Title Banner - Alcohol and Highway Safety 2001: A Review of the State of Knowledge

3 - Alcohol Effects on People

The prior chapter has summarized the current state of knowledge of the size and magnitude of the alcohol-crash problem in the U.S. at the millennium. This chapter discusses the current state of knowledge of the more basic interactions between alcohol and various parts of the body, and of more direct interest to the topic of alcohol-safety, how alcohol affects human behavior related to driving.


Absorption, Metabolism, and Elimination of Alcohol

The 1978 update briefly discussed the elementary aspects of how the body processes alcohol. No significant changes in our understanding of the fundamentals of these processes have occurred since then, although significant new knowledge of interest to specialists has been gained.

Processing of alcohol by the body begins with absorption by the stomach and small intestines, a process that generally requires some one to three hours, depending on the type and quantity of the alcoholic beverage, and the presence of food in the stomach.

Alcohol enters the bloodstream by simple diffusion, and does not have to be digested. The presence of food in the stomach slows the rate of alcohol absorption, but absorption is also influenced by other factors including the type of alcoholic beverage, the drinker's gender, body temperature, the presence of certain medications in the body, and the types of spices in the food. Distribution to various parts of the body then occurs.

Body fat and skeletal mass absorb very little alcohol. Thus, an identical quantity of alcohol per unit of body weight will induce a higher BAC in women than in men because of differences in body constitution (Bode and Bode, 1997). Some recent research suggests that, in a social drinking setting, a shorter time to peak BAC and a faster absorption rate may occur when alcohol is consumed over an extended period. In contrast, earlier studies found longer absorption times (Winek, Wahba, and Dowdell, 1996).

The variability of absorption time is illustrated by a study by Friel, Baer, and Logan (1995). The study examined alcohol disposition in 77 female and 97 male college seniors who were regular drinkers who exceeded legal intoxication levels at least twice a month by history. After receiving a standard alcohol dose (lower for females than for males) over 10 minutes, after a four-hour fast, breath alcohol concentrations (BrACs) were measured for two hours. The time to peak BrAC varied from 10 to 91 minutes after the start of drinking, and mean BrACs were significantly lower in females than in males.

Absorption and peak BAC vary by type of food as well as amount of food. For example, a study of a small sample of women subjects found that the peak BAC was significantly higher in those drinking alcohol and sodium (simulating salty food) than in those drinking alcohol with no sodium (Talbot and La Grange, 1999).

Alcohol is metabolized primarily in the liver, but metabolism occurs also in the stomach and small intestine. Gastric alcohol metabolism, which is significant only at low alcohol concentrations, is more efficient in men than in women, which helps explain why the same amount of alcohol produces higher blood alcohol concentrations in women than in men. There is also evidence that alcohol can be metabolized by bacteria in the large intestine. Bode and Bode (1997) note that alcohol is not only degraded, but also produced in the gastrointestinal tract as a by-product of bacterial breakdown of ingested carbohydrates.

Finally, of the alcohol absorbed, 90-98 % is oxidized, 1-5 % is excreted in an unaltered state in urine, and another 1-5 % is expired via the lungs (Vrij-Standhardt, 1991). The total time to eliminate alcohol from the body is dependent upon the variables that influence absorption (see above).

Measurement of Alcohol Presence

Since alcohol's immediate effects are due to its effect on the brain, it would be desirable to know the alcohol concentration in the brain after drinking. Obviously, direct measurements are impractical for most purposes, and other means must be used for estimating "brain-alcohol concentration."

Chemical tests of blood drawn from a vein or capillary are the preferred indirect way of estimating alcohol concentration in the brain in live humans. Other chemical tests that relate alcohol presence elsewhere in the body to alcohol presence in the blood, have also been used, the most common now being tests of alcohol in air expired from the lungs.9

Breath-alcohol measurement has become more precise and reliable since the 1978 update, and also more convenient and easy to perform, especially in forensic settings. The 1978 update noted that the factor (estimated at 2,100 at that time) for converting breath alcohol measurements to blood alcohol measurements could not be precisely determined, and also presented data from 28 studies on the blood/breath deviation. The data indicated that breath testers typically underestimated BAC by up to 10% or so.

