The Effects of COVID-19 Pandemic Lockdowns on Alcohol Consumption and Tobacco Smoking Behaviour in South Africa: A National Survey

The novel coronavirus disease (COVID-19) has resulted in unprecedented morbidity and mortality across the world [1, 3]. Due to such tremendous increases in morbidity and mortality, almost all countries declared restrictions in the first quarter of 2020 as a way to mitigate the pandemic’s effect [4]. Thus, on March 26, 2020, the South African (SA) government started implementing national lockdowns to slow down and contain the spread of COVID-19 across the country [5]. These national lockdowns included travel bans within and across the provinces, closures of schools, universities, non-essential businesses (tourism, restaurants, sporting events, concerts), as well as social contacts among other restrictions [5]. The SA government also introduced total alcohol and tobacco sale restrictions until the end of August 2020, when controlled and coordinated sales were permitted Monday to Thursday [5].

As a result of lockdown-induced restrictions, people worldwide have experienced social isolation, and general fear [6, 7], potentially resulting in negative dysfunctional responses including unhealthy behaviours such as increased alcohol and tobacco consumption [8, 9]. The World Health Organization (WHO) has already warned of the potential future health risk of increased alcohol consumption and tobacco use during the COVID-19 restrictions [4]. This too was observed with previous pandemics (SARS outbreak in 2003, H1N1 influenza, and EBOLA) influencing people to adopt dysfunctional coping mechanisms such as increased alcohol consumption, tobacco, and marijuana use as well as experiencing more mental health issues [6, 10].

In the USA, approximately 60% of adults reported having increased their alcohol intake as a result of the pandemic lockdown as compared to pre-COVID-19, with only 13% reporting a decreased intake [11]. In Germany, 35.5% of adults reported an increase in alcohol consumption during lockdowns with only 21.3% drinking less [12]. The odds of smoking were also identified to have increased during the lockdown in Germany, with almost 46% of adults reporting increased tobacco use due to stress-related issues [12]. Furthermore, increased alcohol consumption and tobacco use have been attributed to restrictions in movement, increased stress, boredom, loneliness, and availability of alcohol in the house [11, 13].

In South Africa (SA), a study in 2018 reported that in general, 33% of adults consume alcohol with 43% of these binge drinking [14]. Furthermore, excessive alcohol consumption is associated with the burden of infectious diseases, non-communicable diseases, injury and trauma, and poor maternal and child health in SA [15]. With the COVID-19 pandemic lockdowns, alcohol consumption and tobacco use may have increased along with psychological and mental health issues. However, in SA, alcohol and tobacco sale restrictions were implemented to limit social gatherings as well as relieve hospitals of alcohol-related trauma cases. Furthermore, tobacco restrictions were also implemented to prevent the spread of COVID-19 via cigarette sharing and reduce the number of severe cases of COVID-19 among tobacco users [16]. Therefore, we aimed to investigate the self-reported changes in alcohol consumption and tobacco smoking behaviour in a nationally representative population 14 months after total alcohol and tobacco sale restrictions before the Omicron COVID-19 wave 4 and while SA was in a low-level lockdown. We aimed to answer the following research questions: What were the self-reported changes in alcohol consumption and tobacco smoking behaviour among South African adults 14 months after COVID-19 alcohol and tobacco sale restrictions? Furthermore, what were the changes in alcohol consumption and tobacco smoking behaviour which were related to educational level, employment, and wealth status among other factors? Our results are discussed in the context and differences between SA and Europe to draw parallels between the two.

This national cross-sectional survey was conducted on 3,402 South African adult participants and was executed on behalf of the researchers by an international market research company (Fig. 1). Questions included the amount of alcohol consumed and the number of cigarettes, e-cigarettes, and vaping done in a day or week as well as an exploration of the impact of COVID-19 on alcohol consumption and tobacco use. The study was conducted in October 2021 and administered in English and SA’s major local languages (IsiXhosa, IsiZulu, Sesotho, Sepedi, Setswana, and Afrikaans).

