Introduction Health information about COVID-19 has been circulating in social networking sites, including unproven claims that smoking and alcohol drinking could protect against COVID-19. We examined if exposure to such claims was associated with changes in tobacco and alcohol consumption.
Methods We conducted a population-based, landline and mobile phone survey of 1501 randomly sampled adults aged 18 years or older (47.5% male) in Hong Kong in April 2020. Respondents reported if they had ever seen claims that ‘smoking/alcohol drinking can protect against COVID-19’ from popular social networking platforms. Current tobacco and alcohol users reported if they had increased or reduced their consumption since the outbreak. Prevalence data were weighted by sex, age and education of the general adult population.
Results 19.0% (95% CI 16.8% to 21.4%) of all respondents reported having seen claims that ‘smoking/alcohol drinking can protect against COVID-19’ from social networking sites. Multinomial logistic regression showed that exposure to the claims was significantly associated with increased tobacco use (OR 2.37, 95% CI 1.08 to 5.20) in current tobacco users (N=280) and increased alcohol use (OR 4.16, 95% CI 2.00 to 8.67) in current drinkers (N=722), adjusting for sex, age, education level, alcohol/tobacco use status, home isolation, anxiety and depressive symptoms, and survey method.
Conclusion Our results first showed that exposure to health misinformation that smoking/alcohol drinking can protect against COVID-19 was associated with self-reported increases in tobacco and alcohol consumption in Chinese during the pandemic.
- advertising and promotion
- surveillance and monitoring
This article is made freely available for use in accordance with BMJ’s website terms and conditions for the duration of the covid-19 pandemic or until otherwise determined by BMJ. You may use, download and print the article for any lawful, non-commercial purpose (including text and data mining) provided that all copyright notices and trade marks are retained.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Dissemination of health (mis)information about COVID-19 via social networking sites, including mobile instant messaging apps and social media platforms, is rapid and far-reaching. Inaccurate or unproven claims that smoking and alcohol drinking could protect against COVID-19 has been circulating online, despite the WHO’s warnings about the lack of evidence.1 2 Exposure to health misinformation may influence a person’s health beliefs, thereby leading to intention to change and actual change in health behaviours.3 Nevertheless, the impact of health misinformation exposure on behaviours has remained understudied.
The smartphone penetration rate in Hong Kong is among the highest in the world (91.5% in 2019),4 and social networking sites have become a major medium for communicating health information.5 We examined if exposure to misinformation that smoking/alcohol drinking can protect against COVID-19 was associated with self-reported changes in tobacco and alcohol consumption in current users in Hong Kong.
Since the first confirmed COVID-19 case on 23 January 2020, Hong Kong had border restrictions, social distancing measures, quarantine and nearly universal voluntary mask-wearing but no enforced lockdown.6 The outbreak peaked in late March with 1035 confirmed cases by April 23. We drew cross-sectional data from the Hong Kong COVID-19 Health Information Survey (CoVHInS) conducted from 9 to 23 April 2020. The design of the survey was adapted from our Family and Health Information Trend Survey (FHInTS) published elsewhere.5 7 All respondents provided informed consent before completing the questionnaires.
We randomly sampled Chinese-speaking Hong Kong residents aged 18 years or above by landline telephone (random digital dialling) and mobile phone. The landline survey used a two-stage sampling strategy. First, a list of landline telephone numbers was generated by using the Government’s numbering plan for telecommunication services and randomised for telephone contacts. Second, on successful contact with an eligible household, an eligible resident whose next birthday was closest to the interview date was invited to participate in the telephone interview. The mobile survey randomly sampled respondents from a population-representative panel of over 100 000 mobile phone users formed by a leading research agency in Hong Kong, with no second-stage sampling. We planned to recruit 1500 respondents (500 through landline, 1000 through mobile phone), which would give a margin of error of ±2.5% for prevalence estimates of 50% at a 95% confidence level.
Respondents of the landline survey were interviewed directly via telephone by trained interviewers using a computer-assisted telephone interviewing system. Respondents of the mobile survey received a private link via text messages or email to access a web-based computer-assisted personal interviewing system and self-administered the questionnaire. Cognitive interviewing was conducted with 10 persons to refine the questionnaire design. We checked a random fifth of the landline interview record to confirm the quality.
Respondents were asked if they had ever seen claims that ‘smoking/alcohol drinking can protect against COVID-19’ from WhatsApp/WeChat, Facebook, Instagram, Twitter and online forums (including Telegram). Current tobacco (including cigarette, electronic cigarette and heated tobacco products) and alcohol (including social and regular drinking) users were asked if they had changed their tobacco and alcohol consumption since the COVID-19 outbreak, with response categorised into ‘no change’, ‘reduced’ and ‘increased’. Other information collected included sociodemographic variables, the practice of home isolation (number of days staying at home at most times in the past week) and psychological distress in the past 2 weeks assessed by the Patient Health Questionnaire-4 (PHQ-4), which gives a summary score of 0 to 16 with a higher score indicating greater anxiety and depressive symptoms.8
Prevalence or proportions were weighted by the sex, age and education to be representative of the Hong Kong general adult population. We applied multinomial logistic regression to model self-reported increase and reduction in tobacco and alcohol consumption (‘no change’ as the base outcome) by exposure to the claims that ‘smoking/alcohol drinking can protect against COVID-19’ (yes vs no). In addition to survey methods (landline vs mobile), we adjusted for factors that might be associated with health information exposure via social networking sites and/or changes in tobacco and alcohol consumption. These included sex, age, education,5 home isolation, and anxiety and depressive symptoms.9 10 Since tobacco use and alcohol drinking are often linked, tobacco and alcohol use status was also mutually adjusted. We repeated the regressions by modelling ‘drinking tea can protect against COVID-19’ as the exposure variable, which should show no association with the outcomes (negative exposure control). Complete case analyses were used because there were no missing data. All analyses were done in Stata/MP V.15.1. A two-sided p<0.05 indicates statistical significance.
