Choice preference and willingness to pay for COVID-19 vaccination in Namibia

Highlights

• •The study assesses vaccine uptake behavior of the general population in Namibia.
• •The study uses stated choice experiment and latent class discrete choice models.
• •Provides insights into preference for vaccine characteristics on uptake.
• •Provides insights into psychological factors driving vaccine hesitancy and risk perceptions.
• •Enumerates willingness -to-pay and willingness-to-wait measures for vaccines.
• •Provides insights for government strategy to promote future vaccine uptake.

Abstract

Background

Namibia has not been spared from the coronavirus (COVID-19) pandemic, and as intervention the Namibian government has rolled out vaccination programmes. This study was conducted before the roll out of these vaccines to assess the preference for COVID-19 vaccinations. Stated preference studies provide information about social demand, access, willingness-to-pay and financing for future COVID-19 vaccination.

Methods

A stated choice experiment (SCE) survey was administered to a sample of 506 participants from Namibia's general population between October 2020 and December 2020. Participants were asked to make a series of hypothetical choices and estimate their preference for different attributes of a vaccine. A latent class model was used to analyse the SCE data. The study also assessed anti-vaccination behaviour, past vaccination behaviour, impacts of COVID-19 on mental and physical health and Willingness-To-Pay (WTP) measures. The WTP measures were captured as out-of-pocket and further calculated using the marginal rate of substitution method in SCE.

Results

Data from 269 participants was included in the analysis. Vaccine side effects (40.065), population coverage (4.688), payment fee to receive vaccine immediately (3.733) were the top three influential attributes for vaccine preferences. Accordingly, increases in mild and severe side effects of vaccine options had negative impacts on utility; with an average WTP of N$728.26 to reduce serious side effects. The average WTP to receive a high-quality vaccine with 90% efficient was found to be N$233.11 (US$15.14). Across classes, there was a strong preference for vaccines with high effectiveness over longer durations of time.

Conclusions

The results provide useful information for the Namibian government to improve the current strategies for vaccine rollout interventions.

1. Introduction

The coronavirus disease 2019 (COVID-19) which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has reached all countries and devastated the livelihoods of people, including Namibia. The socio-economic impacts of the COVID-19 pandemic are varied and amongst many, include: a reduction of routine child and maternal services [39], additional child and maternal deaths [8], increased food insecurity particularly in low and middle-income countries [4], reduced services from routine essential health services [1]; and increased morbidity and mortality from other diseases due to diversion of resources [7]. The impacts of the COVID-19 pandemic on health and development is expected to last for a prolonged period [30].

Namibia has been hit hard by COVID-19 starting from the second wave. By the third wave which was the worst, the country’s daily new cases peaked at approximately 2500 during the month of July 2021 (worldometers.com). Consequently, the government of Namibia, through the Ministry of Health and Social Services (MoHSS) applied different interventions in place, including a total lockdown of the country during the third wave [6], [32]. To date, Namibia reported 171,000 positive cases and 4,090 deaths. Immunization against COVID-19 has been prioritized to minimize the impact of the pandemic, however there has been a general low vaccine uptake in the country. The MoHSS projected to vaccinate a target population of 1,501,041 (60%) of the total population to reach herd immunity [31]. Although vaccination started around March 2021, just over 500 000 Namibians have been vaccinated with the first dose thus far representing 33% of the population (worldometers.com).

Immunization is one of the most cost-effective and successful prevention interventions for infectious diseases [36]. While some studies have shown a willingness to receive COVID-19 vaccination amongst groups such as healthcare workers, students and pregnant women [10], [35], [40], others show mass uncertainty regarding vaccines and the influence of modern political movement [5], [11], [34]. In particular, the fast-tracked vaccine development and approval process for COVID-19 vaccines known as Operation Warp Speed is perceived to be for political gain rather than science [27]. The scepticism about the safety of the COVID-19 vaccine [37]; and other social factors such as the lack of knowledge leading to misinformation and rumor mongering about COVID-19 [33], [40] have huge implications for coronavirus vaccine acceptance [13]. Vaccine refusal and delays are contributing to an increasing number of vaccine-preventable diseases outbreaks globally. For this reason, WHO named vaccine hesitancy as one of the top ten threats to global health in 2019 [12]. Thus, studies on the preference of COVID-19 vaccines are considered vital to inform governmental strategies for vaccine take-up.

