Abstract
Aims
This study examined differential changes in alcohol use during the COVID-19 pandemic among adults with unhealthy alcohol use.
Methods
Among 62 924 adults identified with unhealthy alcohol use in primary care prepandemic (1 January 2019 to 29 February 2020), changes in alcohol use during the pandemic (1 March 2020 to 30 June 2022) were examined using electronic health record data from Kaiser Permanente Northern California. Outcomes were changes in heavy drinking days in the past three months (HDDs) and overall consumption (drinks/week), including continuous and categorical measures. Differences in outcomes by sex, age, race/ethnicity, and alcohol use disorder (AUD) were examined.
Results
On average, drinking was reduced by 3.0 HDDs (in the past three months) (SD = 18.4) and 4.1 drinks/week (SD = 12.2), but women, certain age groups, White patients, and patients without AUD had smaller decreases than their counterparts. Overall, 9.1% increased, 34.4% maintained, and 56.5% decreased HDDs, and 20.2% increased, 19.8% maintained, and 60.1% decreased drinks/week. Women, patients aged ≥35 years, White patients, and patients with AUD had higher odds of increasing versus decreasing HDDs, and maintaining versus decreasing, compared to their counterparts. Patients aged 18–20 years, White patients, and patients without AUD had higher odds than their counterparts of increasing versus decreasing drinks/week. Women, patients aged 18–20 years, Asian/Pacific Islander, and Latino/Hispanic patients had higher odds of maintaining versus decreasing drinks/week.
Conclusions
While alcohol use decreased overall among this sample of primary care patients with unhealthy drinking prepandemic, certain subgroups were more likely to increase drinking, suggesting a greater risk of alcohol-related problems.
Keywords: alcohol screening, unhealthy alcohol use, changes in alcohol use, COVID-19 pandemic, primary care
Introduction
The COVID-19 pandemic has contributed to increases in stress, negative affect related to social isolation, and the prevalence of depression and anxiety disorders (Park et al. 2020; COVID-19 Mental Disorders Collaborators. 2021). At the same time, other societal stressors occurred (e.g. high-profile cases of ongoing racial injustice, economic inflation), potentially contributing to additional stress (American Psychological Association 2022; Wu et al. 2023). Stress, anxiety, and negative affect are risk factors for increased alcohol consumption (Gonçalves et al. 2020; Guinle and Sinha 2020; Palzes et al. 2020a).
Several systematic reviews suggest that there has been substantial heterogeneity in changes in alcohol consumption during the pandemic (Roberts et al. 2021; Schmidt et al. 2021; Sohi et al. 2022). Studies in the USA have generally found that women increased alcohol use across a variety of measures (Boschuetz et al. 2020; Pollard et al. 2020; Barbosa et al. 2021; Sharma et al. 2023). Changes by age, race, and ethnicity have been less consistent; however, some studies have seen increased consumption among younger, and Black and Hispanic, adults (Czeisler et al. 2020; Barbosa et al. 2021; Metz et al. 2022). Changes among individuals who already drank over the recommended guidelines prepandemic, including those with alcohol use disorder (AUD), have rarely been reported, though one study among young adults showed significant reductions in drinking during the pandemic (Creswell et al. 2024). Another limitation is that most studies have retrospectively assessed alcohol use before and during the pandemic through self-report interviews, which could be prone to recall bias. Additionally, most studies have focused on the pandemic’s early stage when disease mitigation strategies such as stay-at-home orders were common. Prospective, longitudinal studies with longer observation periods are needed to clarify how alcohol consumption changed during the pandemic.
This study aimed to examine changes in alcohol use from before to during the pandemic among adults identified with unhealthy alcohol use at prepandemic primary care visits, by age group, sex, race, ethnicity, and AUD diagnosis. We hypothesized that younger adults, women, Black and Latino/Hispanic patients (compared to White patients), and those with AUD diagnoses would have greater increases in alcohol use. This study focused on individuals with prepandemic unhealthy alcohol use, as they may have been particularly vulnerable to escalated alcohol use during the pandemic. Findings may help in identifying subgroups at greatest risk of increased alcohol use and related harms and understanding the potential demands on providers and healthcare systems, which could inform practice in future crises.
Materials and methods
Setting
Kaiser Permanente Northern California (KPNC) is a large, integrated healthcare system that serves over 4.6 million members, about one-third of the population in the region. The membership is diverse and reflects the insured population in the USA (Davis et al. 2023), with coverage provided through employer-based plans, Medicare, Medicaid, and health insurance exchanges. The healthcare system began systematically screening for alcohol use in adult primary care in June 2013 and has screened ~81% of all adult members. During screening, medical assistants ask patients an adaptation of the evidence-based single-item screening question in the National Institute on Alcohol Abuse and Alcoholism (NIAAA) Clinician’s Guide (Willenbring et al. 2009) (‘How many times in the past three months have you had 5 or more drinks containing alcohol in a day?’ for men aged 18–65 years, or ‘4 or more drinks’ for men ≥66 years and women ≥18 years), and two questions about typical quantity and frequency of drinking to calculate average weekly consumption (‘On average, how many days a week do you have an alcoholic drink?’ and ‘On a typical drinking day, how many drinks do you have?’). Patients exceeding daily or weekly drinking guidelines for their age and sex are considered to have screened positive for unhealthy alcohol use (≥5 drinks/day or > 14 drinks/week for men 18–65 years or ≥ 4 drinks/day or > 7 drinks/week for men ≥66 years and women). The electronic health record (EHR) issues annual reminders for medical assistants to conduct screening during the primary care clinic rooming process or every 6 months if the prior screening was positive. During the COVID-19 pandemic until around the end of 2021, screenings were primarily conducted via telehealth rather than office visits. The study was approved by the KPNC Institutional Review Board, which granted a waiver of informed consent.
Sample
We identified 151 262 adults (≥18 years) with unhealthy alcohol use between 1 January 2019 to 29 February 2020 from the KPNC Adult Alcohol Registry; the last positive screening indicating unhealthy alcohol use was defined as the index ‘pre-pandemic’ screening. The registry is a distributed data structure of comprehensive, longitudinal EHR-based data collected during care delivery (Palzes et al. 2020b). Patients were followed from 1 March 2020 up to the first completed alcohol screening during follow-up (‘postpandemic’ screening), or until 30 June 2022. Patients were excluded if they no longer met eligibility criteria by the start of follow-up (i.e. had a negative alcohol screening, died, or disenrolled from the health plan), were health plan members for <70% of the year before the index date, or had an incomplete index alcohol screening (n = 52 962) (Fig. S1).
