Skip to main content
NCBI home page
Wiley - PMC COVID-19 Collection logoLink to Wiley - PMC COVID-19 Collection
. 2021 Mar 3;45(4):773–783. doi: 10.1111/acer.14575

Trends in US Alcohol Consumption Frequency During the First Wave of the SARS‐CoV‐2 Pandemic

Sarah McKetta 1,✉, Christopher N Morrison 1,2, Katherine M Keyes 1
PMCID: PMC8014717  NIHMSID: NIHMS1676655  PMID: 33587290

Abstract

Background

The SARS‐CoV‐2 pandemic created disruptions and stressors which may have influenced alcohol consumption frequency trends. Varying COVID‐19 health burden and alcohol policies may have contributed to different consumption trends between states. The aim of this study is to assess trends in alcohol consumption and moderation by state of residence.

Methods

We examined trends in adult drinking days, during the first wave of the pandemic (March 10 to June 8) using longitudinal data from the Understanding America Study (N = 6,172 unique participants; N = 28,059 observations). Because state mandates were responsive to disease burden, we modeled the interaction of time by COVID‐19 burden, defined as whether the state had the median (or higher) daily incidence of COVID‐19 cases on the survey date, and state random effects. We controlled for individual sociodemographics, perceived personal/familial COVID‐19 burden, mental health symptomology, and risk avoidance.

Results

Drinking days increased throughout the duration (incidence risk ratio [IRR] for drinking per increase in one calendar day = 1.003, 95% CI 1.001, 1.004); trends were heterogeneous by disease burden, with individuals living in states with lower COVID‐19 burden increasing (IRR = 1.005, 95% CI 1.003, 1.007) faster than those living in states with higher COVID‐19 burden (IRR = 1.000, 95% CI 0.998, 1.002). Trends were heterogeneous between states, but there was no evidence of systematic geographic clustering of state trends.

Conclusions

Drinking days increased during the first months of the COVID‐19 pandemic, particularly among residents of states with lower disease burden.

Keywords: Alcohol, SARS‐CoV‐2, Trends

During the first wave of the pandemic, Americans consumed alcohol at higher frequencies. However, we found heterogeneity in consumption by US state, specifically state‐level COVID‐19 burden. People living in states with higher COVID‐19 burdens did not increase alcohol consumption; the national increases were concentrated only among people living in states with lower COVID‐19 burdens.

graphic file with name ACER-45-773-g001.jpg

Community spread of SARS‐CoV‐2, the virus causing coronavirus disease 2019 (COVID‐19), in the United States led to unprecedented disturbances in the day‐to‐day lives of US residents. Millions of lives were impacted directly by health concerns related to the virus, and the majority of US residents also experienced school and daycare closures, job loss, economic insecurity, and stay‐at‐home orders restricting typical movement, consumption, and social routines (Alon et al., 2020; Bartik et al., 2020; Bayham and Fenichel, 2020; Shanthakumar et al., 2020; Tull et al., 2020). People experiencing financial, social, and family stress, as well as major disasters, may increase alcohol consumption as a form of coping (Brown et al., 2014; Diggs and Neppl, 2018; Fergusson et al., 2014; Keyes et al., 2011; Rospenda et al., 2010; Sillaber and Henniger, 2004); thus, the pandemic may have led to increases in alcohol consumption and problematic alcohol use. Indeed, in the early months of the pandemic, academic editorials warned that the virus and subsequent state and municipal responses may lead to increases in heavy drinking, relapses for those struggling with alcohol use disorder, and subsequently a spike in alcohol‐related diseases and disorders, such as liver disease (Clay and Parker, 2020; Da et al., 2020). Alternatively, many contexts and locations that promote alcohol consumption (e.g., social events, college campuses, bars) were disrupted due to the pandemic, and alcohol consumption may have in fact declined either at the population level or in select demographics as a result as opportunities for consumption dwindled.

Alcohol consumption is a leading cause of death and disability both globally and in the United States (GBD, 2016 Alcohol Collaborators, 2018; Stahre et al., 2014). Even at relatively low levels, alcohol consumption has been associated with adverse health outcomes, in particular for heart health (Braillon and Wilson, 2018; Gallagher et al., 2017). At the population scale, as average population alcohol consumption increases, so do higher‐risk alcohol consumption patterns like binge drinking, heavy drinking, and alcohol use disorders. Alcohol consumption as a form of coping is considered maladaptive and is associated with increased risks for depression and subsequent substance use disorders (Hussong et al., 2011; Thompson et al., 2010; Wardell et al., 2020). Acute increases in alcohol consumption are concerning for their increased risks of injuries and poisoning (Rehm and Imtiaz, 2016; Skog, 2001). If alcohol consumption increases as a response to COVID‐19, both short‐ and long‐term health effects may increase as well. The epidemic has persisted in the United States into late 2020, and widely available vaccines are not expected in the United States until late in 2021 (Goldhill, 2020; Kane et al., 2020). Social and economic disruptions may therefore continue for some time, in which case epidemic‐related increases in alcohol consumption could be anticipated to lead to a variety of morbidity and mortality outcomes including heart disease, metabolic diseases, liver disease, and cancers (Bagnardi et al., 2015; Kuypers et al., 2012; Mellinger, 2019; Rehm and Imtiaz, 2016; Skog, 2001; Tramacere et al., ,2010, 2011). While the quantity of alcohol individuals consume is directly related to disease risk, so is consumption frequency (e.g., number of drinking days per week or month). Higher frequencies of drinking days are related to both all‐cause mortality (Hartz et al., 2018) and chronic disease risk (Elgendy et al., 2019), even among those who only drink one or two drinks daily (Hartz et al., 2017).

Evidence from other countries suggests that the pandemic and its impact on social systems have influenced alcohol consumption. However, the direction of change has not been consistent with the hypothesized increases. In Poland, for example, 30% of adults reported changing their drinking patterns during the COVID‐19 pandemic; but among those respondents, 14% increased and 16% reduced alcohol consumption (Chodkiewicz et al., 2020). In the United States, many economic indicators suggest alcohol consumption may be increasing: for example, a Nielson report early in the pandemic found sales of alcohol increased by 55% in the third week of March 2020 (Brenmer, 2020), and online liquor sales increasing to 243% compared with the same time period a year prior (House, 2020). It is unclear, however, to what extent these national sales trends correspond to consumption—early in the pandemic, many consumers were purchasing goods in excess for fear of shortages, and online sales may simply have supplanted on‐site or in‐person purchases at liquor stores or bars (Loxton et al., 2020). Limited research on self‐reported alcohol consumption in the United States during this time period does suggest, however, that alcohol consumption increased as a result of the pandemic: compared with the spring of 2019, adults in May and June of 2020 on average reported greater frequencies of past‐month alcohol consumption (Pollard et al., 2020). US residents who reported feelings of anxiety and fear related to COVID‐19 consumed more drinks in a typical drinking occasion, and on more days in a typical month (Rodriguez et al., 2020). A sample of US college students—a population normally at elevated risks of alcohol consumption—endorsed higher levels of alcohol consumption after university closure due to COVID‐19 than in the weeks before, with students reporting anxiety and depression symptomology reporting the highest increases (Lechner et al., 2020). Further, solitary drinking (i.e., drinking alone rather than with other people) increased among people who reported social distancing—that is, decreasing the frequency of contact and increasing the physical space between individuals—compared with retrospective recall of their drinking behaviors prior to social distancing (McPhee et al., 2020). Regarding coping, findings from a Canadian sample showed that alcohol consumption frequency and solitary drinking increased during the pandemic relative to the month prior to the pandemic and that drinking as a way to cope with stress was a major contributor to these changes (Wardell et al., 2020). Stressors during the COVID‐19 pandemic—for example, depressive symptoms, lower social connectedness, childcare stress—increased alcohol use and were mediated through individual coping drinking motivation. Alcohol consumption may have continued to increase as the pandemic progressed, based on evidence comparing consumption rates in March 2020 to April 2020, as mental health has worsened (French et al., 2020).

