Event-level correlates of drinking events characterized by alcohol-induced blackouts

Abstract

Background:

Prior research identifies a range of potential predictors of blackouts and suggests that blackouts increase risk for additional negative consequences. However, these studies are based on epidemiological work that allows us to draw conclusions about groups of people but not within-person processes. The present study examined within-person, event-level correlates of blackouts.

Methods:

Ninety-six heavy drinking college students (52% female) completed 28-days of daily reports of alcohol use and consequences, including blackouts. Thirty-three participants reported 56 blackouts. Hierarchical linear modeling compared morning reports of drinking events on which participants did vs did not report a blackout, controlling for total drinks at the event.

Results:

Blackout likelihood increased as a function of total drinks consumed, and of crossing thresholds for heavy episodic drinking (4+/5+ drinks for women/men) and high intensity drinking (8+/10+). Participants reported a higher total number of additional negative consequences on blackout events. Specific consequences that were more likely included embarrassing oneself and hangover. Blackouts were associated with morning ratings of less positive mood and a less favorable drinking event. Motives for drinking and simultaneous use of marijuana were not associated with blackouts.

Conclusions:

Event-level findings of this study document that events leading to alcohol-induced memory loss are associated with other adverse experiences relative to drinking events that do not result in blackout, and offer potentially motivational levers for preventive interventions.

Introduction

Over two-thirds of college students report consuming alcohol in the past 30 days and approximately one third of these students reports engaging in heavy episodic drinking (HED) (4+/5+ drinks in a single sitting for females/ males) at least once in the past 2 weeks (Johnston et al., 2015). This type of risky drinking is associated with substantial physical, social, and emotional negative consequences (Hingson et al., 2009). One negative consequence, alcohol-induced memory loss (i.e., blackout) is an absence of memory for at least some part of a drinking session (Wetherill and Fromme, 2016b, White, 2003). Blackouts are frequently reported among young adults (Chartier et al., 2011, Hallett et al., 2013, Jennison and Johnson, 1994), particularly college students (Barnett et al., 2014, Hingson et al., 2016). Individuals who report blackouts are at increased risk for numerous other negative alcohol-related consequences, including physical injury and overdose (Hingson et al., 2016). In the present study, we sought to extend the prior research that has primarily examined person-level predictors (e.g., personality, drinking history) and consequences of blackouts, by focusing on correlates of specific drinking events characterized by a blackout. In particular, we examined potential blackout predictors (e.g., alcohol use levels, co-use of marijuana) and other alcohol consequences (e.g., aggression, nausea) that co-occur with blackouts that evening or the morning following.

By definition, alcohol consumption is necessary for an alcohol-related blackout. Studies examining individual-level drinking patterns as a correlate of self-reported blackouts show that higher typical quantity and frequency of drinking are associated with blackouts (Marino and Fromme, 2015, Schuckit et al., 2015, White et al., 2002). Few studies provide data on how drinking influences blackouts at the event- (versus person-) level. One study found that the average drinks consumed on blackout nights ranged from 7 to 11 (vs 2–3 on nights without blackouts) among women and 10 to 20 (vs 3–4 on nights without blackouts) among men (Labhart et al., 2018). “High intensity drinking” (HID) in particular, defined as twice the traditional “binge” or heavy episodic drinking (HED) level of 4+/5+, has been shown to be associated with negative alcohol consequences (Linden-Carmichael et al., 2017, Hingson et al., 2017). In a recent study, individuals who retrospectively reported that they drank beyond the HID threshold at the last party they attended (relative to those who drank at HED levels) also reported higher risk for a range of specific negative consequences, including blackouts (Cox et al., 2019). Yet, the extent to which HID uniquely increases the risk for blackouts relative to the risk conferred by HED, has not yet been examined within an intensive longitudinal study.