Alcohol use outcomes
Changes in the number of heavy drinking days in the past 3 months (HDDs) and average weekly alcohol consumption (drinks/week) were calculated as continuous measures by subtracting the prepandemic measure from the postpandemic measure. We also created categorical variables indicating the direction of change (maintained, increased, or decreased).
Covariates
Demographic characteristics included sex (male, female), age group at index (18–20, 21–34, 35–49, 50–64, ≥65 years), race (American Indian/Alaska Native, Asian/Pacific Islander, Black, White, or unknown), and ethnicity (Latino/Hispanic). Type of insurance included Medicaid, Medicare, commercial, or other/unknown. As a measure of socioeconomic status, we used the neighborhood deprivation index (NDI) from geocoded census-tract data from the 2019 American Community Survey and created a categorical variable based on sample distribution quartiles (Messer et al. 2006). Smoking status was identified at the most recent primary care screening in the year before index. Prior-year AUD, drug use disorder (DUD, excluding nicotine), depression, and anxiety disorder were identified using ‘International Classification of Diseases, Version 10’ diagnosis codes. As a measure of medical comorbidity, we calculated the Charlson comorbidity index and created a categorical variable (0, 1–2, and ≥ 3) (Deyo et al. 1992). The index month and number of days between the pre- and postpandemic alcohol screenings were also included.
Inverse probability of attrition weighting
Of the 98 300 eligible patients during follow-up, 593 died (0.6%), 11 240 (11.4%) disenrolled from the health plan, 20 096 (2.4%) had no alcohol screening, and 3447 (3.5%) had incomplete screenings. We compared characteristics of the 62 924 (64.0%) patients retained and 35 376 (36.0%) patients lost to follow-up by calculating standardized mean differences (SMDs) (Yang and Dalton 2012). Greater proportions of patients aged 21–34 years, patients with commercial insurance, and patients with Charlson comorbidity indices of 0 were lost to follow-up than retained (SMD > 0.2) (Table S1). To account for potential selection bias due to attrition, we estimated and applied inverse probability of attrition weights (IPAW) for each patient. Logistic regression was used to estimate the probability of no attrition for each patient, conditional upon a set of covariates hypothesized to be associated with attrition and the outcomes. The distribution of the weights for the retained sample was reasonable (mean = 1.56, median = 1.58, min = 1.07, max = 3.03) and were relatively stable (Seaman and White 2013).
Statistical analysis
Multivariable generalized linear models were weighted by IPAW and fit to examine associations of sex, age group, race, ethnicity, and AUD diagnoses with alcohol use outcomes. Models were adjusted for insurance type, NDI quartile, smoking, depression/anxiety disorder, DUD, Charlson comorbidity index, index month, and the number of days between screenings. Linear models were fit for continuous outcomes. Multinomial logistic models were fit for categorical outcomes to compare the odds of increasing or maintaining alcohol use rather than decreasing. Variance was estimated using the conservative robust sandwich standard error estimator approach (Austin 2016). Small proportions of patients had unknown NDI (0.1%) and smoking status (1.5%); thus, missing data were imputed to the mean and mode, respectively.
Since not all patients were screened at the same time during the pandemic period, the changes in alcohol use observed could have been related to the timing of screening. We conducted two sensitivity analyses to help put the main findings in context. The first sensitivity analysis examined whether patient characteristics were associated with the timing of the postpandemic alcohol screening, and the second sensitivity analysis examined whether there was variation in changes in alcohol use over the postpandemic period. We divided the cohort into five subgroups based on the timing of the postpandemic alcohol screening, roughly in 6-month intervals. Subgroup characteristics were compared with chi-square and Analysis of Variance tests. To examine variation in unadjusted changes in alcohol use, means and 95% CIs were calculated within each subgroup, weighted by IPAW, and plotted. Then, an ordinal variable representing the timing of the postpandemic screening was included in multivariable linear models to examine potential variation in alcohol use changes over time, adjusting for all covariates included in the main analyses and weighting by IPAW. P-values and 95% CIs were not corrected for multiple comparisons.
Analyses were conducted using SAS software, Version 9.4 (SAS Institute Inc., Cary, NC). The analyses presented in the current manuscript are a subset of all analyses performed. Two related follow-up studies with different research aims are published elsewhere (Satre et al. 2024; Kline-Simon et al. Under Review).
Results
Sample characteristics
The analytical sample (n = 62 924) was 39.2% female, 60.8% male, 0.8% American Indian/Alaska Native, 9.8% Asian/Pacific Islander (API), 5.8% Black, 16.3% Latino/Hispanic, 64.9% White, and 2.3% had unknown race (Table 1). The mean age was 52.3 (SD = 18.0) years. Most patients had commercial insurance (63.2%), and 3.4% had an AUD diagnosis. On average, patients had 574.9 days (SD = 231.3), or 1.6 years, between pre- and postpandemic alcohol screenings.
Table 1.
Demographic and clinical characteristics of adults with unhealthy alcohol use between 1 January 2019 and 29 February 2020 with observed outcome data, n = 62 924a
| Characteristic | N (%) |
|---|---|
| Sex | |
| Female | 24 676 (39.2) |
| Male | 38 248 (60.8) |
| Age (years), mean (SD) | 52.3 (18.0) |
| Age group (years) | |
| 18–20 | 1052 (1.7) |
| 21–34 | 12 515 (19.9) |
| 35–49 | 14 278 (22.7) |
| 50–64 | 14 046 (22.3) |
| ≥65 | 21 033 (33.4) |
| Race and ethnicity | |
| American Indian/Alaska Native | 512 (0.8) |
| Asian/Pacific Islander | 6196 (9.8) |
| Black | 3666 (5.8) |
| Latino/Hispanic | 10 285 (16.3) |
| White | 40 825 (64.9) |
| Unknown | 1440 (2.3) |
| Type of insurance | |
| Medicaid | 1662 (2.6) |
| Medicare | 21 226 (33.7) |
| Commercial | 39 785 (63.2) |
| Other/Unknown | 251 (0.4) |
| Neighborhood deprivation index | |
| First quartile (lowest) | 16 475 (26.2) |
| Second quartile | 16 636 (26.4) |
| Third quartile | 15 665 (24.9) |
| Fourth quartile (highest) | 14 085 (22.4) |
| Unknown | 63 (.1) |
| Smoking | |
| No | 54 351 (86.4) |
| Yes | 7620 (12.1) |
| Unknown | 953 (1.5) |
| Charlson comorbidity index | |
| 0 | 47 430 (75.4) |
| 1–2 | 11 887 (18.9) |
| ≥3 | 3607 (5.7) |
| Depression/anxiety disorder | 12 281 (19.5) |
| Anxiety disorder | 7451 (11.8) |
| Depression | 7573 (12.0) |
| Alcohol use disorder | 2163 (3.4) |
| Drug use disorder | 703 (1.1) |
| Days between pre- and postpandemic alcohol screenings | |
| Mean (SD) | 574.9 (231.3) |
| Median (IQR) | 548.0 (350.0) |
| Minimum; maximum | 111.0; 1272.0 |
Unhealthy alcohol use was defined as exceeding either the daily or weekly drinking limit recommended by the NIAAA for the patient’s age and sex (≥5 drinks/day or > 14 drinks/week for men aged 18–65 years, and ≥ 4 drinks/day or > 7 drinks/week for men aged ≥66 years and women).