Alcohol policy is largely determined at the state‐ or local‐ (i.e., county‐, municipal‐) level, and throughout the pandemic in the United States, individual states and jurisdictions within states reacted with heterogeneous changes to alcohol policies responses, which may have contributed to changes in alcohol consumption. For example, the timing, enforcement, and details of state stay‐at‐home orders varied from state to state, all of which led to varying social and economic environments for residents. In addition, some states responded to the pandemic by expanding alcohol access; for example, New York State expanded alcohol sales to include off‐premises consumption and sales of “to go” drinks. However, a handful of states—including Pennsylvania and New Mexico—restricted alcohol access by not deeming liquor stores essential businesses (Rieper, 2020). Both state‐level and individual‐level responses have been sensitive to COVID‐19 burden—for example, local disease prevalence, rates of hospitalization, etc.—suggesting that variability in alcohol consumption may also be related to state‐level COVID‐19 burden (Bergquist et al., 2020; Capano et al., 2020).

While data are emerging to indicate that both alcohol consumption frequency and volume have increased during the COVID‐19 pandemic, it is unknown how these alcohol consumption patterns relate to state‐level trends during this time period. Based on the existing evidence linking alcohol consumption during the pandemic to stress and coping pathways, we anticipate that residents of states with higher COVID‐19 burdens have increased the frequency of alcohol consumption relative to those in states with lower COVID‐19 burdens. On the other hand, those in states with lower COVID‐19 burden may have fewer limitations to accessing alcohol, especially if venues for purchasing and consuming alcohol remained open during the first wave of the pandemic (i.e., in the spring and summer when cases initially peaked and then declined nationwide before increasing again). To better understand alcohol consumption behaviors during the COVID‐19 pandemic, we compared temporal associations between alcohol consumption frequency and state‐level COVID‐19 burden during the first wave of the pandemic in the United States.

Materials and Methods

Sample

We used self‐reported data from the Understanding America Study (UAS), a national, longitudinal survey of roughly 7,400 US households administered online by University of Southern California (Alattar et al., 2018). This is an ongoing survey series with continuous recruitment beginning in 2014 and currently has 22 sample batches being monitored longitudinally. Batch recruitment refers to the process by which respondents are requested to join the survey; briefly, the unweighted demographics of the current sample are analyzed, and new respondents are recruited to re‐balance the sample to maintain national representativeness. This is done through targeting ZIP codes based on demographic composition, and contacting households within those targeted ZIP codes via US mail with survey login information. Response rate varies by sample batch (i.e., changes across each new recruitment period), with approximately 15% of surveyed households ultimately participating in the online panels, that is, joining the UAS survey respondent pool.

Beginning in March 2020, adults over the age of 18 received questions specific to COVID‐19 and behavioral health. These surveys were administered in five overlapping periods between March 10 and June 8, 2020. At the first survey, UAS invited 8,815 respondents who were active in their respondent pool to respond to surveys querying behaviors during the COVID‐19 pandemic; of these, 7,145 (81%) completed the initial survey. Subsequent surveys continued to query samples from the entire active respondent pool, such that those who did not respond in the first survey remained eligible to respond to subsequent follow‐ups; Table S1 contains information regarding participation rates for all five surveys, and links to resources made available by UAS. To be eligible for the current study, participants must have participated in at least one survey and provided a current state of residence (including Washington, D.C.). To measure trends over time, we also required that only those for whom the date of survey response was available could be included in this study. The final eligible sample was N = 7,397 eligible unique participants with N = 31,244 eligible observations over 5 surveys. Dates were coded according to when respondents completed the survey: the first survey corresponds to March 10–March 31, and the fifth survey corresponds to May 13–June 8. Figure S1 shows the start and end dates, by survey, for the analytic sample.

Measures

Alcohol Consumption Frequency

Alcohol consumption frequency was assessed via the question, “Out of the past 7 days, what is your best estimate of the number of days that you did each of the following activities? Consumed alcohol.” This response was measured as a discrete count variable, with range 0 to 7.

State COVID‐19 Burden

The state‐level burden of COVID‐19 was operationalized as whether or not, on the interview date, a state had at or above the national median daily incidence of COVID‐19 cases. We chose to use this state‐level measure for two reasons. The first was pragmatic—states are the highest resolution for geographic references in the UAS. The second was theoretical—while smaller geographies (i.e., counties or municipalities) may have been more salient to individual disease risk, for policies regarding quarantines, mask‐wearing, mobility and accessibility of goods and services, states policies frequently preempted municipal policies (McDonald III et al., 2020; Treskon and Docter, 2020).

We chose to use disease burden as a state‐level indicator of risk, rather than COVID‐19 policies such as the date of state stay‐at‐home orders, because stay‐at‐home orders did not actually correspond closely to quarantine orders or alcohol policy changes; rather, various disease burden statistics were utilized as benchmarks for states in decision making around opening of businesses and schools long after stay‐at‐home orders had expired (Harris, 2020). For example, New York State’s state closing order expired on May 15, but phased re‐openings were conditional on meeting certain disease‐related metrics including testing and incidence rates (New York State Governor’s Office, 2020).

Time‐varying state‐level COVID‐19 data were made available by USA Facts which collected state‐ and county‐level COVID‐19 data directly from state and county public health websites. This publicly available database tracked cumulative incidence of COVID‐19 at the state level. Based on the changes in cumulative incidence, we derived daily incidence (i.e., N new cases each day) for use as a proxy for state‐level disease burden during the first wave of the pandemic. During the March 10 to June 8 study period, the daily national incidence ranged from 0 cases (e.g., for the majority of states at study onset) to 13,262 (New York State, April 3). The distribution was positively (right) skewed (Fig. S2), and we chose to use the median incidence over the study period (N = 147 cases) as a cutoff for dichotomizing states into those with relatively higher (i.e., median or above) or lower disease (i.e., below median) burden at each observation day. There is no established “best” cutoff for high vs. low state COVID‐19 burden, so to confirm that our findings were robust to the choice of cutoff (median), we show sensitivity analyses of our main models using two different criteria—cutoffs at the first quartile (N = 28) and third quartile (N = 458). While absolute case count drove state responses during the early days of the pandemic, we also considered relative (i.e., per capita) measures of disease burden in sensitivity analyses, dichotomizing states as high or low disease burden based on the median daily incidence per capita (N = 36 cases/1 million residents).

