The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic.

OBJECTIVES: The SARS-CoV-2 (COVID-19) pandemic has had profound physical and mental health effects on populations around the world. Limited empirical research has used a gender-based lens to evaluate the mental health impacts of the pandemic, overlooking the impact of public health measures on marginalized groups, such as women, and the gender diverse community. This study used a gender-based analysis to determine the prevalence of psychosocial symptoms and substance use (alcohol and cannabis use in particular) by age, ethnicity, income, rurality, education level, Indigenous status, and sexual orientation.
METHODS: Participants in the study were recruited from previously established cohorts as a part of the COVID-19 Rapid Evidence Study of a Provincial Population-Based Cohort for Gender and Sex (RESPPONSE) study. Those who agreed to participate were asked to self-report symptoms of depression, anxiety, pandemic stress, loneliness, alcohol use, and cannabis use across five phases of the pandemic as well as retrospectively before the pandemic.
RESULTS: For all psychosocial outcomes, there was a significant effect of time with all five phases of the pandemic being associated with more symptoms of depression, anxiety, stress, and loneliness relative to pre-COVID levels (p < .0001). Gender was significantly associated with all outcomes (p < .0001) with men exhibiting lower scores (i.e., fewer symptoms) than women and gender diverse participants, and women exhibiting lower scores than the gender diverse group. Other significant predictors were age (younger populations experiencing more symptoms, p < .0001), ethnicity (Chinese/Taiwanese individuals experiencing fewer symptoms, p = .005), and Indigenous status (Indigenous individuals experiencing more symptoms, p < .0001). Alcohol use and cannabis use increased relative to pre-pandemic levels, and women reported a greater increase in cannabis use than men (p < .0001).
CONCLUSIONS: Our findings highlight the need for policy makers and leaders to prioritize women, gender-diverse individuals, and young people when tailoring public health measures for future pandemics.

Introduction

In the first eighteen months of the SARS-CoV-2 (COVID-19) pandemic, there have been over 150 million cases and over 3 million deaths attributable to the upper respiratory virus [1]. More specifically, Canada has reached a stark milestone of one million cases and over 20,000 deaths in a little over a year (May 2021) [2]. Although the physical health effects of the virus tend to dominate the literature and the media, it is well established that outbreaks, including the current pandemic, have significant impacts on the mental health of those involved. For example, healthcare workers and patients affected by previous outbreaks such as SARS-CoV-1 [3], H1N1 influenza [4], and Ebola [5] have poorer psychosocial outcomes during the onset of societal alarm.

Public health measures put in place due to the COVID-19 pandemic have had a negative impact on the mental health of peoples worldwide [6,7]. Levels of depression [8], anxiety [8], loneliness [9], alcohol use [10,11], and cannabis use [12] have all increased relative to pre-pandemic levels. Additionally, there is mounting evidence highlighting the secondary effects of public health measures on specific populations during the pandemic [13]. For example, younger populations [14] and those of lower income [14] have experienced disproportionate psychosocial outcomes because of the COVID-19 pandemic.

There is a growing realization that a gender lens needs to be applied to COVID-19 research, not only regarding biomedical outcomes, but for psychosocial outcomes as well [15]. This aligns with increasing efforts, across North America, to include sex and gender based analyses in all research. Sex is defined as birth assignment and is usually established by genital anatomy at birth with female, male, and intersex as typical response options in queries about sex. Gender identity is defined as one’s personal feelings about being a woman, man, transgender, gender-diverse individual, or another expression of gender that does not align with that person’s birth assigned sex. When sex is considered in the context of psychosocial issues, it is well established that females are more likely to present symptoms of depression and anxiety in general [16], and face greater job losses than males during the COVID-19 pandemic [17]. Thus, it is not surprising that studies to date have found that females reported more anxiety, depressive symptoms, and post-traumatic stress symptoms relative to males during the COVID-19 pandemic [14,18-20]. Age also plays a large role in sex differences in the risk for neuropsychiatric disorders [21], but thus far the interaction between age and sex has received little attention with regards to how age may interact with sex to impact psychosocial outcomes throughout COVID-19.

In addition to the paucity of sex-based analyses, studies examining psychosocial outcomes from the standpoint of participants’ self-identified gender are sparse. Most of the research on gender and the COVID-19 pandemic have compared responses between women and men, while ignoring the experiences of individuals who experience gender on a spectrum, beyond the binary classification of man and woman. A recent cross-sectional survey by Hawke et al. [22] found that despite no clear significant differences in mental health between cisgender, transgender and gender diverse youth before the pandemic, those identifying as gender diverse were two times more likely to report experiencing mental health challenges relative to the cisgender group during COVID-19. These findings were associated with an unmet need for mental health and substance use services. However, this study was limited to those aged 14–28, thereby reducing the generalizability of the findings to the larger population. To our knowledge, no studies have taken a gender-based approach to understanding the mental health sequelae of COVID-19 pandemic control measures across a general population sample. Given the current data as well as previous findings on the poor mental health outcomes of gender diverse individuals [23], focused empirical attention on this population is critical.

Other social determinants of health, including education, ethnicity, and income, impact physical and mental health outcomes [24] and have shaped the risk and consequences of COVID-19 in communities across North America [25]. Additionally, minority stress theory has posited that those who are part of more than one marginalized societal group may experience even greater health disparities [26]. Given this, we posited that it would be crucial to explore how gender interacted with these social determinants to influence mental health, in particular also because these social factors might moderate the effects of gender.

Many governments have tailored public health interventions throughout the pandemic based on infection incidence and hospitalization rates, resulting in a series of lockdowns (and prescribed regulations), followed by periods of relaxed restrictions, which have generated defined “phases” of the pandemic. While it is now widely known that lockdowns impact mental health [27], what remains unclear is how the tightening and easing of these social restrictions impacts psychosocial factors, by gender. As such, we sought to assess our psychosocial outcomes, cross-sectionally, across various phases of the pandemic retrospectively aligned with provincial changes in public health orders.

We were particularly interested in self-reported symptoms of depression, anxiety, stress, and loneliness, given recent reports of how these have worsened over the pandemic [8,9], and in addition to our primary interest in how gender impacted these outcomes, we also examined the interaction of gender with age, ethnicity, income, education, rurality, and sexual orientation. We predicted higher scores (i.e., more symptomatic) on depression, anxiety, stress, and loneliness in women and gender diverse individuals compared to men, and that this would be influenced by age, ethnicity, and income. We also hypothesized that during phases of increased social restriction psychosocial symptoms would increase. Our secondary interest was in self reports of alcohol and cannabis use, again given data on the pandemic’s effects of substance use behaviors [10-12], and again, we explored the impact of gender, as well as how gender interacted with various social variables, on alcohol and cannabis use. Our analyses of gender and alcohol and cannabis use were exploratory. Using a large cohort of the general population in British Columbia (BC), we assessed participants cross-sectionally, and asked them to retrospectively report on these outcomes across different phases of the pandemic, which again corresponded to stages of pandemic control changes in the province of BC.

Materials and methods

Participant recruitment and study design

Participants, aged 25–69 years, were invited to participate in this study from previously established cohorts from research teams at the Women’s Health Research Institute, representing both general and priority populations of BC who had consented to being contacted for future research [28]. The original cohorts represented healthy women aged 25–65, women living with HIV/AIDS, men and women over age 18 living with a complex chronic disease, and individuals over age 18 with pelvic pain and/or endometriosis [28]. The largest cohort was recruited to reflect a broad and representative sample of BC women. Participants were stratified into nine five-year age strata, and using a SARS-CoV-2 population seroprevalence of 2% (±1, 95% CI), the target recruitment for each stratum was 750. The seroprevalence statistic was used to target recruitment for analyses in a separate manuscript.

Those identified as potentially eligible from the established cohorts (Index Participants) were sent an email invitation to participate via an online survey. To improve the sample size and gender diversity of the study sample, Index Participants were asked to pass the invitation on to one household member who identified as a different gender as the respondent (Household Participants). All potential participants were sent up to three email reminders to participate in the study. The inclusion criteria were: current residents of BC, aged 25–69, any gender, and able and willing to fill out the online survey in English. Ethics approval was obtained from the BC Children’s and Women’s Research Ethics board, and all participants provided consent to participate. Survey responses were collected anonymously, with the exception of three-digit postal codes, which were used to determine rurality for analyses.

