alcohol problems among young adult emergency department patients: making predictions using routine...

Alcohol Problems Among Young Adult Emergency Department Patients: Making Predictions Using Routine

Sociodemographic Information Kimberly Horn, Luba Leontieva, Janet M. Williams, Paul M. Furbee, James C. Helmkamp,

and William G. Manley III

Objectives: The study aims were to explore the process of identifying young adult Emergency De- partment (ED) patients at risk for alcohol problems, examine the sociodemographic predictors of patient alcohol problems, and determine differences between patients who screened positive and those who screened negative for alcohol problems. Implications for ED practice are discussed. Method:As part of a larger study, the Alcohol Use Dis- orders Identification Test was administered to con- senting patients ages 18 to 29. A score of greater than 5 (out of 40) was considered a positive screen for alcohol problems. Sociodemographic information was collected from a generic Individual Information Form. Results: Approximately 48% (1,855 of 3,890} of patients screened positive for alcohol problems. Among screen-positive patients, 91% (1,689) had scores that corresponded to mild to moderate alcohol problems. Compared with screen-negative patients, screen-

positive patients were almost 3 times more likely to be men (odds ratio [OR] = 2.70); 2.5 times more likely to use tobacco (OR = 2.43); 2 times more likely to be single (OR = 2.03); and 1.5 times more likely to have some college education (OR = 1.41), be an 18- to 20- year-old adult (OR = 1.61), be a college student (OR = 1.60), be unemployed (OR = 1.39), and cohab- itate with friends (OR = 1.19}. Screen-positive patients also were more likely to have made at least one past attempt to change their drinking behavior (OR = 3.21). Conclusion: Routine patient information presented an accurate risk profile for alcohol problems among patients in this study. If ED-based health care professionals routinely screened patients for alcohol problems or could predict potential alcohol problems by using routinely collected information, then intervention or referral for additional services may increase. Copyright 2002, Elsevier Science (USA). Aft rights reserved.

A LCOHOL IS THE MOST commonly used drug in the United States, with approximately

47% (104 million) of people ages 12 years and older who currently drink alcohol.1 The 2000 Na- tional Household Survey on Drug Abuse reported that 18.7% (6.6 million) of people between the ages of 12 and 20 engaged in binge drinking and 6% (2.1 million) drank heavily (5 or more drinks on each of 5 or more days) in the" preceding 30 days. Health care costs of untreated alcoholic patients are at least 100% higher that those of nonalcoholic individuals. 2 In 1992 the health care cost for alcohol abuse treatment was estimated at $4.0 billion,

From the Department of Community Medicine, Center for Rural Emergency Medicine, Office of Drug Abuse Intervention Studies, West Virginia University, Morgantown, WE.

Presented at 2001 Research Day, West Virginia University, Morgantown, WV,, April 2001.

Supported by the Centers for Disease Control and Prevention (grant # R49/CCR308469-06).

Address reprint requests to Dr. Kimberly Horn, EdD, West Virginia University, Robert C. Byrd Health Sciences Center, De- partment of Community Medicine, Office of Drug Abuse Inter- vention Studies, PO Box 9190, Morgantown, WV 26506. E-mail: kho rn @ wvu. edu

Copyright 2002, Elsevier Science (USA). All rights reserved. 0883-9441/02/1704-0002535.00/0 doi: i O.l O53/jcrc.2002.37231

whereas prevention costs were almost 4 times less, at approximately $1.1 billion) Early detection, referral, and timely alcohol treatment are crucial to reduce individual and national medical costs, and improve health and treatment impact.

Identification of individuals at risk for alcohol problems is the first step to early detection and intervention. Emergency Departments (EDs) have the potential to be front-line sources of health care for people with alcohol problems .4,5 Many of these people, who are predominantly young or middle aged, present to the ED with medical conditions unrelated to alcohol use but report problems clearly associated with harmful drinking. However, screening for alcohol problems is not conducted routinely in EDs, despite research showing high patient prevalence of alcohol problems and alcohol-related medical conditions. 6-1~ For example, one ED-based study revealed that 43% of patients ages 18 to 29 screened positive for alcohol problems. 5 Similarly, other researchers found that 43% of ED patients were at risk for alcohol problems. 11 Another study described 21% of the subcritically injured ED patient population as alcohol positive, lO A British ED-based study found that 40% of all evening patients had been drinking alcohol] 2 Similar research from Thailand found that 61% of patients who were ad- mitted to the ED used alcohol before admission. 13

