![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD 20892-7003
 |
ABSTRACT |
|---|
 TOP ABSTRACT THE EVOLUTION OF THE...METHODS OF ASSESSMENTSTRATEGIES FOR IMPROVEMENTREFERENCES |
|---|
KEY WORDS: • alcohol • moderate drinking • alcohol epidemiology • alcohol consumption • self-report • survey measurement
|
THE EVOLUTION OF THE ALCOHOL GUIDELINE |
|---|
|
TOPABSTRACTTHE EVOLUTION OF THE...METHODS OF ASSESSMENTSTRATEGIES FOR IMPROVEMENTREFERENCES |
|---|
The 1985 alcohol guidance
Although the phrasing of the alcohol guideline itself has changed
only slightly over the years, the accompanying explanatory text has
evolved substantially. The issue of drinking and driving appears for
the first time in the 1985 edition of the Dietary Guidelines
(1)
. Following the introductory material, the seven
guidelines were listed with the alcohol guideline reading "If you
drink alcoholic beverages, do so in moderation—and don’t drive." On
page 4 in the section on adequate diet, alcohol was added to a
discussion of the dietary needs of elderly people. "Elderly people
may eat relatively little food. Thus, they should pay special attention
to minimizing foods high in calories and low in essential
nutrients—for example, fats, oils, sugars, and alcohol. The latter is
not usually regarded as a food, but is high in calories." The
narrative text elaborating on the alcohol guideline was quite short
compared with those of the other guidelines and focused primarily on
energy and nutritional aspects of alcohol consumption. The issue of
alcohol consumption during pregnancy was discussed in a cautionary
tone. "Excessive consumption of alcoholic beverages by pregnant women
may cause birth defects or other problems during pregnancy. It has not
been established whether there are risks to the fetus from minimum
consumption of alcohol. Until this is known, pregnant women should
abstain or limit alcohol intake to an occasional standard-size
drink of beer, wine or liquor (not more than one drink per day)." By
today’s standards, this paragraph was vague and perhaps not very
informative. "Excessive consumption" was not defined (unless the
subsequent injunction to limit intake to an occasional
standard-size drink is considered definitive). The types of birth
defects and pregnancy problems were not specified. Nevertheless, the
1985 Dietary Guideline Committee took a bold public health step by
acknowledging that although a safe threshold for alcohol consumption
during pregnancy had not been established, pregnant women should either
abstain from alcohol or limit their consumption to an occasional drink
but no more than one drink a day. Although "moderate drinking" and
a "standard-size drink" were not defined, the guideline did begin
to quantitate moderate drinking by stating that "one or two
standard-size drinks daily appear to cause no harm in normal
healthy adults." The guideline concluded by stating "If you drink,
be moderate in your intake and DO NOT DRIVE!"
The 1990 alcohol guidance
The Report of the Dietary Guidelines Advisory Committee on the
Dietary Guidelines for Americans, 1990 was considerably longer and more
detailed than that of the 1985 Committee (2)
. One of the
eight major text changes highlighted in the Executive Summary states:
"Moderate drinking is defined and certain groups, including women who
are pregnant or trying to conceive, are cautioned not to drink at
all." In the introductory text of the Guidelines, it was noted that
"Other guidelines suggest moderation in the use of sugars, salt, and
if used at all, alcoholic beverages." Alcohol was mentioned as a
source of empty calories for inactive individuals and those trying to
lose weight. In the text box "To Decrease Calorie Intake" one of
the tips listed was "Drink little or no alcoholic beverages." The
Committee expressed concern that more space was needed for new material
for several guidelines including the alcohol guideline. Indeed, the
text of the alcohol guideline is expanded and considerably more
specific than previously. In terms of the narrative, the Committee
specified that the new material must achieve the following:
1) make clear that alcohol has no net benefit to health;
2) define moderate drinking; 3) identify people
who should not drink; 4) add inflammation of the pancreas
and damage to heart and brain as complications of drinking; and
5) refer to studies that link alcohol to lower risk for
heart attacks and offsetting higher risk for hypertension and
hemorrhagic stroke. Moderate drinking was defined as no more than one
drink a day for women and no more than two drinks a day for men. A
drink was defined as 12 oz of regular beer, 5 oz of wine or 1.5 oz of
distilled spirits (80 proof). The 1990 alcohol guideline clearly stated
for the first time that "some people should not drink
alcoholic beverages." The statement about alcohol and pregnancy was
strengthened, i.e., women who are pregnant or trying to conceive was
the first group listed among those who should not drink. Others who
should not drink included "Individuals who plan to drive or engage in
other activities that require attention or skill; individuals using
medicines, even over-the-counter kinds; individuals who cannot keep
their drinking moderate; and children and adolescents."
The wording of the alcohol guideline was discussed at great length by the Committee. Some members favored a stronger caution such as: "Alcoholic beverages are not recommended. If you drink at all, do so in moderation." Reasons for this recommendation were that drinking alcoholic beverages had no net benefit to health and could be addictive. In addition, at the time, it was understood that alcohol was a significant contributor to three of the top 10 causes of death, i.e., cirrhosis of the liver, motor vehicle and other "accidents" and suicide. Although all committee members agreed that excess alcohol consumption had serious consequences, the majority of members felt that this stronger alternative might be interpreted as overly judgmental and restrictive. In addition, such a change in wording would have required considerable additional space on the cover of the bulletin, thus drawing undue attention to this single guideline. Limiting ones drinking to moderate levels, if one chooses to drink, was felt to be consistent with recommendations found in both The Surgeon General’s Report on Nutrition and Health of 1988 and the 1989 National Research Council’s report Diet and Health: Implications for Chronic Disease Risk.
The 1990 edition of the alcohol guideline set several precedents. Not only was moderate drinking defined for the first time, but gender differences in the distribution and effect of alcohol were also recognized. By 1990, a substantial body of scientific evidence had demonstrated that the risk of fetal abnormalities increased with an increase in alcohol consumption during pregnancy. Although a threshold level of safety for alcohol consumption had not yet been established, the Committee recommended that pregnant women or those trying to conceive should avoid drinking during pregnancy. Early research was beginning to suggest that alcohol consumption by mothers who were breast-feeding might have a negative effect on their infants, but the Committee did not consider this evidence strong enough to warrant advice that mothers should abstain from drinking while breast-feeding. A caution about the slow elimination of alcohol from the blood even after moderate drinking was also added, including a concrete time frame ("Most people retain some alcohol in the blood 3 to 5 h after even moderate drinking.") Alcohol was not recommended as a means of preventing coronary heart disease because the Committee felt that it had not been shown to be effective and because of the many potentially harmful effects of alcohol consumption, including higher risk for hypertension and hemorrhagic stroke.
