Doomscrolling: Prospective associations between daily COVID news exposure, internalizing symptoms, and substance use among sexual and gender minority individuals assigned female at birth

Abstract

Cross-sectional studies have demonstrated associations between COVID news exposure, anxiety, and depression. However, longitudinal research examining the directionality of these associations is extremely limited. Further, most studies have focused on the general population and neglected sexual and gender minority individuals (SGM), a population disproportionately impacted by the pandemic. To address these limitations, the current study utilized data from a 30-day diary study of SGM assigned female at birth (N = 429). We examined concurrent and prospective associations between COVID news exposure, depressed and anxious affect, COVID distress, substance use, and motives for use. Further, we examined associations between cumulative COVID news exposure across the diary period and prospective changes in anxiety, depression, and COVID distress. When participants were exposed to more COVID news, they experienced more depressed and anxious affect, more COVID distress, less positive affect, and were more likely to use alcohol and cannabis to cope. Further, when participants were exposed to more COVID news, they experienced subsequent increases in depressed affect, decreases in positive affect, and increases in the likelihood of using cannabis to cope. Findings also provided evidence of bi-directional prospective associations between COVID news exposure and COVID distress and of a cumulative impact of COVID news exposure on anxiety, depression, and COVID distress. Findings suggest that individuals should balance the need to remain informed about the pandemic and their own mental health when considering how much COVID news to consume.

Introduction

Internalizing symptoms (i.e., anxiety and depression) have increased in the context of the COVID-19 pandemic (Czeisler et al., 2020) and individuals who used substances prior to the pandemic have experienced increases in substance use and related physical and psychosocial consequences, such as detrimental effects of substance use on social relationships or work/school productivity (Dumas et al., 2020; Dyar et al., 2021; Gritsenko et al., 2020; Janulis et al., 2021). A number of factors have been proposed to explain these elevations in the general population, including social isolation and concerns about impacts of the pandemic (Banerjee & Rai, 2020; Burtscher et al., 2020). Overexposure to news about COVID has also been proposed to contribute to increases in internalizing symptoms (Garfin et al., 2020), and cross-sectional research has linked high COVID news exposure to internalizing symptoms (e.g., Huckins et al., 2020; Olagoke et al., 2020; Riehm et al., 2020; Stainback et al., 2020) and, to a lesser extent, substance use (Sokolovsky et al., 2021; Taylor et al., 2021). However, the few longitudinal studies to examine the association between COVID news exposure and mental health have almost exclusively examined the association between COVID news exposure and COVID anxiety (Chu et al., 2021; Schmidt et al., 2021), neglecting outcomes like affect and substance use (for an exception see Shaw et al., 2021).

Further, existing research has focused on the general population (e.g., Huckins et al., 2020; Olagoke et al., 2020; Riehm et al., 2020; Schmidt et al., 2021; Shaw et al., 2021; Stainback et al., 2020), rather than examining COVID news exposure among populations that have been disproportionately impacted by the pandemic. Sexual and gender minority individuals (SGM) were at heightened risk for SU disorders (Kerridge et al., 2017) and internalizing symptoms pre-pandemic. Further, SGMs’ mental health and substance use have been more negatively impacted by the pandemic compared to cisgender, heterosexual individuals (Fish, Salerno, et al., 2021; Slemon et al., 2022) and they may be disproportionately burdened by psychosocial impacts of the pandemic, such as COVID anxiety (Kamal et al., 2021; Ruprecht et al., 2021).

Among sexual minority individuals, disparities in SU are particularly pronounced for those assigned female at birth, with sexual minority individuals assigned female at birth being 2.3 times and sexual minority individuals assigned male at birth 1.5 times more likely to have a substance use disorder compared to same-sex heterosexuals (Kerridge et al., 2017; Krueger et al., 2020). Sexual minority individuals assigned female at birth also experience high rates of mood disorders, with nearly 1 in 3 sexual minority individuals assigned female at birth having experienced a mood disorder in the past year compared to approximately 1 in 5 sexual minority individuals assigned male at birth, 1 in 10 heterosexual individuals assigned male at birth, and 1 in 6 heterosexual individuals assigned female at birth (Kerridge et al., 2017). Further, recent research indicates that the pandemic may have a more detrimental impact on the mental health and SU of sexual minority individuals assigned female at birth compared to those assigned male at birth, with sexual minority individuals assigned female at birth experiencing larger increases in depression and alcohol use during the pandemic compared to those assigned male at birth (Chang et al., 2021; Salerno et al., 2021). The current study aims to address limitations in the existing literature by examining concurrent and prospective associations between COVID news exposure, internalizing symptoms, and substance use utilizing daily diary data from a sample of sexual and gender minority individuals assigned female at birth (SGM-AFAB).

COVID News Exposure and Mental Health

Research examining exposure to news of traumatic events prior to the COVID-19 pandemic has demonstrated associations between news exposure and symptoms of post-traumatic stress, acute stress, and depression (Pfefferbaum et al., 2014). For example, Silver and colleagues (2013) demonstrated prospective effects of exposure to news coverage of the September 11^th terrorist attacks on post-traumatic stress symptoms 2–3 years later. Similar associations have been demonstrated for other events, including the Boston Marathon bombing (Garfin et al., 2015; Holman et al., 2014) and Pulse nightclub shooting (Thompson et al., 2019).

The effect of exposure to COVID news on mental health has been a focus of research since early in the pandemic. Researchers have cited the high volume of news about COVID and the ambiguity arising from conflicting information provided by different news outlets, individuals, and organizations as being likely to contribute to a particularly negative impact of overexposure to COVID news on mental health (Garfin et al., 2020; Stainback et al., 2020). A number of cross-sectional studies have examined the association between COVID news exposure and internalizing symptoms in the general population in samples from the United States (Olagoke et al., 2020; Riehm et al., 2020; Stainback et al., 2020), other western countries (Bendau et al., 2021), the Middle East (Shabahang et al., 2020), and Asia (Gao et al., 2020; Liu & Liu, 2020; Liu & Tong, 2020; Mongkhon et al., 2021). Studies demonstrated that following COVID news more closely, spending more time viewing COVID news, or consulting a wider range of media sources were cross-sectionally associated with more psychological distress and internalizing symptoms (Bendau et al., 2021; Gao et al., 2020; Hong et al., 2021; Olagoke et al., 2020; Riehm et al., 2020; Stainback et al., 2020). These findings clearly demonstrate that individuals who consumed more COVID news experienced more internalizing symptoms. However, the directionality and temporality of this association remains unclear given the small amount of longitudinal research on the topic.