After two months of data collection from existing research cohorts, recruitment was expanded in order to meet our target sample of n = 750 per age cohort. This expanded recruitment included participants obtained from public recruitment through the REACH BC platform, social media (i.e., Facebook, Twitter and Instagram), posts on the Women’s Health Research Institute website (www.whri.org) and engagement of community groups who are affiliated with the Women’s Health Research Institute, a provincial research institute focused on gender and women’s health. All respondents in the study were invited to enter a draw to win a $100 e-gift card for completing the survey. Recruitment was continued until a target of n = 750 was reached for each of the nine age-based strata, with the exception of the 25–29 year age group. Recruitment was open from August 20, 2020 –March 1, 2021.

Survey design and measures

The survey was tested for face validity, pilot tested, and a final version was designed using REDCap (Research Electronic Data Capture) [29]. While the survey consisted of multiple modules, this study focuses solely on the outcomes from the psychosocial module, which included questions about mental health outcomes such as depression, anxiety, stress, loneliness, alcohol use, and cannabis use.

Demographic information was collected from all respondents including age, sex, gender, sexual orientation, ethnicity, Indigenous status, income, education level, if the participant was currently a student and rurality by postal code. Sex referred specifically to the sex assigned at birth and included the option of male, female or intersex. Gender referred to the respondent’s current gender identity and included the options man, woman, or another option grouping non-binary, transgender, GenderQueer, agender or any other similar identity together. Sexual orientation options included asexual, bisexual, demisexual, gay/lesbian, heterosexual, or pansexual. Participants were given the option to identify as the following ethnicities: White, Chinese/Taiwanese, Black (African, Caribbean, or Other), South Asian (e.g., Indian, Bangladeshi, Pakistani, Punjabi, and Sri Lankan), and several other ethnicities who were analyzed an “Other” category. Indigenous status was assessed separately from ethnicity. Self-reporting of Indigenous status provided participants the option to identify as First Nation, Metis, Inuit, non-status First Nations, other Indigenous or not Indigenous, and they were then asked about Two Spirit identity. Rurality was determined based on three-digit postal codes and were classified into one of the follow categories: census metropolitan area, strong metropolitan influence zone, moderate metropolitan zone, or weak to no metropolitan influence zone.

The study design was cross-sectional in nature, whereby participants completed the survey at one time point. However, several questions asked participants to retrospectively refer to different periods of time: pre-pandemic (before March 2020) as well as across five different phases of the pandemic in BC that corresponded with changes in the public health orders regarding social distancing in the province of BC. In the original version of the survey, Phase 1 lasted from mid-March 2020 to mid-May 2020, Phase 2 lasted from mid-May 2020 to mid-June 2020, and Phase 3 lasted from mid-June 2020 until the end of November 2020. Given that data collection continued past November 2020, we added Phases 4 and 5, as well as modified dates for Phases 2 and 3 (mid-May to end of August 2020; denoted by Phase 2/3_2). Phase 4 lasted from September 2020 to the end of October 2020 and Phase 5 lasted from November 2020 to the date our survey closed (March 1, 2021). We have included a S1 Table that explains the public health recommendations in more detail, through every phase of the pandemic in BC.

Depression

Depression was measured across the phases of the pandemic using the Patient Health Questionnaire (PHQ-9). The PHQ-9 questionnaire was used to measure self-reported symptoms of depression on a Likert scale from zero (not at all) to three (nearly everyday). Scores for this questionnaire range from 0–27 with a score of 0–4 indicating minimal depression, 5–14 indicating mild to moderate depression and 15–27 indicating moderately severe to severe depression [30]. The PHQ-9 has been validated across age and gender, as well as among diverse populations [30,31]. Internal consistency across data collection and Cronbach’s alpha for the PHQ-9 in the current sample was very good at α = 0.848.

Anxiety

Anxiety was measured across the phases of the pandemic using the Generalized Anxiety Disorder questionnaire (GAD-7). The GAD-7 was provided to respondents to self-report feelings of anxiety on a Likert scale from zero (not at all) to three (nearly everyday). Scores for this questionnaire range from 0–21 with scores above 10 indicating a clinical diagnosis for anxiety [32]. The GAD-7 has been validated in the general population and is frequently used in primary care settings to screen for anxiety symptoms[33]. Internal consistency across data collection and Cronbach’s alpha for the GAD-7 questionnaire in the current sample was very good at α = 0.889.

Pandemic stress

General pandemic stress was measured across the phases of the pandemic using the CoRonavIruS Health Impact Survey (CRISIS) V0.3. This survey was developed and validated early in the COVID-19 pandemic to provide a general measure of mental distress and resilience [34]. The CRISIS is found to have strong validity and reliability, and has been recommended for use in population-based studies of mental health during COVID-19. Participants were asked to self-report feelings of stress on a Likert scale from one (not at all) to five (extremely). Scores for this questionnaire range from 10–50 with higher scores indicating greater COVID-related stress. Internal consistency across data collection and Cronbach’s alpha for CRISIS in the current sample was very good at α = 0.882.

Loneliness

Loneliness was also measured across the phases of the pandemic where respondents were asked to self-report feelings of loneliness on a Likert scale from one (not lonely at all) to five (extremely lonely). This item was taken from the validated Coronavirus Health and Impact Survey (CRISIS), where individual items on the CRISIS have been found to have high Intraclass Correlation Coefficients [34]. Previous studies have found loneliness to be positively correlated with both PHQ-9 and GAD-7 scores [35] and to be a significant predictor of suicide [36,37].

Alcohol use

Change in alcohol use was asked for all post-COVID time points (i.e., Has your consumption of alcohol changed since March 2020?). Change in alcohol use was defined as “none” (which included no alcohol use, decreased alcohol use, and same alcohol use) vs. increased alcohol use. Therefore, a single, non-time-varying alcohol change variable was created and used to compare the retrospective responses across the different time points, with time.

Cannabis use

Change in cannabis use was asked for all post-COVID time points (Has your consumption of cannabis changed since March 2020?). As with alcohol, change in cannabis use was defined as “none” (which included no cannabis use, decreased cannabis use, and same cannabis use) vs. increased cannabis use. A single, non-time-varying cannabis change variable was created and used in a longitudinal model with time.

Statistical analyses

Analyses were carried out using R v.4.0.3. Analyses of psychosocial outcomes across the pandemic control phases were conducted using mixed-effects linear regression models with individual and household IDs as random effects. This allows for correlations among individuals in the same household, and separately, correlations over time among responses within the same individual, allowing for a longitudinal assessment. We included pairwise interactions to assess non-additive effects between age and gender, and age and ethnicity, sexual orientation, income, and Indigenous status. Significance was assessed using likelihood-ratio tests, and interactions were removed from the models if non-significant at p < .05. Post-hoc pairwise tests were conducted to further explore main or interaction effects with Bonferroni correction for multiple tests.

To explore associations between increase in alcohol and cannabis use with sex/gender and other demographic variables we used mixed-effects logistic regressions with household ID as a random effect. We also examined increase in alcohol and cannabis use and psychosocial outcomes across the phases as described above. Interactions and post-hoc tests were handled as above. Missing data were excluded from analyses.

Results

Survey participants

Between August 2020 and March 2021, 16,056 survey invites were emailed to prospective Index Participants and 1,872 participants were recruited from the public, for a total of 17,928 prospective participants. Of these participants, a total of 5,415 responded to the invitation to participate in the study and met the analysis inclusion criteria (Fig 1). Of these participants, 1,434 forwarded the survey invitation to a household member of a different sex or gender and we received 661 participants via this method. The present analyses includes the 6,076 Index and Household participants who completed psychosocial measures of anxiety, depression, stress, and loneliness.

Fig 1

A flow diagram of prospective participants and respondents to the study.

Demographic characteristics of participants

A total of 6,426 individuals responded to the question about sex; there were n = 820 males (12.7%) and n = 5,606 females (87.1%). A total of 6,076 responded to the question about gender; including men (n = 750; 12.3%), women (n = 5,254; 86.4%), and gender diverse (n = 72; 1.2%) individuals. Table 1 presents the demographic characteristics of the sample by gender, according to women, men, and gender diverse.

Table 1

Demographic information of survey respondents.

