When Is a Percent Not a Percent? When You've Been Drinking: A Letter to the Editor Regarding BACs and BrAC Terminology.

Dear Editor:

There is a significant problem with the terminology used to describe blood/breath alcohol concentration (BAC and BrAC) test results. The alcohol in BAC results is often measured in grams per deciliter or 100 milliliters, or grams per cubic centimeter, for blood tests, and as grams per 210 liters of breath, for breath tests.

In many places this is called a "percentage," or a "percent" of alcohol in the blood, which is flatly wrong. This "percent" is jargon toxicologists use to describe what they measure: the weight of alcohol in a volume of blood (or air, for a breath test). This measurement is a basic ratio of weight to volume. Because the two units of measurement are different, it is impossible to develop a percentage from this ratio (you must divide weight by weight or volume by volume to get a percentage). Toxicologists call this ratio "percent weight per volume," sometimes abbreviated as "%weight/volume," or "percent grams per liter." Their shorthand eroded to just "percent." They "knew" it wasn't a true percentage. A BAC result was usually rounded and reported as a two-digit number/decimal such as .10, which "looks" like a number just begging to be converted to a percentage.

The terminology is complicated by a coincidence: Ethyl alcohol, the drinkable part of the alcohol chemical family, is significantly lighter than water or blood. Its specific gravity is 0.789, while water is 1.0 and blood is variable at about 1.0506 at body temperature. However, only about 80 percent of blood is made up of water, and the alcohol you drink is dissolved only into that 80 percent of your blood. That 80 percent times 1.0506 equals .84048, which is a lot closer to the alcohol number, 0.789, with the result that the actual weight of alcohol is close to the weight of the water in the equivalent quantity of blood. (Ignore the "false precision" ridiculousness of a variable number carried out to five decimal places inside an idiosyncratic living human body with the Heisenberg problem of a BAC always increasing or decreasing, never stable short of death.) So, on a weight per weight basis, the weight of alcohol and the weight of water in the blood are very nearly equal--so saying that a BAC of. 10 equals a percentage (in the proper sense of the word) is very nearly right--but not quite. Since alcohol is significantly lighter, it takes a lot more of it to fill up the same volume as an equal weight of water. Saying a BAC of. 10 equals a percentage of alcohol on a volume per volume basis is therefore significantly wrong, by about 33 percentage points (the difference between .0789 and 1.0506, if 1.0506 is divided by .789), or a ratio of 3 to 4. (A BAC of .015 is 15 parts of alcohol per 10,000 parts of blood by weight, but 20 parts per 10,000 by volume).

Liquids are most often measured not by weight but by volume--liters, milliliters, deciliters, pints, quarts, gallons, shots, ponies, fifths, flagons, magnums and kegs. Nobody buys a pound of milk or a kilo of 18-year-old single malt. That's why .10 isn't .10 percent, if you're speaking about volume.

Many style guides and reference books require this kind of number have a zero before the decimal point, to read, for instance, as 0.09. You often see BAC and BrAC results this way, but more often the preceding zero is omitted: .09. The first major precedent for not using the zero placeholder is gun caliber: .22, .357 Magnum, .38 Police Special, .380 ACP, etc. The second is a century and a half of baseball statistics: In 1941 Ted Williams hit .406, not 0.406; ask any baseball fan. You can safely lose the zero. Like batting averages, no BAC will ever reach 1.000.

Percent was imbedded into BAC testing language from the start. In 1936 the National Safety Council established its Committee on Tests for Intoxication, and in 1938 it collaborated with the American Medical Association's Committee to Study Problems of Motor Vehicle Accidents to set up the first chemical standards for the legal interpretation of "under the influence of alcohol," as follows:

* Below 0.05 percent alcohol in the blood: no influence of alcohol within the meaning of the law.

* Between 0.05 and 0.15 percent: alcoholic influence usually is present, but courts of law are advised to consider the behavior of the individual and circumstances leading to the arrest in making their decision.