Pre- and postpandemic alcohol use
Overall, the sample had a mean number of HDDs of 5.1 (SD = 16.5) prepandemic and 2.1 (SD = 12.5) during the pandemic and a mean number of drinks/week of 10.6 (SD = 12.2) prepandemic and 6.5 (1.7) during the pandemic (Table 2). Men had more HDDs and drinks/week compared to women across both periods, on average. Patients aged 35–49 years had the highest mean number of HDDs prepandemic and patients aged 50–64 years had the highest mean number of HDDs during the pandemic. Patients aged ≥65 years had the highest mean number of drinks/week both pre- and during the pandemic. Latino/Hispanic patients had the highest mean number of HDDs prepandemic, and White patients had the highest mean number of HDDs during the pandemic. White patients had the highest mean number of drinks/week both pre- and during the pandemic. Patients with AUD had more heavy drinking days and drinks/week than patients without AUD across both periods.
Table 2.
Changes in heavy drinking days and drinks/week by patient subgroups
| Outcome |
Prepandemic
Mean (SD) |
Postpandemic
Mean (SD) |
Post − Pre
Mean (SD) |
aMD (95% CI) a | P-value |
|---|---|---|---|---|---|
| Heavy drinking days | |||||
| Overall | 5.1 (16.5) | 2.1 (12.5) | −3.0 (18.4) | ||
| By sex | |||||
| Female | 3.1 (11.8) | 1.3 (9.1) | −1.9 (13.7) | 1.93 (1.70, 2.16) | <.001 |
| Male | 6.4 (18.8) | 2.7 (14.3) | −3.8 (20.9) | (Reference) | |
| By age group (years) | |||||
| 18–20 | 5.6 (13.0) | 1.4 (8.0) | −4.2 (14.4) | −0.48 (−1.19, 0.22) | .177 |
| 21–34 | 5.5 (14.9) | 1.7 (1.3) | −3.8 (16.5) | (Reference) | |
| 35–49 | 6.0 (17.8) | 2.3 (13.3) | −3.7 (19.7) | 0.20 (−0.13, 0.54) | .230 |
| 50–64 | 5.7 (18.8) | 2.4 (14.1) | −3.2 (21.1) | 0.37 (0.01, 0.74) | .050 |
| ≥65 | 3.4 (14.7) | 2.1 (12.2) | −1.3 (16.7) | 2.17 (0.89, 3.44) | <.001 |
| By race and ethnicity | |||||
| Asian/Pacific Islander | 5.2 (14.5) | 1.6 (10.3) | −3.6 (15.8) | −0.38 (−0.74, −0.02) | .040 |
| Black | 5.0 (15.6) | 1.6 (1.6) | −3.4 (17.9) | −0.16 (−0.68, 0.36) | .541 |
| Latino/Hispanic | 6.4 (17.7) | 2.1 (12.0) | −4.3 (2.0) | −0.78 (−1.15, −0.40) | <.001 |
| White | 4.8 (16.5) | 2.2 (13.1) | −2.5 (18.4) | (Reference) | |
| Other/unknown | 5.4 (16.7) | 2.1 (13.2) | −3.3 (19.5) | −0.09 (−0.78, 0.60) | .805 |
| By AUD diagnosis | |||||
| No | 4.9 (15.6) | 2.0 (12.1) | −2.9 (17.4) | (Reference) | |
| Yes | 13.2 (31.9) | 5.3 (21.8) | −7.9 (36.1) | −4.50 (−5.84, −3.16) | <.001 |
| Drinks/week | |||||
| Overall | 10.6 (12.2) | 6.5 (1.7) | −4.1 (12.2) | ||
| By sex | |||||
| Female | 8.5 (8.7) | 5.1 (7.8) | −3.4 (9.1) | 0.92 (0.77, 1.08) | <.001 |
| Male | 12.0 (13.7) | 7.5 (12.0) | −4.5 (13.8) | (Reference) | |
| By age group (years) | |||||
| 18–20 | 5.7 (12.4) | 3.1 (7.6) | −2.6 (12.5) | 0.52 (−0.10, 1.13) | .101 |
| 21–34 | 7.3 (12.1) | 4.2 (9.5) | −3.1 (11.9) | (Reference) | |
| 35–49 | 9.9 (14.0) | 5.9 (11.7) | −4.0 (13.8) | −0.83 (−1.06, −0.59) | <.001 |
| 50–64 | 12.5 (13.6) | 7.7 (11.8) | −4.8 (13.9) | −1.77 (−2.02, −1.52) | <.001 |
| ≥65 | 13.4 (8.0) | 8.8 (9.1) | −4.6 (9.4) | −1.16 (−1.87, −0.45) | .001 |
| By race and ethnicity | |||||
| Asian/Pacific Islander | 7.5 (11.5) | 4.1 (8.8) | −3.4 (1.8) | 0.19 (−0.05, 0.44) | .122 |
| Black | 8.9 (12.0) | 4.9 (9.8) | −4.0 (12.3) | 0.21 (−0.14, 0.57) | .238 |
| Latino/Hispanic | 9.8 (14.4) | 5.0 (10.9) | −4.7 (14.1) | −0.80 (−1.06, −0.55) | <.001 |
| White | 11.6 (11.5) | 7.6 (10.7) | −4.0 (11.8) | (Reference) | |
| Other/unknown | 9.6 (12.1) | 5.5 (10.7) | −4.1 (12.2) | −0.42 (−0.85, 0.01) | .057 |
| By AUD diagnosis | |||||
| No | 10.3 (11.5) | 6.4 (10.3) | −3.9 (11.4) | (Reference) | |
| Yes | 20.4 (21.8) | 10.4 (17.3) | −1.1 (24.1) | −5.38 (−6.29, −4.47) | <.001 |
Legend: AUD, alcohol use disorder.
