National Institute on Alcohol Abuse and Alcoholism No. 7 PH 278 January 1990
Alcohol Use and Abuse: Where Do the Numbers Come From?
Newspapers and magazines regularly publish stories that report estimates of the number of people who are problem drinkers, the number of traffic deaths that involve alcohol, or the amount of alcohol that Americans consume. Where do these numbers come from, and how realistic are they?
Estimates such as these often come from research in epidemiology, the study of the distribution and determinants of disease, injuries, and other health-related conditions such as alcohol use, abuse, and associated consequences (MacMahan & Pugh 1960; Mausner & Bahn 1974). These studies can tell us how many people suffer from alcohol-related problems and what groups (e.g., according to sex, age, race, occupation, place of residence) are at greater risk, thereby providing information for developing treatment and prevention programs. Studies in alcohol epidemiology address the following areas: drinking levels and patterns, prevalence of alcohol dependence, and problems arising from alcohol abuse. Our information in these areas is generally good, but there are some limitations based on the types of data available.
First, the question of how much and what types of alcoholic beverages (beer, wine, spirits) Americans are drinking and whether overall drinking levels are increasing or decreasing can be studied by monitoring the sales of alcoholic beverages (Brooks et al. 1989). Because sales data are the basis for tax collection, such reports are carefully monitored and are generally accurate. Average or per capita consumption based on overall alcohol sales is particularly useful for international comparisons (Smart 1989), but does not tell us anything about what groups may be at risk for excessive drinking.
The most common way to assess group-specific drinking patterns (e.g., typical quantity and frequency, beverage preferences, drinking contexts) is to conduct population surveys in which respondents report drinking behavior. Research has demonstrated that properly constructed questionnaires can improve reliability and validity for self-reports of drinking behavior (Williams et al.1985; Hilton 1989), and there have been many surveys of drinking behavior (e.g., Malin et al. 1986; Williams et al 1986; Hilton 1987; Brooks et al. 1989).
Although useful, surveys do have limitations. First, their costs often limit sample size, which can increase the margin of error--a special problem when studying subpopulations, such as racial or ethnic minorities and certain age groups, unless the survey plan provides for oversampling of these special groups. Even then, sample sizes are seldom large enough to identify small changes in drinking over time. Second, comparisons from one survey to another can be handicapped by differences in definitions and questions asked. Finally, self-reports of consumption in surveys tend to underestimate consumption compared with beverage sales data (Harford et al. 1988). Household and telephone surveys by their very nature tend to undersample heavier drinkers who may not live in households or who are frequently unavailable for interview. Also, some people underreport their consumption for fear that the interviewer will judge them negatively; yet individual respondents do report consuming amazingly large quantities of alcohol.
A second major area of research involves the question asked most frequently of alcohol epidemiologists: How many people are alcoholic? Early attempts at estimating the prevalence of alcoholism used variations of the Jellinek formula, based on the number of cirrhosis deaths (Brenner 1960; Schmidt & deLint 1970), or formulas based on per capita consum ption (Ledermann 1956; deLint & Schmidt 1968). Other researchers used survey methods to measure the prevalence of alcohol-related problems as an indication of the prevalence of alcoholism (Cahalan et al. 1969; Cahalan 1970; Cahalan & Room 1974), but these same studies also demonstrated that not all alcohol problems are limited to persons who are alcohol dependent.
Newer and improved estimates of the prevalence of alcoholism are based on the concept of an "alcohol dependence syndrome" (Edwards & Gross 1976), defined in the International Classification of Diseases (ICD) and in the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association. Diagnostic criteria for this syndrome have been used to develop operational definitions of dependence based on self-reported symptoms; surveys with questions based on these definitions can provide estimates of the numbers of persons whose alcohol dependence makes them appropriate candidates for treatment (Williams et al. 1987; Regier et al. 1988; Moore et al. 1989; Williams et al. in press). Because this is a relatively new area of research, methods are still evolving and refinements in diagnostic criteria and related survey questions (Grant 1989) will lead to new estimates.
