Strengthening Polio Vaccine Demand in Ghana: Understanding the Factors Influencing Uptake of the Vaccine and the Effectiveness of Different Message Frames

"Ghana has experienced recent polio outbreaks. Behavioral insights can be used to understand behavior and create demand for the polio vaccine."

The first type 2 novel oral polio vaccine (nOPV2) vaccine campaigns started in Ghana in August 2022. This new vaccine is less likely to revert into a form that can cause paralysis than the OPV that has previously been used to fight the wild poliovirus. The introduction of new vaccines can be challenging due to public questions and concerns, which can potentially lead to vaccine hesitancy and even refusal. This cross-sectional study explores factors influencing the uptake of the polio vaccine among Ghanaian mothers with children younger than five years old, as well as their intention to vaccinate in the future. It also includes an experiment with short polio vaccine voice message nudges to identify the most effective message frames for encouraging vaccination.

The Behavioral Drivers Model (BDM) was used as a conceptual framework for the study. The BDM categorises factors influencing behaviour into individual factors, social factors, and environmental factors, which in turn are divided into a number of dimensions. The use of social norms to change behaviour has been long established. Descriptive social norms are based on a principle that most people want to bring their behaviour in line with what they perceive to be the behaviour of others, whereas injunctive social norms tell people what is socially acceptable. Only a few vaccination-related message framing studies have been conducted in Africa, including Nigerian research demonstrating that using a social-norm-related messaging frame in the context of good parenting increased the intention of parents to immunise their children. Understanding how different types of frames may influence vaccination intentions may facilitate development of an effective vaccine demand creation plan.

The United Nations Children's Fund (UNICEF) together with its partners Ghana Health Services (GHS), Cogent, and VIAMO conducted an interactive mobile phone-based survey, based on an interactive voice response (IVR) system, from October to December 2021. Part one of the survey measured various drivers of vaccination behavior and intentions, wherein respondents rated their perceptions on each of these drivers. The BDM was used as a conceptual framework for the predictor development. Part two of the survey focused on measuring the effectiveness of four short audio message nudges against a control message on two dependent variables: willingness to give the polio vaccine to children and willingness to recommend the polio vaccine to others. The nudges represented different conditions - namely, social norms, perception of safety, communicating adoption (messenger message), cautionary (fear message), and a fact-based message (control).

Data from 708 caregivers (mothers with children younger than five years old from all 16 administrative regions of Ghana) were assessed. Of the sample, 35% (n = 250) had not vaccinated their children against polio, around 8% (n = 53) stated they did not plan to do so, and 28% expressed intent to do so during the next polio vaccination campaign. Using the linear probability model (LPM), the researchers report on the effects of each category of factors within the BDM:

• Psychological factors: Higher vaccination of children against polio, i.e., better uptake of the polio vaccine, appeared to be associated with children's caregivers knowing that polio causes paralysis (with a coefficient of 0.13 (95% confidence interval (CI): 0.02, 0.24), i.e., 13% more likely than not to have their child vaccinated). Higher vaccine uptake also appeared to be associated with the perception that the polio vaccine is safe (with a coefficient of 0.11 (95% CI: 0.01, 0.22), i.e., 11% more likely than not to have their child vaccinated).
• Sociological factors: Respondents who perceived support from healthcare workers were 11% more likely to vaccinate their children than those who either did not perceive the support or were not sure about it, with statistically significant coefficients of 0.11 (95% CI: 0.02, 0.20).
• Environmental factors: Difficulty accessing the polio vaccine appeared to be associated with a negative change in vaccine uptake (with a coefficient of -0.16 (95% CI: -0.23, -0.08), i.e., 16% less likely to have their child vaccinated). Satisfaction with the information provided by vaccinators was also associated with better vaccine uptake (with a coefficient of 0.12 (95% CI: 0.05, 0.20) i.e., 12% more likely than not to have their child vaccinated). Notably, those who had seen or heard something negative about the polio vaccine were 10% more likely - with a statistically significant coefficient of 0.10 (95% CI: 0.03, 0.17) - to have vaccinated their child than those who had not heard anything negative about the polio vaccine.