The 1995 alcohol guidance
The introductory text of the 1995 edition of the Dietary
Guidelines included the statement that "Alcohol, although not a
nutrient, also supplies energy—about 7 calories per gram"
(3)
. It also mentioned that "fat and alcohol are high in
calories." In the discussion of location of body fat, the following
statement appeared: "Smoking and too much alcohol increase abdominal
fat and the risk of diseases related to obesity." The text box
containing tips for decreasing energy intake included the item "Drink
less or no alcohol." The 1995 Committee spent considerable time
discussing the fact that the alcohol guideline was unique among the
Dietary Guidelines because the substance referred to was both a food
that is an energy-rich beverage and a drug, which is subject to
abuse and can cause dependence. The Committee considered whether
alcohol-related consequences including abuse and dependence
represented more of a public health issue than a dietary issue and thus
whether alcohol would be better dealt with elsewhere than in the
Dietary Guidelines. Nevertheless, they also agreed that alcoholic
beverages were a part of the diet as well and, when used in moderation,
were safe and pleasurable. The Committee was also reluctant to remove
the alcohol guideline because members were concerned that the absence
of an alcohol guideline after its presence in the three previous
editions of the Guidelines would send a confusing message to the
public. Therefore the Committee ultimately decided to retain the
alcohol guideline with the same heading as in the 1990 version. Changes
in the body of the text of the alcohol guideline were not substantial.
The introductory paragraph of the guideline referred specifically to the physiologic or drug effects of alcohol, including the capacity to alter judgment. A statement that "Alcoholic beverages have been used to enhance the enjoyment of meals by many societies throughout human history" was also added to the opening paragraph. The definition of moderation was identical to that of 1990, but the box defining moderation appeared earlier in the text. In the same box, a statement was also added referring to the calories in alcoholic beverages (also included in the discussion of the weight guideline) and their possible contribution to weight gain. The list of problems associated with heavy drinking was expanded to include high blood pressure, stroke, heart disease, certain cancers, accidents and violence. The 1990 Committee had been concerned that the beneficial effects of moderate alcohol consumption on heart attack were offset by higher risk for hypertension and hemorrhagic stroke. The 1995 Committee concluded that on the strength of further studies on moderate drinkers in which neutral or beneficial effects on risk of stroke or cerebrovascular disease had been recorded, such concerns were no longer warranted.
The list of those who should not drink was similar to that in the 1990 edition but was rearranged to place children and adolescents at the top of the list. In addition, the Committee intended the term adolescents to include all individuals younger than the legal drinking age of 21 y, although this was not specifically stated. Despite the publication of some studies showing no apparent detrimental effects of low or moderate alcohol consumption during pregnancy, because an absolutely safe level of alcohol intake during early pregnancy had not been established, the Committee decided to continue the prudent advice of the 1990 alcohol guideline that women who are pregnant or trying to conceive should not drink. The Committee also acknowledged that including women trying to conceive was done primarily to ensure that women who might already be in the first few weeks after conception and not yet realize that they are pregnant would also be included; however, this was not explicitly stated in the guideline. The 1995 Committee, like that of 1990, felt that evidence was still insufficient to make a recommendation on alcohol consumption during lactation. In the advice for today section, the text was expanded to stress the food use of alcoholic beverages. The Committee felt that including the mention of drinking with meals recognized that consumption of alcohol with food slows consumption and absorption of alcohol, although this is not explicitly stated. It was also noted in this section that alcohol should not be consumed in situations in which others might be put at risk.
The 2000 alcohol guidance
The 2000 Committee endorsed maintaining the wording that appeared
in the 1995 alcohol guideline (4)
. Revisions in the text
of the guideline were primarily to place more emphasis on adverse
effects of excess consumption and to update statements based on more
current scientific evidence. The committee proposed that the
introductory text begin with a paragraph that documents adverse effects
of excess alcohol intake and mentions no beneficial effects. The second
paragraph specifies the age groups for whom moderate drinking may
reduce risk of coronary heart disease ("mainly men over age 45 and
women over age 55"). This greater specificity stemmed from age- and
gender-specific rates of coronary heart disease as well as
age-specific relative risks related to moderate alcohol consumption
obtained from extensive prospective data based on three very large
cohorts of men and women in the U.S. (5
6
7)
. In the third
paragraph, it is clearly stated that moderate consumption provides
little, if any, health benefit for younger people. The Committee also
suggested adding the caution that among young people, risk of alcohol
abuse increases with decreasing age of initiation of drinking.
In the 1995 guideline, the list of health risks associated with alcohol consumption included reference to "certain cancers" without specifying the types of cancer. Based on intervening scientific evidence, for the 2000 guideline, the Committee recommended adding breast cancer ("Even one drink per day can slightly raise the risk of breast cancer") because members felt that although the reported level of risk was quite small, the issue was of serious concern to some women.
As mentioned earlier, the opening paragraph of the 1995 guideline included the statement "Alcoholic beverages have been used to enhance the enjoyment of meals by many societies throughout human history." The Committee recommended that this sentence be deleted because, although the statement was factually correct, a similar statement could have been made for many other foods and nutrients such as sugar or salt. Although the Committee recognized the intent of the earlier Committee to place moderate alcohol consumption in the context of a healthy diet, they felt that such a statement was inconsistent with the text in the other guidelines and, therefore, best omitted.
A sentence was added to the "What is Drinking in Moderation?" text box to make it clear that the different limits for men and women are based on both metabolism and body size/body water. The text of the section on "Who Should Not Drink?" was expanded to be more explicit about the categories of individuals who should not drink.
In particular, the bullet pertaining to pregnancy was strengthened. The phrase "women who are trying to conceive" was changed to "women who may become pregnant" in recognition of the many completed pregnancies in the U.S. that are unplanned. Acknowledging that fetal alcohol syndrome is a well-recognized clinical entity, the text for the 2000 edition states "Major birth defects, including fetal alcohol syndrome, can be caused by heavy drinking by the pregnant mother. Other fetal alcohol effects may occur at lower levels." This replaced the weaker text of the 1995 edition which stated that "birth defects have been attributed to heavy drinking." The statement "While there is no conclusive evidence that an occasional drink is harmful to the fetus or the pregnant woman" was also dropped. Although the text is scientifically accurate, it was felt to weaken the message, especially because, according to survey data, young women tend to underestimate the adverse effects of drinking during pregnancy. The phrase "including the first few weeks" was added to the statement "A safe level of alcohol intake has not been established for women at any time during pregnancy, including the first few weeks " to make it clearer that adverse effects may begin before the woman knows she is pregnant.
The phrase "operate machinery" was added as a second example of activities requiring attention or skill. In discussing the issue of alcohol-medication interactions, the committee felt that the text of the 1995 guideline was too broad-ranging and not supported by scientific evidence; they therefore recommended that the bullet covering medications be changed to be less restrictive. The bullet was modified to suggest that many medications can interact with alcohol and to advise that individuals taking medications should ask their health care professional for guidance. Finally, in the advice for today section, the text was expanded to further specify the meaning of moderation and to reinforce the importance of not drinking before or when driving.
