Fear of COVID-19 and illicit drug use during COVID-19 pandemic in Japan: a case-control study

Data collection

We obtained data from the Japan COVID-19 and Society Internet Survey (JACSIS). The JACSIS is an Internet-based, self-reported questionnaire survey conducted by Rakuten Insight, Inc., a leading Japanese Internet-based research agency (Rakuten Insight Inc, 2002). The JACSIS questionnaire consists of items related to the daily living, health, and economic activities of the population—including matters related to COVID-19. This survey is conducted once per year, between August to November, with the first one taking in August 2020 (i.e., shortly after the start of the COVID-19 pandemic). The survey instrument was sent by the survey agency to panelists who were randomly selected based on the Japanese population distribution in the 2019 Comprehensive Survey of Living Conditions (CSLC), in order to avoid sample bias introduced by age, sex, and prefecture of residence. The panelists who agreed to participate in the survey accessed a designated website and completed the questionnaire. Participants could choose not to complete the survey and withdraw at any point. In this study, we conducted a retrospective longitudinal study using data from the 2020 JACSIS as the baseline and data from the 2021 or 2022 JACSISs as the follow-up data. The baseline survey was conducted from August 25, 2020 to September 30, 2020 (The Japan COVID-19 and Society Internet Survey, 2023). The follow-up surveys were conducted between September 27 and October 29 of 2021, and between September 12 and October 19 of 2022.

Participants

Figure 1 presents the flowchart of participant selection. In the 2020 survey, the survey instrument was distributed online to 224,389 panelists. The respondents were randomly sampled so that their composition was representative of the Japanese population, yielding a sample of 28,000. The response rate was 12.5%. This is lower than the 37.9% Internet response rate for the Census (a survey administered by the Japanese government) but no definitive benchmark for response rates in online surveys has been defined. Some studies have suggested that whether the respondents are representative of the broader population is more significant than high response rates (Ministry of Internal Affairs and Communications, 2021; Wu, Zhao & Fils-Aime, 2022). Since the population of this study were all Japanese and the participants were sampled to fit the distribution of the overall Japanese population, the sample was deemed adequate despite the low response rate. To improve data quality, inconsistent or unnatural responses were considered invalid and excluded from the analysis. Specifically, respondents were considered invalid if they did not select the response that they were instructed to select in Q12 of the survey (n = 1,955) or if they answered that they had all diseases in Q75 (n = 187). The number of participants who gave incorrect responses and were excluded was 2,109, and the number of valid responses was 25,891. Additionally, participants with a history of illicit drug use at the baseline (n = 1,410) and those who did not participate in the follow-up survey (n = 6,681) were excluded, leaving 17,800 participants in the final analysis.

Outcome variables

The primary outcome measure was the occurrence of illicit drug use. Those who had used illicit drugs at least once at the time of the follow-up survey (n = 837) were included in the outcome group, and those who reported no history of illicit drug use at the follow-up (n = 16,963) were included in the no-outcome group. There are various types of illicit drugs, with varying effects on the human body (Lundqvist, 2005). Although overdoses of many of these can lead to death, there are currently no reports in the literature of cannabis overdose as a direct cause of death (National Institute on Drug Abuse, 2019; CDC, 2024). Moreover, although the use of cannabis for any purpose is illegal in Japan, some countries and states have legalized its recreational use—in contrast to other drugs (Wilkinson et al., 2016; Volkow et al., 2016; Government of Canada, 2022). Therefore, we categorized illicit drugs into two groups: cannabis and non-cannabis. Those who used only cannabis among the illicit drugs (n = 104) and those who used other drugs (n = 733) were considered the secondary outcome group. Because of the “gateway drug” aspect of cannabis and the inclusion of those who used stronger drugs after having used cannabis, the latter group was not asked whether they had used cannabis or not (Secades-Villa et al., 2015; Balon, 2018). The no-outcome group included patients who did not use any illicit drugs.

In the question about whether they had used illicit drugs, the answers “No (at least once but did not habitually),” “No (used habitually),” “Sometimes,” and “Almost daily” were defined as “Yes, I have used,” and “Never” was defined as “No history of use.” Illicit drugs were defined as substances such as morphine and other narcotics obtained without a prescription from a physician; organic solvents such as thinner and toluene; designer drugs such as law-evading herbs and magic mushrooms; cannabis; methamphetamine; cocaine; and heroin.

Explanatory variable

We used the Fear of COVID-19 Scale (FCV-19S) as an explanatory variable. The Fear of COVID-19 Scale is a short questionnaire developed to assess fear and anxiety associated with COVID-19 (Midorikawa et al., 2021; Ahorsu et al., 2022). FCV-19S comprises seven questions in total: “I am very afraid of the novel coronavirus,” “I get uncomfortable when I think about the novel coronavirus,” “My hands sweat when I think about the novel coronavirus,” “I am afraid of losing my life due to the novel coronavirus,” “I get nervous and anxious when I see news and topics about the novel coronavirus on the Internet,” “I cannot sleep because I worry about COVID-19,” and “My heart beats faster and I have heart palpitations when I think about the COVID-19.” Each question was rated on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The total score, which ranged from 7 to 35, was calculated by summing the scores for each question. A higher score indicates a stronger fear of COVID-19. To the best of our knowledge, no studies have clearly defined the severity of fear of COVID-19 using the FCV-19S and there is no cutoff score. Therefore, we divided the severity of fear into four levels, as in a previous study (none, <17 points; mild, 17–22 points; moderate, 23–28 points; and severe, 29–35 points) (Yoshida et al., 2023).

