Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 11.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2023 May 3;47(6):1119–1131. doi: 10.1111/acer.15091

Trends in alcohol-impaired crashes in California, 2016 to 2021: A time series analysis for alcohol involvement and crash distribution among demographic subgroups

Christina A Mehranbod 1, Ariana N Gobaud 1, Charles C Branas 1, Qixuan Chen 2, Daniel P Giovenco 3, David K Humphreys 4, Andrew G Rundle 1, Brady R Bushover 1, Christopher N Morrison 1,5
PMCID: PMC10858975  NIHMSID: NIHMS1960075  PMID: 37095075

Abstract

Background:

In 2020, the COVID-19 pandemic and control measures changed alcohol consumption in the United States (US) and globally. Before the pandemic, alcohol-impaired crashes contributed to approximately one-third of all road traffic crash injuries and fatalities nationally. We examined the impact of the COVID-19 pandemic on crashes and examined differences in alcohol-involved crashes across various subgroups.

Methods:

The University of California Berkeley Transportation Injury Mapping Systems provided information on all crashes reported to the California Highway Patrol from January 1, 2016 through December 31, 2021. Using autoregressive integrated moving average (ARIMA) models applied to weekly time series data, we estimated the effect of California’s first mandatory statewide shelter-in-place order (March 19, 2020) on crashes per 100,000 population. We also examined crash subgroups according to crash severity, sex, race/ethnicity, age, and alcohol involvement.

Results:

In California, the mean crash rate per week before the pandemic (January 1, 2016–March 18, 2020) was 9.5 crashes per 100,000 population, and 10.3% of those were alcohol-involved. After the initiation of the COVID-19 stay-at-home order, the percentage of crashes that were alcohol-involved rose to 12.7%. Overall, the crash rate across California decreased significantly (−4.6 crashes per 100,000; 95% CI: −5.3, −3.9), including across all examined subgroups, with the greatest decrease among the least severe crashes. However, there was a 2.3% absolute increase in the proportion of crashes that were alcohol-involved (0.02 crashes per 100,000; 95% CI: 0.02, 0.03).

Conclusions:

The initiation of a COVID-19 stay-at-home ordinance in California was associated with a substantial decrease in overall crash rates. While crashes have returned to pre-pandemic levels, alcohol-involved crashes remain elevated. The initiation of the stay-at-home order significantly increased alcohol-impaired driving, which has remained elevated.

Keywords: alcohol-impaired driving, crashes, pandemic

INTRODUCTION

Motor vehicle crashes are one of the leading causes of mortality in the United States, with death rates attributable to motor vehicle crashes notably higher than in most other high-income countries (Centers for Disease Control and Prevention National Center for Injury Prevention and Control, 2021; Yellman, 2022). In 2020, approximately 110 people died per day from crashes, costing the United States over $430 billion resulting from the high prevalence of injury (Centers for Disease Control and Prevention National Center for Injury Prevention and Control, 2021). One of the leading causes of motor vehicle crashes is alcohol-impaired driving (World Health Organization, 2019). In the United States, approximately 28% of all motor vehicle crash fatalities are attributable to alcohol-impaired driving, with an economic cost of $44 billion for the US (National Highway Traffic Safety Administration National Center for Statistics and Analysis, 2017). While alcohol-impaired driving decreased by 20% from 2007 to 2016, the burden on population health remains high as a major cause of preventable injury in the United States.

The beginning of the COVID-19 pandemic changed the trajectory of public health, transportation, injury, and alcohol consumption overall. Throughout the pandemic timeline, researchers detected changes in a wide range of health outcomes, such as post-traumatic stress disorder and overall mental health (Carmassi et al., 2020; Liu et al., 2021; Marques de Miranda et al., 2020), educational attainment and learning (Bonal & González, 2020; Stanistreet et al., 2020), and substance use (Czeisler et al., 2020; Ornell et al., 2020). Whether through increased psychological distress, changes in employment patterns or increased alcohol availability, among other factors, there is strong evidence of increases in alcohol consumption as a result of the COVID-19 pandemic (Biddle et al., 2020; Chodkiewicz et al., 2020; Grossman et al., 2020; Schmits & Glowacz, 2022). A nationally representative survey in the United States found increases in the average number of days people consumed alcohol and alcohol-related problems comparing before and during the COVID-19 pandemic (Pollard et al., 2020). An analysis of alcohol sales data, often used as a proxy for at-home drinking, found increases in retail alcohol sales in the beginning of the pandemic (Castaldelli-Maia et al., 2021). While trips to on-premise alcohol outlets like bars and restaurants decreased, another study found increases in trips to liquor stores (Hu et al., 2021). Levels of alcohol consumption and changes in those levels in a population have immense implications for alcohol-impaired driving (National Academies of Sciences et al., 2018). There is evidence that greater levels of binge or heavy drinking increase the likelihood of alcohol-impaired driving in a population (Hingson et al., 2017; Naimi et al., 2009; National Academies of Sciences et al., 2018). Further, some interventions aimed at reducing overall alcohol consumption, not necessarily alcohol-impaired driving, have also reduced alcohol-impaired driving (Naimi et al., 2018; National Academies of Sciences et al., 2018; Wagenaar et al., 2010; Xuan et al., 2015). Overall, if the beginning of the COVID-19 pandemic and subsequent stay-at-home orders increased rates of heavy drinking, and binge drinking has been documented as a risk factor for alcohol-impaired driving, a critical examination of the alcohol-related harm and injury that may result is imperative.

The primary goal of stay-at-home orders, according to the Centers for Disease Control and Prevention, was to reduce population mobility as a community mitigation strategy to combat COVID-19 spread (Moreland, 2020). Arguably one of the most common policy interventions enacted across the United States, the stay-at-home order would theoretically increase social distancing by limiting the functioning of non-essential business or mass gatherings and more in order to reduce people’s travel outside the home (Bourassa, 2021; Fowler et al., 2021; Moreland, 2020). For example, stay-at-home orders generally limited bars and restaurants to either close or transition to take-out services only (Borjas, 2020; Lee, 2020). However, with increases in drinking overall, there were still potential implications of stay-at-home orders on alcohol-impaired driving. While overall crash rates decreased due to the greater number of people staying at home and the subsequent decrease in vehicle miles traveled (Doucette et al., 2021; Yasin et al., 2021), when people did drink outside their home, options for safe transportation home after alcohol consumption could have decreased. This decrease may be due to risk aversion of drivers of ride-hailing companies, thus decreasing the number of ridesharing cars on the road and increasing the likelihood of alcohol-impaired driving (Wang et al., 2022). Further, if individuals were less comfortable with sharing spaces with others due to the increased risk of COVID-19 transmission, they might theoretically be less likely to use public transportation, usually associated with decreased rates of alcohol-impaired driving, thereby increasing rates of alcohol-impaired driving at the start of the pandemic (Aloi et al., 2020; Jabbari & MacKenzie, 2020; Sahraei et al., 2021). With more closure of bars and restaurants where alcohol could be consumed, many individuals moved drinking events to other individuals’ homes. After socializing and drinking at a person’s residence, people may have driven their own vehicle home, further increasing the likelihood of alcohol-impaired driving (Grossman et al., 2022; Mohan et al., 2020). With changes in transportation patterns and alcohol use, the pandemic likely left a lasting impact on patterns of alcohol-impaired driving.

