Advertisement

Binge Drinking and Alcohol Problems Among Moderate Average-Level Drinkers

      Introduction

      A significant amount of binge drinking among adults escapes public health scrutiny because it occurs among individuals who drink at a moderate average level. This observational study examined the role of a binge pattern of drinking in predicting alcohol problems among moderate drinkers in a U.S. national sample of adults.

      Methods

      Participants were 1,229 current drinkers aged ≥30 years from 2 waves of the study of Midlife Development in the United States, with a 9-year time lag (2004–2015) (analyzed in 2021‒2022). Negative binomial regression analyses were used to examine the number of alcohol problems, and binary logistic regression analyses were used to examine multiple (≥2) alcohol problems.

      Results

      Independent of the average level of drinking, binge drinking was linked with an almost 3 times increase in the number of concurrent alcohol problems and a 40% increase in the number of alcohol problems prospectively 9 years later. Moderate average level drinkers accounted for most cases of binge drinking and multiple alcohol problems. Among moderate drinkers, binge drinking was linked with a close to 5 times increase in concurrent multiple alcohol problems and a >2 times increase in multiple alcohol problems prospectively 9 years later.

      Conclusions

      These results substantially broaden an increasing recognition that binge drinking is a public health concern among adults. Moderate average-level drinkers should be included in efforts to reduce alcohol problems in adults. These findings are applicable to primary and secondary prevention of alcohol problems with the potential to advance population health.

      INTRODUCTION

      Although research on binge drinking (≥5 drinks on the same occasion) has focused primarily on adolescents and college students, a pattern of binge drinking is prevalent across the adult lifespan. In fact, most binge drinking occurs among adults aged ≥30 years.
      • Kanny D
      • Naimi TS
      • Liu Y
      • Lu H
      • Brewer RD.
      Annual total binge drinks consumed by U.S. adults, 2015.
      Moreover, the prevalence of binge drinking is increasing among adults, including adults aged ≥50 years.
      • Azagba S
      • Shan L
      • Latham K
      • Manzione L.
      Trends in binge and heavy drinking among adults in the United States, 2011-2017.
      ,
      • Grucza RA
      • Sher KJ
      • Kerr WC
      • et al.
      Trends in adult alcohol use and binge drinking in the early 21st-century United States: a meta-analysis of 6 national survey series.
      Stressors associated with the coronavirus disease 2019 (COVID-19) pandemic are increasing excess daily drinking in adults.
      • Daly M
      • Robinson E.
      High-risk drinking in midlife before versus during the COVID-19 crisis: longitudinal evidence from the United Kingdom.
      Adults who engage in a pattern of binge drinking have an increased risk of alcohol problems.
      • Greenfield TK
      • Ye Y
      • Bond J
      • et al.
      Risks of alcohol use disorders related to drinking patterns in the U.S. general population.
      ,
      • Linden-Carmichael AN
      • Vasilenko SA
      • Lanza ST
      • Maggs JL.
      High-intensity drinking versus heavy episodic drinking: prevalence rates and relative odds of alcohol use disorder across adulthood.
      However, because a binge drinking pattern is confounded with an average level of consumption,
      • Naimi TS
      • Xuan Z
      • Brown DW
      • Saitz R.
      Confounding and studies of ‘moderate’ alcohol consumption: the case of drinking frequency and implications for low-risk drinking guidelines.
      ,
      • Rehm J
      • Gmel Sr, GE
      • Gmel G
      • et al.
      The relationship between different dimensions of alcohol use and the burden of disease-an update.
      it is not clear how binge drinking uniquely contributes to alcohol problems, independent of the average level of drinking.
      • Holahan CJ
      • Brennan PL
      • Schutte KK
      • Holahan CK
      • Hixon JG
      • Moos RH.
      Late-life drinking problems: the predictive roles of drinking level vs. drinking pattern.
      Focusing only on moderate average consumption masks the divergent underlying patterns of drinking, from regular low consumption to irregular binge drinking.
      • Naimi TS
      • Xuan Z
      • Brown DW
      • Saitz R.
      Confounding and studies of ‘moderate’ alcohol consumption: the case of drinking frequency and implications for low-risk drinking guidelines.
      In fact, most epidemiologic studies focus solely on average level of drinking, overlooking occasions of high consumption.
      • Naimi TS
      • Xuan Z
      • Brown DW
      • Saitz R.
      Confounding and studies of ‘moderate’ alcohol consumption: the case of drinking frequency and implications for low-risk drinking guidelines.
      ,
      • Rehm J
      • Gmel Sr, GE
      • Gmel G
      • et al.
      The relationship between different dimensions of alcohol use and the burden of disease-an update.
      Similarly, mass media uncritically espouse the health advantages of moderate alcohol consumption.
      • Oppenheimer GM
      • Bayer R.
      Is moderate drinking protective against heart disease? The science, politics and history of a public health conundrum.
      Thus, many drinkers assume that moderate average consumption is safe, regardless of drinking pattern.
      Especially troubling, much binge drinking among adults escapes public health scrutiny because it occurs among individuals who drink at a moderate average level (average of not >1 drink a day for women and 2 drinks a day for men).
      • Naimi TS
      • Brewer RD
      • Mokdad A
      • Denny C
      • Serdula MK
      • Marks JS.
      Binge drinking among U.S. adults.
      Drinking within U.S. National Institute on Alcohol Abuse and Alcoholism low-risk guidelines for average consumption (not >1 drink a day for women and 2 drinks a day for men) leaves sufficient opportunity for risky drinking on particular occasions.
      • Naimi TS
      • Xuan Z
      • Brown DW
      • Saitz R.
      Confounding and studies of ‘moderate’ alcohol consumption: the case of drinking frequency and implications for low-risk drinking guidelines.
      Using data from the U.S. Behavioral Risk Factor Surveillance System, Naimi and colleagues
      • Naimi TS
      • Brewer RD
      • Mokdad A
      • Denny C
      • Serdula MK
      • Marks JS.
      Binge drinking among U.S. adults.
      found that moderate drinkers accounted for half of the occasions of binge drinking. Furthermore, using data from the Canadian Institute for Health Information, Sherk et al.
      • Sherk A
      • Thomas G
      • Churchill S
      • Stockwell T.
      Does drinking within low-risk guidelines prevent harm? Implications for high-income countries using the International Model of Alcohol Harms and Policies.
      found that individuals drinking within average consumption guidelines accounted for a substantial proportion of alcohol's social harms.
      This study examined the role of a binge pattern of drinking in predicting alcohol problems among moderate average-level drinkers in a U.S. national sample of adults. Participants were 1,229 current drinkers aged ≥30 years from 2 waves of the study of Midlife Development in the United States (MIDUS) with a 9-year time lag.