212 Journal of Critical Care, Vol 17, No 4 (December), 2002: pp 212-220

ED PATIENT ALCOHOL PROBLEMS 213

Although research highlights ED patient alcohol prevalence, existing literature on patient alcohol screening lacks detailed descriptions of ED populations appropriate for screening or intervention. In fact, only a few studies concentrate on target population details. 5'11'14-1s Among those studies, data

are conflicting. According to some studies, there

were more men than women who screened positive for risky drinking, subsequently undergoing counseling. 14'15'17'1s However, other studies showed that

women tended to drink more than men. 11,16 Some

studies suggested a relationship between depressive symptoms and drinking, 14,15 tobacco use and drinking, 5'14'15 and other drug use and drinking. 14 As

important, one study suggested that sociodemographic characteristics contributed to the correlations found between alcohol consumption and injury status. 8 Although the literature is limited, there

is evidence to support the importance of under-

standing the sociodemographic predictors of alcohol problems among ED patients, 19 and the need for ED-based alcohol screening. 5's'11'2~

Given the reported prevalence of ED patients who present with alcohol problems and whose problems likely go unnoticed, a response is warranted. There are many health risk conditions for which EDs can provide screening, including obe- sity, tobacco use, illicit drug use, domestic vio-

lence, hypertension, polypharmacy in the elderly,

and so forth. However, the concept of health pro- motion and prevention is relatively new in emergency medicine and not all EDs embrace the concept, especially because there is limited (if any) reimbursement for such services. 2t Moreover, al-

cohol screening methods that meet the time and resource constraints of EDs are not widely available, nor are they promoted as part of ED provider train- ing or education. 22'23 Even among EDs that have

begun alcohol screening, the current health care system often caters to end-stage/severe alcohol problems, in which symptomology is obvious to providers. These patients may be referred else-

where if services are available, but it is unlikely

that treatment is sought by the patient. Further, ED providers may have the perception that it is a waste

of money and time to screen for a disease that they can do nothing about during their period of care or treatment. 4,22

If ED-based health care professionals routinely screened patients for alcohol problems or could reliably predict potential alcohol problems based on

routinely collected information, then intervention or referral for intervention may increase. As part of a larger ED-based alcohol screening and intervention study, the purpose of this investigation was to (1) examine the extent to which individuals at risk

for alcohol problems could be identified; (2) explore predictors of alcohol problems by using routinely collected sociodemograhic information; and (3) determine differences between patients who

screened positive for alcohol problems and those who screened negative. The study hypotheses were that at-risk patients could be identified and that routine patient information could be used to reliably predict alcohol problems. Moreover, it was predicted that there would be significant sociodemographic differences between screen-positive and screen-negative patients.

METHOD

Par t i c i pan ts

Participants were patients aged 18 to 29 who presented for care at a university-affiliated ED from August, 1998 to Decem- ber, 2000. Because epidemiologic studies show that alcohol problems, whether abuse, dependence, or otherwise defined, and excessive drinking increases as age decreases (at least un- til the 20s), the study targeted a young adult populationY -25 A university-based ED provided an adequate setting to access the target population.

As part of a larger alcohol screening and intervention research project, patients were approached by trained project staff who had substance abuse or social work backgrounds. 5'al Patients consuming any amount of alcohol in the previous 12 months were eligible for study inclusion. Patients were excluded if they had life- or limb-threatening conditions; were mentally incom- petent or combative; arrived in police custody; did not speak English; had a severe communication deficit or were intoxicated (according to blood alcohol content validation). Eligibility was determined during face-to-face interviews by research staff in the ED waiting area or treatment rooms. Participants signed In- stitutional Review Board-approved consent forms.

P r o c e d u r e

For research pu~oses, 2 primary screening tools were used: (1) a generic Individual Information Form and (2) an Alcohol Use Disorder's Identification Test (AUDIT). Time was documented as part of the screening process (minutes from start to end). The Individual Information Form included information equivalent to that routinely collected in the ED (eg, age, race, marital status, educational status, tobacco use, and so forth).

The AUDIT, a 10-question inventory with a score range between 1 and 40, T M queried patients about amount of drinking, frequency, severity, alcohol-related injuries, and social consequences of drinking. The AUDIT assesses patients' experiences with alcohol dependence symptoms (questions 4-6), alcohol- related harm (questions 7-10), and alcohol intake, including amount and frequency (question 1-3). The intake domain allows

214

the AUDIT to identify patients who are at high risk for, but have not yet experienced, alcohol-related harm. As such, the AUDIT particularly is useful for identifying patients at early stages of problem drinking.

