Changes in hazardous drinking pre, during and post 70-day alcohol sales ban during COVID-19 pandemic in Botswana

Abstract

Background

Absolute alcohol sales bans instituted in countries like Botswana provide a rare opportunity for a quasi-natural experiment on how such strict policies influence users’ behaviours during the COVID pandemic and beyond. From March 2020 to September 2021, Botswana banned the sales of alcohol on four separate occasions spanning a cumulative 225 days. We studied changes in retrospectively recalled hazardous drinking following the longest and last alcohol sales ban in Botswana.

Methods

This online cross-sectional study, carried out following a 70-day alcohol sales ban in 2021, comprised a convenience sample of 1326 adults who completed the AUDIT-C and had to recall their alcohol use during three points: pre alcohol sale ban (before 28th June 2021), during alcohol sales ban (28th June 2021 to 5th September 2021), and post alcohol sales ban (after 5th September 2021).

Results

The prevalence of hazardous drinking (defined by an AUDIT-C score of 3 or 4 for females and males, respectively) prior, during and post the alcohol sales ban was 52.6% (95%CI=49.8–55.3), 33.9% (95%CI=31.3–36.5), and 43.1% (95%CI=40.4–45.8), respectively.

Conclusion

The findings from this study showed that reduced alcohol availability by way of the fourth alcohol sales ban was associated with reductions in self-reported hazardous drinking, albeit at a lesser degree compared to during an earlier sales ban.

Introduction

Reducing alcohol availability during the COVID pandemic was one of the recommendations made to decrease alcohol related harms which burden health systems (Stockwell et al., 2021). Such reduction in availability included partial or absolute sales bans of alcoholic beverages (Rehm et al., 2020). Botswana implemented absolute alcohol sales bans on four separate occasions, covering 225 days of sales bans between March 2020 and September 2021— 43% of the period was marked by the sales bans. Understanding the role of such prohibition on actual alcohol consumption is needed to assess policy effectiveness, and to contrast with consumption changes during the pandemic in other countries which did not institute prohibitions. The limited research from countries which implemented absolute bans indicates reduction in alcohol use (Wichaidit et al., 2021) and some gains in reducing the strain on health care systems by reductions in trauma cases (Chu et al., 2022). The widely available data on alcohol consumption during this COVID pandemic is from countries that did not implement absolute bans, and indicates a heterogeneity in changes (Acuff et al., 2022), albeit a tendency for reduction is noted in most countries (Kilian et al., 2022). A previous retrospective study on the changes in alcohol use during the second (30 day) alcohol ban in Botswana found a 46% decrease in hazardous drinking compared to before the ban was implemented (Maphisa & Mosarwane, 2022). While the latter finding is instructive, it remains to be determined how longer periods of prohibition influenced hazardous drinking patterns. Therefore, the current study sought to replicate the previous study in Botswana following the fourth (70 day) ban with the aim of establishing differences in recalled hazardous drinking prior, during, and post the prohibition.

Methodology

This study served to replicate an earlier study (Maphisa & Mosarwane, 2022) but focusing on later time points. The earlier study assessed alcohol use around a 30 day alcohol sales ban from 5th August to 3rd September 2020, while this present study focused on alcohol use around a 70 day sales ban running from 28th June 2021 to 5th September 2021. This study utilized an identical protocol to the previous and reported study (Maphisa & Mosarwane, 2022) with a minor adjustment to the instrument to fit the time periods of interest. A 15-item online survey was distributed on a highly subscribed social media platform in Botswana from 21st October period to 7th November 2021. A final sample of 1326 adults was attained from the 1908 survey link clicks. Participants had a mean age of 30.1 years, and 51.4% were males, 49.5% were employed, 60.8% were single, and 55.6% earned less than P3000 (about $228).

The survey required participants to retrospectively recall their drinking patterns, as measured by a modified Alcohol Use Disorder Identification Test-Consumption (AUDIT-C) during three points: pre alcohol sales ban (before 28th June 2021), during alcohol sales ban (28th June 2021 to 5th September 2021), and post alcohol sales ban (after 5th September 2021). When introducing the AUDIT-C to participants, the following statements preceded the repeated test items: a. “Think back to your drinking before June 28th 2021 when alcohol sales were not banned,” b. “Think back to your drinking between June 28th 2021 and September 5th 2021 when alcohol sales were banned”, c. “Think about your drinking since/after September 6th 2021 to date when alcohol sales were allowed”. The prevalence of hazardous drinking (defined as an AUDIT-C score of 3 or 4 for females and males, respectively; Bush et al., 1998) was estimated with respective 95% confidence intervals. Cochran's Q Test and post-hoc analyses (McNemar's Test) were conducted. Logistic regression models were used to determine the associates of hazardous drinking at three time points. Analyses were conducted using the Statistical Package for Social Sciences- Version 27 (IBM Corp, 2020), and significance testing was set at 0.05. Ethical clearance was received from the Botswana's Ministry of Health and Wellness (HPDME:13/8/1).