Daily diary and other intensive longitudinal methods are well-suited to determining the directionality and temporality of associations. Such methods allow for examination of how the association between COVID news exposure and internalizing symptoms unfold on a day-to-day basis and can help to determine which theorized directionality is supported by data. Does exposure to COVID news drive increases in internalizing symptoms, does anxiety drive some people to seek out more COVID news in an attempt to reduce anxiety-causing ambiguity about the pandemic, or is there evidence in support of both hypotheses? Longitudinal methods also allow us to determine whether the association between COVID news exposure and internalizing symptoms is a between-person effect, with people who tend to be more anxious also tending to seek out more COVID news, or a within-person effect, with increases in COVID news exposure predicting increases in internalizing symptoms. We are aware of only two longitudinal studies to examine the association between COVID news exposure and mental health. Israeli ex-prisoners of war were found to experience larger increases in news consumption during the COVID-19 pandemic compared to a control group (Solomon et al., 2021). These increases in news exposure were associated with concurrent increases in PTSD symptomology. In another study conducted early in the pandemic (May/June 2020), researchers found that higher news exposure (a blend of general and COVID-specific news) prospectively predicted increases in negative affect, but not positive affect (Shaw et al., 2021). Three daily diary studies have specifically examined associations between COVID news exposure and COVID anxiety, with one finding a concurrent association (Chu et al., 2021) and two demonstrating a bi-directional association (Schmidt et al., 2021; Shaw et al., 2021). Given that only one study has examined prospective associations between COVID news exposure and negative affect, further research is needed that examines these prospective associations to understand how they unfold on a day-to-day basis.

COVID News Exposure and Substance Use

While the association between COVID news exposure and internalizing symptoms has received a good deal of attention, few studies have examined the association between COVID news exposure and substance use. Individuals who used substances prior to the pandemic have experienced increases in substance use and related consequences during the pandemic (Dumas et al., 2020; Gritsenko et al., 2020; Janulis et al., 2021). These increases have been attributed to using substances to cope with negative affect arising from COVID stressors (Lechner et al., 2020; McPhee et al., 2020). Given that overexposure to COVID news has been conceptualized as a COVID stressor, it is expected that overexposure to COVID news may also help to explain increases in substance use and coping motives for use. One of the few studies to examine this association found that higher COVID news exposure was cross-sectionally associated with a reduced likelihood of tobacco cessation among US college students in May and June 2020. Taylor et al. (2021) also found that compulsively checking for COVID news was associated with retrospectively reported increases in alcohol and drug use since before the pandemic and more severe current alcohol and drug use problems. In a daily diary study, Shaw et al. (2021) found that COVID news exposure did not prospectively predict substance use. However, this study utilized a sample of the general population and did not focus on people who use substances regularly. This current study aims to build on this research by longitudinally examining the association between COVID news exposure and quantity of substance use, consequences of use, and coping motives for use in daily diary data from individuals who use alcohol and/or cannabis regularly, as this group has experienced documented increases in substance use and consequences during the pandemic (Dumas et al., 2020; Gritsenko et al., 2020; Janulis et al., 2021).

Traumatic News Exposure among SGM and POC

Some populations, especially sexual and gender minority individuals (SGM) and people of color (POC), may be disproportionately impacted by COVID news exposure. Garfin et al. (2015) found evidence of an accumulation of the effects of exposure to traumatic events in the news on mental health, with individuals exposed to news coverage of multiple traumatic events experiencing worse mental health. They concluded that individuals who are exposed to news of multiple traumas that are relevant to their community may be sensitized to the impact of exposure to subsequent news of traumatic events. Consistent with their conclusion, Solomon et al. (2021) found that Israeli ex-prisoners of war experienced larger increases in news exposure during the pandemic than matched controls, and these increases helped to explain larger increases in PTSD symptomology experienced by Israeli ex-prisoners of war during the same period.

The past two decades have been peppered by news stories of discriminatory, stigma-based politics and legislation that target SGM and POC (e.g., anti-marriage equality and anti-transgender legislation; racist political rhetoric) as well as news of violence against SGM and POC in the US and abroad (e.g., hate crimes, Pulse nightclub shooting, police violence against POC). Studies have demonstrated the detrimental impact of these kinds of events on SGM and POC. Frost and Fingerhut (2016) found that daily exposure to campaign messages against marriage equality in the 2012 general election was associated with increases in negative affect among same-sex couples. Similarly, exposure to race-related traumatic events in online media has been cross-sectionally linked to post-traumatic stress disorder and depressive symptoms among Black and Latinx adolescents (Tynes et al., 2019), and exposure to negative transgender-related media has been cross-sectionally linked to symptoms of anxiety and depression among transgender adults (Hughto et al., 2021). One recent study also indicated that sexual minority women perceived COVID to be more threatening to themselves, their families, and their communities, were more likely to know someone who had COVID, and were more likely to perceive their workplace as increasing their risk for exposure to COVID compared to heterosexual women (Potter et al., 2021). These findings suggest that exposure to COVID news may be particularly distressing for sexual minority women. Given SGM and POC’s prior experiences with exposure to stressful messages through the news and other media and the disproportionate effects of the COVID pandemic on these populations (Chang et al., 2021; Goldman et al., 2021; Millett et al., 2020; Salerno et al., 2021), it is necessary to examine the association between COVID news exposure, internalizing symptoms, and substance use among these populations, as they may be disproportionately impacted by further exposure to media focused on traumatic events. In the current study, we examined these associations among a racially diverse sample of SGM individuals assigned female at birth to reduce this gap in the literature.