A number of studies have looked at the acceptance and willingness-to-pay for COVID-19 vaccines in African countries [3], [26], [42], and in particular a few have focused on Namibia [47], [48]. However, none of these studies have considered trade-offs between vaccine attributes. Hess et al, (2022) employed advanced discrete choice models on stated choice data collected from 18 countries including Namibia, which compared the influence of vaccine attributes across countries [46]. Other applications of SCE in Namibia have been applied to mode choice in freight transport [2], [28].

The present paper considers the trade-offs between COVID-19 vaccine attributes, that people make when choosing to be vaccinated, and associate factors that drive vaccine uptake. The study employs stated choice surveys and discrete choice models to assess the choices that people make when faced with a discrete set of alternatives [21]. The study furthermore employs the models developed to compute willingness-to-pay (WTP) and willingness-to-wait (WTW) measures for different aspects of the vaccines.

2. Methods

2.1. Survey design

The study employed a Stated Choice Experiment (SCE) design that was developed by the Centre for Choice Modelling [46] and adopted locally by the authors. The objective of the SCE was to find value of the attributes (or characteristics) of a product or service, where the product or service comprises several attributes [43]. Participants were faced with several hypothetical vaccination choice scenarios. They were asked to imagine a situation where several vaccines for COVID-19 have been developed and have undergone all required testing and received regulatory approval for use in humans from the health authorities. Participants were informed that vaccination protects those who received it against infection and against passing the virus on to others. In addition, participants were informed that while vaccination cannot completely eliminate the risk of contracting the virus, it would also reduce the risk of serious illness should a vaccinated person become infected.

The choice design presented respondents with two sets of six choice sets of each choice task, where there were two hypothetical profiles of unknown vaccines against COVID-19 (labelled as Vaccine A or Vaccine B). For each choice set, respondents had three options: to indicate their preferred vaccine profile between Vaccine A or B; or remain unvaccinated. The details of the attributes and levels are displayed in Table 1. An example choice scenario is shown in Fig. 1.

Table 1.

Levels used in experimental design for SC scenarios.

Attribute	Level 1	Level 2	Level 3	Level 4	Level 5	Level 6	Level 7
Risk of infection out of 100,000 people	500 (0.5%)	1,500 (1.5%)	3,000 (3%)	4,000 (4%)	5,000 (5%)	–	7,500 (7.5%)
Risk of serious illness:	2,000 (2%)	4,000 (4%)	6,000 (6%)	10,000 (10%)	15,000 (15%)	–	20,000 (20%)
Estimated protection duration	five years	two years	one year	6 months	Unknown	–	–
Population coverage	More than 80%	60%	40%	20%	Fewer than 10%	–	–
Risk of mild side effects out of 100,000 people	100 (0.1%)	500 (0.5%)	1,000 (1%)	5,000 (5%)	10,000 (10%)	–	–
Risk of severe side effects out of 100,000 people	1 (0.001%)	5 (0.005%)	10 (0.01%)	15 (0.015%)	20 (0.02%)	–	–
Exemption from international travel restrictions	no restrictions	no exemptions	–	–	–	–	Restrictions on international travel
Waiting time (for free option)	2 weeks	1 month	2 month	3 month	6 months		–
Fee (for paid option)	N$186	N$464	N$928	N$1,624	N$2,088	N$3,931	–

Fig. 1.

Example of SC scenario.

