Dietary self-monitoring is associated with increased likelihood of problematic alcohol use among college students

Abstract

Objective:

To assess the relationship between dietary self-monitoring and problematic alcohol use including binge drinking, increased body confidence while drinking, and restricting food intake to compensate for alcohol consumption, among college students.

Participants:

Undergraduate and graduate students from 12US colleges participating in the Healthy Bodies Study in 2013–2014 and 2014–2015 school years (n = 10,133).

Methods:

Weighted prevalence was calculated for dietary self-monitoring and problematic alcohol use. Gender-stratified logistic regressions were used to assess relationships.

Results:

Knowing nutrition facts was associated with restricting to compensate for alcohol consumption among women (OR = 2.42, p < .0001) and men (OR = 1.64, p = .002). Among women, knowing and counting calories predicted all problematic alcohol use behaviors. Among men, knowing calories was associated with restricting to compensate (OR = 2.69, p < .0001) and counting calories was associated with restricting to compensate (OR = 5.10, p < .0001) and increased body confidence while drinking (OR = 2.25, p < .0001).

Conclusions:

Dietary self-monitoring predicts problematic alcohol use among college students, particularly women.

Introduction

Alcohol use is highly prevalent on college campuses, with approximately two-thirds of undergraduate college students drinking alcohol in the last month.^1,2 Moreover, about one-third of college students report binge drinking in the past month, defined as four or five or more drinks in a given period.^1,2 Binge drinking is associated with poor academic performance³ alcohol dependence or alcohol use disorders,⁴ as well as a number of negative health consequences such as being injured,^2,5 and even death.^5,6 Because alcohol is high in calories, it would be assumed that those concerned with their weight would not engage in binge drinking. However, research to support this assumption has been mixed. While restrained eating and disordered eating behaviors are associated with binge drinking among women^7,8 weight concern is not.⁹ Moreover, it could be that individuals who are more likely to engage in disordered eating are drinking alcohol because they have increased body confidence while drinking. This could be particularly harmful as it they would receive positive reinforcement for drinking alcohol which may lead to dependence, as we know that eating disorders and alcohol use disorders often co-occur.^10,11

Another problematic behavior that often accompanies binge drinking is skipping meals or eating less to compensate for drinking alcohol, a practice often referred to as “drunkorexia.”^12–14 When such dietary restriction occurs before drinking alcohol, it can lead to increased risk of intoxication, thus increasing the likelihood of experiencing negative consequences of alcohol use including memory loss, getting injured, or getting into a fight.¹⁵ Approximately 38% of college drinkers report restricting food before drinking alcohol, and this practice is more common among those with weight concern.¹²

Restrictive eating commonly involves knowing or tracking the nutrition facts and/or calories in the foods and drinks one consumes, behaviors broadly known as dietary self-monitoring. The use of dietary self-monitoring has long been dictated as a primary behavioral intervention for the treatment of higher body weights,¹⁶ but in recent years it has become mainstream with the advent of smartphone and Web applications aimed at dietary self-monitoring. As of 2018, MyFitnessPal, the most common dietary self-monitoring smartphone application, had 19.1 million users a month.¹⁷ College students are likely among the most frequent users of this technology, as young adults and those with some college education are most likely to use applications for dietary self-monitoring.^18,19 One study found that 13.8% of college students count their calories, though the prevalence of other forms of dietary self-monitoring have not been explored to our knowledge.²⁰ Prior studies examining calorie counting among college students have found that counting calories is associated with higher rates of eating disorder symptomology.^20–22 Further, because individuals engaging in restriction to compensate for alcohol consumption are more likely to be weight concerned,¹² they may be more likely to engage in dietary self-monitoring for weight control purposes. Additionally, because higher alcohol consumption is seen among college students with low body satisfaction and those who say drinking gives them more confidence,^23–25 using alcohol to increase body confidence may be particularly common among students who engage in dietary self-monitoring.

