Are Protective Behavioral Strategies Associated with Fewer Negative Consequences on High-Intensity Drinking Days? Results from a Measurement-Burst Design

Abstract

Heavy episodic drinking (HED) is prevalent among college students and associated with harms. However, many students engage in “high-intensity drinking” (HID) by drinking at thresholds beyond HED. HID relative to HED-only is associated with elevated risk for acute and severe negative consequences. When used, protective behavioral strategies (PBS) are shown to help drinkers limit their drinking and lower their odds of experiencing consequences. This study assessed whether PBS use was associated with reduced consequences on occasions college students engaged in HID relative to HED-only. Data were from a longitudinal measurement-burst design (14-day bursts across four semesters) from 256 college students who engaged in HID on at least one reported day, yielding 2,352 daily drinking reports. Participants reported the number of standard drinks consumed on each day and, on days with 1+ drinks, whether they used PBS and experienced a variety of negative drinking-related consequences. Three-level multilevel models revealed that on days when students used manner of drinking PBS there was a weaker association between HID and passing out from drinking. On days students used serious harm reduction PBS there was a weaker association of HID with having no one sober enough to drive and experiencing regretted sexual behaviors. Use of some PBS may help college students reduce harms on HID occasions relative to HED-only occasions. Interventions should promote use of planning strategies to minimize harm, especially on HID occasions.

College students who engage in heavy episodic drinking (HED), defined as consuming at least 4 drinks (for women) or 5 drinks (for men) on a given occasion, are at elevated risk for a number of serious alcohol-related harms and health outcomes (White & Hingson, 2013). Examples include academic and occupational problems, passing out from drinking too much, injuries, and driving after drinking. Many college students exceed standard HED thresholds by drinking more than two or three times this number of drinks (Patrick & Terry-McElrath, 2017; Patrick, Terry-McElrath, Kloska, & Schulenberg, 2016a; White, Kraus, & Swartzwelder, 2006), often referred to as “high-intensity drinking” (HID; Patrick, 2016). HID has been defined as 8+/10+ or 12+/15+ drinks in a day for women/men, respectively (Patrick, Cronce, Fairlie, Atkins, & Lee, 2016b). Though engagement in HED and HID has decreased among 18- to 24-year-olds in the U.S. (Patrick, Terry-McElrath, Miech, Schulenberg, O’Malley, & Johnston, 2017), prevalence remains high and nearly equivalent by age during young adulthood (Linden-Carmichael, Vasilenko, Lanza, & Maggs, 2017). An increase in frequency of HID after high school has been documented, but only for those who attend college (Patrick et al., 2016a). High-intensity drinkers are at greater odds of reporting legal problems, emergency department visits (Hingson et al., 2017), driving-related consequences (Evans-Polce, Patrick, & O’Malley, 2017), and past-year alcohol use disorders (Linden-Carmichael et al., 2017) than individuals drinking at lower levels. Comparisons of HED-only (4–7 or 5–9 drinks in a day for women/men) and HID convey the relative harms that can be experienced by individuals drinking at such high levels, highlighting the need for distinguishing between HED-only and HID.

The acute and serious alcohol-related consequences of HID among U.S. college students highlight a need for developing prevention and intervention efforts targeting this higher-risk group of drinkers. Brief motivational interventions, such as the Brief Alcohol Screening and Intervention for College Students (BASICS; Dimeff, Baer, Kivlahan, & Marlatt, 1999) use multiple components, including guidance on how to use strategies to protect oneself from harm. Although designed to reduce heavy drinking and related harms in general, it is unknown whether such strategies are helpful in decreasing the harms associated with HID specifically. Drinking motivations/reasons for drinking and contexts of drinking differentiate people who engage in HED-only from those who engage in HID (Patrick, Evans-Polce, Kloska, Maggs, & Lanza, 2017; Terry-McElrath, Stern, & Patrick, 2017; White, Anderson, Ray, & Mun, 2016), which may suggest that drinking serves different purposes between these groups and potentially for the same individuals but on different drinking occasions. Thus, different strategies may be needed for reducing harms associated with drinking at HID levels.