      Ryff CD, Almeida DM, Ayanian JS, et al. Midlife in the United States (MIDUS 2), 2004-2006 (ICPSR 4652). Ann Arbor, MI: Inter-University Consortium for Political and Social Research. https://doi.org/10.3886/ICPSR04652.v8. Published September 15, 2021.

      Preliminary analyses examined the unique contribution of a regular versus binge pattern of drinking, independent of a moderate versus high average level of drinking, in predicting alcohol problems in the full sample of drinkers. Next, the association of a pattern of binge drinking with multiple alcohol problems was examined among moderate average drinkers. Each set of analyses included cross-sectional analyses and prospective analyses across 9 years.

      METHODS

       Study Sample

      The MIDUS study investigated health and well-being in a U.S. national sample of non-institutionalized, English-speaking adults. The study was initiated by the MacArthur Midlife Research Network with later funding from the National Institute on Aging. Baseline data for the current analyses are from Wave 2 of the MIDUS study, which introduced the measure of binge drinking in the survey. Follow-up data are from Wave 3 of the MIDUS study, 9 years later. Baseline data were collected between 2004 and 2006, and follow-up data were collected between 2013 and 2015.
      The MIDUS sample of U.S. adults was selected by a random-digit-dialing procedure with metropolitan oversampling.

      Ryff CD, Almeida DM, Ayanian JS, et al. Midlife in the United States (MIDUS 2), 2004-2006 (ICPSR 4652). Ann Arbor, MI: Inter-University Consortium for Political and Social Research. https://doi.org/10.3886/ICPSR04652.v8. Published September 15, 2021.

      Each wave of data collection used a phone interview followed by a self-administered questionnaire. Oral consent for participation was obtained by telephone at first contact. Participation rate (adjusted for mortality) was 75% at baseline and 77% at the 9-year follow-up.
      • Radler BT
      • Ryff CD.
      Who participates? Accounting for longitudinal retention in the MIDUS national study of health and well-being.
      ,

      Ryff C, Almeida D, Ayanian J, et al. Midlife in the United States (MIDUS 3), 2013-2014. Ann Arbor, MI: Inter-University Consortium for Political and Social Research. https://doi.org/10.3886/ICPSR36346.v7. Published April 30, 2019.

      Baseline participants in this study were 1,229 current drinkers who provided complete data on predictor variables and covariates. In this study, follow-up data on alcohol problems at 9 years (adjusted for mortality) were available for 70% (767/1,229) of baseline participants. This study did not require IRB approval because it involved secondary analyses of a publicly available, fully deidentified data set.