* 0.15 percent: definite evidence of "under the influence," since every individual with this concentration would have lost to a measurable extent some of the clearness of intellect and control of himself that he would normally possess. (National Safety Council, 1997)

People who aren't toxicologists (just about all of us) understand "percent" in the normal way, one part out of a hundred, sometimes with a twist. People usually intuit that .06 does NOT mean "6 percent of your blood" (that's a lot of booze in your system). Part of the problem is simple innumeracy; many people usually don't know what to do with .08. Left alone, it generally implies 8 percent of something. Slap the word percent after it and legislate it as the nationwide per se limit of drunk driving and the innumeracy problem evaporates. Anyone who ever tipped a waiter, used a discount coupon at Bed, Bath and Beyond, paid income tax, or studied Bryce Harper's batting average knows what a percentage is. So a drunk-driving BAC of .08% is dead simple: 8 percent of your blood is Pabst Blue Ribbon, and 8 percent of the air you blow into the officer's breathalyzer right before he/she takes your keys and puts you in the back seat of his/her squad car. Or maybe it's not 8 percent, it's 8 hundredths of a percent, same difference. Only it isn't.

The National Highway Traffic Safety Administration itself recognized and acknowledged this language problem. In October 1994 NHTSA's Office of Program Development and Evaluation published Computing a BAC Estimate describing the chemistry and mathematics involved in computing a BAC (NHTSA, 1994). This NHTSA report became known as the "NHTSA formula" for calculating BAC. The main text itself contained no description of the language problem, but the very last footnote--the very last paragraph of the entire report--reads:
```6. "Percent" in U.S. toxicological circles means grams per 100
milliliter; this is a weight per volume measure and does not carry the
usual meaning of percentage [emphasis added]. In this report, BAC is
defined as either blood alcohol concentration, stated as grams per 100
milliliters of blood or as breath alcohol concentration, stated as
grams per 210 liters of breath, and is reported without a "%" sign
```

Having published this advice about omitting the % sign, NHTSA proceeded to ignore it in many documents, as did everyone else; the jargon had been in use for more than 60 years at that point. For instance, a NHTSA document published three years later said, "The [BAC] measurement is based on grams per deciliter (g/dl) [sic; read g/dL with an uppercase L], and in most states a person is considered legally intoxicated if his or her BAC is .10 g/dl or greater; that is, alcohol makes up one-tenth of one percent of the person's blood" (emphasis added) (NHTSA, 1997). The first two clauses of that sentence are fine, and it is arguable whether the definition, grams per deciliter (g/dL), and the example of a reading of. 10 is self-sufficient and clear, or whether it required even further "interpretation." But having made the decision to add a further interpretation of what the BAC means, the writer made an incorrect statement of fact, and clearly used the word "percent" in its ordinary, common meaning, apparently unaware that it was a piece of toxicological jargon.

The web site of the Substance Abuse and Mental Health Services Administration, a division of the Department of Health and Human Services, once offered a publication called Blood Alcohol Concentration: Understanding How Drinking Affects You (Samber, 1997). SAMSHA's write-up at the bottom of its first paragraph read, "BAC is expressed as the percentage of alcohol in deciliters of blood." Anyone reading this would take percentage at face value--and would be wrong. A web page of the Centers for Disease Control and Prevention also displays the % symbol (see www.cdc.gov/motorvehiclesafety/impaired_driving/impaired-drv_factsheet.html). It gives as its source The ABCs of BAC, a 2005 NHTSA brochure that contained the % symbol. That NHTSA brochure was revised and corrected a year ago (NHTSA, 2016).

The web site of the University of North Carolina's Highway Safety Research Center (www.hsrc.unc.edu/pubinfo/alc_bac.htm) until recently had a page that read, "BAC refers to the amount of alcohol contained in a person's blood. It is measured as weight per unit of volume. Typically this measurement is converted to a percentage such as 0.10%, which indicates that one-tenth of a percent of a person's blood is alcohol."

This HSRC language now appears all over the Internet, often verbatim. The Quizlet "Flashcards" web site describes itself as "the world's largest student and teacher online learning community. Every month, over 20 million active learners from 130 countries practice and master more than 140 million study sets of content on every conceivable subject and topic" (see https://quizlet.com/77472513/bgyyyyyy-flash-cards/). And what they learn is wrong where BACs are concerned. The Brainly "OpenStudy" web site of "THE [sic] Journal," is billed as "dedicated to informing and educating K-12 senior-level district and school administrators, technologists, and tech-savvy educators within districts, schools, and classrooms to improve and advance the learning process through the use of technology. Launched in 1972, THE [sic] Journal was the first magazine to cover education technology," and has the HSRC paragraph verbatim (see http://openstudy.com/updates/567457cee4b032ed60dc2263). The HSRC language even crossed the Atlantic fully intact into the Glossario Enologico, a multilingual online wine glossary and tutorial of the University of Genoa, which gives as its source that now removed UNC Highway Safety Research Center web page (www.farum.it/glos_enol/show.php?glos_enol=b9875fffbd83a5b9d5ac38b556e34cf5&id=2328).