Dietary self-monitoring and problematic alcohol use are both ubiquitous among college students. Prior research has found that both dietary self-monitoring and binge drinking are associated with eating disorder symptomology among college students.^7,8,20–22 Because a large proportion of college students engaging in dietary self-monitoring are doing so for weight control,²² individuals who engage in dietary self-monitoring may be at increased risk for problematic alcohol use due to elevated desire to control their weight that arises with the use of dietary self-monitoring. To our knowledge, however, no research has examined the relationship between these two common behaviors. The objective of this study was therefore to examine relationships between different forms of dietary self-monitoring and problematic alcohol use among US college students. We hypothesized that dietary self-monitoring would be associated with increased risk of problematic alcohol use.

Materials and methods

Data

Data for the present study came from the Healthy Bodies Study (HBS). HBS was a Web-based, cross-sectional survey administered during the 2013–2014 and 2014–2015 academic years to undergraduate and graduate students at participating institutions. HBS data comes from 12 US colleges and universities. HBS participants were recruited via email which was sent to up to 4,000 randomly selected graduate and undergraduate students per participating institution. The only exclusion criteria to participate in HBS was that students had to be 18 years of age to participate, there were no additional exclusionary criteria for the present study. In the email, participants were invited to take the survey via Qualtrics which took approximately 15 min to complete. Students that did not respond were sent reminder emails. The final sample size was 10,133 students. The Institutional Review Board (IRB) approved HBS at the participating institutions. More information about the HBS methods can be found elsewhere.²⁶

Nonresponse analysis

The response rates for HBS were 19% and 27%, respectively, in the 2013–2014 and 2014–2015 school years. This response rate is comparable with other online surveys distributed to college students.^15,27,28 However, there is the potential that those that opted to participate in the survey may differ from the original random sample recruited to participate. To correct for potential difference between responders and non-responders, sample probability weights were calculated using characteristics obtained from the institutional Registrars including gender, academic level, race/ethnicity, and grade point average. To ensure that estimates were reflective of the original population, respondents who were underrepresented in the response to the original survey had larger weights.

Measures

Outcomes

Binge drinking.

As part of the original HBS, three aspects of problematic alcohol use were assessed and were used for the present study. Binge drinking was assessed with the following survey questions: “Over the last 2 weeks, did you drink any alcohol?,” those that answered “yes” were asked, “Over the last 2 weeks, about how many times did you have four alcohol drinks in a row?” for women, and “Over the last 2 weeks, about how many times did you have five alcoholic drinks in a row?” for men. The response options ranged from “10 or more times” to “0 times” and “I don’t know.” This definition was adapted from the National Institute on Alcohol Abuse and Alcoholism (NIAAA)’s definition of binge drinking.²⁹ We created a binary variable of one or more times binge drinking in the last 2 weeks.

Body confidence while drinking alcohol.

Next, we assessed increased body confidence while drinking alcohol using the item, “When I drink alcohol, I feel better about my body.” Response options were a 4-point Likert scale ranging from “very like me” to “not like me.” Those who said it was somewhat, like, or very like them were considered those who feel more confident about their body when drinking alcohol.

Restricting food to compensate for alcohol consumption.

The third problematic alcohol use measured was restricting food to compensate for alcohol consumption, which was assessed using the item, “If I know that I will be drinking alcohol, I skip meals or eat less on that day or the next day.” The response options consisted of the same 4-point Likert scale, and those who said that this was somewhat like them, like them, or very like them, were considered positive for restricting to compensate for alcohol consumption.

Predictors

Three aspects of dietary self-monitoring were assessed in the original HBS and were utilized for the current study: knowing nutrition facts, knowing calories, and counting calories. To assess these constructs, the following measures were used: “How often do you typically know the nutrition facts (eg, fat, fiber, carbohydrates, protein) about the foods and drinks you consume before you consume them?,” “How often do you typically know the number of calories in the foods and drinks you consume before you consume them?,” and “How often do you typically count the calories that you consume?.” Response options were: “always,” “usually,” “sometimes,” “rarely,” and “never.” Binary responses were created for each predictor with those indicating usually or always considered frequently engaging in the respective dietary self-monitoring behavior.