       Measures

      Alcohol consumption was measured at baseline. Initially, respondents were asked: During the past month, have you had at least one drink of any alcoholic beverage such as beer, wine, wine coolers, or liquor? Respondents answering yes were defined as current drinkers. Respondents answering no were defined as noncurrent drinkers and were excluded from this study. For other research using these MIDUS measures of alcohol consumption, see the studies by Goldwater and colleagues
      • Goldwater D
      • Karlamangla A
      • Merkin SS
      • Seeman T.
      Compared to non-drinkers, individuals who drink alcohol have a more favorable multisystem physiologic risk score as measured by allostatic load.
      and Richardson and Boutwell.
      • Richardson GB
      • Boutwell BB.
      Decomposition of mean sex differences in alcohol use within a genetic factor model.
      To index a moderate versus high average level of alcohol consumption, respondents were first asked: During the past month, how often did you drink any alcoholic beverages, on the average? Responses were scaled to the average number of drinking days per week. Next, to measure the average number of drinks on a drinking day, respondents were asked: On the days when you drank, about how many drinks did you drink on the average? One drink was defined as “a bottle of beer, a wine cooler, a glass of wine, a shot of liquor, or a mixed drink.” To derive a composite index of average daily level of drinking, average drinking days per week and average drinks on a drinking day were multiplied, and the product was divided by 7.
      Moderate drinking was defined following the National Institute on Alcohol Abuse and Alcoholism cut off for low-risk drinking.
      Alcohol Research: Current Reviews Editorial Staff. Drinking patterns and their definitions.
      Specifically, moderate drinking (score=0) was defined as an average level of drinking ≤1 drink/day for women and ≤2 drinks/day for men; high drinking (score=1) was defined as an average daily level of drinking >1 drink per day for women and >2 drinks/day for men. This definition is commonly used in alcohol studies to define a group of moderate as well as low-to-moderate drinkers.
      • Chiva-Blanch G
      • Badimon L.
      Benefits and risks of moderate alcohol consumption on cardiovascular disease: current findings and controversies.
      ,
      • Zhang R
      • Shen L
      • Miles T
      • et al.
      Association of low to moderate alcohol drinking with cognitive functions from middle to older age among U.S. adults.
      A separate question indexed a regular versus binge pattern of drinking on one occasion. Participants were asked: Considering all types of alcoholic beverages, how many times during the past month did you have 5 or more drinks on the same occasion? An occasion was defined as “drinks in a row, or in a short period of time.” For both women and men, the MIDUS surveys indexed a regular pattern of drinking as <5 drinks on the same occasion and binge drinking as ≥5 drinks on the same occasion. This is consistent with the definition used at the time in the U.S. National Survey on Drug Use and Health. Following previous research,
      • Holahan CJ
      • Brennan PL
      • Schutte KK
      • Holahan CK
      • Hixon JG
      • Moos RH.
      Late-life drinking problems: the predictive roles of drinking level vs. drinking pattern.
      ,
      • Goldwater D
      • Karlamangla A
      • Merkin SS
      • Seeman T.
      Compared to non-drinkers, individuals who drink alcohol have a more favorable multisystem physiologic risk score as measured by allostatic load.
      a binary measure was used to contrast no instances of binge drinking (score=0) with 1-or-more instances of binge drinking (score=1).
      Alcohol problems in the past 12 months were indexed at baseline and at the 9-year follow-up by the alcohol Dependence scale of the Composite International Diagnostic Interview-Short Form.
      • Kessler RC
      • Andrews G
      • Mroczek D
      • Ustun B
      • Wittchen H.
      The World Health Organization composite international diagnostic interview short-form (CIDI-SF).
      The scale consists of 7 items: experiencing effects of alcohol in a situation that increased the chances of getting hurt; emotional or psychological problems from alcohol; a desire to use alcohol that you could not resist; a month or more when you spent a great deal of time using alcohol; having to use more alcohol to get the same effect; using much larger amounts of alcohol than intended; experiencing the effects of alcohol at work, school, or caring for children. Following Magidson et al.,
      • Magidson JF
      • Robustelli BL
      • Seitz-Brown CJ
      • Whisman MA.
      Activity enjoyment, not frequency, is associated with alcohol-related problems and heavy episodic drinking.
      for each item, no problems was given a score of 0, and any problems was given a score of 1. The MIDUS surveys included 6 alcohol problems at baseline (use in hazardous situations was not included) and all 7 problems at follow-up. For other research using the MIDUS measure of alcohol problems, see the study by Magidson and colleagues
      • Magidson JF
      • Robustelli BL
      • Seitz-Brown CJ
      • Whisman MA.
      Activity enjoyment, not frequency, is associated with alcohol-related problems and heavy episodic drinking.
      and Glei and Weinstein.
      • Glei DA
      • Weinstein M.
      Drug and alcohol abuse: the role of economic insecurity.
      To increase construct validity, 2 measures of alcohol problems were developed. Following Magidson et al.,
      • Magidson JF
      • Robustelli BL
      • Seitz-Brown CJ
      • Whisman MA.
      Activity enjoyment, not frequency, is associated with alcohol-related problems and heavy episodic drinking.
      a continuous measure indexed the total number of alcohol problems. Following Richardson and Boutwell
      • Richardson GB
      • Boutwell BB.
      Decomposition of mean sex differences in alcohol use within a genetic factor model.
      and Glei and Weinstein,
      • Glei DA
      • Weinstein M.
      Drug and alcohol abuse: the role of economic insecurity.
      a binary measure indexed multiple alcohol problems. To link our definition of multiple alcohol problems to diagnostic criteria for alcohol use disorder,
      • Kranzler HR
      • Soyka M.
      Diagnosis and pharmacotherapy of alcohol use disorder: a review.
      a cut off of 2 was used to index multiple alcohol problems, contrasting ≤1 problem (score=0) with ≥2 problems (score=1).

       Statistical Analysis

      Analyses used Mplus, version 8.1.
      • Muthén LK
      • Muthén BO.
      Mplus Statistical Analysis With Latent Variables: User’s Guide.
      Following previous research predicting alcohol problems,
      • De Bruyn S
      • Wouters E
      • Ponnet K
      • et al.
      Problem drinking among Flemish students: beverage type, early drinking onset and negative personal & social consequences.
      • Anderson Goodell EM
      • Johnson RM
      • Latkin CA
      • Homish DL
      • Homish GG
      Risk and protective effects of social networks on alcohol use problems among Army Reserve and National Guard soldiers.
      • Martins JS
      • Joyner KJ
      • McCarthy DM
      • Morris DH
      • Patrick CJ
      • Bartholow BD.
      Differential brain responses to alcohol-related and natural rewards are associated with alcohol use and problems: evidence for reward dysregulation.
      negative binomial regression analyses were used to examine the association between binge drinking and the continuous measure of the number of alcohol problems, controlling for the average level of drinking. Binary logistic regression analyses were used to examine the association between binge drinking and the binary measure of multiple alcohol problems among moderate average-level drinkers. Following Goldwater and colleagues,
      • Goldwater D
      • Karlamangla A
      • Merkin SS
      • Seeman T.
      Compared to non-drinkers, individuals who drink alcohol have a more favorable multisystem physiologic risk score as measured by allostatic load.
      all analyses controlled for age, sex, years of education, household income, White/non-White, and marital status at baseline. In statistical analyses, age was scaled in 10-year units, and household income was scaled in 10,000-dollar units to facilitate interpretation. The analyses followed the key recommendations for prospective designs: (1) alcohol consumption at baseline predicted subsequent alcohol problems, and (2) alcohol problems at baseline were controlled, ensuring that the predictive effect was independent of reverse effects and stability in alcohol problems.
      • Khazanov GK
      • Ruscio AM.
      Is low positive emotionality a specific risk factor for depression? A meta-analysis of longitudinal studies.

