US Emergency Department Visits for Alcohol-Related Diseases and Injuries Between 1992 and 2000

Background  Alcohol-related diseases and injuries pose a significant burden on hospital emergency departments (EDs). Recognized limitations of self-reported data suggest that previous single-year national studies may have underestimated the magnitude of this burden.

Methods  Data were obtained from the National Hospital Ambulatory Medical Care Survey for 1992 through 2000. Thirty-seven alcohol-related diagnoses and their corresponding alcohol-attributable fractions (AAFs) were used to estimate the number of ED visits attributable to alcohol. Diagnoses with an AAF of 1 were analyzed by age, sex, and race. Disposition to inpatient settings and alcohol screening also were examined.

Results  During these 9 years, there were an estimated 68.6 million (95% confidence interval [CI], 65.6 million to 71.7 million) ED visits attributable to alcohol, a rate of 28.7 (95% CI, 27.1-30.3) per 1000 US population. The number of alcohol-related visits increased 18% during this period. Visit rates for diagnoses with AAFs of 1 were highest for those who were aged 30 through 49 years, male, and black. From 1992 to 2000, these disparities remained stable for age group but significantly changed for sex (+22%) and race (−76%). Most patients with diagnoses with AAFs of 1 were not admitted to an inpatient unit, and the percentage of patients who underwent blood alcohol concentration testing was substantially lower than corresponding AAFs.

Conclusions  Alcohol-related ED visits are approximately 3 times higher than previous estimates determined by physician documentation or patient disclosure of alcohol involvement. Rising trends, changing disparities, and suboptimal ED management of such visits are a call to action.

Healthy People 2010,¹ the national health promotion and disease prevention agenda, cites reducing alcohol-related emergency department (ED) visits as 1 of its 467 objectives for this decade. Indeed, US public health officials recognize that alcohol misuse is a major public health issue faced by our nation and poses a significant burden on hospital EDs. Emergency departments frequently encounter alcohol-related injuries as consequences of short-term exposure to alcohol as well as alcohol-related diseases as consequences of long-term alcohol use. In addition, public health officials recognize that such alcohol-related ED visits offer opportunities for early intervention and referral of patients for appropriate care, which may reduce subsequent injury, illness, or death.² Therefore, policy measures that help reduce alcohol-related ED visits and increase early intervention and appropriate referrals at the time of ED visits may subsequently reduce the prevalence of alcohol-related health problems in the United States.

To create such policies and set objectives, one might begin by examining the magnitude of the burden of alcohol-related visits on EDs in recent years. Previous national studies^3,4 in this area have estimated the number of annual alcohol-related ED visits in the United States based on medical record documentation of physician diagnoses, patient complaints, or relevant laboratory testing results, which indicate that a visit is alcohol related. However, numerous studies² have shown that patients often fail to disclose their drinking habits to physicians and ED physicians often fail to identify signs of alcohol misuse or order the appropriate screening tests. Gordon et al,⁵ for example, found that interns staffing a hospital ED failed to identify 84% of heavy drinkers. Solomon et al⁶ found that ED physicians recognized only half of patients who would have been classified as alcohol misusers by the Alcohol Abuse Scale. Therefore, estimates of the number of alcohol-related ED visits in the current literature are likely conservative. In addition, previous national studies have examined ED visits for only 1 year at a time, thus limiting the ability to examine trends over several years.^3,4 There also is limited knowledge of ED screening and visit disposition of alcohol-related visits on a national scale.

This study examines alcohol-related ED visits at a national level, without relying on physician medical record documentation or patient disclosure of alcohol involvement in ED visits. By examining data during a 9-year period, from 1992 through 2000, we also discuss trends in alcohol-related ED visits as well as ED visit disposition and screening during such visits in this period.

Data from the 1992-2000 National Hospital Ambulatory Medical Care Survey (NHAMCS) were combined to generate national estimates of alcohol-related ED visits.^7-16 The NHAMCS was a 4-stage probability sample of visits to noninstitutional general and short-stay hospitals, excluding federal, military, and Veterans Affairs hospitals, in the United States. The NHAMCS was conducted annually and covers geographic primary sampling units, hospitals within primary sampling units, EDs within hospitals, and patients within EDs. At most participating hospital EDs, hospital staff completed the information requested on the NHAMCS patient record forms, including physicians' diagnoses, during a randomly assigned 4-week data period for each of the sampled hospitals during each year of the study period. At a few hospital EDs, trained field representatives abstracted data from medical records and recorded the data on the NHAMCS patient record forms. When the data collection forms were completed, they were sent to the National Center for Health Statistics, where the physicians' diagnoses were coded using the International Classification of Disease, Ninth Revision, Clinical Modification. National estimates were obtained through use of assigned patient visit weights and were rounded to the nearest thousand. A multistage estimation procedure consisted of inflation by reciprocals of the sampling selection probabilities, adjustment for nonresponse, and a population weighting ratio adjustment. A detailed description of the NHAMCS data collection and estimation procedures is available for review in the technical notes section of each year's NHAMCS Emergency Department Summary.^7-15

