Perceptions of Cannabis in sub-Saharan Africa: Findings from a general population survey in six African countries with different Cannabis policies.

Background Recently, legislative shifts in cannabis regulation away from exclusively prohibitionist controls have been seen in sub-Saharan African countries. However, the extent to which public perceptions are aligned with policy trends is unclear. Given that views concerning cannabis may be inuenced by associated legislation we aimed to compare public opinion towards cannabis amongst countries with legislatively different cannabis policies. We conducted an online survey of the general adult population in six purposively selected countries representative of maximum policy variation across sub-Saharan Africa i.e., Ghana, Nigeria, Sierra Leone, South Africa and Zimbabwe. Survey questions covered perceptions towards users, risks and benets, regulations and public health impacts as well as sociodemographic characteristics. Chi square tests were used to analyse the association between categorical variables and explore country level differences. were in responses towards user stereotypes, risks and preferred In relation to users being honest (X 2 =43.316, P<0.0001), most in Sierra Leone(52.9%), Nigeria (45.6%) and Uganda(48.6%) disagreed whereas in Ghana(39.4%), South Africa(43.3%) and Zimbabwe(40.1%) most agreed. Differences in smoking cannabis and smoking tobacco products(X 2 = 27.712, P=0.002) saw respondents from Ghana(51.4%), Nigeria(58.8%), Uganda(50.9%) and Zimbabwe(46.8%) agreeing that smoking cannabis is more harmful to health, whereas majorities in Sierra Leone(45.7%) and South Africa(49.8%) disagreed. Apart from South Africa were the greater proportion opted to allow cannabis for all purposes(28.8%), majorities in other countries supported only medicinal legalisation(X 2 = 96.631, P<0.0001). be inuenced by legislation in the respective country of residency(Mendiburo-Seguel et al., 2017; Steigerwald Sznitman we aimed to compare public opinion towards cannabis amongst African countries with legislatively different cannabis policies. Our objective was to assess how general adult populations in countries with different policies perceive cannabis in terms of users, risks and benets, regulations and health impact. We hypothesize that cannabis-related stigma is greater in more stringent policy settings, in other words restrictive measures and negative stereotypes would be more common among less-liberal countries than among those with more liberal policies.


Introduction
Cannabis, the most widely produced and consumed illicit drug, globally, is a generic term used to denote the several psychoactive preparations of the plant Cannabis sativa (Jain et al., 2018). The cultivation, trade and consumption of cannabis has been prohibited under international drug conventions for many years. Over the past decade, however, there have been calls to explore legal, decriminalised and commercial systems for production and use. The global drift towards more lenient approaches away from traditional prohibitionist controls are re ective of increased evidence supporting its e cacy, greater interest in cannabis-based therapeutics as well as, of its potentials for revenue generation and job creation. A number of countries and a growing cadre of US states have taken to the adoption of tailored approaches to cannabis regulation premised upon evidence-based and humane considerations (Manu et al., 2021).
Africa is a central hub for cannabis use with an estimated 13.2% annual prevalence (Owusu et al., 2021). According to industry estimates, the continent annually produces a quarter of the global total (over 38,000tonnes) with countries such as Ghana, Nigeria and Eswatini being the most notable primary sources and tra cking points (Partners, 2019). Notwithstanding the increasing cannabis production and use in Sub-Saharan Africa, it has not been an exception to the wave of global reforms concerning policies on cannabis. Nations including, Lesotho, Malawi, Morocco, South Africa, Zambia, and Zimbabwe (Beadle, 2021) have taken to restructuring their policies by adopting legalization and commercialization approaches, with the expectations of reaping medical and economic bene ts. Most recently, on June 28, 2021, Rwanda under Ministerial Order No 003/MoH/2021, legalized growing, processing, exporting and medical use of cannabis (Beadle, 2021). Even so, across Africa there is no harmonized drug law and given that nations largely possess autonomy over respective cannabis legislation, there is little consensus amongst countries on cannabis laws. Policies in respective countries have been an amalgamation of local mandates and stipulations of the International Drug Control Treaties of the 1961 Single Convention on Narcotics Drugs. Among other things, this treaty classi es cannabis as addictive, of little or no therapeutic value and illegal for all signatories states (Putri, 2020). Policies arising thereby have oscillated from liberal to punitive. , we aimed to compare public opinion towards cannabis amongst African countries with legislatively different cannabis policies. Our objective was to assess how general adult populations in countries with different policies perceive cannabis in terms of users, risks and bene ts, regulations and health impact. We hypothesize that cannabis-related stigma is greater in more stringent policy settings, in other words restrictive measures and negative stereotypes would be more common among less-liberal countries than among those with more liberal policies.

