Longitudinal Associations of 12th-Grade Binge Drinking With Risky Driving and High-Risk Drinking

Data were from all 7 waves (W1 to W7) of the NEXT Generation Health Study, a nationally representative longitudinal study starting in 10th grade (2009–2010 year [W1]; sampling details previously published).^15,24  From W1 to W7, 91%, 88%, 86%, 78%, 79%, 84%, and 83% of the full sample (N = 2785) completed the survey during the spring time each year. Parent consent and participant assent were obtained for those under age 18. After turning age 18, participants were consented as adults. African American participants were oversampled. The study protocol was approved by the Institutional Review Board of the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

DWI was assessed with 3 items asking participants how many days they drove after drinking alcohol or using marijuana or illicit drugs (past 30 days). Questions were collapsed and dichotomized as DWI ≥1 day versus no DWI (past 30 days).

RWI was measured by asking participants, “During the last 12 months, how many times did you ride in a vehicle driven by someone who had been drinking alcohol?” The same question was repeated for “smoking marijuana” and “using illicit drugs other than alcohol or marijuana.” Questions were collapsed and dichotomized as RWI ≥1 time versus no RWI (last 12 months).

Blackout was measured by asking participants, “In the last 6 months, how often has your drinking caused you to black out?” with 5 options: not at all, once, twice, 3 times, and 4+ times. Options were dichotomized as blackout ≥1 time versus no blackout (last 6 months).

Extreme binge drinking was measured with an open-ended question, “During the past 12 months, what was the largest number of drinks that you drank on a single day?” with 2 options: either “write in number of drinks” or check “I do not drink.” Answers were recoded as a 5-level categorical variable including no drinking, drinking but no binge drinking (males: 1 to 4; females: 1–3 drinks), binge (males: 5–9; females: 4–7 drinks), 2 times binge drinking (males: 10–14; females: 8–11 drinks), and 3 times binge drinking or more (males: 15+; females: 12+ drinks) in the past 12 months.²⁵ 

Risky driving was measured by using 21 questions derived from the validated self-reported Checkpoints Risky Driving Scale (C-RDS)^26–28  to measure unsafe driving behavior (eg, on how many days in the last 30 days have you exceeded the speed limit in residential or school zones?). Internal consistency of the C-RDS was good (Cronbach α = 0.90). We dichotomized responses on each of the 21 questions (1 = at least 1 day; 0 = none) and summed them to make a continuous variable, with possible scores ranging from 0 to 21.

Binge drinking was measured with 1 question from the Monitoring the Future national survey²⁹ : “Over the last 30 days, how many times (if any) have you had 4 (for females), 5 (for males), or more drinks in a row within 2 hours?” Response options ranged from 1 (none) to 6 (≥10 times). Scores were dichotomized: binge ≥1 time versus no binge. We used binge drinking reported in 12th grade (W3) to predict outcome variables in W4 to W7.

To explore dose-response effects on the association between binge drinking and outcome variables in W4 to W7, the sum of high school binge drinking in W1 to W3 (grades 10–12) was created by summing the 3 dichotomous variables (ie, binge ≥1 time versus no binge in W1, W2, and W3), creating a 4-level categorical variable (none, once, twice, or 3 times in the last 30 days). This sum reflects the chronicity of binge drinking while in high school.

Adolescents reported perceptions of their father’s and mother’s monitoring knowledge (separate items) from a 5-item scale³⁰  including who their friends were, how they spent their money, what they did with their free time, where they were after school, and where they went at night, with 4 response options (1 = do not have or see father, mother, or guardian; 2 = he or she does not know anything; 3 = he or she knows a little; and 4 = he or she knows a lot). The means of these 5 items were calculated for each wave. Because monitoring knowledge in W1 to W3 is highly correlated, we calculated a grand mean of father’s and mother’s monitoring knowledge separately across W1 to W3.

One question derived from the National Survey on Drug Use and Health³¹  asked participants how important it was to their parents or guardians that they do not use alcohol (response options: 1 = not at all; 7 = extremely) to measure student-perceived parental support of not using alcohol. Like parental monitoring knowledge, the grand mean of father’s and mother’s support of not using alcohol was calculated from W1 to W3.

Driving licensure (domain variable) was generated from students reporting if they had a license allowing independent, unsupervised driving and used in the domain analyses with DWI as the outcome variable.

