|Home | About | Journals | Submit | Contact Us | Français|
Excessive alcohol consumption is most widespread among young adults. Practice guidelines recommend screening and physician advice, which could help address this common cause of injury and premature death.
To assess the proportion of persons ages 18–39 who, in the past year, saw a physician and were asked about their drinking and advised what drinking levels pose health risk, and whether this differed by age or whether respondents exceeded low-risk drinking guidelines [daily (>4 drinks for men/>3 for women) or weekly (>14 for men/>7 for women)].
Survey of young adults selected from a national internet panel established using random digit dial telephone techniques.
Adults age 18–39 who ever drank alcohol, n=3,409 from the internet panel and n=612 non-panel telephone respondents.
Respondents were asked whether they saw a doctor in the past year; those who did see a doctor were asked whether a doctor asked about their drinking, advised about safe drinking levels, or counseled to reduce drinking.
Of respondents, 67% saw a physician in the past year, but only 14% of those exceeding guidelines were asked and advised about risky drinking patterns. Persons 18–25 were the most likely to exceed guidelines (68% vs. 56%, p<0.001) but were least often asked about drinking (34% vs. 54%, p<0.001).
Despite practice guidelines, few young adults are asked and advised by physicians about excessive alcohol consumption. Physicians should routinely ask all adults about their drinking and offer advice about levels that pose health risk, particularly to young adults.
Evidence supporting the effectiveness of screening and brief intervention for unhealthy alcohol use among adults is strong and the need for screening compelling. Unhealthy use is a spectrum of consumption, including abuse and dependence, that risks health,1 and its prevalence is highest among adults age 25 and younger. Proactive guidelines recommend screening and physician advice, which could help address this common cause of injury and premature death.2–7 Recent reviews of clinical trials found alcohol brief interventions to be effective in adults in the United States,8,9 among college students,10–13 and cross-nationally.14
In the United States, unhealthy alcohol use is the third leading preventable cause of death,15 causing over 71,000 deaths annually and shortening the lives of those who die by approximately 30 years on average.16 Unhealthy alcohol use is the fifth leading cause of disability for men in the United States and eleventh for women.17 Binge drinking, defined by NIAAA18 as a male consuming five or more and a female four or more drinks on an empty stomach over a two-hour period, is responsible for more than half of the deaths associated with unhealthy alcohol use.16 Binge drinking is associated with alcohol poisoning, unintentional injuries, suicide, hypertension, pancreatitis, sexually transmitted diseases, meningitis, interpersonal violence, drunk driving, and lost economic productivity.19 Further, cardiovascular benefits attributed to moderate alcohol use disappear when binge drinking occurs at least monthly.20
Worldwide, unhealthy alcohol use is the eighth leading cause of death, accounting for 2.5 million (3.8%) deaths, including 320,000 people between ages 15 and 29.21 Unhealthy alcohol use is the third leading cause of disability and accounts for 4.6% of disability adjusted life years (DALYs).21 The burden is particularly high in the European region where unhealthy alcohol use accounts for 11% of deaths and 12% of DALYs.22
In the United States, the age groups with the highest percentages who meet Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) alcohol dependence criteria are ages 18–20 (12%) and 21–24 (11%).23 Analyses of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) indicate that, of persons ages 18 and older (mean age 44) who ever in their life met alcohol dependence criteria, two-thirds did so before age 25.24 Of those with current dependence in the year preceding the survey, only 16% received alcohol treatment during that year.25
A national survey of 853 physicians practicing general internal medicine, family medicine, obstetrics-gynecology, and psychiatry revealed that 88% usually or always ask new patients about alcohol use, and 82% offer some sort of intervention to problem drinkers.26 However, in contrast less than one-third of 7,371 adults in a 1998 national survey who visited a general medical provider in the past year reported being asked about their alcohol or drug use.27 Recent young adult survey data have not been reported, and NESARC did not ask respondents about physician inquiry or advice about alcohol.
This study explores what proportions of young adults nationwide in the United States saw a physician in the past year and, of those, were asked by their physician about their drinking practices and advised about drinking levels that pose health risk or counseled to reduce their drinking. It also examined whether those proportions differed by age and whether respondents exceeded low-risk drinking guidelines.
In 2006, we conducted a U.S. national survey of current or former drinkers aged 18 to 39 focusing on drinking practices and alcohol-related problems. This study was approved by the Boston Medical Center Institutional Review Board.