      RESULTS

      Table 1 shows the sample characteristics for the total sample and for moderate and heavy average-level drinkers.
      Table 1Baseline Sample Characteristics for the Total Sample and for Moderate and Heavy Average-Level Drinkers
      CharacteristicsTotal sample

      (N=1,229)
      Average level of drinking
      Moderate

      (n=1,107)
      Heavy

      (n=122)
      Age55.52 (12.35)55.51 (12.37)55.52 (12.28)
      Sex
       Male640 (52.1)579 (52.3)61 (50)
       Female589 (47.9)528 (47.7)61 (50)
      Years of education14.95 (2.57)14.94 (2.58)15.12 (2.45)
      Household income (median)65,50066,50060,875
       White1,134 (92.3)1,020 (92.1)114 (93.4)
       Non-White95 (7.7)87 (7.9)8 (6.6)
      Marital status
       Not married356 (29.0)316 (28.5)40 (32.8)
       Married873 (71.0)791 (71.5)82 (67.2)
      Drinking pattern
       Regular972 (79.1)925 (83.6)47 (38.5)
       Binge
      Moderate average-level drinkers accounted for most total cases of binge drinking at baseline and most total cases of multiple alcohol problems at baseline and 9 years.
      257 (20.9)182 (16.4)75 (61.5)
      Baseline multiple alcohol problems
       No1,114 (90.6)1,030 (93.0)84 (68.9)
       Yes
      Moderate average-level drinkers accounted for most total cases of binge drinking at baseline and most total cases of multiple alcohol problems at baseline and 9 years.
      115 (9.4)77 (7.0)38 (31.1)
      9-year multiple alcohol problems
       No666 (86.8)606 (88.3)60 (74.1)
       Yes
      Moderate average-level drinkers accounted for most total cases of binge drinking at baseline and most total cases of multiple alcohol problems at baseline and 9 years.
      101 (13.2)80 (11.7)21 (25.9)
      Note: Multiple alcohol problems are shown at baseline and 9 years; for multiple alcohol problems at 9 years, n=767. Means (with SDs in parentheses) are shown for continuous variables (age and years of education), except for household income, for which the median is reported. The number of participants (with the percentage in parentheses) is shown for categorical variables.
      a Moderate average-level drinkers accounted for most total cases of binge drinking at baseline and most total cases of multiple alcohol problems at baseline and 9 years.
      A negative binomial regression analysis was used to examine the association of average level of drinking and a binge pattern of drinking, controlling for one another, with the number of alcohol problems at baseline (Table 2). The analysis controlled for all sociodemographic covariates. A high compared with a moderate average level of drinking was independently and significantly associated with a 2.16 times increase in the number of alcohol problems. A binge compared with a regular pattern of drinking was independently and significantly associated with a 2.76 times increase in the number of alcohol problems.
      Table 2Negative Binomial Regression Analyses With the Level and Pattern of Drinking Predicting the Number of Alcohol Problems
      Predictors at baselineNumber of alcohol problems at baseline

      (n=1,229),
      Number of alcohol problems at 9 years

      (n=767),
      OR (95% CI)OR (95% CI)
      Age
      In the analysis, age is scaled in 10-year units to facilitate interpretation.
      0.75 (0.70, 0.82)0.81 (0.73, 0.90)
      Sex (male=0, female=1)0.70 (0.56, 0.86)0.76 (0.58, 0.99)
      Years of education1.03 (0.99, 1.07)1.03 (0.98, 1.08)
      Household income
      In the analysis, income is scaled in 10,000-dollar units to facilitate interpretation.
      1.01 (0.99, 1.02)1.01 (0.99,1.02)
      Non-White (White=0, non-White=1)1.25 (0.87, 1.68)2.46 (0.64, 1.65)
      Married (no=0, yes=1)0.91 (0.73, 1.13)1.08 (0.83, 1.45)
      Baseline number of alcohol problems1.74 (1.53, 2.02)
      Average level of drinking (moderate=0, high=1)2.16 (1.71, 2.71)1.06 (0.71, 1.46)
      Pattern of drinking (regular=0, binge=1)2.76 (2.19, 3.41)1.40 (1.05, 1.83)
      Note: OR in boldface indicates statistical significance (p<0.05).
      Analyses were unadjusted for covariates. At baseline, the OR for average drinking=1.94 (95% CI=1.55, 2.41), and the OR for binge drinking=3.70 (95% CI=3.06, 4.42). At 9 years, the OR for average drinking=0.90 (95% CI=0.61, 1.22), and the OR for binge drinking=1.62 (95% CI=1.23, 2.07).
      a In the analysis, age is scaled in 10-year units to facilitate interpretation.
      b In the analysis, income is scaled in 10,000-dollar units to facilitate interpretation.
      A negative binomial regression analysis was also used to examine the predictive association of average level of drinking and a binge pattern of drinking at baseline, controlling for one another, with the number of alcohol problems at the 9-year follow-up (Table 2). The analysis controlled for all sociodemographic covariates as well as for the number of alcohol problems at baseline. Prospectively, a binge compared with a regular pattern of drinking was independently and significantly associated with a 40% increase in the number of alcohol problems 9 years later. Average level of drinking was not prospectively associated with the number of alcohol problems 9 years later.
      Although negative binomial regression is commonly used in predicting alcohol problems, binomial logistic regression provides an alternative procedure appropriate for count data with an upper bound.
      • Britt CL
      • Rocque M
      • Zimmerman GM.
      The analysis of bounded count data in criminology.
      To further validate these findings, binomial logistic regression was used to reexamine the models predicting the number of alcohol problems in Table 2. Results were similar to those reported in Table 2. At baseline, high average drinking (OR=2.58; 95% CI=1.90, 3.50) and binge drinking (OR=3.12; 95% CI=2.47, 4.05) significantly predicted the number of alcohol problems. Prospectively, only binge drinking (OR=1.43; 95% CI=1.03, 1.98) significantly predicted the number of alcohol problems.
      Moderate drinkers accounted for 70.8% (182/257) of cases of binge drinking at baseline and 67.0% (77/115) and 79.2% (80/101) of cases of multiple alcohol problems at baseline and 9 years, respectively (Table 1). For men, moderate drinkers accounted for 73.3% (143/195) of cases of binge drinking at baseline and 70.1% (54/77) and 84.9% (62/73) of cases of multiple alcohol problems at baseline and 9 years, respectively. For women, moderate drinkers accounted for 62.9% (39/62) of cases of binge drinking at baseline and 60.5% (23/38) and 64.3% (18/28) of cases of multiple alcohol problems at baseline and 9 years, respectively.
      Among moderate average-level drinkers, a binary logistic regression analysis was conducted to examine the association of a binge pattern of drinking with multiple alcohol problems at baseline (Table 3). The analysis controlled for all sociodemographic covariates. At baseline, a binge compared with a regular pattern of drinking was independently and significantly associated with a 4.85 times increase in odds of multiple alcohol problems among moderate average-level drinkers.
      Table 3Logistic Regression Analyses Among Moderate Drinkers With Pattern of Drinking Predicting Multiple Alcohol Problems
      Multiple alcohol problems at baseline