The Chronbach's c~ reliability and item-total correlations for AUDIT are typically in the 0.80s. 5'24'2s A detailed description

of the AUDIT, as used in this study, and its psychometric prop- erties are described elsewhere. 5'11 The AUDIT has been vali- dated for use in medical settings 26 and extensively used in a research context. It has shown high correlation with other well-accepted alcohol screening tools: the Michigan Alcoholism Screening Test; MacAndrew alcoholism screening test; hepatic enzyme levels (reflect recent heavy drinking); subjects' hypo- thetical vulnerability to alcoholism (eg, familial alcoholism and degree of sociopathy)27; and measures of somatic and affective consequences of drinking. 26 Research has shown the use of the AUDIT for identification of hazardous or harmful drinking in drug-dependent patients. 27

It is important to note that the AUDIT is not a diagnostic instrument. Although statistically it compares very favorably with other screening instruments and there is some research on which thresholds are optimal for different purposes and in different populations, there is a dearth of information related to the AU- DIT and different clinical manifestations of alcohol abuse or alcohol problems.

Based on previous studies, an AUDIT score of greater than 5 was considered a positive screen for self-reported mild to moderate alcohol problems 5'11.28-3o; a threshold of 6 (out of 40 points)

was selected to increase the screening sensitivity to include a wide range of alcohol problems and severity. ]t'24'25-a8'3~ Recom-

mended cut-off scores for the AUDIT range between 5 and 11. The World Health Organization, the developers of the AUDIT, originally recommended a cut-off score of greater than 11 (sensitivity 0.84; specificity 0.71). However, with a cut-off score of greater than 8, which was commonly used, sensitivity ranged between 0.61 to 0.96, specificity was between 0.84 to 0.96, 26'32-34 and with a cut-off score greater than 5 the sensitivity was 0.61 and specificity was 0.84Y Thus, decreasing the cut-off score increases sensitivity and specificity, and, in turn, enhances early identification of alcohol problems, especially among young adults. 15,32,36 Additional past research by the current investigators has shown the prognostic significance of a score of 6 pointsJ 1,37

Data Analysis The initial analysis included a calculation of alcohol use sta-

tus: number of patients who screened positive for alcohol problems (AUDIT > 5) versus those who screened negative (AU- DIT <-- 5). Next, 9 sociodemographic factors routinely available in ED patient intake charts (eg, ED registration databases) were selected as variables in the alcohol problem prediction model (see Table 1). Screen-negative patients constituted the comparison group. Initially, general comparisons were made between screen-positive and screen-negative patients on the factors shown in Table 1. X 2 analysis was used to further explore significant differences between the explanatory variables and alcohol use status. Next, correlation analysis was conducted on all factors to rule out the occurrence of multicolinearity. The data were examined to determine that no evidence of a prob- lematic level of multicolinearity existed among predictors and to confirm that all Pearson correlation coefficients among predictors were less than 0.50. 38

Table 1. Factors Explored

HORN ET AL

Dependent var iable

Alcohol status (screen posit ive - AUDIT > 5;

screen negative - AUDIT < 5)

Explanatory variables

Age

Sex

Employment status

Marital or romantic cohabitat ion status

Living arrangements (nonmarital)

Education level

Current student status

Tobacco use

Previous alcohol treatment/help

To determine the extent to which the study variables explained and predicted alcohol problems when the other variables were controlled, multiple logistic regression by using forward and backward step-wise methods was conducted. Continuous data were transformed into dichotomous data by using dummy or in- dicator coding to enhance the model reliability and facilitate meaningful clinical interpretation. Factors were scaled as dichotomous, by using 0 or 1 to define the absence or presence of a particular attribute, respectively. For instance, the item for education was recoded from a 5-option response (<high school, high school diploma, some college, college graduate, postgrad- uate) to 2 exclusive categories of "no college (high school degree or less)" and "some college (at least some college)." Fur- ther, tobacco use status included past 30-day use of either cigarettes or smokeless tobacco. Smoking and smokeless tobacco use were collapsed because almost 92% of patients did not use smokeless tobacco; a greater than 90%/10% split in any dichotomous category is not acceptable for regression analy- ses. 39 Regression model coefficients for explanatory variables were transformed to odds ratios (ORs). This approach provided a comparison of the unique contribution of each factor to the model. An alternative analytic approach could have been re- ceiver operating characteristic curve analysis. However, because the purpose was to explore a risk profile rather than the specificity or sensitivity of the sociodemographic factors, regression was selected. Finally, the mean screening time was calculated in minutes from real-time recordings documented by the ED providers. Screening start/end times were recorded on each patient file. SAS software (SAS Institute, Cary, NC) was used to analyze data. 4~

RESULTS

The study patient consent rate was 92% (3,963 of 4,308). A total of 3,890 patients were screened (73 patients withdrew, became ill, or were called away by the doctor or nurse); 48% (1,855) screened positive. The screening process used to confu'm alcohol problem status in this study lasted 15 minutes or less in 87% of patients, excluding consent for research participation. Among screen-positive patients, most (91%, n = 1,689) had AUDIT scores ranging between 6 and 20, which corresponded to

ED PATIENT ALCOHOL PROBLEMS

mild to moderate self-reported alcohol problems. Ap- proximately 9% (166) had an AUDIT score higher than 20, which corresponded to severe self-reported alcohol problems. Mean AUDIT scores for screen- positive (mean = 11.78, SD = 5.74) and screen- negative patients (mean = 2.90, SD = 1.40) were significantly different (P < .05). General sociodemographic comparisons between screen-positive and screen-negative patients are presented in Table 2. There were significant differer~ces on all factors.