With regard to the text of the 2000 Dietary Guidelines outside the alcohol guideline, alcohol appeared far less often than in the 1995 edition. Except for the first page of text where all the guidelines were outlined, alcohol appeared once in "Box 22: Steps That May Help Keep Blood Pressure In A Healthy Range" where it was stated "If you drink alcoholic beverages, do so in moderation. Excessive alcohol consumption has been associated with high blood pressure."
|
METHODS OF ASSESSMENT |
|---|
|
TOPABSTRACTTHE EVOLUTION OF THE...METHODS OF ASSESSMENTSTRATEGIES FOR IMPROVEMENTREFERENCES |
|---|
Epidemiology has its roots in the study of infectious diseases. Such
research first used the public health model, which involves defining a
disease agent, host and environment, to study and control epidemics.
During the early and mid-20th century, medical advances and
improvements in sanitation led to the mastery of many infectious
diseases, such as tuberculosis. Soon thereafter, cancer and heart
disease gained recognition as common killers. Thus a new discipline
emerged—chronic disease epidemiology. In addition the focus of
epidemiology shifted to assessing the general health of communities
under normal conditions rather than only under specific epidemic
circumstances. Alcohol epidemiology as a unique discipline is a
relative newcomer to the field, and it marks a logical progression in
the evolution of the science of epidemiology (8)
. Classic
infectious disease epidemiology is relatively straightforward, having a
specific, well-defined, easily identifiable etiologic agent (e.g.,
a bacterium or virus), a clear-cut case identification, generally a
short clinical course and a quickly effective prevention (vaccination)
or treatment (antibiotics). The epidemiology of chronic diseases such
as cancer and heart disease is more complex. Alcohol epidemiology,
however, is even more intricate and challenging because of the
multiplicity of contributing factors. In addition, it encompasses a
broad domain, i.e., alcohol use, abuse and dependence as well as
countless medical, psychological, social, legal and economic
consequences. One of the basic tenets in epidemiology is that in order
to count something accurately, one must first be able to define it. The
large number of varying definitions of terms such as "abstainer,"
"a drink," "moderate drinking" or "heavy drinking" in
alcohol research further adds to the complexity of interpreting its
epidemiologic findings. This lack of uniform, standard definitions
stems in large part from the diverse origins of the discipline, the
range of alcohol-related conditions and consequences, and the
necessary diversity of purposes and methods required (9)
.
Alcohol epidemiology has developed as a conglomeration of four major
epidemiological perspectives, i.e., psychosocial epidemiology,
psychiatric epidemiology, chronic disease epidemiology and
epidemiologic sociology (10)
. Research conducted
using each perspective seeks slightly different information from study
participants and defines measures and concepts differently. Although
this diversity is an advantage in that the four disciplines complement
each other in revealing drinking patterns and problems among the U.S.
population, it is also a serious handicap in that information collected
about alcohol consumption often is not comparable across studies. For
example, one survey may ask questions in a way that permits the
derivation of a diagnosis of alcohol dependence. Another study,
however, may ask questions about alcohol consumption and alcohol
problems without including specific diagnostic criteria, and thus a
diagnosis cannot be made. A study’s proposed research goals dictate
the particular measurement approach. Furthermore, techniques that are
ideally suited for one population subgroup may not work equally well
for other subgroups (e.g., teenagers vs. senior citizens).
Consequently, scientists conducting large and expensive
population-based surveys to answer multiple research questions and
elucidate drinking behavior in diverse population subgroups must weigh
many factors in deciding which alcohol consumption measures will best
meet their needs.
Over the years, the focus of alcohol epidemiology also has changed.
When alcohol-use surveys of the general population were first begun
in the United States, they focused mainly on the distinction between
drinkers and nondrinkers (11)
. For example, in 1939, the
Gallup surveys, which probably were the first surveys to measure
drinking on a national level in the United States, included the
following question about alcohol consumption: "Do you have occasion
to use alcoholic beverages such as liquor, wine, or beer or are you a
total abstainer?" (11)
. As alcohol survey research
progressed and investigators became interested in assessing the
consequences of various levels and patterns of alcohol consumption,
they found it necessary to develop methods to quantify consumption more
accurately. Early on, volume of consumption was often the sole measure
of alcohol intake. In recent years, with greater interest in better
assessment of such alcohol-related consequences as motor vehicle
crashes, there has been a corresponding increase in interest in
assessing drinking patterns as well. This obviously has bearing on our
examination of how American drinking compares with the alcohol dietary
guideline. Two men may both report having 14 drinks a week. If we
average this volume over a week it appears that both men consume an
average of two drinks a day, in keeping with the recommendation of the
guideline. If drinking pattern is assessed as well and one man reports
having two drinks every evening and others reports 14 drinks on Friday
night or two occasions of seven drinks or four occasions of three or
four drinks, for example, we now see that only one of these individuals
is actually following the guideline recommendation. Over the 50 years
that alcohol researchers in the U.S. and abroad have conducted surveys
of alcohol consumption and alcohol-related problems in
representative population samples, investigators have made great
progress in survey research methodology, including the quantification
of drinking levels, despite the previously mentioned definitional
differences. To address the issues surrounding how best to measure and
model alcohol consumption, the National Institute on Alcohol Abuse and
Alcoholism convened the International Workshop on Consumption Measures
and Models for Use in Policy Development and Evaluation which was held
in May, 1997 (12)
. This meeting provided a valuable
overview of the state-of-the-science as well as identifying future
research opportunities. Two additional such efforts include the
International Guidelines for Monitoring Alcohol Consumption: Report to
the World Health Organization Program on Substance Abuse
(13)
and the recent consensus conference "Measuring
Drinking Patterns, Problems, and Their Connection: An International
Research Conference" held recently in Sweden (14)
.
Extensive research is also currently underway to identify a biochemical marker or develop a laboratory test that will provide an accurate, reliable and objective chemical indicator of quantity and duration of alcohol consumption, but no single test meets these criteria to date. Thus, in the vast majority of epidemiologic surveys, information about alcohol consumption has been based on self-report. Why is self-report of alcohol consumption any more of a concern than that of carrot or rye bread consumption? Just how accurately and reliably is alcohol consumption reported?
One reason why self-report of alcohol consumption is sometimes
considered suspect is that one of the hallmarks of alcohol dependence
is denial. Alcoholic individuals often deny that they are having
problems or that these problems stem from alcohol; therefore, they may
lie or grossly underestimate the reported quantity and frequency of
their alcohol consumption. A second reason why self-report of
alcohol consumption may be called into question is the discrepancy
between alcoholic beverage sales data and survey reports of alcohol
consumption. In general, comparison studies in the alcohol literature
(15)
have shown that self-reported alcohol consumption
accounts for only 40–60% of alcoholic beverages sold as measured by
sales and tax data. Several factors, when taken into consideration,
account for part of this discrepancy. For example, alcohol used in
cooking and losses of alcohol due to bottle breakage are not deducted
from sales totals. An assumption is also made that alcoholic beverages
purchased in a given year will be consumed in that same year, which may
or may not be the case.