      (n=1,107),
      Multiple alcohol problems at 9 years

      (n=686),
      Predictors at baselineOR (95% CI)OR (95% CI)
      Age
      In the analysis, age is scaled in 10-year units to facilitate interpretation.
      0.59 (0.46, 0.73)0.70 (0.53, 0.87)
      Sex (male=0, female=1)0.61 (0.34, 1.07)0.33 (0.16, 0.57)
      Years of education1.01 (0.90, 1.14)1.05 (0.95, 1.17)
      Household income
      In the analysis, income is scaled in 10,000-dollar units to facilitate interpretation.
      1.03 (0.99, 1.07)1.00 (0.96, 1.04)
      Non-White (White=0, non-White=1)1.33 (0.53, 2.57)0.63 (0.12, 1.77)
      Married (No=0, yes=1)0.75 (0.43, 1.45)0.75 (0.42, 1.45)
      Baseline multiple alcohol problems5.72 (2.81, 12.73)
      Pattern of drinking (regular=0, binge=1)4.85 (2.81, 8.86)2.11 (1.11, 4.01)
      Note: OR in boldface indicates statistical significance (p<0.05).
      Analyses were unadjusted for covariates: OR for binge drinking=7.18 (95% CI=4.48, 11.80) at baseline and 3.06 (95% CI=1.62, 5.63) at 9 years.
      a In the analysis, age is scaled in 10-year units to facilitate interpretation.
      b In the analysis, income is scaled in 10,000-dollar units to facilitate interpretation.
      Among moderate average-level drinkers, a binary logistic regression analysis was also conducted to examine the predictive association of a binge pattern of drinking at baseline with multiple alcohol problems at the 9-year follow-up (Table 3). The analysis controlled for all sociodemographic covariates as well as for multiple alcohol problems at baseline. Prospectively, a binge compared with a regular pattern of drinking was independently and significantly associated with a 2.11 times increase in the odds of multiple alcohol problems 9 years later.
      Potential moderation of the binge drinking effect by sex and (separately) age was examined in each of the models mentioned earlier. The interaction of sex with a binge pattern of drinking was not significant (p>0.05) in predicting either the number of alcohol problems or, among moderate average-level drinkers, multiple alcohol problems either at baseline or at the 9-year follow-up. Similarly, the interaction of age with a binge pattern of drinking was not significant (p>0.05) in predicting either the number of alcohol problems or multiple alcohol problems among moderate average-level drinkers either at baseline or the 9-year follow-up.
      Potential moderation of the binge drinking effect by average level of drinking was also examined in the models in Table 2. The interaction of average level of drinking with a binge pattern of drinking in predicting the number of alcohol problems was significant (p<0.05) in cross-sectional analyses but not in prospective analyses (p>0.05). Cross-sectional analyses by drinking-level subgroups indicated that binge drinking significantly (p<0.05) predicted more alcohol problems among both moderate and high average-level drinkers, with the effect stronger for moderate drinkers.
      Several baseline variables showed statistically significant (p<0.05), although small, correlations with 9-year attrition: age (r=0.07), years of education (r = −0.14), household income (r = −0.13), non-White (r=0.12), and number of alcohol problems (r= −0.09). In a binary logistic regression with these variables controlling for one another, only non-White showed a noteworthy effect size (OR=2.34; 95% CI=1.47, 3.96) in predicting increased 9-year attrition.
      In addition, the baseline models predicting alcohol problems described in Tables 2 and 3 were reexamined restricting the analyses to participants who provided complete data across both waves. Results were similar to those reported in Tables 2 and 3. High average drinking (OR=1.97; 95% CI=1.46, 2.52) and binge drinking (OR=2.60; 95% CI=2.04, 3.36) significantly predicted the number of alcohol problems; among moderate drinkers, binge drinking (OR=5.39; 95% CI=2.83, 11.87) significantly predicted multiple alcohol problems.