Protective behavioral strategies (PBS) are one commonly used intervention component with demonstrated efficacy in helping individuals reduce harms from drinking (see Pearson, 2013 for a review). Several sub-types of strategies exist, including strategies for limiting the amount of alcohol consumed (e.g., pre-plan the number of drinks to consume before going out), modifying the way of drinking (e.g., avoid drinking games, drink more slowly), reducing the likelihood of experiencing serious harms (e.g., use a designated driver), or planning activities to avoid drinking (e.g., avoid situations where alcohol is likely) (Martens, Ferrier, Sheehy, Corbett, Anderson, & Simmons, 2005; Novik & Boekeloo, 2011; Sugarman & Carey, 2007). More frequent use of such strategies is generally found to be negatively associated with amount of alcohol consumed and alcohol-related harms in cross-sectional comparisons (e.g., Araas & Adams, 2008; Benton, Schmidt, Newton, Shin, Benton, & Newton, 2004; Linden, Lau-Barraco, & Milletich, 2014; Pearson, Kite, & Henson, 2012) and as a mechanism of change within an intervention context (e.g., Barnett, Murphy, Colby, & Monti, 2007). Sub-types of strategies tend to have differential associations with alcohol use outcomes. Though findings are mixed, modifying one’s drinking style (“manner of drinking” strategies) and attempts to reduce serious harms (“serious harm reduction” strategies) are generally more negatively correlated with experienced harms than strategies involving limiting the amount one drinks (Frank, Thake, & Davis, 2012; Martens et al., 2005; Napper, Kenney, Lac, Lewis, & LaBrie, 2014; Pearson et al., 2012).

Intensive longitudinal designs, such as measurement-burst designs, are an ideal approach for gaining a better understanding of the interaction between PBS use and HID on alcohol-related harms. Such designs can compare drinking occasions (e.g., HID occasions vs. HED-only occasions) within individuals and across semesters of college. Some recent, albeit limited, research has used daily or weekly diary studies to examine within-person variability as well as predictors and consequences of HID (Patrick et al., 2016b) and, separately, PBS (Braitman, Linden-Carmichael, & Henson, 2017; Lewis, Patrick, Lee, Kaysen, Mittman, & Neighbors, 2012; Linden-Carmichael, Braitman, & Henson, 2015; Pearson, D’Lima, & Kelley, 2013; Sell, Turrisi, Scaglione, Cleveland, & Mallett, 2018). Results from these repeated measures studies suggest that the experience of negative consequences differs not only between individuals (e.g., HID drinkers relative to non-HID drinkers), but also differs within an individual on their own less versus more intense drinking occasions.

Our study seeks to extend knowledge of harms associated with HID in several key ways. First, although HID days are associated with greater harms than drinking days not involving HID (Patrick et al., 2016b), previous studies defined “non-HID” days as drinking fewer than 8/10 drinks for women/men. This operationalization combines much lighter drinking days (e.g., 1 or 2 drinks) with heavier drinking days (e.g., 7 drinks) and compares these to HID days (e.g., 11 drinks). Second, it remains unclear whether prior cross-sectional findings comparing HID and HED are due to between-person differences or within-person differences. For prevention implications, it is useful to determine whether, within an individual, HID occasions are associated with more harms than HED-only occasions. Comparisons examining whether and which harms are more likely to occur on HID occasions relative to HED-only occasions for the same persons would elucidate this issue. Third, findings related to daily PBS and alcohol use outcomes have been mixed with respect to daily associations between types of PBS used and negative consequences experienced (e.g., Lewis et al., 2012; Pearson et al., 2013). Several of these daily diary studies have relied on reports from short time windows (e.g., 14 days) or from small samples. Use of a measurement-burst design examining PBS and consequences across many days, semesters, and individuals may provide a richer perspective on the association between PBS and consequences. Finally, the extent to which PBS use may be associated with a weaker association between HID and alcohol-related harms in a daily context has not yet been tested. Findings indicating whether and which types of strategies are associated with experiencing fewer harms associated with HID would have clear implications for prevention and intervention work targeting HID-related risks.

The current study used a longitudinal measurement-burst design following emerging adults across the college years to answer the following research questions: (1) Are HID days associated with greater odds of experiencing acute negative consequences than HED-only days? (2) Is PBS use (i.e., manner of drinking strategies, serious harm reduction strategies) on a given day associated with lower odds of experiencing acute negative consequences? (3) Is PBS use more strongly linked with fewer consequences on HID days compared to HED-only days? Toward this end, we compared days in which college students reported drinking at levels of HID (8+/10+ drinks for women/men) relative to days involving HED-only (4–7/5–9 drinks for women/men) in terms of negative consequences experienced, as well as the interaction between PBS and drinking level predicting consequences. We hypothesized that HID days would be associated with greater odds of consequences and that PBS use would be associated with fewer consequences, particularly on HID days.

As our focus was on negative consequences, we focused solely on PBS designed to directly limit negative consequences (i.e., manner of drinking, serious harm reduction strategies) rather than only to limit amount of alcohol consumed (e.g., limit number of drinks, avoid drinking any alcohol). Furthermore, to test specific hypotheses, we matched PBS to consequences they could logically reduce. Manner of drinking PBS focus on speed of drinking and may reduce blood alcohol concentration or intoxication despite consuming the same amount during a drinking occasion. Thus, manner of drinking strategies were matched to physiological consequences (e.g., hangover, passing out from drinking). As serious harm reduction PBS are aimed at precautions while drinking but not directly related to consumption, serious harm reduction PBS were matched to consequences of arranging a designated driver or regretted sexual behavior.