A few studies have examined trends in alcohol-impaired driving in the context of the pandemic. An AAA Foundation for Traffic Safety research brief examined self-reported risky driving during the pandemic and found report an increase in “risky driving,” such as speeding and driving under the influence of alcohol. Although 60% of drivers reduced driving, those who reduced driving were already part of a population with a lower-than-average crash rate (Tefft et al., 2022). Some individuals increased their driving, such as those who were younger and male, at risk for greater crash rates. Another report by the National Highway Traffic Safety Administration (NHSTA) documented an increase of 14.3% from 2019 to 2020 in alcohol-impaired crashes (National Highway Traffic Safety Administration National Center for Statistics and Analysis, 2022). Similarly, a study of seriously or fatally injured road users involved in motor vehicle crashes found increases in alcohol prevalence when comparing pre- and post-March 16, 2020 (National Highway Traffic Safety Administration Office of Behavioral Safety Research, 2021). Crash data from Virginia pointed to increases in rates of crashes involving alcohol use after the start of the COVID-19 pandemic (Dong et al., 2022). However, although most trends have been documented through research briefs and surveys emphasizing population-level changes in injury, fatalities, and substance use, these analyses do not control for temporal structure or time-varying confounders that accompany a sudden decrease in mobility due to the start of the pandemic and subsequent increase in alcohol consumption patterns. Using an epidemiological design informs future policy and intervention initiatives to reduce alcohol-impaired driving, improve traffic safety, and better population health overall.

To examine the impact of the COVID-19 pandemic on alcohol-involved crash rates, we conducted interrupted time series analyses of crashes in California between 2016 and 2021. ARIMA models are used for time series forecasting and to examine the impact of interventions or events, interrupted in time, commonly in public health (Bernal et al., 2016; McDowell et al., 1980). Interrupted time series analyses estimate the causal effect of observational data when randomization into intervention arms is not feasible (Bernal et al., 2016; Kontopantelis et al., 2015). This interrupted time series design tested whether the initiation of a statewide shelter-in-place order reduced crashes using the pre- and post-event time series segments and examined changes in alcohol involvement in crashes. Given increased alcohol consumption resulting from the statewide shelter-in-place order during the COVID-19 pandemic, we hypothesized that rates of alcohol-impaired driving would increase. We also examined the impact of the shelter-in-place order on demographic subgroups and crash types to identify those most severely impacted by the changing landscape of injury and alcohol use.

MATERIALS AND METHODS

Study population

California has a population of 39,029,342 in 2022, making up approximately 12% of the population of the United States (United States Census Bureau, 2021). Over the last 5 years, California experienced an average of 176,305 motor vehicle crashes a year, resulting in various injuries and fatalities (“Transportation Injury Mapping System (TIMS)”, 2022). On January 25, 2020, California saw its first case of COVID-19, and by March 4, 2020, the governor had declared a state of emergency. On March 19, 2020, a statewide shelter-in-place order was issued (Procter, 2021).

Data and variables

We obtained vehicular crash data through the Transportation Injury Mapping System (TIMS), a data source managed by the University of California, Berkeley (“Transportation Injury Mapping System (TIMS)”, 2022). TIMS uses information from crashes reported to the California Highway Patrol through Statewide Integrated Traffic Records System (SWITRS; California Highway Patrol, 2023). Information on the crash level, party level and victim level are available in this dataset. The crash dataset contains information on the date and time of the crash, day of week, population size of area where crash occurred, severity of collision, alcohol involvement, and the longitude and latitude of the crash, among many other variables. The party dataset includes information for all parties that were involved in a crash, such as multiple motorists, pedestrians, etc. They are considered “the major players in traffic crashes” (“Transportation Injury Mapping System (TIMS)”, 2022). For a given crash in the crash-level dataset, there could be multiple parties. Information on parties include but are not limited to the party type (such as driver, pedestrian, parked vehicle, etc.), whether the given party was the one at fault for the crash, party race, party sex, and party age. For each party, there could be multiple victims whose individual-level characteristics were included in the dataset, such as victim role, sex, or age. Personal identifiers are not included in the dataset such as name or date of birth, so identifiability is not a major concern. Due to its relational database structure, we merged crash-level and party-level data by common variable “case_id.”

Variables

Variables in the crash-level dataset necessary for our analysis included: (1) crash date, (2) collision severity (“fatal”), “injury (severe)”, “injury (other visible)”, “injury (complaint of pain)”, and (3) alcohol involvement. Variables in the party-level dataset necessary for our analysis included: (1) at fault (“yes”, “no”), (2) party sex (“male”, “female”, “nonbinary”), (3) party age (continuous variable 0–125+), and (4) party race (“Asian”, “Black”, “Hispanic”, “Other”, “White”).

For this analysis, we first used crash date to obtain weekly counts of crashes. Weekly time series were the most appropriate interval for this study due to the within-week patterning in alcohol-involved crashes (e.g., alcohol-involved crashes are most common on weekend nights; National Highway Traffic Safety Administration National Center for Statistics and Analysis, 2011). Each week began Monday 12 a.m. through Sunday 11:59 p.m. The first week of January 2016 only included 3 days and was therefore excluded from the analysis; similarly, we excluded the last week of December 2021, which only had 5 days. This ensured that there was even distribution of weekends across the study sample. To calculate crashes per 100,000 population, we divided weekly counts of crashes by the population of California using the American Community Survey 1-year Estimates for California for each year included in the study. We used a similar approach when examining crash severity. For the analyses examining demographic subgroups, information on total population to serve as the denominator for the crashes per population estimates was similarly obtained from the American Community Survey 1-Year Estimates (United States Census Bureau, 2021), such as the percent of each racial/ethnic group in California, percent of each sex in California, and percent of each age category in California. For the racial categories, the denominator for White was the variable aggregating “White alone or in combination with one or more races,” Black was “Black or African American alone or in combination with one or more other races,” Asian was “Asian alone or in combination with one or more other races,” and Hispanic was “Hispanic or Latino Origin.” We did not examine associations among the Other category due to potential mismatch between how the CHP identifies Other and the categories included on the American Community Survey. Given the way California reports age in their Census summaries, we grouped variables into commonly used age groups into 0–19, 20–39, 40–59, and 60 years of age and over.

The second outcome we examined was the change in proportion of crashes that were alcohol involved. We filtered crashes using the alcohol involvement variable and divided the total crashes per week that were alcohol involved by the total number of crashes that occurred that week, obtaining values between 0 and 1.