Current Study

The goal of the current study was to further our understanding of the associations between COVID news exposure, internalizing symptoms, and substance use by using daily diary data from a sample of racially diverse SGM-AFAB. We examined concurrent and prospective within-person associations between COVID news exposure, affect (depressed, anxious, and positive affect), COVID distress, and substance use (alcohol/cannabis use; coping motivation for substance use). We made the following hypotheses about these associations:

• 1We hypothesized that at the within-person level, SGM-AFAB would report more depressed affect, anxious affect, and COVID distress as well as less positive affect when they were exposed to more COVID news than usual.
• 2Further, we expected that SGM-AFAB would be more likely to report alcohol and cannabis use, use more alcohol and cannabis, and be more likely to report using alcohol and cannabis to cope on days when they were exposed to more COVID news than usual.
• 3We also expected that these effects would be present prospectively, with COVID news exposure on one day predicting subsequent increases in anxious and depressed affect, COVID distress, and substance use and decreases in positive affect from that day to the next.

Given that anxiety, depression, and COVID distress may also lead to increased COVID news consumption, we also tested prospective models in which anxious and depressed affect and COVID distress during one assessment predicted subsequent increases in COVID news exposure.

• 4We hypothesized that we would find evidence of bi-directional associations, with anxiety, depression and COVID distress prospectively predicting COVID news exposure, and vice versa.

Finally, we examined potential effects of cumulative COVID news exposure on changes in mental health (anxiety and depression), substance use (alcohol and cannabis use consequences), and COVID distress over a one-month period.

• 5We expected that individuals who were exposed to more COVID news over the course of the diary study would experience increases in anxiety, depression, COVID distress, and alcohol and cannabis use consequences from prior to the diary period (baseline) to after the diary period (follow-up).

Method

Participants and Procedures

The current analyses used data from a longitudinal study of substance use among SGM-AFAB. Participants were recruited via online advertisements on social media (e.g., Facebook). The majority of participants reported hearing about the study via advertisements on Facebook or Instagram (88.3%) and the remained heard about it from a friend or indirectly (e.g., via a twitter post). Recruitment took place between August 2020 and May 2021. The study included a baseline assessment (day 0), a 30-day ecological momentary assessment study (days 1–30), and a follow-up assessment (completed within two weeks of day 30). This study predominately uses data from the 30-day ecological momentary assessment. Data from the baseline and follow-up assessments are also included in a subset of analyses. During the ecological momentary assessment, participants were invited (via text message or email) to complete one survey in the morning (at 8:00am in their time zone) and one in the evening (at 6:00pm in their time zone). Participants who had not completed the survey by three hours after the survey invitation were sent a reminder. Participants had from 8:00am–1:00pm to complete the morning survey and between 6:00pm–12:00am to complete the evening survey. Participants who missed more than three surveys in a row were contacted by study staff to check in and re-engage participants. Surveys took approximately 2 minutes to complete. The study received IRB approval at Northwestern University.

Eligible participants were U.S. residents, 18–25 years old, identified as lesbian, bisexual, pansexual, or queer, were assigned female at birth, and met alcohol or cannabis use criteria (i.e., reported having 4 or more drinks at least twice and/or using cannabis on at least three days in the past month).¹ Participants were paid up to $150 based on completion rates: $20 for baseline, $20 for follow-up, $1 for each daily diary survey, and $5 bonus for each 6 surveys completed in a row.

There were 429 participants. See Table 1 for demographics. The sample was comprised predominately of people of color, with 33.6% of the sample identifying exclusively as non-Latinx White. There was a sizeable number of gender minority participants (26.8%).

Table 1.

Demographics of Analytic Sample at Baseline (N = 429)

Demographic Variable	n	%
Sexual Identity
Lesbian	112	26.1%
Bisexual	111	25.9%
Pansexual	112	26.1%
Queer	94	21.9%
Race/Ethnicitya
White	235	54.8%
Black	102	23.8%
Latinx	129	30.1%
Asian	53	12.4%
Other Race/Ethnicity	34	7.9%
Gender Identity
Cisgender Women	314	73.2%
Non-Binary	63	14.7%
Genderqueer	20	4.7%
Gender Non-Conforming	19	4.4%
Another Identity	13	3.0%
Substance Use Criteria Met
Alcohol Only	110	25.6%
Cannabis Only	107	24.9%
Alcohol and Cannabis	212	49.4%
Age (M, SD)	22.27 (2.01)

^aPercentages add up to more than 100% because participants could select multiple racial/ethnic identities.

Measures

Daily Diary Measures

COVID News Exposure

COVID News Exposure was assessed using a single item administered during the evening surveys. Specifically, participants were asked “Since this time yesterday, how many hours were you exposed to coronavirus information (radio, TV, Twitter, Facebook, Instagram, newspapers, etc.)?” Response options ranged from 1 (none at all) to 5 (more than 2 hours). This is a face-valid item developed for the current study.

COVID Distress

COVID Distress was assessed using a single item administered during evening surveys. Participants were asked “Since this time yesterday, how much distress have you experienced related to the coronavirus pandemic?” Response options ranged from 1 (no distress) to 5 (extreme distress). This item was developed for the current study, and the convergent and divergent validity of this item is demonstrated by its moderate between-person correlations with anxious and depressed affect (r = .41–.42).

Affect.

Affect. Depressed and anxious affect were each assessed using two items from the Profile of Mood States Short Form’s depressed (discouraged; sad) and anxious (anxious; on edge) affect subscales (Cranford et al., 2006; Curran et al., 1995). Positive affect was assessed using two items from the Positive and Negative Affect Scale’s positive affect subscale (happy; cheerful). Participants were asked to indicate how much they felt each emotion since the last survey. Response options ranged from 0 (not at all) to 4 (extremely). Affect was assessed prior to any other items on the daily survey to avoid order effects and was assessed in the morning and evening. Internal consistency was estimated using procedures recommended by Nezlek (2017) for longitudinal data and was acceptable (.63–.71). These measures are well-validated, have been demonstrated to be reliable measures of mood at the daily level, and are widely used in EMA studies in a wide range of populations (Cranford et al., 2006; O’Hara et al., 2014; Röcke et al., 2009). Studies with SGM provide evidence of the predictive validity of these measures via significant within- and between-persons associations between minority stressors and anxious and depressed affect (Feinstein et al., 2022; Kiekens & Mereish, 2022; Livingston et al., 2020; Mereish et al., 2021).