Total N = 6,076	Women n = 5,254	Men n = 750	Gender Diverse n = 72
Sex N(%)
Male	9 (0.2%)	744 (99.2%)	6 (8.3%)
Female	5,243 (99.8%)	5 (0.7%)	62 (86.1%)
Age M (SD)	48.5 (±12.0)	48.5 (±12.2)	42.4 (±11.6)
25–29	348 (6.6%)	68 (9.1%)	8 (11.1%)
30–39	1,075 (20.5%)	133 (17.7%)	29 (40.3%)
40–49	1,291 (24.6%)	164 (21.9%)	14 (19.4%)
50–59	1,333 (25.4%)	212 (28.3%)	11 (15.3%)
60–69	1,207 (23.0%)	173 (23.1%)	10 (13.9%)
Sexual Orientation (%)
Heterosexual	4,480 (85.3%)	674 (89.9%)	7 (9.7%)
Non-Heterosexual	757 (14.4%)	74 (9.9%)	65 (90.3%)
Ethnicity (%)
White	4,265 (81.2%)	609 (81.2%)	44 (61.1%)
Black	28 (0.5%)	6 (0.8%)	1 (1.4%)
Chinese/Taiwanese	311 (5.9%)	40 (5.3%)	2 (2.8%)
South Asian	123 (2.3%)	9 (1.2%)	1 (1.4%)
Other ethnicity	504 (9.6%)	81 (10.8%)	20 (27.8%)
Indigenous Status (%)
Indigenous	166 (3.2%)	27 (3.6%)	11 (15.3%)
Not Indigenous	4,830 (91.9%)	696 (92.8%)	59 (81.9%)
Current Student (%)
Yes	311 (5.9%)	37 (4.9%)	14 (19.4%)
No	4,939 (94.0%)	711 (94.8%)	58 (80.6%)
Level of education (%)
High school or less	641 (12.2%)	92 (12.3%)	13 (18.1%)
More than high school	4,605 (87.6%)	658 (87.7%)	58 (80.6%)
Current Household Income (%)
<$10,000 –$20,000	128 (2.4%)	14 (1.9%)	7 (9.7%)
$20,000 –$40,000	269 (5.1%)	23 (3.1%)	14 (19.4%)
$40,000 –$60,000	458 (8.7%)	32 (4.3%)	10 (13.9%)
$60,000 –$80,000	488 (9.3%)	56 (7.5%)	5 (6.9%)
$80,000 –$100,000	642 (12.2%)	67 (8.9%)	8 (11.1%)
$100,000 –$150,000	1,104 (21.0%)	156 (20.8%)	14 (19.4%)
>$150,000	1,273 (24.2%)	304 (40.5%)	6 (8.3%)
Rurality
Census metropolitan area	4,957 (94.3%)	716 (95.5%)	70 (97.3%)
Strong metropolitan influence zone	79 (1.4%)	7 (0.9%)	0 (0.0%)
Moderate metropolitan influence zone	116 (2.0%)	8 (1.1%)	1 (1.4%)
Weak to No metropolitan influence zone	46 (0.8%)	5 (0.67%)	0 (0.0%)

Note: Values do not add up to 100% due to missing data.

M refers to the mean age and (SD) standard deviation of participants.

Effect of pandemic phase, age, ethnicity and gender and sex on psychosocial outcomes

Controlling for household income, we found no significant interactions between age and gender, age and sex, age and ethnicity or rurality on any of the psychosocial measures. For all psychosocial outcomes, there was a significant relationship with pandemic phase (all p < .0001, Table 2), with the greatest increases in mental health symptoms in Phase 1 compared to pre-COVID. The scores in all subsequent phases remained significantly higher (i.e., more symptoms) than in the pre-COVID phase across all outcomes (Figs 2–5, Table 2). Gender was significantly associated with all outcomes (all p < .0001, Figs 2–5, Table 2), and pairwise comparisons showed that men had lower scores than both women and gender-diverse participants, while women had lower scores than the gender-diverse participants. Age was significantly negatively associated with all the outcomes, with older participants having lower scores on average (i.e., fewer psychosocial symptoms) (p < .0001, Table 2). Finally, there was a significant relationship between ethnicity and all outcomes (GAD-7 and PHQ-9 p < .0001, CRISIS and Loneliness p = .005, Table 2), with scores lower in Chinese/Taiwanese participants compared to the White, South Asian, and Other ethnicity participants. When sex was included in the model in place of gender, there were no differences to the findings, indicating the overlap in our participants self-reported sex and gender. Given our intention to explore outcomes separately for gender-diverse individuals, all subsequent analyses were done by gender (not sex).

Table 2

Impact of sociodemographic factors and pandemic phase on psychosocial outcomes.

	Anxiety (GAD-7)		Depression (PHQ-9)		Pandemic Stress (CRISIS)		Loneliness
Predictors	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI
(Intercept)	8.72**	7.97 – 9.47	9.58**	8.81 – 10.35	29.35**	28.18 – 30.51	2.56**	2.40 – 2.73
Age (per year increase)	-0.1**	-0.10 –-0.09	-0.08**	-0.09 – -0.07	-0.15**	-0.16 – -0.13	-0.01**	-0.01 – -0.01
Phase
Pre-COVID	Reference **		Reference **		Reference **		Reference **
Phase 1	2.92	2.82 – 3.01	2.38	2.28 – 2.48	7.31	7.13 – 7.49	0.62	0.59 – 0.64
Phase 2/3	2.09	1.98 – 2.19	1.98	1.88 – 2.09	5.34	5.15 – 5.53	0.47	0.45 – 0.50
Phase 2/3_2	1.61	1.40 – 1.83	1.56	1.34 – 1.78	4.42	4.03 – 4.81	0.42	0.37 – 0.48
Phase 4	2.16	1.95 – 2.37	2.44	2.23 – 2.66	6.08	5.69 – 6.47	0.63	0.58 – 0.69
Phase 5	2.53	2.32 – 2.74	2.99	2.78 – 3.21	7.31	6.92 – 7.71	0.91	0.85 – 0.96
Household Income
<$10K to $20K	Reference **		Reference **		Reference **		Reference **
$20K to $40K	0.51	-0.23 – 1.25	-0.56	-1.33 – 0.20	-0.29	-1.45 – 0.86	-0.2	-0.37 – -0.04
$40K to $60K	0.04	-0.66 – 0.73	-0.93	-1.65 – -0.21	-1.12	-2.20 – -0.04	-0.31	-0.46 – -0.15
$60K to $80K	-0.62	-1.31 – 0.07	-1.93	-2.65 – -1.22	-2.56	-3.63 – -1.49	-0.47	-0.62 – -0.32
$80K to $100K	-1.03	-1.70 – -0.35	-2.43	-3.13 – -1.74	-2.99	-4.03 – -1.95	-0.57	-0.71 – -0.42
$100K to $150K	-1.13	-1.78 – -0.49	-2.66	-3.33 – -1.99	-3.47	-4.47 – -2.47	-0.69	-0.83 – -0.55
>$150K	-1.55	-2.19 – -0.91	-3.38	-4.04 – -2.71	-4.42	-5.41 – -3.42	-0.8	-0.94 – -0.66
Gender
Women	Reference **		Reference **		Reference **		Reference **
Men	-1.44	-1.75 – -1.13	-1.25	-1.57 – -0.93	-2.28	-2.75 – -1.80	-0.22	-0.29 – -0.15
Gender Diverse	1.67	0.71 – 2.64	2.09	1.11 – 3.07	3.21	1.72 – 4.71	0.34	0.12 – 0.55
Ethnicity
White	Reference **		Reference **		Reference *		Reference *
Black	-0.4	-1.74 – 0.95	0.51	-0.87 – 1.89	-0.19	-2.32 – 1.94	0.06	-0.24 – 0.36
Chinese/Taiwanese	-0.98	-1.45 – -0.52	-1.36	-1.84 – -0.88	-1.14	-1.87 – -0.42	-0.18	-0.28 – -0.08
South Asian	0.43	-0.33 – 1.18	0.08	-0.69 – 0.85	1.01	-0.18 – 2.19	0.09	-0.07 – 0.26
Other	0.38	0.02 – 0.73	0.14	-0.22 – 0.51	0.34	-0.21 – 0.89	0.03	-0.05 – 0.11
Indigenous Status Across Phases
Pre-COVID	Reference **		Reference **		Reference **		Reference **
Phase 1	2.87	2.77–2.97	2.33	2.23–2.44	7.26	7.08–7.45	0.60	0.58–0.63
Phase 2/3	2.06	1.95–2.17	1.94	1.83–2.05	5.28	5.08–5.47	0.46	0.43–0.49
Phase 2/3_2	1.57	1.34–1.79	1.46	1.23–1.69	4.31	3.90–4.72	0.41	0.35–0.46
Phase 4	2.14	1.91–2.36	2.36	2.13–2.59	5.98	5.57–6.39	0.61	0.55–0.67
Phase 5	2.53	2.30–2.75	2.94	2.71–3.17	7.34	6.93–7.75	0.88	0.82–0.94
Non-Heterosexual Orientation Across Phases
Pre-COVID	Reference **		Reference **		Reference **		Reference **
Phase 1	2.81	2.70–2.92	2.28	2.17–2.38	7.21	7.01–7.40	0.58	0.55–0.61
Phase 2/3	2.00	1.89–2.12	1.89	1.77–2.00	5.25	5.04–5.46	0.45	0.42–0.48
Phase 2/3_2	1.56	1.33–1.80	1.45	1.21–1.69	4.36	3.93–4.79	0.42	0.36–0.48
Phase 4	2.17	1.93–2.40	2.40	2.16–2.64	6.13	5.70–6.56	0.63	0.57–0.70
Phase 5	2.58	2.35–2.81	2.96	2.72–3.20	7.41	6.97–7.84	0.91	0.85–0.97
Marginal R2 /Conditional R2	0.173 / 0.716		0.164 / 0.720		0.255 / 0.674		0.149 / 0.623