Furthermore most general population surveys that measure alcohol
consumption use household sampling frames and thus do not include many
of the heaviest drinking segments of the population. For example, in a
large county study, Weisner and colleagues (16)
estimated
the prevalence of problem drinking in samples drawn from the
nonhousehold population (individuals reporting that they were homeless
or were living in rooming houses, hotels or prisons, for example) to be
49% compared with 11% in the general population of the county as
defined by households. It should be noted, however, that because these
groups are relatively small compared with the general population, the
bias introduced into general population estimates by excluding such
individuals is also relatively small.
On the other hand, sales data also do not include untaxed alcohol such
as alcohol purchased in duty-free shops and homemade alcohol
whether produced legally or otherwise. Legal home production of wine
and beer is potentially a significant source of alcoholic beverages
that will not be captured by sales and taxation data. Cider is thought
to be another source of unregistered alcohol consumption in some parts
of the United States. In addition, the increasingly popular home
beer-brewing kits provide a further source of beer that will not be
represented in alcohol sales data (17)
.
Nevertheless, taking all of these factors into account, some degree of undercoverage of survey data on alcohol consumption with respect to sales data remains, suggesting downward recall bias and perhaps other response and measurement biases as well. Do some groups of drinkers, defined either by demographics or drinking level, bias their reports differently? Do heavy drinkers deny, forget or otherwise underreport drinking amounts and/or occasions more than do lighter drinkers?
There is mixed empirical evidence regarding differential under- or
overreporting. Despite the previously mentioned phenomenon of denial,
Polich and colleagues (18)
reported that alcoholics in
clinical services research seldom have been found to diminish their
self-reported drinking greatly compared with measurements of blood
alcohol concentrations. Alcoholics in treatment, however, may be more
likely to report drinking accurately compared with heavy drinkers in
general. Compared with diary-based reporting, using summary
measures based on graduated frequencies, Hilton (19)
reported some underreporting of the number of occasions on which larger
numbers of drinks were consumed and a compensatory overreporting of
occasions on which only 1 or 2 drinks were consumed.
Underrepresentation of heavier amounts also was reported by Perrine and
colleagues (20)
, comparing daily telephone report of
consumption with retrospective quantity-frequency self-report. Thus
it appears that heavier drinkers do, in fact, underreport their
drinking more than light drinkers. In addition, some questions ask only
about usual and not atypical heavy drinking, or gather frequency of
consuming at some level such as "5 or more drinks on an occasion."
The latter may result in imputational problems because the threshold
represents a lower number than actually consumed. In other words, an
individual who drinks 8 drinks on these days will select the "5 or
more" category. Researchers may then elect to count this as "5
drinks" when they are calculating index measures because they do not
know where in the spectrum of "or more" the drinker falls.
Just exactly what is the magnitude of underreporting by heavier
drinkers? What is the distribution of alcohol consumption across the
American population? Greenfield and Rogers (21)
recently
investigated this issue. Using a dataset consisting of four large
pooled surveys having a common design and using the graduated
frequencies approach (explained in greater detail below), they found
that 67.3% of the U.S. population were current drinkers (defined as
respondents who reported at least one drink in the 12 mo before
survey). Among these current drinkers, the top 2.5% of drinkers
accounted for over one fourth (27%) of adult drinkers’ total reported
alcohol consumption. The top 15% of drinkers accounted for nearly
three fourths (72.6%) of the reported alcohol consumption, whereas the
top 20% of drinkers accounted for 
80% (80.6%). These results were
confirmed using a separate national survey conducted the same year.
Thus relatively small biases by heavy drinkers may have a large effect
on population consumption estimates.
Alcohol abuse, dependence and alcohol-related consequences such as
alcohol-related motor vehicle crashes are major public health
problems. To better address these problems as well as to address the
issue of accurate reporting of alcohol consumption, survey researchers
have made major advances in alcohol epidemiology. It is now possible to
classify individuals as alcohol abusers and/or alcohol dependent
according to strict established criteria in general population surveys
employing lay interviewers. Because researchers believed early on that
most alcohol abuse, dependence and other consequences generally
occurred at high levels of consumption, because of the highly skewed
distribution of alcohol consumption and the above-mentioned issues
about underreporting of heavy drinking, it makes sense that more
attention has been given to fine-tuning quantification on the high
end of the drinking spectrum. As research has progressed, it has been
established that lower risk drinkers, because of their sheer numbers,
actually account for a significant number of alcohol-related
consequences (22
,23)
. Clearly, patterns of drinking are as
important as volume of consumption. Analyses from the 1992 National
Longitudinal Alcohol Epidemiologic Survey
(NLAES;2described in detail in the Surveillance section) disclose that just
over half of all individuals in this sample who are alcohol dependent
had average daily volumes of ethanol intake <1 oz of
ethanol (two drinks), that is, within the alcohol guideline for male
drinkers (23)
. On the other hand, only 6% of dependent
drinkers never drank 5+ drinks. The effects of drinking pattern have
even been found to affect all-cause mortality rates
(24)
.
The effectiveness of a survey questionnaire in accurately measuring
alcoholic beverage consumption is influenced by the way in which
questions are phrased, the order in which the questions are asked and
the way in which the answers to the questions are compiled into a
summary measure. The types of survey questionnaires most commonly used
fall into the following five categories: frequency measures;
quantity-frequency (QF) measures; graduated frequency measures;
short-term recall methods; and diary methods (25)
. All
of these measures are best suited to measuring consumption over the
past year or some more recent period. When the focus is on consumption
over a longer period, another technique for assessing alcohol
consumption is the time line followback (TLFB) method
(26)
. The TLFB is a structured interview in which
participants receive calendar-based memory cues to assist them in
constructing a chronological report of their alcohol consumption.
Although the procedure is widely employed in research on the efficacy
of alcohol treatment, the required interviews are highly individualized
and thus extremely impractical for large-scale population-based
surveys. Other more efficient lifetime measures suitable for use in
surveys are now becoming available (27)
.
As the term implies, the most basic frequency measure captures only
frequency of drinking. Because the amount of alcohol consumed on each
drinking occasion is not assessed, it is not possible to calculate
total volume of consumption. The semiquantitative food-frequency
measure, frequently used in diet and nutrition studies, often employs a
single-frequency item including ordinal number of times drinking in
a day at the high end of the scale (25)
. This method of
measurement also has limitations including inability to assess drinking
patterns. QF measures ascertain both quantity and frequency of
consumption. At present, QF measures are probably the most widely used
instruments with which to measure drinking in most countries, including
the United States. In some QF surveys, respondents are asked how often
and how much they usually consumed different types of alcoholic
beverages over the past year. Although overall quite effective, one
disadvantage of this method is that respondents, especially those
having irregular drinking habits, may have trouble providing accurate
answers because they must mentally average their alcohol consumption
over the entire year (25)
.