Being a user of both alcohol and other drugs also is associated with more consequences, including blackouts (Haas and Smith, 2012, Schuckit et al., 2015). College students with repeated blackouts over two years are more likely than those with few blackouts over time to report baseline use of cigarettes, marijuana, and other illicit drug use (Merrill et al., 2016a). It stands to reason that co-use of alcohol and marijuana would also increase blackout likelihood at the event level. Animal research shows memory impairment due to alcohol combined with THC (Ciccocioppo et al., 2002). In one recent study of day-level behavior (assessed retrospectively), researchers observed increased blackout risk on days where alcohol was combined with marijuana or stimulants (Mallett et al., 2017). However, prospective studies that examine the influence of simultaneous alcohol and marijuana use on blackouts are lacking.

There is also some evidence for a potential link between drinking motives and blackouts. In both cross-sectional (Merrill and Read, 2010) and longitudinal (Merrill et al., 2014) studies, enhancement motives for drinking were positively related to several consequences including any blackouts and repeated blackouts during the first two years of college (Merrill et al., 2016a). Coping motives have also been linked to a range of alcohol-related consequences above and beyond level of drinking (Merrill et al., 2014, Merrill and Read, 2010). Drinking motives such as these are not mutually exclusive and in fact may differ not only between individuals, but also within an individual, from one drinking event to another (O’Hara et al., 2015, Arbeau et al., 2011). Yet, no studies have explored whether event-level motivations for drinking increase blackout risk.

Blackouts co-occur with other negative alcohol-related consequences (White et al., 2002). Drinking to the point of blackouts is linked to a higher likelihood of injury and emergency department care than drinking without blacking out (Mundt and Zakletskaia, 2012, Mundt et al., 2012) and is prospectively associated with sexual revictimization in college women who had been victimized as adolescents (Valenstein-Mah et al., 2015). Hingson and colleagues tested whether endorsement of a blackout in the past 6 months was associated with other negative consequences in that same time period in a young adult sample (Hingson et al., 2016). After controlling for drinking levels, those with blackouts were ~8x more likely to also report a hangover or argue with friends, 10x more likely to do something regretted; 11.5x more likely to miss class, 17x more likely to get behind in school work, and 144x more likely to have an alcohol overdose. However, again, these were all between-person examinations (whether individuals with blackouts have a higher likelihood of experiencing other problems over a specified timeframe). This work did not clarify whether blackout drinking per se is associated with elevated risk for other consequences at the event-level.

Blackouts may also impact other cognitive and affective outcomes. For example, blackouts were associated with perceptions that the drinking experience was less worth it and overall less positive (Fairlie et al., 2016). Given this finding, and the evident co-occurrence between blackouts and negative consequences, it is also plausible that mornings followed by blackouts would be associated with less positive mood. Yet, no research to date has examined event-based associations between blackouts and mood subsequent to the event.

The Present Study

We sought to examine event-level correlates – both potential predictors and outcomes – of alcohol-induced blackouts. We hypothesized that the risk for blackouts would be greater at higher levels of alcohol quantity, on events characterized by either HED or HID, and when simultaneous use of marijuana was endorsed. Also, we explored distinct drinking motives as potential predictors of blackouts. We tested whether blackouts were associated with an increased number of alcohol consequences experienced and explored which specific consequences were most likely to co-occur with blackouts. Further, we hypothesized that blackouts would be associated with more negative subjective evaluations of the overall drinking event. Finally, we explored whether blackouts were associated with next day mood but had no formal hypotheses given the lack of literature on this topic.

Materials and Methods

Participants

A total of 101 participants were enrolled in the study, and one participant withdrew. Of the 100 participants who completed 28-days of daily reports, an additional four participants did not report any alcohol use and were removed for analyses. Thus, participants were 96 (52% female) college student drinkers (Table 1). Eligibility criteria included age 18–20, access to a smartphone with a data plan, enrollment in a local 4-year college or university, and either (a) weekly HED or (b) experience of at least 1 (of 10 assessed) negative alcohol-related consequence in the past two weeks. Exclusion criteria included past two-week illicit drug use (other than marijuana) or current treatment for a substance use disorder.