Results

As shown in Table 1 , the prevalence of hazardous drinking prior, during and post the alcohol sales ban was 52.6% (95%CI=49.8–55.3), 33.9% (95%CI=31.3–36.5), and 43.1% (95%CI=40.4–45.8), respectively. There were significant differences in hazardous drinking across the three time points χ² (2) =132.6, p <0.001. The prevalence estimates at each of the time points differed statistically from each other, with the greatest difference being between pre-ban and during ban hazardous drinking. The highest prevalence of hazardous drinking was observed prior the alcohol sales ban among high income earners (60.2%) and males (57.6%). Gender was the only variable statistically associated with increased odds of hazardous drinking across all the three time points. Males had nearly a twofold increased odds of hazardous drinking prior the ban (AOR= 1.47; CI: 1.17–1.81), during the ban (AOR= 1.56; CI: 1.12–1.97) and post the ban (AOR=1.52; CI: 1.24–1.89).

Table 1.

Prevalence of hazardous drinking across demographic characteristics of the sample (n = 1326).

	n (%)	Pre Ban Hazardous Drinking (n = 697) % (CI)	During Ban Hazardous Drinking (n = 449) % (CI)	Post Ban Hazardous Drinking (n = 571) % (CI)
Whole Sample	1326	52.6 (49.9–55.3)	33.9(31.4–36.4)	43.1(40.4–45.8)
Gender Female Male	644 (48.6) 682 (51.4)	47.2 (43.3–51.1) 57.6 (53.9–61.3)	28.7(25.2–32.2) 38.7(35.0–42.4)	37.9 (34.2–41.6) 47.9 (44.2–51.6)
Employment Status Unemployed Employed Student	303 (22.8) 819 (61.8) 204 (15.4)	52.1 (46.5–57.7) 54.7 (51.3–58.1) 44.6 (37.8–51.4)	37.3 (31.9–42.7) 33.6 (30.4–36.8) 29.9 (23.6–36.2)	45.9 (40.3–51.5) 43.3 (39.9–46.7) 37.7(31.0–44.4)
Monthly Income Range <= P3000 P3001-P6000 P6001-P9000 P9001-P12000 >P12000	737 (55.6) 152 (11.5) 105 (7.9) 98 (7.4) 234 (17.6)	49.1 (45.5–52.7) 51.9 (44.0–59.8) 56.2 (46.7–65.7) 60.2 (50.5–69.9) 59.0 (52.7–65.3)	33.2 (29.8–36.6) 31.6 (24.2–39.0) 37.1 (27.9–46.3) 36.7 (27.2–46.2) 34.6 (28.5–40.7)	42.9 (39.3–46.5) 39.5(31.7–47.3) 39.4 (30.1–48.7) 50.0 (40.1–59.9) 44.9(38.5–51.3)
Relationship Status In Relationship Not in relationship	478 (36.1) 848 (63.9)	51.5 (47.0–56.0) 53.2 (49.8–56.6)	32.8 (28.6–37.0) 34.4 (31.2–37.6)	43.5 (39.1–47.9) 42.8 (39.5–46.1)

Discussion

This study aimed to assess retrospectively recalled hazardous drinking pre, during and post a 70-day alcohol sales prohibition in Botswana during the COVID-pandemic in 2021. Hazardous drinking remains highly prevalent in the current sample, particularly among higher income earners and males, as found in earlier studies in Botswana during the COVID pandemic (Maphisa & Mosarwane, 2022) and even before the pandemic (Lama et al., 2016). The prevalent hazardous drinking in Botswana during and outside the sales ban further highlights the regional trend of a low prevalence of current drinkers (e.g. 30% in Botswana) but a high prevalence of heavy episodic drinking among current drinkers (Manthey et al., 2019). The reduction in alcohol use during the sales ban is consistent with results from Thailand which also instituted absolute sales bans (Wichaidit et al., 2021). Like the earlier study in Botswana (Maphisa & Mosarwane, 2022), the period of the fourth sales ban was marked by a self-reported reduction in hazardous drinking (from 52.3% to 33.9%), albeit the reduction during the second ban was greater by 11%. It is likely that public anticipation of bans at the time of the fourth ban may have led to stockpiling of alcohol prior the bans thus enabling hazardous drinking. Additionally, reported illicit alcohol sales of unquantified magnitude during the periods of the latter bans (Business Weekly, 2021; WHO, 2022) may have also contributed to the lower reduction in hazardous drinking during the period of the fourth ban reported in this study.

Furthermore, the results of this study support the predicted mechanism of decreases in alcohol use as a function of decreases in alcohol availability during the COVID pandemic (Rehm et al., 2020). The results are further consistent with other findings on the positive relationship between stricter alcohol policies and reductions in alcohol use during the COVID pandemic (Mangot-Sala et al., 2022; Vandenberg et al., 2021). Consequently, these results give support for the implementation of WHO's best buys around alcohol availability within LMICs like Botswana. However for LMICs, as was found during the COVID pandemic in Botswana but not unique to the time or country, the public health imperatives of alcohol policy contend with other interests. Firstly, there is the economic pressures of Governments to raise funds from alcohol sales (Mahadevan et al., 2021), and secondly, there is a well-resourced alcohol industry which can use various means (e.g. litigation and publicity campaigns against the sales ban in Botswana) to resistance policy enactment (Morojele et al., 2021). Therefore, there is a need, especially in LMICs, for continued research on responses to alcohol policy to provide robust evidence which can compete with the aforementioned pressures facing policy makers.

While instructive, this study must be viewed considering its limitations of being cross-sectional in design and reliant on participant retrospective recall.

Ethics approval

Ethical clearance was received from the Botswana’s Ministry of Health and Wellness (HPDME:13/8/1).

Declarations of Interest

None.