More recent studies using improved technology indicate that the conversion factor may be closer to 2,400 than 2,100, (Jones and Anderson, 1996). This means that, on average, using a conversion factor of 2,100 would underestimate BAC by about 10%. Jones and Anderson note the fairly high variability of the conversion factor and discuss some of the factors that may influence the variability. Jones and Pounder (1998) discuss current practices for measuring alcohol concentration in clinical and forensic laboratories and recommend methods for assuring quality in laboratory procedures.

Two major advances in instrumentation of interest in the drinking-driving field are: much more precise and less expensive portable breath testers for operational use, and the development of "passive" breath testers that test the breath of expired air near the mouth without the need for collecting air directly from the mouth (Farmer, Wells, Ferguson et al., 1999). Other measurement techniques now under study in this country are the use of saliva (Flores, Spicer, and Frank, 1992) and "sweat patches" (Deveaux and Gosset, 2000) for estimating BAC. Saliva measurement devices are being used more often outside the United States and have been found to perform favorably for rapid estimation of BAC (Keim, Bartfield, and Raccio-Robak, 1996; Kiesow, Simons, and Long, 1993). Practical self-testing devices have also been developed and are being used in Australia (Haworth and Bowland, 1995; Haworth, Bowland, Vulcan et al., 1997). Exploratory studies of the use of laser technology to detect alcohol presence in a closed vehicle have also been conducted, but no formal reports of their results were found in our literature search.

In addition to chemical tests, improved behavioral tests for alcohol impairment are now being employed widely to assist police officers in identifying alcohol impairment among drivers suspected of a drinking-driving law violation. The standardized field sobriety test (SFST) of one's performance in a set of three sub-tests is being used in jurisdictions in all 50 States (Burns, 1999). The sub-tests are: horizontal gaze nystagmus (HGN), walk-and-turn (WAT), and one-leg-stand (OLS). HGN requires the subject to visually follow a moving object, and the angle of onset and degree of nystagmus (an involuntary jerking of the eye) is observed. Alcohol-impairment causes an earlier onset and a greater degree of nystagmus.10 HGN has been found to be the best index of alcohol of the three tests.

Subjective estimates of BAC by persons (e.g., police officers, physicians, and bartenders) who deal with drinkers in various settings have been shown to be notoriously inaccurate (Hansen, Popkin, Campbell et al., 1991). In another study of police officers's ability to detect even the odor of alcohol at various BACs up to .13, researchers found that odor strength estimates were unrelated to BAC levels and that estimates of BAC level "failed to rise above random guesses" (Moskowitz, Burns, and Ferguson, 1999).

Methods for calculating one's own BAC after consuming a given amount of an alcoholic beverage have been published in various forms, including formulas, procedures, tables, computer programs, and nomograms. South (1992) summarized factors affecting BAC and presented a formula calculating it, deeming the formula "complex to use and not very accurate." This assessment holds true for self-determination methods in general, which give only a rough idea of one's BAC after drinking. South, a resident of Australia, recommended that those wanting to know much they can drink and drive legally use a combination of counting drinks and using a coin-operated breath testing device.

Alcohol measurement techniques are discussed in more detail in Chapter 4 in conjunction with their use in alcohol-crash countermeasures.

Acute and Chronic Effects of Alcohol

The short-term or acute effects of alcohol of interest here are those related to alcohol's depressant effect on the brain. The exact nature of the mechanisms involved is not known. Fromme and D'Amico (1999) discuss basic knowledge of the neural systems that are implicated in alcohol's acute and chronic effects and suggest two relatively distinct neuroanatomical and neurochemical response systems to account for the subjective and behavioral effects of alcohol: (1) a simple reinforcement/motivation system, and (2) a complex neurochemical system that mediates higher-order cognitive functions and conditioned effects of alcohol. The U.S. Department of Health and Human Services's Ninth Special Report to Congress on Alcohol and Health (1997) provides an extensive discussion of the neuromolecular actions of alcohol on the brain and the ability of alcohol to influence many cellular functions.

It is known that the depressant effect increases with BAC; hence, the importance of BAC as an index of impairment. Extreme amounts of alcohol (e.g., BAC &asymp .40) can paralyze the respiratory system and cause death, but some persons can survive and even drive at these and still higher concentrations. Jones (1999) examined 81 drinking drivers in Sweden who had unusually high blood alcohol concentrations (BAC= .40+) when apprehended. He concluded that "drinking alcohol to reach a BAC of .40 or more and attempting to drive a motor vehicle indicates an exceptionally high cellular tolerance to the impairment caused by this drug. The alcohol burn-off rate [mean= .023 per hour] was relatively high in these heavy drinkers, which probably reflects the development of metabolic tolerance as well."