CI refers to confidence intervals for the adjusted estimates. Pre-COVID: Prior to mid-March 2020; Phase 1: Mid-March 2020 to mid-May 2020; Phase: 2/3: Mid-May 2020 to November 2020; Phase 2/3_2: Mid-May 2020 to August 2020; Phase 4: September 2020 to October 2020; Phase 5: November 2020 to March 1, 2021. GAD-7: Generalized Anxiety Disorder measure; PHQ-9: Patient Health Questionnaire; CRISIS: CoRonavIruS Health Impact Survey; Loneliness was measured using a single item.

*p = .005

**p < .0001.

Fig 2

Depressive symptoms, as measured by the PHQ-9, across phases of the pandemic.

Data points refer to mean scores of the given psychosocial measure, error bars refer to the standard error. Pre-COVID: Prior to mid-March 2020; Phase 1: Mid-March 2020 to mid-May 2020; Phase: 2/3: Mid-May 2020 to November 2020; Phase 2/3_2: Mid-May 2020 to August 2020; Phase 4: September 2020 to October 2020; Phase 5: November 2020 to March 1, 2021.

Fig 5

Loneliness symptoms across phases of the pandemic.

Data points refer to mean scores of the given psychosocial measure, error bars refer to the standard error. Pre-COVID: Prior to mid-March 2020; Phase 1: Mid-March 2020 to mid-May 2020; Phase: 2/3: Mid-May 2020 to November 2020; Phase 2/3_2: Mid-May 2020 to August 2020; Phase 4: September 2020 to October 2020; Phase 5: November 2020 to March 1, 2021.

Anxiety symptoms, as measured by the GAD-7, across phases of the pandemic.

Pandemic stress symptoms, as measured by the CoRonavIruS Health Impact Survey (CRISIS), across phases of the pandemic.

Psychosocial outcomes by indigenous status

Controlling for household income, there was no significant interaction between Indigenous status and age or gender. There was a significant interaction between Indigenous status and time for all four psychosocial outcomes (p < .0001, Table 2) and follow-up post-hoc pairwise tests suggest that at all time points except pre-COVID, those who identified as Indigenous had significantly higher GAD-7, PHQ-9, CRISIS, and loneliness scores (i.e., more mental health symptoms) than those who did not identify as Indigenous.

Psychosocial outcomes by sexual orientation

Across all outcomes, the non-heterosexual group (which included asexual, bisexual, demisexual, gay/lesbian, pansexual, and other) had significantly more mental health symptoms than the heterosexual group for all phases, and the magnitude of the difference between the groups was largest in Phase 1 of the pandemic.

Associations between psychosocial outcomes and alcohol by gender

A total of 23.3% of the sample reported an increase in alcohol use. Increased alcohol use was negatively associated with age (p < .001, Table 3), with older participants having lower odds of increased alcohol use. There was no significant difference among genders in the odds of increased alcohol use, but there was a trend of increasing odds as household income increased. Additionally, those residing in census metropolitan areas were found to have increased their alcohol use relative to those outside of these dense urban areas (p = .03, Table 3).

Table 3

Changes in alcohol and cannabis use across sociodemographic factors.

	Change in Alcohol Use		Change in Cannabis Use
Predictors	OR	95% CI	OR	95% CI
Age	0.98**	0.97 – 0.98	0.97**	0.96 – 0.99
Household income
<$10K to $20K	Reference *		Reference
$20K to $40K	1.16	0.64 – 2.11	1.5	0.67 – 3.37
$40K to $60K	1.19	0.68 – 2.08	1.32	0.60 – 2.91
$60K to $80K	1.13	0.66 – 1.95	1.67	0.76 – 3.70
$80K to $100K	1.28	0.75 – 2.18	1.39	0.64 – 3.03
$100K to $150K	1.45	0.86 – 2.43	1.24	0.59 – 2.61
>$150K	1.62	0.97 – 2.70	1.17	0.55 – 2.48
Rurality
Census metropolitan area	Reference *		Reference
Strong metropolitan influence zone	0.93	0.48 – 1.82	0.61	0.19 – 2.00
Moderate metropolitan influence zone	1.42	0.85 – 2.36	0.84	0.37 – 1.91
Weak to No metropolitan influence zone	0.24	0.07 – 0.81	2.15	0.48 – 9.61
Gender
Women	Reference		Reference *
Men	0.86	0.69 – 1.06	0.66	0.43 – 1.02
Gender Diverse	0.72	0.32 – 1.61	0.84	0.34 – 2.11
Ethnicity
White	Reference		Reference
Black	1.17	0.49 – 2.82	2.42	0.66 – 8.89
Chinese/Taiwanese	0.68	0.45 – 1.02	0.57	0.24 – 1.33
South Asian	0.89	0.52 – 1.53	0.38	0.10 – 1.40
Other	0.94	0.72 – 1.22	1.56	1.02 – 2.37

OR refers to the odds ratio, CI refers to the confidence intervals for the OR.

* p = .03

** p < .001.

Controlling for household income, and across all psychosocial outcomes, there was no interaction between gender and increased alcohol use, suggesting that the differences among genders in these psychosocial variables was the same between those who did and did not increase alcohol use since the start of the pandemic (Table 4). There was a significant interaction between increased alcohol use and pandemic phase (all p < .0001, Table 4). Pairwise tests indicated that at all phases, with the exception of pre-COVID, those who reported increased alcohol use had significantly more psychosocial symptoms on all measures (p < .0001, Table 4).

Table 4

Psychosocial outcomes and sociodemographic factors for those that reported an increase in alcohol use.

