H. J. HARWOOD
M. B. REICHMAN
The Lewin Group, Falls Church, Virginia, United States of America
It is widely recognized that alcohol and drug abuse by workers can adversely affect their performance and the productivity of the workplace. The specific ways in which substance abuse can be harmful are well understood. Major elements of the costs incurred (for example, in lost productivity and earnings of workers and in deaths at the workplace) are captured in the most recent cost studies, as well as in the international guidelines for estimating the economic costs of substance abuse. However, no studies have rigorously measured the full economic burden on the workplace alone, because of the theoretical and empirical difficulties arising from the spread of the impact of substance abuse, via the markets, among employers (through lost profits), workers (through lost earnings and benefits) and consumers (through higher prices for goods and services). Both employers and workers recognize the nature of the problem and have worked together through bodies such as the International Labour Organization to find common solutions and formulate multilateral policies. Data for the United States of America show that policies are frequently established at the workplace to reduce alcohol and drug abuse by workers.
It is widely recognized that alcohol abuse by workers can adversely affect their performance and the productivity of the workplace. The specific ways in which alcohol abuse can be harmful are fairly easy to understand and their impact can be readily analysed and quantified.
The cost methodology proposed under the international guidelines for estimating the economic costs of substance abuse recognizes and takes into account the impact of alcohol abuse, without providing for a detailed breakdown of workplace costs. However, recent studies of the cost of alcohol abuse to society have covered major elements of such costs, including lost productivity and earnings of workers and deaths at the workplace. In general, published studies that attempt to estimate the impact of alcohol abuse on the workplace alone have not been available.
The lack of such studies may be due to the major theoretical and empirical problems associated with attempts to estimate the costs incurred at the workplace as a result of alcohol abuse. Employers, as actors in markets for labour, other inputs, capital and goods and services, play a key role in spreading the impact of alcohol abuse among employers (lost profits), workers (lost earnings and benefits) and consumers (higher prices for goods and services). The distribution of costs between the parties cannot be determined by using theoretical concepts, but is subject to market conditions and the bargaining ability of employers, workers and consumers.
Despite uncertainty about the causal link between alcohol abuse and its apparent economic consequences, both employers and workers recognize the nature of the problem. They have worked together through bodies such as the International Labour Organization (ILO) to formulate multilateral policies to address the problem. Data for the United States of America show that workplaces have largely embraced policies to reduce alcohol abuse by workers.
The present paper is based on a literature search conducted in the following databases: ETOH, HealthSTAR, Information on Drugs and Alcohol (IDA), MEDLINE and Substance Abuse Information Database (SAID). The publications of the following bodies were also searched: the Canadian Centre on Substance Abuse (CCSA), the United States Department of Labor, ILO, the National Institute on Drug Abuse (NIDA), the National Institute on Alcohol Abuse and Alcoholism (NIAAA) and the United States Substance Abuse and Mental Health Services Administration (SAMHSA). The databases and lists of publications were searched using various combinations of the following search terms: cost, economics, employers, employees, workers, workplace, alcohol and alcoholism.
Economic theory may be directly applied to analyse the impact of alcohol abuse by workers in the workplace, as illustrated in figure I. Worker productivity can be described in terms of the amount and value of the work done. In a market economy, the wages paid to workers are expected to equal the value of their productivity to the enterprise. Earnings are by definition the amount of time worked multiplied by the wages paid per amount of time.
Figure I. Impact of alcohol abuse on the amount and quality of production
A review of the literature reveals that alcohol abuse can affect both the amount of time worked (through absenteeism, tardiness or extended sick leave) and the productivity of workers (by making them less quality-conscious and more mistake-prone). The problems experienced by the affected workers may also have an adverse impact on the productivity of their co-workers and managers. In figure I, lost work time is reflected in a shift from H0 to H1, and reduced or impaired productivity in a shift in wages from W0 to W1. Expected productivity is W0 x H0; actual productivity is W1 x H1. The difference in value represents the performance decline due to alcohol abuse.
The challenge to management arising from alcohol abuse by workers is to achieve levels of productivity as close as possible to the expected levels. That means minimizing lost work and poor productivity due to alcohol abuse. Matters are complicated, however, by the fact that employees miss work and underproduce for a number of reasons.