      DISCUSSION

      A binge compared with a regular pattern of drinking predicted the number of alcohol problems independent of the average level of drinking. Extending previous research,
      • Greenfield TK
      • Ye Y
      • Bond J
      • et al.
      Risks of alcohol use disorders related to drinking patterns in the U.S. general population.
      ,
      • Linden-Carmichael AN
      • Vasilenko SA
      • Lanza ST
      • Maggs JL.
      High-intensity drinking versus heavy episodic drinking: prevalence rates and relative odds of alcohol use disorder across adulthood.
      ,
      • Holahan CJ
      • Brennan PL
      • Schutte KK
      • Holahan CK
      • Hixon JG
      • Moos RH.
      Late-life drinking problems: the predictive roles of drinking level vs. drinking pattern.
      this association persisted prospectively across 9 years and generalized to both women and men throughout the adult life span. Independent of a high average level of drinking, a pattern of binge drinking was uniquely linked with an almost 3 times increase in the number of concurrent alcohol problems and a 40% increase in the number of alcohol problems prospectively 9 years later. These findings underscore that average consumption by itself does not sufficiently reflect alcohol risk.
      • Roerecke M
      • Greenfield TK
      • Kerr WC
      • Bondy S
      • Cohen J
      • Rehm J.
      Heavy drinking occasions in relation to ischaemic heart disease mortality–an 11–22 year follow-up of the 1984 and 1995 U.S. National Alcohol Surveys.
      Alcohol research and policy need to address patterns as well as average level of consumption.
      Extending previous research,
      • Naimi TS
      • Brewer RD
      • Mokdad A
      • Denny C
      • Serdula MK
      • Marks JS.
      Binge drinking among U.S. adults.
      ,
      • Sherk A
      • Thomas G
      • Churchill S
      • Stockwell T.
      Does drinking within low-risk guidelines prevent harm? Implications for high-income countries using the International Model of Alcohol Harms and Policies.
      moderate average-level drinkers accounted for most cases of binge drinking at baseline and of multiple alcohol problems at baseline and follow-up. Moreover, binge drinking was linked to multiple alcohol problems among moderate average-level drinkers. This association persisted prospectively across 9 years and generalized to both women and men throughout the adult life span. Among moderate average-level drinkers, a binge pattern of drinking was associated with a close to 5 times increase in concurrent multiple alcohol problems and a >2 times increase in multiple alcohol problems prospectively 9 years later.
      The present results that moderate average-level drinkers accounted for many cases of binge drinking and multiple alcohol problems have public health relevance.
      • Rossow I
      • Mäkelä P.
      Public health thinking around alcohol-related harm: why does per capita consumption matter?.
      These findings are consistent with the prevention paradox, which emphasizes the need for alcohol interventions targeting moderate average drinkers in addition to conventional strategies focusing on the higher-risk but smaller population of habitually high-level drinkers.
      • Rossow I
      • Romelsjö A.
      The extent of the ‘prevention paradox’ in alcohol problems as a function of population drinking patterns.
      In a cross-sectional design, O'Dwyer et al.
      • O'Dwyer C
      • Mongan D
      • Millar SR
      • et al.
      Drinking patterns and the distribution of alcohol-related harms in Ireland: evidence for the prevention paradox.
      addressed this issue among Irish adults with alcohol problems who were not alcohol dependent. Consistent with the present findings, the authors emphasized the role of binge drinking in alcohol problems among nonalcohol-dependent drinkers.
      These results support including moderate average drinking adults in public health efforts to reduce binge drinking. Alcohol control policies focused on the price and availability of alcohol are especially effective in reducing binge drinking at a population level.
      • Chisholm D
      • Moro D
      • Bertram M
      • et al.
      Are the “best buys” for alcohol control still valid? An update on the comparative cost-effectiveness of alcohol control strategies at the global level.
      ,
      • Xuan Z
      • Blanchette J
      • Nelson TF
      • Heeren T
      • Oussayef N
      • Naimi TS.
      The alcohol policy environment and policy subgroups as predictors of binge drinking measures among U.S. adults.
      A significant contribution to such efforts can also come from primary care providers, who interact with a broad spectrum of adults who consume alcohol.
      • Schuckit MA.
      Screening and brief behavioral counseling interventions to reduce unhealthy alcohol use in adults 18 years and older, including pregnant women.
      At present, binge drinking among moderate drinkers is largely undetected in primary care settings.
      • Sahker E
      • Arndt S.
      Alcohol use screening and intervention by American primary care providers.
      ,
      • Tan CH
      • Hungerford DW
      • Denny CH
      • McKnight-Eily LR.
      Screening for alcohol misuse: practices among U.S. primary care providers, DocStyles 2016.

       Limitations

      The present findings are correlational and do not provide evidence of causality. In addition, the data were self-report and are subject to recall bias, common method variance, and social desirability. In addition, because the MIDUS sample underrepresented non-White participants and because non-White status was linked to increased 9-year attrition, caution is needed in generalizing these findings to minority populations. Furthermore, the MIDUS surveys indexed binge drinking as 5 or more drinks on an occasion for both women and men. Because binge drinking in women is now more commonly indexed as 4 or more drinks on an occasion, these findings may have underreported binge drinking in women.

      CONCLUSIONS

      In a U.S. national sample of adults, moderate drinkers accounted for most cases of both binge drinking and multiple alcohol problems. Among moderate drinkers, binge drinking was predictively linked to multiple alcohol problems. These results substantially broaden an increasing recognition that binge drinking is a public health concern among adults.
      • Kanny D
      • Naimi TS
      • Liu Y
      • Lu H
      • Brewer RD.
      Annual total binge drinks consumed by U.S. adults, 2015.
      Moderate average level drinkers should be included in efforts to reduce binge drinking and alcohol problems in adults. These findings are applicable to primary and secondary prevention of alcohol problems with the potential to advance population health.