In this study, ED visits with 1 of 37 alcohol-related diagnoses were considered (Table 1). Each of these 37 diagnoses was assigned an alcohol-attributable fraction (AAF), computed and published in the Alcohol-Related Disease Impact Software developed, at the request of the Centers for Disease Control and Prevention, to enable health care officials to estimate the health impact of alcohol use and misuse.¹⁷ The AAF for each diagnosis, which was calculated based on extensive review of clinical case series studies, injury surveillance studies, and available epidemiologic investigations, represents the proportion of disease cases, injuries, or deaths that are causally linked to alcohol use or misuse. Therefore, the number of alcohol-related ED visits for each of the 37 diagnoses was calculated by multiplying the total number of ED visits for each diagnosis by its AAF. For example, from 1992 through 2000 there were approximately 1.2 million ED visits during which a diagnosis of acute pancreatitis was documented. The AAF of acute pancreatitis was estimated at 0.42; therefore, the number of alcohol-related ED visits for acute pancreatitis was approximately 500 000 during this period. The total number of alcohol-related ED visits from 1992 through 2000 was calculated by adding the number of alcohol-related visits for all 37 diagnoses.

The ED visit rates also were reported per 1000 per year for the US population, using midyear age-, sex-, and race-specific population estimates for 1992 through 2000 from the US Census Bureau.^18,19 The overall alcohol-related ED visit rates were calculated for each year and the entire 9-year period.

The 37 diagnoses were divided into 3 groups for further analysis: (1) diagnoses that were 100% attributable to alcohol by definition (AAF of 1), such as alcohol-dependence syndrome or alcoholic cirrhosis; (2) diseases with an AAF of less than 1 (weighted average of AAFs of 0.08, calculated by dividing the total number of ED visits with one of these diseases attributable to alcohol by the total number of ED visits with one of these diseases); and (3) injuries with an AAF of less than 1 (weighted average of AAFs of 0.34, calculated by dividing the total number of ED visits with one of these injuries attributable to alcohol by the total number of ED visits with one of these injuries). For those with diagnoses with an AAF of 1 only, ED visit rates were analyzed by age group (15-19, 20-29, 30-39, 40-49, 50-59, 60-74, and ≥75 years), sex (male and female), and race (white and black).

Trends were analyzed for the 9-year study using STATA 7.0 software (StataCorp, College Station, Tex). To assess a change in trend over time, 95% prediction bands were calculated by fitting a line to all years except 2000. The 2000 data point was considered significantly different from the established trend when outside the 95% prediction band. Confidence intervals (CIs) were calculated using the relative SE of the estimate.

In addition, 2 aspects of ED management (visit disposition to an inpatient setting and screening for alcohol use by blood alcohol concentration [BAC] testing) were considered. The percentages of those with diagnoses that were 100% attributable to alcohol with disposition to an inpatient setting (admitted to the hospital or transferred to other facilities) from 1992 through 2000 were determined and analyzed by age, sex, and race. Data were not available on disposition to treatment for alcohol abuse, in particular. The percentages of those within each of the 3 groups of diagnoses who received BAC testing from 1995 through 2000 (data not available for 1992-1994) were also determined and analyzed by age, sex, and race for those with diagnoses that were 100% attributable to alcohol.

The estimated total number of ED visits attributable to alcohol from 1992 through 2000 was 68.6 million (95% CI, 65.6-71.7 million), which averages to 7.6 million alcohol-related ED visits per year. Alcohol-related visits accounted for 7.9% of the total 866.5 million ED visits from 1992 through 2000. Overall, the estimated annual rate of ED visits attributable to alcohol was 28.7 (95% CI, 27.1-30.3) visits per 1000 population (Table 2).