Costs to the workplace arise when the actual productivity of workers is less than expected for a given wage level. However, if a worker that abuses alcohol is less productive in a job that requires less productivity and pays less, there may be no difference between expected and actual employee productivity; hence, there is no cost to the workplace. It is the hidden alcohol abusers that cost the workplace money.
Economic theory notes two further complications (figure II). First, workplaces generally establish wages based on the value of productivity of all workers in a group, the average productivity. A group of workers with hidden alcohol abusers has a lower average productivity because of the few abusers, and the entire group therefore receives a lower wage. Thus, the workplace balances expected and actual productivity, although at the cost of paying the majority of workers less than the value of their individual productivity.
Figure II. Redistribution of alcohol-related costs in the workplace
Secondly, should the workplace misread (overestimate) average worker productivity, the costs would be higher because higher average wages would be paid. The increased costs may translate into reduced profits for the workplace; however, they may also cause an increase in the price of the goods and services provided. Thus, management failure to recognize productivity problems due to worker alcohol abuse (or any other factor) will mean either reduced profits or increased prices, or some combination of the two. The costs of worker alcohol abuse are therefore borne by workers, consumers and the workplace, and their respective shares are indeterminate, theoretically.
A review of publications and other materials available from bodies such as the Department of Labor, ILO and NIDA revealed a range of workplace behaviours possibly related to alcohol. Employers should be aware of the consequences of alcohol abuse by workers at the firm level, including : increased absenteeism and tardiness; increased insurance claims for treatment and services for alcohol abuse and its various sequelae; increased use of workers compensation and sick leave; developing and implementing a substance abuse policy; testing for alcohol and drugs; development and administration of an employee assistance programme; accidents and damage; theft and fraud; increased turnover and replacement; diverted supervisory, managerial and co-worker time; friction among workers; poor decision-making; damage to a companys public image; and increased liability.
The above-mentioned consequences of alcohol abuse for the workplace are all confirmed by the substantial literature on the effects of alcohol on cognition and performance. The literature is compiled and summarized by NIAAA in publications such as the tri-yearly report entitled Alcohol and Health and the NIAAA peer-review journal Alcohol Research and Health (formerly Alcohol Health and Research World).1
Recent evidence showing the relationship of alcohol abuse to adverse workplace outcomes such as those enumerated above was collected in the United States from the National Household Survey on Drug Abuse (NHSDA). Table 1 presents data from the 1997 NHSDA on the workplace outcomes and behaviours that are associated with heavy alcohol use (generally not in the workplace) by workers.
Table 1. Adverse workplace outcomes reported for 1997 by full-time workers aged 18 to 49
|Workplace outcome||Employees reporting use heavy alcohol use (percentage)||Employees not reporting use heavy alcohol use (percentage)|
|Worked for three or more employers in the past year||8.0||4.4|
|Missed two or more days of work in the past month due to illness or injury||12.4||8.5|
|Skipped one or more days of work in the past month||11.3||5.1|
|Voluntarily left an employer in the past year||19.7||15.8|
|Fired by an employer in the past year||0.9||1.4|
|Had a workplace accident in the past year||8.5||5.3|
|Source: Substance Abuse and Mental Health Services Administration, Office of Applied Studies, National Household Survey on Drug Abuse (Washington, D.C., 1997).|
A clear increase in adverse workplace outcomes was reported by employees who were heavy users, defined as drinking five or more drinks on five or more occasions in the past month. Such outcomes are not unique to alcohol, however. The higher rates reported by heavy users indicate that they are more likely to have problems, but it is necessary to study whether the heavy users have other characteristics that may also be linked to those outcomes, such as their demographic characteristics. In fact, the population of heavy alcohol users is disproportionately composed of young males, who are also at risk of experiencing adverse workplace outcomes. An analysis will need to adjust for other factors (the attribution factor) to develop meaningful estimates of the impact of heavy alcohol consumption on workplace problems.
Alcohol-related job performance problems can be caused not only by alcohol consumption while on the job (often a cause for dismissal), but also by heavy drinking when not at work . A 1986 study analysing the hangover effect in airline pilots is frequently cited in the literature [3, 4]. Using a flight simulator, Yesavage and Leirer studied the performance of airline pilots under three different alcohol conditions . A single episode of moderate drinking (leading to an increase in blood alcohol concentration (BAC) to 0.10/100 millilitres) produced a dramatic increase, from 10 to 89 per cent, in the number of pilots unable to perform all operations correctly. Fourteen hours later, after the BAC of the pilots had returned to 0.0/100 millilitres, 68 per cent still could not perform all operations correctly.