A third major area of interest in alcohol epidemiology is assessing the scope of various health and social problems resulting from alcohol abuse. Many problems (e.g., with marriage or family, job, or the law) can be recognized by respondents as related to drinking and can be measured by self-reports from population surveys, with the same survey-related limitations mentioned earlier. For data on alcohol-related mortality and morbidity, however, researchers often rely on more objective sources. The three primary sources are death certificate data, the Fatal Accident Reporting System (FARS), and the National Hospital Discharge Survey (NHDS).
Alcohol-related mortality is estimated using death certificate data collected by the National Center for Health Statistics (NCHS) for all U.S. deaths (1988). The major methodological issue for studies of mortality lies in the determination of which deaths are alcohol related. Certain causes of death are by definition alcohol related (e.g., alcoholic psychoses, alcohol dependence syndrome, nondependent abuse of alcohol, certain categories of cirrhosis, excessive blood level of alcohol, and accidental poisoning by alcohol). There may be some underreporting of alcohol-related codes in mortality data to the extent that "sympathetic" certifiers of the cause of death may believe they are protecting an individual's reputation by recording a cause of death that does not directly implicate alcohol.
Individual studies have reported that a certain proportion of other diseases (e.g., certain cancers) and many accidents also are alcohol related (Ravenholt 1984; Roizen 1988). Researchers apply such estimates to overall cause-specific mortality to estimate additional alcohol-related mortality. The Public Health Service has proposed use of two new codes in the 10th Revision of the ICD for identifying alcohol involvement in mortality based on measures of blood alcohol content or other evidence of intoxication (Grant et al. 1987). These new codes should improve our ability to identify Alcohol-related mortality.
The National Highway Traffic Safety Administration operates FARS, which counts every death from a traffic crash occurring on public roads within 30 days of the accident and records alcohol-related deaths based upon blood alcohol measurement, issuance of a citation for driving under the influence, or the determination of the responding police officer (USDOT 1985). The lack of uniformity among jurisdictions in the use of blood alcohol testing results in some underreporting of alcohol involvement in FARS data, and there is no information on the race of decedents.
As in the case of mortality, studies of alcohol -related morbidity or illness depend upon the specification of diseases that are associated with alcohol abuse. Data for such studies come from NHDS, an ongoing NCHS survey of discharges from short-stay community hospitals that provides codes for conditions diagnosed during the hospital stay (NCHS 1985; Stinson 1989). The availability of multiple diagnostic codes on each record allows for the examination of comorbidity of alcohol-related and other diseases. Sample sizes are large (typically in excess of 200,000 records per year) but not always large enough for reliable estimates when examining specific disease codes or when studying smaller age or race groups.
Even with the limitations described above, studies in alcohol epidemiology give us the data we need to assess the problems resulting from alcohol abuse. Such information is useful in estimating the societal costs of various health problems (Harwood et al. 1985), often providing the basis for justifying program development and implementation. Surveillance, an ongoing form of descriptive epidemiology characterized primarily by uniformity of methods for repeated observations over time, is conducted to monitor change and provides one means of evaluating the effectiveness of treatment and prevention programs (Stinson et al. 1987; Brooks et al. 1989; Grant &. Zobeck 1989; Zobeck et al. 1989). In addition, through the identification of nonrandom variations in the distribution of alcohol-related problems. epidemiologic research can generate hypotheses for testing in analytical studies.
Alcohol Use and Abuse: Where Do the Numbers Come From?--A Commentary
by NIAAA Director Enoch Gordis, M.D.
There are many different and sometimes conflicting statistics on the nature and extent of alcohol-related problems. Consequently, many alcohol professionals and the lay public are skeptical about the legitimacy of alcohol-related data, and often ask "How do you know what you know about alcohol-related problems?". We hope that the general description of alcohol epidemiological research in this issue of Alcohol Alert sheds some light on this question and helps alcohol program personnel. policymakers, and others to interpret better the many alcohol related facts they see and hear.