The interruption variable was the date that shelter-in-place was ordered by the governor of California. A dummy variable was created for the week statewide shelter-in-place was ordered in California on March 19, 2020 (Procter, 2021). The variable was coded “0” for the time period before the given event and coded “1” for the time period after. For example, the proportion of crashes that were alcohol involved conducted on March 1, 2020, would fall in week 218 and be coded “0” for exposure to the pandemic, while the proportion of crashes that were alcohol involved on April 1, 2020 would fall in week 223 and be coded “1” for exposure to the pandemic.

Statistical analysis

We used an ARIMA interrupted time series to examine the impact of the statewide shelter-in-place order on alcohol-impaired crashes in California over 6 years. The time series included January 1, 2016 to December 31, 2021. There were 220 pre-interruption weeks and 94 post-interruption weeks.

Stationarity of a time series can be determined by examining the autocorrelation function (ACF) and the partial autocorrelation function (PACF) for each city (McDowell et al., 1980; Nau, 2020). We considered the time series to be stationary when the data is not dependent on the time in which the data is observed (Nau, 2020). Seasonal ARIMA models take the form of ARIMA(p,d,q)(P,D,Q)s. The P, D, and Q terms represent seasonal structural parameters while the p,d,q terms represent nonseasonal parameters (McDowell et al., 1980; Nau, 2020). In this study, the s represents the number of weeks in a year: 52. All ARIMA models were adjusted for seasonality to account for confounding due to the multiple years present in the dataset. Because the data spanned 6 years, differencing the data in the time series model created a dataset that does not trend nor drift (McDowell et al., 1980). To address persistent autocorrelation of data in each city, the models included nonseasonal structural parameters, where p is the autoregression term, d is the differencing term, and q is the moving average term for each city, applied only if necessary.

We used the “auto.arima” function in the R package forecast (Hyndman et al., 2023; Hyndman & Khandakar, 2008). The auto. arima function works to conduct various differencing techniques and then fit models according to the optimal p, q, P, and Q values. The function looks to choose the best fitting models by optimizing the Akaike Information Criterion, Corrected Akaike Information Criterion, or Bayesian Information Criterion. To ensure the most parsimonious fit, we also manually compared models and used a cutoff of 0.05 to determine whether a constant is included and the order of each of the integrated terms in the final ARIMA model. Last, we examined the distribution of model residuals using the Ljung-Box test to examine whether any temporal autocorrelation persisted in our time series (Burns, 2002).

After fitting an appropriate seasonal ARIMA model for each analysis where the residual series would be considered non-significant, the residuals can be considered background noise (McDowell et al., 1980). Seasonal ARIMA models were fitted to examine the impact on crashes per 100,000 population overall and separated by crash severity. Models also examined the impact of the statewide shelter-in-place order on the parties involved in crashes, separated by race/ethnic group, age, and sex. For all sub-analyses, we additionally examined the impact of the statewide shelter-in-place order on the proportion of crashes that were alcohol involved.

RESULTS

The dataset includes 314 weeks from January 2016 to December 2021. The total number of crashes included in this time period was 1,090,422 with 10.8% of those crashes having alcohol involvement (n = 117,999). Each week included in our dataset there was a mean of 3486 crashes (SD = 509) with a mean of 377 crashes being alcohol-involved. Before the initiation of the stay-at-home order there was a weekly mean of 9.53 crashes per 100,000 (SD = 0.86) and after the initiation of the stay-at-home order there was a weekly mean of 7.33 crashes per 100,000 population (SD = 1.17). Approximately 0.97 crashes per 100,000 (SD = 0.07) were alcohol-involved before the initiation of the stay-at-home order and 0.92 crashes per 100,000 (SD = 0.16) were alcohol-involved after. Weekly summary statistics are presented in Table 1 for both pre-interruption and post-interruption.

TABLE 1.

Weekly summary statistics of crashes and parties involved pre-interruption and post-interruption.

Variable Description January 1, 2016 - march 18, 2020
 March 19, 2020 - December 31,2021
Mean (n) Std. dev. Mean (n) Std. dev.
Crash level
 Crashes per 100,000 per week 9.53 0.86 7.33 1.17
 Alcohol involved crashes per 100,000 per week 0.98 0.08 0.92 0.16
 Severity Fatal 67.06 9.68 71.85 12.99
Injury (Severe) 246.89 37.94 274.91 47.03
Injury (Visible) 1120.26 140.80 1000.75 187.96
Injury (Complaint of Pain) 2287.30 233.36 1542.81 255.32
Party level
 Age 40.04 16.77 39.81 16.62
 Sex Female 2899.16 273.45 1958.67 407.32
Male 4327.86 334.38 3348.50 549.58
Nonbinary 4.31 2.49 5.45 3.59
 Race/ethnicity Asian 543.58 57.58 341.47 87.73
Black 689.72 65.7 586.48 90.07
Hispanic 2699.35 269.63 2205.84 393.93
White 2482.77 271.52 1628.5 291.05
Open in a new tab

Abbreviation: Std. Dev, Standard Deviation.

To determine the best fit for models, the R package auto.arima examined both AIC and Ljung Box Test results (Table 2). Overall crashes after the initiation of the stay-at-home order decreased by approximately 48% (−4.6 crashes per 100,000, 95% CI: −5.3, −3.9; Figure 1). However, the proportion of those crashes that were alcohol-involved increased by 2.3% (Table 3, Figure 2). When the crashes were separated by severity, the models found that crashes resulting in the least severe type of injury [injury (complaint of pain)] had the greatest decrease of 51% (−3.0 crashes per 100,000; 95% CI: −3.5, −2.5; Figure 3). The proportion of alcohol-involved crashes increased for the least two severe groups (injury [visible] and injury [complaint of pain]), while the proportion of alcohol-involved crashes resulting in fatal and severe injuries were unchanged (Figure 4).

TABLE 2.

Model parameters (p,d,q)×(P,D,Q)s for seasonal autoregressive integrated moving average models investigating the impact of COVID-19 on crashes and party involvement.