Considering only those diagnoses that were 100% attributable to alcohol (AAF of 1), the estimated rate of ED visits from 1992 through 2000 was 5.1 (95% CI, 4.6-5.5) visits per 1000 population (Table 3). The ED visit rates for those with such diagnoses varied according to age, sex, and race. The visit rate for those aged 30 through 49 years was approximately 2 times the rates for those aged 15 through 29 years and 50 years or older. The visit rate for males with such diagnoses during this period was approximately 3 times that for females. The visit rate for blacks with such diagnoses was approximately 2 times that for whites from 1992 through 2000. These differences between sexes and races were most pronounced for those aged 40 through 49 years (Figure 1).

From 1992 through 2000, there were overall upward trends in the total number of alcohol-related ED visits and the annual rate of ED visits attributable to alcohol (Figure 2). During this period, the total number of alcohol-related ED visits increased approximately 18% (P<.001), whereas the overall rate for alcohol-related ED visits increased approximately 8% (P = .002). The percentage of the total number of ED visits that were attributable to alcohol remained fairly constant from 1992 through 2000 (Table 2).

Considering the ED visit rates for those with diagnoses that were 100% attributable to alcohol (AAF of 1), the differences between age groups remained fairly constant from 1992 through 2000 (Table 3, Figure 3). By contrast, the difference between sexes increased approximately 22% from 1992 through 2000 (Table 3, Figure 3). The difference between races more than doubled between 1992 and 1999, but declined 89% between 1999 and 2000 (Table 3, Figure 3). The 2000 visit rate for blacks was significantly different from the established trend from 1992 through 1999. Comparing 1992 to 2000, the difference between visit rates for whites and blacks with such diagnoses decreased approximately 76%.

Approximately 20% of those with diagnoses that are 100% attributable to alcohol were admitted to the hospital, whereas approximately 11% were transferred to another facility. There were significant differences between age groups for those admitted to the hospital. Approximately 14% (95% CI, 8%-20%) of those aged 15 through 29 years, 19% (95% CI, 15%-23%) of those aged 30 through 49 years, and 26% (95% CI, 19%-33%) of those 50 years and older were admitted. However, there were no major differences by sex or race, with admission rates of 20% for males, 20% for females, 20% for whites, and 18% for blacks. Approximately 11% of those with diagnoses that are 100% attributable to alcohol were transferred to another facility. Transfer was more likely among patients aged 15 through 49 years (12%) compared with those 50 years and older (7%, P = .007). As with admission, there were no significant differences in transfers according to sex or race (data not shown).

Among those visiting the ED for diagnoses that were 100% attributable to alcohol (AAF of 1) from 1995 through 2000, approximately 45% (95% CI, 41%-49%) underwent BAC testing, with no significant difference between age groups. Approximately 46% (95% CI, 41%-51%) of those aged 30 through 49 years underwent such testing, whereas 44% (95% CI, 36%-52%) of those 50 years and older underwent such testing. For those aged 15 through 29 years, approximately 45% (95% CI, 36%-54%) underwent BAC testing. There were no significant differences by sex or race for those undergoing BAC testing when visiting the ED for such diagnoses, with rates of 44% for males, 48% for females, 45% for whites, and 46% for blacks. During this same period, approximately 1.5% (95% CI, 1.1%-1.9%) of those visiting the ED for diseases with AAFs of less than 1 (weighted average of AAFs was 0.08, or 8%) underwent BAC testing, and approximately 2.8% (95% CI, 2.5%-3.1%) of those with injuries with AAFs of less than 1 (weighted average of AAFs was 0.34, or 34%) underwent BAC testing. The number of visits studied was too small to analyze differences by age, sex, or race for these groups of diagnoses. However, for all 3 groups of diagnoses, the percentages of patients who underwent BAC testing were significantly lower than the groups' weighted average AAFs.

Although US public health officials recognize that EDs throughout the United States face an enormous burden from alcohol-related diseases and injuries, this study shows that the current literature significantly underestimates the magnitude of this burden. Although previous studies rely on patient disclosure and/or physician documentation of the alcohol-relatedness of presenting diseases or injuries, this study does not rely on such disclosure or diagnosis, thus suggesting a likely reason for previous underestimation. Our 9-year study also reveals a rising trend in the number and rate of alcohol-related ED visits and a widening gap between sexes and a shrinking gap between races among those seen in the ED with certain alcohol-related diagnoses. In addition, ED screening seems to be nonconcordant with current recommendations, and disposition to inpatient settings occurs in a few visits with specific alcohol-related diagnoses.