Method

Data were from the University Life Study (ULS), which assessed the daily activities and risk behaviors of college students at a large, state university in the Northeastern United States (Patrick, Maggs, & Lefkowitz, 2015). The ULS used a longitudinal measurement-burst design in which students completed 14 consecutive daily web-based surveys, in addition to a longer web-based survey, in each of seven consecutive semesters from Fall of their first year of college through Fall of their fourth year. A stratified, random sampling procedure was used to attain a diverse sample with respect to gender and race/ethnicity. Eligible participants were first-time, full-time, first-year students who were younger than 21 years of age, were U.S. citizens or permanent residents, and resided within 25 miles of the university campus at the start of the study. The current study was approved by the university’s internal review board and complied with all APA ethical guidelines.

In total, 744 students (65.6% of those invited to participate) provided informed consent and completed surveys in Semester 1. Measurement of PBS began in Semester 4 so only Semesters 4 through 7 were used in analyses. Across these semesters, participants completed daily surveys on a maximum of 56 days (14 days × 4 semesters). Students who reported at least one HID day (n = 256) were selected for the analytic sample. Moreover, given the focus on differentiating between levels of heavy drinking, study analyses included days with reports of at least 4 drinks for women and at least 5 drinks for men, in order to compare HED-only days (4–7/5–9 drinks on a given day for women/men) and HID days (8+/10+ drinks). The analytic sample consisted of 2,352 HED or HID person-days within 770 person-semesters reported by 256 persons. On average, each student provided data on 5.30 HED-only days (SD = 3.76, range = 1–22 days) and 4.57 HID days (SD = 4.60, range = 1–24 days). The sample was 50.0% female and was 34.4% European American non-Hispanic/Latinx (NHL), 32.0% Hispanic/Latinx, 14.1% Asian American/Pacific Islander NHL, 9.8% African American NHL, and 9.8% multi-racial NHL.

Measures

High-intensity drinking (HID) and heavy episodic drinking (HED).

On each daily survey, students were asked, “How many drinks of alcohol did you drink?” in reference to the previous day from the time they woke up until the time they went to sleep. Students used a pull-down menu, with options ranging from 0 to 25+ to indicate the total number of drinks they consumed the previous day. A single drink was defined as, “Half an ounce of absolute alcohol, for example, 12-ounce can of beer or cooler, 5-ounce glass of wine, or a drink containing 1 shot of liquor or spirits” (NIAAA, 2003). As noted previously, only days with 4+/5+ drinks for women/men were used in analyses. The number of drinks consumed was recoded into a binary variable. HED-only days (coded 0) were those in which women reported consuming 4 to 7 drinks and men reported 5 to 9 drinks (Wechsler, Dowdall, Davenport, & Rimm, 1995). HID days (coded 1) were those in which women/men reported consuming 8+/10+ drinks (Linden-Carmichael et al., 2017; Patrick, 2016; White et al., 2006).

Protective behavioral strategies (PBS).

Items were a brief version of the Protective Behavioral Strategies Survey (PBSS; Martens et al., 2005), adapted to reduce participant burden given daily reporting. On each day participants reported drinking they were asked, “Did you use any of the following strategies when using alcohol or ‘partying’ on [the previous day]?” Manner of drinking was presented as, “Avoid drinking fast (e.g., avoid drinking games, not doing shots, not drink fast)” and serious harm reduction was operationalized as,”Avoid serious consequences (e.g., watch your drink, go home with a friend, use a designated driver).”Each day, students could select “No” (0) or “Yes” (1) for each question.

Negative consequences of alcohol use.

On each daily survey, participants were asked “As a result of drinking on [the previous day], did you…” with possible responses of “No”(0) and “Yes” (1) to each of eight negative consequences of alcohol use (adapted from Patrick & Maggs, 2008, 2014). Specifically, students indicated whether they had a hangover, passed out, ended up in bad physical shape the next day, did or said something embarrassing, lost control of themselves, had their coordination affected, found themselves in a situation in which no one was sober enough to drive, or had a sexual experience they regretted. Each of these consequence items served as a dichotomous outcome variable in separate analyses.

Weekend day.

Social weekend days (Del Boca, Darkes, Greenbaum & Goldman, 2004; Maggs, Williams & Lee, 2011) were specified as Thursdays, Fridays, and Saturdays (coded as 1). Weekdays were specified as Sundays, Mondays, Tuesdays, and Wednesdays (0).

Gender.

In Semester 1, students indicated their gender as “Female” (0) or “Male” (1).

