Highway Safety Programs; Model Specifications for Screening Devices To Measure Alcohol in Bodily Fluids

DEPARTMENT OF TRANSPORTATION

National Highway Traffic Safety Administration

[NHTSA Docket No. 94-004; Notice 2]

Highway Safety Programs; Model Specifications for Screening

Devices To Measure Alcohol in Bodily Fluids

AGENCY: National Highway Traffic Safety Administration (NHTSA), DOT.

ACTION: Notice.

-----------------------------------------------------------------------

SUMMARY: This notice establishes Model Specifications for the

performance and testing of alcohol screening devices. These devices

test for the presence of alcohol, and may use breath or other bodily

fluids, such as saliva, to do so. NHTSA is establishing these

specifications to support State laws that target youthful offenders

(i.e., ``zero tolerance'' laws) and the Department of Transportation's

regulations on Alcohol Misuse Prevention, and in recognition of

industry efforts to develop new technologies (e.g., non-breath devices)

that measure alcohol content from bodily fluids.

A Conforming Products List (CPL) will be published identifying the

devices that meet NHTSA's Model Specifications. The CPL can serve as a

guide for those interested in purchasing devices that screen for the

presence of alcohol.

DATES: The Model Specifications established by this notice become

effective August 2, 1994.

FOR FURTHER INFORMATION CONTACT:

Ms. Lori A. Miller, Office of Alcohol and State Programs, NTS-21,

National Highway Traffic Safety Administration, 400 Seventh Street,

SW., Washington, DC 20590. Telephone (202) 366-9835.

SUPPLEMENTARY INFORMATION: On December 15, 1992 (57 FR 59382), the U.S.

Department of Transportation (DOT) published a notice of proposed

rulemaking (NPRM) to implement the ``Omnibus Transportation Employee

Testing Act of 1991,'' which requires alcohol testing programs in the

aviation, motor carrier, rail, and mass transit industries. The

Research and Special Programs Administration (RSPA) proposed similar

regulations for the pipeline industry. In general, the NPRM proposed to

prohibit covered employees from performing safety-sensitive functions

when test results indicate alcohol concentration levels of 0.04 or

greater. The NPRM proposed to apply slightly different consequences to

employees having alcohol concentration levels of 0.02 or greater but

less than 0.04.

To determine alcohol concentration, the NPRM proposed to use breath

as measured by those evidential breath testing devices (EBTs) listed on

NHTSA's Conforming Products List (CPL) which are capable of providing a

printed result, sequentially numbering the tests conducted, and

distinguishing alcohol from acetone at the 0.02 BAC level. EBT's listed

on NHTSA's CPL have been tested and determined to meet the agency's

Model Specifications for EBTs, which were last amended on September 17,

1993 (58 FR 48705).

In a final rule published on February 15, 1994 (59 FR 7340), DOT

amended its regulations and added procedures for conducting alcohol

testing in transportation workplaces (49 CFR Part 40). This final rule

differed from the NPRM in a number of respects. The final rule required

the use of breath testing devices listed on the CPL for EBTs. For

screening devices, it permitted the use of EBTs on the CPL that do not

print the result, but only if confirmation tests are conducted using

EBTs listed on the CPL which are capable of providing a printed result.

(These devices must also be capable of distinguishing alcohol from

acetone at the 0.02 BAC level and sequentially numbering the tests

conducted.)

NHTSA published a separate notice in the same issue of the Federal

Register (59 FR 7372) proposing to adopt Model Specifications and a CPL

that would permit additional alcohol testing devices to be used for

screening purposes. In its notice, NHTSA proposed to establish Model

Specifications for alcohol screening devices, which differ from the

Model Specifications for Evidential Breath Testing devices in a number

of important respects. It stated that the proposed Model Specifications

are designed to test whether devices are suitable for screening, not

evidential, purposes and that they are designed to test the performance

of devices that may use bodily fluids other than breath (such as

saliva) to determine the presence of alcohol.

NHTSA requested comments on these proposed Model Specifications.

Comments Received

The agency received twenty comments in response to the notice.

Comments were received from manufacturers of screening devices and

related equipment, persons representing sectors of the transportation

industry subject to the DOT regulations (including rail, transit, motor

carriers and pipelines) and substance abuse program administrators, an

interested individual and a health professional.

General Comments

The comments, in general, were supportive of the agency's proposed

Model Specifications. Some of the comments praised the notice for

proposing to increase flexibility, stimulate development and reduce

barriers and cost for those charged with implementing DOT's new alcohol

testing rules.

A number of commenters raised concerns about the schedule NHTSA

would following publishing the final Model Specifications. DOT's final

rule becomes effective for large employers (in general, with 50 or more

safety-sensitive employees) on January 1, 1995. The commenters,

therefore, urged the agency to issue the Model Specifications and

approve conforming devices prior to that date. Two commenters

recommended that if final rules and product evaluations are not

completed within a specified period of time (one commenter suggested

August 1 1994, another mid-1994), the effective date of DOT's final

rule should be delayed.