The study participants were recruited from all nine provinces of SA using a 6-phase stratified random probability sampling method. Phase 1, stratification (Fig. 2), was used to select provinces, community size (metropolitan areas; city; large town; small town, a large village, and rural area), and gender and to ensure adequate representation. Phase 2, selection of sampling units (with a population of greater than 500), was randomly selected for six interviews per unit. The geographic information system (GIS) mapping technology was used to determine the starting point for the survey in each sampled unit. From the identified starting point, the first household to be interviewed would be randomly selected and thereafter, interviewers would skip five houses and conduct the next interview on the sixth household. Only adults, 18 years and older, were included in the interviews.

The survey included assessments of sociodemographic variables (age, gender, education level, and employment status), wealth index (which was calculated using household assets, with a score of 1 [classified as poor] to 5 [classified as wealthy]) and changes in both alcohol consumption and tobacco use during the COVID-19 pandemic lockdown. The assessments on both alcohol consumption and tobacco use focused on the period from the beginning of the lockdown towards the end of March 2020 until October 2021. These self-reported changes in this period were being considered to the pre-COVID-19 levels of consumption and use. The survey’s exact items/questions were from the US Centre for Disease Control (CDC) Global Adult Tobacco Survey and the WHO Alcohol Use Disorders Test (WHO-AUDIT) [17, 18]. These questionnaires were administered in person (face to face) by the research field workers. The dependent variables for both alcohol consumption and tobacco use were categorised into two groups each as follows: hazardous or harmful alcohol consumption and moderate-severe alcohol use disorder; and light and heavy smoking (Table 1).

Statistical analyses were performed using Stata version 16 (Stata Corp Ltd, Texas, USA). The general characteristics of the study respondents were analysed using descriptive statistics. χ² with confidence intervals was used to determine the association between the binary scores and the independent variables of interest (socio-demographic characteristics, COVID-19 effect on alcohol consumption, province, and community size). Using simple logistic regression, dependent variables that were significant in χ² analysis and the independent variables of interest were computed. Odds ratios were used to determine the strength of association between socio-demographic characteristics and provinces with moderate-severe alcohol use disorder and heavy smoking. All the analyses were two-sided, and the p value was highly significant at <0.001 and moderately significant at < 0.05.

The mean age of the participants was 37.7 years, 52.5% were female, 59.2% were never married, and 66.8% had secondary and/or high school education (Table 2). The largest proportion of participants (57.10%) was employed, with 37.2% not in any employment. In addition, most participants were from Gauteng province (28.92%).

Thirty-three per cent (33.2%) of the respondents reported drinking alcohol, with half of these (50.3%) classified as having hazardous/harmful alcohol consumption or moderate-severe alcohol use disorder according to the alcohol audit scores. Twenty-two per cent (22.0%) of participants reported that COVID-19 pandemic lockdowns changed their alcohol drinking, with 38.1% indicating that their consumption had decreased a lot, or they had quit during the pandemic (between March 2020 and October 2021 and this included the period where alcohol sales were restricted) (Table 2).

Only 19.2% of the participants indicated that they had ever smoked, with 82.6% of those still currently smoking at the time of the survey. Of those currently smoking, 87.8% were light smokers, smoking 10 or fewer cigarettes per day. As a result of the COVID-19 pandemic lockdowns, almost one-third (27.2%) of respondents changed their use of tobacco products or vaping, with over 60% indicating that their smoking has decreased a lot or they had quit during the pandemic and one-third (29.1%) having increased or initiated their tobacco use during the pandemic.

Moderate-severe alcohol use disorder (likelihood of alcohol dependence) did not vary by age (p = 0.11), marital status (p = 0.68), educational level (p = 0.17), wealth index (p = 0.75), or community size (p = 0.32). Moderate-severe alcohol use disorder was more frequently observed in men, respondents who were not employed (including students), and those who reside in the Gauteng province (Table 3). Additionally, those with a higher likelihood of alcohol dependence less frequently reported decreasing or stopping alcohol consumption as a result of the pandemic.

Heavy smoking differed by older age (p = 0.02), wealth index (those classified as wealthy) (p < 0.001), tobacco smoking changes due to COVID-19 lockdowns (p = 0.01) and those residing in Gauteng province (p = 0.04). The reported degree of tobacco uses among smokers did not, however, differ by gender (p = 0.08), marital status (p = 0.10), educational level (p = 0.21), employment status (p = 0.20), or community size (metropolitan area, city of rurality) (p = 0.09) (Table 4). Those who smoked more (heavy smokers) each day less frequently reported decreasing or quitting tobacco as a result of the pandemic.

