variations across BI message types. Message frames that included fear, altruism, social norms were all followed by positive responses toward the vaccine, as were messages with three distinct messengers: GHS, a doctor, and religious leaders. The effectiveness of the BI message types was similar for both vaccinated and unvaccinated respondents.

Notably, the sample may have been biased, as 75% of the respondents reported having received at least one vaccination, while, according to GHS, only 22% of the population had received the first dose of the vaccine by end of January 2022. One explanation for this discrepancy could be that the mobile phone survey database was based on people registered to receive health messages, that they may be more interested in health interventions including vaccinations than the general population. The sample may also have been biased because the majority of the respondents were young male from rural areas.

According to the researchers, this study highlighted the usefulness of behavioural analytics, which identified unconscious drivers affecting vaccine uptake and intentions by comparing it with respondents' conscious perceptions of factors influencing the uptake of the vaccine. Also, the  study demonstrates that the use of mobile phone-based surveys may prove to be an efficient methodology to reach people from rural areas, who tend to be underrepresented in many studies examining behaviour. 

The findings of the study are expected to result in the development of interventions that aim to encourage uptake of COVID-19 vaccines. Future research should focus on measuring the impact of these interventions.