In response to these comments, NHTSA has sought to publish the

final Model Specifications as quickly as possible. As described further

below, we intend to begin testing immediately, and hope to publish

within 30 days from today's date a Conforming Products List (CPL) of

screening devices that have been tested to date and conform to these

Model Specifications. The CPL will be updated and published

periodically, as further testing is completed.

A number of commenters raised issues that pertain to other notices

that were published in the Federal Register on February 15, 1994, such

as DOT's final rule (59 FR 7340) on Procedures for Transportation

Workplace Drug and Alcohol Testing Programs (49 CFR Part 40) or the

final rules and common preamble (59 FR 7302) on the Limitation on

Alcohol Use by Transportation Workers. Others raised issues that are

also outside the scope of NHTSA's notice and request for comments. For

example, one respondent commented that all alcohol testing should be

performed by law enforcement representatives. Another respondent urged

the Department to permit testing to be conducted only using evidential

breath testing devices. Other commenters suggested that the use of non-

breath alcohol tests (which use blood, saliva or urine samples) as a

condition for employment is an invasion of privacy and a violation of

individual rights.

NHTSA's Model Specifications contain the performance criteria and

methods for the testing of alcohol screening devices. It does not

address whether such devices are permitted to be used to perform

screening tests, who is authorized to administer such tests or who is

subject to them. These issues are addressed instead in DOT's final

rules.

Other commenters raised questions or concerns regarding the Model

Specifications for Evidential Breath Testing Devices, last revised on

September 17, 1993 (58 F.R. 48705), or the Model Specifications for

Breath Alcohol Ignition Interlock Devices (BAIIDs), published on April

7, 1992 (57 F.R. 11772).

Issues such as these are outside the scope of the notice published

in February proposing Model Specifications for alcohol screening

devices, and therefore have not been addressed in this notice adopting

Model Specifications.

Specific Comments on Model Specifications

No comments were received regarding some portions of the proposed

Model Specifications. These portions have been adopted without change.

For further discussion regarding these portions, interested persons

should review the February notice. Portions of the proposed Model

Specifications that generated comment, and the issues raised in the

comments, are discussed below.

1. Purpose, Scope, Classification and Definitions

In its February 15, 1994 notice, NHTSA proposed to define an

alcohol screening device as a device that is used to detect the

presence of 0.020 or more BAC, and that indicates the test result by

numerical read-out or by other means, such as by the use of lights or

color changes. All comments addressing these aspects of the Model

Specifications supported the definition. They have been adopted without

change.

The notice proposed that the Model Specifications would provide

that devices may measure any bodily fluid (including blood, breath or

saliva), but that the output must be in blood alcohol concentration

(BAC) units. It explains that NHTSA believes the relationship between

BAC and the bodily fluid being measured be properly established so that

a means for evaluating the device can be devised, and that NHTSA

considers use of a one-to-one conversion factor between blood and

saliva to be appropriate. NHTSA requested comments in the February 15

notice on the proposed use of a one-to-one conversion factor for

saliva, and on what may constitute acceptable criteria for bodily

fluids other than saliva, blood and breath.

All comments regarding the one-to-one conversion factor and the

applicability of the proposed Model Specifications to blood, breath and

saliva were supportive of NHTSA's proposal. These aspects of the Model

Specifications have been adopted without change.

Comments were received from the manufacturer of an alcohol

screening device that uses ocular vapor analysis. The type of analysis

used by this device measures alcohol using vapors from the surface of

the eye. The commenter requested that the model specifications include

the ocular vapor analysis technique as an acceptable and recognized

method.

The Model Specifications, as proposed in the agency's February 15

notice and as finally adopted in today's Federal Register notice,

define an alcohol screening device as a device that may measure ``any

bodily fluid'' for the purpose of detecting the presence of 0.020 or

more BAC. This definition is clearly broad enough to include use of the

ocular vapor analysis technology.

NHTSA did not include in its proposal, however, testing procedures

for all conceivable types of screening technologies. Rather, it

proposed testing procedures for the types of screening technologies

currently most commonly available. The notice explained that the agency

would modify and improve the Model Specifications as new data and test

procedures become available, and that it would alter the test

procedures, if necessary, to meet unique design features of specific

devices. If the test procedures need to be altered to test the ocular

vapor analysis technology, NHTSA would make such alterations. Any

needed alterations would be published in the Federal Register.

One commenter, a manufacturer of alcohol breath testing devices,

raised concerns about devices that are not capable of detecting ethyl

alcohol and isopropyl alcohol. The commenter stated that if devices

cannot identify all three of these alcohols, they will produce false

negative alcohol readings.