Knowledge Networks used list-assisted random digit dialing (RDD) telephone techniques to identify a national panel sample. For our study, we randomly selected a sample of panel households, recruited in the past 2 years, with one adult age 18–39 randomly selected per household. Screening e-mails were sent to 6,200 panel members inviting them to participate in a study of alcohol-related behavior and experiences; those not responding to the invitation within a month were offered a short version of the survey and 189 responded. We defined current or former drinkers as individuals who ever consumed at least 12 drinks in a 12-month period, the criterion used to define drinkers in the 1992 National Longitudinal Alcohol Epidemiologic Survey. Of the 5,788 respondents to the screening e-mail (93% response rate), 4,012 were identified as current or former drinkers. Six-hundred and three eligible respondents were randomly selected for a separate sub-study and not included in the analysis. The remaining 3,409 respondents screening positive continued on to take the survey.
Surveying a pre-recruited panel raises the concern that persons who agreed to serve on a survey panel may not be representative of the general population. To minimize selection bias and improve the generalizabilty of the study sample, the same survey was administered by phone to a stratified, random sample of n=612 persons from the RDD sample underlying the panel (the telephone survey included three groups not in the panel), those who had not been reached for panel recruitment (28% response rate), those who refused to join the panel (63% response rate), or those who had dropped off the panel (58% response rate). The combined sample (n=4,021) was weighted to reflect sampling probabilities and the national population by age, sex and race/ethnicity. By American Association for Public Opinion Research (AAPOR) definitions, the co-operation rate for the Internet survey sample was 93% (e.g., 93% of panel respondents sent an e-mail screening invitation responded, and all those eligible continued on to the survey). The overall population response rate for the combined sample, accounting for the acceptance rate into panel, response rates for the telephone survey and the sample weighting was 37%.
Both the online and telephone surveys used the same instrument. Questions measured demographic characteristics (age, sex, race/ethnicity), education, and drinking practices.
Questions taken from NESARC asked respondents their frequency, usual and maximum amount of alcohol consumption during the past year, and how often men consumed 5 or more drinks or women consumed 4 or more in a single day. NIAAA has established low-risk drinking limits as no more than four drinks in a day for men or three for women (not binge drinking) and no more than fourteen drinks per week for men and seven for women. Healthy adults who do not exceed these limits are at lower risk of alcohol-related injury or illness. These limits do not apply to pregnant women, persons using medications, driving motor vehicles, operating machinery, under age 21, or with medical conditions that contraindicate drinking.28 Respondents were categorized as 1) current drinkers not exceeding the NIAAA guidelines, 2) drinkers exceeding either the daily or the weekly guidelines, 3) drinkers exceeding both the daily and weekly guidelines, and 4) former drinkers, those who did not drink more than 12 drinks in the past 12 months but did so in at least one previous year.
Respondents were also asked: “Did you see a doctor for any reason in the past year?” Those responding yes were asked additional separate questions: Did the doctor ask you about how much you drink, offer advice about what level of drinking is a risk factor for your health, and advise you to reduce your drinking?
Analyses used SUDAAN,29 a statistical software program that adjusts for survey design and weighting. Chi-square tests evaluated bivariate associations between study outcomes, age, and whether respondents exceeded weekly or daily limits.
Logistic regression analyses explored whether respondent age (categorized 18–20, 21–25, and 26–39) and whether drinking in excess of daily or weekly low-risk drinking limits or both were associated with whether respondents saw a physician in the past year, or whether a physician asked about their drinking, advised about drinking levels that pose health risk, and counseled to reduce drinking. Analyses controlled for gender, race/ethnicity, employment status, education, and family history of alcoholism. An additional regression controlled for whether respondents were Internet panel members or from the phone supplement sample. Analyses of whether respondents saw a physician in the past year included all respondents. Analyses exploring whether respondents were asked or counseled by a doctor about their drinking focused on those seeing a physician in the past year. Odds ratios and 95% confidence intervals (CIs) were calculated. Of the 4,021 respondents, 189 were slow responders given the short version of the survey which did not include data on doctor visits. These respondents were excluded from these analyses, leaving a final sample of 3,799.
Table 1 describes respondent characteristics. In the past year, 16% did not drink alcohol, 24% drank at or below daily or weekly limits, 47% exceeded either daily or weekly limits (almost all of them daily), and 13% exceeded both. Those ages 18–20 and 21–25 were more likely to exceed weekly or daily drinking limits (Table 2). Those given the short vs. full version of the survey were somewhat younger (31% vs. 22% below 26 years) but did not differ on sex, race, education level, or past year drinking pattern.