However, alcohol's effects begin to occur at much lower BACs. The 1978 update found that alcohol impairment of both simple processes involving the ability to perform relatively uncomplicated tasks not requiring high degrees of motivation and understanding begins to occur at BACs as low as .03. A recent review of literature published from 1981-1997 concluded that the majority of studies reported significant impairment in driving skills by BACs of .05, and that "alcohol impairs driving skills beginning with any significant departure from zero BAC" (Moskowitz and Fiorentino, 2000). The effects of alcohol on behavior are discussed further in the next section of this report.

Finally, of particular interest to this review is a study by Waller, Stewart, and Hansen (1986) which used data from North Carolina crash reports, driver records, and medical examiner reports to estimate the effects of alcohol on increasing the severity of injuries suffered in traffic accidents. They concluded that alcohol increases vulnerability to injury in any given crash. A more recent case-control study examined the risk of injury of any cause after the recent consumption of alcohol (McLeod, Stockwell, Stevens et al., 1999). The 797 cases were injured patients from a hospital emergency unit. The 797 controls were matched on residence location and were interviewed at home regarding activities leading up to the time of their matched case's injury. Cases and controls were breath tested and questioned about the injury event and alcohol and other drug use consumed in the six hours prior to the injury. Analysis of the data produced an odds ratio of 3.4 of sustaining an injury from any cause after consuming more than 60 grams of alcohol in a 6-hour period, after controlling for demographic variables.

Study of the chronic effects of alcohol used over a long period of time has generated a large body of literature since the 1978 update. Much of this literature in existence circa 1997 is reviewed in U.S. Department of Health and Human Services (1997). More recently, Dawson (2000) examined the effects of alcohol consumption and alcohol dependence on the overall risk of mortality in the United States using data from the 1988 National Health Interview Survey Alcohol Supplement matched to the National Death Index for the years 1988 to 1995. The author found that very heavy drinkers had a significantly increased risk relative to past-year abstainers and a risk of 1.65 relative to lifetime abstainers.

Hart, Smith, Hole et al. (1999) studied the relationship between alcohol consumption and mortality from all causes of 5,766 Scottish men, aged 35-64. The subjects entered the study in 1970-1973 and were followed for 21 years. The study found a similar relative risk for all-cause mortality for nondrinkers and for those drinking up to 14 units a week; and increasing risk with consumption, amounting to 1.34 for 15-21 units a week, 1.49 for 22-34 units, and 1.74 for 35 or more units. The authors concluded that "the overall association between alcohol consumption and mortality is unfavorable for those drinking more than 22 units a week," and that "there is no evidence for any protective effect at any level of consumption."

Of foremost concern has been the effects of alcohol on the liver which bears the major burden in metabolizing alcohol. Liver cirrhosis (a degeneration of liver tissue, resulting in fibrosis and nodule formation) has received particular attention. The path toward cirrhosis starts within the liver as inflammation (hepatitis), and progresses to fatty liver, and cirrhosis. The epidemiology of cirrhosis is complicated by the fact that heavy drinking is not its only cause, and that not all heavy drinkers develop cirrhosis. Other conditions that lead to cirrhosis include viral hepatitis, inherited diseases, diseases of the bile duct, and diseases of the blood. While it has been estimated that the incidence of cirrhosis is 3 out of 10,000 people, only about 10% to 15% of alcoholics have cirrhosis at the time of death.

DeBakey, Stinson, Grant et al. (1995) estimated that, during 1970 - 1992, age-adjusted death rates from alcohol-related liver cirrhosis dropped by 24.1% (5.4 deaths per 100,000 in 1970 to 4.1 deaths per 100,000 in 1992). An analysis of the relationships between cirrhosis mortality and per capita consumption of distilled spirits in the United States in the years from 1949-1994 found that there is a consistent long-term trend relationship between mortality from cirrhosis and per capita consumption of distilled spirits, but could not establish a direct causal link between consumption of distilled spirits and long-term cirrhosis mortality (Roizen, Kerr, and Fillmore, 1999). Kernochan and Yee (1999) even suggest that societal changes could be partially responsible for the development of serious liver disease in populations, and that spirits consumption may serve as marker for some societal event that occurred many years earlier and affected cirrhosis mortality.