Model Parameter and fit Q* statistic Ljung-box test (p-value)
Crashes per 100,000 (2,1,0) × (1,0,0)52 38.427 0.9826
Alcohol-Involved Crash Proportion (1,1,2) × (0,0,2)52 40.351 0.9535
Crashes per 100,000 by Severity
 Fatal (1,0,1) × (0,0,0)52 57.659 0.5618
 Injury (Severe) (0,1,1) × (0,0,0)52 51.516 0.8014
 Injury (Visible) (3,1,2) × (1,1,0)52 73.733 0.05624
 Injury (Complaint of Pain) (2,1,0) × (1,0,0)52 41.125 0.963
Alcohol-involved crash proportion by severity
 Fatal (1,1,2) × (0,0,0)52 60.537 0.4201
 Injury (Severe) (1,0,0) × (2,0,0)52 64.737 0.2834
 Injury (Visible) (2,1,1) × (1,0,0)52 38.544 0.9771
 Injury (Complaint of Pain) (0,1,1) × (0,0,1)52 71.389 0.1491
Crashes per 100,000 by race
 Asian (0,1,1) × (0,0,1)52 49.423 0.8333
 Black (2,1,0) × (0,0,1)52 40.78 0.9661
 Hispanic (0,1,2) × (1,0,0)52 39.961 0.9728
 White (2,1,0) × (1,0,0)52 54.873 0.6283
 Other (3,1,0) × (1,0,2)52 40.758 0.9372
Alcohol-involved crash proportion by race
 Asian (1,1,3) × (0,0,0)52 59.855 0.4534
 Black (0,1,1) × (0,0,0)52 80.078 0.05122
 Hispanic (1,1,1) × (0,0,2)52 49.494 0.7791
 White (0,1,1) × (0,0,1)52 50.983 0.7901
Crashes per 100,000 by sex
 Females (0,1,1) × (1,0,0)52 59.382 0.4983
 Males (2,1,0) × (1,0,0)52 42.467 0.9485
Alcohol-involved crash proportion by sex
 Females (0,1,3) × (0,0,2)52 79.594 0.03114
 Males (0,1,1) × (0,0,0)52 53.57 0.7393
Crashes per 100,000 by age
 0–19 (0,1,1) × (1,0,0)52 57.577 0.5648
 20–39 (2,1,0) × (1,0,0)52 45.163 0.9077
 40–59 (0,1,2) × (1,0,0)52 34.825 0.9949
 60+ (0,1,1) × (0,0,2)52 56.218 0.5787
Alcohol-involved crash proportion by age
 0–19 (0,1,2) × (0,0,1)52 48.679 0.8289
 20–39 (1,1,1) × (0,0,2)52 51.032 0.7405
 40–59 (0,1,2) × (0,0,0)52 68.652 0.236
 60+ (0,1,1) × (0,0,0)52 77.96 0.1122
Open in a new tab

FIGURE 1.

FIGURE 1

Open in a new tab

Total crashes in California, 2016–2021.

TABLE 3.

Results of seasonal ARIMA models examining the impact of the initiation of COVID-19 stay-at-home orders on crash rate, proportion of crashes that were alcohol-involved, party involvement in crashes rate, and proportion of parties that were in alcohol-involved crashes.

Coefficient 95% CI
Crashes per 100,000 −4.563 −5.291 −3.835
Alcohol-involved crash proportion 0.023 0.019 0.028
Crashes per 100,000 by severity
 Fatal −0.011 −0.061 0.040
 Injury (Severe) −0.210 −0.310 −0.109
 Injury (Visible) −1.226 −1.477 −0.975
 Injury (Complaint of Pain) −3.015 −3.485 −2.544
Alcohol-involved crash proportion by severity
 Fatal 0.041 −0.005 0.087
 Injury (Severe) −0.007 −0.014 0.000
 Injury (Visible) 0.017 0.006 0.028
 Injury (Complaint of Pain) 0.012 0.007 0.017
Crashes per 100,000 by race
 Asian −4.998 −5.901 −4.095
 Black −10.890 −13.743 −8.037
 Hispanic −8.931 −10.442 −7.419
 Other −3.386 −4.131 −2.640
 White −5.112 −6.048 −4.175
Alcohol-involved crash proportion by race
 Asian 0.018 0.001 0.034
 Black 0.015 0.007 0.023
 Hispanic 0.019 0.013 0.026
 White 0.016 0.009 0.024
Crashes per 100,000 by sex
 Females −8.543 −9.738 −7.347
 Males −10.446 −12.362 −8.529
Alcohol-involved crash proportion by sex
 Females 0.018 0.013 0.023
 Males 0.019 0.011 0.026
Crashes per 100,000 by age
 0–19 −2.849 −3.466 −2.232
 20–39 −14.980 −17.414 −12.545
 40–59 −10.766 −12.457 −9.075
 60+ −8.042 −9.505 −6.579
Alcohol-involved crash proportion by age
 0–19 0.020 0.010 0.029
 20–39 0.019 0.011 0.027
 40–59 0.020 0.015 0.024
 60+ 0.008 0.004 0.013
Open in a new tab

FIGURE 2.

FIGURE 2

Open in a new tab

Proportion of crashes that were alcohol-involved from 2016 to 2021.

FIGURE 3.

FIGURE 3

Open in a new tab

Total crashes in California by crash severity, 2016–2021.

FIGURE 4.

FIGURE 4

Open in a new tab

Proportion of crashes that were alcohol-involved by crash severity, 2016–2021.

In the overall examination of the parties involved in crashes (Table 3), the statewide shelter-in-place order was significantly related to a decrease in crash rate among each racial group with the greatest decrease among parties who police offiers identified (California Highway Patrol, 2003) as Asian with 58% decrease (−5.0 crashes per 100,000, 95% CI: −5.9, −4.1) and who officers identified as White with a 51% decrease (−5.1 crashes per 100,000; 95% CI: −6.0, −4.2) and the lowest decreases among parties who officers identified as Black, a 43% decrease (−10.9 crashes per 100,000; 95% CI: −13.7, −8.0) and Hispanic, a 50% decrease (−8.9 crashes per 100,000; 95% CI: −10.4, −7.4; Figure S1).

However, among all racial/ethnic groups there was a uniform increase (approximately a 2% increase) in the proportion of crashes that were alcohol-involved. Males experienced a crash decrease of 47% while females experienced a crash decrease of 58% (Figure 5), but the proportion of alcohol-involved crashes was similar for both sexes (females: 0.02, 95% CI: 0.01, 0.02; males: 0.02, 95% CI: 0.01, 0.03; Figure 6). Although information on nonbinary parties were included in the data, the numbers were small and therefore we could not identify a stable, fitted ARIMA model to analyze results. In the examination of the ages of the parties involved in the crashes, while all age groups experienced a significant decrease, the group aged 0–19 experienced the greatest decrease of 57% (−2.9 crashes per 100,000, 95% CI: −3.5, −2.2) and the group 20–39 experienced the lowest decrease of 50% (−15.0 crashes per 100,000, 95% CI: −17.4, −12.5; Figure S2). All age groups experienced an approximately 2% increase in alcohol-involved crashes except the over 60 group, which had less than a 1% increase in alcohol-involved crashes (0.008 crashes per 100,000, 95% CI: 0.004, 0.01).

FIGURE 5.

FIGURE 5

Open in a new tab

Parties involved in crashes in California by sex, 2016–2021.

FIGURE 6.

FIGURE 6

Open in a new tab

Proportion of parties involved in crashes that were in an alcohol-involved crash by sex, 2016–2021.