Analytic Plan

Generalized linear mixed models (GLMMs) nesting days within semesters within persons predicted the likelihood that students experienced specific negative consequences of alcohol use on a given day as a function of HID, PBS use, day of the week, and gender, using SAS 9.4. Because the eight outcomes were dichotomous, models used logistic GLMMs based on the Bernoulli distribution (Raudenbush & Bryk, 2002). A daily-level (level-1) random effect was used to account for under- and over-dispersion, that is, the presence of less or more variability in the data than would be expected under model assumptions (Raudenbush & Bryk, 2002). The general equation for the GLMMs was:

$$\begin{gathered} \text{logit}{\left(\text{Y}\right)}_{\text{ijk}}={\gamma }_{000}+{\gamma }_{001}{\left(\text{Person-mean HID}\right)}_{\text{k}}+{\gamma }_{002}{\left(\text{Person-mean PBS use}\right)}_{\text{k}}+{\gamma }_{003}{\left(\text{Male gender}\right)}_{\text{k}} \\ +{\gamma }_{010}{\left(\text{Semester-mean HID}\right)}_{\text{jk}}+{\gamma }_{020}{\left(\text{Semester-mean PBS use}\right)}_{\text{jk}}+{\gamma }_{100}{\left(\text{HID day}\right)}_{\text{ijk}} \\ +{\gamma }_{200}{\left(\text{PBS use day}\right)}_{\text{ijk}}+{\gamma }_{300}{\left(\text{Weekend day}\right)}_{\text{ijk}}+{\mu }_{00\text{k}}+r_{0\text{jk}} \end{gathered}$$

Each negative consequence outcome was matched with one of the two PBS based on theoretical and logical considerations: the manner of drinking strategy was used in models predicting the likelihood students experienced a hangover, passed out, ended up in bad physical shape the next day, did or said something embarrassing, lost control of themselves, and had their coordination affected. Serious harm reduction strategies were used in models predicting the likelihood students ended up in a situation in which no one was sober enough to drive and had a sexual experience they regretted.

Intercepts in the GLMMs reflected the odds of experiencing a particular negative consequence of alcohol use on HED weekdays in which no PBS were used for females who had average levels of HID and PBS use at both person- and semester-levels. Person- and semester-mean HID and PBS use variables were created by calculating means of these variables for each person (i.e., the overall proportion of days individuals engaged in each behavior) and for each person-semester (i.e., the proportion of days individuals engaged in each behavior in each semester), respectively. Person-mean HID, γ₀₀₁, was grand-mean centered by subtracting the sample-wide HID mean (.50) from each individual’s HID mean. Person-level means for each of the two PBS use variables, γ₀₀₂, were grand-mean centered in the same fashion. Semester-mean HID, γ₀₁₀, was person-mean centered by subtracting each individual’s person-mean HID from their HID mean in each semester. Semester-level means for each of the two PBS use variables, γ₀₂₀, were person-mean centered in the same fashion. Daily HID, daily PBS use, weekend day, and gender were left uncentered as they were dichotomized.

Results

Descriptive statistics

Of the 2,352 daily reports used in this study, students reported HED-only on 1,183 (50.30%) days and HID on 1,169 (49.70%) days. On the 1,183 HED-only days in the analytic sample, students reported using manner of drinking strategies on 330 (27.90%) days, using serious harm reduction strategies on 618 (52.24%) days, and experiencing at least one negative consequence of alcohol use on 548 (46.32%) days. On the 1,169 HID days in the analytic sample, students used manner of drinking strategies on 189 (16.17%) days, used serious harm reduction strategies on 512 (43.80%) days, and experienced a consequence on 853 (72.97%) days.

Aim 1: Are HID days associated with greater odds of experiencing acute negative consequences than HED-only days?

Models in Tables 1 and 2 tested whether negative consequences of alcohol use were more likely on HID days in comparison to HED-only days, after controlling for PBS use. At the daily level, students were more likely to experience all negative consequences of alcohol use, except for having a sexual experience they regretted, on HID days than on HED-only days (range of odds ratios (ORs) = 3.47 – 7.57), γ₁₀₀. For example, the odds of passing out on HID days were over seven times greater than the odds of passing out on HED-only days. In contrast, students had 39% lower odds of having a sexual experience they regretted on HID days than on HED-only days. Regarding time of the week, students were less likely to end up in bad physical shape the next day after drinking or to have a sexual experience they regretted on weekend days compared to weekdays, γ₃₀₀.

Table 1.