Statistical analysis

All analyses were stratified by gender because prior research has shown that prevalence of binge drinking and dietary restriction prior to drinking alcohol differs by gender,^12,30 as does confidence while drinking alcohol^23–25 among college students. We were not powered to assess gender minorities; therefore, gender minority students were included in the overall sample but not gender-stratified analyses. Weighted prevalence was calculated for all demographic, predictor, and outcome variables. Chi-square tests were used to assess differences in prevalence of predictor and outcome variables by gender. To examine the individual relationships between predictors and outcomes, we conducted logistic regressions to determine odds ratios (OR) and 95% confidence intervals (CI). Race/ethnicity, age, body mass index (BMI), and parental education were included as covariates in all relevant models to increase model precision because alcohol use differs by race/ethnicity, age, and parental education.² BMI is included because public health recommendations for dietary self-monitoring are based on BMI. Parental education was accounted for because parental education has been shown to be associated with binge drinking. Results were considered statistically significant if p < .05. All analyses were conducted using SAS 9.4.

Results

The study sample was comprised of 10,133 students, with an average age of 23.4 years (Table 1). The sample was 54.4% women, 44.4% men, and 1.2% gender minorities. Over two-thirds identified as white (66.8%), 11.4% as Asian, 9.4% as Hispanic or Latinx, 4.4% as African American, and 8.1% as another racial identity. Approximately, one-fourth of students were first-generation college students, with 10.1% having parents with a high school degree or less, 17.2% had some college or an associate’s degree, and nearly three-fourths of students had at least one parent with a bachelor’s degree (30.3%) or graduate degree (42.4%). The average BMI was 23.8, with 31.5% of students having a BMI ≥25 and may have been given clinical recommendation to engage in dietary self-monitoring.

Table 1.

Sample characteristics, weighted prevalence %.

	Overall (n = 10,133)	Women (n = 6,961)	Men (n = 3,049)
Age, mean (SE)	23.4 (0.1)	23.1 (0.1)	23.7 (0.1)
Gender
Women	54.4	–	–
Male	44.4	–	–
Gender minority	1.2	–	–
Race/ethnicity
White	66.8	67.5	66.2
Latinx	9.4	9.6	9.0
African American	4.4	4.6	4.3
Asian	11.4	10.2	12.8
Other	8.1	8.1	7.7
Parental education
High school or less	10.1	10.1	10.0
Some college or associates	17.2	19.1	14.9
Bachelor’s degree	30.3	30.0	30.6
Graduate degree	42.4	40.8	44.5
BMI, mean (SE)	24.0 (0.1)	23.6 (0.1)	24.6 (0.1)

As shown in Figure 1, 44.2% of the sample indicated that they usually or always knew the nutrition facts of the foods and drinks they consumed before they consumed them, 33.7% knew the calories of the foods and drinks they consumed before they consumed them, and 12.8% count calories. The prevalence of knowing nutrition facts did not differ significantly by gender, with 44.8% of women and 43.3% of men endorsing this form of dietary self-monitoring (p = .24). Knowing calories, however, did differ by gender (p < .0001) with women having a higher prevalence of knowing the calories in the foods/drinks they consume compared to men (35.7% compared to 31.1%, respectively). Counting calories also differed by gender (p < .0001), being more prevalent among women (14.8%) than men (10.4%).

Figure 1.

Weighted prevalence (%) of exposure and outcome variables.

Prevalence of the problematic alcohol use was also high, particularly binge drinking, with 41.9% of the sample engaging in the behavior in the past two weeks. Additionally, 18.8% endorsed feeling more confident about their body while drinking, and 19.1% reported restricting their food intake to compensate for their alcohol consumption. Among those that endorsed drinking at all in the two weeks prior to survey, fewer women (63.5%) than men (66.7%) reported binge drinking (p = .02). Feeling more confident about their body when drinking was more common among women (20.4%) than men (16.9%) (p < .01), as was restricting to compensate for alcohol consumption (p < .0001), with 22.9% of women and 14.2% of men reporting engaging in this behavior.