2.2. COVID-19 specific questions

Following completion of the SC scenarios, several questions were used to collect additional information on preferences in relation to COVID-19 vaccination. These questions related to the preferred location for being vaccinated (for example hospital vs special COVID-19 vaccination centres), and the preferred person for administering the vaccine (for example doctors vs nurses). This was followed by questions relating to the reason for vaccination, for example the relative importance of protecting oneself vs protecting others, and the influence of recommendations received by friends and family, medical experts, and politicians. The same respondents were next faced with a willingness-to-pay question, where they were asked how much they would be willing to pay to avoid a six-month waiting period for a COVID-19 vaccine with desirable characteristics (over 90% efficacy, five-year protection, and low side effects). Separate willingness-to-pay levels were also captured for the same vaccine with and without vaccines providing exemptions from travel restrictions. Finally, using a 5-point Likert scale, respondents were asked how likely they thought it was that an infected person a) develops symptoms; b) develops a serious illness that does not require hospitalization; and c) develops a serious illness that requires hospitalization, and c) dies.

2.3. Socio-demographics, risk exposure and past vaccination choices

Key socio-demographic measures were captured, including gender, age, income, ethnicity, education, and employment status (and change therein since the pandemic start). Respondents were also asked to indicate if they had previously been vaccinated against influenza, whether they have had any other elective vaccines, and if they have had all recommended vaccines. Finally, they were asked for a personal rating of their physical and mental health (separately), and how these had changed since the start of the COVID-19 pandemic.

2.4. Ethical approval and data collection

Ethical approval was obtained from the Ministry of Health and Social Services (MoHSS; ref: 17/3/3/AK). The survey was conducted across all the 14 regions of Namibia between October-December 2020. The questionnaire was administered through Qualtrics (https://www.qualtrics.com) and presented in all the major languages of Namibia including: Oshiwambo, English, Afrikaans, Otjiherero, Khoe-khoe (Damara/Nama), Silozi and Rukwangali. Survey participants were recruited through Ndatara Surveys (https://www.ndatara.com). Participants were recruited via social media platforms (online) or through field interviewers (CAPI). To encourage participation, the main incentive for participating was a prize draw given to four lucky participants, while each participant recruited was also awarded a small token for their time after completing the survey. Only participants older than 18 who could consent took part in the study.

3. Data analysis

3.1. Initial data analysis

The core of the analysis was concerned with modelling the choices from the SC component of the survey, however prior to this, we undertook a detailed analysis of other parts of the data. We analysed the responses for questions used to gauge respondents’ views in relation to the risk of COVID-19 infection, past vaccination behaviour, as well as the impact COVID-19 has had on their physical and mental health.

3.2. Model estimation

Given the likely high levels of heterogeneity in preferences in the sample, we used Latent Class (LC) models (cf. [25]) to analyse the data, where the discrete choice model in each class was of the Multinomial Logit (MNL) type [22]. In general terms, for person n in study area c, we write the deterministic component of utility for alternative i in choice scenario t and in class c as:

$$V_{\mathit{ncjt}}={\delta }_{\mathit{cj}}+{{\beta }^{\prime }}_{\mathit{cj}}{X}_{\mathit{ncjt}}$$ (1)

where ${\delta }_j$ is an alternative specific constant (ASC) used in class c for alternative j; ${\beta }_c$ is a vector of coefficients to be estimated and $X_{\mathit{ncj}}$ the vector of attributes of alternative j in choice scenario t faced by respondent n. Some differences arise across alternatives and across attributes as follow:

• •For the constants, we used the no vaccine option as the base, normalising its constant to zero. Separate constants were estimated for free and paid vaccine options, along with an effects coded position constant to distinguish between the left and right vaccine in the survey.
• •For the no vaccine option, the only attributes that entered the utility function were the risk of infection and the risk of illness, using the baseline levels from Table 1.
• •After initial tests for non-linearity, all attributes were treated as continuous, with two exceptions: For protection duration, a separate term was estimated for unknown protection duration, alongside the continuous term for known durations, while the travel exemption attribute, which has only two levels, was dummy coded (with no exemption as the base).

The LC model was estimated using maximum likelihood routines in Apollo v0.2.5 [23]. No weighting was used in estimation, and the results were instead reweighted after estimation.