	Anxiety (GAD-7)		Depression (PHQ-9)		Pandemic Stress (CRISIS)		Loneliness
Predictors	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI
(Intercept)	8.64*	7.91 – 9.38	9.39*	8.63 – 10.15	29.39*	28.25 – 30.54	2.56*	2.40 – 2.72
Age (per year increase)	-0.09*	-0.10 – -0.08	-0.07*	-0.08 – -0.06	-0.14*	-0.16 – -0.13	-0.01*	-0.01 – -0.01
Phase
Pre-COVID	Reference *		Reference *		Reference *		Reference *
Phase 1	2.64	2.53 – 2.75	2.13	2.01 – 2.24	6.74	6.53 – 6.94	0.58	0.55 – 0.60
Phase 2/3	1.9	1.78 – 2.02	1.77	1.65 – 1.89	4.91	4.69 – 5.13	0.44	0.41 – 0.47
Phase 2/3_2	1.42	1.17 – 1.66	1.36	1.11 – 1.61	4.01	3.56 – 4.47	0.38	0.32 – 0.45
Phase 4	1.91	1.67 – 2.16	2.26	2.01 – 2.51	5.62	5.16 – 6.07	0.59	0.52 – 0.65
Phase 5	2.22	1.98 – 2.47	2.71	2.46 – 2.96	6.78	6.33 – 7.23	0.89	0.82 – 0.95
Household Income
<$10K to $20K	Reference *		Reference *		Reference *		Reference *
$20K to $40K	0.51	-0.23 – 1.25	-0.58	-1.34 – 0.18	-0.31	-1.46 – 0.84	-0.21	-0.37 – -0.05
$40K to $60K	-0.02	-0.71 – 0.67	-0.97	-1.69 – -0.25	-1.17	-2.25 – -0.09	-0.31	-0.46 – -0.16
$60K to $80K	-0.69	-1.37 – 0.00	-1.97	-2.68 – -1.26	-2.63	-3.70 – -1.56	-0.48	-0.63 – -0.33
$80K to $100K	-1.07	-1.74 – -0.40	-2.46	-3.15 – -1.77	-3.02	-4.06 – -1.98	-0.57	-0.72 – -0.43
$100K to $150K	-1.22	-1.86 – -0.57	-2.72	-3.39 – -2.05	-3.58	-4.58 – -2.58	-0.7	-0.84 – -0.56
>$150K	-1.68	-2.32 – -1.04	-3.49	-4.16 – -2.83	-4.58	-5.57 – -3.58	-0.81	-0.95 – -0.67
Gender
Women	Reference *		Reference *		Reference *		Reference *
Men	-1.39	-1.70 – -1.08	-1.19	-1.51 – -0.87	-2.26	-2.73 – -1.78	-0.22	-0.29 – -0.15
Gender Diverse	1.8	0.86 – 2.74	2.29	1.34 – 3.24	3.44	1.99 – 4.90	0.34	0.13 – 0.54
Increased alcohol use	0.08	-0.20 – 0.36	0.07	-0.21 – 0.36	-0.38	-0.83 – 0.08	-0.03	-0.10 – 0.03
Phase x Increased alcohol use
Pre-COVID	Reference *		Reference *		Reference *		Reference *
Phase 1	1.12	0.89 – 1.35	1.06	0.83 – 1.29	2.36	1.94 – 2.78	0.16	0.10 – 0.22
Phase 2/3	0.79	0.54 – 1.03	0.9	0.66 – 1.15	1.76	1.31 – 2.21	0.13	0.07 – 0.20
Phase 2/3_2	0.79	0.31 – 1.27	0.77	0.29 – 1.26	1.67	0.80 – 2.55	0.14	0.02 – 0.27
Phase 4	0.98	0.51 – 1.46	0.72	0.24 – 1.21	1.93	1.05 – 2.80	0.15	0.03 – 0.28
Phase 5	1.23	0.75 – 1.70	1.07	0.58 – 1.55	2.18	1.30 – 3.05	0.08	-0.04 – 0.20
Marginal R2 /Conditional R2	0.175 / 0.719		0.165 / 0.722		0.259 / 0.678		0.148 / 0.624

CI refers to confidence intervals for the adjusted estimates. Pre-COVID: Prior to mid-March 2020; Phase 1: Mid-March 2020 to mid-May 2020; Phase: 2/3: Mid-May 2020 to November 2020; Phase 2/3_2: Mid-May 2020 to August 2020; Phase 4: September 2020 to October 2020; Phase 5: November 2020 to March 1, 2021. GAD-7: Generalized Anxiety Disorder measure; PHQ-9: Patient Health Questionnaire; CRISIS: CoRonavIruS Health Impact Survey; Loneliness was measured using a single item.

*p < .0001.

Associations between psychosocial outcomes and cannabis use by gender

A total of 5.9% of the sample reported an increase in cannabis use since the start of the pandemic. Increased cannabis use was negatively associated with age (p < .001, Table 3), with older participants having lower odds of increased use. There was a significant relationship with gender (p = .02, Table 3, Fig 6), where women had a significantly higher odds of increased cannabis use compared to men, and there was no significant difference between men and gender diverse, and women and gender diverse groups.

Fig 6

Cannabis use across different genders.

Controlling for household income, there was a significant interaction between change in cannabis use and pandemic phase (p < .0001 for GAD-7, PHQ-9, and CRISIS, p = .04 for Loneliness, Table 5). Post-hoc pairwise tests suggest that across all phases, including pre-COVID, those who increased cannabis use had significantly higher anxiety, more depressive symptoms, and higher COVID-stress scores than those who did not have increased cannabis use. Loneliness scores were significantly higher across all phases of the pandemic for those who increased cannabis use compared to those who did not. There was no interaction between gender and increased cannabis use for GAD-7, PHQ-9, or CRISIS scores. However, there was a significant interaction between gender and increased cannabis use on Loneliness (p = .008, Table 5). For both men and women, those who increased cannabis use had more loneliness symptoms than those who did not have increased cannabis use. Conversely, among the gender-diverse participants, there was no difference in loneliness between those who increased cannabis, and those who did not.

Table 5

Psychosocial outcomes and sociodemographic factors for those that reported an increase in cannabis.

	Anxiety (GAD-7)		Depression (PHQ-9)		Pandemic Stress (CRISIS)		Loneliness
Predictors	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI	Estimates	95% CI
(Intercept)	8.29***	7.56 – 9.03	9.02***	8.26 – 9.77	28.81***	27.67 – 29.95	2.51***	2.35 – 2.67
Age (per year increase)	-0.09***	-0.10 – -0.08	-0.07***	-0.08 – -0.06	-0.14***	-0.15 – -0.12	-0.01***	-0.01 – -0.01
Phase
Pre-COVID	Reference ***		Reference ***		Reference ***		Reference ***
Phase 1	2.82	2.72 – 2.93	2.29	2.19 – 2.39	7.2	7.02 – 7.39	0.6	0.58 – 0.63
Phase 2/3	2	1.89 – 2.11	1.9	1.79 – 2.01	5.23	5.03 – 5.43	0.46	0.44 – 0.49
Phase 2/3_2	1.5	1.27 – 1.72	1.44	1.21 – 1.67	4.2	3.79 – 4.61	0.41	0.35 – 0.47
Phase 4	2	1.78 – 2.23	2.26	2.03 – 2.49	5.85	5.44 – 6.26	0.62	0.56 – 0.67
Phase 5	2.38	2.16 – 2.61	2.81	2.59 – 3.04	7.14	6.72 – 7.55	0.89	0.83 – 0.95
Household Income								0.
<$10K to $20K	Reference ***		Reference ***		Reference ***		Reference ***
$20K to $40K	0.44	-0.30 – 1.17	-0.68	-1.44 – 0.08	-0.43	-1.57 – 0.72	-0.21	-0.37 – -0.05
$40K to $60K	-0.04	-0.72 – 0.65	-1	-1.71 – -0.29	-1.23	-2.31 – -0.16	-0.31	-0.46 – -0.16
$60K to $80K	-0.65	-1.34 – 0.03	-1.95	-2.65 – -1.24	-2.61	-3.67 – -1.55	-0.48	-0.63 – -0.33
$80K to $100K	-1.02	-1.68 – -0.35	-2.42	-3.11 – -1.74	-2.98	-4.02 – -1.95	-0.57	-0.71 – -0.42
$100K to $150K	-1.11	-1.75 – -0.47	-2.62	-3.28 – -1.96	-3.45	-4.44 – -2.45	-0.69	-0.83 – -0.55
>$150K	-1.51	-2.14 – -0.87	-3.33	-3.98 – -2.67	-4.36	-5.34 – -3.37	-0.79	-0.93 – -0.65
Gender
Women	Reference ***		Reference ***		Reference ***		Reference ***
Men	-1.42	-1.73 – -1.12	-1.22	-1.54 – -0.91	-2.27	-2.75 – -1.80	-0.24	-0.31 – -0.17
Gender Diverse	1.6	0.67 – 2.52	2.08	1.14 – 3.02	3.17	1.72 – 4.62	0.34	0.12 – 0.57
Increased cannabis use	1.2	0.71 – 1.70	1.29	0.78 – 1.80	1.69	0.88 – 2.50	0.06	-0.06 – 0.18
Phase x Increased cannabis use
Pre-COVID	Reference ***		Reference ***		Reference ***		Reference *
Phase 1	1.4	1.00 – 1.81	1.47	1.05 – 1.88	1.64	0.89 – 2.39	0.15	0.04 – 0.26
Phase 2/3	1.43	0.97 – 1.89	1.53	1.07 – 2.00	1.85	1.01 – 2.69	0.14	0.02 – 0.26
Phase 2/3_2	1.38	0.68 – 2.08	1.34	0.62 – 2.06	2.48	1.19 – 3.77	0.16	-0.03 – 0.34
Phase 4	1.8	1.11 – 2.50	2	1.28 – 2.71	2.62	1.34 – 3.91	0.18	-0.01 – 0.36
Phase 5	1.73	1.03 – 2.43	1.94	1.23 – 2.66	2.17	0.88 – 3.46	0.22	0.03 – 0.40
Gender x Increased cannabis use
Women							Reference**
Man							0.45	0.16 – 0.73
Gender Diverse							-0.14	-0.70 – 0.42
Marginal R2 /Conditional R2	0.183 / 0.718		0.175 / 0.722		0.262 / 0.675		0.151 / 0.623

CI refers to confidence intervals for the adjusted estimates. Pre-COVID: Prior to mid-March 2020; Phase 1: Mid-March 2020 to mid-May 2020; Phase: 2/3: mid-May 2020 to November 2020; Phase 2/3_2: mid-May 2020 to August 2020; Phase 4: September 2020 to October 2020; Phase 5: November 2020 to March 1, 2021. GAD-7: Generalized Anxiety Disorder measure; PHQ-9: Patient Health Questionnaire; CRISIS: CoRonavIruS Health Impact Survey; Loneliness was measured using a single item.