      CRediT AUTHOR STATEMENT

      Charles J. Holahan: Conceptualization, Formal analysis, Methodology, Project administration, Writing - original draft, Writing - review and editing. Carole K. Holahan: Conceptualization, Methodology, Writing - review and editing. Rudolf H. Moos: Conceptualization, Methodology, Writing - review and editing.

      ACKNOWLEDGMENTS

      No financial disclosures were reported by the authors of this paper.

      REFERENCES

        • Kanny D
        • Naimi TS
        • Liu Y
        • Lu H
        • Brewer RD.
        Annual total binge drinks consumed by U.S. adults, 2015.
        Am J Prev Med. 2018; 54: 486-496https://doi.org/10.1016/j.amepre.2017.12.021
        • Azagba S
        • Shan L
        • Latham K
        • Manzione L.
        Trends in binge and heavy drinking among adults in the United States, 2011-2017.
        Subst Use Misuse. 2020; 55: 990-997https://doi.org/10.1080/10826084.2020.1717538
        • Grucza RA
        • Sher KJ
        • Kerr WC
        • et al.
        Trends in adult alcohol use and binge drinking in the early 21st-century United States: a meta-analysis of 6 national survey series.
        Alcohol Clin Exp Res. 2018; 42: 1939-1950https://doi.org/10.1111/acer.13859
        • Daly M
        • Robinson E.
        High-risk drinking in midlife before versus during the COVID-19 crisis: longitudinal evidence from the United Kingdom.
        Am J Prev Med. 2021; 60: 294-297https://doi.org/10.1016/j.amepre.2020.09.004
        • Greenfield TK
        • Ye Y
        • Bond J
        • et al.
        Risks of alcohol use disorders related to drinking patterns in the U.S. general population.
        J Stud Alcohol Drugs. 2014; 75: 319-327.https://doi.org/10.15288/jsad.2014.75.319
        • Linden-Carmichael AN
        • Vasilenko SA
        • Lanza ST
        • Maggs JL.
        High-intensity drinking versus heavy episodic drinking: prevalence rates and relative odds of alcohol use disorder across adulthood.
        Alcohol Clin Exp Res. 2017; 41: 1754-1759https://doi.org/10.1111/acer.13475
        • Naimi TS
        • Xuan Z
        • Brown DW
        • Saitz R.
        Confounding and studies of ‘moderate’ alcohol consumption: the case of drinking frequency and implications for low-risk drinking guidelines.
        Addiction. 2013; 108: 1534-1543https://doi.org/10.1111/j.1360-0443.2012.04074.x
        • Rehm J
        • Gmel Sr, GE
        • Gmel G
        • et al.
        The relationship between different dimensions of alcohol use and the burden of disease-an update.
        Addiction. 2017; 112: 968-1001https://doi.org/10.1111/add.13757
        • Holahan CJ
        • Brennan PL
        • Schutte KK
        • Holahan CK
        • Hixon JG
        • Moos RH.
        Late-life drinking problems: the predictive roles of drinking level vs. drinking pattern.
        J Stud Alcohol Drugs. 2017; 78: 435-441https://doi.org/10.15288/jsad.2017.78.435
        • Oppenheimer GM
        • Bayer R.
        Is moderate drinking protective against heart disease? The science, politics and history of a public health conundrum.
        Milbank Q. 2020; 98: 39-56https://doi.org/10.1111/1468-0009.12437
        • Naimi TS
        • Brewer RD
        • Mokdad A
        • Denny C
        • Serdula MK
        • Marks JS.
        Binge drinking among U.S. adults.
        JAMA. 2003; 289: 70-75https://doi.org/10.1001/jama.289.1.70
        • Sherk A
        • Thomas G
        • Churchill S
        • Stockwell T.
        Does drinking within low-risk guidelines prevent harm? Implications for high-income countries using the International Model of Alcohol Harms and Policies.
        J Stud Alcohol Drugs. 2020; 81: 352-361https://doi.org/10.15288/jsad.2020.81.352
      1. Ryff CD, Almeida DM, Ayanian JS, et al. Midlife in the United States (MIDUS 2), 2004-2006 (ICPSR 4652). Ann Arbor, MI: Inter-University Consortium for Political and Social Research. https://doi.org/10.3886/ICPSR04652.v8. Published September 15, 2021.

        • Radler BT
        • Ryff CD.
        Who participates? Accounting for longitudinal retention in the MIDUS national study of health and well-being.
        J Aging Health. 2010; 22: 307-331https://doi.org/10.1177/0898264309358617
      2. Ryff C, Almeida D, Ayanian J, et al. Midlife in the United States (MIDUS 3), 2013-2014. Ann Arbor, MI: Inter-University Consortium for Political and Social Research. https://doi.org/10.3886/ICPSR36346.v7. Published April 30, 2019.