3.3. Willingness-to-pay calculations

A very important concern that face vaccine uptake is whether the public is willing to accept and willing to purchase the vaccine [18]. In this study, two MRS measures of willingness-to-pay (WTP) and willingness-to-wait (WTW) were computed substitution from the ratio of the parameter of interest over the parameter for payment fee or waiting time [29]. For example, taking the relative importance of risk of mild side effects on waiting time, the MRS in class s is:

$${\mathit{MRS}}_{s,mildvswait}\left(\theta \right)=\frac{{\beta }_{\mathit{mildsideeffects},s}}{{\beta }_{\mathit{waitingtime},s}}$$ (2)

Subsequent to this, another value of WTP was captured from the survey through the iterative bidding method of the out-of-pocket expense. The survey asked respondents to specify their willingness-to-pay (WTP) for two paid vaccination options. The first option was to avoid a six month wait for a high-quality vaccine (over 90% efficacy, five-year protection, and low side effects). The second option was to pay for a vaccine to avoid travel restrictions. This was done to find the hypothetical value of WTP that respondents are willing to pay for a vaccine [24]. The mean WTP was then calculated by summing the bidding prices, and averaging them across the sample.

4. Results

4.1. Participant demographics

The survey registered 506 complete responses, of which 212 were male, 281 females; and 13 preferred not to state their sex (Fig. 2, Table 2). Sample sizes varied substantially across regions, with the majority of the respondents from Erongo (21.3%; n = 120), Oshana (26.9%; n = 137), Khomas (14.9%; n = 76) and ||Karas (22.7%; n = 115). The age group between 18 and 30 years old made up 56% of the respondents. Compared to the overall Namibian population, the resulting data was generally balanced in relation to age and gender (Table 2).

Fig. 2.

Survey respondents in the 14 regions of Namibia.

Table 2.

Distribution of age in sample versus population (source: Namibia Statistics Agency).

Female		age 18–30		age 31–50		age 51 and over
Data	Adult pop	Data	Pop	Data	Pop	Data	Pop
54.01%	52.7%	56.13%	41.0%	38.34%	39.3%	5.52%	19.7%

4.2. Anti-Vaccination behaviour

Due to the choice options presented in the survey (see Fig. 1), the overall vaccine uptake in the study was made of three groups, namely: respondents who always choose a vaccine across the six SCE scenarios, respondents who choose a vaccine in some but not all scenarios, and respondents who never choose a vaccine (vaccine hesitant), where this is not due to the characteristics of the vaccines presented to them. Table 3 shows the share of vaccine uptake in the sample, and reveals that individuals who always choose a vaccine, regardless of the characteristics of the vaccine, formed the largest group at 58.7%, while the anti-vaccination group formed 18.8% of the sample.

Table 3.

Share of Vaccine uptake in the sample.

Overall vaccine uptake	72.0%
Share likely to accept any reasonable vaccine	58.7%
Share open to vaccination depending on characteristics	22.5%
Share of vaccine − hesitant individuals (anti-vaccine)	18.8%

The anti-vaccination group of individuals is of particular concern in relation to herd immunity as it relates to a growth in vaccine hesitancy in many countries [12]. Fig. 3 shows the reasons submitted by the vaccine hesitant group for vaccine refusal. The reasons: Vaccine requires more testing before I can trust, I prefer natural immunity, I don’t believe in vaccines, and Options presented are not good enough; informed most of the choices. We also observed differences across reasons in terms of the share of respondents indicating that they choose not to be vaccinated due to ‘options presented are not good enough’ compared to other ‘anti-vax’ reasons. The high share of vaccine-hesitant individuals in the sample can thus be attributed to a lack of trust in the vaccination procurement and dissemination process, the presence of underlying beliefs that fuel the anti-vaccination behaviour, and a lack of information and clarity surrounding vaccination [11], [44]. These are key areas that require intervention by the MoHSS.

Fig. 3.

Reasons for anti-vaccination behaviour.

4.3. Past vaccination behaviour

Past vaccination choices often persist into the future. An initial indication of differences in vaccination preferences is thus given by looking at past vaccination choices made by the respondents. As shown in Fig. 4, about 45% of the sample reported to have taken all recommended government vaccinations, 25% report to have taken past influenza vaccination and 17% have additionally taken other elective vaccines. When these results are compared to the 58.8% of respondents who indicated willingness to take any reasonable vaccine (see Table 3); it indicates that that general vaccination rates are likely to increase over time.