Substance Use

Substance Use was assessed by asking participants to indicate whether or not they had used alcohol or cannabis since the last survey. Participants who reported alcohol/cannabis in the morning or evening survey were coded as 1 and those who had not were coded as 0. When participants endorsed drinking, they were asked how many drinks they consumed since the last survey. Participants could indicate any number between 1 and 49 drinks or indicate having consumed 50 or more drinks. On cannabis use days, participants were asked to indicate how many hours they were high since the last survey and could indicate any number between 0 and 11 hours or indicate being high for 12 or more hours. Daily composites of number of drinks and hours of intoxication were created by averaging morning and evening assessments. These are face-valid items that have been used in a number of EMA studies of substance use (e.g., Gunn et al., 2021; Patrick et al., 2020; Shrier et al., 2013).

Coping Motives for Substance Use

Coping Motives for Substance Use were assessed in both morning and evening surveys. Participants who indicated any alcohol or cannabis use since the last survey were asked to indicate why they drank or used cannabis. Three items adapted from a daily measure of motives for alcohol and cannabis use (Patrick et al., 2019) were used to assess coping motives for alcohol and cannabis use (“to reduce my anxiety”). Participants responded on a scale of 0 (no) and 1 (yes). Participants who endorsed at least one coping motive for cannabis use in the morning or evening survey received a score of 1 and those who did not received a score of 0. A similar composite score was created for drinking motives. While the Patrick et al. (2019) measure of coping motives has not been subjected to thorough psychometric analyses to date, several closely related measures of daily coping motives for alcohol use have demonstrated concurrent (via moderate associations with negative affect), divergent (via small to moderate associations with other daily drinking motives), and predictive validity (via associations with number of drinks and alcohol consequences; Votaw & Witkiewitz, 2021).

Baseline/Follow-Up Measures

Anxiety

Anxiety was assessed using the Generalized Anxiety Disorder 7 item (Spitzer et al., 2006). Participants were asked to indicate how often they had been bothered by seven symptoms of anxiety (α=.90–.91) over the past 30 days (“worrying too much about different things”) on a scale of 0 (not at all) to 3 (nearly every day). The GAD-7 has strong psychometric properties, including high reliability as well as strong criterion, factorial, and procedural validity (Spitzer et al., 2006). Its psychometric properties have been examined in a range of populations, including among SGM (Borgogna et al., 2021; Löwe et al., 2008; Mills et al., 2014; Rutter & Brown, 2017).

Depression

Depression was assessed using the 8-item Patient Health Questionnaire (Kroenke et al., 2009). Participants were asked to indicate how often they had been bothered by eight symptoms of depression (α=.88–.89) over the past 30 days (“feeling down, depressed or hopeless”) on a scale of 0 (not at all) to 3 (nearly every day). The PHQ-8 has strong psychometric properties including high reliability as well as strong criterion, factorial, and procedural validity (Kroenke et al., 2009). Its psychometric properties have been demonstrated among SGM (Borgogna et al., 2021).

COVID Distress

COVID Distress was measured using five-items (α=.81) that included questions about how worried participants had been over the past 30 days about: being infected by the coronavirus, friends or family being infected, having one’s own or friends’/family member’s physical or mental health affected by the pandemic (Merikangas et al., 2020). Items were measured on a scale of 1 (not at all) to 5 (extremely). A single factor structure has been established for this measure as well as the convergent validity via associations with impact of COVID-19 on one’s family (Nikolaidis et al., 2022). The current study is the first study we are aware of to use this measure with SGM. An EFA indicated that this measure has single factor structure in this sample. Moderate correlations between this measure and the GAD-7 and PHQ-8 (r = .35–.38) also provide support for this measure’s convergent validity among SGM.

Alcohol and Cannabis Use Consequences

Alcohol and Cannabis Use Consequences were assessed using the Brief Young Adult Alcohol Consequences Questionnaire (24 items; Kahler et al., 2005) and the Brief Marijuana Consequences Measure (21 items; Simons et al., 2012). Participants were asked to indicate whether they had experienced each consequence (e.g., “I have taken foolish risks when I have been drinking” and “I haven’t been as sharp mentally because of my marijuana use”) in the past 30 days on a scale of 0 (no) and 1 (yes). Cronbach’s alphas indicated high internal consistency (α=.86–.87). The psychometric properties of the Brief Young Adult Alcohol Consequences Questionnaire have been thoroughly investigated, providing evidence of its internal consistency, sensitivity to changes in drinking over time, and convergent validity (Kahler et al., 2008; Kahler et al., 2005). This measure has been previously used with SGM samples, including samples of sexual minority women, and has demonstrated reliability and convergent validity in these samples (Kelley et al., 2018; Schipani‐McLaughlin et al., 2022). The Brief Marijuana Consequences Questionnaire has also demonstrated strong psychometric properties, including reliability, convergent and discriminant validity, and metric and scalar invariance across countries (Bravo et al., 2019; Simons et al., 2012). This measure has not been previously used with SGM samples but demonstrates internal consistency and convergent validity in the current sample via a moderate association with frequency of cannabis use (r = .41).

Analytic Plan

Analyses were conducted in Mplus version 8.4. There was a total of 19,186 completed surveys from 429 participants. The median completion rate was 88.3% (M = 74%, SD = 28%). Individual completion rates ranged from 1.7% to 100%. All participants completed the baseline assessment and 408 (95.1%) completed the follow-up assessment. Within completed surveys, less than 1% of data were missing. Missing data were handled using Bayesian methods (Asparouhov & Muthén, 2010). Multilevel structural equation modeling (MSEM) with a Bayesian estimator and diffuse (non-informative) priors was used. MSEM utilizes latent variables, rather than group- and grand-mean centering, to separate within- from between-person variance (Lüdtke et al., 2008). By removing the between-person variance from the within-person variance, the within-person variables indicate the extent to which an individual was experiencing more/less of a construct than usual (above/below their person mean) on a particular day (e.g., experiencing more/less anxiety than usual). We used Markov Chain Monte Carlo (MCMC) algorithms to generate a series of 10,000 random draws from the multivariate posterior distribution of our sample for each model. Trace plots and the Gelman-Rubin potential scaling reduction (PSR) were used to determine whether convergence was achieved (Depaoli & Clifton, 2015; Muthén, 2010). A probit link was used for binary outcomes (e.g., alcohol use). Probit regression coefficients represent the variance shared by the predictor and the latent continuous response variables underlying each binary observed item (Agresti, 2003).