Covariates

We present main models without adjustment for individual‐level covariates. However, because individual covariates may be related both to differential social/economic experiences during the COVID‐19 pandemic and to frequency of alcohol consumption, we additionally show main models with control for these covariates to confirm that the variation in state‐level and marginal trends persisted above and beyond the effects on these individual‐level characteristics. Time‐invariant covariates were gender (Grucza et al., 2018), having children under the age of 18 at home (Wardell et al., 2020), race (categorized as White, Black, Asian, or other), and Hispanic ethnicity (Meyers et al., 2017). Time‐varying covariates were respondents’ age (Grucza et al., 2018), employment status (Collins, 2016; Frone, 2008), and a dichotomous measure of whether or not they reported avoiding public spaces to avoid disease risk at the time of each survey, as social distancing may be related to frequency of consumption (McPhee et al., 2020). We additionally include two mental health covariates, depression and anxiety, both of which are associated with alcohol consumption (Marmorstein, 2009; McHugh and Weiss, 2019; Smith and Randall, 2012). Both were measured using the Patient Health Questionnaire‐4 (PHQ‐4) (Kroenke et al., 2009), which consists of ascertaining the frequency of two anxiety items (feeling nervous, anxious, or on edge; being unable to stop or control worrying) and two depression items (little interest or pleasure in doing things; feeling down, depressed, hopeless). Responses are asked how frequently they have been bothered by any of the items over the past two weeks, with scores for each item ranging from 0 (“not at all”) to 3 (“nearly every day”). Respondents were coded as having anxiety or depression if they received sum scores at or above 3 for either set of items.

Finally, because individuals’ perceived presence of COVID‐19 among the network of people they know may influence both stress, coping, and frequency of alcohol consumption, we also included a measure of individuals’ perceived personal/familial COVID‐19 burden. At each survey, respondents were asked whether they had received a COVID‐19 diagnosis, if they thought they had COVID‐19, and how many people they knew who they thought might be infected. If they endorsed a positive diagnosis, or a positive belief that they have COVID‐19, or a nonzero value for the number of people in their life whom they believe to be infected, they were classified as having a high perceived personal/familial COVID‐19 burden; if otherwise, they were classified as low.

Statistical Analysis

We examined marginal trends in number of drinking days during the study period as well as state‐level variation in trends. We used survey‐weighted multilevel growth models with both a state‐level random intercept and random slope for time (to allow for state‐level variation in time trends) and a participant‐level random intercept to account for nonindependence across multiple observations. Because respondents occasionally moved states during the study duration, individuals were not fully nested in states and random effects for both individuals and states were modeled to allow cross‐classification. For all models, we used negative binomial regression to model number of drinking days as discrete count data. Time was measured as count of t days beginning on March 10, 2020, so the model estimates reflect the incident risk ratio (IRR) for count of drinking days given a 1‐day increase in calendar time. Effect modification by COVID‐19 presence was assessed using interaction between linear time and an indicator variable for whether or not the state had reported the median cumulative incidence of disease cases on that date. Time trends were modeled using model‐based marginal predicted count of drinking days from interaction models. Models included a uniform offset of 1. Global Moran’s I assessed spatial autocorrelation for the incidence risk ratio of US states and DC. Analyses were performed using SAS 9.4, and figures were produced using R version 4.0.

Among the 7,397 eligible participants with 31,244 observations, outcome information was missing for 43 (0.1%) of observations and covariate information was missing for 3,142 (9.9%) observations. Main models were analyzed using complete case analysis (N = 6,172 (83.4%) unique participants; N = 28,059 (89.8%) observations) due to computational barriers to imputing hierarchical, multilevel data. To explore the possibility of bias due to missingness, we examined distributions predictors of outcome and covariate missingness, and how those varied compared with the complete case sample.

Sensitivity Analyses

As described above, we performed sensitivity analyses with different cut‐points to dichotomize states with relatively high or low COVID‐19 burdens, including using relative (i.e., per capita) rather than absolute measures.

We performed two other sensitivity analyses. Eligible participants responded to at least one survey of the possible five. To confirm that observed trends were not a spurious function of those who responded only at earlier surveys having meaningfully different drinking patterns than those who responded at only later surveys, we performed a sensitivity analysis restricting the sample to only those who responded to all five surveys (N = 4,874 participants and N = 23,518 observations).

We also examined trends in alcohol consumption frequency in the spring of 2019. In 2019, a sample of the UAS active respondent pool received a single survey administration asking a comparable alcohol measure (“In the last three months, on average, how many days per week have you had any alcohol to drink? [For example, beer, wine, or any drink containing liquor.]”) to the one used in the current study. Among those queried, 914 respondents are also represented in our current eligible sample. The survey was administered over the period April 29 through July 1, 2019, with each participant responding once. Trends over time in this subsample in the spring of 2019 (from April 29 to June 8, corresponding to the time period in the current study) were examined to determine the trend in alcohol consumption frequency during a comparable time period one year before.

Results

Table 1 shows demographic characteristics and number of drinking days for respondents according to whether or not a state had a high COVID‐19 burden at the time of the survey. For observations during periods of higher state COVID‐19 burden, we observed higher average frequencies of alcohol consumption (drinking days = 1.57 vs. 1.43, p < 0.001), younger respondents, a higher proportion of respondents with children in the house, a lower proportion of White respondents, and higher proportions of respondents reporting avoiding public spaces, experiencing unemployment, and reporting high perceived personal/familial COVID‐19 burdens.

Table 1.

Demographic and Outcome Characteristics of Respondents, Stratified by Whether or not State had Median Cumulative Incidence of COVID‐19 Cases at the Time of the Survey, March 10, 2020–June 8, 2020

State had < median COVID‐19 cases

(N = 8,021 observations)

State had ≥ median COVID‐19 cases

(N = 20,038 observations)
Number of drinking days (mean, SD) 1.43 (2.15) 1.57 (2.20) p < 0.001
Male gender 3,373 (42.1%) 8,490 (42.4%) p = 0.63
Age (mean, SD) 52.2 (15.7) 50.8 (16.1) p < 0.001
Children live in the house 2,948 (36.8%) 7,913 (39.5%) p < 0.001
White race 6,595 (82.2%) 15,739 (78.6%) p < 0.001
Hispanic 738 (9.2%) 3,263 (16.3%) p < 0.001
Respondent avoids public spaces 5,215 (65.0%) 17,535 (87.5%) p < 0.001
Respondent has a job 4,572 (57.0%) 10,102 (50.4%) p < 0.001
Respondent has high personal/familial perceived COVID‐19 burden 749 (9.3%) 3,862 (19.3%) p < 0.001
Depression PHQ4 ≥ 3 791 (9.9%) 2,427 (12.1%) p < 0.001
Anxiety PHQ4 ≥ 3 1,159 (14.5%) 3,270 (16.3%) p < 0.001
Open in a new tab

Figure 1 shows the predicted count of drinking days from marginal (top) and interaction (bottom) models without adjustment for individual covariates. Overall, the number of drinking days increased during the study period; at the beginning of the study period, respondents on average reported 1.30 (95% CI: 1.26, 1.34) drinking days. By the end of the study period, respondents on average reported 1.65 (95% CI: 1.41, 1.92) drinking days.