Fig. 4.

Past Vaccination choices.

4.4. Impact of COVID-19 on mental and physical health

The COVID-19 pandemic and associated restrictions are often said to have caused substantial harm to physical and mental health of people; even for those not directly affected by the pandemic [16], [20]. Our data shows that majority of respondents did not in fact report a change in their mental or physical health (cf. Fig. 6). Slightly more respondents reported a deterioration in mental health (12.5%) than physical health (5%). Smaller numbers of respondents’ report improvements, potentially because of home working and reduced stress levels [41]. In Namibia, the prohibition of alcohol sales during the initial lockdown has also led to dramatic reductions in-patient admissions due to assault and accidents (cf. [38]).

Fig. 6.

Risk perception in relation to COVID-19.

4.5. Public risk perception of COVID-19 pandemic

Experience from past pandemics shows that the success of policies to slow down the rapid transmission of highly infectious disease rely, in part, on the public having accurate perceptions of personal and societal risk factors [14], [15], [17]. Fig. 5 shows the risk perceptions from the sample, which measures four areas: likelihood of infection leading to symptoms, likelihood of infection leading to serious illness, likelihood of infection leading to hospitalisation and likelihood of infection leading to death. We plotted the mean risk perception scores from the data (Fig. 6), and the results show that risk perception across the four areas varied between 3.1 and 3.2 on a 5-point scale. The proportionally large standard deviation measured across all four questions furthermore indicate that risk perceptions are spread out and diverse in the sample, but the general risk perception remains fairly high.

Fig. 5.

Perceived impact of COVID-19 on own Physical and Mental health.

4.6. Latent class analysis

4.6.1. Model formulation

The core of the empirical work was concerned with the analysis of vaccine uptake behaviour using the latent class model. As mentioned in s.3, the overall vaccine uptake was composed of three patterns of preferences, namely respondents who always choose a vaccine across their six scenarios, respondents who choose a vaccine in some but not all the scenarios, and respondents who never choose a vaccine (vaccine-hesitant). These groupings translated to the respective class allocations, namely: class 1, class 2 and class 3 in the latent class model.

A key aim of the study in this part was to understand the impact of vaccine characteristics on uptake. Thus, to achieve this aim, individuals classified as ‘vaccine-hesitant’ were excluded from the data during estimation. This reduced the sample used for estimation to 256 respondents. The result of the latent class model are presented in Table 4.

Table 4.

Results of Latent Class Model.

	Class 1			Class 2			Class 3
Attribute	coeff	r.s.e	r.t-r	coeff	r.s.e	r.t-r	coeff	r.s.e	r.t-r
ASC position ($\mathrm{\delta }$ pos) [base]	−0.006170	0.063317	−0.09745	0.056055	0.047496	−1.18021	−0.031935	0.276709	−0.11541
ASC free ($\mathrm{\delta }$ free)	1.603373	0.500102	3.20609	1.585343	0.506360	3.13086	−9.027259	2.098749	−4.30126
ASC paid ($\mathrm{\delta }$ paid)	2.297221	0.478364	4.80224	−0.926700	0.485035	−1.91058	−8.358209	2.61881	−3.19159
ASC no vaccination ($\mathrm{\delta }$ nv)	0.000000	NA	NA	0.000000	NA	NA	0.000000	NA	NA
Vaccine attributes
Risk of infection (βinfection)	−0.005172	0.044605	−0.11595	−0.022839	0.037242	−0.61326	0.0000	NA	NA
Risk of illness (βillness)	0.0000	NA	NA	−0.043294	0.014203	−3.04818	−0.355028	0.124661	−2.84795
Protection duration unknown (βdu)	0.0000	NA	NA	0.00000	NA	NA	−0.396795	1.494417	−0.26552
Protection duration (βdur)	0.002493	0.003813	0.65371	0.003003	0.003426	0.87673	0.00000	NA	NA
Vaccine side effects: mild	−0.022488	0.022348	−1.00631	−0.025652	0.017865	−1.43589	0.000000	NA	NA
Vaccine side effects: severe	−40.06515	9.042884	−4.43057	0.0000	NA	NA	0.0000	NA	NA
Vaccine administration variables
Waiting time to receive vaccine	−0.055851	9.042884	−4.43057	−0.007521	0.008425	−0.89271	−0.010271	0.036214	−0.28361
Fee [to receive vaccine immediately]	−3.7334e-04	8.791e-05	−4.24713	−1.6909e-04	1.4503e-04	−1.16584	−3.5512e-04	4.2202e-04	−0.84148
Population coverage of vaccinated	4.6879e-04 04	0.012488 04	0.03754	0.009904	0.009198	1.07678	0.006133	0.012074	0.50798
Exempt from international travel (ΔOT_RL)	0.428344	0.499399	0.85772	0	NA	NA	0	NA	NA
Delta (ΔSHIP_Freq_RD) [base]	0	NA	NA	0.239298	0.166410	1.43800	−1.575445	0.252125	−6.24866
Model Statistics
Observations	1614
Parameters	31
Class allocation: mean probability	0.40367			0.51280			0.08353
Model LL (Start)	−2597.633			−2597.633
Model specific LL (final)	−2890.17			−3026.383			−5928.619
Overall LL (Start)	−2135.333
Overall LL(0)	−2597.633
Overall LL(final)	−2072.276
Pseudo R2	0.2022
Adjusted R2	0.1903
AIC	4206.55
BIC	4373.53