Adjusted mean differences (aMD) and 95% confidence intervals (CI), for a specific group as compared to the reference group, were estimated by fitting multivariable linear models on the change in alcohol use outcome including all patient covariates (sex, age group, race, ethnicity, NDI quartile, type of insurance, smoking, Charlson comorbidity index, AUD, drug use disorder, depression/anxiety disorder, index month, and number of days between screenings). Models were fit among the sample with observed outcome data, weighted by inverse probability of attrition weights.
Mean changes in alcohol use
In univariate analyses weighted by IPAW, overall alcohol use decreased during the pandemic. Specifically, the cohort reduced drinking by an average of 3.0 HDDs (in the past 3 months) (SD = 18.4) and 4.1 drinks/week (SD = 12.2) (Table 2).
All patient subgroups decreased the number of HDDs and drinks/week, on average, during the pandemic. However, women had smaller decreases in the number of HDDs (adjusted mean difference [aMD] [CI] = 1.93 [1.70, 2.16]) and in drinks/week (aMD [CI] = 0.92 [0.77, 1.08]) compared to men (Table 2). Compared to patients aged 21–34 years, patients aged 50–64 years and ≥65 years had smaller decreases in HDDs (aMD [CI] = 0.37 [0.01, 0.74], and 2.17 [0.89, 3.44], respectively). Patients aged 35–49 years, 50–64 years, and ≥65 years had larger decreases in drinks/week compared to patients aged 21–34 years (aMD [CI] = −0.83 [−1.06, −0.59], −1.77 [−2.02, −1.52], and − 1.16 [−1.87, −0.45], respectively). API patients had larger decreases in HDDs compared to White patients (aMD [CI] = −0.38 [−0.74, −0.02]). Latino/Hispanic patients had larger decreases in both HDDs and drinks/week compared to White patients (aMD [CI] = −0.78 [−1.15, −0.40], and −0.80 [−1.06, −0.55], respectively). Patients with AUD diagnoses, compared to those without, had larger decreases in both HDDs and drinks/week (aMD [CI] = −4.50 [−5.84, −3.16] and − 5.38 [−6.29, −4.47]).
Direction of changes in alcohol use
Overall, smaller proportions of patients increased or maintained rather than decreased alcohol use. Specifically, 9.1% increased, 34.4% maintained, and 56.5% decreased HDDs (in the past 3 months), and 20.2% increased, 19.8% maintained, and 60.1% decreased drinks/week (Table 3). Distributions of these proportions varied by patient subgroups. For example, the unadjusted proportions who increased, maintained, and decreased the number of HDDs were 8.0%, 41.7%, and 50.3%, respectively, for women, and 9.9%, 29.8%, and 60.4%, respectively, for men.
Table 3.
Direction of changes in heavy drinking days and drinks/week
| Outcome | Increased, % | Maintained, % | Decreased, % | Increased vs decreased, aOR (95% CI)a | Maintained vs decreased, aOR (95% CI)a |
|---|---|---|---|---|---|
| Heavy drinking days | |||||
| Overall | 9.1 | 34.4 | 56.5 | ||
| By sex | |||||
| Female | 8.0 | 41.7 | 50.3 | 1.07 (1.02, 1.13) | 2.34 (2.26, 2.42) |
| Male | 9.9 | 29.8 | 60.4 | (Reference) | (Reference) |
| By age group (years) | |||||
| 18–20 | 9.2 | 10.0 | 80.9 | 0.92 (0.78, 1.09) | 0.66 (0.57, 0.78) |
| 21–34 | 9.3 | 13.5 | 77.2 | (Reference) | (Reference) |
| 35–49 | 9.8 | 21.5 | 68.7 | 1.17 (1.10, 1.24) | 1.85 (1.76, 1.94) |
| 50–64 | 9.5 | 36.9 | 53.7 | 1.39 (1.30, 1.48) | 3.60 (3.44, 3.78) |
| ≥65 | 8.2 | 68.3 | 23.6 | 2.55 (2.09, 3.10) | 12.52 (11.05, 14.19) |
| By race and ethnicity | |||||
| Asian/Pacific Islander | 7.4 | 19.0 | 73.6 | 0.59 (0.55, 0.64) | 0.56 (0.52, 0.59) |
| Black | 7.7 | 26.4 | 66.0 | 0.61 (0.55, 0.68) | 0.61 (0.56, 0.65) |
| Latino/Hispanic | 9.3 | 20.8 | 69.9 | 0.77 (0.73, 0.82) | 0.67 (0.64, 0.70) |
| White | 9.5 | 42.1 | 48.4 | (Reference) | (Reference) |
| Other/unknown | 9.3 | 25.3 | 65.4 | 0.83 (0.74, 0.94) | 0.79 (0.72, 0.86) |
| By AUD diagnosis | |||||
| No | 9.0 | 34.3 | 56.7 | (Reference) | (Reference) |
| Yes | 12.5 | 37.9 | 49.6 | 1.48 (1.31, 1.68) | 1.15 (1.04, 1.27) |
| Drinks/week | |||||
| Overall | 20.2 | 19.8 | 60.1 | ||
| By sex | |||||
| Female | 18.0 | 21.1 | 60.9 | 0.78 (0.75, 0.81) | 1.11 (1.07, 1.15) |
| Male | 21.6 | 18.9 | 59.5 | (Reference) | (Reference) |
| By age group (years) | |||||
| 18–20 | 24.1 | 28.1 | 47.8 | 1.37 (1.22, 1.53) | 1.79 (1.61, 2.01) |
| 21–34 | 21.7 | 19.4 | 58.9 | (Reference) | (Reference) |
| 35–49 | 21.9 | 17.8 | 60.4 | 0.96 (0.92, 1.00) | 0.91 (0.87, 0.96) |
| 50–64 | 20.3 | 17.8 | 61.9 | 0.86 (0.82, 0.90) | 0.89 (0.85, 0.94) |
| ≥65 | 16.6 | 23.2 | 60.2 | 0.73 (0.63, 0.84) | 1.02 (0.88, 1.17) |
| By race and ethnicity | |||||
| Asian/Pacific Islander | 18.2 | 23.5 | 58.3 | 0.79 (0.75, 0.84) | 1.33 (1.26, 1.40) |
| Black | 19.0 | 19.6 | 61.4 | 0.87 (0.81, 0.93) | 1.03 (0.96, 1.11) |
| Latino/Hispanic | 18.9 | 19.9 | 61.3 | 0.80 (0.76, 0.84) | 1.08 (1.03, 1.13) |
| White | 21.0 | 19.2 | 59.8 | (Reference) | (Reference) |
| Other/unknown | 20.0 | 18.6 | 61.4 | 0.84 (0.77, 0.91) | 1.00 (0.91, 1.09) |
| By AUD diagnosis | |||||
| No | 20.3 | 20.0 | 59.8 | (Reference) | (Reference) |
| Yes | 18.6 | 13.4 | 68.0 | 0.85 (0.77, 0.94) | 0.62 (0.56, 0.69) |
Legend: AUD, alcohol use disorder.