Multivariate logistic regression showed that males were more likely to have moderate-severe alcohol use disorder (OR = 1.69, CI: 1.2–2.4). Furthermore, those who started or increased their alcohol drinking due to COVID-19 lockdowns had a higher chance (OR = 2.1, CI: 1.3–3.4) of reporting moderate-severe alcohol use disorder as compared to those who quit or decreased alcohol intake due to COVID-19 lockdowns. Employment status had no predictive effect on moderate-severe alcohol use disorder, employed (OR = 1.4, CI: 0.9–2.0) and retired (OR = 2.2, CI: 0.7–6.3) when compared to those unemployed.

Those in the Eastern Cape (OR = 0.5, CI: 0.2–0.9), KwaZulu Natal (OR = 0.5, CI: 0.2–0.8), North West (OR = 0.2, CI: 0.04–0.6), and Gauteng (OR = 0.5, CI: 0.3–0.8) provinces were found to lower the odds of moderate-severe alcohol use disorder when compared to those in the Western Cape. Although those in the Northern Cape had a higher proportion (above 60%, see Fig. 3) of moderate-severe alcohol use disorder, there was no significant difference when compared to the Western Cape (OR = 1.8, CI: 0.2–20.9). There were also no significant differences between Western Cape and Free State (OR = 0.5, CI: 0.2–1.0) provinces, nor for Mpumalanga (OR = 0.4, CI: 0.2–1.1) and Limpopo (OR = 0.6, CI: 0.3–1.1) provinces.

Multivariate logistic regression showed that those in the age group, 25–34 years had lower odds (OR = 0.3, CI: 0.1–0.9) of smoking heavily when compared to the reference group of 18–24 years age. There was no significant difference between the 35–44 years (OR = 0.7, CI: 0.3–1.6), 45–54 years (OR = 1.6, CI: 0.8–3.6), and 55–64 years (OR = 0.8, CI: 0.3–2.5) when compared to those 18–24 years. Those with a higher wealth index, wealth index 3 (OR = 3.1, CI: 1.0–9.6), wealth index 4 (OR = 3.3, CI: 1.1–10.4), and wealth index 5 (OR = 7.0, CI: 2.4–20.6), had higher chances of being heavy smokers when compared to those in the wealth index 1 (poorest). There was no significant difference between those in wealth index 2 (OR = 0.2, CI: 0.02–1.4) and wealth index 1.

Only those in the Eastern Cape (OR = 0.1, CI: 0.02–0.9) were found to lower the odds of heavy smoking when compared to those in the Western Cape (reference group). There were no significant differences between Free State (OR = 0.5, CI: 0.1–2.4), KwaZulu Natal (OR = 0.5, CI: 0.2–1.2), Gauteng (OR = 0.8, CI: 0.5–1.5), Mpumalanga (OR = 0.4, CI: 0.09–1.9), Limpopo (OR = 0.2, CI: 0.02–1.2), and Western Cape provinces. Despite those in the North West having a higher proportion (20%) of smoking heavily, there was no significant difference when compared to Western Cape (OR = 1.3, CI: 0.3–6.3) (see Fig. 3).

This study aimed to investigate self-reported changes in alcohol consumption and tobacco smoking behaviour in a nationally representative population during the COVID-19 pandemic lockdowns in SA. We further explored factors associated with the reported changes in alcohol consumption and tobacco smoking behaviour. On average, more than 62,000 people in SA die from alcohol or intoxication-related causes, where alcohol-related medical cases contribute to a significant proportion of the hospital trauma burden [16, 19]. The study findings show that 33.2% and 19.2% of adults in SA reported consuming alcohol and smoking, respectively. These findings are consistent with what has been reported in other studies [20, 23] thus not indicating much overall change. However, despite enforced sales bans on tobacco and alcohol products, 24.0% of the 33.2% of adults who drank alcohol reported increasing or starting the use of alcohol and 29.1% of those who smoked had increased or started the use of tobacco/vapes during the pandemic. At the time of this national survey in Oct 2021, SA had moved into lower levels of movement restrictions and products were again being sold, with 50.3% of adults who drank reporting potentially harmful alcohol use and 12.2% of those smoking reported heavy smoking.