The definition of alcohol included in the proposed Model

Specifications permits alcohol screening devices to detect different

types of alcohol (including ethyl alcohol, methyl alcohol and isopropyl

alcohol), but does not require that devices must be capable of

distinguishing between each type. To determine compliance with the

Model Specifications, the agency proposed that it would conduct tests

using ethanol.

NHTSA does not disagree that the potential for false negative

results may exist should be a technology be employed in a screening

device that is specific to ethanol only and an individual has consumed

methyl or isopropyl alcohol. However, the agency is aware of no

screening devices using such a technology. Rather, the screening

devices available today on the market generally employ technologies

that are not specific to any single type of alcohol, and, therefore,

are capable of detecting (but not distinguishing between) ethanol and

the other alcohols.

As a result, and since ethanol is the alcohol most often consumed,

we believe that the probability of obtaining false negative results by

screeners that conform to these Model Specifications is extremely low.

The proposed definition has been adopted without change.

2. Statistical Accuracy

In its February 15 notice, NHTSA proposed to test alcohol screening

devices at 0.008 and 0.032 BAC under normal laboratory conditions to

determine their precision and accuracy at detecting the presence of

0.020 or more BAC (Test 1), and at 0.000 BAC to determine the

performance of these devices when providing blank readings (Test 2).

The notice explained that the .008 and .032 BAC levels were

selected based on criteria for precision and accuracy that are

equivalent to those used for EBTs. The criteria require that devices

perform at a level of accuracy within 0.005 of 0.020 BAC

(thereby establishing target valves within 0.015 and 0.025 BAC), and a

level of precision which yields a standard deviation not greater than

0.0042. To achieve a confidence rate of approximately 95% in the

results of these 20 tests, we proposed to establish measurement points

at 1.73 standard deviations (or 0.007 BAC) below and above the lower

and upper values, respectively (i.e., 0.015-0.007=0.008 BAC and

0.025+0.007=0.032 BAC).

One commenter expressed the opinion that the proposed method of

testing does not truly reflect the accuracy standard of

0.005 BAC with standard deviation not to exceed .0042 BAC.

This commenter recommended that instruments should be tested instead at

the .020 BAC level, that results should fall within the 0.15 and .025

BAC range, and that a deviation of not more than .0042 should be

maintained. The commenter's response further stated that, to achieve a

confidence rate of 95%, only 5% of the tests conducted should be

outside the .015 to .025 BAC range.

The method proposed by this commenter would require that devices

identify the precise BAC level detected by the instrument. The Model

Specifications do not include such a requirement. Rather, they simply

require that devices are capable of detecting the presence of alcoholic

at the 0.020 or greater BAC level. To accommodate the use of non-

numerical as well as numerical alcohol screening devices, the Model

Specifications use two test points which are 1.73 times the maximum

allowed standard deviation on either side of 0.020 0.005

BAC (0.008 and 0.032). The number of false positives and false negative

allowed were obtained based on the use of Student's distribution (a

small sample approximation to the normal distribution).

One commenter illustrated a range of error that would be permitted

under the proposed Model Specifications, and suggested that the Model

Specifications be amended to permit a smaller range of error. Another

commenter, addressing the same concern, proposed that the Model

Specifications be amended to provide for the adjustment of the test at

.032. This commenter recommends that we conduct 20 tests at .025 with

no more than one false negative result and 20 tests at .015 with no

more than two false positives. NHTSA believes these proposals would

require that screening devices perform at a higher level of precision

than is required for EBTs. The procedures contained in the proposed

Model Specifications have been adopted without change.

3. Test Methods

NHTSA proposed to use a Breath Alcohol Sample Simulator (BASS),

non-alcoholic human breath, and a calibrating unit to test breath

devices. For non-breath devices, the agency proposed to use

preparations of bodily fluids or scientifically acceptable substitutes.

For example, the agency proposed to use aqueous alcohol test solutions

equivalent to blood or saliva on a one-to-one basis to test saliva

devices.

One commenter, a manufacturer of a saliva device, expressed its

view that there are no fluids that are scientifically acceptable

equivalents to bodily fluids. The commenter asserted that aqueous

alcohol test solutions lack the viscosity, solid content and inhibitors

that are present in bodily fluids such as saliva, and recommended that

the agency instead collect saliva specimens from individuals known to

be alcohol-free. According to the commenter, the non-alcohol saliva

pool could then be spiked with various alcohol solutions for device

evaluation.

NHTSA disagrees with this respondent's comment. The agency has data

finding that aqueous alcohol test solutions are acceptable substitutes

for saliva-alcohol testing purposes.\1\ In addition, while we agree

that aqueous solutions and saliva do have different characteristics, we

have no reason to believe that these difference would interfere with

the agency's ability to test the capability of saliva screening devices

to detect alcohol content. The final Model Specifications continue to

provide that aqueous alcohol test solutions will be used.