        • Goldwater D
        • Karlamangla A
        • Merkin SS
        • Seeman T.
        Compared to non-drinkers, individuals who drink alcohol have a more favorable multisystem physiologic risk score as measured by allostatic load.
        PLoS One. 2019; 14e0223168https://doi.org/10.1371/journal.pone.0223168
        • Richardson GB
        • Boutwell BB.
        Decomposition of mean sex differences in alcohol use within a genetic factor model.
        Behav Genet. 2020; 50: 320-331https://doi.org/10.1007/s10519-020-10004-0
      3. Alcohol Research: Current Reviews Editorial Staff. Drinking patterns and their definitions.
        Alcohol Res. 2018; 39: 17-18
        • Chiva-Blanch G
        • Badimon L.
        Benefits and risks of moderate alcohol consumption on cardiovascular disease: current findings and controversies.
        Nutrients. 2019; 12: 108https://doi.org/10.3390/nu12010108
        • Zhang R
        • Shen L
        • Miles T
        • et al.
        Association of low to moderate alcohol drinking with cognitive functions from middle to older age among U.S. adults.
        JAMA Netw Open. 2020; 3e207922https://doi.org/10.1001/jamanetworkopen.2020.7922
        • Kessler RC
        • Andrews G
        • Mroczek D
        • Ustun B
        • Wittchen H.
        The World Health Organization composite international diagnostic interview short-form (CIDI-SF).
        Int J Methods Psychiatr Res. 2006; 7: 171-185https://doi.org/10.1002/mpr.47
        • Magidson JF
        • Robustelli BL
        • Seitz-Brown CJ
        • Whisman MA.
        Activity enjoyment, not frequency, is associated with alcohol-related problems and heavy episodic drinking.
        Psychol Addict Behav. 2017; 31: 73-78https://doi.org/10.1037/adb0000220
        • Glei DA
        • Weinstein M.
        Drug and alcohol abuse: the role of economic insecurity.
        Am J Health Behav. 2019; 43: 838-853https://doi.org/10.5993/AJHB.43.4.16
        • Kranzler HR
        • Soyka M.
        Diagnosis and pharmacotherapy of alcohol use disorder: a review.
        JAMA. 2018; 320: 815-824https://doi.org/10.1001/jama.2018.11406
        • Muthén LK
        • Muthén BO.
        Mplus Statistical Analysis With Latent Variables: User’s Guide.
        8th ed. Muthén & Muthén, Los Angeles, CA2017
        • De Bruyn S
        • Wouters E
        • Ponnet K
        • et al.
        Problem drinking among Flemish students: beverage type, early drinking onset and negative personal & social consequences.
        BMC Public Health. 2018; 18: 234https://doi.org/10.1186/s12889-018-5120-7
        • Anderson Goodell EM
        • Johnson RM
        • Latkin CA
        • Homish DL
        • Homish GG
        Risk and protective effects of social networks on alcohol use problems among Army Reserve and National Guard soldiers.
        Addict Behav. 2020; 103106244https://doi.org/10.1016/j.addbeh.2019.106244
        • Martins JS
        • Joyner KJ
        • McCarthy DM
        • Morris DH
        • Patrick CJ
        • Bartholow BD.
        Differential brain responses to alcohol-related and natural rewards are associated with alcohol use and problems: evidence for reward dysregulation.
        Addict Biol. 2022; 27: e13118https://doi.org/10.1111/adb.13118
        • Khazanov GK
        • Ruscio AM.
        Is low positive emotionality a specific risk factor for depression? A meta-analysis of longitudinal studies.
        Psychol Bull. 2016; 142: 991-1015https://doi.org/10.1037/bul0000059
        • Britt CL
        • Rocque M
        • Zimmerman GM.
        The analysis of bounded count data in criminology.
        J Quant Criminol. 2018; 34: 591-607https://doi.org/10.1007/s10940-017-9346-9
        • Roerecke M
        • Greenfield TK
        • Kerr WC
        • Bondy S
        • Cohen J
        • Rehm J.
        Heavy drinking occasions in relation to ischaemic heart disease mortality–an 11–22 year follow-up of the 1984 and 1995 U.S. National Alcohol Surveys.
        Int J Epidemiol. 2011; 40: 1401-1410https://doi.org/10.1093/ije/dyr129
        • Rossow I
        • Mäkelä P.
        Public health thinking around alcohol-related harm: why does per capita consumption matter?.
        J Stud Alcohol Drugs. 2021; 82: 9-17https://doi.org/10.15288/jsad.2021.82.9
        • Rossow I
        • Romelsjö A.
        The extent of the ‘prevention paradox’ in alcohol problems as a function of population drinking patterns.
        Addiction. 2006; 101: 84-90https://doi.org/10.1111/j.1360-0443.2005.01294.x
        • O'Dwyer C
        • Mongan D
        • Millar SR
        • et al.
        Drinking patterns and the distribution of alcohol-related harms in Ireland: evidence for the prevention paradox.
        BMC Public Health. 2019; 19: 1323https://doi.org/10.1186/s12889-019-7666-4
        • Chisholm D
        • Moro D
        • Bertram M
        • et al.
        Are the “best buys” for alcohol control still valid? An update on the comparative cost-effectiveness of alcohol control strategies at the global level.
        J Stud Alcohol Drugs. 2018; 79: 514-522https://doi.org/10.15288/jsad.2018.79.514
        • Xuan Z
        • Blanchette J
        • Nelson TF
        • Heeren T
        • Oussayef N
        • Naimi TS.
        The alcohol policy environment and policy subgroups as predictors of binge drinking measures among U.S. adults.
        Am J Public Health. 2015; 105: 816-822https://doi.org/10.2105/AJPH.2014.302112
        • Schuckit MA.
        Screening and brief behavioral counseling interventions to reduce unhealthy alcohol use in adults 18 years and older, including pregnant women.
        JAMA Psychiatry. 2019; 76: 5-6https://doi.org/10.1001/jamapsychiatry.2018.3278
        • Sahker E
        • Arndt S.
        Alcohol use screening and intervention by American primary care providers.
        Int J Drug Policy. 2017; 41: 29-33https://doi.org/10.1016/j.drugpo.2016.11.013
        • Tan CH
        • Hungerford DW
        • Denny CH
        • McKnight-Eily LR.
        Screening for alcohol misuse: practices among U.S. primary care providers, DocStyles 2016.
        Am J Prev Med. 2018; 54: 173-180https://doi.org/10.1016/j.amepre.2017.11.008