Table 1.

Description of the sample of 96 drinkers

	N/% or Mean (SD)
Age	18.7 (0.66)
Biological sex
Male	46 (48%)
Female	50 (52%)
Year in school
Freshman	77 (80%)
Sophomore	15 (16%)
Upperclass	4 (4%)
Race (check all that apply)
White	69 (72%)
African American	7 (7%)
Asian	22 (23%)
Native American/Alaskan	1 (1%)
Hawaiian/Pacific Islander	1 (1%)
Other	5 (5%)
Hispanic/Latino ethnicity	14 (15%)
Baseline drinking behavior
Drinks per typical week	10.48 (6.34)
Drinking events per typical week	2.30 (0.93)
Heavy episodic drinker	88 (92%)
High intensity drinker	30 (31%)
During or after drinking, have you forgotten what you did?	37 (39%) Never 59 (61%) Ever 20 (21%) Past month
Have you been unable to remember large stretches of time while drinking heavily?	47 (49%) Never 49 (51%) Ever 14 (15%) Past month
Have you woken up in an unexpected place after heavy drinking?	78 (81%) Never 18 (19%) Ever 1 (1%) Past month

Procedures

Recruitment and orientation.

All procedures were reviewed and approved by the Brown University Institutional Review Board. Flyers and social media advertisements were used for recruitment. Interested volunteers completed an online screener, and those eligible viewed an online informed consent page. Upon consent, participants were directed to an online baseline survey (~20 minutes). Group orientation sessions were conducted to describe the procedures and obtain consent for the longitudinal phase of the study.

Daily assessment protocol.

All study participants began the 28-day daily assessment protocol on the same day. They received daily notifications for an application-triggered morning report and, on days following drinking, 5pm surveys. Each day a morning report was triggered at 7am and remained available for completion throughout the day; however, participants were instructed to complete it as close in time to waking up as possible. Blackouts were again assessed at the 5pm survey, in the event that a participant did not realize he/she had blacked out during the morning report. Participants were paid based on percent compliance with daily reports, earning from $5 (for less than 20% compliance) to $45 in Week 1. Potential payments increased slightly each week, to a maximum of $51 at Week 4.

Measures

Demographics.

Demographics assessed at baseline included age, gender, year in school, race, and ethnicity.

Baseline drinking behavior.

At baseline, participants reported on alcohol use and consequences, used in the present study for descriptive purposes. Specifically, they reported their alcohol use during a typical week in the past 30 days using a grid modeled after the daily drinking questionnaire (DDQ) (Collins et al., 1985). Additionally, they were asked to “Think of the one day you consumed the most alcohol during the last 30 days: how many standard drinks did you consume on that day?” This variable was used to determine past 30-day endorsement of HED (4+/5+ drinks for women/men) and HID (8+/10+ drinks for women/men). Participants reported on experience of alcohol-related consequences with the Brief Young Adult Alcohol Consequences Questionnaire (BYAACQ; Kahler et al., 2005). This measure includes two items suggestive of blackouts: “Have you been unable to remember large stretches of time while drinking heavily” and “Have you woken up in an unexpected place after heavy drinking.” An additional single item asked “During or after drinking, have you forgotten what you did?” (never; yes, but not in the last 30 days; yes, in the last 30 days). This item was added to map onto the way in which blackouts were assessed daily (see below).

Daily morning report measures.

Alcohol use.

When prior-day drinking was endorsed, participants indicated the total number of standard drinks consumed. This variable was used to create dichotomous indicators of whether each day was an HED (4+/5+ drinks for women/men) or HID day (8+/10+ drinks for women/men).

Alcohol Consequences.