In this study, even if the majority of respondents quit or drank less alcohol due to COVID-19 lockdowns, slightly above a fifth of drinkers reported that they had started or increased their alcohol consumption. Changes in alcohol consumption could have been due to the restrictions placed on alcohol sales and social interactions in SA [24, 25], as well as the ability to purchase alcohol at inflated prices as seen by the association between potentially harmful alcohol use and heavy smoking with increasing wealthy status. These changes in alcohol consumption are similar to those observed in previous studies around the world, where more than 11.0% and 20.0% reported an increase in alcohol in France, and Australia, respectively [26, 27]. These similarities with European and other first world countries indicate that COVID-19 restrictions might have had a universally negative impact on people’s risky behaviours, which prompted increases in smoking or excessive alcohol use. Therefore, if such universality exists, it will also be encouraged for SA and first world countries to learn from each other about the mechanisms and measures to address these risky behaviours and their effects on health.

Being male, employed, living in certain provinces (Free State, Limpopo, Mpumalanga), and having started or increased alcohol consumption due to COVID-19 lockdowns were found to be significantly associated with moderate-severe alcohol use disorder. In a previous study conducted in SA, men were significantly more likely to exhibit problem drinking [28]. The negative impact of lockdowns on alcohol consumption was also reported in Australia and Poland although these were associated with psychological distress (stress, anxiety, and depression) [27, 29]. However, in Belgium, the increase in alcohol consumption was found to be associated with unemployment, being young and hazardous alcohol use among other things [30, 31]. Such differences between SA and Belgium might be due to drinking patterns, cultural factors as well as the imposed total restriction on alcohol sales that occurred between March and August 2020 in SA but did not happen in the other mentioned countries. Although parallels can be drawn between SA and Europe as to the reasons behind the changes in alcohol and tobacco consumption, it is worth pointing out that the potential effects of such risky behaviours on health are universal and should be addressed as such through coordinated efforts.

Although living in only three provinces (Free State, Limpopo, and Mpumalanga), was found to be associated with moderate-severe alcohol use disorder in our study, the South Africa Demographic and Health Survey (SADHS) 2016 Key Indicator Report stated risky drinking was pervasive across all provinces and wealth quintiles [32]. Therefore, in future assessments, it will be worthwhile to understand why these three provinces, Free State, Limpopo, and Mpumalanga were associated with moderate-severe alcohol use disorder.

Overwhelmingly, respondents reported a change in their smoking behaviour (reduction or quit smoking) since the onset of the COVID-19 pandemic lockdowns though this was less likely in heavy smokers. This reduction/quitting change was different to what was reported in Australia, where there was no noticeable change in smoking behaviour among adults during lockdowns [27]. These differences might be due to the total tobacco sales ban which was executed during the initial lockdowns in SA but not in Australia. In another study in SA, evidence of the decline in smoking prevalence was seen during the lockdowns with between 16% and 49% of smokers reporting not smoking during the sales ban, nationally [33]. However, the true extent of this fall in tobacco use remains uncertain and post-lockdown research will be useful to quantify this change.

Having a higher wealth index of 4 or 5 (relatively wealthy) was associated with heavy smoking when compared to those with a wealth index of 1 (poorest). With total restrictions on tobacco sales, those relatively wealthy might have purchased bulk stocks of cigarettes before the restrictions kicked in or they might have managed to buy expensive cigarettes through illegal sales. In France, heavy smoking during the COVID-19 pandemic lockdown was reported to have been associated with individuals with a high level of education or high socio-professional category although there were no tobacco sales restrictions [26]. In as much as SA and most of Europe and the rest of the world have increased tobacco tax, this study has shown that more can be done to further restrict tobacco use in the world and limit the negative health implications that smoking has.

Middle-aged (those aged between 25 and 34 years) respondents had lower odds of smoking heavily when compared to the young adults (18–24 years) and this was also reported in a previous study in SA where the age of initiation of tobacco use and uncontrolled smoking was below the age of 23 years [34]. This was also consistent with what was reported in Belgium where younger respondents had increased odds of smoking heavily during the lockdown [30]. This finding is important as it highlights the need to have this young group as a target for creative anti-smoking strategies and interventions that reduce smoking. Those living in Eastern Cape were found to have lower odds of heavy smoking when compared to those in Western Cape and this is consistent with previously reported findings which placed the Western Cape as the province with the highest smoking prevalence in SA [22].