*p = .04

**p = .008

***p < .0001.

Discussion

This large Canadian study recruited 6,076 women, men, and gender diverse people across the province of British Columbia. Our main findings indicated that age, sex, gender, ethnicity, Indigenous status, sexual orientation, and phase of the pandemic have distinct effects on psychosocial outcomes. Across outcomes, women had more symptoms of depression, anxiety, loneliness, and stress than men, regardless of their age or ethnicity, while the gender diverse group (n = 72) had even more symptoms than women. An analysis by sex revealed the same findings as for gender, except that the gender diverse group was now absorbed into one of the two binary sex categories and obscuring their findings.

Our results highlight the greater negative outcomes on all psychosocial variables in gender diverse individuals, which would have been obscured in an analysis by sex alone and adds to the literature highlighting the value in analyzing data by gender. It is important to underscore that being a woman was a significant factor that determined higher anxiety, depression, stress, and loneliness—a finding mirrored in the literature across all continents [38,39]. The novelty of this study, however, is that this effect of being a woman was not impacted by participants’ age, ethnicity, or other sociodemographic variables. In other words, having a woman gender was sufficient to place individuals at higher risk for depression, anxiety, stress, and loneliness over the pandemic. Given that women and gender diverse individuals are more likely to be diagnosed with mood disorders or score lower on mood surveys outside of a pandemic [16,40,41], it is not surprising that these populations are experiencing mental health inequities during COVID-19. However, our results should be interpreted with some caution as our gender-diverse cohort accounted for only 1% of the sample. Nonetheless, our results are striking and consistent with many other studies focused on gender using larger cohorts [42,43].

Our study also benefited from examining the effects of other sociodemographic variables, such as age, to determine how they might play a role in the effect of sex and gender on mental health. Across all the psychosocial measures, younger participants were more likely to have anxiety, depression, pandemic stress, and loneliness, irrespective of their gender. These findings are consistent with others in smaller cohort studies that indicated younger ages were associated with more psychosocial symptoms [44]. There may be several reasons for these findings such as restricted social engagements, barriers to employment, and living conditions. Lockdowns across the globe have resulted in restricted social gatherings, closing of restaurants, bars and clubs, as well as recreational sporting activities (gyms, sports clubs, exercise classes, yoga and dance). In addition, younger adults are more likely to either live on their own, or with unrelated roommates and have greater perceived lack of social support. Indeed, findings from a larger cohort in China found that greater loneliness was associated not only with younger age (16–29) but also in unmarried individuals [36]. Physical activity is another important factor as a large survey across fourteen countries found that decreased physical activity during restrictions and lockdowns, as well as high physical activity pre-pandemic, were associated with poorer mental health scores [45]. Other studies have also noted that suicide and suicidal ideation have increased during the pandemic in younger adults [46], related partially to job losses. Taken together, the underlying reasons for this significant effect of age are of great importance and require further study. At a minimum, these findings suggest that mental health resources tailored to younger individuals are required in any pandemic relief measures taken by government. It might not be sufficient to increase all mental health supports, but rather have tailored ones to young adults that are cost effective and accessible.

In addition to age, ethnicity was associated with psychosocial outcomes with Chinese/Taiwanese participants reporting significantly lower scores (i.e., fewer psychosocial symptoms) on anxiety, depression, pandemic stress, and loneliness. These data are consistent with findings from other studies, such as a survey of more than 46,000 Canadians which found that Chinese individuals were less likely to report symptoms consistent with moderate to severe generalized anxiety disorder than other visible minority groups during the COVID-19 pandemic [47]. It is possible that the lower rates of mental disorders seen in Asian or Chinese immigrants [48] may be due to cultural stigma associated with mental illness leading to lower rates of disclosure of psychological symptoms [48,49]. It is also possible that the lower rates of psychological symptoms may be due to differences in the validity of these measures cross culturally [34,50], leaving open the possibility of a measurement bias [51], although it was concluded to be a reliable measure across some cultural groups [51,52]. In sum, our findings suggest that our Chinese/Taiwanese sample experienced fewer psychosocial symptoms throughout the pandemic relative to other groups, and of note, ethnicity did not interact with gender or income to impact these outcomes. Similarly, gender did not interact with income to impact these outcomes.

We found that those who self-identified as Indigenous had significantly more psychosocial symptoms than non-Indigenous participants across all four scales for all phases of the pandemic in BC. Importantly, there was no difference in psychosocial outcomes between Indigenous and non-indigenous groups pre-COVID, which underscores the disproportionate impact of the pandemic on this community. While investigations on the mental health impacts on Indigenous peoples during the COVID-19 pandemic have been limited, our results are consistent with the available data. For example, other data from Australia (Aboriginal or Torres Strait Islander) [53] as well as Canada [54] showed more psychosocial symptoms among Indigenous respondents during the COVID-19 pandemic. The lack of interaction between Indigenous status and gender suggests that the higher psychosocial symptoms occur regardless of an Indigenous persons’ gender, standing in contrast to another study finding that Indigenous women were particularly impacted by mental health issues (severe generalized anxiety, worse mental health, and stress) during COVID-19 [54]. Future studies should explore the extent to which variables such as rurality (which can contribute to barriers accessing care) and income may account for these higher rates of psychological symptoms among Indigenous communities [44]. In the meantime, these findings point to the need for culturally-safe mental health resources being made available to Indigenous communities in any COVID relief efforts.

Findings on the relationship between anxiety, depression, pandemic stress, and loneliness, with increased alcohol and cannabis use, align with previous studies [12]. Given the poorer self-reported mental health among younger populations, it was not surprising to observe an increase in alcohol and cannabis use among this group, which suggests that alcohol use may be a form of coping for younger persons. We cannot attribute directionality to this association, nor eliminate the possibility that increased alcohol and cannabis use may be contributing to the increased psychosocial symptoms observed among younger populations during the pandemic. The lack of a gender difference in increased alcohol use is in contrast with a previous American study [55] which found that females had increased their alcohol use compared to males. It may be that differences in the samples accounts for these contrasting findings. It is also possible that the increase seen in men in our sample was higher than in previous studies, thus rendering the gender difference void. In contrast, we saw a gender effect on increased cannabis use, which was expressed by women, but not by men or gender diverse persons. In recent surveys, 28% of British Columbians had engaged in cannabis use in the past twelve months, compared to the Canadian average of 11%, suggesting that British Columbians are more likely to engage in cannabis use, and therefore may be more likely to use cannabis as a form of coping [56,57]. Although cannabis use has been associated with male typicality and may go against gender norms typical to women [58], it may be that the social isolation disrupted these social norms and facilitated women’s more active engagement in additional cannabis use, relative to pre-pandemic levels.

Our findings align with the global trend of increased substance use, as recent studies have demonstrated that alcohol, cannabis, and opiate use changed during and post-lockdown [59]. Alcohol use has remained elevated relative to pre-pandemic levels, and though opiate use seemed to have dropped during lockdowns, a return to regular dosage post-lockdown has helped to drive overdoses, due to diminished tolerance [59]. It is possible that deteriorated mental health could be attributed to overuse of certain substances, though studies with multiple follow-up points are needed to determine a causal pathway for the increase in psychosocial symptoms demonstrated in our study. Future studies should aim to elucidate potential mechanisms by which substance use can influence mental health in the context of a pandemic and lockdowns to mitigate the consequences of public health interventions on well-being.