The author guidelines for the Journal of Studies on Alcohol and Drugs (2017) published by the Rutgers Center of Alcohol Studies, reads, "Abbreviations, Symbols, and Nomenclature: Blood alcohol concentration (BAC) should be expressed in percent for whole blood [emphasis added] and in mg/dl for plasma. ... The forensic standard for BAC (e.g., driving while intoxicated = .08%) is measured in whole blood... ."

BrAC numbers are, literally, 2,100 times more problematic. Testing has shown that after drinking alcohol, your blood has in it approximately 2,100 times as much alcohol as your breath. Your body absorbs about 90 to 95 percent of the alcohol you drink, and as blood carrying it passes through your lungs a tiny amount of that alcohol evaporates to mix with the air you exhale; the ratio for this is about 2,100 to 1.

This ratio is called the "blood/breath partition ratio," or just plain "partition ratio," and actually varies from about 1:1,900 to 1:2,400, depending on a person's body temperature, recent vigorous exercise, the recent presence of alcohol in the mouth and whether the BAC is rising or falling. It was created by the National Safety Council's Committee for Tests on Intoxication more than six decades ago (Turner, Bennett, Cestaric, & Muehlberger, 1952), and has been controversial ever since. Some sources--usually DUI defense attorneys--even put the ratio range as far apart as 1:1,100 to 1:3,000 (Dubowski, 1985) or more. Other nations have different rules for the partition ratio, but in the United States it is set at 1:2,100 by statute, so 2,100 is the number used here (usually expressed as grams of alcohol per 210 liters of breath), with the understanding that it is a controversial, variable average, not a hard-and-fast number. (Once again, consider the absurdity of a human variable that ranges from 1,900 to 2,400 carried out to several decimal places.)

When you blow into a breathalyzer, the device measures the alcohol content and multiplies it by 2,100 to obtain equivalence to a blood test reading. Thus, by definition, it would be impossible for you to have a blood result of .09 g/dL and a breath test also be .09 g/dL; in fact, the breath test quantity is about 2,100 times smaller as a "true" measurement.

The percent toxicological jargon was deeply imbedded into legislation States (delete "States" and add "states") used to implement laws about BAC tests, breathalyzers, regulations for testing and test equipment and related matters. For example, the Indiana Administrative Code on breathalyzer equipment stipulated "[t]he instrument must analyze breath samples and the numerical blood alcohol concentration value it reports shall be in units of percent by weight (%weight/volume) of alcohol in blood" (Ind. Admin. Code title 260, r. 1.1-5-1(a)(1)). However, that code--and similar codes in many other states--neglected to define exactly what the word "percent" meant, nor did it indicate the special meaning given to it by toxicologists. That omission was a lawsuit just waiting to happen, and sure enough, it did.

Virtually from the date of inception--New York State passed the first drunk driving law in 1910--every kind and type of alcohol-related traffic ticket has been tested, challenged and appealed in any and every court that had jurisdiction. The literature search is enormous. Understandably, defense attorneys attack these tests on behalf of their clients; that is how the adversarial legal system works. In 1998 there came a case that was markedly different from all others. Sales v. Indiana didn't dispute "the facts," the tests, the science, the biology, the variability, the physiology, the equipment, or anyone's conduct or training. Sales v. Indiana challenged the word "percent."

After failing to yield the right-of-way, a man named Mark Sales was pulled over by a police officer, who administered a breath test; the results showed a BrAC of ".14 grams of alcohol per 210 liters of breath." Sales was charged with three counts: operating a vehicle while intoxicated, operating a vehicle with .10 percent or more by weight of alcohol in his blood and operating a vehicle with at least .10 percent of alcohol by weight in his breath. Sales' attorney filed a motion to suppress the breath test on the basis of the language problem, noting that the law said the test result should be expressed as previously mentioned, "the numerical blood alcohol concentration value it reports shall be in units of percent by weight (%weight/volume) of alcohol in blood," whereas the breathalyzer itself issued a printout that read "grams of alcohol per 210 liters of breath."