Multiple logistic regression revealed that all of 9 potential predictors explored in this study were independently associated with alcohol use problems. Coefficients, ORs, and related statistics are shown in Table 3. The overall percent accuracy classification (ie, concordance) was 76.2%. The likelihood ratio also was significant (X 2 = 876.2968, P < .000l). Compared with screen- negative patients, screen-positive patients were (1) almost 3 times more likely to be men (OR = 2.70); (2) 2.5 times more likely to use tobacco (OR = 2.43); (3) 2 times more likely to be single (OR = 2.03); and (3) 1.5 times more likely to have some

215,

college education (OR = 1.41), be a young adult ages 18 to 22 (OR = 1.61), be a college student (OR = 1.60), be unemployed (OR = 1.39), and co- habitate with a roommate (nonromantic or marital) (OR = 1.19). In addition, screen-positive patients were over 3 times more likely than screen-negative patients to have made at least one past attempt to change their drinking behavior (OR = 3.21).

DISCUSSION

Importantly, one of the primary intents of this study was to determine red flags that may alert ED health care providers to patient needs for further screening or intervention for alcohol problems. This study revealed several important findings: (1) screening and identification of ED patients with alcohol problems were possible; (2) screen-positive and screen-negative patients were significantly different sociodemographically; and (3) routine sociodemographic information was predictive of alcohol problems among ED patients. All of these issues have implications for ED-based screening and intervention for alcohol problems.

Table 2. General Comparisons of Screen-Positive Patients and Screen Negative Patients

Screen-Positive Screen-Negative Variables % (n) % (n) P

Age 18-22 56% (1,086) 44% (870) 23-29 40% (769) 60% (1,165) <.0001

Sex Men 61% (1,085) 39% (690) Women 36% (768) 64% (1,345) <.0001

Employment status Employed 43% (956) 57% (1,265) Unemployed 54% (898) 46% (769) <.0001

Marital/cohabitation status Married/cohabitating 33% (460) 67% (928) Not married/cohabitating 56% (1,395) 44% (1,107) <.0001

Roommate status Live with roommate/family 50% (1,435) 50% (1,422) Do not live with roommate/family 41% (417) 59% (606) <.0001

Education Some college 50% (1,232) 50% (1,212) No college (high school or less) 43% (618) 57% (820) <.0001

Student status Student 60% (585) 40% (396) Non student 44% (1,270) 56% (1,639) <.0001

Tobacco use status Tobacco user 56% (1,178) 44% (944) Nonuser 38% (677) 62% (1,091) <.0001

Part alcohol treatment Past treatment/help received 69% (713) 31% (322) No past treatment/help 40% (1,442) 60% (1,713) <.0001

216 HORN ET AL

Table 3 Results From the Multiple Logistic Regression Analysis for Screen-Positive Patients

Risk Factors OR* 95% Cl ~ P

Education level 1.41 1.18-1.68 <.0001 (has some college)

Sex 2.70 2.34-3.13 <.0001 (is a man)

Tobacco use 2.43 2.08-2.84 <.0001 (uses tobacco)

Past help 3.21 2.71-3.79 <.0001 (has sought past help)

Marital/romantic cohabit status 2.03 1.71-2.40 <.0001 (is single)

Student status 1.60 1.33-1.93 <.0001 (is a college student)

Employment 1.39 1.20-1~61 <.0001 (is unemployed)

Age 1.61 1.38-1.88 <.0001 (is a young adult age 18-22)

Living arrangements 1.19 1.01-1.40 <.0434 (is living with friend or nonspouse family member)

*OR using coefficients obtained from logistic regression. t95% confidential Wald intervals.

Identification of ED Patients With Alcohol Problems

Results of this study indicate that almost half of the ED study patients screened positive for alcohol problems during the study period. Patients were screened between 12 vM and 1 A~, times of heav- iest patient flow. This finding is important for 2 reasons: (1) it reveals a high frequency of patients with alcohol problems presenting for care in the ED during the study period; and (2) it shows that it is possible to identify ED patients with alcohol problems through brief self-reported information. The screening process used to confirm alcohol problem status in this study lasted less than 15 minutes in most cases (excluding research consent). In addition, there was a consent rate of almost 92%, con- firming that when approached and queried, patients were willing to report on their alcohol use patterns and behaviors to ED staff. Together, these findings suggest that if we ask, even briefly, they will tell.