We examined the associations between COVID news exposure and affect, substance use, and coping motives for substance use. In each of model, within- and between-person components of COVID news exposure were modeled as predictors of within- and between-person components of an affect, substance use, or coping motives variable. In all models, the linear association between day of assessment (e.g., day 1 of 30) and the outcome variable as well as the first-order autocorrelation of the outcome’s residuals (i.e., the correlation between the residual at day t−1 with the residual at day t) were included at the within-person level. Including first-order autocorrelations effectively controls for the prior timepoint of the outcome and controlling for the day of assessment removes any changes in substance use that may result from participating in an EMA study (i.e., reactivity). All within-person associations were allowed to vary across individuals. Age, sexual identity, gender identity, race/ethnicity, and months between the start of the COVID-19 pandemic (March 11, 2021; World Health Organization, 2020) and the completion of the first daily survey were included as covariates at the between-person level. Prior research with SGM-AFAB has consistently demonstrated differences in substance use and internalizing symptoms based on age, race, ethnicity, gender identity, and sexual identity (Dyar et al., 2015; Dyar et al., 2019; Fish, Bishop, et al., 2021; Ross et al., 2017; Shokoohi et al., 2022). Therefore, we controlled for age, race, ethnicity, gender identity, and sexual identity at the between-person level. Months since the start of the pandemic was included as a covariate in order to remove any between-person differences in the outcomes that may have resulted from when individuals participated in the EMA. Only observations during which participants reported drinking or using cannabis were included in analyses of coping motives for substance use as motives for use were only assessed and are only relevant on days when participants use substances.

We examined both concurrent (COVID news exposure at time t predicting outcome at time t) and prospective (COVID news exposure at t predicting outcome at t+1) within-person associations. Given that COVID news exposure was only assessed once per day and most variables were assessed twice a day (except for COVID distress), we took the following approach to analyses. For analyses involving substance use and coping motives for substance use, we examined whether COVID news exposure reported on one day (reported at t covering period of t−1 and t) predicted any substance use during t−1 or t (i.e., same day; concurrent) or during t+1 and t+2 (i.e., next day; prospective). Given that affect had substantial within-day variability (40–43% of variance at the within-day level), an average daily affect score was not appropriate, and we took a different approach. For analyses involving affect, we examined whether COVID news exposure reported during an evening assessment predicted affect reported during the same evening assessment (concurrent) or during the next morning assessment (prospective). Given that the association between COVID news exposure and anxious/depressed affect may be bi-directional, we also tested prospective effects in the opposite direction. In these analyses, affect reported during an evening assessment predicted COVID news exposure reported during the next evening assessment. Using affect during the prior evening assessment and not the prior morning assessment was necessary as COVID news exposure asks about news exposure over the past 24 hours.

To assess the potential cumulative effect of exposure to COVID news, we also conducted regression models in which the sum of COVID news exposure across all days of the daily diary period represented cumulative COVID news exposure. Cumulative COVID news exposure predicted internalizing symptoms, COVID distress, and alcohol and cannabis use consequences at follow-up (completed after the daily diary period). Covariates included outcome at baseline, proportion of daily surveys completed (to allow for the use of a sum score rather than a mean score for cumulative COVID news exposure), age, sexual identity, gender identity, race/ethnicity, and months between the start of the pandemic and completion of the first daily diary study.

Results

Participants indicating drinking on 3,353 days (26% of days) and using cannabis on 4,152 days (32% of days). On drinking days, participants consumed an average of 2.91 drinks (SD=2.25) and endorsed coping motives for use on 57% of drinking days. On cannabis use days, participants reported being high for an average of 4.34 hours (SD=3.17) and endorsed coping motives for use 81% of cannabis use days. See Table 2.

Table 2.

Correlations, Means, Variances, and Intraclass Correlations

	1	2	3	4	5	6	7	8	9	10	11
1. COVID new exposure	-	.39*	.06*	.08*	−.03*	−.01	.002	−.02	−.02	.08*	.10*
2. COVID-Related Distress	.62*	-	.11*	.13*	−.06*	.001	−.02	.00	−.04*	.02	.07*
3. Depressed Affect	.33*	.41*	-	.56*	−.47*	−.04*	−.04*	−.01	−.02	.22*	.18*
4. Anxious Affect	.33*	.42*	.84*	-	−.39*	−.04*	−.03	−.01	−.03*	.21*	.21*
5. Positive Affect	−.05	−.16*	−.33*	−.30*	-	.20*	.11*	.12*	.12*	−.21*	−.13*
6. Alcohol Use	−.01	−.06	−.09	−.06	.17*	-	.25*	a	.10*	a	−.17*
7. Cannabis Use	−.03	.06	.02	.04	−.03	−.04	-	.03	a	−.09*	a
8. Number Drinks	.09	−.08	.23*	.14*	−.09	a	.03	-	b	.08*	b
9. Hours High (Cannabis)	.04	.14*	.16*	.18*	−.10	−.14*	a	b	-	b	.17*
10. Drinking to Cope	.26*	.15*	.34*	.32*	−.25*	a	−.16*	.20*	b	-	b
11. Using Cannabis to Cope	.12	.12	.26*	.26*	−.21*	−.14	a	b	.32*	b	-
Mean	2.20	2.45	1.22	1.44	1.95	.31	.38	2.91	4.34	.57	.81
Standard Deviation	1.04	1.06	1.00	1.04	.99	.46	.49	2.25	3.17	.49	.39
Range	1–5	1–5	0–4	0–4	0–4	0–1	0–1	1–34	0–21	0–1	0–1
Intraclass Correlation	.51	.64	.47	.48	.50	.38	.70	.35	.42	.46	.48

Correlations above the diagonal are within-person correlations; those below the diagonal are between-person correlations;

^aindicates an association that cannot be estimated as one variable (e.g., number of drinks) was only assessed during observations when the other variable was at a certain value (e.g., alcohol use was endorsed);

^bindicates that both variables were only assessed on a relatively small number of unique days (i.e., when cannabis and alcohol were used) - these correlations are not presented.

^*p < .05.