Table 4 presents the ARIMA results for the parties that were declared at fault for the crash by the California Highway Patrol. The impact of the stay-at-home order had the biggest decrease for individuals who officers identified as Asian with a 51% decrease (−1.8 crashes per 100,000, 95% CI: −2.2, −1.4) and Hispanic with a 49% decrease (−4.0 crashes per 100,000, 95% CI: −4.7, −3.3) declared at-fault by the CHP. However, examining the proportion of crashes that were alcohol-involved for party-level data, the greatest increases are in individuals who officers identified as White and Hispanic. The relative decrease for parties at fault is less than the decrease found for all parties involved in crashes. In the examination of the sex of parties at fault for crashes, females have a larger decrease of 56% in the proportion of crashes that are alcohol involved (−3.5 crashes per 100,000, 95% CI: −4.1, −3.0) than males whose decrease was only 42% (−4.6 crashes per 100,000, 95% CI: −5.5, −3.7). There is a greater increase in the proportion of crashes that had alcohol-involvement post-interruption for both males and females that were at fault for crashes than the increase post-interruption when analyzing all parties involved in crashes.

TABLE 4.

Results of seasonal ARIMA models examining the impact of the initiation of the COVID-19 stay-at-home orders on AT FAULT party involvement in crashes rate, and proportion of parties that were in alcohol-involved crashes.

Coefficient 95% CI
Crashes per 100,000 by race
 Asian −1.789 −2.163 −1.416
 Black −4.468 −5.905 −3.031
 Hispanic −4.014 −4.725 −3.302
 White −2.295 −2.676 −1.914
Alcohol-involved crash proportion by race
 Asian 0.018 0.013 0.023
 Black 0.020 0.010 0.030
 Hispanic 0.026 0.017 0.034
 White 0.023 0.011 0.033
Crashes per 100,000 by sex
 Females −3.535 −4.064 −3.006
 Males −4.611 −5.524 −3.697
Alcohol-involved crash proportion by sex
 Females 0.034 0.022 0.045
 Males 0.021 0.011 0.031
Crashes per 100,000 by age
 0–19 −1.552 −2.053 −1.050
 20–39 −3.166 −3.583 −2.749
 40–59 −3.976 −4.713 −3.239
 60+ −3.440 −4.135 −2.745
Alcohol-involved crash proportion by age
 0–19 0.035 0.020 0.051
 20–39 0.024 0.013 0.036
 40–59 0.028 0.024 0.031
 60+ 0.007 0.002 0.012
Open in a new tab

DISCUSSION

Previous studies provide strong evidence that the start of the COVID-19 pandemic changed mobility patterns in the United States greatly (Aloi et al., 2020; de Palma et al., 2022; Engle et al., 2020; Jay et al., 2020), in addition to altering population-level alcohol consumption patterns (Biddle et al., 2020; Chodkiewicz et al., 2020; Grossman et al., 2020; Schmits & Glowacz, 2022). Given the potential implications of mobility changes and increased levels of heavy drinking, information on the quantifiable impact of alcohol-impaired driving was warranted. This interrupted time series analysis using data from California demonstrated a decrease in overall crashes per capita immediately following the initiation of the stay-at-home order in that state, with an eventual return to pre-pandemic levels. However, the proportion of crashes that were alcohol-involved increased significantly compared to the pre-pandemic rate and has remained elevated. With differing magnitude of effects in population subgroups in the analyses of parties involved in crashes, this study demonstrates the mixed impact of alcohol-involved driving in relation to the pandemic.

Overall, the findings of this study noting decreasing rates of overall crashes and increasing rates of alcohol-involved driving mostly support work conducted by researchers, federal agencies, and other organizations. Our findings corroborate findings from federal agencies that alcohol-impaired driving increased after the pandemic. NHSTA identified changes in certain “risky” behaviors like increased alcohol-impaired driving, lower seat belt use, higher speeds, and cell phone distraction associated with the pandemic (National Highway Traffic Safety Administration National Center for Statistics and Analysis, 2021; National Highway Traffic Safety Administration Office of Behavioral Safety Research, 2021). Further, NHTSA also found varying patterns of these behaviors when examining subgroups. For example, most ejection rates associated with lower seatbelt use were among men and those aged 18–49. It appears that alcohol-impaired driving may potentially be related changes in overall risk aversion patterns and warrants further investigation. When examining changes in the rates of severe crashes based on people treated at the sites of motor vehicle crashes by Emergency Medical Services, researchers found an increase in the most severe of crashes (people with <36.1% probability of survival) from 2019 through 2020–2021 (National Highway Traffic Safety Administration National Center for Statistics and Analysis, 2021). Our study detected changes in crash severity, however, mostly in the least severe crashes. The National Institute on Alcohol Abuse and Alcoholism examined the change in per capita sales of ethanol in March 2020 compared to March 2017–2019 in 13 states and found certain changes as high as 25% in some states (National Institute on Alcohol Abuse and Alcoholism (NIAAA) Division of Epidemiology and Prevention, 2022). As alcohol sales data have been used as an proxy for alcohol consumption at home in previous studies, (Castaldelli-Maia et al., 2021; Foster & Ferguson, 2012), and alcohol sales increased in the period following many stay-at-home orders (Castaldelli-Maia et al., 2021), this may support theories of greater alcohol consumption once the pandemic began and could indicate the reason for increases in alcohol-involved crashing.

Some studies did not produce results as clear as many of the federal agency reports that the COVID-19 pandemic increased alcohol-impaired driving. Watson-B rown et al., in an individual-level study using three surveys conducted across Queensland, Australia found decreases in alcohol-impaired driving, but the likelihood of driving while impaired was heavily dependent on previous experiences driving after drinking (Watson-Brown et al., 2021). Another study in Bucharest retrospectively analyzed blood alcohol concentration data and found no statistically significant difference in overall blood alcohol concentration from tests before and during the pandemic (Hostiuc et al., 2021) Methodological limitations in study designs and a different COVID-19 and drinking contexts than that of the United States likely contributed to different findings (Hostiuc et al., 2021; Watson-Brown et al., 2021). An interrupted time series is an overall stronger study design than many previous studies given its ability to control trends in crash rate and alcohol-impaired driving that were taking place before the start of the stay-at-home orders (Soumerai et al., 2015). Controlling for trends is vital to quantify the independent effect of the event.

Our findings of demographic subgroup differences in decreases in crash rates and increases in the proportion of crashes that were alcohol-involved are also supported by previous literature. There is evidence that reduced mobility due to COVID-19 stay-at-home orders was not uniform across demographic groups given that physical distancing was lower in low-income neighborhoods than in high-income neighborhoods (e.g., differences in work-from-home ability; Jay et al., 2020). If this is the case, then the differences in crash rate decreases in racial/ethnic subgroups could be attributed to the inequitable distribution of social distancing ability. Another study in Queensland, Australia, found that younger, male drivers were more likely to increase driving while alcohol-impaired and found that a group of individuals who drove alcohol-impaired for the first time likely emerged during the period of COVID-19 lockdowns (Watson-Brown et al., 2021), potentially supporting some of the gender differences found in the study described in this paper. It is also important to further analyze the types of vulnerable road users affected by the changes in crash distribution and alcohol involvement. The Governors Highway Safety Association projected an increase in pedestrian fatalities even with the decrease in vehicle miles traveled (Governors Highway Safety Association, 2021), which may explain differences we found in the effect estimates examining all parties involved in crashes compared to the parties that were at fault.