When prior day-drinking was endorsed, participants were asked whether they had experienced any of 18 negative and positive consequences of drinking. Items were derived from several measures including the BYAACQ (Kahler et al., 2005), the Positive Drinking Consequences Questionnaire (Corbin et al., 2008), and a daily study of consequences (Lee et al., 2017, Lee et al., 2015). Additionally, we previously conducted qualitative research to assess the consequences experienced by contemporary college students (Merrill et al., 2018), and a subsequent small pilot (n=11) of our measures and methods, to assess the need for further adaptations. Lee et al. assessed blackouts via the item “I couldn’t remember what I did while drinking” which was modified slightly in the present study for brevity to “forget what you did?” We also examined experiences that are viewed negatively (embarrass yourself, become rude/obnoxious, hurt or injure yourself by accident, feel nauseated or vomited, behave aggressively, neglect school-related obligations, have a hangover).

Motives.

When prior-day drinking was endorsed, participants were also asked to report their reason(s) for drinking including,: to feel less nervous/anxious; to feel less depressed; to make the day/night more fun; to get high, buzzed or drunk; and to not be left out. These items were chosen and modified to map onto the 5 factor model of drinking motives (Grant et al., 2007) and were based in part on a prior daily assessment study (O’Hara et al., 2015).

Mood.

Participants were asked “How would you describe your mood right now?” with a scale from −3 (very negative) to +3 (very positive).

Marijuana use.

Participants indicated whether they used marijuana the prior day and if so, whether they were asked if they were “under the influence of marijuana and alcohol at the same time yesterday?”

Event evaluations.

Two items assessed overall evaluation of the drinking event (Fairlie et al., 2016). First, we asked “Thinking about your overall drinking experience yesterday, how would you rate the experience?” (−3=extremely negative to +3=extremely positive). Second, we asked “Thinking about the whole drinking experience yesterday, the positive and the negative, how much was it worth it?” (0=not at all to 6=very).

Analytic Plan

Hierarchical linear modeling (HLM), with event day (Level 1) nested within person (Level 2), was used for primary analyses. First, models were used to examine the impact of drinking behaviors (HED, HID, total drinks), drinking motives, and simultaneous use of alcohol and marijuana on blackout likelihood. Next, blackouts were examined as predictors of other negative consequences, next-day mood, and drinking event evaluations. We examined sex at Level 2 and day-of-study at Level 1 as potential covariates; however, neither was associated with likelihood of blackout and were therefore not included in final models for parsimony. The corresponding Level 2 aggregate of the Level-1 variable of interest was always included in the model. For example, when examining event-level (Level 1) HED and HID as predictors of event-level blackouts, the proportion of drinking days across the full study that were HED and HID were controlled at Level 2. Consistent with recommended procedures for centering (Enders and Tofighi, 2007), continuous predictors at the daily-level (Level 1, e.g., total drinks) were person mean-centered and those at the person-level (Level-2, e.g., proportion of drinking days that were HID) were grand-mean centered. These methods allowed us to isolate truly within-person effects of interest, after partialling out variability in outcomes due to between-person differences. All models also included controls for weekend (Friday, Saturday) vs weekday at Level 1, and, with the exception of the model where HED and HID were predictors of blackouts, we controlled for total drinks.

Results

Descriptives.

See Tables 1 and 2. Across a total possible 2,688 morning reports (96 participants x 28 days), 2653 (98.7%) were submitted. Surveys sent at 5pm to reassess blackouts following drinking days were completed 413 out of 479 times that they were triggered (86%). During daily assessment, thirty-three participants (34%) reported a total of 56 blackouts (range 1–5 per person), with no gender differences in frequency (χ²(5)=3.87, p=.57). Forty-one blackouts were reported on morning reports, and the additional 15 were captured on 5pm surveys. Average drinks reported on events characterized by blackouts was 7.98 (SD=2.73) compared to 4.87 (SD=2.67) on events without a blackout.

Table 2.