Main Effects of High-Intensity Drinking and Manner of Drinking Strategy Use on the Odds of Experiencing Alcohol-Related Consequences

	Hangover	Pass Out	Bad Phys. Shape	Embarrass Self	Lose Control	Affect Coord.
Fixed Effects	OR [CI]	OR [CI]	OR [CI]	OR [CI]	OR [CI]	OR [CI]
Daily Level
n (days)	2,347	2,344	2,343	2,345	2,343	2,346
Intercept, γ000	0.24 [0.17, 0.35]***	0.01 [0.01, 0.02]***	0.06 [0.04, 0.09]***	0.12 [0.08, 0.19]***	0.02 [0.01, 0.03]***	0.15 [0.09, 0.24]***
HID, γ100	3.87 [2.99, 5.02]***	7.57 [5.59, 10.3]***	5.83 [4.36, 7.78]***	3.47 [2.65, 4.55]***	7.21 [5.32, 9.78]***	6.11 [4.68, 7.97]***
MOD use, γ200	0.48 [0.34, 0.68]***	0.34 [0.22, 0.53]***	0.47 [0.32, 0.70]***	0.36 [0.25, 0.53]***	0.13 [0.08, 0.22]***	0.31 [0.21, 0.44]***
Social weekend, γ300	0.80 [0.62, 1.03]	1.12 [0.83, 1.50]	0.57 [0.43, 0.75]***	0.91 [0.69, 1.20]	1.34 [0.98, 1.83]	0.83 [0.63, 1.09]
Semester Level
Mean HID, γ010	0.83 [0.51, 1.33]	0.54 [0.25, 1.18]	1.07 [0.56, 2.02]	0.66 [0.37, 1.16]	0.32 [0.15, 0.70]**	0.61 [0.34, 1.08]
Mean MOD use, γ020	1.71 [0.94, 3.12]	0.91 [0.33, 2.46]	0.87 [0.36, 2.09]	1.34 [0.64, 2.77]	2.29 [0.83, 6.34]	1.23 [0.59, 2.57]
Person Level
Mean HID, γ001	1.09 [0.51, 2.33]	1.10 [0.33, 3.61]	0.40 [0.15, 1.06]	0.68 [0.26, 1.79]	0.50 [0.15, 1.66]	0.57 [0.21, 1.59]
Mean MOD use, γ002	2.27 [1.05, 4.91]*	1.20 [0.33, 4.34]	0.61 [0.22, 1.74]	0.94 [0.34, 2.61]	1.47 [0.39, 5.49]	0.95 [0.33, 2.71]
Male gender, γ003	1.25 [0.86, 1.82]	1.80 [0.97, 3.33]	1.33 [0.81, 2.18]	0.87 [0.53, 1.43]	1.25 [0.67, 2.33]	2.13 [1.26, 3.61]**

Note. Between-person n = 256 individuals. OR = odds ratio, CI = 95% confidence interval, HID = high-intensity drinking (consuming 8+/10+ drinks in a day for women/men, MOD = manner of drinking protective behavioral strategy (avoid drinking fast, playing drinking games, and doing shots). Bad Phys. Shape = Bad Physical Shape. Affect Coord. = Affect Coordination.

^*p < .05,

^**p< .01,

^***p < .001.

Table 2.

Main Effects of High-Intensity Drinking and Serious Harm Reduction Strategy Use on the Odds of Experiencing Alcohol-Related Consequences

	No Sober Driver	Regret Sex
Fixed Effects	OR [CI]	OR [CI]
Daily Level
n (days)	2,336	2,335
Intercept, γ000	0.00 [0.00, 0.01]***	0.00 [0.00, 0.00]***
HID, γ100	6.00 [4.54, 7.94]***	0.61 [0.46, 0.81]***
SHR use, γ200	0.89 [0.66, 1.20]	0.40 [0.28, 0.57]***
Social weekend, γ300	1.16 [0.87, 1.54]	0.57 [0.43, 0.77]***
Semester Level
Mean HID, γ010	0.67 [0.22, 1.99]	1.90 [0.35, 10.27]
Mean SHR use, γ020	2.07 [0.63, 6.75]	7.13 [1.13, 44.93]*
Person Level
Mean HID, γ001	2.88 [0.52, 16.00]	4.49 [0.76, 26.61]
Mean SHR use, γ002	0.61 [0.17, 2.21]	2.28 [0.60, 8.69]
Male gender, γ003	1.76 [0.69, 4.49]	1.10 [0.43, 2.81]

Note. Between-person n = 256 individuals. OR = odds ratio, CI = 95% confidence interval, HID = high-intensity drinking (consuming 8+/10+ drinks in a day for women/men, SHR = serious harm reduction protective behavioral strategy (avoid serious consequences by watching your drink, going home with a friend, or using a designated driver).

^*p < .05,

^**p< .01,

^***p < .001.

At the semester level, students were less likely to lose control as a result of drinking during semesters in which they engaged in HID more frequently, compared to semesters they did so less frequently, γ₀₁₀. The frequency of HID in a given semester was not associated with the likelihood of experiencing negative consequences of alcohol use in any of the other six models. At the person level, there was no association between the total number of times a student engaged in HID across the sampled days across college and the likelihood they experienced any of the individual negative consequences of alcohol use on a given day in any of the eight models, γ₀₀₁. Males were more likely to have their coordination affected by drinking than were females, γ₀₀₃, but gender was not a significant predictor of experiencing any other negative consequence.