Nevertheless, statistics regarding the impact of alcohol abuse on workplace outcomes must be interpreted with caution. While certain types of behaviour and outcome (for example, increased accidents) may be attributed, in part, to employee alcohol abuse, other factors undoubtedly contribute as well. For example, a 1989 study by Bernstein and Mahoney  found that up to 40 per cent of industrial fatalities and 47 per cent of industrial injuries can be linked to alcohol consumption and alcoholism. Researchers must disentangle the effects of alcohol from the effects of other factors such as gender, race, ethnicity, age, educational attainment, occupation and industry. The United States Bureau of Labor Statistics, in its census of fatal occupational injuries, reported 28,068 fatal occupational injuries in the United States from 1992 through 1996. The risk of occupational injury and death varies significantly across industries, with mining and quarrying, agriculture, construction, transportation and public utilities all reporting substantially higher rates of unintentional injuries and deaths .
Another example is provided by a large-scale study concerning illicit drug use and job performance in the United States Postal Service . The authors concluded that drug use as exhibited by a positive pre-employment test was significantly associated with turnover, absenteeism and other negative workplace outcomes. Butler cites the following results from another study conducted within the United States Postal Service: Applicants who had tested positive for illicit drugs were 2.67 times as likely to experience problems requiring employee assistance programme intervention, 2.44 times as likely to be formally disciplined, and 1.88 times as likely to accumulate an elevated dollar amount of medical claims than those who had tested negative for drug use . All such outcomes incur costs for the employer, the employee and others. A portion of the costs may be directly attributed to employee substance abuse. However, a cost-of-illness study that immediately assigns the full cost to drug abuse is misleading.
NHSDA, conducted annually since 1990 by SAMHSA, is the primary source of statistical information on the use of alcohol, tobacco and illicit drugs in the United States. The 1994 and 1997 NHSDA included a special module designed to collect data on the use of illicit drugs and alcohol by United States workers and on company policies on drug and alcohol use. An analysis of the 1997 NHSDA was released in 1999 by the SAMHSA Office of Applied Studies . The analysis included the following findings regarding the prevalence of current heavy alcohol use2 among workers employed in various occupations and establishments of different sizes (see table 2):
(a) Among full-time workers, 6.2 million, or 7.6 per cent, reported current heavy alcohol use;
(b) Heavy alcohol use was more common among 18- to 25-year-olds (11.7 per cent), males (11.1 per cent), Hispanics (9.8 per cent) and whites (8.1 per cent), those with less than a high school education (14.7 per cent), and those who reported personal income of $9,000-$19,999 (9.4 per cent);
(c) Workers in medium-size establishments (25-499 employees) reported the highest rate of heavy alcohol use (8.3 per cent);
(d) Current heavy alcohol use varied across occupational categories, the highest rates being accounted for by food preparation workers, waiters, waitresses and bartenders (15 per cent); handlers, helpers and labourers (13.5 per cent); and construction workers (12.4 per cent).
Table 2. Current heavy alcohol use among full-time United States workers aged 18 to 49, 1994 and 1997
Percentage of heavy alcohol users
|Over 500 employees||7.3||7.4|
|Food preparation, waiters, bartenders||12.2||15.0|
|Handlers, helpers, labourers||15.7||13.5|
|Precision production and repair||13.1||11.6|
|Transportation and materials moving||13.1||10.8|
|Machine operators and inspectors||13.5||9.0|
|Executive, administrative, managerial||6.5||7.1|
|Extraction and precision production||12.9||5.5|
|Technicians and related support||6.2||5.1|
|Source: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse (Washington, D.C., 1994 and 1997).|
It is clear that many people who abuse alcohol are employed. Though the rate varies across establishment size, occupational categories and regions, all employers are vulnerable to the adverse effects of employee alcohol abuse. The first step that employers can take to mitigate the impact of alcohol abuse in the workplace is to understand and recognize what the impact is.