Our results reveal that there were approximately 68.6 million alcohol-related ED visits in the United States between 1992 and 2000, averaging 7.6 million visits per year. The numbers of alcohol-related ED visits found for particular years in our study are approximately 3 times higher than those previously reported in single-year studies that relied on patient disclosure or physician documentation of alcohol involvement in ED visits. For instance, Nelson and Stussman³ report that in 1992 there were an estimated 2.8 million alcohol-related ED visits, accounting for 3.1% of all ED visits, whereas our study estimates that figure to be 7.1 million alcohol-related ED visits or 7.9% of all ED visits. In addition, the rate of alcohol-related ED visits is estimated at 27.7 visits per 1000 population for 1992 in our study, whereas Nelson and Stussman³ estimate the rate for both alcohol and drug-related ED visits combined at 16.0 visits per 1000 population for 1992. Similarly, Li et al⁴ report that there were approximately 2.6 million alcohol-related ED visits in 1995, accounting for 2.7% of all ED visits, whereas our study estimates that total at 7.6 million or 7.9% of all ED visits. The rate for alcohol-related ED visits is estimated at 28.9 visits per 1000 population for 1995 in our study, whereas Li et al⁴ estimate the rate at 10.0 visits per 1000 population for 1995. Thus, our analysis suggests that the burden of alcohol-related diseases and injuries on EDs is likely much greater than previously reported.

The 9-year period of this study allows us to observe trends in alcohol-related ED visits. The total number and rate of alcohol-related ED visits in the United States increased significantly from 1992 to 2000. However, the volume of total ED visits increased approximately 20% from 1992 to 2000. Further analysis reveals that the percentage of total ED visits that were alcohol related remained relatively constant throughout this period, suggesting that the increase in alcohol-related ED visits reflects an increase in ED visits in general. However, as the total number of ED visits overall between 1992 and 2000 increased, the case mix consisted of a greater percentage of visits for illness rather than injury conditions.¹⁴ Therefore, perhaps further research that addresses trends for specific alcohol-related diagnoses may elucidate specific types of alcohol-related problems that should be targeted by prevention programs or policies.

Trends in the demographics of alcohol-related ED visits can be observed for those visits with diagnoses that were 100% attributable to alcohol. Those patients aged 30 through 49 years consistently had the highest visit rate among those with such a diagnosis, but the gap between age groups studied did not change substantially from 1992 to 2000. The rate of such alcohol-related ED visits was higher in males and blacks than in females and whites from 1992 to 2000. The difference between sexes increased between 1992 and 2000, whereas the difference between races decreased significantly during this period. However, from 1992 to 1999, the difference between races increased by more than 100%, whereas that difference sharply declined from 1999 to 2000. Such changing gaps could represent changes in sex and race disparities in alcohol misuse, diagnosis of such alcohol misuse, or ED utilization. Further study of factors that have led to these specific changes might help shape policies that seek to reduce such disparities. Additionally, policy measures that target those in their 30s and 40s, males, and blacks may effectively slow the increase in, or perhaps even decrease, the total number of alcohol-related ED visits.

In addition to reducing the rate of alcohol-related ED visits, policies also should seek to improve ED management of such visits. Few other specialists see more of the negative consequences of alcohol abuse than the emergency physician, thus rendering ED visits valuable opportunities to initiate early screening, intervention, and referral.² Indeed, patients are often more receptive to intervention in moments of crisis, such as an illness or injury that requires acute care.²⁰ Thus, intervention at the time of ED visits may be crucial. Our study shows that most patients visiting the ED with diagnoses that are 100% attributable to alcohol were not admitted to an inpatient unit and were therefore discharged with a risk of no follow-up medical care near the time of their illness or injury. Although we do not expect that admission to an inpatient facility would be warranted in all such cases, we believe that more such admissions would have allowed further necessary interventions at a time when patients were likely most receptive. In addition, although those aged 30 through 49 years more frequently visit the ED with such diagnoses, those 50 years and older are more frequently admitted to the hospital. Thus, interventions that are performed in the ED are critical for alcohol-related visits, considering that often the only health care received at the time of a crisis is in the ED.

To intervene, effective screening for alcohol use in the ED is required. Laboratory testing (eg, BAC) may be helpful either when patients are not forthright about alcohol use or when clinical suspicion is high and is recommended in most cases of trauma.² For instance, a positive BAC test result in injured patients after a road crash increases the chance that the final diagnosis will include more injuries than initially documented, indicating a need for more careful monitoring of alcohol-positive trauma patients.²¹ This study reveals that BAC testing among patients with alcohol-related diagnoses was substantially lower than the AAFs of such diagnoses. Although we expect that BAC testing would not be indicated in all alcohol-related visits, we expect that the frequency of such testing would be closer to the frequency of alcohol-related visits studied. Thus, the BAC screening practices studied are nonconcordant with current recommendations, although no significant differences in such testing among age groups, sexes, and races are revealed in this study.