Of respondents, 67% saw a physician in the past year but only 14% of those exceeding daily or weekly guidelines were asked and advised about risky drinking patterns (Table 3). Only 26% of those ages 18–20 and 42% ages 21–25 who saw a doctor recalled being asked about their drinking, compared to 54% ages 26–39 (p<0.001) (Table 3). Of respondents who saw a physician and exceeded either daily or weekly guidelines, 49% recalled being asked about their drinking, but only 14% were advised about low-risk drinking guidelines and 7% were advised to cut down. Neither measure differed significantly by age. Among those who exceeded both limits, only 24% were advised about low-risk guidelines and 21% to reduce drinking. Even though respondents ages 18–25 were most likely to exceed drinking limits (68% vs. 56%, p<0.001), they were least often asked about drinking (34% vs. 54%, p<0.001).
Regression analyses (Table 4) indicated seeing a doctor in the past year did not significantly vary by age or exceeding drinking guidelines. Females, those with a family history of alcoholism, some college education, and persons who were retired, disabled, or a homemaker more often saw a doctor.
Among respondents who saw a doctor in the past year, those age 18–20 (adjusted odds ratio [AOR], 0.3; 95% confidence interval [CI], 0.2-0.6) and 21–25 (AOR, 0.6; 95% CI, 0.4-0.8) were significantly less likely than those 26–39 to have recalled being asked about their drinking, controlling for drinking pattern. Drinking pattern did not predict whether someone recalled being asked about drinking. Those with a family history of alcoholism more often recalled being asked, and those with less than a high school education less often recalled being asked about their drinking.
Compared with drinkers who saw physicians and did not exceed the daily or weekly low-risk guidelines, those exceeding both were more often advised about drinking levels that pose health risk (AOR, 2.4; 95% CI: 1.4-3.9) and to reduce their drinking (AOR, 7.0; 95% CI, 3.0-16.2). Males and Hispanics were more frequently advised about drinking levels. Males and persons with family history of alcoholism were more often encouraged to cut down.
Those who exceeded the daily or weekly guidelines were not significantly more likely to be advised about drinking limits or to cut down than those not exceeding either guideline. Further, being advised about drinking guidelines or cutting down did not differ significantly by age. Respondents under age 21 were half as likely as those ages 26 and older to receive advice to reduce their drinking independent of their drinking pattern, but this was not statistically significant. Results did not change when we analyzed whether respondents were part of the Internet panel or in the phone sample. Phone sample respondents more often recalled being asked about their drinking (AOR 1.9; 95% CI 1.4 -2.6) and given advice about drinking guidelines (AOR 2.0; 95% CI 1.4-3.1).
This national survey indicated two-thirds of adults ages 18–39 saw a physician in the past year. Only 14% of those young adults who saw a physician and exceeded daily or weekly limits remembered being asked about their drinking and advised of risky drinking levels. Although respondents ages 18–25 were more likely than those 26–39 to exceed low-risk drinking daily and weekly guidelines, they were less likely if they saw a doctor to recall being asked how much alcohol they drank.
Per clinical practice recommendations, alcohol screening should be universal. Most patients with alcohol disorders are not identified and unhealthy alcohol use is less likely to be detected when screening is not done universally.
Given the health burden posed by alcohol among young adults, it is particularly important to explore why only one in seven persons ages 18–39 in this national sample who exceeded drinking guidelines and saw a physician recalled being asked and advised of drinking levels that pose risk to health and why this less often occurred among people ages 18–25, those most likely to exceed low-risk drinking limits.
Several explanations can be offered. First, this finding is derived from survey self-report. Some respondents seen by a physician in the past year may have been asked and advised about potentially unhealthy drinking levels but failed to recall such interaction. Respondents’ answers were not corroborated by their physicians. However, even if some respondents were counseled, if they could not recall this, it is doubtful that their knowledge, attitudes, beliefs, behavior, and health were influenced.
Second, our survey’s 37% response rate is low, raising questions about internal validity and generalizability. However, the response rate was for participating in both the panel and this survey. Of our random sample of panel members, 93% were surveyed. A previous paper30 comparing this combined sample with NESARC (overall response rate 81%) indicated the two surveys’ respondents ages 18–39 had similar demographic characteristics, drinking patterns, proportions who exceeded NIAAA guidelines, and associations between lifetime alcohol dependence and current risky drinking and gender, education, family history of alcohol problems, and age of drinking onset. Also, an internet panel may be less likely to represent low SES and rural populations with limited internet availability. However, the Knowledge Networks panel RDD sample oversampled telephone numbers from phone banks with higher concentrations of Blacks and Hispanics. Subjects agreeing to participate in the panel who did not have internet access were given WebTV and Internet access for free.