Most study patients were experiencing mild to moderate alcohol problems (eg, increased alcohol consumption, not doing something that was ex- pected from them because of drinking, unable to remember what they have done during some of their drinking episodes, have injuries to themselves or someone else because of drinking, and have some-

body telling them that they should cut down on their drinking). Per AUDIT scoring, these are not indi- cators of severe alcohol dependence. This finding is important for ED-based clinical practice because there is increasing evidence that patients with low dependence symptoms can be impacted success- fully by brief counseling in the ED and other clinical settings. 15'41-47 Brief counseling is defined as minimal interaction with a medical or mental health professional focusing on the health risks associated with drinking. The brief counseling in this study included feedback of AUDIT results, explanation of how the patient's score in AUDIT related to nor- mative group scores, emphasizing responsibility to change, giving advice on what to change, provid- ing a menu of different change strategies, commu- nicating empathy to the patient, and encouraging the patient's self-efficacy. 5J 1 Brief counseling ap- pears to be a suitable way to arrest the progression of alcohol dependence or reverse the early stages of dysfunctional drinking. 4s The fact that 91% of screen-positive patients exhibited mild to moderate problems rather than dependence lends hope to early and brief intervention conducted in the E D - - an alternative public health care site where patients might receive help that saves them from the de- strnctive consequences of alcoholism.

ED .PATIENT ALCOHOL PROBLEMS

Sociodemographic Differences Between Screen-Positive and Screen-Negative Patients

There were significant sociodemographic differences between screen-positive and screen-negative patients on all selected variables. Screen-positive patients tended to be tobacco-using, young single men who were in college, "unemployed, and living with a roommate. In addition,,screen-positive patients had made at least one past attempt to change their drinking behavior. These factors essentially constituted a risk profile for alcohol problems among the ED patient population in this study. So- ciodemographic protective factors for alcohol problems included being older (ages 23-29); married; women; employed; being less educated (having a high school education or less); not being a student; and not having attempted to change drinking behavior in the past.

These fundamental differences provide a risk and protective profile of the young adult ED patients in this study. The researchers are not suggesting that sociodemographic markers be used as a substitute for screening instruments. However, a risk profile may provide ED professionals with a framework to further investigate potential alcohol problems, especially if diagnostic screening tools are not available or feasible in a critical care circumstance. As intended, this study had the advantage of a screening tool to determine alcohol use status for research purposes. The use of such a tool may not be time or resource feasible to use with all patients or in all EDs, or in other critical care environments. Screen- ing all patients may not be necessary. However, sociodemographic data such as those collected in this study may alert ED staff to high-risk patients, po- tentially bringing greater efficiency to an alcohol screening process.

Sociodemographic Predictors of Alcohol Problems

The most striking predictors of alcohol problems were previous attempts to cut back on drinking, being a man, using tobacco, and being single, respectively. These were closely followed by being a college student. These findings beg the question: What are the implications of these factors for ED providers? Knowing and understanding these sociodemographic predictors may help ED providers

217

identify patients who may benefit from further query about their alcohol use. Moreover, these factors have implications for ED clinical practice related to alcohol intervention and warrant further discussion. The relevance of each of these factors to ED practice is discussed because the prediction model revealed unique and significant contribu- tions of each predictor variable.

Similar to other studies, being a male college student was a predictor of alcohol problems. For instance, one study defined 3 factors associated with increased quantity of drinking: male sex, residence in a fraternity or sorority, and a history of conduct problems. 41 As another group of researchers pointed out, heavy drinking among students has been a major health concern over the past decade and early onset of drinking poses an increased risk for lifetime alcohol-related problems. 49 One recent study found that screening in the ED identified a considerable number of young people who might benefit from brief intervention.29 The current study findings confirmed that men were prone to drink more excessively than women, especially among college students. Moreover, the screen-positive patients tended to be single and in their undergradu- ate years of study.

These findings should alert ED personnel who have frequent contact with college students about the seriousness of the risks they face. Attention to these issues is essential, particularly during the first years in college. This is a period when many students are away from home for the first time. It also is a period marked with increased feelings of lone- liness, worry, uncertainty about oneself, increased need to belong, and, consequently, attempts to es- tablish oneself in a social group. College social life frequently involves excessive alcohol consumption. Taking these facts into account, ED personnel can briefly and effectively administer intervention that addresses recognition of social reasons for drinking, positive coping skills, simple stress-relief tech- niques, and education about socialization and positive relationship building without the use of alcohol. Problem drinking should not be dismissed as normal college behavior. ED personnel may have the opportunity to intervene before problems be- come severe. This also may provide an opportunity for university-based hospital EDs to partner with Student Assistance Programs and other student counseling services for intervention efforts.