Adjusted odds ratios (aOR) and 95% confidence intervals (CI) were estimated by fitting multivariable multinomial logistic models on the direction of change in alcohol use outcome including all patient covariates (sex, age group, race, ethnicity, NDI quartile, type of insurance, smoking, Charlson comorbidity index, AUD, drug use disorder, depression/anxiety disorder, index month, and number of days between screenings). Models were fit among the sample with observed outcome data, weighted by inverse probability of attrition weights.
In multivariable analyses, certain patient subgroups had higher odds of increasing rather than decreasing or maintaining rather than decreasing alcohol use. Compared to men, women had higher odds of increasing rather than decreasing the number of HDDs (adjusted odds ratio [aOR] [CI] = 1.07 [1.02, 1.13]) and of maintaining rather than decreasing the number of HDDs (aOR [CI] = 2.34 [2.26, 2.42]). Women also had higher odds of maintaining rather than decreasing the number of drinks/week (aOR [CI] = 1.11 [1.07, 1.15]), but lower odds of increasing rather than decreasing the number of drinks/week (aOR [CI] = 0.78 [0.75, 0.81]), compared to men.
Patients aged 18–20 years, compared to patients 21–34 years, had lower odds of maintaining rather than decreasing the number of HDDs (aOR [CI] = 0.66 [0.57, 0.78]), but higher odds of increasing rather than decreasing the number of drinks/week (aOR [CI] = 1.37 [1.22, 1.53]) and higher odds of maintaining rather than decreasing drinks/week (aOR [CI] = 1.79 [1.61, 2.01]). Patients aged ≥35 years had higher odds of increasing and maintaining rather than decreasing the number of HDDs compared to patients aged 21–34 years (aORs = 1.17 to 12.52). Patients aged 35–64 years had lower odds of maintaining rather than decreasing the number of drinks/week (aORs = 0.89 to 0.91), and patients aged ≥50 years had lower odds of increasing rather than decreasing the number of drinks/week (aORs = 0.73 to 0.86), compared to patients 21–34 years.
Compared to White patients, all other racial/ethnic groups had lower odds of increasing rather than decreasing the number of HDDs (aORs = 0.59 to 0.83) and the number of drinks/week (aORs = 0.79 to 0.84). All other racial/ethnic groups also had lower odds of maintaining rather than decreasing the number of HDDs compared to White patients (aORs = 0.56 to 0.79); however, API and Latino/Hispanic patients had higher odds of maintaining rather than decreasing the number of drinks/week (aOR [CI] = 1.33 [1.26, 1.40], and 1.08 [1.03, 1.13], respectively).
Patients with AUD diagnoses, compared to those without, had higher odds of both increasing and maintaining rather than decreasing the number of HDDs (aOR [CI] = 1.48 [1.31, 1.68], and 1.15 [1.04, 1.27], respectively). Patients with AUD diagnoses, compared to those without, had lower odds of both increasing and maintaining rather than decreasing the number of drinks/week (aOR [CI] = 0.85 [0.77, 0.94], and 0.62 [0.56, 0.69], respectively).
Sensitivity analyses
Patients with earlier screenings (e.g. 1 March 2020 to 31 August 2020), compared to those with later screenings (e.g. 1 March 2022 to 30 June 2022), were relatively older and had higher proportions with Medicare insurance, Charlson comorbidity indices of ≥3 (indicating more comorbidities), depression/anxiety disorder, and AUD diagnoses (Table S2). The plots of the unadjusted changes in alcohol use by the timing of the postpandemic screening suggested a potential negative linear relationship, with greater reductions in drinking observed for patients with later screenings (Fig. S2). In multivariable analyses, reductions in the number of HDDs (P = 0.780) and drinks/week (P = 0.764) did not significantly vary by timing of the postpandemic screening, after adjusting for differences in patient characteristics.
Discussion
Our study contributes to the literature by examining changes in alcohol consumption during the COVID-19 pandemic among adults with prior unhealthy alcohol use with a prospective cohort study design. While continuous measures are informative for understanding potential harms at a population- or group-based level, means can mask higher-risk groups; thus, we also included categorical measures to understand who was more likely to increase or maintain unhealthy levels of drinking, indicating higher risk and greater need for intervention. Contrary to our hypotheses, we found that most patients decreased alcohol use during the pandemic, leading to overall and subgroup mean decreases; however, there was significant heterogeneity by sex, age group, race, ethnicity, and AUD diagnoses.
Some subgroups had smaller decreases in alcohol use than others, on average. Specifically, we found that women, White patients, and patients without AUD diagnoses generally had smaller decreases in drinking compared to their counterparts, though women and patients without AUD generally had lower levels of drinking across both periods. Additionally, older adults aged ≥50 years had smaller mean decreases in HDDs, but larger mean decreases in overall consumption, compared to adults aged 21–34 years. Older adults aged ≥65 years had the lowest mean number of HDDs and the highest mean number of drinks/week prepandemic of any age group; these baseline levels may have affected the magnitude of change. Analyses of the direction of change in drinking suggest that certain subgroups may have been at greater risk for increased alcohol use. We found that women (compared to men), patients aged ≥35 years (compared to patients aged 21–34 years), White patients (compared to other racial and ethnic groups), and patients with AUD diagnoses were more likely to increase rather than decrease HDDs and that patients aged 18–20 years and White patients were more likely to increase rather than decrease overall consumption. We also found that women and certain age groups were more likely to maintain prepandemic levels of unhealthy alcohol use during the pandemic. However, the findings regarding larger decreases in HDDs, yet greater likelihood of increasing rather than decreasing, for patients with AUD diagnoses, suggests that the increases observed among those who increased did not outweigh the decreases observed among those who decreased. Overall, these findings suggest that women, White patients, patients with AUD diagnoses, and certain age groups may be at greater risk of alcohol-related problems. Prior studies have found that women and older age groups were less likely to receive brief alcohol interventions following screening (Chen et al. 2020; Lu et al. 2021; Parthasarathy et al. 2023); thus, improving universal delivery of brief intervention to patients identified with unhealthy drinking will be important for addressing disparities in alcohol-related harm.