Methods
The countries covered in this study, Ghana, Nigeria, Sierra Leone, South Africa and Zimbabwe, are signatories to all UN drug control treaties (Putri, 2020), and were purposively selected as representative of the maximum variation in cannabis policy across sub-Saharan Africa (Table 1). Adult residents (18years and older) of any of the selected countries, were eligible to participate. Participant sampling was strati ed by respective country's cannabis policy alignment. Public consumption is liable to a ne or between two and fteen years of imprisonment.

Minimum of ve years in imprisonment
Liable to a ne and or one to ve years imprisonment Possession or use are subject to a ne and or at least six months imprisonment or court ordered treatment.
On a spectrum from liberal to punitive, we placed the South Africa on the liberal side and Nigeria on the punitive side. According to South African legislation, cannabis was approved for medical purposes in 2017 and following a 2018 constitutional court ruling, which ruled cannabis criminalisation as breach to the right to privacy (Tharoor, 2018), private consumption and cultivation was decriminalised(AFRICA, 2020) (Veldman, 2020b). Ghana, passed the Narcotics Control Commission Act (Act 1019), 2020. This act essentially decriminalizes cannabis use for commercial and health purposes (Owusu et al., 2021), representing a country systematically aiming at liberalization and the next most liberal country in our study.
Zimbabwe and Sierra Leone take intermediate positions, as policies in both countries legalise cultivation for medicinal and research purposes but do not have provisions for legal recreational use or for patient access (Veldman, 2020a). However, in the former country there has been greater implementation as evidenced by government support and issuance of licences and ongoing medicinal cannabis activity (Partners, 2019;Sowoya et al., 2020), which for our purposes places it on the liberal half of the spectrum. Whereas for the latter, we labelled as more conversative owing to the fact that, plans to initiate cannabis related activities have been stagnated by multiple policy re-evaluations and widespread condemnation from religious, political and conservative groups(BENDU, 2017; Veldman, 2020a).
In Uganda, government institutions have taken a public prohibitionist position on drugs(Eligh, 2019) however the country can be classi ed as transitioning system, yet still closer to the punitive end. Recreational consumption and cultivation are stringently restricted but, the Narcotic Drugs and Psychotropic Substances Act, 2015, legalises production and exportation of medical cannabis, although guidelines for such activities have not been rati ed(AFRICA, 2021; experts used to tailor questioning to country speci c relevance. To further ensure validity and reliability a pilot study was conducted. Ten individuals from each country were invited to complete the survey and as a result, some questions were rephrased for clarity, grammatical errors corrected and additional options added. Data from the pilot study were excluded from the main study. Our nal questionnaire had six sections; 1)demographics: including age, sex, country, marital status, religion, educational level and employment status.; 2)cannabis use: considering ever cannabis use and awareness of a current user; 3) risks and bene ts: including two multi-selection questions to pick from a list of potential risks and bene ts and 5 point Likert statements comparing cannabis use against alcohol and tobacco; 4)cannabis laws: considering preferential choice of cannabis legal status and 5 point Likert scale statements on cannabis as a source of problems and as easily accessible; 5)cannabis users: 5 point Likert scale statements used to consider user stereotypes of gender, morality /mental health status, work ethic and criminality and a single option question on how users should be addressed under the law and; 6) legalisation of cannabis: all 5 point Likert scale statements considering associated medical e cacy; impact on crime levels, employment opportunities and medical options; age restrictions; provision of information on safe use; and tolerance towards sale and public use.

Data Collection
The online cross-sectional survey was 'live' from October 22 to December 10, 2021. Data were collected via KoboToolbox, a secure web-based platform allowing respondents to directly enter responses online. The link was promoted using popular social media platforms i.e., Facebook and WhatsApp. Respondents read and agreed to an informed consent statement before lling in the survey. All of the data provided by respondents were anonymous and kept con dential.

Statistical Analyses
Analyses were conducted using Statistical Package for Social Sciences(SPSS) Statistics v.27. Likert-scale responses were aggregated into three categories i.e., agree, disagree and not sure. We compared countries using proportional analysis to summarize the demographic data. Cross tabulations and proportional analysis were used for multi-response questions. Chi square tests were used to analyse the association between categorical variables and explore the differences in perceptions towards cannabis use of among residents in countries with different policies. A P value<0.05 was considered to be statistically signi cant.

Ethical Considerations
The study received ethics clearance from the Health Science Centre of Xi'an Jiaotong University. The survey tool is available as supplementary material (Supplementary File1).