Bivariate correlations among event-level variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Blackout	1
2. Total Drinks	.35**	1
3. HED	.23**	.89**	1
4. HID	.28**	.62**	.43**	1
5. Alc + MJ	−.01	.06	.04	.00	1
6. Coping-Dep motive	.12**	.03	.01	−.00	−.13**	1
7. Coping-Anx motive	.02	−.02	−.06	−.02	−.00	.42**	1
8. Fun motive	.04	.20**	.21**	.12**	.00	−.16**	−.10*	1
9. Drunk motive	.11*	.34**	.32**	.13**	.25**	.08	.05	.03	1
10. Conform motive	.08	−.03	−.01	−.02	−.14**	.13**	.12**	.05	−.07	1
11. Negative cons	.33	.29**	.24**	.18**	.04	.05	.07	.10*	.16**	.09*	1
12. Worth it	−.12**	.03	.05	.01	.06	−.08	−.05	.08	.05	−.17**	−.17**	1
13. Overall eval	−.13**	.01	.02	−.01	.05	−.17**	−.09**	.12**	.07	−.12**	−.20**	.77**	1
14. Mood	−.14**	−.01	−.01	−.03	.04	−.16**	−.13**	.00	−.01	−.11*	−.29**	.33**	.37**	1

Note:

^*p<.05

^**p<.001

HED=heavy episodic drinking (0=no, 1=yes), HID=high intensity drinking (0=no, 1=yes), Alc + MJ = simultaneous use of alcohol and marijuana (0=no, 1=yes)

Level 1 correlations are derived from data representing (non-independent) repeated measures within person

Alcohol use quantity as a correlate of blackouts.

Two dichotomous Level-1 variables indicating whether an event crossed the HED threshold and whether it crossed the HID threshold (0=no, 1=yes) were entered into a single model as predictors of blackout (Table 3). In this coding scheme, events that crossed the HID threshold were also coded as 1 for HED. Results indicated that an event that crossed the HED threshold was associated with 5.56 times the likelihood of a blackout, and crossing the HID threshold was associated with an additional 3.31 increase in the likelihood. In other words, relative to light drinking events (under 4/5), HID events were associated with over 18 times the likelihood of a blackout (5.56 × 3.31). In a second model, blackout likelihood also increased as a function of total drinks consumed at the event-level (OR=1.61, 95% CI=1.32–1.97), controlling for total drinks at Level 2 and weekend vs weekday at Level 1.

Table 3.

Hierarchical linear models predicting odds of a blackout from event-level high intensity drinking (HID) and heavy episodic drinking (HED)

	OR	95% CI
Intercept	0.01	(0.00,0.03)
HID event (L1)	3.31	(1.49,7.34)
HED event (L1)	5.56	(1.34,23.15)
Weekend day (L1)	3.08	(1.01,9.42)
Proportion HID days (L2)	1.28	(0.28,5.86)
Proportion HED days (L2)	1.99	(0.27,14.88)

Note: L1= modeled at Level 1; L2=Level 2; Proportion of HID and HED days sample-mean centered; Bold effects are significant (CI does not include 1.00).

Two post-hoc models were run to examine the potential for (1) a moderating effect of gender on the association between total drinks and blackouts, and (2) a curvilinear association between total drinks and blackouts. Although there were no gender differences in blackout odds (OR=2.17, 95% CI = 0.40–11.71), there was an interaction between total drinks and gender in predicting odds of blackout (OR= 0.73, 95% CI = 0.56 – 0.95) such that the strength of the association between total drinks and blackout odds is stronger for women than for men. In a separate model, the additional quadratic term for total drinks also was significant (OR=0.95, 95%CI = 0.92–0.98).

Potential predictors of blackout.

Controlling for total drinks and weekend vs weekday at Level 1, and whether or not an individual endorsed simultaneous use of alcohol and marijuana ever during the study, daily-level simultaneous use of alcohol and marijuana (vs alcohol alone) was not associated with blackout likelihood. Additionally, none of the 5 motive items were associated with blackout likelihood when entered in a single model, controlling for total drinks and weekend at Level 1 as well as proportion of days on which each motive was endorsed across 28 days at Level 2. Of note, when testing each Level 1 motive individually (no covariates), “to get high, buzzed, or drunk” was significantly associated with blackout likelihood (OR=2.10, 95% CI=1.08–4.10).