Fig. 1.

Fig. 1

Open in a new tab

Predicted count of drinking days over time and state‐level COVID‐19 presence, March 10, 2020–June 8, 2020; marginal estimates and with interaction by disease burden.

We observed heterogeneity by state‐level COVID‐19 case burden. At the beginning of the observation period, respondents living in states with higher cumulative incidence of COVID‐19 cases had a predicted 1.51 (95% CI: 1.42, 1.59) drinking days per week, whereas those in states with lower COVID‐19 presence had a predicted 1.21 (95% CI: 1.17, 1.26) drinking days per week. By the end of the study period, respondents living in states with higher COVID‐19 presence had a predicted 1.46 (95% CI: 1.24, 1.71) drinking days per week, whereas those in states with lower COVID‐19 presence had a predicted 1.83 (95% CI: 1.54, 2.17) drinking days per week. Statistical tests for interaction confirmed that slopes over time evidenced differences according to COVID‐19 presence (F = 27.86, p < 0.01). Figure S3 shows the number of states with less than median COVID‐19 presence over the study duration, which varied over time. At the beginning of the study, all states reported less than the median cumulative incidence of COVID‐19 cases; by the end of the study, 19 states reported below the median.

Table 2 shows the model estimates for both the unadjusted (corresponding to Fig. 1) and adjusted models, including both parameter estimates (i.e., beta estimates and standard errors) and incident risk ratios (IRRs) for count of drinking days over time. Model 1 is the average, marginal change without interaction for COVID‐19 presence; on average, the number of drinking days increased between March 10 and June 8, with every day of calendar time associated with a.3% increase in number of drinking days (IRR: 1.003, 95% CI: 1.001, 1.004). Model 2 in Table 2 is the average, marginal change over time with interaction by COVID‐19 presence. We observed heterogeneity in trends in alcohol consumption frequency according to state‐level COVID‐19 presence. States with higher COVID‐19 burden had an increased risk of drinking days (IRR: 1.244, 95% CI: 1.163, 1.3330) at baseline (intercept) compared to those with fewer than the median cases. However, though states with a lower caseload evidenced fewer drinking days at baseline, the trends over time increased relative to states with a higher caseload. The average number of drinking days in states with lower caseloads increased by 0.5% (IRR: 1.005, 95% CI: 1.003, 1.007) each day, whereas drinking days in states with higher caseloads remained stable during this time period (IRR: 1.000, 95% CI: 0.998, 1.002). Models 3 and 4 are adjusted for individual‐level covariates, and the estimates and interpretation are not meaningfully different than those without covariate adjustment.

Table 2.

Model Parameters and Incident Risk Ratios of Alcohol Consumption Over Time Among Adults in the UAS, March 10, 2020–June 8, 2020; N = 6,712 Unique Participants and N = 28,059 Observations

Predictor Estimate

Model 1

Time trend only, unadjusted

Model 2

Time trend with interaction, unadjusted

Model 3

Time trend only, adjusted a

Model 4

Time trend with interaction, adjusted a

Time

Β (SE)

p‐value

0.0027 (0.0009)

p < 0.01

0.0046 (0.0010)

p = 0.01

0.0021 (0.0009)

p = 0.02

0.0039 (0.0010)

p = 0.03
IRR (95% CI) 1.003 (1.001, 1.004) 1.005 (1.003, 1.007) 1.002 (1.000, 1.004) 1.004 (1.002, 1.006)
High state‐level COVID‐19 presence

Β (SE)

p‐value

0.2180 (0.0340)

p < 0.01

0.2171 (0.0354)

p < 0.01
IRR (95% CI) 1.244 (1.163, 1.330) 1.242 (1.159, 1.332)
Time × high COVID‐19 presence

Β (SE)

p‐value

−0.0050 (0.0009)

p < 0.01

−0.0047 (0.0010)

p < 0.01
IRR (95% CI) 0.995 (0.993, 0.997) 0.995 (0.993, 0.997)
IRR (95% CI), states with low COVID‐19 burden 1.005 (1.003, 1.007) 1.004 (1.002, 1.006)
IRR (95% CI) states with high COVID‐19 burden 1.000 (0.998, 1.002) 0.999 (0.997, 1.001)
Open in a new tab
a

Adjusted for gender, age, race, Hispanic ethnicity, presence of children in the house, current employment status, whether or not respondent currently avoids public spaces to reduce disease risk, anxiety, depression, and perceived personal/familial COVID‐19 burden.

In sensitivity analyses, we examined whether this heterogeneity varied according to the threshold at which we dichotomized states into high or low COVID‐19 presence. Trends using the first quartile (daily incidence = 28 cases) and third quartile (daily incidence = 458 cases) did not meaningfully impact the estimates (Table S2). We additionally examined whether this heterogeneity was sensitive to relative rates, by using median daily incidence per capita as the threshold for dichotomizing states. Trends using incidence per capita did not evidence different results or interpretation than those using absolute measures (Table S3). Among those who responded to all five surveys (N = 4,874 participants and N = 23,518 observations), trends over time were indistinguishable from the full eligible complete case sample (Table S4).

We examined trends in alcohol consumption frequency among the subsample of respondents who were also queried in 2019 during a similar time period (i.e., from April 29, 2019, to June 8, 2019). We show the unadjusted marginal trends for 2019 in Fig. S4. Trends in alcohol consumption in 2019 during this time period evidenced average increases by day (IRR: 1.025, 95% CI 1.010, 1.040). This incidence rate ratio was higher than the daily incident rate ratio for those same respondents in 2020, suggesting alcohol consumption frequency during the same time period 2020 increased at a slower rate than in the previous year.