Aim 2: Is PBS use on a given day associated with lower odds of experiencing acute negative consequences?

Models in Tables 1 and 2 also tested whether individual negative consequences of alcohol use were less likely on drinking days in which students used a manner of drinking strategy (Table 1) or serious harm reduction strategy (Table 2), after controlling for HID. At the daily level, on days they used a manner of drinking strategy students were less likely to experience all six strategy-relevant physical and behavioral negative consequences of alcohol use (range of ORs = 0.13 – 0.48), γ₂₀₀. For example, students had 66% lower odds of passing out on days they used the manner of drinking strategy compared to days they did not use this strategy. At the semester level, the number of times the manner of drinking strategy was used was not associated with the likelihood of experiencing any of the six physical and behavioral negative consequences of alcohol use on a given drinking day, γ₀₂₀. At the person level, students who used the manner of drinking strategy more often across college were more likely to experience a hangover on a given drinking day than students who used the manner of drinking strategy less often, γ₀₀₂. The number of times the manner of drinking strategy was used across college was not associated with the likelihood of experiencing any of the other negative consequences of alcohol use on a given drinking day.

Models in Table 2 tested whether specific negative consequences of alcohol use were experienced less frequently on drinking days in which students used the serious harm reduction strategy. Students were 2.50 times less likely to have a sexual experience they regretted on drinking days they used serious harm reduction strategies than on days they did not, γ₂₀₀. However, there was no association between students’ use of a serious harm reduction strategy and the likelihood of finding themselves in a situation in which no one was sober enough to drive. In semesters in which students used a serious harm reduction strategy more often, students were more likely to have a sexual experience they regretted on a given drinking day compared to semesters this strategy was used less often, γ₀₂₀. The number of times students used a serious harm reduction strategy in a given semester was not associated with the likelihood of no one being sober enough to drive on a given drinking day that semester. At the person level, there were no associations between the frequency with which students used a serious harm reduction strategy and the likelihood of having no one sober to drive or having a sexual experience they regretted on a given drinking day, γ₀₀₂.

Aim 3: Is PBS use more strongly linked with fewer consequences on HID days compared to HED-only days?

To test whether associations between PBS use and negative consequences of alcohol use differed between HED-only and HID days, a daily-level PBS Use × HID Day interaction term, γ₄₀₀, was added to the models presented in Tables 1 and 2. Across the eight models, there were three statistically significant interactions, which are described below and presented visually in Figures 1 and 2.

Figure 1.

Use of manner of drinking strategies attenuates the association between high-intensity drinking (relative to heavy episodic drinking) and students’ likelihood of passing out.

Figure 2.

Use of serious harm reduction strategies attenuates the association between high-intensity drinking (relative to heavy episodic drinking) and students’ likelihood of having (a) no one sober enough to drive and (b) a regretted sexual experience.

The use of a manner of drinking strategy was negatively associated with the likelihood of passing out, and this association was stronger on HID days than on HED-only days (OR = 0.39, 95% CI = [0.19, 0.80], p < .01). On HED-only days, students had 48% lower odds of passing out when they used a manner of drinking strategy, compared to days when they did not (OR = 0.52, 95% CI = [0.31, 0.87], p < .05). On HID days, students had 80% lower odds of passing out when they used a manner of drinking strategy, compared to days when they did not (OR = 0.20, 95% CI = [0.11, 0.37], p < .001; Figure 1).

The association between using a serious harm reduction strategy and ending up in a situation in which there was no one sober to drive differed between HED-only and HID days (OR = 0.58, 95% CI = [0.35, 0.95], p < .05). On HED-only days, the odds of having no sober driver were greater when a serious harm reduction strategy was used, compared to days one was not used, but this positive association was non-significant (OR = 1.23, 95% CI = [0.81, 1.88], ns). On HID days, the odds of having no sober driver were less when a serious harm reduction strategy was used, compared to days one was not used, but this negative association was non-significant (OR = 0.71, 95% CI = [0.50, 1.02], ns). Therefore, although the magnitude and direction of the association between using a serious harm reduction strategy and no sober driver differed between HED-only and HID days, the use of a serious harm reduction strategy was not significantly associated with having no one sober to drive (Figure 2, plot a).

The use of a serious harm reduction strategy was negatively associated with the likelihood of having a regretted sexual experience, and this association was stronger on HID days than on HED-only days (OR = 0.13, 95% CI = [0.07, 0.24], p < .001). On HED-only days, students had 35% lower odds of having a sexual experience they regretted when they used a serious harm reduction strategy, compared to days when they did not (OR = 0.65, 95% CI = [0.44, 0.95], p < .05). On HID days, students had 92% lower odds of having a sexual experience they regretted when they used a serious harm reduction strategy compared to days when they did not (OR = 0.08, 95% CI = [0.04, 0.16], p < .001; Figure 2, plot b).