Daily Diary Models

At the within-person level (Table 3), COVID news exposure was significantly² associated with concurrent COVID distress, depressed affect, anxious affect, and positive affect. Specifically, when participants were exposed to more COVID news than on a typical day, they reported experiencing more distress about COVID (b = .32, p < .001), depressed affect (b = .05, p < .001), and anxious affect (b = .07, p < .001) as well as less positive affect than usual (b = −.03, p =.04). COVID news exposure was also significantly associated with drinking/using cannabis to cope. On cannabis use days when participants were exposed to more COVID news than usual, they were more likely to indicate that they were motivated to use cannabis to cope with negative affect (b = .18, p < .001). Similarly, on drinking days when participants were exposed to more COVID news than usual, they were also more likely to indicate that they were motivated to drink to cope (b = .11, p = .02). COVID news exposure was not significantly associated with the likelihood of drinking or using cannabis, number of drinks consumed, or duration of cannabis intoxication on the same day.

Table 3.

Concurrent and Prospective Associations between COVID News Exposure, Affect, and Substance Use

		Within-Person Associations
		Concurrent Associations			Prospective Associations			Prospective Predicting COVID News Exposure			Between-Person Associations
Predictor	Outcome	b	95% CI	p	b	95% CI	p	b	95% CI	p	b	95% CI	p
COVID news exposure	COVID-Related Distress	.32	.29, .35	< .001	.06	.04, .09	< .001	.09	.06, .13	< .001	.70	.58, .81	< .001
	Depressed Affect	.05	.02, .08	< .001	.03	.001, .06	.04	.02	−.004, .04	.11	.27	.15, .39	< .001
	Anxious Affect	.07	.04, .10	< .001	.03	−.01, .06	.12	.01	−.01, .03	.19	.33	.21, .46	< .001
	Positive Affect	−.03	−.05, −.001	.04	−.03	−.06, −.002	.04	-	-	-	−.06	−.18, .06	.32
	Alcohol Use	−.02	−.07, .03	.50	.02	−.03, .07	.50	-	-	-	−.02	−.17, .13	.80
	Cannabis Use	−.01	−.02, −.005	.66	.03	−.02, .09	.19	-	-	-	−.13	−.42, .17	.39
	Number Drinks	.04	−.08, .16	.48	.02	−.12, .16	.76	-	-	-	.36	−.66, −.07	.02
	Hours High (Cannabis)	−.01	−.12, .11	.88	.03	−.09, .16	.56	-	-	-	.14	−.36, .63	.57
	Drinking to Cope	.11	.03, .20	.02	.07	−.01, .17	.10	-	-	-	.36	.07, .69	.01
	Using Cannabis to Cope	.18	.08, .27	< .001	.17	.07, .27	< .001	-	-	-	.20	−.11, .52	.21

In all models, the linear association between the day of the assessment (e.g., day 1 of 30) and the outcome variable as well as the first-order autocorrelation of the outcome’s residuals (i.e., the correlation between the residual at day t−1 with the residual at day t) were included at the within-person level. All within-person associations were allowed to vary across individuals. Age, sexual identity, gender identity, race/ethnicity, and months between the start of the COVID-19 pandemic and the completion of the first daily survey were included as covariates at the between-person level.

At the within-person level, COVID news exposure was prospectively associated with COVID distress, depressed affect, and positive affect. Specifically, when participants were exposed to more COVID news than usual on one day, they experienced significant increases in COVID distress (b = .06, p < .001) and depressed affect (b = .03, p = .04) as well as significant decreases in positive affect from that day to the next (b = −.03, p =.04). Only one model testing prospective effects in the opposite direction (affect predicting subsequent COVID news exposure) was significant, with participants who experienced more COVID distress on one day experiencing a significant increase in COVID news exposure from that day to the next (b = .09, p < .001). However, neither depressed (b = .02, p = .11) nor anxious affect (b = .01, p =.19) prospectively predicted COVID news exposure. COVID news exposure also prospectively predicted using cannabis to cope (but not drinking to cope). When participants were exposed to more COVID news than on a typical day, they were significantly more likely to use cannabis to cope on the next day (b = .17, p < .001). COVID news exposure was not significantly associated with the likelihood of drinking or using cannabis, likelihood of drinking to cope, nor with the number of drinks consumed or the duration of cannabis intoxication on the next day.

At the between-person level, COVID news exposure was significantly associated with COVID distress, depressed affect, and anxious affect. This indicates that participants who tended to be exposed to more COVID news on average (i.e., across all observations) also tended to have more COVID distress (b = .70, p < .001), depressed affect (b = .27, p < .001), and anxious affect (b = .33, p < .001) on average. Further, COVID news exposure was associated with drinking to cope and quantity of alcohol consumption at the between-person level. Specifically, participants who tended to be exposed to more COVID news on average also endorsed coping motives for drinking more frequently (b = .36, p = .01) but consumed a lower quantity of alcohol on each day on average (b = −.36, p = .02). COVID news exposure was not significantly associated with any other substance use variables (e.g., likelihood of alcohol or cannabis use, using cannabis to cope) at the between-person level.

There were also several demographic differences at the between-person level (Table 4). There were several differences by age, with older participants reporting lower average depression, anxiety, quantity of alcohol consumed per day, and number of hours of intoxications per day, but older participants also drank more frequently during the EMA period. Gender minority individuals reported a higher average level of COVID news exposure, COVID-related distress, and depressed and anxious affect compared to cisgender women. There were relatively few differences by sexual identity. Queer-identified individuals reported lower drinking frequency and endorsed coping motives for drinking less often than lesbian individuals. Pansexual individuals reported less COVID news exposure than lesbian individuals. There were also some sporadic differences by race/ethnicity, with most differences between individuals who identified as Black and those who did not. Specifically, Black individuals reported lower COVID news exposure, lower positive affect, less frequent drinking, and lower average number of drinks per day than individuals who did not identify as Black. In addition, Latinx individuals reported less frequent alcohol use than non-Latinx individuals; Asian individuals reported less positive affect than those who did not identify as Asian; and individuals who identified with racial or ethnic labels other than Black, White, Latinx, or Asian reported coping motives for drinking less often than individuals who did not identify with these labels.

Table 4.