One thousand five hundred and one respondents completed the survey, with a response rate of 61.3% (500 of 816) for the landline survey and 61.7% (1001 of 1623) for the mobile survey. Of the 1501 respondents, 282 (19.0%; 95% CI 16.8% to 21.4%) reported having seen claims that ‘smoking/alcohol drinking can protect against COVID-19’ from social networking sites. Leading sources of exposure were WhatsApp/WeChat (n=149) and Facebook (n=138), followed by online forums (n=64), Instagram (n=37) and Twitter (n=6; not widely used in Hong Kong). Bivariate analysis showed that exposure was associated with male sex, younger age and higher education (all p<0.001) and current tobacco use (p=0.045) (online supplementary table 1). Home isolation, PHQ-4 score and survey method were not associated with exposure.
Since the outbreak, 15.6% (48 of 280) of current tobacco users and 5.5% (42 of 722) of alcohol users reported having increased their consumption. The corresponding prevalences of reduced consumption were 19.1% (49 of 280) and 36.8% (244 of 722). Table 1 shows that exposure to the claims and longer home isolation was significantly associated with self-reported increase, but not reduction, in tobacco and alcohol consumption. Higher PHQ-4 scores were significantly associated with both increase and reduction in tobacco and alcohol consumption.
In the sensitivity analysis, exposure to claims that ‘drinking tea can protect against COVID-19’ (17.0%; 242 of 1501) showed no significant association with self-reported change in tobacco and alcohol consumption (all p>0.11).
In this population-based survey of Hong Kong adults, nearly a fifth of the respondents reported having seen claims that smoking/alcohol drinking can protect against COVID-19, and exposure to these claims was associated with self-reported increases in tobacco and alcohol use in current users. The results remained significant after adjusting for other factors that may contribute to changes in tobacco and alcohol consumption under the COVID-19 outbreak, including home isolation and mental distress.9 10
The study also had some notable results. First, more men than women were exposed to such misinformation, which may reflect the male predominance in tobacco and alcohol use in Chinese cultures. The higher rate of exposure in younger and higher educated respondents can be explained by the greater use of social networking tools,5 which facilitated dissemination of health (mis)information. Second, the proportion of reduction in alcohol drinking (36.8%) was much larger than that of increase (5.5%). This could be attributed to enforced closures of bars, pubs and nightclubs from April 3 through the entire data collection period. Since Chinese often drink alcohol during social gatherings and dinning, the practice of social distancing might have contributed to alcohol reduction. Third, increased psychological distress was associated with both increase and reduction in tobacco use. While mental distress is a known risk factor for increased smoking,11 it is plausible that some tobacco users reduced their consumption out of distress induced by the fear of adverse COVID-19 outcomes. This suggests that the COVID-19 outbreak could be a teachable moment for promoting smoking cessation.
Claims about smoking protecting against COVID-19 might have been facilitated by a widely publicised article hypothesising a protective role of nicotine in COVID-19 infection,12 wherein the lead author had previously received funding from the tobacco industry.13 The hypothesis was based on tentative data showing an under-representation of smokers in hospitalised patients with COVID-19.14 However, these data might have been misinterpreted because of misclassification of smoking status (eg, under-reporting, reverse causation) and sampling error (eg, over-representation of healthcare workers). Similar claims about smoking could prevent severe acute respiratory syndrome had also been widely spread in 2003,15 which has later been proven unfounded.16 Growing evidence has also suggested that smoking is associated with poor prognosis and deaths from COVID-19, although more evidence on the effect of smoking on COVID-19 incidence is needed.17 Our results have highlighted the potential of unproven claims in misguiding the public and undermining public health efforts against tobacco use and risky drinking, not to mention adding burden to the already overstrained healthcare system during the pandemic.
Study limitations included the cross-sectional design and self-reported measures. The relatively small sample precluded examination of the associations in demographic subgroups. Although we assessed exposure to health misinformation from several popular social networking sites, exposure via other outlets (eg, traditional media) was not accessed. This might have misclassified some respondents as unexposed and biased the observed associations towards the null. Unmeasured or residual confounding could not be excluded, but the null association between the negative exposure control and outcomes provided support for the validity of the results.
What this paper adds
Unproven health information about the COVID-19 has been circulating through online outlets, but little is known about the impact of exposure to health misinformation on behaviours.
In a population-representative sample of Chinese adults in Hong Kong, nearly a fifth of the respondents reported having seen claims that smoking/alcohol can protect against COVID-19 from social networking sites.
Exposure to such misinformation was associated with self-reported increases in smoking and alcohol drinking in current users, after adjusting for sociodemographic characteristics, home isolation and mental distress.
Patient consent for publication
Ethical approval was granted by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW 20-238).
We thank the Social Policy Research for conducting the study and the respondents for their participation in the study.
Twitter @TT_Luk, @KelvinW95143430
Contributors SZ, JY-HW and MPW conceptualised the study. SZ implemented the study. TTL conducted statistical analyses and drafted the first version of the manuscript. All authors reviewed the manuscript, interpreted the data and approved the final version of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.