Figure 2 shows, for each state, the slope of the increase in drinking days during the study period (from March 10, 2020, through June 8, 2020), with adjustment for individual‐level covariates. The color of the line indicates whether the state had greater than or equal to the median number of COVID‐19 cases at any point during the study period. These estimates correspond to state random effects from model 4. Table S5 shows the adjusted state‐level estimates that correspond to Fig. 2, as well as the range of respondents per survey and total number of observations per state. Figure S5 shows a map with the geographic distribution of IRRs by state, corresponding to both Fig. 2 and Table S5. The majority of states (N = 35, including D.C.) did not evidence changes in number of drinking days throughout the study period. Eight states (Kentucky [IRR = 0.991], Arkansas [IRR = 0.992], Iowa [IRR = 0.992], West Virginia [IRR = 0.992], New Hampshire [IRR = 0.994, Indiana [IRR = 0.995], Virginia [IRR = 0.996], and Tennessee [IRR = 0.997] evidenced significant decreases in trends during this time period; all of these states experienced at or above the median cumulative incidence of COVID‐19 during the study. Eight states (Alaska [IRR = 1.014], Maine [IRR = 1.011], Nevada [IRR = 1.009], Missouri [IRR = 1.006], Colorado [IRR = 1.007], Oregon [IRR = 1.006], Washington [IRR = 1.004], and Georgia [IRR = 1.003]) evidenced increases in trends during this time period. Among these, four (West Virginia, Alaska, Maine, and Oregon) experienced less than the median cumulative incidence of COVID‐19 cases, while the remaining twelve experienced at or above the median. Incidence risk ratio was not spatially autocorrelated (I = 0.03). To confirm that there was no further underlying geographic patterning, we tested a final model with interaction between time and US region based on US census criteria (US Department of Commerce, n.d.); tests of interaction between time and US region showed no meaningful heterogeneity (F = 0.81, p = 0.59).

Fig. 2.

Fig. 2

Open in a new tab

Incident rate ratio for the count of drinking days as a function of time, from March 10, 2020 to June 8, 2020, by US state, ranked by magnitude of slope.

Finally, regarding missing data, we examined distributions of covariates comparing complete cases to those with missing outcome or missing control variables. Outcome missingness was unrelated state COVID‐19 burden, though it was associated with age and race (Table S6). The covariate distribution of the complete case sample was different than that of the eligible sample (i.e., those excluded were on average younger, less likely to be White, more likely to have children in the house, more likely to be depressed or anxious), though inclusion of covariates in the main models did not change results, suggesting that bias from missingness was not a contributor to the overall observed effects.

Discussion

In this study, we analyzed trends in alcohol consumption frequency during the first wave of the COVID‐19 pandemic in the United States. The number of drinking days increased between March 15 and June 8, 2020, with a model‐based 0.3% increased risk of increase in drinking days each day. We observed heterogeneity in this trend by both COVID‐19 burden. Namely, while individuals living in states with higher COVID‐19 burdens reported a higher average number of drinking days at the beginning of the epidemic, there was no evidence of further increases as the epidemic progressed; however, respondents living in states with lower COVID‐19 burdens increased the number of drinking days throughout the first wave of the pandemic. These findings were robust to multiple specifications of COVID‐19 burden, whether it was measured in terms of absolute or per capita incidence.

Regarding state‐level heterogeneity, we observed that these trends differed by state, with residents in eight states evidencing decreases in the number of drinking days and residents in eight states evidencing increases. Among the states with increases, three (Alaska, Maine, Oregon) experienced consistently lower COVID‐19 presence than the national median cumulative incidence. The majority of US states disallowed on‐premise alcohol consumption at some point during the pandemic, though they varied in when they re‐opened drinking establishments (Alcohol Policy Information System, 2020). Alaska re‐opened bars and restaurants with restrictions in April 2020, Oregon re‐opened bars and restaurants in May 2020, and Maine re‐opened bars and restaurants in June; all three allowed curb‐side and take‐out alcohol purchases during the pandemic. However, overall, the majority of states did not evidence meaningful changes in trends in drinking days. The limited heterogeneity in trends suggests that the observed trends were largely invariant to other state‐level features, and we additionally observed no apparent variation in trends according to geographic region.

Though literature regarding alcohol consumption during this time period suggests that it is overall increasing, our findings suggest that the general increases mask meaningful heterogeneity according to state disease burden. The increases in alcohol consumption were exclusively among those living in states with a relatively low disease burden, whereas those living in states with a relatively high burden did not increase alcohol consumption frequency. In fact, compared with trends in alcohol consumption frequency in 2019, trends in 2020 were much less pronounced, suggesting that residents of states with higher COVID‐19 burden consumed alcohol less frequently than in the same season in the previous year. Limitations of the 2019 data are that respondents were queried only once throughout the spring; therefore, these analyses are not strictly comparable. They do, however, suggest that trends in alcohol consumption frequency during the first wave of the pandemic are aberrant from what might be expected in an otherwise normal year during this season.

The mechanisms determining these trends remain unknown. Respondents living in states with higher disease burden could be less willing to leave their homes to purchase alcohol because of fear of disease risk, or during this time period these respondents may have experienced more supply‐chain disruptions during COVID‐19 (Hobbs, 2020). Enforcement of policies, including stay‐at‐home orders, has not been uniform; states with a higher COVID‐19 burden may have both implemented more stringent policies and enforced them more strictly, which may have contributed to variation in alcohol consumption (Bergquist et al., 2020; Cain, 2020; Capano et al., 2020; Jennings and Perez, 2020). Alternatively, respondents living in states experiencing a high COVID‐19 burden early in the pandemic may have increased drinking days before the survey onset, and then remained at the increased levels as respondents in states with relatively lower burdens eventually “caught up” to these higher rates. As the pandemic progresses, further inquiry into the patterning of alcohol behaviors across axes of state‐level differences may shed light on the contextual drivers of these health behaviors.

While we had hypothesized that respondents in states with higher COVID‐19 burden would evidence higher rates of increase in alcohol consumption, due primarily to pathways related to stress and coping, we did not find support for this hypothesis in our findings. The role of stress and coping in relation to state‐level features like policies and state‐level pandemic response is unknown, and these mechanisms may still play a large role in the increases among states with relatively lower COVID‐19 burdens. However, these findings point to the importance of state‐level, rather than individual‐level, mechanisms that facilitate increases in alcohol consumption.

State‐level variation in alcohol policy impacts when, where, and under what circumstances individuals can purchase and consume alcohol (Naimi et al., 2014). These include permissive policies like drink special laws as well as alcohol taxes and state‐mandated opening or closing times. These contextual features are central drivers of alcohol consumption and alcohol‐related harms, including heavy drinking, impaired driving, and injury (Naimi et al., 2014, 2018; Nelson and McNall, 2016; Xuan et al., 2015). Aside from being a mediator of alcohol policy’s impact on consumption, alcohol availability—including alcohol outlet density—is positively associated with increases in alcohol consumption (Campbell et al., 2009). Limiting outlet density is regarded as among the most effective policies for reducing adult and youth binge drinking (Campbell et al., 2009; Nelson et al., 2013). New COVID‐19 policy changes that increase alcohol availability—including legalizing take‐out alcoholic beverages and alcohol delivery—are likely contributors to the increased alcohol consumption frequency evidenced in this study. While there was limited evidence of geographic patterning of drinking days during the COVID‐19 pandemic to link state‐level variation to clear geographic or policy differences, we expect that as the pandemic progresses we may see long‐term impacts in the states with the most permissive policy changes.