Looking at the subsample of caregivers who had not yet vaccinated their child against polio (N = 250, 35% of the full sample) but who expressed an intent to do so, one sociological factor - receiving support from healthcare workers - was found to have a statistically significant effect, with a coefficient of 0.12 (95% CI: 0.01, 0.24). Likewise, one environmental factor - being satisfied with the information provided by vaccinators - was found to be statistically significant, with a coefficient of 0.12 (95% CI: 0.01, 0.23). This means that those who received support from health workers and those who felt satisfied with the information provided by vaccinators were each 12% more likely to intend to vaccinate their child in the next polio vaccination campaign.

The effectiveness of messaging frames was measured in the full sample (N = 708) by the proportion of respondents reporting their intention to give the polio vaccine to their child or intention to recommend the polio vaccine to others. Table 5 in the paper shows the results of this assessment, which indicates that only the social norms message frame was statistically significant, with a coefficient of 0.06 (95% CI: -0.004, 012). Thus, messaging does not appear to affect intention to vaccinate, which is consistent with the finding that, out of those who say that their children are not vaccinated (N = 250), a majority (N = 197) say they already intend to vaccinate their children.

In order to assess the effectiveness of nudge messages against vaccine hesitancy, the nudges were also tested in a truncated sample of those caregivers who had not vaccinated their children and also stated they do not intend to do so in the future (N = 53). Select findings:

• Both the social norms frame ("Most people in your community are getting their children vaccinated against polio. Get your child vaccinated against polio!") and safety frame ("Polio vaccines are safe. Get your child vaccinated against polio!") were 45% more likely than the control message to make respondents report intent to give polio vaccine to their child than those hearing the control frame ("Protect your child against polio. Get your child vaccinated against polio!").
• Those who heard the messenger frame ("Polio vaccines are recommended by health professionals. Get your child vaccinated against polio!") were 42% more likely to get their child vaccinated than respondents who heard the control frame.
• Only the fear message frame ("Polio causes paralysis and sometimes death. Get your child vaccinated against polio!") appears to have an effect in predicting whether a respondent would recommend the polio vaccine to others. It has a statistically significant coefficient of 0.37 (95% CI: 0.02, 0.72), which means that the fear message was 37% more likely than the fact-based control message to compel people to recommend vaccination to others.

Overall, the findings indicate that:

• Vaccine demand creation in Ghana should focus on communicating the safety of the vaccine, as well as the negative consequences of poliovirus (namely, paralysis).
• Healthcare workers play a central in vaccine demand generation in Ghana, where these personnel appear to be strongly preferred and trusted information sources for polio- and vaccine-related information. Efforts should be made to build their capacity to communicate polio- and vaccine-related messages.
• Polio campaigns in Ghana have been typically implemented by vaccinators making house-to-house visits, yet this study revealed that difficulties to access the vaccine were considered as a barrier to taking the polio vaccine. It would be important to investigate what kind of logistical problems are linked with house-to-house vaccination, including the quality of the visits.
• The study also pointed out that mothers perceived the permission of a male member of their family to vaccinate their children as a barrier, which highlights the need to include male members of the family in the vaccine demand activities. Evidence-based and multi-faceted communication campaigns that take a whole society approach have been found effective in addressing male members of families with polio messaging.
• All the message frames tested should be considered in encouraging (i.e., "nudging") people to vaccinate their children against polio, but the use of a norm-based frame may be a particularly good fit with the cultural context of Ghana, where people emphasise communal values such as family, respect for the elderly, and honour in traditional rulers. Future research could test different types of social norm messages (e.g., "mothers like you are increasingly vaccinating their children.").

The researchers indicate that use of an automated mobile-phone-administered survey based on voice response (IVR system) generally worked well. However, although Ghana has the highest mobile penetration in West Africa, data shows that women in low- and middle-income countries are 40% less likely to use a cell phone with internet access than men. These women are often the most vulnerable ones in society. Field-based face-to-face surveys could be conducted in the future to compensate for the gap.

In conclusion, the behaviour and the intention to vaccinate are both driven by a number of factors that must be addressed to create demand for the polio vaccine. More research is needed to understand how targeted message frames impact vaccine behaviour and future intention for vaccination.