Workplaces in the United States generally take specific measures to prevent or discourage employee alcohol abuse. Recent workplace surveys have found that employers (in particular medium- and large-scale employers) have several policies or services intended for that purpose. They include the distribution of basic educational and information materials, the provision of employee assistance programmes and insurance for treatment of alcoholism, policies to counter alcohol abuse and even the testing of workers for improper use of alcohol.
The United States Department of Labor  suggests a five-part programme to control substance abuse in the workplace. Each component is also found in the recommendations established by ILO, including the following: a written substance abuse policy; an employee education and awareness programme; a supervisor training programme; an employee assistance programme; and drug- and alcohol-testing, as appropriate.
Employer efforts to prevent or reduce the incidence of substance abuse among employees vary by establishment size and industry (see table 3). In general, the larger the establishment, the more likely it is: (a) to provide information about alcohol and drug use; (b) to have a written policy about drug and alcohol use; (c) to offer an employee assistance programme; and (d) to test current employees or applicants for alcohol use [1, 10, 11]. In 1997, in the occupational categories in which employers were most likely to offer information about, and to have a written policy on, drug and alcohol use, the percentage that did so for drugs and alcohol, respectively, were: in protective services, 95.2 and 92.6 per cent; in transportation and the moving of materials, 85.9 and 83.3 per cent; and in extraction and precision production, 85.5 and 81.5 per cent . The occupational categories in which employers were most likely to offer an employee assistance programme were protective services (79.7 per cent), extraction and precision production (65 per cent) and technicians and related support (61.7 per cent).
Table 3. Workplace policies covering full-time United States workers
|Item||Information about drug and alcohol use||Written policy about drug and alcohol use||Employee assistance programme|
|Over 500 employees||92.3||90.4||75.4|
|Transportation and materials moving||85.9||83.3||46.5|
|Extraction and precision production||85.5||81.5||65.0|
|Technicians and related support||78.6||75.3||61.7|
|Machine operators and inspectors||79.6||76.3||52.7|
|Precision production and repair||75.3||72.4||48.4|
|Handlers, helpers, labourers||71.7||70.0||51.3|
|Executive, administrative, managerial||71.5||70.3||49.7|
|Food preparation, waiters, bartenders||66.8||63.4||26.5|
|Source: SAMHSA, Office of Applied Studies, National Household Survey on Drug Abuse (Washington, D. C., 1997).|
Workplace testing for alcohol use, though still less common than other policies for dealing with alcohol abuse, is increasing in prevalence. A national survey of workplaces conducted in 1995 found that 21.7 per cent tested applicants and 28.4 per cent tested current employees (see table 4).
Table 4. National estimate of alcohol testing among private non-agricultural workplaces by size, industry and census region, 1995
Workplace with testing programmes, by test group
|Characteristics||Applicants||Current employees||Applicants and|
|Over 1,000 employees||33.3||40.0||55.2|
|Communications, utilities, transportation||30.3||47.4||52.7|
|Finance, insurance, real estate||5.8||7.3||11.0|
|Mining and construction||29.3||39.0||45.5|
Alcohol testing programmes can be divided into two main types: conditional and unconditional . Conditional testing consists of assessments conducted for certain employees, taking into account factors such as an accident or the determination of reasonable cause. Unconditional testing includes random and regular assessments that any employee may be required to undergo regardless of job performance or behaviour. On the whole, conditional testing is more prevalent than unconditional testing. Within workplaces that conduct conditional testing, 73.9 per cent reported testing employees following an accident and 77.9 per cent reported testing after establishing reasonable cause. Among workplaces conducting unconditional testing, random testing (48.4 per cent) is far more prevalent than regular testing (12.4 per cent).
Though alcohol testing continues to increase in prevalence, very few alcohol-testing programmes exist in workplaces that do not also test for illicit drugs. Furthermore, alcohol testing rarely occurs as a single strategy to reduce employee alcohol abuse. Testing programmes almost always occur in conjunction with formal policies and employee assistance programmes [10, 11].
Alcohol testing, on its own, is considered to be an ineffective and misleading way for an employer to combat employee alcohol abuse. A positive alcohol test does not necessarily indicate impairment. Testing programmes should be implemented in conjunction with a formal written policy that establishes the purpose for testing, rules, regulations, rights and responsibilities of all the parties concerned [3, 10, 12].