Sales also argued that it was impossible to have a blood content of .14 as well as a breath content of .14, since the one is supposed to be 2,100 times larger than the other. The trial court upheld the first two counts, but sua sponte ("of its own accord") dismissed the third count as being "physically and medically impossible." Both sides appealed to the Indiana Court of Appeals, whose three-judge panel upheld the first two counts and dismissed the third. This decision was appealed to a four-judge panel of the Indiana Supreme Court, who ultimately re-instated the third count. Along the way, some excerpts from the discussions by the three court rulings are instructive:
```"The need to convert alcohol in the breath to alcohol in the blood can
be eliminated by stating the prohibition in terms of alcohol in the
breath as well as alcohol in the blood. Prior to 1997, 38 of the 50
states had done so by adopting a standard to include grams of alcohol
per 210 liters of breath as well as grams of alcohol per 100
milliliters of blood."

--Trial Court (Sales v. Indiana, 1998)
"The term 'percent' is unambiguous. Thus, we must apply the plain,
ordinary and usual meaning of the word.... The plain, ordinary and
usual meaning of 'percent' is 'out of each hundred; per hundred....
[W]hen words in a statute are unambiguous, a court must apply the plain
and obvious meaning and not resort to other rules of construction.
[Emphasis added.]... There is no indication in either Indiana Code
Section 9-30-5-1 or elsewhere that the word 'percent' is to be given a
technical or scientific meaning. Therefore, we must apply the plain,
ordinary and usual meaning of 'percent.'"

--Court of Appeals
```

and the coup de grace from the Indiana Supreme Court:
```"Under the trial court's calculations and rationale, a person would
violate Indiana Code [section] 9-30-5-1 (a)(2) only if his or her
'alcohol blood ratio' were 210%, which would long since have produced
not an impaired driver but a corpse, indeed one perhaps needing no
embalming."

--Indiana Supreme Court
```

It took three courts, eight judges and an "unusually expedited" session of the Indiana General Assembly to do basically what at least 38 other states had done--eliminate the "somewhat arcane" use of the word percent from BAC and BrAC results. As of December 31, 2015, only 4 states and Puerto Rico still use legal language that describe BACs as the "percentage" of alcohol in the blood; 31 states and the District of Columbia have completely abandoned the old toxicology shorthand and simply use "grams per 100 milliliters." Fifteen states retain part of the old language, "percent of weight," meaning percent (properly, "percentage") of weight per volume, but immediately add in the additional language defining it (NHTSA, 2017). It seems clear that continued use of the word "percent" in any future discussions of BACs and BrACs is nothing but trouble.

REFERENCES

Dubowski, K. M. (1985). Absorption, distribution and elimination of alcohol: highway safety aspects, Journal of Studies on Alcohol and Drugs, Suppl., 98.

National Highway Traffic Safety Administration. (1994, October). Computing a BAC estimate. (Unnumbered report). Washington, DC: Author.

NHTSA. (1997). Setting limits, saving lives: The case for 0.08 BAC laws (Report No. DOT HS 808 524). Washington, DC: Author.

NHTSA. (2016, July, revised from 2005). The ABCs of BAC (Report No. DOT HS 809 844). Washington, DC: Author.

NHTSA. (2017, March). Digest of impaired driving and selected beverage control laws, 30th edition, Current as of December 31, 2015 (Report No. DOT HS 812 394). Washington, DC: Author.

National Safety Council. (1997). A history of the committee on alcohol and other drugs (CAOD). Itasca, IL: Author.

Sales v. Indiana, Cause No. 08D01-9801-CM-22, 1998. [Carroll County Superior Court, a.k.a. Trial Court]; Sales v. Indiana, 714 N.E.2d 1121, 1128 (Ind.Ct.App.1999) [Indiana Appeals Court, No. 08A02-9806-CR-515. September 10, 1999], also cited as Sales v. Indiana, 714 N.E.2d 416, 421 (Ind. 2000). Also cited as Sales v. State, 714 N.E.2d 416, 421 (Ind. 2000); Sales v. Indiana, No. 08S02-0001-CR-29. February 7, 2000. [Indiana Supreme Court].

Samber, S. (1997, September 4). Blood alcohol concentration: Understanding how drinking affects you. Rockville, MD: NCADI [National Clearinghouse for Alcohol and Drug Information] Reporter.

William C. Swanson