Notes: coeff = coefficient, rob.s.e = robust standard error, rob.t-r = robust t-ratio, *insignificant.

4.6.2. Model outcomes

Payment fee to receive vaccine immediately, vaccine side effects, and population coverage emerged as the top three influential attributes for vaccine preferences in the model and across classes (Table 4). Overall, we also found that:

• •Risk of infection and risk of illness have negative impacts on utility, meaning that vaccines with a higher efficacy obtain a greater utility. The (per percentage point) impact of changes in the risk of infection is larger than the impact of changes in the risk of illness.
• •Increases in the length of time that a vaccine protects from infection/illness have a positive impact on the utility of vaccination, where there is an additional disutility if the protection duration is unknown.
• •Increases in mild and severe side effects have negative impacts on utility.
• •Increases in waiting time reduce the utility of vaccines, as do increases in the cost for paid vaccine options.
• •Increases in the share of the population already vaccinated have a positive impact on the utility of vaccination. This behaviour could be explained based on risk averseness, by being less willing to accept vaccination when it has not been ‘tested’ on a large share of the population.
• •If vaccination implies an exemption from travel restrictions, then this has a positive impact on the utility of vaccination.
• •Accurate public risk perception is also critical to effectively manage public health risks.

In addition to the above description of the overall effects, the models also uncovered substantial heterogeneity in preferences across individuals, with different sensitivities obtained in the different classes of the LC structures.

4.7. Willingness-to-pay analysis

The average out-of-pocket expense to obtain a high-quality vaccine was found to be N$233.11 (equivalent of US$12.83), while the out-of-pocket expense to avoid travel restrictions was N$377.39 (equivalent of US$20.77). The value obtained for the high-quality vaccine of N$233.11, was found to be higher than values obtained in other African countries such as Ethiopia at US$4.9 [42]) and Nigeria at US$1.9 [3]; but was much lower than WTP values obtained in Chile at US$232 [9], Indonesia at US$ 57.20 [19]and Malaysia at US$30 [45].

Marginal rates of substitution analysis from the model are presented in Fig. 6, Fig. 7. We see that while vaccine performance was presented to respondents as the percentage point risk of infection and illness in the survey, for the MRS calculations, we translated the values into efficacy, looking at how many additional weeks respondents are willing to wait for a vaccine with a 10% increase in efficacy. Some factors were found to have an impact on the WTP of respondents. Overall, we found the following:

• •There are substantial differences across the different vaccine attributes, where these differences vary also between WTW and WTP measures.
• •Protection against illness is seen as more important than protection against infection.
• •There is a high willingness to wait for a vaccine until it becomes clear how long the vaccine will offer protection.
• •We also present the MRS for a reduction from the highest to the lowest risks presented in the SC scenarios, which implies going from 10% to 0.1% for mild side effects, and 0.02% to 0.001% for severe side effects. On this, we see that there is both marginally high sensitivities to go both from serious and mild side effects to the low or no side effects.
• •The attribute relating to exemptions from travel restrictions was also found to have a meaningful role. We note that respondents would be willing to wait several months for a vaccine that would exempt them from such restrictions (see Fig. 7, and similarly a higher WTP value to obtain exemption for international travel (see Fig. 8).