The data clearly show that employers take the problem of alcohol abuse among employees very seriously. The majority of employers, especially in large workplaces, have some form of substance abuse policy or programme. Employers have not had the opportunity to base their decisions to develop and implement substance abuse policies on empirical evidence. There is an abundance of material showing the prevalence of alcohol abuse among employees and its impact on the workplace. Such data are available at both the national and industry level [3, 9, 12]. In general, however, there is a lack of data for an analysis of the impact of alcohol abuse on specific workplaces and on the employers concerned.
While employers have access to a wealth of information regarding the prevalence of alcohol and other drug use among workers, most estimates of the prevalence and consequences (including the cost) of alcohol and other drug use are derived from large surveys of households, such as NHSDA . While information at the national level is useful, employers must base policy decisions regarding employee substance use and abuse on assessed needs in their own workplaces. Estimates of prevalence and cost at the national level may be of little use to employers in deciding on policy at the workplace or company level. Table 2 shows that rates of employee alcohol abuse vary widely across occupational categories, regions and demographic characteristics. At the very least, employers should take into account the basic demographic make-up of their workforce, determined by such factors as gender, age and educational attainment, as well as industry and occupational category, when assessing data at the national level.
The literature provides few examples of the cost of alcohol abuse in specific companies and workplaces. The vast majority of the literature on alcohol and other drug abuse in the workplace pertains to incidence and prevalence rather than evaluations of cost [3, 9, 13, 14]. Several studies have relied on self-reporting of alcohol and drug use and related workplace outcomes through employee surveys [13, 14]. Employee surveys can be valuable because workers are in the best position to assess the problem, provided that steps are taken to ensure the anonymity and confidentiality of reported results .
A 1995 study by French and others not only measured the prevalence of smoking, alcohol, illicit drug use and prescription drug use, but also reported workplace consequences such as reduced performance and absenteeism . The study used an employee survey administered to more than 1,200 randomly selected employees at five different workplaces (manufacturing services, manufacturing, municipal Government, financial services and health-care services). The survey asked employees whether their drinking and related problems (such as a hangover, an illness or an accident) had caused any of the following workplace consequences: poor performance, tardiness or leaving work early, absence, accident and injury, being high at work and needing emergency care. At three of the five workplaces, approximately 20 per cent of the drinkers reported poor performance attributable to the effects of alcohol during the past year. Fewer than 10 per cent of drinkers at each site indicated that they had been tardy or absent, or that they had left work early because of alcohol use. None of the respondents reported being hurt at work in an accident due to alcohol use, and less than 1 per cent at each site revealed that they had been high at work during the past year.
While the authors do not attempt to estimate the costs of employee drinking, they do use an employee survey to assess the workplace consequences. From a policy perspective, the value of the work of French and others is that it presents empirical results from a unique and relatively small data sample. More studies of that kind would assist employers in measuring the impact on performance and possibly the costs of substance abuse in the workplace.
In 1988, the American Productivity and Quality Center, a non-profit body that works with organizations to improve productivity and quality, undertook a research project funded by the United States Department of Labor . The purpose of the project was to determine whether the costs associated with substance abuse in the workplace could be tracked. On the basis of discussions with and reviews by experts from organizations such as the American Federation of Labor and Congress of Industrial Organizations, the National Academy of Sciences, NIAAA, NIDA and the Social Security Administration, the research team developed a total costs model that organizations could use to calculate both direct and indirect costs. The model requires the collection both of objective financial data and of more subjective performance data. The data collected were the costs incurred in dealing with the consequences of alcohol abuse listed above in the section entitled Behaviours and outcomes associated with employee alcohol abuse. The only items not covered in the data collected were the increased use of workers compensation and sick leave; friction among workers; poor decision-making; damage to a firms public image; and increased liability. The performance data were collected via two subjective employee surveys. Sample questions included the following: From your observations, how do substance abuse problems interfere with job performance?; and Have you observed substance abusers having any of the following problems: missed deadlines, excessive mistakes, decreased efficiency, etc.?
Three companies agreed to conduct a pilot test of the model, without having full access, in each case, to all of the data included in the model. However, the research team refined the model to make it adaptable to different types of organization with different data capabilities. The companies that tested the model indicated that it provided valuable information upon which to base future cost-benefit estimates. Two of the three companies indicated that they would continue to use the model. On the basis of the pilot test of the model, the research team concluded that the costs typically tracked by employers are only the tip of the iceberg. While the model proved to be a good way for the three companies that tested it to assess the impact of substance abuse on their workplaces, certain types of impact and costs still need to be measured. The data collected usually fail to capture the so-called ripple effect, that is, costs incurred when mistakes are made by substance abusers and problems and/or errors then occur down the line or ripple throughout the rest of the organization (, p. 65).