In addition to variation in state‐level alcohol climate, major economic downturns—for example, the Great Recession of 2008—and mass traumatic events, like natural disasters, are sources of stress and despair that can increase alcohol consumption and exacerbate problematic consumption patterns (Bor et al., 2013; Cerdá et al., 2011). However, certain consumption patterns may be more sensitive to these events than others; for example, in the case of the Great Recession, average alcohol consumption declined (presumably due to decreased flexible resources to spend on alcohol use) but the frequency of binge episodes—that is, drinking multiple alcoholic beverages in a short period of time—increased (Bor et al., 2013). Problematic alcohol consumption after mass traumas can also develop in the years after the trauma, rather during the event or in the immediate aftermath (North et al., 2011); therefore, it is vital to continue to closely monitor these trends as the COVID‐19 pandemic continues and ultimately resolves.

Among the limitations in our study were limited alcohol measures. Our primary outcome—the number of drinking days in the past week—is an incomplete measure of overall consumption, as the volume consumed on drinking days is critical information. While more quantity‐based measures later became available in the UAS, these were not introduced until later surveys that corresponded to the peak of the pandemic in many states. However, the number of drinking days has been shown to be highly correlated with quantity‐based measures of alcohol consumption like volume of alcohol consumed and average quantity consumed in a typical drinking day (Leigh, 2000); further, frequency of alcohol consumption itself is an important determinant of health which has been implicated in mortality risk, even among those who consume only one or two beverages per drinking occasion (Hartz et al., 2017, 2018).

We use complete case analysis to examine these trends, and approximately 10% of our sample were missing covariate information. The patterning of missing suggests that those missing data had different distributions of individual covariates, which may have impacted nonresponse. However, given that outcome missingness was unrelated to COVID‐19 burden and that main analyses were invariant to covariate control, we feel reassured that selective missing was unlikely to produce biases in our estimated trends and selection was not a major contributor to our findings and interpretation.

Finally, we did not have information about preexisting alcohol consumption disorders among this sample. Drinking after disasters predominantly occurs among those who already have preexisting alcohol conditions, and these disasters exacerbate their symptoms (North et al., 2011). However, state‐level differences in underlying alcohol consumption disorders are unlikely to contribute to the observed trends in the study, as they are unlikely to be related to state‐level COVID‐19 burden.

Conclusion

The COVID‐19 pandemic in the United States resulted in mass disruption, including via public health measures to curb the spread of the virus. We find that during the first wave of the pandemic in the United States, the average number of drinking days increased, and these increases were primarily among individuals living in states with relatively lower COVID‐19 burden. Health behaviors, including alcohol consumption, are sensitive to contextual and psychosocial influences, and this study underscores the necessity of screening not just for COVID‐19 symptoms, but for risky health behaviors as well. Alcohol consumption is a normalized health behavior, but nevertheless is a major contributor to morbidity and mortality, and care providers and policy‐makers ought to continue to remain mindful that focusing resources on the current epidemic must not come at the expense of screening and management of other health risks.

Funding

This work was supported by R01‐AA026861 from National Institute on Alcohol Abuse and Alcoholism (NIAAA, https://www.niaaa.nih.gov/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The collection of UAS COVID‐19 tracking data is supported in part by the Bill & Melinda Gates Foundation and by grant U01AG054580 from the National Institute of Aging.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Fig. S1 UAS surveys and dates submitted among study sample, March 10 to June 8, 2020.

Fig. S2 Histogram of cumulative incidence of state COVID‐19 cases between March 10 and June 8, 2020.

Fig. S3 Number of states with less than the median cumulative incidence of COVID‐19 cases, over time, from March 10 to June 8, 2020.

Fig. S4 Predicted count of drinking days over time among 2019 UAS respondents, April 29, 2019–June 8, 2019; marginal estimates.

Fig. S5 Geographic distribution of incidence risk ratio by US states and DC (n = 51).

Table S1 UAS COVID‐19 surveys and responses rates.

Table S2 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020 using alternative cut‐points for dichotomizing high and low COVID‐19 burden, N = 28,059 observations.

Table S3 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020 using per capita incidence measures for high and low COVID‐19 burden, N = 28,059 observations.

Table S4 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020, only among respondents who participated in all 5 waves, N = 4,874 participants and N = 23,518 observations.

Table S5 Incident risk ratios of alcohol consumption over time among adults in the UAS, March 10, 2020–June 8, 2020, by state, adjusted for individual covariates.

Table S6 Predictors of missing outcome data and predictors of any item missingness in eligible sample.

Click here for additional data file. (1MB, docx)

Acknowledgements

All authors contributed equally to this work. The project described in this paper relies on data from survey(s) administered by the UnderStanding America Study, which is maintained by the Center for Economic and Social Research (CESR) at the University of Southern California. The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of USC or UAS.