218

In addition, unlike other health care providers, ED personnel may have the opportunity to relate alcohol use to presenting medical problems; these teachable moments have been shown to be effective for alcohol reduction. 2~176 One group of researchers conducted a patient education and referral services study among adult patients at an urban, inner-city trauma center. 14 Patients identified as drinkers were administered a brief interview for which the goal was to explore conflicting motiva- tions and to negotiate possible strategies for change depending on the patient's readiness. This strategy resulted in favorable change. To illustrate, some patients were referred to the appropriate treatment fa- cility. Patients who kept their appointments reported significant reductions in alcohol use and the amount of drinks consumed at one sitting.

A majority of screen-positive patients in the current study also used tobacco. ED providers may simply consider tobacco use a potential predictor of alcohol problems or they may consider address- ing it as part of a brief intervention. In fact, ad- dressing tobacco use may enhance the likelihood of alcohol reduction. 51 As one group of researchers pointed out, dual-substance intervention by using a brief intervention approach may be feasible in the ED. 5 Moreover, another group of investigators stressed that ignoring nicotine addiction can un- dermine efforts to help patients understand their chemical dependency problems. 52 There is evidence to show that when individuals quit using tobacco in concert with changing alcohol use behavior, they improve chances of success with both behaviors. 51 In addition, because most insurance companies continue to refuse payment for treatment of nicotine dependence, coupling it with alcohol intervention provides a valuable service. 52 Nicotine addiction leads to substantial morbidity and mortality, and the achievement of abstinence from tobacco can be life saving. Because of com- mon psychologic and physiologic issues (eg, emo- tional stress and physical withdrawal) about tobacco and alcohol use, it is reasonable that ED personnel could address tobacco and alcohol use conjointly, with minimal time or resource intrusion on routine care. 5

Limi ta t ions

The use of sociodemographic characteristics as the sole predictors of patient alcohol problems is a potential limitation. Although one of the purposes

HORN ET AL

of this investigation was to show that routine information could be used to predict alcohol problems, future studies should examine information such as individual difference variables and comor- bid psychologic or physical health problems, other drug use, and readiness to change behavior. Such factors would not only strengthen the prediction model, but would help determine patients' potential responsiveness to intervention.

A second potential limitation is that the study was conducted in a university-based ED with a ho- mogenous sample of young adults, ages 18 to 29. As noted previously, research shows that alcohol problems increase as age decreases, thus the young adult population is an appropriate and necessary population for the study of alcohol use. Nonethe- less, some of the patients in this study were college students and not all EDs or critical care units serve a college population. As such, it is possible that sociodemographic predictors of alcohol problems may differ with an older population of adults or among noncollege students. Consistent with previous research, however, the researchers aver that the risk profile found in this study will hold for the general young adult population. Further research is needed to assess potential differences. Moreover, the field may benefit from additional research to incorporate the sociodemographic red flags found in the present study in a screening tool having high K values for interobserver agreement and high specificity and sensitivity levels for fast detection of alcohol problems in the ED.

CONCLUSION

In conclusion, several sociodemographic factors routinely collected in EDs may be useful for identifying young adult patients at risk for alcohol problems, particularly factors such as being a single male, tobacco-using young adult, who already tried to quit drinking at least once. Effective identification of at-risk patients is the first step in improving the health care system's response to alcohol problems and may increase further screening, appropriate treatment, or referral for specialized care. Although not shown by data in this study, programs designed to help patients reduce or totally stop alcohol consumption at early stages may help de- crease the health care strain and resource use and the potential harmful behavior associated with alcohol misuse . 53 Further research is warranted to elu- cidate the feasibility, efficacy, and cost effective-

ED PATIENT ALCOHOL PROBLEMS 219

ness of such an approach and whether improved screening leads to intervention.

Many ED providers may have the opportunities to identify and intervene with patients in early stages of alcohol abuse, without the use of time- consuming diagnostic tools. This type of effort is essential for reducing patient risk for alcohol- related illness, injury, and other addictions. 2~176 The present study findings shed light on the sociodemographic profile of ED patients with alcohol problems. In summary, several important findings emerged: (1) patients with alcohol problems, along with the severity of their problems, were identifi-

able in less than 15 minutes; (2) alcohol problems were predicted by using routine ED patient information, in turn, serving as red flags for ED-based providers; and (3) the majority of the patients were in the early stages of alcohol problems, suggesting that brief intervention may be appropriate and effective. The bottom-line message is that if ED- based providers and other critical care providers routinely screened patients for alcohol problems or could predict potential alcohol problems based on routinely collected information, then further screening, intervention, or referral for intervention may increase.