Discussion

High-intensity drinking is a critical public health concern in the U.S. (Patrick, 2016; White et al., 2006). Compared to the literature on risks associated with HID, far less is known about potential protective factors drinkers can use to reduce the likelihood of experiencing alcohol-related harms on HID occasions. The purpose of our study was to determine whether (1) HID days were associated with higher odds of negative consequences than HED-only days, (2) PBS were associated with lower odds of consequences, and (3) associations between PBS use and consequences differed between HED-only and HID days. Findings from our longitudinal, measurement-burst study of college students revealed that HID occasions were associated with greater risk for all harms measured (e.g., affected coordination, lost control) with the exception of engaging in sexual behavior that students later regretted. HID may be related to lower odds of regretted sex given links between heavy alcohol consumption and inhibited sexual response (see McKay, 2005 for a review). Findings generally converge with Patrick and colleagues’ (2016a) daily diary study comparing HID days (8+ drinks for women/10+ drinks for men) with non-HID (1 to 7 drinks for women/1 to 9 drinks for men) days but results here clarify that this increased risk also occurs relative to HED-only days. These results complement the broader literature on HID by highlighting that odds of experiencing consequences are much higher when drinking beyond the HED-only threshold, even contrasting different drinking occasions within the same individuals (Patrick et al., 2016a). In other words, results suggest that not only are high-intensity drinkers at greater risk as a group but also that occasions involving HID are associated with greater risk of acute harm.

Despite somewhat consistent associations between HID, PBS, and consequences assessed at the day level, differences emerged when examining these effects at the person- or semester-levels. First, although consequences were more likely on HID days compared to HED-only days in within-person daily comparisons, between-person comparisons between students who, on average across college, had more HID days and those who had fewer showed no differences. Although individuals who engage in HID relative to those who do not represent a vulnerable group of drinkers at risk for serious, potentially long-term consequences for alcohol use such as alcohol use disorder (e.g., Evans-Polce et al., 2017; Hingson et al., 2017; Linden-Carmichael et al., 2017), HID-related risk for the more acute, immediate types of negative consequences we measured are experienced within hours of drinking and thus better captured at the daily level. Further, statistically, students who engage in HID more often may still be at risk for more consequences than those who engage in HID less often, but once day-level risk is included in the multilevel model (i.e., where the risks occur), higher-level effects are no longer significant. Associations between HID and consequences were also generally non-significant at the semester-level as well, with the exception of students being less likely to lose control from drinking during semesters in which they engaged in HID more often. This finding may reflect tolerance to excessive drinking such that individuals lose control less often during periods of heavy drinking. Alternatively, if interpreted in the opposite direction, this finding may suggest that when students engaged in fewer HID days during a semester, they lost control more easily when they did engage in HID. Losing control is also subjective, and students who engage in HID more often and/or those who intend to have a ‘wild’ time may have a higher threshold.

Second, although PBS and the experience of a hangover were negatively associated on a day-level, individuals who used PBS more often on average across college were more likely to experience a hangover on a given day. This finding is somewhat consistent with prior work examining person-level PBS versus day-level PBS used (Pearson et al., 2013). It is plausible that more frequent PBS use may be indicative of more frequent heavy drinking. Thus, although it appears that PBS is protective on days when used, people who use these strategies more often may be more susceptible to experiencing acute consequences, such as hangovers, if they do not use PBS on a given day. Relatedly, during semesters in which students used more serious harm reduction strategies, they reported more regretted sex. These higher-level findings may indicate that students who experience more hangovers or regretted sex during a semester may be more likely to use PBS to try to mitigate these negative consequences from drinking.

Lastly, although not directly related to the main study aims, an interesting finding emerged when examining day of the week and negative consequences. Students were at greater odds of reporting being in bad physical shape the next day or engaging in sexual activity they regretted as a result of drinking on weekdays compared to drinking on weekends. Though alcohol use and consequences are typically higher on weekends (Del Boca et al., 2004; Maggs et al., 2011), this finding may reflect differential drinking patterns (e.g., drinking to cope) that are exhibited more strongly on weekdays (Studer, Baggio, Mohler-Kuo, Dermota, Daeppen, & Gmel, 2014). Also, specific to being in bad physical shape, the greater freedom to sleep late on weekends may help students avoid suffering from some physical effects of heavy drinking. Few daily diary studies have examined specific negative consequences across many weekdays and weekend days, suggesting future research is needed in this area.