To overcome this mental averaging problem for people having varying
quantities of intake, researchers have developed graduated frequency
measures, which begin with a question eliciting the largest number of
drinks consumed by the respondent on any one drinking occasion in the
past year. Subsequent questions then ask about the number of occasions
on which progressively lower quantities of alcohol were consumed. The
benefit of this approach is that the reports of alcohol consumption are
highly accurate because, according to some researchers, not as much
recall and mental arithmetic are required. A degree of mental
calculation is still required, nevertheless, because respondents must
convert their usual size drink into a "standard-size" drink. In
addition, this approach captures volume due to hazardous drinking
better than QF and short-term measures discussed next
(28)
. Other researchers feel that it is cognitively very
demanding to figure out how often one has consumed a given number of
standard drinks exclusive of other amounts, but that the demand
characteristics of the graduated QF favor the reporting of larger
quantities of alcohol in a way that may overcome the tendency for
people to underestimate heavier drinking. This tendency plus the
opportunity to report intake in more detail would improve accuracy. The
main disadvantage, however, is that the greatly increased length of the
questionnaire requires more time for respondents to answer which, in
turn, not only increases research costs but decreases time available to
ask about other items of interest (25)
.
Short-term recall methods ask about actual alcohol consumption over a
short period of time, such as 1 or 2 wk. This approach is based on the
assumption that respondents are more likely to remember the actual
amounts they drank over a short time period more accurately than over a
longer period such as a year. One drawback of this methodology is that
many infrequent or occasional drinkers may report no alcohol
consumption during the time studied. Consequently, short-term
recall measures may overestimate the number of abstainers relative to
other survey methods (25)
. In addition, such
short-term methods may well entirely miss intermittent heavy
drinking (28)
.
In diary methods, participants record each drink consumed over a given
time frame (e.g., 1 wk), ideally shortly after consumption. Researchers
have recently introduced an automated variation of the diary method. In
this approach, participants report their daily alcohol intake by
calling a dedicated toll-free number and activating, through a
touch-tone telephone, an automated, interactive
voice-simulation system (29)
.
Why are the differences among assessment instruments relevant to our discussion of the alcohol guideline? One reason is that, for a given drinker, different questions may elicit different responses and therefore lead to varying estimates of alcohol consumption for that person. Furthermore, even if a respondent provides identical answers, differences in the scientific assumptions and calculations associated with the particular survey method may produce variations in reported results.
For example, questionnaires using the graduated frequency approach
consistently produce higher estimates of volume of drinking than do
basic QF measures, especially among heavier drinkers. Survey
researchers have discovered that more questions (and consequently, more
answers) may lead to higher consumption estimates, which are generally
considered to be more accurate (28
,30)
. Whether this
assumption actually holds, however, is open to question, and further
research in this area is clearly warranted (14)
.
Diary methods produce higher estimates than do either QF or
short-term recall methods for frequent drinkers. For example, in
the previously mentioned study using an automated interactive telephone
reporting system (29)
, 50 volunteers reported their daily
alcohol intake for 112 consecutive days. Other data collected by
traditional means immediately after study completion demonstrated that
drinkers, particularly heavier drinkers, retrospectively underreported
their alcohol consumption.
More detailed and specific questions also elicit higher estimates of
alcohol consumption. For example, QF questions for different periods
within a given time frame produce higher estimates when summed than do
global QF questions. Similarly, beverage-specific questions or
questions asking about consumption in different contexts (e.g., at
home, in bars, or at parties and celebrations) produce higher estimates
than do global questions asking about total alcohol consumption
(25)
.
With the increasing focus on research on moderate drinking has come
greater interest in refining measurement of alcohol consumption across
the range of consumption levels and patterns. Studies in the alcohol
epidemiology literature examining the reliability and validity of
survey measures of alcohol consumption indicate high levels of
reliability. Validity estimates, although not uniformly as high, were
generally within the upper range of validity coefficients found in much
of psychometric research (31)
. Therefore, research
suggests that carefully crafted questions elicit reliable and valid
self-reported estimates of alcohol consumption (32)
.
In epidemiologic research on the role of alcohol intake as a risk factor for chronic disease, however, information about alcohol intake may be derived from data on dietary intake (i.e., nutritional epidemiology) rather than from a method designed specifically for the purpose of measuring alcohol intake (alcohol epidemiology). As mentioned earlier, as a part of the diet, alcoholic beverages also fall into the domain of nutritional epidemiology, a discipline that has evolved quite apart from that of alcohol epidemiology. The field of nutritional epidemiology has developed from interest in the concept that aspects of the diet may influence the occurrence of human disease. Early on, investigations focused primarily on deficiency states. More recently the focus has switched to conditions such as cancer and heart disease in which a whole host of factors including genetics, occupational factors, psychological factors, infectious factors and behavioral factors (smoking, drinking, level of physical activity) play a role in addition to that of diet. In addition to multiple potential determinants, alone or in combination, for most diseases, the relevant period of exposure is often not known. Long latent periods may separate exposure from manifestation of disease. The critical exposure may be a short exposure many years before diagnosis or the cumulative exposure over many years. In addition, these conditions may arise from excessive as well as insufficient intake of dietary factors.
Diet represents an unusually complex set of exposures that are strongly
intercorrelated. Such exposures are not usually present or absent, but
rather are continuous variables. Individuals rarely make clear changes
in diet at identifiable points in time; rather, these changes happen
gradually over many years. In addition, people are not generally aware
of the precise content of the foods they eat, making it necessary to
ascertain the composition of nutrients indirectly on the basis of the
reported use of foods or on the level of biochemical measures. Another
serious issue in nutritional epidemiology has been the lack of
practical methods to measure diet. Because these studies are large,
involving hundreds or thousands of participants, methods of dietary
assessment must not only be reasonably accurate but also inexpensive
and not extremely time consuming. Fortunately, estimates of total
alcohol intake from dietary interviews agree with estimates derived
from recorded or measured dietary intake better than do estimates of
most other nutrients (33)
. In multiple studies of the
reliability and validity of self-report of food intake in a wide
variety of populations ranging from highly educated health
professionals to ordinary subjects not selected on the basis of their
education or interest in diet, alcohol intake was found to be
particularly reproducible compared with other nutrients
(34
35
36)
.
Measures available to monitor progress
The following large national surveys collect some basic
information on alcohol consumption (37)
: the State
Behavioral Risk Factor Surveillance System, the National Health and
Nutrition Examination Survey series (NHANES), The National Health
Interview Survey series and the National Household Survey of Drug
Abuse. Two additional national survey series that focus exclusively on
youth are the Youth Behavior Surveillance System and the Monitoring the
Future Study (38)
. The National Household Survey of Drug
Abuse collects alcohol information on children beginning at age 12 y as well as adults. With the exception of NHANES, these surveys are
conducted annually. Information from these data sets could be used to
calculate average daily alcohol consumption. The proportion
of individuals whose average daily consumption exceeds levels
recommended in the alcohol guideline would provide a very crude
estimate of compliance because the measure derived would reflect
average rather than actual measures of daily consumption. For example,
a woman with an average daily consumption of one drink could drink one
each day and thus be within the recommendations of the alcohol
guideline or she could have three drinks on Tuesday, two on Wednesday
and two on Saturday. The latter pattern clearly does not follow the
guideline but still adds up to seven drinks a week with an average of 1
per day.