Co-occurring consequences of blackouts.

Controlling for total drinks at Level 1 and whether the participant reported a blackout ever during the study at Level 2, participants reported a higher total number of additional negative consequences on blackout events (B=0.71, SE=0.16, p<.001). As shown in Table 4, specific consequences that were more likely on blackout events included embarrassing self and hangover, controlling for total drinks. At Level 2, participants who reported a blackout during the study did not differ from those who did not. For descriptive purposes, Table 5 shows overlap at the day-level between blackouts and other consequences.

Table 4.

Hierarchical linear models predicting specific consequences by event- and person-level blackouts

	Event-level blackout (Level 1)		Person-level blackout (Level 2)
Consequence	Odds ratio	Confidence Interval (95%)	Odds ratio	Confidence Interval (95%)
Embarrassing self	5.35	(2.41,11.89)	0.12	(0.38,3.29)
Hurt/injured self	2.73	(0.95,7.85)	3.14	(0.72,13.80)
Felt nauseous/vomited	1.87	(0.77,4.52)	1.58	(0.74,3.35)
Had a hangover	3.24	(1.31,7.98)	0.64	(0.30,1.35)
Neglected school work	1.69	(0.80,5.56)	0.14	(0.47,2.82)
Was rude/obnoxious	1.86	(0.29,11.86)	0.69	(0.15,3.15)
Behaved aggressively	3.29	(0.69,15.67)	0.73	(0.09,5.89)

Note: Bold effects are significant (CI does not include 1.00). All models control for total drinks consumed (person-centered) and (with the exception of hurt/injury due to collinearity) weekend vs weekday at Level 1 (event-level). Level 2 blackouts were modeled as 1=ever and 0=never during the 28-day period.

Table 5.

Overlap between events characterized by blackouts and other consequences

Consequence	Total # of events	% of blackout nights	% of non-blackout nights	% times consequence co-occurred with blackout
Embarrassing self	44	28.6	6.5	36.4
Hurt/injured self	14	12.5	1.6	50.0
Felt nauseous/vomited	57	30.4	9.3	29.8
Had a hangover	93	42.9	16.0	25.8
Neglected school work	86	25.0	16.7	16.3
Was rude/obnoxious	22	8.9	3.9	22.7
Behaved aggressively	12	7.1	1.9	33.3

Post-blackout morning mood and cognition.

Controlling total drinks, reporting a blackout at the event-level was marginally associated with a lower next-morning mood (B= - 0.40, SE=0.21, p=.052). Additionally, controlling total drinks, blackout was associated with rating the event as less worth it (B=−0.50, SE=.25, p=.047), but not with ratings of the overall event as positive vs negative.

Discussion

This was the first study to use intensive longitudinal data collected on an event-level to examine predictors and outcomes associated with blackouts. Participants increased their risk for blackouts when they crossed the HED threshold, with even further risk upon crossing the HID threshold. Events characterized by blackouts were associated with more total negative consequences and specific consequence types, controlling for total drinks. Moreover, blackouts were significantly associated with evaluations of the overall drinking event as less “worth it” and marginally significantly associated with less positive mood the next morning. Importantly, our results also indicated that the impact of blackouts on other outcomes is specific to the event- (rather than person-) level.