Demographics
Overall, 1216 responses were recorded with South Africa (18.6%) registering the most responses, followed by Zimbabwe(18.3%), Uganda(17.9%), Ghana(17.1%), Nigeria(16.8%) and Sierra Leone(11.3%). Most respondents were male, except in South Africa and Zimbabwe. The majority of respondents in all countries were aged between 18-30 years(56.0%) and Christians(65.5%). In Nigeria most participants were married (49.0%) whereas, in all other countries they were single, with Sierra Leone recording the highest number of single participants (76.8%). Generally, most respondents were educated to the postgraduate level (50.7%), with Uganda found to have the highest proportion of postgraduate degree holders(66.1%), closely followed by Sierra Leone(63.8%) and Nigeria(52.9%). Students(26.5%) made up the bulk of our sample group and were the largest represented group in Ghana(26.0%), Sierra Leone(52.2%) and Zimbabwe(25.7%).In Nigeria and Uganda, respondents were mainly employed, (34.3% and 38.1% respectively), whereas in South Africa they were mostly unemployed (27.4%).
An evident difference was shown in ever cannabis use(X 2 = 16.148, P =0.006) and in knowing a current user(X 2 = 51.402, P <0.0001). Nearly two-thirds(62.7%) had never used cannabis however, a nearly similar proportion reported knowing a current user(63.1%). On a country level, South Africa(48.2%) had the highest proportion of people who have used cannabis and Sierra Leone had the lowest(29.7%). Sierra Leone had the highest proportion of respondents aware of a current user(83.3%), whilst over half of Zimbabweans(51.8%) reported not knowing a current user. (Table 2)  Prefer not to say 6 (0.5)

Risks and Bene ts
Proportional analyses showed that perceptions of associated health, social and substance abuse risks were similar across countries (Figure 1(a)). The belief that use leads to substance abuse, addiction and polysubstance use was recognised by nearly a quarter of all respective country participants, with Uganda(31.1%) and Nigeria(31.0%) having the highest proportions. In terms of perceived bene ts of use, Figure 1(b) shows that Ghana(55.4%) and South Africa(55.4%) had larger proportions of participants who endorsed the belief that cannabis use has health bene ts. Cross tabulation also revealed that, compared to other nationalities, South Africans(35%) more greatly considered cannabis to have social bene ts. The belief that cannabis use is non-bene cial was most supported by Sierra Leoneans(25.9%) and least supported by South Africans(8.4%). (Figure 1

