Overexploitation of wildlife is pervasive in many tropical regions, and in addition to being a significant conservation and sustainability concern, it has received global attention given discussions over the origins of zoonotic disease outbreaks. Where unsustainable, consumption of wild meat by urban residents has been identified as a major socio-environmental challenge, given it is a significant driver of wildlife declines. Yet, information on urban wild meat consumers and possible ways to target conservation interventions remains lacking. Using one of the largest datasets of urban wild meat consumers (1,391) collected through structured questionnaires in 20 towns and cities, we model the demographic, psychographic, and spatial factors associated with wild meat consumption patterns in Cameroon. We find that nearly half of the sampled consumers ate wild meat once per week or more, and find that the probability of being a frequent consumer was greater among men, those living in smaller towns, and those who do not consider there to be a link between eating wild meat and disease. Threatened pangolin species are highly preferred among urban consumers, and most consumers do not consider there to be a link between COVID-19 and pangolins. Most respondents had eaten wild meat since the beginning of COVID-19 and had not reduced their wild meat consumption due to COVID-19. For the first time, we show that consumers with beliefs against a link between wild meat consumption and disease and those with greater income were less likely to have decreased their wild meat consumption. We identified stakeholders including teachers and religious/community leaders as potentially appropriate messengers for demand-reduction campaigns, with television and radio being the most trusted communication channels among wild meat consumers. Crucially, our study advances the current scientific understanding of the factors that influence wild meat consumption frequency and change due to COVID-19 by urban consumers (particularly health beliefs and settlement size). We discuss how our results could be used to inform the design of wild meat demand-reduction interventions to bring the consumption of wild meat towards sustainability in Cameroon, and how our approach is applied pan-tropically.

We conducted a structured questionnaire study in 20 towns and cities located in four regions (Centre, East, Littoral, and South) across the forested southern half of Cameroon in August 2021. We formed teams of three researchers each, and each team was assigned to one region. Whilst teams were allocated to a given region, they also sampled towns and cities strategically if they were located along roads between allocated regions. In small towns, respondents were recruited to the study using the street-intercept method (Buschmann, 2019), whereby pedestrians were approached in the street. A semi-random sampling method was adopted, whereby researcher teams randomly approached pedestrians near central market areas and actively attempted to recruit men and women of all age groups evenly to improve sample representativeness. The same method was used in medium, large, and very large towns/cities, but we targeted our sampling efforts to areas where wild meat is sold to maximise the likelihood of identifying people who consumed wild meat given the population size of these towns/cities. 

Questionnaires were administered in French but were translated into English where necessary, covering both official languages in Cameroon. The research team facilitated the completion of the questionnaire by reading the questions aloud to respondents and inputting their responses on the phones/tablets on their behalf. On average, the questionnaire took less than 15 minutes to complete with each respondent. Each team of three separately administered the questionnaire with different respondents at the same time to maximise the number of respondents that could complete the questionnaire at once. We excluded those who stated they did not eat wild meat from our sample. The questionnaire was developed and deployed using KoBoToolbox (2018) and loaded onto tablets and smartphones.