The effects of alcohol consumption on the risk of various types of cancers has also been studied extensively. A meta-analysis of 123 studies found not only higher risks for cirrhosis, but also "weaker but significant" relationships for colorectum, liver, and breast cancers (Corrao, Bagnardi, Zambon et al., 1999). The authors found that: "For all these conditions, low intakes, corresponding to daily consumption of two drinks or two glasses of wine (25 g/day), have shown significant risks." The authors concluded:

"The small number of sufficiently reliable studies, the strong indications of heterogeneity across them, and the suspicion of publication bias suggest a great need for well-conducted epidemiological studies in several countries to examine the dose-response relationship between alcohol intake/drinking pattern and the risk of several alcohol-related conditions."

Finally, an extensive recent study on carcinogens in general (U.S. Department of Health and Human Services, 2000a) concluded that "consumption of alcoholic beverages is known to be a human carcinogen based on sufficient evidence of carcinogenicity from human studies that indicate a causal relationship between consumption of alcoholic beverages and cancer in humans," and, specifically, that:

"Consumption of alcoholic beverages is causally related to cancers of the mouth, pharynx, larynx, and esophagus. Cohort and case control studies in a variety of human populations are notable for their consistency in reporting the presence of moderate to strong associations with dose-response relationships for these four sites. Evidence also supports a weaker but possibly causal relation between alcoholic beverage consumption and increased risk of cancers of the liver and breast."

By contrast, numerous studies have found a protective effect of moderate drinking on heart disease. The American Heart Association (AHA) issued an advisory in 1997, stating that:

More than a dozen prospective studies have demonstrated a consistent, strong, dose-response relation between increasing alcohol consumption and decreasing incidence of coronary heart disease (CHD). The data are similar in men and women in a number of different geographic and ethnic groups. Consumption of one or two drinks a day is associated with a reduction in risk of approximately 30-50 percent. Studies of coronary narrowings defined by cardiac catheterization or autopsy show a reduction in atherosclerosis in persons who consume moderate amounts of alcohol. In general, the inverse association is independent of potential confounders such as diet and cigarette smoking. Concerns that the association could be an artifact due to cessation of alcohol consumption in persons who already have CHD have largely been disproved. No clinical trials have been performed to test the alcohol-CHD relation. However, the large numbers of observational studies support a true protective effect of moderate consumption of alcohol. While 100,000 excess deaths have been attributed to alcohol-related diseases each year, approximately 80,000 excess deaths would occur if all current consumers of alcohol abstained from drinking." (Pearson, 1997)

Many other, but not all, studies and reviews have arrived at similar conclusions, including a large-scale study of the responses of over 43,000 respondents to the 1988 National Health Interview Survey (Hanna, Chou, and Grant, 1997). However, the study by Hart and associates cited above found no strong association between alcohol consumption and mortality from coronary heart disease after adjustment of the data. Puddey and associates (1997) cautioned that balanced public health advice based on studies should take into account the full spectrum of alcohol's effects on the cardiovascular system, particularly its well documented potential to increase blood pressure and the prevalence of hypertension. An editorial by Criqui (1997) offered stronger advice, concluding that, "while it is clear that a modest intake of alcoholic beverages affords some protection against CHD [coronary heart disease], a general public health recommendation endorsing drinking is contraindicated."

The above discussion merely touches on the extensive literature on the acute and chronic effects of alcohol on the human body. A more detailed discussion of the literature can be found in recent reports by the U.S. Department of Health and Human Services (1997; 2000b). All in all, heavy drinking has been found to adversely affect bodily functions and general well-being. Light to moderate drinking does not appear to have adverse long-term effects on the bodies of healthy persons, and seems even to have a protective effect in some instances. However, even light drinking can have adverse short-term effects on behaviors that lead to such harmful events as traffic crashes, and can increase the severity of injuries that result from those events.