The National Alcohol Survey (NAS) includes consumption, pattern and
problem variables as well as attitudes, cognitive and drinking context
variables. The NAS is conducted approximately every 5 y by the
Alcohol Research Group and collects extensive data, which include
graduated quantity/frequency measures of total alcohol consumption as
well as individual beverage type, including beer, wine and spirits,
binge drinking, attempts to reduce drinking, attitudes/opinions on
drinking levels in different drinking situations, treatment status and
drinking consequences. Consumption and problem trend studies have been
a major contribution of the NAS series (39
40
41)
.
Although the NAS surveys have contributed significantly to alcohol epidemiology over the years, it is not possible to use this information to track compliance with the alcohol guideline in all cases. For respondents who are drinkers, the lowest category of drinks per occasion that can be selected is "1–2." Men who drink at this level clearly would be within the recommendations of the guideline. Thus we could track compliance among men with this data set. Because one drink per occasion is not offered as a choice, compliance among women cannot be tracked.
The National Longitudinal Alcohol Epidemiologic Survey (NLAES), which
was conducted in 1992, was designed and sponsored by the National
Institute on Alcohol Abuse and Alcoholism. This survey gathered
information on alcohol consumption and alcohol-related problems
from 42,862 U.S. adults 
18 y old, one of whom was selected at random
from each household in a sample that was representative of the
noninstitutionalized population of the conterminous United States.
Variables in this data set include use of alcohol and other drugs
(illicit drugs as well as illicit use of prescription
medications), specific abuse and dependence, family history of
alcoholism, alcohol-related medical conditions and alcohol
treatment utilization. As discussed in the surveillance section, these
survey data can be used to estimate the degree with which drinking
patterns are congruent with the guidelines for both men and women. The
measures in NLAES do not permit calculation of exact frequency of
drinking in excess of the dietary guideline and may miss some cases in
which there is beverage overlap (people consuming more than one
beverage type on a drinking occasion) that leads to exceeding the
guideline. The National Epidemiologic Survey of Alcohol-Related
Conditions (NESARC), the successor of NLAES, is scheduled to be fielded
in 2001 with a longitudinal follow-up in 2003.Thus, although it is
currently not possible to track long-term trends in adherence, data
on short-term trends from 2001 to 2003 will be available in time to
inform the development of the 2005 Dietary Guidelines.
The 2000 alcohol guideline contains a section entitled "Who should
not drink?" which states that some people should not drink
alcoholic beverages at all. These include the following: children
and adolescents; people who cannot keep their drinking moderate; women
who may become pregnant or who are pregnant; individuals who plan to
drive, operate heavy machinery or take part in other activities that
require attention or skill; and individuals taking certain prescription
or over-the-counter medications that interact with alcohol. Taken
together, the individuals in each of these five groups probably add up
to a substantial proportion of the population. Tracking these
individuals is just as important as following those who have no
contraindications to drinking and stay within or exceed the drinking
levels recommended in the guideline. Measuring "no drinking" on a
population basis is difficult. In alcohol surveys, the lowest category
of drinking is often called "abstainer." This term, however, is not
always defined as no drinking. For example, in NLAES, people
who respond that they have not had 12 drinks in the past year are
labeled past-year abstainers. Further, assuming that the
definitional issue is addressed, at present, identification and
delineation of the five above-mentioned groups in one survey do not
occur. As mentioned earlier, under-age drinking is currently
tracked through the Monitoring the Future, the Youth Risk Behavior
Surveillance System and the National Household Survey of Drug Abuse.
The NLAES provides information on individuals 18 y old. Special
surveys examining alcohol use among pregnant women are conducted
periodically. Obviously women who are very early in their pregnancies
and do not yet realize that they are pregnant, would not be included in
such studies. Clearly, extensive work remains to be done before we are
in a position to provide national surveillance related to the "no
drinking" aspects of the alcohol guideline.
Healthy People 2010
Chapter 26: Substance Abuse of Healthy People 2010
(42)
contains the following three objectives related to
levels of alcohol consumption.
26–11. Reduce the proportion of persons engaging in binge drinking of alcoholic beverages.
26–12. Reduce average annual alcohol consumption.
26–13. Reduce the proportion of adults who exceed guidelines for low-risk drinking.
The first two objectives are not new. Although the wording differs
slightly, both objectives appeared in Healthy People 2000. Because
Objective 26–11 deals with binge drinking, it is not directly relevant
to tracking how drinking levels correlate with the recommendations in
the Dietary Guidelines, although clearly binge drinking
exceeds the recommendations of no more than one or two drinks a day.
Objective 26–12 on average annual alcohol consumption is also not
directly related to the Dietary Guidelines. It is worth
noting at this juncture, however, that annual estimated per capita
alcohol consumption is a crude but robust and valuable means for
monitoring trends in U.S. alcohol consumption. These estimates are
based on population figures as they relate to information on beverage
sales, tax receipt data or both. The data come primarily from States
with some data provided by beverage industry sources. Because, as
mentioned earlier, alcoholic beverage sales and tax data represent the
gold standard against which self-reported alcohol
consumption is compared, this objective relates indirectly to the
alcohol guideline. Obviously, however, because per capita consumption
is calculated by dividing total sales data by the population 14 y
old, this measure cannot be interpreted on a per drinker basis without
an independent estimate of the proportion of drinkers within the
denominator.
Objective 26–13 is new and was developed specifically to track lower
level drinking in a range similar to that found in the Dietary
Guidelines. In Objective 26–13, low risk drinking is defined as
no >14 drinks a week or no >4 drinks per occasion for men and no >7
drinks per week or no >3 drinks per occasion for women
(43)
. The definition of a drink is identical to that found
in the Dietary Guidelines. The baseline data for tracking
this objective are derived from the 1992 NLAES. The analyses reported
here were restricted to the 18,352 individuals identified as
past-year drinkers, i.e., those who in response to screening
questions reported having consumed at least 12 alcoholic drinks in the
year preceding interview. In 1992, among current drinkers 21 y old,
72% of women and 74% of men exceeded the above-mentioned
recommended levels of alcohol consumption (42)
. Most
persons who exceeded these recommendations did so by drinking more than
the specified maximum number of drinks per occasion at least once a
year. The target for the year 2010 is to reduce to 50% the proportion
of adults, both men and women, who exceed the Healthy People 2010
guidelines for low risk drinking (42)
. This objective also
lists several categories of people who should not drink at all; these
are basically the same as those appearing in the Dietary
Guidelines alcohol guideline: children and adolescents; women who
are pregnant or considering pregnancy; persons who are alcohol
dependent; people taking prescription or over-the-counter drugs that
interact with alcohol; and people who plan to drive or engage in other
activities requiring attention or skill. A sixth category not listed in
the alcohol guideline also appears, i.e., persons with health problems
that may be made worse by drinking alcohol.