Knowing nutrition facts

Results from the logistic regression analyses can be found in Table 2.

Table 2.

Adjusted odds ratios with 95% confidence intervals.

	Overall	Women	Males
Knowing nutrition facts
Binge drinking	1.12 (1.02, 1.24)	1.18 (1.05, 1.32)	1.04 (0.88, 1.25)
Confidence	1.04 (0.90, 1.20)	1.20 (1.02, 1.41)	0.81 (0.61, 0.97)
Restrict to compensate	2.18 (1.81, 2.45)	2.42 (2.07, 2.84)	1.64 (1.20, 2.24)
Knowing calories
Binge drinking	1.10 (0.99, 1.22)	1.22 (1.09, 1.37)	0.94 (0.78, 1.14)
Confidence	1.32 (1.14, 1.54)	1.41 (1.20, 1.66)	1.15 (0.86, 1.55)
Restrict to compensate	3.36 (2.89, 3.91)	3.79 (3.23, 4.45)	2.69 (1.97, 3.66)
Counting calories
Binge drinking	1.22 (1.06, 1.40)	1.33 (1.14, 1.55)	1.02 (0.77, 1.35)
Confidence	1.91 (1.58, 2.32)	1.74 (1.41, 2.14)	2.25 (1.51, 3.35)
Restrict to compensate	5.28 (4.40, 6.33)	5.36 (4.41, 6.51)	5.10 (3.47, 7.50)

Bold text indicates statistical significance (p < .05).

Knowing nutrition facts was associated with binge drinking in women (OR = 1.18, 95% CI: 1.05, 1.32) but not men. Increased body confidence while drinking alcohol had similar results, as it was associated with knowing nutrition facts among women (OR = 1.20, 95% CI: 1.02, 1.41) but not for men. Though nonsignificant, the direction of the association between knowing nutrition facts and increased body confidence when drinking alcohol differed by gender. Knowing nutrition facts did significantly predict restricting to compensate for alcohol consumption for both women and men, though the relationship was stronger for women (women OR = 2.42, 95% CI: 2.07, 2.84; men OR = 1.64, 95% CI:1.20, 2.24).

Knowing calories

Among women, knowing calories significantly predicted binge drinking (OR = 1.22, 95% CI: 1.09, 1.37), body confidence while drinking alcohol (OR = 1.41, 95% CI: 1.20, 1.66), and restricting to compensate for alcohol consumption (OR = 3.79, 95% CI: 3.23, 4.45). However, knowing calories only predicted restricting to compensate for alcohol consumption among men (OR = 2.69, 95% CI: 1.97, 3.66).

Counting calories

Among men, counting calories was also associated with body confidence (OR = 2.25, 95% CI: 1.51, 3.35) and restricting to compensate for alcohol consumption (OR = 5.10, 95% CI: 3.47, 7.50), but was not associated with binge drinking. Conversely, counting calories was associated with binge drinking (OR = 1.33, 95% CI: 1.14, 1.55) among women as well as body confidence while drinking (OR = 1.74, 95% CI: 1.41, 2.14) and restricting to compensate for alcohol consumption (OR = 5.36, 95% CI: 4.41, 6.51). Body confidence while drinking alcohol had a larger odds ratio for men than women, which was the only relationship where this held true.

Discussion

The objective of this multi-campus study was to examine the cross-sectional associations between dietary self-monitoring and problematic alcohol use among college students. Overall, we found that all forms of dietary self-monitoring and problematic alcohol use were common in this sample. Further, knowing calories and counting calories were more common among women than men. Prevalence of binge drinking did not differ by gender, though body confidence while drinking alcohol and restricting to compensate for alcohol consumption were more common among women. Overall, adjusted logistic regressions revealed that dietary self-monitoring was associated with problematic alcohol use for both men and women.