These changes in alcohol consumption and tobacco use cannot be fully explained by the restrictions imposed during the lockdowns since we did not collect data on pre-lockdown consumption levels. Besides the sales restrictions that were imposed during the lockdowns, the reported changes in alcohol consumption and tobacco use might be partly due to differences in socioeconomic status and the potential impact of working from home. In Denmark, the confirmed polarised alcohol consumption in the first few months of the COVID-19 pandemic was characterised by age, consumption levels, and educational status [35]. Therefore, in our study, the more socially advantaged or wealthy groups might have had unfavourable changes because they could afford to buy alcohol and cigarettes in bulk and/or have stocks at home. In Norway, after the COVID-19 restrictions, self-reported reasons for consuming less alcohol pertained to less on-premises drinking and fewer social occasions, whereas the reasons for heavy drinking pertained to fewer consequences of drinking more and treating oneself to something good [36]. However, further research might be needed to confirm these findings, including the understanding of the difference between pre-and post-lockdown consumption levels. To fully understand if the reported changes persist in the long term, assessing the consumption/smoking during and after the Omicron wave reported in November 2021 as well as the post-lockdown alcohol consumption and smoking levels will be important. This research can also be carried out in Europe because of the similarities that have been reported in our study and what has been found in different European countries.

The validity of the data presented in this study may have been influenced by recall bias since a pre-lockdown assessment was not conducted. Furthermore, the use of the cross-sectional study design and non-probabilistic sample might have resulted in underrepresented groups not being included in the survey although the sample for this study was nationally represented. Therefore, the reported changes in alcohol consumption and tobacco might need to be interpreted with some caution. The strength of this study is that the findings can be used to better target specific at-risk populations for harmful practices and encourage them to adopt positive healthy lifestyle behaviours, such as physical activities instead of increased alcohol consumption and tobacco use. The findings could also be used to facilitate health promotion and education toward reducing alcohol consumption and smoking cessation in future pandemics or subsequent waves of COVID-19. Health promotion and education activities might contribute to fighting unhealthy behaviours in societies. This is because health promotion and education essentially improve people’s access to better information and health services that play vital roles in empowering them to have more control over their well-being and health.

Despite restrictions on the sale of alcohol and tobacco in SA between 27 March and August 17, 2020, during the COVID-19 pandemic, a similar proportion of respondents reported the use of both substances 14 months after sale restrictions were lifted. The overall decline in alcohol consumption and tobacco use might suggest that the regulatory restrictive strategies on sales had some effect on the reported changes but may be inadequate, especially during times when individuals are likely to experience high-stress levels. Therefore, encouraging and promoting people to adopt positive healthy lifestyles such as physical activities and reducing alcohol consumption and quitting smoking might have a long-lasting effect than restrictions on sales. The results of this study should encourage the government to carry out targeted health promotion and education towards reduction in alcohol consumption and smoking cessation. These targeted health promotions and education campaigns (which can be universal) can be provided through the media on platforms such as national television channels, radios, and social media (Twitter, Instagram, and Facebook) [37].

We thank IPSOS for their contributions in conducting the research.

This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study respondents were approved by the University of the Witwatersrand Human Research Ethics Committee (Clearance number H21/06/36). Written informed consent was obtained from all respondents after the objectives of the questionnaire were explained to them.

The authors declare no competing interests.

Funding for this study was provided by the National Research Foundation (NRF) of South Africa.

Witness Mapanga made substantial contributions to the conception, design of the work, acquisition, data analysis, and interpretation of data, drafted the work, and substantively revised it. Shane A. Norris was responsible for oversight of data collection. Made substantial contributions to the conception, design of the work, acquisition, and interpretation of data, drafted the work, and substantively revised it. Ashleigh Craig, Asanda Mtintsilana, Siphiwe N. Dlamini, Justin Du Toit, and Lisa J. Ware made substantial contributions acquisition in writing the manuscript and substantively revised the work. All authors approved the final manuscript.

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.