Surveillance
Parallel analyses of the 1992 NLAES disclosed that among women,
98.7% of current drinkers (as defined above) exceeded the alcohol
dietary guideline recommendation of no more than one standard drink on
any day. Among men, 91.2% exceeded the level of no more than two
standard drinks on any day (44)
. The majority of drinkers
who exceeded the dietary guideline limits exceeded the daily limit as
opposed to an overall volume of intake that exceeded the weekly limits.
In other words, most women who exceeded the guidelines did so because
they had >1 drink on any day but not >7 drinks in a given week.
Similarly, most men who exceeded the guidelines reported consuming >2
drinks on a given day but no >14 drinks a week. Among individuals who
exceeded the daily limits, 21.1% did so less than once a month and
54.7% did so less than once a week (44)
. Thus, in 1992,
the vast majority of drinkers, both men and women, drank more than
recommended in the alcohol guideline on at least one occasion in the
year before interview with about a fifth of these individuals exceeding
the guidelines less than once a month and slightly more than half doing
so less than once a week.
As mentioned earlier, the successor to NLAES, the National
Epidemiologic Survey of Alcohol-Related Conditions (NESARC) will be
fielded in 2001 with a longitudinal follow-up in 2003. Because this
follow-up survey data will not be available for a few years, we are
currently unable to track trends over time but will be able to do so in
the future. In the interim, trends in estimated per capita alcohol
consumption suggest that behavior probably has not changed dramatically
in the intervening years. In 1992, estimated per capita alcohol
consumption for the U.S. was 2.31 gal of ethanol per person 14 y old
(45)
. Over the years since 1992, per capita consumption
has decreased gradually, reaching a low of 2.17 gallons in 1995, and
then increased to 2.19 gallons in 1996. In 1997, the most recent year
for which per capita consumption data are available, the level was
similar at 2.18 gal ethanol (45)
.
|
STRATEGIES FOR IMPROVEMENT |
|---|
|
TOPABSTRACTTHE EVOLUTION OF THE...METHODS OF ASSESSMENTSTRATEGIES FOR IMPROVEMENTREFERENCES |
|---|
The fact that Healthy People 2010 offers alcohol recommendations slightly different from those in the Dietary Guidelines serves to highlight the need for further research. The current guideline states "Limit intake to one drink/d for women or two/d for men ..." How do readers interpret this statement? Do women conclude that they should never have more than one drink on any day or do they do mental arithmetic to average consumption, i.e., three drinks today and none for the next two is O.K.? Public health guidelines are designed to be "short and sweet" and protect as many people as possible. Yet clearly "one size" may not fit all. A 6-ft tall 300-lb woman may well be able to safely drink more than 1 drink/d, whereas for a small man, 2 drinks/d may be too much. How can this information be conveyed most effectively? The various Dietary Guideline committees over the years have all been composed of nationally known figures with extremely strong subject matter expertise. These committees have invested a great deal of time and effort in functioning as "wordsmiths" for the language in the alcohol guideline. What is the evidentiary basis for specific wording recommendations? Clearly, this is an important research agenda item for the future. Similarly, we know little of a scientific nature about the possible (mis)interpretations of guidelines and the potential for social movement backlash if a guideline is perceived as overly restrictive.
Identification of definitive biomarkers and the development of highly specific and sensitive laboratory tests to track alcohol consumption over time would also advance the field. Finally, creation of some type of miniaturized "wearer friendly" alcohol sensor or other telemetry device that would objectively quantify and record drinking patterns over time would revolutionize research in this area. Such developments would enable researchers to collect the data necessary to advance the science and best inform the American people.
|
FOOTNOTES |
|---|
2 Abbreviations: NAS, National Alcohol Survey;
NESARC, National Epidemiologic Survey of Alcohol-Related
Conditions; NHANES, National Health and Nutrition Examination Survey
series; NLAES, National Longitudinal Alcohol Epidemiologic Survey; QF,
quantity-frequency; TLFB, time line followback.
|
REFERENCES |
|---|
|
TOPABSTRACTTHE EVOLUTION OF THE...METHODS OF ASSESSMENTSTRATEGIES FOR IMPROVEMENTREFERENCES |
|---|
1. Dietary Guidelines Advisory Committee Report of the dietary guidelines advisory committee on the dietary guidelines for Americans 1985 U.S. Department of Agriculture Human Nutrition Information Service Washington, DC.
2. Dietary Guidelines Advisory Committee Report of the dietary guidelines advisory committee on the dietary guidelines for Americans 1990 U.S. Department of Agriculture Human Nutrition Information Service Washington, DC.
3. Dietary Guidelines Advisory Committee Report of the dietary guidelines advisory committee on the dietary guidelines for Americans 1995 U.S. Department of Agriculture Human Nutrition Information Service Washington, DC.
4. Dietary Guidelines Advisory Committee Report of the dietary guidelines advisory committee on the dietary guidelines for Americans 2000 U.S. Department of Agriculture Human Nutrition Information Service Washington, DC.
5.
Fuchs C. S., Stampfer M. J., Colditz G. A., Giovannucci E. L., Manson J. E., Kawachi I., Hunter D. J., Hankinson S. E., Hennekens C. H., Rosner B., Speizer F. E., Willett W. C. Alcohol consumption and mortality among women. N. Eng. J. Med. 1995;332:1245-1250
6. Rimm E. B., Giovannucci E., Willett W. C., Colditz G. A., Ascherio A., Rosner B., Stampfer M. J. Prospective study of alcohol consumption and risk of coronary disease in men. Lancet 1991;338:464-468[Medline]
7.
Thun M. J., Peto R., Lopez A. D., Monaco J. H., Henley J., Heath C. W., Doll R. Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N. Engl. J. Med. 1997;337:1705-1714
8. Dufour M. C. What is moderate drinking?. Alcohol Res. Health 1999;23:5-14[Medline]
9. Dawson D. A. Measuring alcohol consumption: limitations and prospects for improvement (editorial with commentaries). Addiction 1998;93:965-968[Medline]
10. Grant B. F. Epidemiology. J 1994 W. Langenbucher B. S., McCrady, W., Frankenstein, P. E. Nathan, Eds., Annual Review of Addictions and Treatment Vol 3. NY Pergamon Press. pp 71–86.
11. Clark W. B., Midanik L. Alcohol use and alcohol problems among U 1982 S. adults results of the 1979 national survey. National Institute on Alcohol Abuse and Alcoholism Monograph no. 1, DHHS Publication no. (ADM)82–1190, pp. 3–52. Bethesda, MD.
12. Dawson D. A. Proceedings: international workshop on consumption measures and models for use in policy development and evaluation. Alcohol. Clin. Exp. Res. 1998;22:1S-3S[Medline]
13. Stockwell T., Dawson D., Holder H., Jernigan D., Medina-Mora M. E., Single E. International guidelines for monitoring alcohol consumption and harm: report to the World Health Organization Programme on Substance Abuse, National Centre for Research into the Prevention of Drug Abuse 2000 WHO Geneva, Switzerland (in press).