Findings from this study build on the existing research on both problematic alcohol use and dietary self-monitoring among college students. The prevalence of binge drinking^1,2,31,32 and restricting to compensate for alcohol intake³³ are commensurate with prior studies among college students. This is the first known study to examine relationships between dietary self-monitoring and problematic alcohol use, although one study found that college students who regularly use nutrition labels had a lower intake of added sugar, alcohol, and solid fat.³⁴ That study looked at alcohol use in general, not specifically problematic alcohol use, and grouped added sugar, alcohol, and fat into one category. It has been shown that college students frequently consume low-fat and low-calorie foods to compensate for drinking,³⁰ therefore, lower alcohol consumption may not be driving the association. Prior studies have also found that elevated weight concern was associated with restricting food intake to compensate for alcohol consumption,¹² but among women binge drinking is not associated with weight concern.^9,35 In one study, calorie counting was associated with eating concern and dietary restraint but not weight concern.²⁰ Thus, the association between dietary self-monitoring and problematic alcohol use may be driven by increases in eating concerns due to dietary self-monitoring rather than weight concerns. This may explain why we found more consistent and stronger findings for the outcome of restricting food intake to compensate for drinking relative to the binge drinking outcome. Additionally, because restricting food intake to compensate for alcohol consumption is a disordered weight control behavior, our findings align with prior studies that have found dietary self-monitoring is associated with eating disorder risk in this population^20–22 and adds to the literature pertaining to potential harms of dietary self-monitoring.

The present study had several strengths including a large sample size that allowed us to examine gender specific associations between dietary self-monitoring and problematic alcohol use. Additionally, dietary self-monitoring was assessed using several items that allowed us to look further at how different forms of dietary self-monitoring including knowing nutrition facts and knowing calories in the foods/drinks that individuals consume and counting calories may differentially be associated with problematic alcohol use. Similarly, various components of problematic alcohol use were examined which allowed us to explore behaviors that are well-known to be present in this population, such as binge drinking and restricting to compensate for alcohol consumption, but also body confidence while drinking, which to our knowledge has not been previously studied. Although the assessment of multiple components of dietary self-monitoring and problematic alcohol use was a strength of the study, the measures used were single-item measures and their validity and reliability have not been established nor were they tested in the present study. Moreover, our measure of restriction asked about restricting before or after alcohol consumption, which, while both highly prevalent behaviors, may have different public health implications.^9,30 The study was also cross-sectional, and therefore, the temporality of the relationships between predictors and outcomes cannot be established. Another limitation is that the response rates across the two academic years were 19% and 27%, although we attempted to correct for this with our probability sampling weights. Additionally, because there was a small number of people in the population that identified as a gender minority, and because it is possible that they may have very different associations and we were not powered to look at them independently, they were excluded from analyses.

Future research is needed to elucidate the relationship between dietary self-monitoring and problematic alcohol use. Longitudinal studies may yield more information on the temporality of these relationships and could provide insight into potential mechanisms. Additionally, larger studies are needed to examine these relationships among understudied populations such as gender minority students, as we were not powered to assess these relationships specifically among gender minorities.

The current study makes an important contribution to the literature surrounding problematic alcohol use and dietary self-monitoring, particularly adding to the literature about the potential consequences of dietary self-monitoring. Specifically, we found that dietary self-monitoring was associated with problematic alcohol use among a large sample of US college students. Our findings suggest that knowing and counting calories may be more harmful for problematic alcohol use than knowing nutrition facts, particularly for women. Colleges and universities employing alcohol interventions to reduce problematic alcohol use may be able to better target at-risk students by aiming their interventions at those who engage in dietary self-monitoring, specifically those that know or count calories. Moreover, colleges engaging in efforts to reduce problematic alcohol use among college students may benefit from including information in existing interventions on the effects of dietary self-monitoring and the dangers of restricting dietary intake while consuming alcohol. For clinicians working on college campuses, asking about dietary self-monitoring may help indicate students who are high risk for problematic alcohol use.