Several other aspects of the treatment of patients visiting the ED for alcohol-related problems offer room for improvement. Referral for further evaluation and treatment of alcohol abuse are often the most appropriate dispositions for the patient. The Society for Academic Emergency Medicine Substance Abuse Task Force recommends that each ED establish a referral system that considers local resources available for alcohol treatment, including detoxification sites, hospital-based treatment programs, community outpatient treatment programs, and information on Alcoholics Anonymous.²² It is the experience of the authors that such referrals and interventions are not occurring frequently enough, and future studies should investigate national trends in such aspects of ED management. Indeed, more ED-based screening and interventions may subsequently reduce the overall adverse health impact of alcohol, as well as the subsequent number of alcohol-related ED visits.

A potential limitation of our study is reliance on the diagnoses and corresponding AAFs described in the Alcohol-Related Disease Impact software.¹⁷ Some diagnoses that could be caused by alcohol may not be included. This possibility would lead us to underestimate the number of alcohol-related ED visits. In addition, considering the limitations involved in the indirect calculations of AAFs, these fractions should be regarded as best estimates.¹⁷ The lack of age-, sex-, and race-specific AAFs also limited our ability to examine demographics of those alcohol-related visits involving diagnoses with AAFs of less than 1 and our ability to fully discern the proportion of alcohol-related visits that included BAC testing or specific visit dispositions. Despite these limitations, data reported in this study shed light on the magnitude of the impact of alcohol-related diseases and injuries on EDs, in addition to specific aspects of ED management.

Other potential limitations include those limitations of the NHAMCS data collection and reporting process itself. Errors in medical record review and coding, for instance, would affect the accuracy of results. In addition, because NHAMCS only accounts for ED visits and not individual patients, our ability to examine the number of patients visiting the ED for alcohol-related health problems is limited. However, again, our results are estimates that reveal a great deal about the frequency and management practices of alcohol-related ED visits.

For US public health officials to continue to assess the effectiveness of policies that seek to reduce the number of alcohol-related ED visits during this decade, it is noteworthy that previously reported totals of alcohol-related ED visits appear to substantially underestimate the magnitude of the impact of alcohol on EDs. Thus, a high priority should be placed on future policies and objectives that address the underlying causes of the increasing number of alcohol-related visits and the widening gap between sexes and the sharp decrease in the gap between races (1999-2000) for specific alcohol-attributable diagnoses. Further research on ways to target prevention programs to certain populations may help direct resources to programs and policies that will slow increasing trends and may begin to decrease the frequency of alcohol-related ED visits. Additionally, improving the frequency of ED screening may lead to more appropriate referrals and interventions during alcohol-related ED visits, with a reduction in subsequent illness and additional visits to the ED. Among possible interventions studied, previous research has concluded that brief counseling by ED staff was more effective than no treatment and compared favorably with more traditional treatments.²³ One study²⁴ has shown that most ED physicians would support the implementation of such brief intervention for alcohol misuse in the ED, although most thought that the lack of sufficient time was an impediment to such intervention. One possible solution was offered by researchers who found that brief counseling intervention in the ED by an alcohol health worker (nurse specialist in alcohol problems), who also served as a liaison with specialist agencies, resulted in 65% of patients reporting decreased drinking 6 months later.²⁵ Thus, perhaps additional staff specializing in alcohol screening and interventions would be a welcomed change in EDs. Future investigations should explore such innovative screening, intervention, and referral practices to improve ED management of alcohol-related visits. Healthy People 2010 aims to reduce such alcohol-related ED visits,¹ and this important objective appears to be a larger task than previously considered.

Corresponding author and reprints: Carlos A. Camargo, MD, DrPH, EMNet Coordinating Center, Department of Emergency Medicine, Massachusetts General Hospital, 55 Fruit St, Clinics Bldg 397, Boston, MA 02114 (e-mail: emnet@partners.org).

Accepted for publication March 31, 2003.

This study was supported by the PASTEUR Educational Program and the Office of Enrichment Programs at Harvard Medical School, Boston, Mass, and the Emergency Medicine Foundation Center of Excellence Award, Dallas, Tex.