Not surprisingly, a higher number of drinks increased the odds of blackout, and post-hoc models suggested (1) the risk conferred by each additional drink was greater form women than for men, and (2) there may be an exponential relationship between drinking quantity and blackouts. Looking at specific cutoffs revealed that crossing the HED threshold increased the odds of experiencing a blackout more than five times. Of note, the risk of a blackout increased by an additional three times if an individual drank at high intensity levels. Taken together, these findings suggest that drinking more than 8 (women) or 10 (men) drinks in a night increases the risk of a blackout by about 18-fold (relative to light drinking nights). This event-level data adds to prior work showing that individuals who drink at high intensity levels report more consequences (Hingson et al., 2017, Linden-Carmichael et al., 2017) and blackouts in particular (Cox et al., 2019). Our prospective data collected across 28 days confirm that blackouts are much more likely on events where these risky levels of use are reached. Taken together, this growing body of work suggests that drinkers should not exceed HID levels if they wish to avoid the negative consequences likely to result.

We also aimed to identify whether blackouts confer risk for other negative outcomes at the event level. Importantly, even when partialling out variance in consequences due to number of drinks consumed, blackout events were characterized by a greater number of additional negative consequences. That is, not only are drinkers who report blackouts also reporting more prior consequences (Hingson et al., 2016), but a single event on which a blackout occurs may also be accompanied by a greater number of other consequences. In particular, likelihood was higher for experiencing embarrassment and/or a hangover.

It is remarkable to observe an effect of blackouts on any consequences beyond quantity consumed, and raises questions about the mechanisms of these effects. Regarding embarrassment, it is possible that students learned from friends about behaviors in which they engaged during a period of memory loss (and likely disinhibition). Our data may also suggest that style of drinking adds to risk of negative consequences over and above the quantity consumed. Both blackout and hangover relate to physiological impacts of alcohol, and could potentially be influenced by not only number of drinks consumed, but speed of alcohol consumption or specific behaviors that influence the rate at which BAC rises (e.g., shots, chugging). Indeed, prior work suggests that speed of drinking may impact blackouts (Perry et al., 2006, Goodwin et al., 1970, Ryback, 1970). Future work using data collected in real-time versus only the next morning may help isolate specific patterns of drinking during an event that increase risk for both blackouts and other consequences.

In addition to negative consequences during the drinking event, there was some evidence that blackouts were associated with next-day cognitive and affective outcomes. In particular, events characterized by a blackout were followed by more negative mood (marginally significant) and perceptions that the drinking event was less “worth it.” The finding that blackout events were less “worth it” is consistent with prior survey work (Fairlie et al., 2016). Additionally, HID (related to blackouts in this study) has been shown to be associated with worse next-day physical and cognitive consequences (excessive tiredness, problems concentrating) (Polak and Conner, 2012). It is possible that these effects, or the other consequences demonstrated to co-occur with blackouts in this study (embarrassment, hangover), are mechanisms that help to explain effects of blackouts on both mood and event evaluations. Indeed, a prior qualitative study revealed that blackouts themselves are rated more negatively when they are accompanied by other negative consequences (Merrill et al., 2019); the same may be true for the overall drinking event.

We did not find evidence that simultaneous use of alcohol and marijuana increased the likelihood of a blackout after adjusting for consumption, inconsistent with prior work looking at use of alcohol and marijuana in the same day (Mallett et al., 2017) and our hypotheses. Of note, we did not collect data on marijuana use quantity, potency, the order in which the two substances were used, or speed of drinking, all of which may have been important. Likelihood of a blackout might depend where during the ascending and descending limbs of the blood alcohol concentration curves that marijuana is consumed. There is also some evidence that one motivation for co-use is to avoid getting too intoxicated from alcohol (Patrick et al., 2018). As such, it is also possible that drinkers consume alcohol in a less risky manner when also smoking (e.g., drinking more slowly, avoiding shots of liquor); this would be the case if marijuana co-use affected consumption speed. These particular aspects of alcohol-marijuana co-use and how they influence blackout likelihood should be examined in future event-level studies.