Between-Person Associations with Demographics

Predictor	COVID News Exposure	COVID Distress	Depressed Affect	Anxious Affect	Positive Affect	Alcohol Use	Cannabis Use	Number Drinks	Hours High	Drinking to Cope	Using Cannabis to Cope
Age	.03	.05	−.13*	−.10*	.02	.25*	−.03	−.22*	−.14*	.06	.07
Gender
Cisgender Woman	ref	ref	ref	ref	ref	ref	ref	ref	ref	ref	ref
Gender Minority	.14*	.11*	.12*	.16*	−.08	−.07	.04	−.02	.07	−.04	−.01
Sexual Identity
Lesbian	ref	ref	ref	ref	ref	ref	ref	ref	ref	ref	ref
Bisexual	−.04	−.01	.05	.04	−.08	.01	−.03	−.08	−.01	−.07	−.08
Queer	−.05	.02	.01	.06	−.04	.03	.02	−.19*	.04	−.16*	−.05
Pansexual	−.14*	−.10	−.03	−.01	.05	.01	−.003	−.03	−.08	−.01	−.02
Race/Ethnicity
White	−.01	.01	−.05	.03	−.05	.05	.07	−.02	−.10	−.09	−.04
Black	−.24*	−.12	−.07	−.07	−.14*	−.18*	.03	−.16*	.10	.03	.01
Latinx	−.05	.04	−.02	−.02	−.08	−.14*	−.06	−.10	−.09	−.11	−.01
Asian	−.03	−.02	.05	.02	−.14*	−.08	.03	−.04	−.06	−.08	−.03
Other	−.05	−.06	−.01	−.07	.02	−.02	.07	−.08	.07	−.17*	−.03

^*p < .05.

Analyses Examining Potential Cumulative Impact of COVID News Exposure

Next, we conducted regression analyses examining the association between cumulative COVID news exposure across the full diary study and subsequent changes in COVID distress, internalizing symptoms, and consequences of substance use (see Table 5). Exposure to more COVID news across the length of the diary study was associated with increases in COVID distress and internalizing symptoms from baseline to follow-up. However, the associations between cumulative COVID news exposure and changes in alcohol and cannabis use consequences were not significant.

Table 5.

Associations Between Cumulative COVID News Exposure and Changes in Internalizing Symptoms and Substance Use

Predictor	Outcome	β	p
Cumulative COVID news exposure	COVID Distress	.25	< .001
	Depression	.15	< .001
	Anxiety	.23	< .001
	Alcohol Consequences	.003	.95
	Cannabis Consequences	−.04	.51

Cumulative COVID news exposure was operationalized as the sum of COVID news exposure across all days in the daily diary study. Control variables included: proportion of daily diary days completed, outcome at baseline, age at baseline, months since the beginning of the pandemic at the first daily diary, sexual identity, gender identity, and race/ethnicity.

Discussion

The current study is the first we are aware of to examine associations between COVID news exposure, internalizing symptoms, and substance use using daily diary data among SGM-AFAB. Results indicate that on days when participants were exposed to more COVID news than usual, they experienced more depressed and anxious affect, less positive affect, more COVID distress, and were more likely to drink or use cannabis to cope. Prospective associations between COVID news exposure, COVID distress, depressed affect, positive affect, and using cannabis to cope were also significant. Further, there was a cumulative impact of COVID news exposure on anxiety, depression, and COVID distress. These findings highlight the potential negative impact of overexposure to COVID news on a daily basis among SGM-AFAB – a high-risk population that may be disproportionately impacted by the pandemic.

At the daily level, we found that COVID news exposure was associated with COVID distress, depressed/anxious affect, and positive affect. This finding extends previous cross-sectional research demonstrating similar associations (Bendau et al., 2021; Gao et al., 2020; Hong et al., 2021; Olagoke et al., 2020; Riehm et al., 2020) by providing evidence that elevations in COVID news exposure, COVID distress, and anxious/depressed affect tend to co-occur on the same days. These associations remained significant in prospective analyses, with higher news exposure predicting higher distress, depressed affect, and lower positive affect. The association between COVID news exposure and depressed affect was unidirectional with COVID news exposure predicting increases in depressed affect, but depressed affect not prospectively predicting higher COVID news exposure. COVID news exposure did prospectively predict increased COVID distress and vice versa suggesting that associations between COVID news exposure and broad COVID-related distress are bi-directional, consistent with two recent daily diary studies (Schmidt et al., 2021; Shaw et al., 2021). This is concerning as such a bi-directional association may lead to the development of a cycle in which individuals who are experiencing distress about COVID seek out more COVID news, likely in an attempt to reduce ambiguity about aspects of the pandemic and distress. However, this news-seeking may lead to exposure to more conflicting and distressing information, which in turn leads to more distress as proposed by Garfin et al. (2020). Notably, data were collected from August 2020 to May 2021 (5–15 months into the pandemic), far later than existing daily diary studies which focused on the first few months of the pandemic (Chu et al., 2021; Schmidt et al., 2021; Shaw et al., 2021). This indicates that COVID news exposure affected individuals far beyond the initial weeks and months of the pandemic.

We also found that COVID news exposure was only significantly associated with two substance use variables at the within-person level – drinking to cope and using cannabis to cope. The association between COVID news exposure and using cannabis to cope was present both concurrently and prospectively, indicating that high COVID news exposure on one day predicted a higher likelihood of using cannabis to cope on the same day as well as increases in the likelihood of using cannabis to cope from that day to the next. However, COVID news exposure only concurrently predicted drinking to cope. Using cannabis to cope and drinking to cope are known risk factors for the development of problematic use (Bresin & Mekawi, 2019), and rates of cannabis and alcohol use disorders are particularly elevated among sexual minority individuals assigned female at birth (Kerridge et al., 2017; Krueger et al., 2020). This furthers the urgent need for cannabis use research with SGM-AFAB populations.

In addition to the daily effects described above, we also found evidence of a cumulative effect of COVID news exposure on internalizing symptoms and COVID distress. This study is the first to demonstrate a cumulative impact of COVID news exposure on internalizing symptoms, and extend prior research demonstrating a cumulative impact of exposure to multiple traumatic events in the news (Garfin et al., 2015). These findings suggest that in addition to temporally limited associations between COVID news exposure, internalizing symptoms, and COVID distress at the daily level, individuals who are chronically exposed to more COVID news may be more likely to develop more lasting negative mental health effects, potentially including development or exacerbation of existing mood disorders.