References

  1. Alattar L, Messel M, Rogofsky D (2018) An introduction to the understanding America study internet panel. Soc Sec Bull 78:13. [Google Scholar]
  2. Alcohol Policy Information System (2020) Emergency for On‐Premise and Off‐Premise Establishments as of August 15, 2020. [Google Scholar]
  3. Alon T, Doepke M, Olmstead‐Rumsey J, Tertilt M (2020) The Impact of COVID‐19 on Gender Equality. National Bureau of Economic Research, Cambridge MA. [Google Scholar]
  4. Bagnardi V, Rota M, Botteri E, Tramacere I, Islami F, Fedirko V, Scotti L, Jenab M, Turati F, Pasquali E (2015) Alcohol consumption and site‐specific cancer risk: a comprehensive dose–response meta‐analysis. Br J Cancer 112:580–593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bartik AW, Bertrand M, Cullen ZB, Glaeser EL, Luca M, Stanton CT (2020) How are small businesses adjusting to covid‐19? early evidence from a survey. National Bureau of Economic Research. [Google Scholar]
  6. Bayham J, Fenichel EP (2020) Impact of school closures for COVID‐19 on the US health‐care workforce and net mortality: a modelling study. Lancet Public Heal 5:e271–e278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bergquist S, Otten T, Sarich N (2020) COVID‐19 pandemic in the United States. Heal Policy Technol 9:623–638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bor J, Basu S, Coutts A, McKee M, Stuckler D (2013) Alcohol use during the great recession of 2008–2009. Alcohol Alcohol 48:343–348. [DOI] [PubMed] [Google Scholar]
  9. Braillon A, Wilson M (2018) Does moderate alcohol consumption really have health benefits? BMJ 362:k3888. [DOI] [PubMed] [Google Scholar]
  10. Brenmer JUS (2020) Alcohol Sales Increase 55 Percent in One Week Amid Coronavirus Pandemic.
  11. Brown RL, Richman JA, Rospenda KM (2014) Economic stressors and alcohol‐related outcomes: exploring gender differences in the mediating role of somatic complaints. J Addict Dis 33:303–313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cain D (2020) COVID‐19 and the Constitution: State Police Powers and Judicial Scrutiny. The University of Missouri. [Google Scholar]
  13. Campbell CA, Hahn RA, Elder R, Brewer R, Chattopadhyay S, Fielding J, Naimi TS, Toomey T, Lawrence B, Middleton JC (2009) The effectiveness of limiting alcohol outlet density as a means of reducing excessive alcohol consumption and alcohol‐related harms. Am J Prev Med 37:556–569. [DOI] [PubMed] [Google Scholar]
  14. Capano G, Howlett M, Jarvis DSL, Ramesh M, Goyal N (2020) Mobilizing policy (In)capacity to fight COVID‐19: understanding variations in state responses. Policy Soc 39:285–308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cerdá M, Tracy M, Galea S (2011) A prospective population based study of changes in alcohol use and binge drinking after a mass traumatic event. Drug Alcohol Depend 115:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chodkiewicz J, Talarowska M, Miniszewska J, Nawrocka N, Bilinski P (2020) Alcohol consumption reported during the COVID‐19 pandemic: the initial stage. Int J Environ Res Public Health 17:4677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Clay JM, Parker MO (2020) Alcohol use and misuse during the COVID‐19 pandemic: a potential public health crisis? Lancet Public Heal 5:e259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Collins SE (2016) Associations between socioeconomic factors and alcohol outcomes. Alcohol Res 38:83–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Da BL, Im GY, Schiano TD (2020) COVID‐19 hangover: a rising tide of alcohol use disorder and alcohol‐associated liver disease. Hepatology 72:1102–1108. [DOI] [PubMed] [Google Scholar]
  20. Diggs ON, Neppl TK (2018) The influence of economic pressure on emerging adult binge drinking: testing the family stress model over time. J Youth Adolesc 47:2481–2495. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Elgendy R, Deschênes SS, Burns RJ, Levy M, Schmitz N (2019) Alcohol consumption, depressive symptoms, and the incidence of diabetes‐related complications. J Diabetes 11:14–22. [DOI] [PubMed] [Google Scholar]
  22. Fergusson DM, Horwood LJ, Boden JM, Mulder RT (2014) Impact of a major disaster on the mental health of a well‐studied cohort. JAMA Psychiatry 71:1025–1031. [DOI] [PubMed] [Google Scholar]
  23. French MT, Mortensen K, Timming AR (2020) Changes in self‐reported health, alcohol consumption, and sleep quality during the COVID‐19 pandemic in the United States. Appl Econ Lett 1–7. 10.1080/13504851.2020.1861197 [DOI] [Google Scholar]
  24. Frone MR (2008) Employee alcohol and illicit drug use: Scope, causes, and organizational consequences. Handb Organ Behav 1:519–540. [Google Scholar]
  25. Gallagher C, Hendriks JML, Elliott AD, Wong CX, Rangnekar G, Middeldorp ME, Mahajan R, Lau DH, Sanders P (2017) Alcohol and incident atrial fibrillation – a systematic review and meta‐analysis. Int J Cardiol 246:46–52. [DOI] [PubMed] [Google Scholar]
  26. GBD 2016 Alcohol Collaborators (2018) Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 392:1015–1035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Goldhill O (2020) Covid‐19 vaccine‐distribution timeline will keep slipping, experts say ‐ STAT. Stat.
  28. Grucza RA, Sher KJ, Kerr WC, Krauss MJ, Lui CK, McDowell YE, Hartz S, Virdi G, Bierut LJ (2018) Trends in adult alcohol use and binge drinking in the early 21st‐century United States: a meta‐analysis of 6 national survey series. Alcohol Clin Exp Res 42:1939–1950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Harris JE (2020) Reopening under COVID‐19: What to Watch For. National Bureau of Economic Research, Cambridge, MA. [Google Scholar]
  30. Hartz SM, Oehlert M, Horton AC, Grucza RA, Fisher SL, Culverhouse RC, Nelson KG, Sumerall SW, Neal PC, Regnier P (2018) Daily drinking is associated with increased mortality. Alcohol Clin Exp Res 42:2246–2255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hartz SM, Oehlert M, Horton AC, Grucza R, Fisher SL, Nelson KG, Sumerall SW, Neal PC, Regnier P, Chen G (2017) Components of alcohol use and all‐cause mortality. bioRxiv 129270.
  32. Hobbs JE (2020) Food supply chains during the COVID‐19 pandemic. Can J Agric Econ Can d’agroeconomie 68:171–176. [Google Scholar]
  33. House A (2020) Online liquor sales up 243% in the US. Drink Trade. Available at: https://www.drinkstrade.com.au/online‐liquor‐sales‐up‐243‐in‐the‐us. Accessed September 20, 2020.
  34. Hussong AM, Jones DJ, Stein GL, Baucom DH, Boeding S (2011) An internalizing pathway to alcohol use and disorder. Psychol Addict Behav 25:390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Jennings WG, Perez NM (2020) The immediate impact of COVID‐19 on law enforcement in the United States. Am J Crim Justice 45:690–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kane PB, Moyer H, MacPherson A, Papenburg J, Ward BJ, Broomell SB, Kimmelman J (2020) Expert forecasts of COVID‐19 vaccine development timelines. J Gen Intern Med 35:3753–3755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Keyes KM, Hatzenbuehler ML, Hasin DS (2011) Stressful life experiences, alcohol consumption, and alcohol use disorders: the epidemiologic evidence for four main types of stressors. Psychopharmacology 218:1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Kroenke K, Spitzer RL, Williams JBW, Löwe B (2009) An ultra‐brief screening scale for anxiety and depression: the PHQ–4. Psychosomatics 50:613–621. [DOI] [PubMed] [Google Scholar]
  39. Kuypers KPC, Legrand S‐A, Ramaekers JG, Verstraete AG (2012) A case‐control study estimating accident risk for alcohol, medicines and illegal drugs. PLoS One 7:e43496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lechner WV, Laurene KR, Patel S, Anderson M, Grega C, Kenne DR (2020) Changes in alcohol use as a function of psychological distress and social support following COVID‐19 related University closings. Addict Behav 110:106527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Leigh BC (2000) Using daily reports to measure drinking and drinking patterns. J Subst Abuse 12:51–65. [DOI] [PubMed] [Google Scholar]
  42. Loxton M, Truskett R, Scarf B, Sindone L, Baldry G, Zhao Y (2020) Consumer behaviour during crises: preliminary research on how coronavirus has manifested consumer panic buying, herd mentality, changing discretionary spending and the role of the media in influencing behaviour. J Risk Financ Manag 13:166. [Google Scholar]
  43. Marmorstein NR (2009) Longitudinal associations between alcohol problems and depressive symptoms: early adolescence through early adulthood. Alcohol Clin Exp Res 33:49–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. McDonald BD III, Goodman C, Hatch ME (2020) Tensions in state‐local intergovernmental response to emergencies: the case of COVID‐19. State Loc Gov Rev. 10.1177/0160323X20979826 [DOI] [Google Scholar]
  45. McHugh RK, Weiss RD (2019) Alcohol use disorder and depressive disorders. Alcohol Res 40:arcr.v40.1.01. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. McPhee M, Keough MT, Rundle S, Heath LM, Wardell J, Hendershot CS (2020) Depression, environmental reward, coping motives and alcohol consumption during the COVID‐19 pandemic. Front Psychiatry 11:574676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Mellinger JL (2019) Epidemiology of alcohol use and alcoholic liver disease. Clin liver Dis 13:136–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Meyers JL, Brown Q, Grant BF, Hasin D (2017) Religiosity, race/ethnicity, and alcohol use behaviors in the United States. Psychol Med 47:103–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Naimi TS, Blanchette J, Nelson TF, Nguyen T, Oussayef N, Heeren TC, Gruenewald P, Mosher J, Xuan Z (2014) A new scale of the U.S. alcohol policy environment and its relationship to binge drinking. Am J Prev Med 46:10–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Naimi TS, Xuan Z, Sarda V, Hadland SE, Lira MC, Swahn MH, Voas RB, Heeren TC (2018) Association of state alcohol policies with alcohol‐related motor vehicle crash fatalities among US adults. JAMA Intern Med 178:894–901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Nelson JP, McNall AD (2016) Alcohol prices, taxes, and alcohol‐related harms: a critical review of natural experiments in alcohol policy for nine countries. Health Policy (New York) 120:264–272. [DOI] [PubMed] [Google Scholar]
  52. Nelson TF, Xuan Z, Babor TF, Brewer RD, Chaloupka FJ, Gruenewald PJ, Holder H, Klitzner M, Mosher JF, Ramirez RL, Reynolds R, Toomey TL, Churchill V, Naimi TS (2013) Efficacy and the strength of evidence of U.S. alcohol control policies. Am J Prev Med 45:19–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. New York State Governor’s Office (2020) NY Forward | New York Forward. Available at: https://forward.ny.gov/ Accessed October 14, 2020.
  54. North CS, Ringwalt CL, Downs D, Derzon J, Galvin D (2011) Postdisaster course of alcohol use disorders in systematically studied survivors of 10 disasters. Arch Gen Psychiatry 68:173–180. [DOI] [PubMed] [Google Scholar]
  55. Pollard MS, Tucker JS, Green HD (2020) Changes in adult alcohol use and consequences during the COVID‐19 pandemic in the US. JAMA Netw Open 3:e2022942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Rehm J, Imtiaz S (2016) A narrative review of alcohol consumption as a risk factor for global burden of disease. Subst Abuse Treat Prev Policy 11:37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Rieper M (2020) MultiState | States Relax Alcohol Rules During COVID‐19 Pandemic. Multistate. Available at: https://www.multistate.us/insider/2020/4/28/states‐relax‐alcohol‐rules‐during‐covid‐19‐pandemic. Accessed September 25, 2020.
  58. Rodriguez LM, Litt DM, Stewart SH (2020) Drinking to cope with the pandemic: the unique associations of COVID‐19‐related perceived threat and psychological distress to drinking behaviors in American men and women. Addict Behav 110:106532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Rospenda KM, Minich LM, Milner LA, Richman JA (2010) Caregiver burden and alcohol use in a community sample. J Addict Dis 29:314–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Shanthakumar SG, Seetharam A, Ramesh A (2020) Understanding the socio‐economic disruption in the United States during COVID‐19’s early days. arXiv Prepr arXiv200405451.
  61. Sillaber I, Henniger MSH (2004) Stress and alcohol drinking. Ann Med 36:596–605. [DOI] [PubMed] [Google Scholar]
  62. Skog O (2001) Alcohol consumption and overall accident mortality in 14 European countries. Addiction 96:35–47. [DOI] [PubMed] [Google Scholar]
  63. Smith JP, Randall CL (2012) Anxiety and alcohol use disorders: comorbidity and treatment considerations. Alcohol Res 34:414–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Stahre M, Roeber J, Kanny D, Brewer RD, Zhang X (2014) Contribution of excessive alcohol consumption to deaths and years of potential life lost in the United States. Prev Chronic Dis 11:E109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Thompson RJ, Mata J, Jaeggi SM, Buschkuehl M, Jonides J, Gotlib IH (2010) Maladaptive coping, adaptive coping, and depressive symptoms: variations across age and depressive state. Behav Res Ther 48:459–466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Tramacere I, Negri E, Pelucchi C, Bagnardi V, Rota M, Scotti L, Islami F, Corrao G, La Vecchia C, Boffetta P (2011) A meta‐analysis on alcohol drinking and gastric cancer risk. Ann Oncol 23:28–36. [DOI] [PubMed] [Google Scholar]
  67. Tramacere I, Scotti L, Jenab M, Bagnardi V, Bellocco R, Rota M, Corrao G, Bravi F, Boffetta P, La Vecchia C (2010) Alcohol drinking and pancreatic cancer risk: a meta‐analysis of the dose‐risk relation. Int J cancer 126:1474–1486. [DOI] [PubMed] [Google Scholar]
  68. Treskon M, Docter B (2020) Preemption and its impact on policy responses to COVID‐19.
  69. Tull MT, Edmonds KA, Scamaldo KM, Richmond JR, Rose JP, Gratz KL (2020) Psychological outcomes associated with stay‐at‐home orders and the perceived impact of COVID‐19 on daily life. Psychiatry Res 289:113098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. US Department of Commerce (n.d.) Census Regions and Divisions of the United States.
  71. Wardell JD, Kempe T, Rapinda KK, Single A, Bilevicius E, Frohlich JR, Hendershot CS, Keough MT (2020) Drinking to cope during COVID‐19 pandemic: the role of external and internal factors in coping motive pathways to alcohol use, solitary drinking, and alcohol problems. Alcohol Clin Exp Res 44:2073–2083. [DOI] [PubMed] [Google Scholar]
  72. Xuan Z, Chaloupka FJ, Blanchette JG, Nguyen TH, Heeren TC, Nelson TF, Naimi TS (2015) The relationship between alcohol taxes and binge drinking: evaluating new tax measures incorporating multiple tax and beverage types. Addiction 110:441–450. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Fig. S1 UAS surveys and dates submitted among study sample, March 10 to June 8, 2020.