There are potentially many reasons why individuals changed alcohol consumption during the pandemic. They may have increased drinking to cope with stress, worries, or loneliness; to manage depression or anxiety symptoms; or because they had more free time due to, for example, loss of employment or fewer activities outside the home (McPhee et al. 2020; Alpers et al. 2021; Tucker et al. 2022). Conversely, they may have decreased drinking due to reduced income, lower alcohol availability, fewer social motives and peer influences due to social distancing, protective effects of having more time spent at home with family, or general health concerns in the context of the pandemic (Grossman et al. 2020; Garnett et al. 2021; Graupensperger et al. 2021). The pandemic also changed where people consumed alcohol; bar and restaurant closures and the rise of alcohol home delivery services could have affected people differently depending on where they typically consumed alcohol prior to the pandemic (Grossman et al. 2022; Sohi et al. 2022).
Prior studies in the USA generally found increases in drinking frequency, prevalence of exceeding drinking limits, and severity of alcohol use problems during the early period of the pandemic in 2020, particularly among women and young adults (Boschuetz et al. 2020; Czeisler et al. 2020; Pollard et al. 2020; Barbosa et al. 2021; Metz et al. 2022), although these studies did not focus specifically on individuals with prepandemic unhealthy alcohol use as we did. Our study findings may differ from the majority of the existing literature for several reasons. First, our sample may have decreased alcohol use in response to screening or receipt of brief alcohol interventions or other alcohol treatment (Kaner et al. 2018; Chi et al. 2022), as our sample consists of adults with prepandemic unhealthy alcohol use. It is possible that attention to alcohol use in healthcare provided an overall protective effect against drinking escalation during the pandemic. Second, our sample was derived from a primary care population who used care. Healthcare utilization decreased during the pandemic, especially among patients in better health (Moynihan et al. 2021), and it is possible that patients in our sample were more likely to have medical comorbidities and therefore more inclined to reduce drinking over time. Third, we included a longer follow-up period to allow alcohol screening workflows to return to prepandemic levels as screening rates dropped significantly during the first few months of the pandemic in this healthcare system, similar to others (Alford et al. 2023; Chi et al. 2023). Most prior studies that found increased drinking examined only the early pandemic period, making it difficult to compare findings. Including a longer follow-up period provided an opportunity to examine how drinking has changed over the course of the pandemic, but it is unclear whether patients maintained the same level of drinking throughout. Our sensitivity analyses suggest that there may have been variation in changes in drinking depending on when screening occurred during the pandemic and that the variation might be due to differences in characteristics of patients who were screened earlier rather than later during the period. Our findings are more similar to a recent study of young adults (aged 21–29 years) with prepandemic unhealthy drinking that found significant reductions in alcohol use quantity and frequency during and after the pandemic (Creswell et al. 2024).
Limitations
This study has limitations. Generalizability may be limited to individuals with stable insurance who use care, though findings from this study are important for understanding how healthcare systems can tailor alcohol interventions. Additionally, because our study was focused on individuals with prepandemic unhealthy alcohol use, findings cannot be generalized to those with lower-risk or no drinking prior to the pandemic. While some prior studies show fluctuations in alcohol use during the pandemic (Pelham et al. 2021), we were unable to conduct a repeated-measures study due to frequency of screening, which typically occurs annually; thus, we may have missed short-term spikes in increased drinking. Screening practices may also have varied due to pandemic-related fluctuations in access to services. While our study examined changes in alcohol use from before to during the pandemic, we cannot say whether the changes were directly attributable to the pandemic. For the small proportion of patients (4.7%) with follow-up screenings during the first few months of the pandemic (March to May 2020), responses to the heavy drinking day question may have included drinking prior to the pandemic as it asks about the prior 3 months. We defined the start of the pandemic period as 1 March 2020 to be consistent with the literature, though changes in drinking related to stress and worry may have occurred earlier, as the first report of COVID-19 infection in the USA was on 20 January 2020 (Centers for Disease Control and Prevention 2023). Discussion of potential mechanisms underlying our findings are only speculative, and further research is needed for understanding reasons for decreased drinking during the pandemic. The alcohol use measures included in our study were self-reported, which might have been affected by social desirability bias due to stigma. Although the healthcare system has a well-established screening workflow to maximize patient comfort for disclosure of alcohol use, workflows changed during the pandemic to accommodate telehealth visits and may have affected patient responses. Additionally, prior research in other healthcare systems suggests that sensitivity of clinical alcohol screening may decrease over time (Lapham et al. 2014; McGinnis et al. 2022), which could affect findings. Having both self-reported and biological alcohol use data for the purposes of validation is ideal but impracticable and costly in the context of real-world healthcare. Therefore, self-reported data alone may still be valuable for understanding patterns in alcohol use over time when no other data are available, especially for a large observational sample during a public health emergency, such as the COVID-19 pandemic. While differences in the magnitude of drinking reductions by patient subgroups are notable clinically, there could be variability in floor or ceiling effects across subgroups; future research to replicate findings and to examine distributions of change as additional outcomes is warranted. Hypothesis tests were not corrected for multiple comparisons as we did not want to impose restrictions on the ability to examine changes across several subgroups (Rothman 1990); thus, some significant findings may be due to Type I error. Lastly, we used inverse probability weighting to adjust for potential selection bias due to attrition; however, residual bias is possible. The covariates for AUD, DUD, depression/anxiety diagnoses, and the Charlson comorbidity index included in this study were based on data recorded at encounters with the healthcare system in the EHR, which do not capture symptoms or duration of illness but serve as proxies for severity of health problems.
Conclusions
This study adds to the literature on how alcohol use changed during the COVID-19 pandemic by focusing on patients with pre-existing unhealthy alcohol use and examining differences by patient subgroups. Certain subgroups may be at greater risk of alcohol-related problems due to a greater likelihood of increasing or maintaining unhealthy alcohol use during the pandemic. Findings may be informative to healthcare systems for understanding service needs and tailoring alcohol interventions.
Supplementary Material
Acknowledgements
We gratefully acknowledge Raye Z. Litten, Ph.D.; Daniel Falk, Ph.D.; Laura E. Kwako, Ph.D.; Brett T. Hagman, Ph.D., Aaron White, Ph.D., and Mark Egli, Ph.D. at the National Institute on Alcohol Abuse and Alcoholism (NIAAA) for their feedback on analyses. We also thank Alison Truman at the Kaiser Permanente Division of Research for administrative support. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of NIAAA.