As predicted, psychosocial symptoms worsened over the course of the pandemic, with some of the highest symptoms observed early on, aligning with previous studies that found a higher prevalence of mental health disorders during the initial COVID-19 lockdown in March 2020 [60,61]. Phases 2 and 3 of COVID restrictions in BC were characterized by an easing of restrictions, permitting outdoor gatherings and small social events, and the summer season. This loosening of public health measures was associated with a slight improvement in mental health, more than likely due to an increase in perceived social support and optimism regarding the state of the pandemic. Mental health outcomes then worsened in Phases 4 and 5 as BC entered wave 2 of the pandemic and public health orders tightened once again. It is important to note the average PHQ-9 and GAD-7 scores did not meet the criteria for clinical depression or anxiety, but that these levels increased relative to pre-pandemic levels as well as over time.

Strengths and limitations

Our study benefitted from a large, population-based sample size, and, despite known mental health disparities by gender, as far as we are aware, was one of the few that sought to explore findings from a gender lens by including gender-diverse groups as well, given known mental health disparities by gender [17]. That said, our sample size for gender diverse individuals was still limited [17]. Future studies should further investigate mental health in the gender diverse community during the COVID-19 pandemic with a focus on people of all ages, in contrast to previous studies [22]. Another limitation of the present study was the retrospective, cross-sectional nature of the survey, where participants completed the survey at only one time point, and were asked to retrospectively recall their mood and anxiety during different time points. This may have increased the likelihood of recall bias and reducing our capacity to examine causality and directionality of poor mental health outcomes. Finally, this study was confined to the general population of BC, and only individuals who had access to email and internet, and therefore results may only be generalizable to the Canadian population, and populations with similar demographics to the present study, and to individuals who have access to email and internet.

Implications

Our study has important implications for public health policy. These findings illustrate that government policies and interventions for future pandemics should place on emphasis on young adults, low-income populations, women, Indigenous, and gender diverse communities. Additionally, our study was one of the first to measure mental health outcomes across different phases of the pandemic, directly examining the effect of increased public health measures on mental health. At the time of writing, the vaccine rollout is well underway in BC with experts predicting an end to the pandemic in the months ahead, however, it is unclear whether mental health will return to pre-pandemic levels, or when life will return to “normal.” Moving forward, policy makers and leaders need to consider our findings when planning future public health measures. In future pandemics, the mental health of marginalized populations needs to be considered proactively. As vaccination efforts continue and case counts fall, it will also be critical to monitor the health status of these populations to ensure that they are not left behind. Additionally, for future pandemics and outbreaks, mobilizing resources to these communities early on can aid in mitigating these inequities from the beginning, rather than as an afterthought.

Public health measures during different phases of the COVID-19 pandemic in British Columbia.

Measures listed are not exhaustive.

(DOCX)

Click here for additional data file.

21 Sep 2021

PONE-D-21-18028The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemicPLOS ONE

Dear Dr. Brotto,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Two reviewers provided feedback on this interesting paper.  Reviewer 1 identified numerous methodological issues and recommended further focus and less breadth in the discussion.  In contrast Reviewer 2 indicated that the scope could be widened.  In appreciation of both reviews, I gave this a careful read.  I agree with Reviewer 1 that the paper should be revised substantially in all areas identified by that reviewer.  Please ensure that your decision is justified on PLOS ONE’s publication criteria and not, for example, on novelty or perceived impact.

Please submit your revised manuscript by November 19, 2021. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Kimberly Page, PhD, MPH

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

2. Peer review at PLOS ONE is not double-blinded (https://journals.plos.org/plosone/s/editorial-and-peer-review-process). For this reason, authors should include in the revised manuscript all the information removed for blind review.

3. Please include your full ethics statement in the ‘Methods’ section of your manuscript file. In your statement, please include the full name of the IRB or ethics committee who approved or waived your study, as well as whether or not you obtained informed written or verbal consent. If consent was waived for your study, please include this information in your statement as well.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: “The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic” applies gender and intersection lenses to an investigation of mental health and substance use. It includes over 6,000 people living in Canada, the majority of whom are white (81%), heterosexual (85%), women (87%), living in an urban area (94%), who earn over $80,000 per year (57%) and have access to email as well as the internet (100% by virtue of recruitment). Authors report that outcomes are significantly associated with pandemic phase, gender, and age.

This is an important topic and little research has been conducted in regard to the COVID-19 pandemic. The authors report interesting findings, which could help inform responses to future pandemics. Three points dampen my enthusiasm for the paper. First, sampling methods appear to be haphazard, making it difficult to understand the results and to whom they can be generalized. Second, the large number of outcomes renders the discussion as rather surface and lacking an anchor or main “thread.” Third, the novel findings discovered here aren’t highlighted in a way that makes them stand out; how they advance the field, and how they should be used need to be further developed.

A few suggestions:

-The title and abstract refer to a gender lens and the introduction refers to an intersectional lens. The introduction then goes on to motivate a gender-based study, but brings back intersectionality in paragraph 5. It would be helpful for the authors to commit and focus the paper on either gender or intersectionality (which includes gender, but is a different topic and needs to be set up differently). If intersectionality is be the focus of the paper, it should be defined and the paper will need some editing. Intersectionality is generally considered the interaction between aspects of social status (gender, race, socioeconomic status), which requires a different introduction (different motivation). For example, ethnicity is a major player in intersectionality, but the authors include only one sentence on minority stress in the introduction (and as an “other social determinant of health), with nothing about how race, sex and socioeconomic status have been shown in prior research to influence outcomes studied here. Also, if the paper is focused on intersectionality, interaction effects should be the main effects of interest.

-The authors introduce a lot of topics in the introduction which makes the the point of the paper unclear. It would be helpful for the authors to streamline the text a bit. Also, while a variety of topics are raised, several key topics are left out. For example, each outcome is a huge topic in itself and to understand how this study advances science, the reader needs to understand what is already known about the outcomes (e.g., why are the questions asked here important for each outcome?) Understanding what results reported here mean will require at least some background on each outcome (or background on considering them collectively). This may be a lot to accomplish for so many outcome variables.

-“Phases” of the pandemic are mentioned, but what they are and why they matter is not specified. Are the dates the authors assigned to phases arbitrary? Exact what does each phase represent? The authors hypothesize that “during phases of increased social restriction psychosocial symptoms would increase,” suggesting that social restriction somehow contributes to the definition of each phase, but exactly how that happens is unclear.

-More detail on inclusion criteria, sampling methods and recruitment venues of the original studies is needed.

-The statistical analysis section states that longitudinal analyses were conducted, but a limitation of the study cited by authors is “cross-sectional nature of the survey, where participants completed the survey at only one time point.” Please clarify.

-There are multiple types of longitudinal data analysis so please be more specific about analytic methods.

-The abstract lists six outcomes, but the last paragraph of the introduction seems to suggest only four mental health outcomes, with analyses that consider how these outcomes are associated with the substance use variables. This may just be a matter of a few edits to the final sentences of the introduction, but please be specific and consistent throughout the manuscript to improve clarity.

-The last paragraph of the introduction also refers to the four “psychosocial domains.” Please explain how the four mental health variables map onto psychosocial domains. Referring to a score intended to identify a mental disorder as a domain is confusing, as is some of the text around it (e.g., “We predicted that these four psychosocial domains would be more symptomatic in women and gender diverse individuals”).

-Given that potential participants were invited to enroll in the current study via email and to take an online survey, people without email or access to the internet were left out. This overlooked low-income individuals, which is a limitation of the study and should be cited as such.

-Authors state “To improve the representativeness of the study sample, Index Participants were asked to pass the invitation on to one household member who identified as a different gender as the respondent.” While this would have increased the sample size, and possibly increase gender variability, it would not necessarily increase representativeness of the population. To the contrary, sampling from the same household results in another person who is very similar to the first in terms of race and socioeconomic factors. In the results section, it becomes clear that this population is in fact composed of many more women (87%) than men. Why the population is so skewed toward women deserves explanation.

-The authors do not appear to have used a measure of loneliness that was previously tested for reliability and validity. It would be helpful for the authors to provide assessment of reliability and validity for the measure they created here.

-A major problem with this study is understanding exactly who it represents. It starts out including people who have participated in prior research (but recruitment criteria are not provided). It then goes on to state that, after two months of data collection, recruitment was expanded. Why expansion occurred is unclear and how the additional sampling venues were chosen is similarly unclear. Public recruitment through social media, and “engagement of community groups and other stakeholders” (without explanation as to what is meant by stakeholder in this specific scenario) make it very difficult to understand who this study applies to and what it means. The number recruited using each sampling mode is not stated in the results, and there is also no attempt to account for recruitment mode in analyses.