REFERENCES

1. Substance Abuse and Mental Health Statistics. Summary of findings from the 2000 National Household Survey on Drug Abuse. Accessed on October 23, 2001. Available: http://www. drugabusestatistics.samhsa.gov/oas.cfm

2. Holder HD: Alcoholism treatment and potential health care cost savings. Med Care 25:52-71, 1987

3. The Economic Cost of Alcohol and Drug Abuse in the United States--1992. National Institute on Drug Abuse. Na- tional Institute on Alcohol Abuse and Alcoholism. Accessed on October 23, 200t. Available: http://www.health.org/govstudy/ bkd265/Chapter4.htm

4. Institute of Medicine: Broadening the base of treatment for alcohol problems. Washington, DC: National Academy Press, 1990

5. Horn K, Xin G, Williams J, et al: Conjoint smoking and drinking: A case for dual substance intervention among young emergency department patients. Acad Emerg Med 7:1126-1134, 2000

6. Cherpitel CJ: Drinking patterns and problems associated with injury status in emergency room admissiorlg. Alcohol Clin Exp Res 12:105-110, 1988

7. Cherpitel CJ: Alcohol and violence-related injuries: An emergency room study. Addiction 88:79-88, 1993

8. Cherpitel CJ: Alcohol consumption among emergency room patients: Comparison of county/community hospitals and HMO. J Stud Alcohol 54:432-440, 1993

9. Cherpitel CJ: Alcohol and injuries: A review of inter- national emergency room studies. Addiction 88: 651-665, 1993

10. Becker B, Woolard R, Nirenberg TD, et al: Alcohol use among subcritically injured emergency department patients. Acad Emerg Med 2:784-790, 1995

11. Hungefford DW, Williams JM, Fnrbee PM, et al: The feasibility of screening and brief intervention for alcohol problems (SBI) among young adults in a rural, university-affiliated emergency department. Am J Emerg Med 2002 (in press)

12. Holt S, Stewart IC, Dixon JM, et al: Alcohol and the emergency service patient. Br Med J 281:638-640, 1980

13. Lapman SC, Brown P, Suriyawongpaisal P, et al: Use of AUDIT for alcohol screening among emergency room patients in Thailand. Subst Use Misuse 34:1881-1895, 1999

14. Bernstein E, Bernstein J, Levenson S: Project ASSERT and ED-based intervention to increase access to primary care,

preventive services, and the substance abuse treatment system. Ann Emerg Med 30:18t-189, 1997

15. Fleming MF, Barry KL, Manwel LB, et al: Brief physician advice for problem alcohol drinkers. JAMA 277:1039-1045, t997

16. Cherpitel CJ, Clark WB: Ethnic differences in per- formauce of screening instruments. Alcohol Clin Exp Res 19:682-634, 1995

17. Bernstein E, Tracey A, Bernstein J, et al: Emergency department detection and referral rates for patients with problem drinking. Subst Abuse 17:69-76, 1996

18. Whiteman PJ, Hoffman RS, Goldfrank LR: Alco- holism in the emergency department: An epidemiologic study. Acad Emerg Med 7:14-20, 2000

19. Betsy T, Herring R, Judd A: Identifying alcohol- related harm in young drinkers: The role of accident and emergency departments. Alcohol Alcohol 34:910-915, 1999

20. Minugh PA, Nirenberg TD, Clifford PR, et al: Analy- sis of alcohol use clusters among subcritically injured emergency department patients. Acad Emerg Med 4:1059-1067, 1997

21. Williams JM, Chinnis AC, Gutman D: Health pro- motion practices of emergency physicians. Am J Emerg Med 18:17-21, 2000

22. Lowenstein SR, Weissberg M, Terry D: Alcohol in- toxication, injuries, and dangerous behaviors--and the re- volving emergency department door. J Trauma 30:1252- 1257, 1990

23. D'Onofrio G, Berbstein E, Bernstein J, et al: Patients with alcohol problems in the emergency department, part 1: Improving detection. Acad Emerg Med 5:1200-1206, 1998

24. Saunders JB, Aaslund OG, Babor TF, et al: Devel- opment of Alcohol Use Disorder Identification Test (AU- DIT): WHO collaborative project on early detection of per- sons with harmful alcohol consumption--II. Addiction 88:791-804, 1993

25. Allen JR Litten RZ, Fertig JB, et al: A review of research on the Alcohol Use Disorders Identification Test (AUDIT). Alcohol Clin Exp Res 21:613-619, 1997

26. Bohn MJ, Babor TF, Kranzler HR: The Alcohol Dis- order Identification Test (AUDIT): Validation of a screening instrument for use in medical setting. J Stud Alcohol 56:423- 432, 1995

220

27. Skipsey K, Burleson JA, Kranzler HR: Utility of the AU- DIT for identification of hazardous or harmful drinking in drag- dependent patients. Drug Alcohol Dependence 45: 157-t63, I997