Driving-Related Performance

The 1978 update reviewed a number of laboratory studies of the effects of alcohol on one's ability to perform various tasks related to driving, summarizing its conclusions as follows:

"With respect to the simpler behavioral processes, there is evidence that neuromuscular responses may be impaired in some individuals at BACs as low as .04% to .05% w/v [weight per volume] and that many more individuals suffer such impairment at BACs in the range of .10% w/v. However, studies indicate that experienced drinkers can, if motivated, overcome these impairing tendencies at BACs as high as .20% w/v. Vision per se is not greatly affected by alcohol at BACs of less than .10% w/v, but above that it becomes impaired in most persons. ‘Simple' tracking performance does not appear to be seriously degraded at BACs of less than .10% w/v, but the performance of ‘complex' tracking tasks has been degraded in many individuals at BACs in the .05% to .10% w/v range. The ability to divide attention between tasks can be impaired at very low BACs (i.e., .02% w/v) and is often impaired at BACs above .08% w/v.

Studies of the more complex behavioral processes indicate that risk taking may be increased at moderate BACs for introverts and light drinkers. Moreover, low doses of alcohol have been observed to improve the intellectual performance of heavy drinkers and alcoholics while having the opposite effect on lighter drinkers. Alcohol has been found detrimental to memory, particularly the long-term memory, of heavy drinkers." (pp. 48-49)

The report went on to conclude that "behavior that has been studied is consistently and significantly impaired in virtually all individuals as BACs approach .10." The 1989 update (and this review, as well) found no study that contradicted this conclusion, but that there was important new research on the impairing effects of alcohol at BACs below .10. The 1989 update cited a review by Moskowitz and Robinson (1988) that concluded that performance of tracking and divided attention tasks is degraded at BACs considerably less than .05, and that information processing, perception, and psychomotor skills are impaired at BACs of less than .10, but generally more than .05.

Thus, the focus of recent experimental research on the behavioral effects of alcohol has been on impairment at low BACs. The report by Moskowitz and Fiorentino (2000) mentioned in the prior section of this report reviewed 87 experimental studies of skills performance at low BACs. The authors made an effort to restrict the behaviors of concern to those clearly related to driving, and factors such as motivation, aggression, and emotion were excluded from the review. The results of 550 tests in 12 behavioral categories were complied. The review was concerned with behaviors at BACs of .08 and lower, but some of the studies also contained the results of tests at higher BACs. Commentary on each behavioral category was largely concerned with the BAC threshold at which impairment was first noticed. The authors reported thresholds as low as .01 for some skills, and as high as .06 for others.

We plotted the percentage of tests showing impairment versus type of test for two groups of BACs, <.05 and >.05. The graphs (Figure 3-1) indicate that, for the lower BAC group, only four behaviors out of the 12 were impaired in more than half of the tests for a given behavior. The four impaired behaviors were, in descending order of percent impaired: drowsiness (not, as the authors note, a behavior, but a condition), vigilance, divided attention, and visual functions. All of these are clearly related to driving, while some of the others showing a lesser percentage of impairment have a less obvious relationship.

Figure 3.1: Percentage of Tests Showing Impairment 
by BAC Group and Type Test

Figure 3.1: Percentage of Tests
Showing Impairment by BAC Group and Type Test

By contrast, for the higher BAC group, all but two of the behaviors (critical flicker fusion and aftereffects) were impaired in more than half of the tests, and two behaviors were impaired (drowsiness and vigilance) in all of the tests. Again, the highest percentages of behaviors impaired were for those most clearly related to driving, while the two behaviors with the lowest percent were the least related to driving.

Other reviews of experimental studies since the 1989 update have also concluded that alcohol can cause significant impairment at low BACs.11 For example, Ferrara, Zancaner, and Georgetti (1994) reviewed the international literature of the effects of low levels of alcohol on driving ability, and found that most authors had concluded that low alcohol levels (apparently BACs in the .025 - .08 range) can cause significant impairment in psychomotor performance, to the extent that driving safety is compromised.

However, an earlier review voiced different conclusions about the effects of low BACs on behavioral skills. In a 1985 review, Mitchell (1985) concluded that alcohol impairment of driving-related behavioral skills is greatest for those tasks that require cognitive functioning, and that simple perception alone is least affected. He found that impairment of tasks requiring cognitive functioning begins to be evident at BACs above .05 and that there was no evidence that BACs below .05 impair any behavior in most individuals. His review is one of the few that addressed the amount of impairment, finding that, for most behavioral skills, the impairment at low BACs is slight, the order of 8-10% in many studies. He concluded that tolerance to central nervous system impairment may develop in regular drinkers, with sensorimotor coordination showing the greatest degree of tolerance, and that divided attention shows relatively little impairment.