No other associations between COVID news exposure and substance use were significant at the within- or between-person levels nor was cumulative COVID news exposure associated with changes in substance use consequences. This suggests that factors other than COVID news exposure are driving changes in substance use experienced by many populations during the pandemic. Future research should explore the roles of other stressors and changes arising from the COVID pandemic, such as isolation, financial difficulties, and reduced access to resources (Abrams & Szefler, 2020; Singu et al., 2020).

Gender identity and race also were associated with differences in COVID news exposure and COVID related distress. In addition to reporting more COVID news exposure – a novel finding, gender minority individuals reported more COVID-related distress, depressed affect, and anxious affect compared to cisgender women, which is consistent with research demonstrating that gender minority individuals’ mental health has been disproportionately impacted by the pandemic due to a range of interpersonal and structural factors (e.g., disruptions to mental health and gender affirming care, reduced support from family members, economic insecurity; Hawke et al., 2021; Jarrett et al., 2021). Additionally, pansexual and Black individuals reported lower COVID news exposure compared to lesbian individuals and those who did not identify as Black. Few studies focused on COVID news exposure have examined demographic differences in rates of news exposure; thus, future research should replicate these findings before any strong conclusions should be drawn based on them. A myriad of factors may contribute to reduced COVID news exposure among Black SGM-AFAB. For example, structural racism leads Black individuals to be more likely to be of lower socioeconomic status and have jobs associated with higher risk for COVID exposure, which may contribute to COVID burnout and news avoidance (Asfaw, 2021; Goldman et al., 2021; US Census Bureau, 2019). Further, the plethora of traumatic news in 2020, including news of anti-Black police violence and the COVID pandemic, may have contributed to news exhaustion or burnout among Black individuals (Campbell & Valera, 2020) and led them to engaged in news avoidance as a form of self-protection (Mannell & Meese, 2022; Ytre-Arne & Moe, 2021). If racial differences in COVID news exposure are replicated, future research should explore which factors contribute to reduced COVID news exposure among Black individuals.

Several additional demographic differences emerged in average levels of affect and substance use, and these were largely consistent with past research. Older participants reported lower depressed and anxious affect and a higher frequency of alcohol use, consistent with past research among SGM samples of a similar age (Birkett et al., 2015; Dyar et al., 2017; Dyar et al., 2020). However, we also found that older participants reported fewer drinks per day and a lower duration of intoxication, which add to a mixed literature on this topic (Fairlie et al., 2018; Lee et al., 2021; Litt et al., 2013). There were few differences based on sexual identity. This finding is consistent with some studies (Dyar et al., 2020; Watson et al., 2020) but not with a recent meta-analysis, which found evidence of higher rates of internalizing symptoms among bisexual compared to lesbian/gay individuals (Ross et al., 2017). Further, we found that Black participants reported lower positive affect, alcohol use frequency, and number of drinks and several other racial/ethnic differences emerged, with POC consistently having similar or lower levels of substance use. These findings are consistent with several prior studies, which have interpreted this as potential evidence of SGM POC’s resilience in the face of experiencing stigma based on their sexual/gender identity and their race or ethnicity (Balsam et al., 2015; Dyar et al., 2019; Kertzner et al., 2009; Talley et al., 2014).

Clinical Implications

Current study results have clinical implications for the current pandemic and for future traumatic major events. Clinicians working with clients who are continuing to experience distress about the COVID pandemic may work with clients to set guidelines for themselves around how frequently they engage with media about COVID. Complete disengagement from the media in the context of continued changes in the state of the pandemic and the WHO, CDC, and government recommendations is not advisable. Rather, it is necessary to balance the need to stay informed with reducing the mental health effects of COVID news exposure by determining a maximum level of healthy engagement with the media, consuming higher quality media over a larger quantity of media, and determining which media outlets may encourage healthy engagement with COVID news. Clinicians may also want to conduct a functional analysis of relationships between media consumption and psychiatric distress and work with clients experiencing bidirectional cycles of escalating consumption and distress to implement alternative coping strategies.

Limitations

Findings should be considered in light of study limitations. First, this study utilized data from a racially diverse sample of sexual and gender minority individuals assigned female at birth. While this is a strength of the study, it is unclear whether findings will generalize to sexual and gender minority individuals assigned male at birth. Second, the study also only included SGM-AFAB who were engaging in regular heavy episodic drinking or cannabis use, and thus it is unclear whether findings generalize to individuals who engage in substance use less frequently. Third, we did not assess whether participants currently or previously had COVID and were thus unable to control for this variable or examine whether this moderated the association between COVID news exposure and COVID-related distress. Fourth, our measure of COVID news exposure was assessed using a single broad item. While this item (“Since this time yesterday, how many hours were you exposed to coronavirus information (radio, TV, Twitter, Facebook, Instagram, newspapers, etc.)?”) provided a range of examples of ways in which individuals could be exposed to COVID-related news, it did not allow for the examination of how different sources of news may have differential impacts. This is an important direction for future research on the impact of types of media consumption on mental health. Fifth, given the current study’s focus on daily associations between COVID news exposure, mental health, and substance use, we did not examine associations with between-person level covariates, like financial difficulties and reduced access to resources.

Conclusion

The current study substantially furthers our understanding of the associations between COVID news exposure, internalizing symptoms, and substance use by examining these associations at the daily level, testing the directionality of these associations using prospective models, and examining the cumulative impact of exposure to COVID news over a 30-day period among SGM-AFAB. Results indicate that on days when participants were exposed to more COVID news than usual, they experienced more depressive and anxious affect, more COVID distress, less positive affect, and were more likely to use substances cope. Findings also provided evidence of bi-directional prospective associations between COVID news exposure and COVID distress as well as evidence of a cumulative impact of COVID news exposure on anxiety, depression, and COVID distress. Findings suggest that individuals should balance the need to remain informed about the pandemic and their own mental health when considering how much COVID news to consume.

Public Significance Statement:

Results indicate that when sexual and gender minority individuals assigned female at birth were exposed to more COVID news, they felt more anxious and depressed, were more worried about COVID, and were more likely to turn to substances to cope. Many of these effects persisted into the next day. These findings suggest that it is necessary to balance the need to stay informed about COVID with reducing the mental health effects of COVID news exposure by determining a maximum level of healthy engagement with the media, consuming higher quality media over a larger quantity of media, and determining which media outlets may encourage healthy engagement with COVID news.