28. Bradley KA: Brief physician advice for problem drinkers. JAMA 278:1059, 1997

29. Chung T, Colby SM, Barnett N, et al: Screening adoles- cents for problem drinking: Performance of brief screening against DSM-1V alcohol diagnoses. J Stud Alcohol 61:579-587, 2000

30. Fleming MF, Barry KL, MacDonald R: The Alcohol Use Disorder Identification Test (AUDIT) in a college sample. Int J Addict 26:1173-1185, 1991

31. Irving LM, Seinder AL, Burling TA, et al: Drug and alcohol abuse in patients' attitudes about smoking cessation. J Subst Abuse 6:267-278, 1994

32. Conigrave KM, Hall WD, Saunders JB: The AUDIT questionnaire: Choosing a cut-off score. Addiction 90:1349- 1356, 1995

33. Seppa K, Makela R, Sillanaukee P: Effectiveness of the Alcohol Use Disorder Identification Test in occupational health screening. Alcohol Clin Exp Res 19:999-1003, 1995

34. Isaacsen JH, Butler R, Zacharek M, et at: Screening with Alcohol Use Disorder Identification Test (AUDIT) in a inner city population. J Gen Intern Med 9:550-553, 1994

35. Schmidt A, Barry KL, Fleming MF: Detection of problem drinkers: The Alcohol Use Disorders Identification Test (AUDIT). South Med J 88:52-59, 1995

36. Reinert DF, Allen JP: The Alcohol Use Disorder Identi- fication Test (AUDIT): A review of recent research. Alcohol Clin Exp Res 26:272-279, 2002

37. Helmkamp JC, Hungerford DW, Williams JM, et al: Screening and intervention for alcohol problems among college students treated in a university hospital emergency department. Am J College Health 2002 (in press)

38. Licht MH: Multiple regression and correlation, in Grimm LG, Yamold PR (eds): Reading and Understanding Multivariate Statistics. Washington, DC: American Psychological Associa- tion, 1994

39. Hutcheson G, Sofroniou N: The Multivariate Social Sci- entist. London: Sage, 1999

40. SAS Institute Inc.: SAS Procedure Guide, Version 7-1. Cary, NC: SAS Institute, 1999

HORN ET AL

41. Baer JS, Kivlahan DR, Marlatt GA. High-risk drinking across the transition from high school to college. Alcohol Clin Exp Res 19:54-61, 1995

42. Babor TF, Treffardier M, Weill J, et al: The early detection and secondary prevention of alcoholism in France. J Stud Alcohol 44:600-616, 1983

43. Babor TF, Ritson EB, Hodgson RJ:Alcohol-relatedprob- lems in the primary health care settihg: A review of early intervention strategies. Br J Addict 81:23-46, 1986

44. Berg G, Skutle A: Early intervention with problem drinkers, in Miller WR, Heather N (eds): The Addictive Behavior: Process of Change. New York: Plenum Press, 1986

45. Parish DC: Another indication for screening and early intervention. JAMA 277:1079-1080, 1997

46. Guarnieri M: Brief physician advice for problem alcohol drinkers: A randomized controlled trial in community-based primary care practices. Comment. JAMA 278:1059, 1997

47. Manwell LB, Fleming MF, Mundt MR et al: Treatment of problem alcohol use in women of childbearing age: Results of a brief intervention trial. Alcohol Clin Exp Res 24:1517-1524, 2000

48. Cohen S: Methods of intervention, in de Luca J (ed): Pre- vention, Intervention, and Treatment: Concerns and Models. Washington, DC: U.S. Government Printing Office, 1982, pp 127-143

49. Chou AR Pickering RP: Early onset of drinking as a risk factor for lifetime alcohol-related problems. Br J Addict 87:1199-1204, 1992

50. Mitka M: "Teachable moments" provide a means for physicians to lower alcohol use. JAMA 279:1767-1768, 1998

52. Hughes JR: Combining behavioral therapy and pharma- cotherapy for smoking cessation: An update, in Onken LS, Blaine JD, Boren JJ (eds): Integrating Behavior Therapies With Medication in the Treatment of Drng Dependence. NIDA Re- search Monograph. Washington, DC: U.S. Government Printing Office, 1995; pp 92-109

52. Bobo JK, Slade J, Hoffman AL: Nicotine addiction counseling for chemically dependent patients. Psychiatr Serv 46:945- 947, 1995

53. Dimeff LA, Baer JS, Kivlahan DR, et al: Brief Alcohol Screening and Intervention for College Students. A Harm Re- duction Approach. New York: The Guilford Press, 1999

alcohol problems among young adult emergency department patients: making predictions using routine...

Documents