Driving Performance

Laboratory studies such as those discussed above cannot determine exactly how or to what extent the behaviors studied are related to driving, for example, whether a decrement of 10% in a given behavioral task will cause what, if any, decrement in driving performance. Tests of actual driving performance, conducted in on-the-road settings or in driving simulators, offer the promise of more realistic estimates of the effects of alcohol. As noted by Linné, Triggs, and Redman (1999), impairments are typically measured by increased lateral deviation, but other measures are sometimes used as well. The limitations of such tests are well-known, but have been reduced in recent years by technological advances that have made simulators and measurement techniques more sophisticated and more sensitive to alcohol effects.

The review by Moskowitz and Fiorentino discussed above also included literature on driving and flying. Included were 25 studies containing 50 behavioral tests, with over 90% of the tests, both at BACs <.05 and at BAC >.05, showing some degree of impairment. This is in contrast to Mitchell's review of earlier studies which concluded that impairment of actual driving is begins at BACs of .05 to .06, but is small, and then increases more rapidly as the BAC exceeds .10.

Several other simulator studies of alcohol effects have been published since the cutoff year of the review by Moskowitz and Fiorentino, 1997. Interestingly, the most recent of these that we located, (Lenne, Triggs, and Redman, 1999), found reduced performance in maintaining lateral position on a simulated road, and also in secondary reaction time, at BAC &asymp .05.


The general principles regarding the processing of alcohol by the body remain essentially unchanged from those established many years earlier. Alcohol is absorbed by diffusion, metabolized mainly in the liver, and the small remaining amount is eliminated in urine and expired air.

Alcohol's immediate effects are due to its depressant effect on the brain, and chemical tests of blood drawn from a vein or capillary are the preferred indirect way of estimating alcohol concentration in the brain in live humans. The most common way of estimating the concentration of alcohol in the blood is testing air expired from the lungs.

At the millennium, breath testing has become more precise, more reliable, and more convenient. Also, other techniques are evolving that measure alcohol presence in alternative substances such as saliva (Flores, Spicer, and Frank, 1992) and sweat. Practical self-testing devices have also been developed and are being used in some countries. Improved behavioral tests are also being employed widely to assist police officers in determining alcohol impairment among drivers suspected of a drinking-driving law violation. Subjective estimates of BAC by persons such as police officers and physicians, and the use of methods for calculating one's own BAC, are not accurate enough for use either in research or operationally.

The acute depressant effect of alcohol increases with BAC, and has been measured in terms of its effects on human performance at BACs as low as .03. Alcohol also has been shown to increase one's vulnerability to injury. Studies of the chronic effects of alcohol used over a long period of time indicate that very heavy drinkers have a significantly increased risk of mortality relative to lifetime abstainers. Studies have found not only higher risks for cirrhosis, but also relationships for colorectum, liver, and breast cancers (Corrao, Bagnardi, Zambon et al., 1999), and that low intakes corresponding to daily consumption of two drinks or two glasses of wine can lead to increased risks. By contrast, a protective effect of light to moderate drinking has been found in some instances, and seems well-established for coronary heart disease.

With respect to alcohol's effect on performance related to driving, recent research has focused on low BACs, it having been clearly established in prior research that performance is substantially impaired in virtually everyone at BACs of .10 and higher. Techniques for testing and measurement have improved markedly in recent years, resulting overall in increased sensitivity to degradations of behaviors due to alcohol as determined both in laboratory experiments and in tests of actual driving performance. As a result, there is evidence that behaviors related to driving are impaired at lower BACs than was previously believed, with increased impairment of many behaviors clearly occurring at BACs in excess of .05. The amount of impairment of these behaviors at lower BACs less than .05, and whether it is associated with increased crash risk, cannot be stated in general terms on the basis of the findings of experimental studies alone, but awaits new evidence from epidemiologic studies. The epidemiologic study by Zador and associates discussed in the prior chapter suggests that risk does not increase for drivers as a whole at very low BACs (less than .02).

  1. Jones (2000) provides an excellent review of the evolution of the technology of blood-alcohol and breath-alcohol testing over the past 50 years.
  2. An excellent discussion of HGN and its use by police officers can be found in a recent NHTSA report prepared by the National Traffic Law Center (Dietrich and Frost, 1999)
  3. Moskowitz and Fiortino (2000) cite some of these reviews in their report and provide pertinent references.