The findings must be interpreted with some limitations in mind. First, although the interrupted time series methods used in these analyses account for temporal autocorrelation and seasonal trends, a limitation is that a comparison group would limit selection bias and confounding due to between-group differences (Bernal et al., 2016). However, the massive interruption in 2020 was essentially universal, stay-at-home orders occurred in some form across the United States and adding a control group is unlikely to reduce the effect sizes found. It also remains unclear whether the denominators used to capture total population in each of the subgroups from the Census data matches the definitions of those subgroups using California Highway Patrol data. This is mostly because while race is self-identified on the Census, the California Highway Patrol makes the determination on an individual’s race/ethnicity. The process of how the California Highway Patrol does this is unclear and therefore, increases the likelihood of a mismatch between how an individual would identify and how the California Highway Patrol documents an individual’s race/ethnicity. Related, we had to remove the “Other” group from this analysis because it is unknown what categories would be in the “Other” category by the California Highway Patrol and “Other” category of the Census. Since we do not know the validity of the racial/ethnic subgroups, it is difficult to be certain if the denominators used were correct. The California Highway Patrol does not indicate a separate category for American Indian and Alaska Native or Native Hawaiian and Other Pacific Islander. We assume in this study that the CHP labeled those who would be categorized as “Asian alone or in combination with one or more races” in the American Community Survey is the same group as those the CHP identify as Asian. These are hefty assumptions, which ultimately draws attention to the tremendous problem of the CHP identifying individuals into one of their five predetermined groups. In addition, previous research has demonstrated significant social inequalities in the distribution of crashes among neighborhoods of various socioeconomic statuses and demographic compositions (Barajas, 2018; Morency et al., 2012). Understanding the impact of the stay-at-home orders on alcohol-involved crashes by neighborhood socioeconomic status would allow local and state governments to allocate resources to reduce alcohol-impaired driving to those specific areas. However, given limitations of the data available, it is not possible to denominate the traffic flow in each of these neighborhoods to obtain the number of crashes per population. Lastly, although California’s population makes up about 12% of the United States population, policies related to the COVID-19 pandemic varied widely across the nation. Research in other states with available crash data can inform whether the changes found in this study are generalizable to the entire country.

CONCLUSION

This study emphasizes that although the initiation of a stay-at-home order in CA may have contributed to lower overall crash incidence, alcohol-involved crashes increased and have remained at these increased levels. This finding underscores the importance of research and policy to reduce alcohol-impaired driving and crashing.

Supplementary Material

Supplementary Fig 2
Supplementary Fig 1

FUNDING INFORMATION

This work was supported, in part, by the following grants: NIH/NIAAA R01AA029112 and K01AA026327 and NIH/NIDA T32DA031099. The findings and conclusions in this presentation are those of the authors and do not necessarily represent the views of the NIH.

Footnotes

CONFLICT OF INTEREST STATEMENT

The authors report no conflicts of interest.