The standing Internet panel from which our sample was drawn was initially recruited to participate in a variety of non-alcohol related surveys. Thus, nonresponse to enrollment in the panel is not related to this survey’s topic, alleviating a major concern regarding nonresponse. Further, to increase the study’s representativeness, we surveyed both Internet panel members and another sample of non-panel members drawn from the RDD sample that generated the panel. Nonetheless, we recommend cautious interpretation of our results.
Lastly, the survey did not specify primary care practices, the setting where alcohol screening is most convincingly recommended. Episodic care and specialist practices clearly have opportunities to screen, and screening is recommended in trauma centers. Although this may partly account for the very low proportions of patients screened, it seems unlikely to be a satisfactory explanation.
Numerous barriers discourage routine alcohol screening and brief interventions. First, screening and advising patients takes time. Although true, brief validated 1–3 item screening tools are now available.3,31,32 Second, many clinicians may feel these practices require specific training they have not received. While also true, easy to use training resources are widely available to meet this need.3 Third, alcohol treatment services may be lacking, particularly for young people and in rural areas. Fourth, reimbursement can be an issue. In 25 states, laws allow insurance companies to withhold treatment reimbursement for persons injured under the influence of alcohol.33 Though infrequently invoked, these laws may discourage screening, particularly in emergency departments and trauma centers where alcohol-related injuries are common. Recently, 12 states repealed these laws.33 Further, recently new CPT billing codes for screening and counseling patients about alcohol were established. If states adopt these codes in Medicaid and private insurers agree to reimburse, an important barrier may be overcome. Fifth, some physicians’ drinking behaviors may influence whether they offer advice.34
While limited clinician time, inadequate systems and heavily loaded clinical agendas shared by patients and physicians are barriers in such settings, these do not seem to be issues for preventive services like blood pressure assessment which is universally done during physician visits despite absence of any recommendation to do so frequently. Furthermore, although screening rates clearly need to be improved across the entire age range we studied, if the observed screening effort is optimal, then it should be redirected to younger adults, those with the higher prevalence of unhealthy alcohol use, and greatest risk of morbid and mortal consequences.
In summary, although most young adults in this national sample saw a physician in the past year, half or fewer were asked about their drinking despite evidence-based practice guidelines that recommend the practice. Furthermore, although the youngest adults were more likely to drink too much, they were least likely to have been asked how much they drank. Since the prevalence of unhealthy alcohol use is highest among youth, and since screening and brief intervention have efficacy, screening young adults should be routine. Screening is a recommended preventive service; in some settings (like the largest health system in the US—the VA) screening rates are close to 100%.35 While future studies are needed to confirm our findings and explore reasons why young adults are least likely to be asked about their drinking, efforts to implement universal screening and intervention (e.g. performance measures, training, and appropriate reimbursement) should continue, particularly with young adult patients who are most at-risk and least likely to be identified.
This article is dedicated to Helen Marie Witty, who at age 16 was fatally injured by a young alcohol-impaired driver. Dr. Ralph Hingson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. This study was supported in part by the National Institute on Alcohol Abuse and Alcoholism Center Grant P60AA13759.
Conflict of Interest Drs. Ralph Hingson, Tim Heeren, and Erika Edwards have no conflicts of interest to report. Dr. Richard Saitz reports having been a consultant for online alcohol-related screening and brief intervention education projects supported by National Institutes of Health (NIH) grants to Medical Directions and Inflexxion and for NIH grants to the RAND corporation, Kaiser Permanente, the University of Massachusetts, and Brandeis University. Dr. Saitz also has been compensated by Beth Israel Deaconess Hospital and the National Institute on Alcohol Abuse and Alcoholism for serving on data and safety monitoring boards. He is compensated for educational work by the Massachusetts Medical Society, the British Medical Journal Group, and the American Society of Addiction Medicine. He has developed educational materials for Fusion medical education and consulted for Saatchi and Saatchi Healthcare on alcohol dependence treatment. He has or anticipates being compensated as a speaker on alcohol and drug topics by multiple government agencies, academic institutions, and professional societies. He has also provided expert opinion on legal cases involving identification and management of alcohol and drug related problems.
Ralph W. Hingson, Phone: +1-301433-2860, Fax: +1-301443-8614, Email: rhingson/at/mail.nih.gov.
Timothy Heeren, Email: tch/at/bu.edu.
Erika M. Edwards, Email: eedwards/at/bu.edu.
Richard Saitz, Email: rsaitz/at/bu.edu.