The model has several limitations. The way in which the performance data are collected could bias individual responses. The model calls for collecting data about the job performance of employees from supervisors in an interview format. Data collected in such a manner are inherently subjective and thus difficult to aggregate and compare across companies. In addition, a small sample size of only three companies, with an unknown number of employees being interviewed at each company for performance data, makes it impossible to generalize from conclusions drawn about the usefulness of the total costs model.
The total costs model was the only one of its kind found in the literature reviewed. Although some organizations may not have all the required data at their disposal, such a model would provide valuable guidance to employers in evaluating the cost of substance abuse in their workplaces.
The international guidelines for estimating the economic costs of substance abuse  are designed to cover the types of cost that result from worker abuse of alcohol. Table 5 shows the relationship between the major cost categories of the guidelines and a range of specific impacts identified in the literature review. Most of the impacts fit into the major cost categories of the guidelines.
Despite the conceptual fit, there are two problems. First, some of the specific impacts have not yet been evaluated in cost-of-illness studies. For a variety of reasons, such as lack of data and limited resources, studies have not yet determined those costs. That applies, in particular, to workplace-specific costs, such as those arising from absenteeism and tardiness, increased turnover and replacement, diverted co-worker and managerial time and friction among workers.
Table 5. Cost categories, international guidelines and estimating the impacts of alcohol abuse
|Impacts dealt with in the literature|
|Consequences to health||Increased insurance claims for alcohol abuse treatment and services and for the various sequelae of alcohol abuse|
|Prevention, research etc.||Development and administration of an employee assistance programme|
|Premature mortality||Accidents and damage|
|Lost employment or productivity||Increased absenteeism and tardiness; increased use of workers compensation and sick leave; increased turnover and replacement; diverted supervisory, managerial and co-worker time; friction among workers; and poor decision-making|
|Property destruction (fire, accident)||Accidents and damage and increased liability|
|Legal consequences||Theft and fraud|
|Outside the guidelines||Developing and implementing a substance abuse policy and testing for alcohol or drug abuse|
Secondly, the costs may have been subsumed under broader estimation categories and dealt with, in whole or in part, in recent cost-of-illness studies. For example, the costs of workplace accidents are generally included in the costs of the following: all accidents, injuries and trauma; health care; lost work due to morbidity, disability and premature death; and property destruction.
Another factor that must be taken into account in carrying out studies in different countries is the difference in economic and social systems. In some countries, for example, health services are paid for in part by insurance provided at the workplace (often with cost-sharing by the workers), while in others, national health systems are financed through payroll taxes and other tax revenues. Thus, issues related to the burden or location of costs need to be understood before workplace costs can be estimated.
There is ample evidence that alcohol problems have major impacts beyond the workplace. A wide range of such impacts have been studied and their costs estimated, but even more remain to be considered. Studies based on the international guidelines for estimating the costs of substance abuse capture many of the costs of workplace alcohol abuse, even though a detailed breakdown and analysis of workplace costs may not be provided.
In general, the available published studies fail to estimate the costs of alcohol abuse specifically related to the workplace, in particular those that fall upon employers. The studies of costs to society include important components of costs beyond the workplace, but they explicitly refrain from attributing specific amounts to workplaces and, in particular, to employers.
Economic theory shows that it is very difficult and potentially meaningless to develop estimates of the cost of alcohol problems for a particular workplace or population of workers. That is because the impacts felt and costs incurred beyond the workplace are shared or redistributed in uncertain proportions between employers, workers, customers and consumers.
Alcohol abuse by workers is clearly a matter of great concern to employers in the United States, most of whom have established policies, services, benefits or initiatives dealing with the issue. Such efforts are designed to counter the potential economic harm caused by alcohol in the workplace and by its negative impact on the efficiency, quality and costs of the business organization.
1See, in particular, vol. 19, No. 2, on alcohol and cognition (1995).
2Heavy alcohol use is defined as consuming five or more alcoholic beverages on the same occasion on five or more occasions in the past 30 days.