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

REFERENCES

  1. Aloi A, Alonso B, Benavente J, Cordera R, Echániz E, González F et al. (2020) Effects of the COVID-19 lockdown on urban mobility: empirical evidence from the City of Santander (Spain). Sustainability, 12, 3870. Available from: 10.3390/su12093870 [DOI] [Google Scholar]
  2. Barajas JM (2018) Not all crashes are created equal: associations between the built environment and disparities in bicycle collisions. Journal of Transport and Land Use, 11, 865–8 82. Available from: 10.5198/jtlu.2018.1145 [DOI] [Google Scholar]
  3. Bernal JL, Cummins S & Gasparrini A (2016) Interrupted time series regression for the evaluation of public health interventions: a tutorial. International Journal of Epidemiology, 46, 348–355. Available from: 10.1093/ije/dyw098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Biddle N, Edwards B, Gray M & Sollis K (2020) Alcohol consumption during the COVID-19 period: May 2020. Canberra: Australian National University. [Google Scholar]
  5. Bonal X & González S (2020) The impact of lockdown on the learning gap: family and school divisions in times of crisis. International Review of Education, 66, 635–655. Available from: 10.1007/s11159-020-09860-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Borjas GJ (2020) Business closures, stay-at-home restrictions, and COVID-19 testing outcomes in New York City. Preventing Chronic Disease, 17, E109. Available from: 10.5888/pcd17.200264 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bourassa KJ (2021) State-level stay-at-home orders and objectively measured movement in the United States during the COVID-19 pandemic. Psychosomatic Medicine, 83, 358–362. Available from: 10.1097/PSY.0000000000000905 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Burns P (2002) Robustness of the Ljung-Box test and its rank equivalent. SSRN Electronic Journal. Available from: 10.2139/ssrn.443560 [DOI] [Google Scholar]
  9. California Highway Patrol. (2003) Collision investigation manual. United States Department of Transportation National Highway Traffic Safety Administration. https://www.nhtsa.gov/sites/nhtsa.gov/files/documents/ca_chp555_manual_2_2003_ch1-13.pdf [Google Scholar]
  10. California Highway Patrol. (2023) SWITRS –Statewide Integrated Traffic Records System. Available from: https://www.chp.ca.gov/programs-services/services-information/switrs-internet-statewide-integrated-traffic-records-system [Accessed 11th March 2023]. [Google Scholar]
  11. Carmassi C, Foghi C, Dell’Oste V, Cordone A, Bertelloni CA, Bui E et al. (2020) PTSD symptoms in healthcare workers facing the three coronavirus outbreaks: what can we expect after the COVID-19 pandemic. Psychiatry Research, 292, 113312. Available from: 10.1016/j.psychres.2020.113312 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Castaldelli-Maia JM, Segura LE & Martins SS (2021) The concerning increasing trend of alcohol beverage sales in the U.S. during the COVID-19 pandemic. Alcohol, 96, 37–42. Available from: 10.1016/j.alcohol.2021.06.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Centers for Disease Control and Prevention National Center for Injury Prevention and Control. (2021) Web-based Injury Statistics Query and Reporting System (WISQARS). Available from: https://www.cdc.gov/injury/wisqars/index.html [Accessed 21st December 2022]. [Google Scholar]
  14. Chodkiewicz J, Talarowska M, Miniszewska J, Nawrocka N & Bilinski P (2020) Alcohol consumption reported during the COVID-19 pandemic: the initial stage. International Journal of Environmental Research and Public Health, 17, 4677. Available from: 10.3390/ijerph17134677 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Czeisler MÉ, Lane RI, Petrosky E, Wiley JF, Christensen A, Njai R et al. (2020) Mental health, substance use, and suicidal ideation during the COVID-19 pandemic –United States, June 24–30, 2020. MMWR. Morbidity and Mortality Weekly Report, 69, 1049–1057. Available from: 10.15585/mmwr.mm6932a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. de Palma A, Vosough S & Liao F (2022) An overview of effects of COVID-19 on mobility and lifestyle: 18 months since the outbreak. Transportation Research Part A: Policy and Practice, 159, 372–397. Available from: 10.1016/j.tra.2022.03.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Dong X, Xie K & Yang H (2022) How did COVID-19 impact driving behaviors and crash severity? A multigroup structural equation modeling. Accident; Analysis and Prevention, 172, 106687. Available from: 10.1016/j.aap.2022.106687 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Doucette ML, Tucker A, Auguste ME, Watkins A, Green C, Pereira FE et al. (2021) Initial impact of COVID-19’s stay-at-home order on motor vehicle traffic and crash patterns in Connecticut: an interrupted time series analysis. Injury Prevention, 27, 3–9. Available from: 10.1136/injuryprev-2020-043945 [DOI] [PubMed] [Google Scholar]
  19. Engle S, Stromme J & Zhou A (2020) Staying at home: mobility effects of COVID-19. SSRN Electronic Journal. Available from: 10.2139/ssrn.3565703 [DOI] [Google Scholar]
  20. Foster JH & Ferguson CS (2012) Home drinking in the UK: trends and causes. Alcohol and Alcoholism, 47, 355–358. Available from: 10.1093/alcalc/ags020 [DOI] [PubMed] [Google Scholar]
  21. Fowler JH, Hill SJ, Levin R & Obradovich N (2021) Stay-at-home orders associate with subsequent decreases in COVID-19 cases and fatalities in the United States. PLoS One, 16, e0248849. Available from: 10.1371/journal.pone.0248849 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Governors Highway Safety Association. (2021) Pedestrian traffic fatalities by state: 2020 preliminary addendum. In: Spotlight on highway safety. Washington, DC: Governors Highway Safety Association. [Google Scholar]
  23. Grossman ER, Benjamin-Neelon SE & Sonnenschein S (2020) Alcohol consumption during the COVID-19 pandemic: a cross-sectional survey of US adults. International Journal of Environmental Research and Public Health, 17, 9189. Available from: 10.3390/ijerph17249189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Grossman ER, Benjamin-Neelon SE & Sonnenschein S (2022) Alcohol consumption and alcohol home delivery laws during the COVID-19 pandemic. Substance Abuse, 43, 1141–1146. Available from: 10.1080/08897077.2022.2060432 [DOI] [PubMed] [Google Scholar]
  25. Hingson RW, Zha W & White AM (2017) Drinking beyond the binge threshold: predictors, consequences, and changes in the U.S. American Journal of Preventive Medicine, 52, 717–727. Available from: 10.1016/j.amepre.2017.02.014 [DOI] [PubMed] [Google Scholar]
  26. Hostiuc S, Radu D, Seretean L, Tirdea C, Siminiuc R & Curcă GC (2021) Driving under the influence of alcohol during the COVID-19 pandemic. Forensic Science International, 329, 111076. Available from: 10.1016/j.forsciint.2021.111076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hu Y, Quigley BM & Taylor D (2021) Human mobility data and machine learning reveal geographic differences in alcohol sales and alcohol outlet visits across U.S. states during COVID-19. PLoS One, 16, e0255757. Available from: 10.1371/journal.pone.0255757 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hyndman R, Athanasopoulos G, Bergmeir C, Caceres G, Chhay L, O’Hara-Wild M et al. (2023) forecast: Forecasting functions for time series and linear models. R package version 8.21, https://pkg.robjhyndman.com/forecast/ [Google Scholar]
  29. Hyndman R & Khandakar Y (2008) Automatic time series forecasting: the forecast package for R. Journal of Statistical Software, 26, 1–22. Available from: 10.18637/jss.v027.i0319777145 [DOI] [Google Scholar]
  30. Jabbari P & MacKenzie D (2020) Ride sharing attitudes before and during the COVID-19 pandemic in the United States. Findings Available from: 10.32866/001c.17991 [DOI] [Google Scholar]
  31. Jay J, Bor J, Nsoesie EO, Lipson SK, Jones DK, Galea S et al. (2020) Neighbourhood income and physical distancing during the COVID-19 pandemic in the United States. Nature Human Behaviour, 4, 1294–1302. Available from: 10.1038/s41562-020-00998-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kontopantelis E, Doran T, Springate DA, Buchan I & Reeves D (2015) Regression based quasi-experimental approach when randomisation is not an option: interrupted time series analysis. BMJ, 350, h2750. Available from: 10.1136/bmj.h2750 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lee A (2020) These states have implemented stay-at-home orders. Here’s what that means for you. CNN. Available from: https://www.cnn.com/2020/03/23/us/coronavirus-which-states-stay-at-home-order-trnd/index.html [Accessed 13 March 2023]. [Google Scholar]