The good news is that alcohol-related data are getting better and better. Refined data collection methodology and analysis and better reporting continue to increase information validity and reliability. Alcohol epidemiological research also has raised important questions for future alcohol research--questions that will help us to better prevent and treat alcohol abuse, alcoholism, and related consequences. For example, epidemiological research findings show that only a minority of alcoholics develop cirrhosis. Thus, a need has been identified for research on risk factors other than alcohol (e.g., genetic, nutritional, viral) that may be involved in cirrhosis development.
BRENNER, B. Estimating the prevalence of alcoholism from vital rates. Quarterly Journal of Studies on Alcohol 21:140-141, 1960. * BROOKS, S.; Williams, G.: Stinson, F.; and Noble, J. Surveillance Report #13: Apparent Per Capita Alcohol Consumption: National, State and Regional Trends. 1977-1987. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism, Division of Biometry and Epidemiology, Alcohol Epidemiologic Data System, Sept. 1989. * CAHALAN, D. & Room. R. Problem Drinking Among American Men. Monograph No. 7. New Brunswick, NJ: Rutgers Center for Alcohol Studies, 1974. * CAHALAN, D. Problem Drinkers. San Francisco: Jossey-Bass, 1970. * CAHALAN, D.; Cisin, l.; and Crossley, H. American Drinking Practices. Monograph No. 6. New Brunswick, NJ: Rutgers Center for Alcohol Studies, 1969. * DELINT, J. & Schmidt, W. The distribution of alcohol consumption in Ontario. Quarterl y Journal of Studies on Alcohol 29:968-973, 1968. * EDWARDS, G. & Gross, M.M. Alcohol dependence: Provisional description of a clinical syndrome. British Medical Journal 1:1058-1061, 1976. * GRANT, B. DSM III-R and ICD- 10 classifications of alcohol use disorders and associated disabilities: A structural analysis. International Review of Psychiatry 1:21-39, 1989. * GRANT. B.F.; Dufour, M.; Stinson, F.; Towle, L.; and Bertolucci, D. Epidemiologic Bulletin No. 17: Proposed coding of alcohol's role in casualties. Alcohol Health & Research World 12(1):48-50, 1987. * GRANT, B. & Zobeck, T. Surveillance Report #11: Liver Cirrhosis Mortality in the United States, 1972-86. Rockville. MD: National Institute on Alcohol Abuse and Alcoholism, Division of Biometry and Epidemiology, Alcohol Epidemiologic Data System, June 1989. * HARFORD, T.C; Zobeck, T.S.; Grant, B.F.; Stinson, F.S.; Aitken. S.S.; Dufour, M.; and Stoil, M.J. Epidemiologic Bulletin No. 19: The alcohol surveillance program of the Division of Biometry and Epidemiology: History, growth, and future directions. Alcohol Health & Research World 12(4):309-313, 1988. * HARWOOD, H.; Kristiansen, P.; and Rachal, J. Social and economic costs of alcohol abuse and alcoholism. Issue Report No. 2. Research Triangle Park, NC: Research Triangle Institute, 1985. * HILTON, M. Drinking patterns and drinking problems in 1984: Results from a general population survey. Alcoholism: Clinical and Experimental Research 11:167-175, 1987. * HILTON, M. A comparison of a prospective diary and two summary recall techniques for recording alcohol consumption. British Journal of Addiction 84:1085- 1092. 1989. * LEDERMANN, S. Alcool, Alcoolism, Alcoolisation; Donnes Scientifiques de Caractere Physiologique, Economique et Social. (Institut National d'Etudes Demographiques, Travaux et Documents, Cahier No. 29.) Paris: Presses Universitaires, 1956. * MACMAHON, B. & Pugh. T. Epidemiologic Methods. Boston: Little, Brown, 1960. * MALIN, H.; Wilson, R.; Williams, G.; and Aitken, S. Epidemiologic Bulletin 10: 1983 Alcohol Health Practices Supplement. Alcohol Health & Research World 10(2):48-50 1986. * MAUSNER, J. & Bahn, A. Epidemiology: An Introductory Text Philadelphia: W.B. Saunders, 1974. * MOORE, R.; Bone. L.; Geller, G.; Mamon, J.; Stokes, E.; and Levine, D. Prevalence, detection, and treatment of alcoholism in hospitalized patients. Journal of the American Medical Association 261 :403-407, 1989. * National Center for Health Statistics. Utilization of short-stay hospitals, United States. 