Fig. 7.

WTW measures.

Fig. 8.

WTP measures.

5. Discussion and conclusion

This is the first study to use DCE to explore the preferences of Namibians towards potential vaccines. Although this study was conducted before vaccines were available, Namibia has to date vaccinated just over 500 000 individuals [31]. Therefore, understanding preferences for COVID-19 vaccines is a major need.

The current study revealed a number of outcomes. Several psychological factors were assessed including reasons for not taking the vaccine, past vaccination behavior, perceived impact of mental and physical health, and public risk perception of COVID-19. The study reveals that vaccine hesitancy is overall decreasing in Namibia, but there is still high share of vaccine-hesitant Namibians who demonstrate a lack of trust in the vaccination procurement and dissemination process. Thus, we suggest government to overcome this by improving information and clarity surrounding vaccination. Local studies of monitoring vaccinated individuals can also increase the trust towards COVID-19 vaccination, where this information is shared to elucidate vaccine safety and efficacy in the local population.

The results of the latent class model show that payment fee to receive vaccine immediately, vaccine side effects, and population coverage are the most influential attributes of the vaccines. Despite the considerable heterogeneity in the study sample, we found that characteristics of vaccines matter, both in terms of the decision to be vaccinated or not, and in the choice between different vaccines. A key implication of this work is that some individuals clearly will accept any reasonable vaccine while others are hesitant to be vaccinated. Importantly, we found a third group who are open to vaccination only if the characteristics of the vaccine are right for them. What defines “right” differs across individuals, but efficacy is of especially great importance and low risk of both mild and severe side effects.

A major output of the research is the ability to make predictions of the uptake of vaccination against COVID-19, and this was demonstrated both by the enumeration of WTP and WTW measures. Overall, there is high willingness to wait for a vaccine until it becomes clear how long the vaccine will offer protection, and a high willingness to pay to reduce vaccine side effects. The average WTP for a high-quality vaccine recorded was N$233.11 (US$ equivalent of US$15.14), which corroborates with the model output. An effective behavioural change strategy for COVID-19 vaccines uptake can now be devised to address multiple beliefs and behavioural determinants, and thereby reduce barriers and leverage enablers identified in this study.

There are a number of potentially fruitful avenues for future research. For example, we explored rationales for vaccine hesitancy but our work does not offer solutions in terms of encouraging vaccination uptake amongst this group, which should be explored in further research. Similarly, given our findings in terms of the impact that vaccine characteristics have on uptake, and previous work on the preference and willingness-to-pay for COVID-19 vaccines in Africa [3], [42], another area of work relates to developing the most effective messaging for encouraging uptake by the most hesitant subgroups of the population. A study in Asia found healthcare workers more willing to be vaccinated due to perceived COVID-19 susceptibility and low potential risk of vaccine harm [40], and hence it will be important to understand the hesitance according to different groups including healthcare workers, the elderly and pregnant women to ensure a targeted approach to addressing vaccine hesitancy.

A key limitation is that this survey was rolled out before Namibia had COVID-19 vaccines, and moreover, before Namibia had its worse COVID-19 wave between March and July 2021, which was fuelled by the Delta variant which resulted in more infections and deaths. Preferences of people might have changed due to these factors, and this warrants another wave of data collection using the same survey tools. Nonetheless, the results of this study can inform government and policy makers in refining the vaccine delivery plan to overcome vaccine hesitancy in Namibia.

6. Disclosures about potential conflict of interests

No potential conflict of interests is declared by the authors.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.