  34. Liu CH, Erdei C & Mittal L (2021) Risk factors for depression, anxiety, and PTSD symptoms in perinatal women during the COVID-19 pandemic. Psychiatry Research, 295, 113552. Available from: 10.1016/j.psychres.2020.113552 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Marques de Miranda D, da Silva Athanasio B, Sena Oliveira AC & Simoes-e-Silva AC (2020) How is COVID-19 pandemic impacting mental health of children and adolescents? International Journal of Disaster Risk Reduction, 51, 101845. Available from: 10.1016/j.ijdrr.2020.101845 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McDowell D, McCleary R, Meidinger E & Hay RA Jr. (1980) Quantitative applications in the social sciences: interrupted time series analysis. Thousand Oaks, CA: SAGE Publications, Inc. [Google Scholar]
  37. Mohan A, Fitzgerald N, Uny I & Begley A (2020) Drink Drive–taking stock, moving forward. Parliamentary Advisory Council for Transport Safety. https://www.pacts.org.uk/news-and-publications/pacts-report-drinkdriving-taking-stock-moving-forward/ [Google Scholar]
  38. Moreland A (2020) Timing of state and territorial COVID-19 stay-at-home orders and changes in population movement — United States, March 1–May 31, 2020, MMWR Morb. Mortal. Wkly. Rep, 69, 1198–1203. Available from: 10.15585/mmwr.mm6935a2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Morency P, Gauvin L, Plante C, Fournier M & Morency C (2012) Neighborhood social inequalities in road traffic injuries: the influence of traffic volume and road design. American Journal of Public Health, 102, 1112–1119. Available from: 10.2105/AJPH.2011.300528 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Naimi TS, Nelson DE & Brewer RD (2009) Driving after binge drinking. American Journal of Preventive Medicine, 37, 314–320. Available from: 10.1016/j.amepre.2009.06.013 [DOI] [PubMed] [Google Scholar]
  41. Naimi TS, Xuan Z, Sarda V, Hadland SE, Lira MC, Swahn MH et al. (2018) Association of state alcohol policies with alcohol-related motor vehicle crash fatalities among US adults. JAMA Internal Medicine, 178, 894–901. Available from: 10.1001/jamainternmed.2018.1406 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. National Academies of Sciences, Engineering, and Medicine, Health and Medicine Division, Board on Population Health and Public Health Practice, Committee on Accelerating Progress to Reduce Alcohol-Impaired Driving Fatalities, Negussie Y, Geller A et al. (2018) Current environment: alcohol, driving, and drinking and driving, getting to zero alcohol-impaired driving fatalities: a comprehensive approach to a persistent problem. Washington, DC: National Academies Press (US). [PubMed] [Google Scholar]
  43. National Highway Traffic Safety Administration National Center for Statistics and Analysis. (2011) Time of day and demographic perspective of fatal alcohol-impaired-driving crashes (No. DOT HS 811 523). Traffic safety facts. Washington, DC: National Highway Traffic Safety Administration. [Google Scholar]
  44. National Highway Traffic Safety Administration National Center for Statistics and Analysis. (2017) Alcohol-impaired driving (No. DOT HS 812 450), traffic safety facts. Washington, DC: National Highway Traffic Safety Administration. [Google Scholar]
  45. National Highway Traffic Safety Administration National Center for Statistics and Analysis. (2021) Continuation of research on traffic safety during the COVID-19 public health emergency: January — June 2021 (behavioral safety research No. DOT HS 813 210), traffic safety facts. Washington, DC: National Highway Traffic Safety Administration. [Google Scholar]
  46. National Highway Traffic Safety Administration National Center for Statistics and Analysis. (2022) Alcohol-impaired driving (No. DOT HS 813 294), traffic safety facts. Washington, DC: National Highway Traffic Safety Administration. [Google Scholar]
  47. National Highway Traffic Safety Administration Office of Behavioral Safety Research. (2021) Update to special reports on traffic safety during the COVID-19 public health emergency: fourth quarter data (No. DOT HS 813 135), traffic safety facts. Washington, DC: National Highway Traffic Safety Administration. [Google Scholar]
  48. National Institute on Alcohol Abuse and Alcoholism Division of Epidemiology and Prevention Research. (2022) Alcohol Sales During the COVID-19 Pandemic (Surveillance Report COVID-19). National Institute on Alcohol Abuse and Alcoholism. https://pubs.niaaa.nih.gov/publications/surveillance-covid-19/COVSALES.htm#fig20 [Google Scholar]
  49. Nau R (2020) Statistical forecasting: notes on regression and time series analysis. Available from: https://people.duke.edu/~rnau/411home.htm [Accessed 24th April 2020]. [Google Scholar]
  50. Ornell F, Moura HF, Scherer JN, Pechansky F, Kessler FHP & von Diemen L (2020) The COVID-19 pandemic and its impact on substance use: implications for prevention and treatment. Psychiatry Research, 289, 113096. Available from: 10.1016/j.psychres.2020.113096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Pollard MS, Tucker JS & Green HD Jr. (2020) Changes in adult alcohol use and consequences during the COVID-19 pandemic in the US. JAMA Network Open, 3, e2022942. Available from: 10.1001/jamanetworkopen.2020.22942 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Procter R (2021) Remember when? Timeline marks key events in California’s year-long pandemic grind. CalMatters. [Google Scholar]
  53. Sahraei MA, Kuşkapan E & Çodur MY (2021) Public transit usage and air quality index during the COVID-19 lockdown. Journal of Environmental Management, 286, 112166. Available from: 10.1016/j.jenvman.2021.112166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Schmits E & Glowacz F (2022) Changes in alcohol use during the COVID-19 pandemic: impact of the lockdown conditions and mental health factors. International Journal of Mental Health and Addiction, 20, 1147–1158. Available from: 10.1007/s11469-020-00432-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Soumerai SB, Starr D & Majumdar SR (2015) How do you know which health care effectiveness research you can trust? A guide to study design for the perplexed. Preventing Chronic Disease, 12, E101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Stanistreet P, Elfert M & Atchoarena D (2020) Education in the age of COVID-19: understanding the consequences. International Review of Education, 66, 627–633. Available from: 10.1007/s11159-020-09880-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Tefft BC, Villavicencio L, Benson A, Arnold L, Kim W, Añorve V et al. (2022) Self-reported risky driving in relation to amount of driving during the COVID-19 pandemic. Washington, DC: AAA Foundation for Traffic Safety. [Google Scholar]
  58. Safe Transportation Research and Education Center, University of California, Berkeley. (2022) Transportation Injury Mapping System (TIMS). https://tims.berkeley.edu/ [Google Scholar]
  59. United States Census Bureau. (2021) American community survey 1-year estimates. Census Report. Profile Page Calif. Available from: http://censusreporter.org/profiles/04000US06-california/ [Accessed 24th October 2022]. [Google Scholar]
  60. Wagenaar AC, Tobler AL & Komro KA (2010) Effects of alcohol tax and Price policies on morbidity and mortality: a systematic review. American Journal of Public Health, 100, 2270–2278. Available from: 10.2105/AJPH.2009.186007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Wang W, Miao W, Liu Y, Deng Y & Cao Y (2022) The impact of COVID-19 on the ride-sharing industry and its recovery: causal evidence from China. Transportation Research Part A: Policy and Practice, 155, 128–141. Available from: 10.1016/j.tra.2021.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Watson-Brown N, Truelove V, Parker E & Davey J (2021) Drink driving during the COVID-19 pandemic. Transportation Research Part F: Traffic Psychology and Behaviour, 78, 369–380. Available from: 10.1016/j.trf.2021.02.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. World Health Organization. (2019) The SAFER technical package: five areas of intervention at national and subnational levels. (No. Licence: CC BY-NC-SA 3.0 IGO). Geneva: World Health Organization. [Google Scholar]
  64. Xuan Z, Blanchette JG, Nelson TF, Heeren TC, Nguyen TH & Naimi TS (2015) Alcohol policies and impaired driving in the United States: effects of driving-vs. drinking-oriented policies. International Journal of Alcohol and Drug Research, 4, 119–130. Available from: 10.7895/ijadr.v4i2.205 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Yasin YJ, Grivna M & Abu-Zidan FM (2021) Global impact of COVID-19 pandemic on road traffic collisions. World Journal of Emergency Surgery, 16, 51. Available from: 10.1186/s13017-021-00395-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Yellman MA (2022) Motor vehicle crash deaths — United States and 28 other high-income countries, 2015 and 2019. MMWR. Morbidity and Mortality Weekly Report, 71, 837–8 43. Available from: 10.15585/mmwr.mm7126a1 [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Fig 2
NIHMS1960075-supplement-Supplementary_Fig_2.docx (95.7KB, docx)
Supplementary Fig 1
NIHMS1960075-supplement-Supplementary_Fig_1.docx (93KB, docx)

RESOURCES