Participants reported age in W1 (mean = 16.3 years; SE = 0.03), sex, race and/or ethnicity, family affluence, and residence at W4 to W7; 1 parent provided the parent education levels and participant’s living areas in W1 (variable details in the Table 1 legend).

Our analyses considered features of complex survey design (ie, stratification, clustering, and sampling weights) with SAS 9.4 (SAS Institute, Inc, Cary, NC). The measures in W1 to W7 were assessed from the 2010–2011 through the 2016–2017 school years. Binary logistic regression was used for DWI, RWI, and blackout; linear regression was used for C-RDS; and ordinal logistic regression was used for extreme binge drinking. To select the variables that would be included in adjusted models, bivariate association of each outcome variable in all 4 waves with any of the potential covariates was examined. The variables that are associated (P = .10 level) with any outcome variable were included in adjusted models. Unadjusted (Supplemental Tables 7–11) and adjusted model analyses were conducted. Missing data were listwise deleted in the analysis for each wave, and domain analysis was applied for analyses when using the driving subsample.

Demographic and outcome variable information in W4 to W7 is provided in Table 1. Binge drinking prevalence in W1 to W3 (27.2%, 23.8%, and 26.8%, respectively) was calculated. Paired t tests (Bonferroni-based adjustment; significance at P = .02) indicate no significant difference between W1 and W3 (P > .05) or W2 and W3 (P = .03). Thus, we used 12th-grade (W3) binge drinking to indicate purposeful versus early-experimental drinking to get drunk.

Table 2 shows that binge drinking in W3 was associated with higher odds of DWI in W4 to W7. Of note, in W5, the odds were nearly 6 times greater, and in W7, the odds were still >2 times greater. Compared with adolescents who reported never binging in high school, those who binged in more years during high school had higher odds of DWI at later waves. Furthermore, although binging in at least 2 high school years was associated with higher odds of DWI in all later waves, a dose-response relationship was not observed after Bonferroni correction (Supplemental Table 12).

Binge drinking in W3 was associated with higher odds of RWI, with odds ranging from >4 times in W4 to >2.5 times in W7 (Table 3). Compared with adolescents who reported never binge drinking while in high school, those who binged at least once while in high school had higher odds of RWI in W4 and W5. Those who binged in at least 2 high school years had higher odds of RWI in all subsequent waves. However, a dose-response relationship was not observed after Bonferroni correction (Supplemental Table 12).

High father monitoring knowledge was related to lower odds of DWI and RWI, decreasing the odds of DWI in W5 and W6 by 30% and 20%, respectively, and decreasing the odds of RWI in W4 and W7 by 20% and 20%, respectively. High mother monitoring knowledge was also related to DWI in W4, lowering the odds by 60%, but the odds were not lowered in any of the waves for RWI. Parental support for not drinking alcohol was related to DWI, lowering the odds in W4, W5, and W7 by 40%, 40%, and 30%, respectively, and similarly lowering the odds of RWI in W4 by 20%.

As shown in Table 4, binge drinking in W3 was associated with ≥2.5 times greater odds of blackout throughout W4 to W7. Compared with adolescents who report never binge drinking while in high school, those who reported binge drinking in at least 3 high school years had higher odds of blackout in all subsequent waves. Moreover, those who reported binge drinking in at least 2 high school years, compared with those who reported never binge drinking while in high school, had higher odds of blackout in W5 and W7. Finally, those reporting binge drinking in at least 1 high school year, compared with those who reported never binge drinking while in high school, had higher odds of blackout in W4 and W5.

Binge drinking in W3 was associated with 2 times greater odds of W7 extreme binge drinking (Table 5). Compared with those who never binged, adolescents who reported binging in at least 2 high school years had higher odds of extreme binge drinking at W7.

There was no association between father’s monitoring knowledge and blackout (Table 4). However, high mother monitoring knowledge was associated with 30% decreased odds of blackout in W5 (Table 4), whereas high parental support for not using alcohol reduced the odds of blackout by 20% in both W4 and W5 (Table 4).

Father’s monitoring knowledge, mother’s monitoring knowledge, and parental support of not using alcohol were not associated with W7 extreme binge drinking (Table 5).

Binge drinking in W3 was positively and significantly associated with W4–W7 risky driving (ie, C-RDS scores; Table 6). Compared with those who never binged, adolescents who binged in at least 2 years of high school had higher C-RDS scores in W4, and those reporting binge drinking in at least 3 years of high school had higher C-RDS scores for W4, W5, and W7. Mother’s monitoring knowledge was associated with lower C-RDS scores only in W4. Otherwise, all other parental monitoring knowledge findings and parental support for not using alcohol were not associated with C-RDS score.