Discussion
Across the included countries, legislative trends are not uniformly re ected in public attitudes towards cannabis. Variations were noted in responses towards user stereotypes, cannabis being a problem, comparisons with tobacco and alcohol use, preferred legislation, and strategies towards addressing users.
Consistent with our hypothesis, countries with more liberal policies were associated with more positive sentiments. Dependent upon the question of focus, at the liberal end of the spectrum are Ghana, South Africa and Zimbabwe whilst Nigeria, Sierra Leone and Uganda were more conservative. Questions were marked consistency across countries were identi ed related to the accessibility, age limits for sales, comfortability with regulated medical use and sale, public use and associated risks and bene ts. Participants were asked if they had ever used cannabis and if they were aware of a current user, to better contextualise interpretation of attitudinal differences between countries. There is a relationship between toleration of cannabis and its consumption and awareness of current users: countries that are more liberal tended to have higher rates of ever cannabis use (above 35%) whilst less liberal countries had higher rates of knowing a current user (above 56%). These results indicate that country differences in cannabis use experience and encounters with users can be re ective in collective attitudes.
Overall, participants from more conservative countries more greatly endorsed common gender (male) and negative typecasts of users (criminals, mentally ill) and disagreed with positive personality attributes (hardworking and trustworthy). Studies have shown that African cultures and religions tend to consider cannabis consumption as taboo (Carrier & Klantschnig, 2017; Mametja & Ross, 2020) and our study found higher levels of religiosity amongst countries considered more conservative-particularly Sierra Leone. Cannabis use is a stigmatized behaviour associated with devaluation of users (Skliamis et al., 2020).
Amongst users this could be linked to the need of concealment to avoid prejudice and punishment, particularly in countries such as Nigeria with harsh penalties for users (Nelson, 2018) and Uganda where harsher repealed law provisions are used to charge and prosecute users ((HRAPF), 2015).Moreover, such divisions of stereotypes can provide an explanatory basis for the fact that participants in more liberal countries mainly considered that users should be left alone whereas sending them to treatment (medical or religious) was more greatly associated with less liberal countries.
Risk (social, health and substance abuse) perceptions were similar across countries. However, participants from more liberal countries had an overall more favourable view towards social and health bene ts of cannabis use. Cannabis is typically introduced to adolescents in social settings such as parties (Mehanović et al., 2020) and moderate use is linked to positive social attributes whereas abstinence is viewed negatively (Robertson et al., 2020). As most survey participants are considered young this may explain the endorsement of social bene ts. However, pressure to conform to socializing norms and engage in risky behaviours ( Regarding smoking cannabis as more harmful to health than tobacco products and driving under the in uence of cannabis as safer than under the in uence of alcohol, a divide was seen in responses with Sierra Leonean and South African respondents being more likely to believe cannabis use as safer in both instances. Although considered to be on different sides of the policy spectrum, these responses may be connected to experiential encounters with cannabis and users, given the fact that South Africa reported the highest proportion of ever cannabis use and Sierra Leone had the large percentage of participants aware of current users. Similar to previous studies that found enhanced tolerance for cannabis when purposed for medical reasons (Mikos & Kam, 2019), majorities in all countries considered the sale and use of medicinal cannabis under doctor authorisation acceptable, safe and effective. This somewhat supports the assumption that public opinion toward cannabis is driven by deliberations related to medical effects, as opposed to current polices and concerns for detrimental public health effects (Sznitman & Bretteville-Jensen, 2015). Evidence suggests that exclusive cannabis smokers, while, exposed to lower levels of harmful constituents than exclusive tobacco smokers and co-users, are still exposed to higher toxicant levels than non-smokers (Meier et al., 2021).
Associations have been found between blood tetrahydrocannabinol(THC) levels and impaired performance with cannabis intoxication related to a statistically signi cant increased risk of vehicular accidents(Jasmine Turna, 2020). As policies evolve, the perspective that cannabis use is safe and of low risk may become a more salient public health discussion requiring targeted campaigns against misconceptions.
Support for legalization of cannabis was mainly associated with medical use. Total legalisation was less supported which may be linked to the idea that full legalisation would increase crime and illicit use as majorly agreed upon in all countries. While the consequences of liberalised cannabis systems on crime trends are not fully understood, there have been reports of increased illicit use and poisoning in jurisdictions with commercialized medical cannabis (Sznitman & Bretteville-Jensen, 2015). This may explain why our participants, regardless of nationality, expressed the need for age restrictions and providing information on laws and safe use. Although contrasting to results from a similar Latin American study where support was largely associated with legalization of recreational use, our ndings showcase the effect of territorial differences on public opinions and the importance of accurate assessments of current situations in respective countries. Support for medical cannabis was further shown by most participants in agreement with legalization leading to greater medical options and employment opportunities. Countries are eyeing the socioeconomic prospect of legally regulating cannabis for medicinal, industrial and scienti c purposes (Bizimungu, 2020; Bloomer, 2019; Partners, 2019). By 2023, a fully legal and regulated African cannabis industry, valued at over $US7.1b (Partners, 2019), is anticipated to boost regional employment particularly to the bene t of South Africa and Nigeria with the world's highest and third highest unemployment rate (Naidoo, 2021 are unable to vote for their favoured policy and changes are made largely without public consideration. The signi cances of policy implementation without formulation of public consensus have not been explored and should stand as a key area for countries considering similar actions.

Limitations
Owing to particular limitations, ndings must be interpreted with caution. Firstly, the African population is over a billion and we included only six out of a total of 54 African countries therefore results cannot be generalized. The national representativeness of our survey also comes into question as we did not consider speci c regional variations and settings which may have affected perceptions. The use of an internet survey may have limited responses to those with internet access and the population who chose to participate may differ from individuals who chose not to or were unable to. In our case most of our respondents were young and well-educated and therefore may have biased our ndings to the opinions of this particular subgroup. We did not consider country level factors such as crime and violence indices associated with cannabis which may be pertinent. Data collection reliant upon self-reporting is subject to social desirability bias which may have made individuals less likely to report substance use behaviors; however, use of an anonymous survey should have mitigated this effect. Lastly, the study was cross-sectional therefore, it is beyond the ambit of this paper to assess if legislative reform drives public opinion, or vice versa. Future studies should consider examining existing perceptions of cannabis using greater experimental variation of question format, as well as incorporating qualitative components.

Conclusion
Amongst sub-Saharan African countries there exist signi cant differences in cannabis use and associated policy attitudes, suggesting that people are questioning the status quo. Respective country's policy and associated stringency appeared to be re ected on participants' responses. Our ndings imply increasing public support for medical cannabis whilst recreational use remains a more contentious issue. This is based on the observation that there is a very clear direction of travel amongst countries towards forms of cannabis liberalisation with the main underpinning argument for reform in both conservative and liberal countries being the potential medicinal and economic bene ts afforded by less stringent regulations. It is essential to gain a more nuanced understanding of policy dynamics as well as the related degree and rationale for public opinion as even conservative countries appeared on certain topics to be divided. This may bene t countries contemplating reformation.   SupplementaryFile1Survey.pdf