Contributor Information
Vanessa A Palzes, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States.
Felicia W Chi, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States.
Derek D Satre, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States; Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, 675 18th Street, San Francisco, CA 94107, United States.
Andrea H Kline-Simon, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States.
Cynthia I Campbell, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States; Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, 675 18th Street, San Francisco, CA 94107, United States; Department of Health Systems Science, Kaiser Permanente Bernard J. Tyson School of Medicine, 98 S Los Robles Ave, Pasadena, CA 91101, United States.
Constance Weisner, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States; Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, 675 18th Street, San Francisco, CA 94107, United States.
Stacy Sterling, Center for Addiction and Mental Health Research, Division of Research, Kaiser Permanente, 4480 Hacienda Drive, Pleasanton, CA 94588, United States; Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, 675 18th Street, San Francisco, CA 94107, United States; Department of Health Systems Science, Kaiser Permanente Bernard J. Tyson School of Medicine, 98 S Los Robles Ave, Pasadena, CA 91101, United States.
Author contributions
Vanessa Palzes (Conceptualization [equal], Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing—original draft, Writing—review & editing [lead]), Felicia Chi (Conceptualization, Methodology, Writing—review & editing [equal], Investigation, Software, Supervision [supporting]), Derek Satre (Conceptualization, Investigation [supporting], Funding acquisition, Methodology, Supervision, Writing—review & editing [equal]), Cynthia Campbell (Funding acquisition, Supervision, Writing—review & editing [equal]), Andrea Kline-Simon (Conceptualization, Methodology, Writing—review & editing [equal], Investigation, Software [supporting]), Constance Weisner (Conceptualization [equal], Methodology, Supervision, Writing—review & editing [equal], Funding acquisition, Project administration [lead], Investigation [supporting]), and Stacy Sterling (Conceptualization, Funding acquisition, Methodology, Project administration, Writing—review & editing [equal], Investigation [supporting], Supervision [lead]).
Conflict of interest: None declared.
Funding
This work was supported by the National Institute on Alcohol Abuse and Alcoholism (75N94021C00003 to C.W.; K24AA025703 to D.D.S.).
Data availability
The analytical datasets from this project consist of aggregate-level data that comply with Kaiser Permanente Northern California (KPNC) electronic medical record and administrative and clinical database policies per HIPAA regulations. External investigators may contact the corresponding author to initiate a request for study data to support new study proposals or manuscripts. Approval of requests will evaluate whether the proposed project is of high scientific merit, is consistent with the overall goals and objectives of the parent study, and is consistent with KPNC policies per HIPAA regulations.
References
- Alford DP, German JS, Bangham C. et al. Alcohol screening during COVID-19 surges in an urban health system the United States Prev Med Rep. 2023;33:102197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Alpers SE, Skogen JC, Mæland S. et al. Alcohol consumption during a pandemic lockdown period and change in alcohol consumption related to worries and pandemic measures Int J Environ Res Public Health. 2021;18:1220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- American Psychological Association . Stress in America 2020: A National Mental Health Crisis. 2022, 1–2. https://www.apa.org/news/press/releases/stress/2020/sia-mental-health-crisis.pdf [accessed 18 July 2022].
- Austin PC. Variance estimation when using inverse probability of treatment weighting (IPTW) with survival analysis Stat Med. 2016;35:5642–55. 10.1002/sim.7084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barbosa C, Cowell AJ, Dowd WN. Alcohol consumption in response to the COVID-19 pandemic in the United States J Addict Med. 2021;15:341–4. 10.1097/ADM.0000000000000767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Boschuetz N, Cheng S, Mei L. et al. Changes in alcohol use patterns in the United States during COVID-19 pandemic WMJ. 2020;119:171–6. [PubMed] [Google Scholar]
- Centers for Disease Control and Prevention . CDC Museum COVID-19 Timeline. Centers for Disease Control and Prevention. 2023. https://www.cdc.gov/museum/timeline/covid19.html [accessed 22 May 2024].
- Chen JA, Glass JE, Bensley KMK. et al. Racial/ethnic and gender differences in receipt of brief intervention among patients with unhealthy alcohol use in the U.S. veterans health administration J Subst Abuse Treat. 2020;119:108078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chi FW, Parthasarathy S, Palzes VA. et al. Alcohol brief intervention, specialty treatment and drinking outcomes at 12 months: Results from a systematic alcohol screening and brief intervention initiative in adult primary care Drug Alcohol Depend. 2022;235:109458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chi FW, Palzes VA, Lu Y. et al. Screening and Brief Intervention Rates before and after COVID-19 Onset in a U.S. Integrated Healthcare System with Systematic Alcohol SBIRT in Adult Primary Care . Greensboro, NC: Invited talk, 2023. [Google Scholar]
- COVID-19 Mental Disorders Collaborators . Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the COVID-19 pandemic Lancet. 2021;398:1700–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Creswell KG, Hisler GC, Lyons G. et al. Changes in alcohol consumption and alcohol problems before and after the COVID-19 pandemic: A prospective study in heavy drinking young adults Nat Mental Health. 2024;2:728–39Nature Publishing Group. [Google Scholar]
- Czeisler MÉ, Lane RI, Petrosky E. et al. Mental health, substance use, and suicidal ideation during the COVID-19 pandemic - United States, June 24-30, 2020 MMWR Morb Mortal Wkly Rep. 2020;69:1049–57. 10.15585/mmwr.mm6932a1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Davis AC, Voelkel JL, Remmers CL. et al. Comparing Kaiser Permanente members to the general population: Implications for generalizability of research Perm J. 2023;7:87–98The Permanente Federation. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases J Clin Epidemiol. 1992;45:613–9. [DOI] [PubMed] [Google Scholar]
- Garnett C, Jackson S, Oldham M. et al. Factors associated with drinking behaviour during COVID-19 social distancing and lockdown among adults in the UK Drug Alcohol Depend. 2021;219:108461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gonçalves PD, Moura HF, Amaral RA. et al. Alcohol use and COVID-19: Can we predict the impact of the pandemic on alcohol use based on the previous crises in the 21st century? A brief review Front Psych. 2020;11:581113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Graupensperger S, Fleming CB, Jaffe AE. et al. Changes in young adults’ alcohol and marijuana use, norms, and motives from before to during the COVID-19 pandemic J Adolesc Health. 2021;68:658–65. 10.1016/j.jadohealth.2021.01.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grossman ER, Benjamin-Neelon SE, Sonnenschein S. Alcohol consumption during the COVID-19 pandemic: A cross-sectional survey of US adults Int J Environ Res Public Health. 2020;17:E9189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grossman ER, Benjamin-Neelon SE, Sonnenschein S. Alcohol consumption and alcohol home delivery laws during the COVID-19 pandemic Subst Abus. 2022;43:1139–44. 10.1080/08897077.2022.2060432. [DOI] [PubMed] [Google Scholar]
- Guinle MIB, Sinha R. The role of stress, trauma, and negative affect in alcohol misuse and alcohol use disorder in women Alcohol Res. 2020;40:05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaner EF, Beyer FR, Muirhead C. et al. Effectiveness of brief alcohol interventions in primary care populations Cochrane Database Syst Rev. 2018;2:CD004148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kline-Simon A, Palzes V, Chi F. et al. Was Receipt of Specialty Alcohol Treatment during the Pandemic Effective at Reducing Drinking for Patients with or at Risk of AUD? Under Review.