Our study examined prospective associations of 12th-grade binge drinking with later risky driving (ie, C-RDS) and high-risk drinking in a nationally representative sample of emerging adults. The timing and magnitudes of these associations reflect several important transitions in emerging adults, including developmental transitions (eg, adolescence into emerging adulthood), environmental transitions (eg, high school, then out of high school), policy context transitions (eg, underage drinking into minimum legal-aged drinking), and youth-parent relational transitions (eg, parental-dependent stage into more independent stage). Our analyses show that 12th-grade binge drinking is associated with a high likelihood of DWI, RWI, blackouts, engaging in more frequent risky driving up to 4 years after high school, and extreme binge drinking 4 years after high school. This is consistent with previous risky-driving research among binge drinking adolescents.⁸  Furthermore, adolescents who binged in at least 3 high school years, compared with those who did not, were more likely to DWI, RWI, blackout, be involved in extreme binge drinking, and report more risky driving. Parental practices appeared to have some enduring effect in protecting emerging adults against DWI, RWI, and blackouts in some but not all waves.

An association between early exposure to substance use and later risk behaviors among youth has been well established.^13,34  Our findings incrementally and importantly advance the understanding of a similar relationship for the extent of chronicity of high school binge drinking in a longitudinal context. Moreover, our results indicate that for those adolescents who do not abstain from alcohol use in high school, reduced frequency and in particular intensity of alcohol use (ie, avoidance of binge drinking and lower chronicity of binge drinking while in high school) can decrease the magnitude of much later major health-risk behaviors. This could reduce major harm due to alcohol-related adverse health outcomes (ie, alcohol-related DWI and/or RWI injury or fatal crash).

Study findings on lasting effects of parental influence on risk behaviors when adolescents transition into emerging adulthood have been somewhat mixed, with more evidence observed for female than male adolescents.¹⁶  However, several studies report that facets of good parenting (eg, parental monitoring knowledge and parental trust and support) provide important protective effects against adolescent risk behaviors, including in the context of DWI and RWI.^15,17,35,36  Our study shows limited but important parenting influence on DWI, RWI, and to some extent blackouts. Although the relationships are not consistent across all assessment waves and all outcomes, there is evidence of positive influence through parenting practices that may last 2 or more years after the adolescent leaves high school.

We recognize that our study has limitations. First, although the NEXT study was longitudinal, we chose to conduct a series of 2–time-point prospective analyses rather than multiple–time-point prospective analyses because of the wave-dependent changing sample size (ie, waves with a sequentially increasing number of participants who were eligible for driver licensure who could then be included in analyses). The changing sample sizes were not consistent across waves, especially for driving-related variables (ie, DWI and C-RDS). Furthermore, only 1 wave of data for extreme binge drinking is available in the NEXT study data set. Therefore, for consistency, we used a longitudinal approach and variables in earlier waves to predict outcomes at later waves. Second, we examined dichotomous outcomes of DWI, RWI, and blackouts because of sparse and skewed responses. As such, the transition probabilities and their correlates are only for “no” DWI, RWI, or blackouts to “yes” DWI, RWI, or blackouts. The severity and frequency of those outcomes were left for future investigation. Third, our study uses participant self-reports instead of direct parent reports in assessment of parental monitoring knowledge and parental support of not using alcohol. This inherently introduces potential for reporting bias.

Binge drinking in 12th grade predicted a significantly higher likelihood of DWI, RWI, blackout, and extreme binge drinking and led to more risky driving when transitioning into early adulthood. To an important but limited extent, parental practices while the teenager is in high school may protect against DWI, RWI, and blackouts as adolescents move into early adulthood. Our findings are relevant to prevention programs that seek to incorporate alcohol screening with intentional inquiry about binge drinking. Moreover, our results may be instructive to programs that seek to leverage facets of parental practices to reduce health-risk contexts for youth. Such prevention activities coupled with strengthening of policies and practices reducing adolescents’ access to alcohol could reduce later major alcohol-related health-risk behaviors and their consequences.

BAC

blood alcohol concentration

C-RDS

self-reported Checkpoints Risky Driving Scale

DWI

driving while impaired

RWI

riding with an impaired driver

W

wave