14. Dawson, D. A. & Room, R. (2000) Towards agreement on ways to measure and report drinking patterns and alcohol-related problems in adult general population surveys: the Skarpo conference overview. J. Subst. Abuse (in press).
15. Midanik L. The validity of self-reported alcohol consumption and alcohol problems: a literature review. Br. J. Addict. 1982;77:357-382[Medline]
16. Weisner C., Schmidt L., Tam T. Assessing bias in community-based prevalence estimates: towards an unduplicated count of problem drinkers and drug users. Addiction 1995;90:391-405[Medline]
17. National Institute on Alcohol Abuse and Alcoholism 3rd ed. Alcohol Epidemiologic Data Reference Manual 1997;1 U.S. apparent per capita consumption of alcoholic beverages based on state sales taxation, or receipt data. NIH Publication no. 97–4263, p. 124. Bethesda, MD
18. Polich J. M., Armor D. J., Braiker H. B. The Course of Alcoholism: Four Years after Treatment 1981 John Wiley & Sons New York, NY.
19. Hilton M. E. A comparison of a prospective diary and two summary recall techniques for recording alcohol consumption. Br. J. Addict. 1989;84:1085-1092[Medline]
20. Perrine M. W., Mundt J. C., Searles J. S., Lester L. S. Validation of daily self-reported alcohol consumption using interactive voice response (IVR) technology. J. Stud. Alcohol 1995;56:487-490[Medline]
21. Greenfield T. K., Rogers J. D. Who drinks most of the alcohol in the U.S.?. The policy implications. J. Stud. Alcohol 1999;60:78-89
22. Kreitman N. Alcohol consumption and the preventive paradox. Br. J. Addict. 1986;81:353-363[Medline]
23. Dawson D. A. Alternative definitions of high risk for impaired driving: the overlap of high volume, frequent heavy drinking and alcohol dependence. Drug Alcohol Depend 1999;54:219-228[Medline]
24. Rehm, J., Greenfield, T. K. & Rogers, J. D. (2000) Average volume of alcohol consumption, patterns of drinking and all-cause mortality. Results from the U.S. National Alcohol Survey. Am. J. Epidemiol. (in press).
25. Rehm J. Measuring quantity, frequency, and volume of drinking. Alcohol. Clin. Exp. Res. 1998;22:4S-14S[Medline]
26. Sobell, L. C. & Sobell, M. B. (1995) Alcohol consumption measures. In: Assessing Alcohol Problems: A Guide for Clinicians and Researchers. National Institute on Alcohol Abuse and Alcoholism Treatment Handbook Series 4, pp. 55–73. The Institute, Bethesda, MD.
27. Lemmens P. H. Measuring lifetime drinking histories. Alcohol. Clin. Exp. Res. 1998;22:29S-36S[Medline]
28.
Rehm J., Greenfield T. K., Walsh G., Xie X., Robson L., Single E. Assessment methods for alcohol consumption, prevalence of high risk drinking and harm: a sensitivity analysis. Int. J. Epidemiol. 1999;28:219-224
29. Searles J. S., Perrine M. W., Mundt J. C., Helzer J. E. Self-report of drinking using touch tone telephone: extending the limits of reliable daily contact. J. Stud. Alcohol 1995;56:375-382[Medline]
30. Greenfield T. K. Ways of measuring drinking patterns, and the difference they make: experience with graduated frequencies 2000 Measuring Drinking Patterns Problems, and Their Connection An International Research Conference, Skarpo, Sweden, April 3–7, 2000. J. Subst. Abuse (in press).
31. Williams G. D., Aitken S. S., Malin H. Reliability of self-reported alcohol consumption in a general population survey. J. Stud. Alcohol 1985;146:223-227
32. Babor T. F., Del Boca F. K. Just the facts: enhancing measurement of alcohol consumption using self-report methods. Litten R. Z. Allen J. P. eds. Measuring Alcohol Consumption: Psychosocial and Biochemical Methods 1992:3-19 Humana Press Totowas, NJ.
33. Flegal K. M. Agreement between two dietary methods in reported intake of beer, wine and liquor. J. Stud. Alcohol 1991;52:174-179[Medline]
34.
Willett W. C., Sampson L., Stampfer M. J., Rosner B., Bain C., Witschi J., Henneken C., Speizer F. E. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am. J. Epidemiol. 1985;122:51-65
35.
Rimm E. B., Giovannucci E. L., Stampfer M. J., Colditz G. A., Litin L. B., Willett W. C. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am. J. Epidemiol. 1992;135:1114-1126
36. Longnecker M. P., Lisner L., Holden J. M., Flack V. F., Taylor P. R., Stampfer M. J., Willett W. C. The reproducibility and validity of a self-administered semiquantitative food frequency questionnaire in subjects from South Dakota and Wyoming. Epidemiology 1993;4:356-365[Medline]
37. National Institute on Alcohol Abuse and Alcoholism Alcohol epidemiologic data directory 1998 CSR Washington, DC.
38. Johnston, L. D., O’Malley, P. M. & Bachman, J. D. (1999) National Survey Results on Drug Use from The Monitoring the Future Study, 1975–1998, Vol. 1, Secondary School Students. National Institute on Drug Abuse, Bethesda, MD.
39.
Greenfield T. K., Midanik L. T., Rogers J. D. A ten-year national trend study of alcohol consumption 1984–1995: is the period of declining drinking over?. Am. J. Public Health 2000;90:47-52
40.
Midanik L., Greenfield T. K. Trends in social consequences and dependence symptoms in the United States: The National Alcohol Surveys, 1984–1995. Am. J. Public Health 2000;90:53-56
41.
Midanik L.T., Clark W.B. The demographic distribution of U.S. drinking patterns in 1990: descriptions and trends from 1984. Am. J. Public Health 1994;84:1218-1222
42. U.S. Department of Health and Human Services (2000) 26: Substance abuse. In: Healthy People 2010 (Conference edition, vol. II), pp. 26–1–26–56. USDHHS, Washington, DC.
43. National Institute on Alcohol Abuse and Alcoholism (1995) The physicians’ guide to helping patients with alcohol problems. NIH Publication no. 95–3769. NIH, Rockville, MD.
44. Dawson D. A. U 2000 S. low-risk drinking guidelines an examination of four alternatives. Alcohol. Clin. Exp. Res. (in press).
45. Nephew, T. M., Williams, G. D., Stinson, F. S., Nguyen, K. & Dufour, M. C. (1999) Surveillance Report #51. Apparent per capita alcohol consumption: national, state, and regional trends, 1977–97. National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD.
This article has been cited by other articles:
|
|
 |
|
  A. H. Friedlander and D. C. Norman Geriatric alcoholism: Pathophysiology and dental implications J Am Dent Assoc, March 1, 2006; 137(3): 330 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 References Circulation, December 17, 2002; 106(25): 3373 - 3421. [Full Text]
|
|
|
|
 |
|
 R. Ballard-Barbash Designing Surveillance Systems to Address Emerging Issues in Diet and Health J. Nutr., February 1, 2001; 131(2): 437S - 439. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||