-On average, the sample seems to be college-educated, straight, white women who live in an urban environment, have access to the internet, and earn over $80,000 a year. This isn’t entirely clear and it would be helpful to distill information like this in a summarized fashion to give the reader an idea of who is included.

-Regarding lines 152 and 153: how was the target sample size of 750 determined?

-Discussing mental health scores is confusing (e.g., lower usually indicates worse health, but in this case, it refers to fewer symptoms and better health). Consider referring to the number of symptoms throughout instead of scores.

-The text refers to symptom scores (suggesting a continuous variable), but the table simply states each mental health condition (suggesting a dichotomous variable). Given that the methods section describes both formats, this is confusing. Consider clarifying the format of the table variables on the table.

-The text refers to sex and gender separately and I’m assuming that the headers in Table 1 also refer to gender, but it would be helpful to specify that “women” and “men” refers to gender and not sex. Given the focus on gender and sex in this paper, it would also be helpful to add a row showing how sex breaks down by gender.

-The methods section states, “We included pairwise interactions to assess non additive effects between pandemic phase, sex, gender, ethnicity, sexual orientation, income, and Indigenous status.” However, the results section leads with a series of interactions based on age. There is a disconnect here. Also, intersectionality is often thought of as being based in race and gender, so the results section was surprising due to its inconsistency with both earlier methods text (which did not emphasize age) and the field of study (which often focuses on gender and race). Age is certainly an important factor; so authors might want to consider (1) developing the introduction in a way that talks about age as a component of intersectionality, and (2) being consistent between methods and results in terms of factors discussed.

-Unless I missed it, there were no significant interactions between race and sex or race and income or sex and income with regard to the mental health and substance use outcomes. It may be helpful to clearly state this in the discussion.

-The discussion is rather surface and fails to highlight what is new here and how, in combination with prior research, results could be used to improve policy or practice (i.e., beyond the fact that it should be used, exactly how should it be used?) For example

●The discussion states that women were more likely to have worse mental health scores and this is consistent with prior studies. How was this study different, novel or unique; how do the novel findings contribute something new here; and how should policy or practice be altered based on this study?

●There is a paragraph devoted to why age may have been associated with study outcomes, but I wonder if the authors have any suggestions for changes to pandemic-related policy or practice based on age? Or how future research could use this finding?

● The discussion states that this study, which found no gender effect on alcohol use, is in contrast to a prior study. Why might that be? Geographic differences? Differences in the population? Differences in terms of the time point (relative to the beginning of the pandemic) studied?

● The authors state, “These findings illustrate that government policies and interventions for future pandemics should place on emphasis on young adults, low-income populations, women, Indigenous, and gender diverse communities.” I wonder what this means and how it would be done. It would be extremely helpful for the authors to cite one or two pandemic policies and suggest exactly how they could have placed an emphasis on the factors identified here.

● Similar to the prior bullet point, the final conclusion of the abstract is, “Our findings highlight the need for policy makers and leaders to proactively consider gender when tailoring public health measures for future pandemics,” which is rather generic. It could have been written (based on prior COVID research) before this study was conducted. What do these results uniquely suggest that policy makers should do specifically?

-The paragraph regarding stigma against Chinese culture and validity of study measures in Chines Canadians is long and seems to be a tangent. Consider a more succinct paragraph on how these results harmonize with prior studies and how findings could be put to practice.

-Authors state that there was a trend for increasing odds of increased alcohol use as household income increased. However, given that none of the income categories were significantly different than the reference, this seems unlikely and the statement needs justification (e.g., was the trend statistically significant?)

-The abstract states that participants were asked questions “across five phases of the pandemic as well as retrospectively before the pandemic,” which seems to suggest multiple measures. The text should be edited to clearly indicate that this is a cross-sectional study.

-Consider “fewer symptoms” rather than “less symptoms.”

-Consider additional editing (e.g., using “have” multiple times in the same sentence; also, regarding Lines 84 and 85, rather than stating that age may interact with sex and psychosocial outcomes [suggesting a 3-way interaction], I think the authors mean that age may interact with sex to influence psychosocial outcomes).

Reviewer #2: Dear Authors,

I have perused with great interest your praiseworthy contribution titled The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic, which aims to elaborate on one of the most consequential aspects of the ongoing pandemic.

The article is well-written and competently assembled. The discussion is broad-ranging and thorough, its conclusions well supported by solid methodology.

Analyzing the influence of sex, gender, age and ethnicity confers an element of novelty to the report which makes it a valuable piece of research of considerable interest for the public at large. Discrepancies in how deeply the pandemic has impacted different segments of the population are well expounded upon, and that itself has value in terms of contributing to more effective policy-making approaches and strategies.

An element of weakness attributable to studies such as this has to do with its being cross-sectional, which makes it all but impossible to establish the actual correlations and causality directions taking into account all the study's variables. In that regard, future longitudinal studies will go a long way towards establishing actual causality. All data were self-reported, hence liable to be affected by well-established method biases, as you hinted.

Since the title mentions "substance use", I believe it would be advisable for the authors to dig a buit deeper than alcohol and cannabis, although their research has only accounted for those.

I would recommend making a brief subchapter or mention in the conclusions as to how substance use worldwide has been profoundly reshaped as a result of the pandemic, which might dovetail with some of the conclusions you arrived at based on your data, both from a substance use prspective and a psychosocial one. That would go a long way toward broadening the scope of your article, allowing for more in-depth comparisons of your own findings versus other sources.

You may want to draw upon the following sources:

Zaami S, Marinelli E, Varì MR. New Trends of Substance Abuse During COVID-19 Pandemic: An International Perspective. Front Psychiatry. 2020 Jul 16;11:700. doi: 10.3389/fpsyt.2020.00700.

Ornell F, Moura HF, Scherer JN, Pechansky F, Kessler FHP, von Diemen L. The COVID-19 pandemic and its impact on substance use: Implications for prevention and treatment. Psychiatry Res. 2020 Jul;289:113096. doi: 10.1016/j.psychres.2020.113096.

Stack E, Leichtling G, Larsen JE, Gray M, Pope J, Leahy JM, Gelberg L, Seaman A, Korthuis PT. The Impacts of COVID-19 on Mental Health, Substance Use, and Overdose Concerns of People Who Use Drugs in Rural Communities. J Addict Med. 2020 Nov 3:10.1097/ADM.0000000000000770. doi: 10.1097/ADM.0000000000000770.

Zaami S. New psychoactive substances: concerted efforts and common legislative answers for stemming a growing health hazard. Eur Rev Med Pharmacol Sci. 2019 Nov;23(22):9681-9690.

Lastly, since your manuscript qualifies as a research article, it would be best to have a structured abstract withObjectives, Methods and study design, Results and Conclusions.

All in all, the manuscript will make for a fine and valuable conytribution of great interest to a large audience.

Congratulations on your commendable effort.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

9 Oct 2021

9 October 2021

Editorial Team

PLOS One

Dear members of the editorial team,

Thank you very much for your email dated September 21, 2021 that our paper, “The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic” was provisionally accepted pending a revised manuscript was reviewed and deemed adequate.

We have carefully considered each of the two reviewers comments and addressed each out, as outlined point by point in the attached response letter. We have also included both a clean copy and a track changed copy.

Thank you very much for your ongoing consideration of our manuscript.

Respectfully submitted,

Sincerely,

Lori A Brotto PHD, R PSYCH

Executive Director, Women’s Health Research Institute

Professor | Department of Obstetrics & Gynaecology, University of British Columbia

Canada Research Chair | Women’s Sexual Health

Allied Staff Member | Vancouver Acute Health Service

Submitted filename: PONE response to reviewers_Oct 9.docx

Click here for additional data file.

25 Oct 2021

The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic

PONE-D-21-18028R1

Dear Dr. Brotto,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Kimberly Page, PhD, MPH

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

I commend the authors on their responses to reviewers critiques and the very well revised manuscript. I do think this articles contributes important data and knowledge regarding gender differences and response to the COVID-19 pandemic.

Reviewers' comments:

11 Nov 2021

PONE-D-21-18028R1

The influence of sex, gender, age, and ethnicity on psychosocial factors and substance use throughout phases of the COVID-19 pandemic

Dear Dr. Brotto:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Kimberly Page

Academic Editor

PLOS ONE