- Lapham GT, Rubinsky AD, Williams EC. et al. Decreasing sensitivity of clinical alcohol screening with the AUDIT-C after repeated negative screens in VA clinics Drug Alcohol Depend. 2014;142:209–15. 10.1016/j.drugalcdep.2014.06.017. [DOI] [PubMed] [Google Scholar]
- Lu Y, Chi FW, Parthasarathy S. et al. Patient and provider factors associated with receipt and delivery of brief interventions for unhealthy alcohol use in primary care Alcohol Clin Exp Res. 2021;45:2179–89. 10.1111/acer.14702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McGinnis KA, Tate JP, Bryant KJ. et al. Change in alcohol use based on self-report and a quantitative biomarker, phosphatidylethanol, in people with HIV AIDS Behav. 2022;26:786–94. 10.1007/s10461-021-03438-y. [DOI] [PubMed] [Google Scholar]
- McPhee MD, Keough MT, Rundle S. et al. Depression, environmental reward, coping motives and alcohol consumption during the COVID-19 pandemic Front Psych. 2020;11:574676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Messer LC, Laraia BA, Kaufman JS. et al. The development of a standardized neighborhood deprivation index J Urban Health. 2006;83:1041–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Metz VE, Palzes VA, Chi FW. et al. Alcohol use among young adults in northern California during the COVID-19 pandemic: An electronic health records-based study Front Psych. 2022;13:883306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moynihan R, Sanders S, Michaleff ZA. et al. Impact of COVID-19 pandemic on utilisation of healthcare services: A systematic review BMJ Open. 2021;11:e045343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Palzes VA, Parthasarathy S, Chi FW. et al. Associations between psychiatric disorders and alcohol consumption levels in an adult primary care population Alcohol Clin Exp Res. 2020a;44:2536–44. 10.1111/acer.14477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Palzes VA, Weisner C, Chi FW. et al. The Kaiser Permanente northern California adult alcohol registry, an electronic health records-based registry of patients with alcohol problems: Development and implementation JMIR Med Inform. 2020b;8:e19081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Park CL, Russell BS, Fendrich M. et al. Americans’ COVID-19 stress, coping, and adherence to CDC guidelines J Gen Intern Med. 2020;35:2296–303. 10.1007/s11606-020-05898-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Parthasarathy S, Chi F, Metz V. et al. Disparities in the receipt of alcohol brief intervention: The intersectionality of sex, age and race/ethnicity Addiction. 2023;118:1258–69. [DOI] [PubMed] [Google Scholar]
- Pelham WE, Tapert SF, Gonzalez MR. et al. Early adolescent substance use before and during the COVID-19 pandemic: A longitudinal survey in the ABCD study cohort J Adolesc Health. 2021;69:390–7. 10.1016/j.jadohealth.2021.06.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pollard MS, Tucker JS, Green HD. Changes in adult alcohol use and consequences during the COVID-19 pandemic in the US JAMA Netw Open. 2020;3:e2022942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roberts A, Rogers J, Mason R. et al. Alcohol and other substance use during the COVID-19 pandemic: A systematic review Drug Alcohol Depend. 2021;229:109150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rothman KJ. No adjustments are needed for multiple comparisons Epidemiology. 1990;1:43–6Lippincott Williams & Wilkins. [PubMed] [Google Scholar]
- Satre DD, Palzes VA, Chi FW. et al. Unhealthy alcohol use among adults with depression or anxiety: Changes during COVID-19 and associations with mental health treatment J Stud Alcohol Drugs. 2024. 10.15288/jsad.23-00373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schmidt RA, Genois R, Jin J. et al. The early impact of COVID-19 on the incidence, prevalence, and severity of alcohol use and other drugs: A systematic review Drug Alcohol Depend. 2021;228:109065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Seaman SR, White IR. Review of inverse probability weighting for dealing with missing data Stat Methods Med Res. 2013;22:278–95. 10.1177/0962280210395740. [DOI] [PubMed] [Google Scholar]
- Sharma P, Nguyen QA, Kurani S. et al. Association of socio-demographic characteristics with alcohol use initiation among never users during the COVID-19 pandemic: A longitudinal study J Public Health (Oxf). 2023;45:442–9. 10.1093/pubmed/fdac099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sohi I, Chrystoja BR, Rehm J. et al. Changes in alcohol use during the COVID-19 pandemic and previous pandemics: A systematic review Alcohol Clin Exp Res. 2022;46:498–513. 10.1111/acer.14792. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tucker JS, Rodriguez A, Green HD. et al. Trajectories of alcohol use and problems during the COVID-19 pandemic: The role of social stressors and drinking motives for men and women Drug Alcohol Depend. 2022;232:109285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Willenbring ML, Massey SH, Gardner MB. Helping patients who drink too much: An evidence-based guide for primary care clinicians Am Fam Physician. 2009;80:44–50. [PubMed] [Google Scholar]
- Wu C, Louie P, Bierman A. et al. Assessment of sociodemographics and inflation-related stress in the US JAMA Netw Open. 2023;6:e2313431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yang D, Dalton JE. A Unified Approach to Measuring the Effect Size between Two Groups Using SAS, SAS Global Forum 2012. SAS Institute Inc, 2012. https://support.sas.com/resources/papers/proceedings12/335-2012.pdf [accessed 27 Oct 2020]. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The analytical datasets from this project consist of aggregate-level data that comply with Kaiser Permanente Northern California (KPNC) electronic medical record and administrative and clinical database policies per HIPAA regulations. External investigators may contact the corresponding author to initiate a request for study data to support new study proposals or manuscripts. Approval of requests will evaluate whether the proposed project is of high scientific merit, is consistent with the overall goals and objectives of the parent study, and is consistent with KPNC policies per HIPAA regulations.