Materials and Methods

Study Population and Data Collection
A cross-sectional survey was sent out over a brief period of time (20 days; between April 9, 2020, and April 29, 2020) during the full COVID-19 lockdown measures in Belgium. The survey was distributed by the communication services of the University Hospital of Brussels (UZ Brussel) and the university (Vrije Universiteit Brussel) and was circulated through online social media (Facebook and Twitter). The target population included adult Belgian residents, irrespective of age, living area, or current substance use. To reach out to an unselected population, the survey was made available by a link to the survey on Qualtrics online and the link was shared through social media, including Facebook (targeting a broad range of individuals), the websites of the university (targeting students) and the university hospital (targeting patients and hospital co-workers), and the website SeniorenNet.be (targeting the senior generation).
Individuals older than 18 years and living or residing in Belgium during the COVID-19 lockdown measures were asked to fill in the survey online. There were no formal exclusion criteria. The ethics committee of the University Hospital Brussels, Belgium, approved this study on April 9, 2020 (No. 143202043).

Measures
The questionnaire consisted of 5 different parts with 20 questions regarding (1) demographics, (2) smoking cigarettes, (3) alcohol use, (4) cannabis use, and (5) factors that influenced substance use during the COVID-19 lockdown measures. The questionnaire was made available in Dutch, French, and English.
Demographic questions included age, gender, and postal code. Age was asked as an age category (per 5 years). Postal code was recoded into the variables “urban” and “rural.” Current living situation (with partner, with children, with others, or alone), social situation (age and number of children and cared for at home or outside of home), and professional situation (profession, education, and currently home-working) were also asked.
Smoking prior to the lockdown measures was noted as the number of cigarettes smoked on an average day in the period before the COVID-19 lockdown. Response options included “not smoking” (0 cigarettes per day), smoking 1–5 cigarettes per day, smoking 6–10 cigarettes per day, smoking 11–19 cigarettes per day, or smoking 20 or more cigarettes per day. Smoking during the current COVID-19 lockdown was assessed as the average number of cigarettes smoked now and by asking whether they currently smoke more, less, or about the same compared to before the lockdown. In case they smoked more or less, they were questioned about the average number of additional or fewer cigarettes on a normal day, using the same response options regarding smoking prior to the lockdown measures.
For questions regarding alcohol use, reference was made to alcohol units per day. A standard drink (one alcohol unit) was defined as 10 g of pure alcohol. This corresponds to one 25 cL glass of beer (5% alcohol), 10 cL of wine (12% alcohol), 5 cL of aperitif (15% alcohol), or 3.5 cL of distilled spirits (35% alcohol). Respondents were asked how much alcohol he/she drank on an average day before the COVID-19 measures and how much he/she is drinking now (during the lockdown measures), with additionally specifying whether he/she drank more, less, or about the same than before. Response options referred to the average quantity of alcohol consumption during the prior week (rather than the frequency) and included the options “I do not drink alcohol” (mean of 0 alcohol units per day) and “on average 1 alcohol unit per day” to “on average >12 alcohol units per day,” with specifically denoting whether they drank more, less, or about the same as usual.
Cannabis use was assessed as mean amount of cannabis joints smoked per day, both before and during the COVID-19 lockdown measures. The questions concerned the average amount of cannabis consumption per day and specifically asked whether the respondent smoked less, more, or about the same.
The final question asked about motives for increased use of any substance during the lockdown with several options (listed further in Table 4). Those options were not mutually exclusive; several options could be chosen together, and there was opportunity for personal additions from the respondents, as “others, please specify: ___.”

Statistical Analyses
In the survey, age, amount of alcohol units, smoking cigarettes, and cannabis use were recorded in categories. For making these outcomes linear, we applied a transformation of the outcomes by using the mean value of the upper and the lower bound of the categories. Profession was recoded taking into account the variables “do you work in healthcare” and “profession,” thereby regrouping free text into 6 categories: working healthcare, working non-healthcare, student, retired, job seeking, and others.
Linear variables were presented as mean ± SD where applicable. Categorical variables were reported using the number of observations and the percentage of observations. Comparison between the amounts of use in both periods (before and during the COVID-19 lockdown measures) was performed using a Wilcoxon signed-rank test. Cohen's d were calculated and represent the effect size. McNemar tests were used to compare substance use before and during the lockdown. In these McNemar tests, a categorization of “using more” or “using less or equal” was used. When >5% of the respondents showed to consume more during the lockdown compared to before, a binary logistic regression was executed. Here, a backwards regression model building technique was used with variables (age, sex, living in a city, family situation, number of children, professional occupation, educational background, and whether or not you work more from home during the lockdown). Removal of a predictor variable or factor was considered appropriate when having a significance of >0.10. Furthermore, a significance level of 0.05 was considered statistically significant for all tests. In order to examine the motives for substance use and the reason why respondents were using more during the COVID-19 measures than before, a descriptive analysis was executed. After reporting the frequencies, the motives were ranked from highest present motive to lowest present motive for each of the substances. Statistics were performed using SPSS version 26.