1983, annual summary. Vital and Health Statistics. Series 13. No. 83. DHHS Pub. No. (PHS)85-1744. Public Health Service. Washington, DC U.S. Govt. Print. Off., 1985. * National Center for Health Statistics. Advance Report of Final Mortality Statistics. 1986. (Monthly Vital Statistics Report, Vol. 37, No. 6) DHHS Pub. No. (PHS)88-1120. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1988. * RAVENHOLT, R. Addiction mortality in the United States. 1980: Tobacco, alcohol, and other substances. Population and Development Review 10:697-724. 1984. * REGIER, D.; Boyd. J.; Burke, J.; Rae, D.: Myers, J.; Kramer, M.; Robins, L.; George. L.; Karno, M.: and Locke, B. One-month prevalence of mental disorders in the United States. Archives of General Psychiatry 45:977-986. 1988. * ROIZEN, J. Alcohol and trauma. In: Giesbrecht, N.; Gonzales, R.; Grant, M.; Osterberg, E.; Room, R.; Rootman. I.; and Towle, L., eds. Drinking and Casualties: Accidents.,Poisonings, and Violence in an International Perspective. London: Routledge, 1988. * SCHMIDT, W. & deLint, J. Estimating the prevalence of alcoholism from alcohol consumption and mortality data. Quarterly Journal of Studies on Alcohol 31 :957-964, 1970. * SMART, R. ls the postwar drink ing binge ending? Cross-national trends in per capita alcohol consumption. British Journal of Addiction 84:743-748, 1989. * STINSON, F. U.S. Epidemiologic Data Reference Manual, Vol. 4: Hospital Discharges with Alcohol-Related Conditions. National Institute on Alcohol Abuse and Alcoholism. Washington, DC: U.S. Govt. Print. Off., 1989. * STINSON, F. & Williams, G. Surveillance Report #4: Trends in Alcohol-Related Morbidity Among Short-Stay Community Hospital Discharges, 1979-84. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism, Division of Biometry and Epidemiology, Alcohol Epidemiologic Data System, Jan. 1987. * U.S. Department of Transportation. Fatal Accident Reporting System 1983. Washington, DC: the Department, 1985. * WILLIAMS, G.; Aitken. S.; and Malin, H. Reliability of self-reported alcohol consumption in a general population survey. Journal of Studies on Alcohol 46:223-227, 1985. * WILLIAMS, G.; Dufour, M., and Bertolucci, D. Drinking levels, knowledge, and associated characteristics, 1985 NHIS findings. Public Health Reports 101:593-598. 1986. * WILLIAMS, G.D.; Stinson, F.S.; Parker, D.A.; Harford, T.C.; and Noble, J. Epidemioiogic Bulletin No. 15: Demographic trends. alcohol abuse and alcoholism, 1985-1995. Alcohol Health & Research World 11 (3):80-83, 1987. * WILLIAMS, G.; Grant, B.; Harford, T.; and Noble, J. Population projections using DSM-III criteria: Alcohol abuse and dependence, 1990-2000. Alcohol Health & Research World. in press. * ZOBECK, T.; Grant, B.; Williams, G.; and Bertolucci, D. Surveillance report #12: Trends in Alcohol-Related Fatal Traffic Crashes, United States: 1977-1987. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism, Division of Biometry and Epidemiology, Alcohol Epidemiologic Data System, Aug. 1989.
All material contained in the Alcohol Alert is in the public domain and may be used or reproduced without permission from NIAAA. Citation of the source is appreciated.
Copies of the Alcohol Alert are available free of charge from the Scientific Communications Branch, Office of Scientific Affairs, NIAAA, Willco Building, Suite 409, 6000 Executive Boulevard, Bethesda, MD 20892-7003. Telephone: 301-443-3860.
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES
Public Health Service * National Institutes of Health
Updated: October 2000