Use of PBS (manner of drinking, serious harm reduction) was associated, for the most part, with lower odds of experiencing acute consequences. Similar to prior work examining differential associations between PBS use and consequences (Frank et al., 2012; Napper et al., 2014), manner of drinking strategies, such as avoiding drinking too quickly, were found to be more consistently associated with fewer consequences while serious harm reduction strategies, such as watching one’s drink, were less consistently associated with consequences. As suggested by Pearson and colleagues (2013) who found positive associations between serious harm reduction strategy use and negative consequences, it is possible that strategies such as using a designated driver is an indicator of intentions to engage in risky drinking behaviors.

Level of drinking moderated associations between PBS use and three of the eight negative consequences of alcohol use. For example, findings revealed that the negative association between using manner of drinking strategies and passing out was stronger on HID days compared to HED-only days. Passing out or blacking out from drinking have been shown to be among the most predictive of problematic drinking behaviors in college students (Read, Beattie, Chamberlain, & Merrill, 2008). Our findings suggest that manner of drinking strategies may be more effective in weakening this negative consequence on HID days than HED days. In part this may reflect the fact that the risk of passing out is higher when drinking is particularly excessive. Although moderate, temperate, or even no use might be preferable for health, harm reduction strategies by definition are targeted to minimize harm among otherwise problematic behaviors. Though serious harm reduction strategy use was not consistently associated with daily negative consequences in our between-person findings or others’ (e.g., Pearson et al.,2013), associations between use of such strategies and engaging in later-regretted sexual behavior were stronger on HID days than on HED-only days. These findings suggest that using PBS may help reduce the likelihood of experiencing some of the harms associated with engaging in HID.

Though the associations between PBS use and some negative consequences were stronger on HID days compared to HED-only days, it is important to note that the strength of the associations did not differ across levels of drinking for many other consequences (e.g., losing control, affecting coordination). Given prior work showing that HID drinkers report stronger drinking motivations and drink in more contexts than HED-only drinkers (Patrick et al., 2017; Terry-McElrath et al., 2017; White et al., 2016), drinking may serve different functions – not only between HID drinkers and HED-only drinkers, but also on occasions when individuals engage in HID. Some PBS may not be conducive to young adults’ drinking goals or plans on HID days and thus fail to impact their likelihood of protecting themselves from harm because they do not use these strategies. Given traditional-aged, residential undergraduate students’ high risk for problematic levels of drinking, there may be a need for researchers to develop strategies that address the elevated risks exhibited by high-intensity drinkers. An important first step would be to better specify the temporal associations between HID and PBS use within drinking occasions and across the transition to adulthood. Some recent longitudinal work has examined links between PBS, drinking intentions, and alcohol use outcomes (e.g., Grazioli et al., 2015) but more fine-grained information is needed. For example, we need to know which behavior occurs first – do individuals plan in advance to avoid drinking too quickly or arrange a designated driver before going out, or do these strategies emerge (or get abandoned) after a drinking occasion is initiated? Answers to such questions could elucidate the amount individuals intend to drink and/or their intended (and followed-through) PBS use when engaging in HID and move us toward the ultimate goal of helping individuals reduce harms while drinking.

Strengths include the measurement-burst design assessing drinking across multiple semesters, yielding many HID and HED-only days and longitudinal data across college. Such a design allowed us to compare drinking occasions at different thresholds of drinking for the same individual. One limitation, however, is that data were collected from a single university in the U.S., limiting generalizability. The study also relied solely on self-reports. Relatedly, our daily measures of PBS were limited to single-items to reduce participant burden. Future research may consider collecting daily (or more frequent) reports using the full PBSS scale (Martens et al., 2005). Also, it should be noted that, similar to prior daily work examining PBS (e.g., Braitman et al., 2017; Pearson et al., 2013), daily PBS use was reported retrospectively and at the same time as the number of drinks and actual consequences experienced. Given the retrospective (i.e., next-day) nature of the data collection, participants may have had difficulty separately reporting their intended strategies, actual drinking, and consequences. Future work would benefit from participants reporting on PBS prior to or while drinking to intended harm reduction strategies. Further, while the daily study is likely to have reduced recall biases, accuracy especially related to number of drinks consumed may be reduced as a drinking event progresses. Wearable alcohol sensors that non-invasively measure actual intoxication through a participant’s skin may provide more nuanced, valid, and continuous data in future work (see Greenfield, Bond, & Kerr, 2014).

Our findings suggest that HID occasions are linked with greater odds of experiencing a variety of serious negative alcohol-related consequences, that PBS use is associated with decreased consequences, and that manner of drinking PBS and serious harm reduction PBS may help college students reduce particular harms on HID occasions. Efforts focused on modifying drinking behavior and planning strategies to minimize harm may be an important point of consideration in intervention work targeting heaviest alcohol consumption levels, but given that PBS did not help reduce many consequences associated with HID, it may be useful for future research to identify additional types of strategies drinkers can use when drinking at such high levels.