Statistical analysis

We used a multivariate binary logistic regression model to analyze the association between the level of fear of COVID-19 in the early stages of the COVID-19 pandemic and the presence of illicit drug use. We adjusted for potential imbalances in the following key associated factors between groups with and without outcome: sex (male, female), age (15–24, 25–34, 35–44, 45–54, 55–64, 65+), occupational status (employed, student, unemployed), number of persons in household (one, two or more), annual household income (>7 million yen, 3–7 million yen, <3 million yen, do not know/do not want to answer), marital status (married, unmarried), education (ISCED ≥6, ISCED<6, others), drinking status (never, rarely, sometimes, almost daily), smoking status (never, used to smoke but quit, current smoker), Kessler Psychological Distress Scale (K6) (negative, positive), and presence of any psychiatric disorder (no, yes).

The K6 is a 6-item questionnaire that assesses stress in the past month (Kessler et al., 2002). Respondents answered each question on a 5-point scale (0 to 4), and the total score was used to evaluate their level of stress. Total scores ranged from 0 (low stress) to 24 (high stress). In this study, a score of 5 or more, representing moderate psychological stress, was defined as stressed (Sakurai et al., 2011).

To verify the uniformity of the impact of fear on drug use, subgroup analyses were conducted post-hoc. Subgroups were defined according to smoking history (yes or no) and drinking status (current or nondrinker). The statistical significance of differences between subgroups was tested using interaction terms. No adjustments were made for multiple tests.

The variance inflation factor (VIF) was used to check for multicollinearity, with a significance level of 5%. SAS software (version 9.4, SAS Institute Inc., Cary, NC, USA) was used for analysis.

Ethics approval

All procedures were conducted in accordance with the ethical standards of the 1975 Declaration of Helsinki. The research protocol was reviewed and approved by the Research Ethics Committee of the Osaka International Cancer Institute (approved June 19, 2020, approval number 20084). The Internet survey agency respects Japanese privacy laws. All participants provided electronic informed consent by reviewing and agreeing to a web-based instructional document detailing the survey prior to completing the online questionnaire. As an incentive, participants were given “E-points,” credit points that could be used for Internet shopping or redeemed for cash. The precise value of “E-points” was not disclosed by the company that administered the Internet survey (Sato et al., 2022).

Participants characteristics

Table S1 presents the participant characteristics. The total number of participants included in this analysis was 17,800, with 837 individuals in the primary outcome group of illicit drug users and 16,963 in the comparison group of non-users. In the secondary outcome group, 104 individuals used only cannabis and 733 used drugs other than cannabis (regardless of cannabis use).

The proportion of men was higher in all of the drug-use groups, while that of women was higher in the non-use group. The drug-use group had a higher percentage of younger individuals, and the most represented age group in the non-use group was those aged ≥65 years (29.1%). In terms of education, the drug-use group had a higher proportion of individuals with ISCED scores of ≥6, whereas the non-use group had more individuals with ISCED scores of <6. Regarding drinking status, the group that used only cannabis had a lower percentage of non-drinkers (10.6%), by ∼50% vs the other groups. Regarding smoking status, the group that used only cannabis had the highest proportion of ex-smokers (46.2%), whereas non-smokers comprised the highest proportion in the other groups. Occupational status, number of persons in the household, household income, marital status, K6 score, and the presence of any psychiatric disorder showed similar distributions across all of the groups.

Primary & secondary outcomes

The results of the logistic regression analysis are presented in Table S2. The distribution of fear levels regarding COVID-19 was also similar across all of the groups, with the highest proportion being “mild” (∼40%), followed by “none”, “moderate”, and “severe.”

In the group that used illicit drug, using “none” as the reference for the degree of fear of COVID-19, the adjusted odds ratios were 1.32 for “mild”, 1.42 for “moderate”, and 1.74 for “severe”—thus showing an increase with rising levels of fear. In the group that used only cannabis, no significant difference was observed in the number of cannabis users, owing to varying levels of fear. In the group that used drugs other than cannabis, the adjusted odds ratios were 1.35 for mild fear, 1.51 for moderate fear, and 1.86 for severe fear. Similarly to the primary outcome, this indicated that the odds ratios increased as the level of fear increased.

Subgroup analysis

No significant interactions were observed between the subgroups in terms of drug use, across all of the groups (Fig. S1). In the groups that used illicit drugs and those that used drugs other than cannabis, consistent odds ratios for drug use were observed across all of the subgroups. In the group that used cannabis only, no statistically significant differences were found, and the results were inconsistent.