Fig. S2 Histogram of cumulative incidence of state COVID‐19 cases between March 10 and June 8, 2020.

Fig. S3 Number of states with less than the median cumulative incidence of COVID‐19 cases, over time, from March 10 to June 8, 2020.

Fig. S4 Predicted count of drinking days over time among 2019 UAS respondents, April 29, 2019–June 8, 2019; marginal estimates.

Fig. S5 Geographic distribution of incidence risk ratio by US states and DC (n = 51).

Table S1 UAS COVID‐19 surveys and responses rates.

Table S2 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020 using alternative cut‐points for dichotomizing high and low COVID‐19 burden, N = 28,059 observations.

Table S3 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020 using per capita incidence measures for high and low COVID‐19 burden, N = 28,059 observations.

Table S4 Incident risk ratios for alcohol consumption over time among adults in the UAS, March 10 2020–June 8, 2020, only among respondents who participated in all 5 waves, N = 4,874 participants and N = 23,518 observations.

Table S5 Incident risk ratios of alcohol consumption over time among adults in the UAS, March 10, 2020–June 8, 2020, by state, adjusted for individual covariates.

Table S6 Predictors of missing outcome data and predictors of any item missingness in eligible sample.

Click here for additional data file. (1MB, docx)

Articles from Alcoholism, Clinical and Experimental Research are provided here courtesy of Wiley

RESOURCES