We also suspected that certain drinking motives at the event-level may place students at greater risk for blackouts, above and beyond quantity of use, due to potentially riskier patterns of drinking during an event. In a single-predictor HLM model, “to get high, buzzed, or drunk” was associated with blackouts, consistent with prior work suggesting that individuals with higher enhancement motives may experience blackouts (Merrill et al., 2016b). However, we did not find that any of the drinking motives examined were uniquely associated with blackout likelihood, when controlling other motive types and alcohol use level. Prior work has only examined motives as a person-level predictor of alcohol consequences such as blackouts. The intensive nature of our assessment protocol made the use of single-item measures preferable, to reduce participant burden; however, a lack of multi-item, validated measures of drinking motives may have impacted null findings. Alternatively, the general tendency to drink for certain reasons may be a more important predictor of blackouts than the motives for drinking reported the morning following the event.

Limitations and Future Directions

As with any study on blackouts, because we are asking participants to report on events that by definition they do not recall (Wetherill and Fromme, 2016a), it is plausible that blackouts were underreported here. Of note, participants became aware of and recorded some of their blackouts at the 5pm assessment, and there may have been instances where students realized even later than this. Additionally, our single item to assess blackouts (forgot what you did) may not have captured the full range of potential blackout experiences. For example, participants who forgot what they said, or where they were, may not have endorsed this item, regardless of whether their experience would actually be classified as a blackout. We also cannot distinguish between en bloc (i.e., total absence of memory for a circumscribed period of time) versus fragmentary (i.e., ability to retrieve partial memory with appropriate cues) types of blackouts (Goodwin et al., 1969) in this study. Future studies should include more comprehensive blackout measures as well as questions on how drinkers learn about their blackouts.

The incidence of other negative consequences that occurred during blackouts may also have been underreported, as it might require learning about behavior from sources other than one’s own memory. Alternatively, it is possible that reported consequences were inaccurately endorsed. For example, participants may have endorsed consequences they were told they experienced by their (also intoxicated) friends, who may have been unreliable sources of information; or, participants may have made assumptions about their consequences (e.g., embarrassing self) due to the blackout, whether or not those consequences actually occurred. Moreover, the present study does not allow us to draw conclusions about whether other alcohol-related consequences are more specifically the result of blacking out, as for example, it is possible that some occurred prior to any memory loss.

Additionally, our sample was characterized by college students from a single university. The extent to which our findings apply to other populations of heavy drinkers (e.g., alcohol dependent individuals; or non-college attending young adults) is yet unknown. Moreover, our results are limited by the use of single-item measures, which, while preferable for assessment-intensive studies, may be less reliable than multi-item measures. However, bivariate correlations suggest convergent and divergent validity of these measures, with associations mapping onto those observed in the extant literature based on multi-item measures. For example, more positive mood was negatively associated with both types of coping motives and negative consequences, and “fun” and “drunk” motives were positively associated with number of drinks consumed. Finally, future studies that use biosensors to detect alcohol levels in the body in real-time would allow for more accurate conclusions to be drawn about the role of both level and rate of consumption in predicting blackouts.

Conclusion and Intervention Implications

This work represents an important extension of prior studies on person-level influences on blackouts, with at least three important implications for intervention. First, the growing body of data supporting the link between blackouts and other negative consequences could inform interventions to help motivated drinkers to avoid blackouts. For example, teaching college drinkers that drinking to the point of blackout also increases the likelihood that they will embarrass themselves or have a hangover may be motivational. Although prior work has indicated that such increases in motivation do not always predict changes in drinking (Marino & Fromme, 2018), our data make a stronger case for a within-person association. Second, our data add another affective dimension to the drinking to blackout that drinkers might consider, potentially triggering a conversation about how nights when one drinks to the point of blackout are perceived as less “worth it” and associated with less positive mood. Thus, an intervention goal might be to shift the positive attitude valence of blackouts (e.g., DiBello et al., 2018).Third, event-based studies such as these suggest opportunities for real-time intervention. By pinpointing which drink levels, behaviors, or cognitions increase risk for blackouts, we could intervene at the time those variables are detected (using mobile, just-in-time adaptive interventions) (Nahum-Shani et al., 2016). Needed is systematic study of the additive value of these suggestions for refinement of alcohol interventions.