---------------------------------------------------------------------------

\1\Flores, A.L., Spicer, A. and Frank, J.F., ``Laboratory

Testing of a Saliva-Alcohol Test Device by Enzymatics, Inc.,''

Washington, D.C., U.S. Department of Transportation, National

Highway Traffic Safety Administration, Technical Report No. DOT-HS

807 893, December 1992.

---------------------------------------------------------------------------

Two commenters recommended that NHTSA use alcohol reference

material 1828, obtained from the National Institute of Standards and

Technology (NIST), to prepare all standard solutions. One of these

commenters also suggested that, following preparation, these solutions

should themselves be analyzed against a referee method (enzymatic or

gas chromatography) which has been calibrated using NIST standards.

NHTSA does not plant to use NIST 1828 material in its standard

solutions. However, the agency presently uses the material for the

purpose for which it was intended, as a reference material for

calibration purposes, and will continue to do so.

The agency proposed to conduct 40 trials under Test 1 (20 at .008

BAC and 20 at .032 BAC) and 20 trials under Test 2 (at .000 BAC). For

reusable devices, these 60 trials would be conducted using a single

unit. For disposable devices, these 60 trials would be conducted using

60 separate units.

NHTSA's notice explained that some alcohol screening devices

indicate the presence of alcohol in a manner that is unambiguous and

requires no interpretation, such as by the use of a light or numerical

reading. For these devices, NHTSA proposed that Tests 1 and 2 (at .008,

.032 and .000 BAC) would be performed by an investigator at the DOT

Volpe National Transportation Systems Center (VNTSC). To conform with

the Model Specifications, the notice stated that the device must

perform with no positive results at .000 BAC, not more than one

positive result at .008 BAC and not more than one non-positive result

at .032 BAC. If the device is capable of providing a reading of greater

than 0.000 BAC and less than 0.020 BAC, the device must perform with

not more than one such result at .000 BAC.

NHTSA's notice explained that other devices indicate the presence

of alcohol in a manner that requires interpretation and may involve

some ambiguity, such as by the use of color changes. For these devices,

NHTSA proposed that Tests 1 and 2 (at .008, .032 and .000 BAC) would be

performed by ten individuals who have no knowledge of test BACs and

qualify as test interpreters. VNTSC would select these individuals

using manufacturer's restrictions, if any. These individuals would be

asked to read the manufacturer's instructions for the interpretation of

the device's read-out, and interpret the test results independently.

To conform with the Model Specifications, the notice proposed that

the device must perform, with each interpreter, with no positive

results at .000 BAC, not more than one positive result at .008 BAC and

not more than one non-positive result at .032 BAC. If the device is

capable to providing a reading of greater than 0.000 BAC and less than

0.020 BAC, the notice proposed that the device must perform, with each

interpreter, with not more than one such result at .000 BAC. These

aspects of the Model Specifications have been adopted without change.

An organization that represents substance abuse program

administrators suggested that, if practical, the ten individuals select

to interpret the devices should have no medical training since it is

likely that the persons who will be administering the tests in the

field will have no such training. The agency plans to select

individuals with varying backgrounds and experience. While we do not

believe there is justification for imposing a restriction on the

selection of individuals who have medical training, it is likely that

few if any of the individuals selected will have such training.

A manufacturer of saliva screening devices suggested that the Model

Specifications should provide for a familiarization period, to ensure

that investigators and individuals who will be evaluating these devices

are familiar with the manner in which the devices should operate.

The preamble to the proposed Model Specifications explained that

individual evaluators will be asked to read the manufacturer's

instructions before they perform their evaluations. These individuals

will be provided sufficient time to become familiar with these

instructions, and will also be given instructions for conducting the

evaluations. Investigators will also provide themselves with sufficient

time to read the manufacturer's instructions and become familiar with

the devices they are testing, as well as the evaluation procedures.

NHTSA stated in the February 15 notice that, through the

independent interpretation of ten individuals, it believed the Model

Specifications would ensure that the results of tested devices are

visible and will remain so for a reasonable period of time and area

likely to be interpreted in a consistent manner. The notice indicated

that the tests would require approximately two hours to run. The agency

requested comments on these aspects of the proposed Model and

Specifications.

The comments were supportive of these aspects of the proposed Model

Specifications, except that two commenters objected to the requirement

that screening results remain visible for two hours. One of the

commenters considered this to be an unreasonable requirement,

particularly when (according to the commenter) the primary basis for

the requirement is the convenience of the testing facility that will be

evaluating the device. The other commenter was concerned that this two-

hour period could invalidate the results, since some devices require

that the user read and record the test result within a specific period

of time (such as two minutes).

Upon further consideration based on these comments, NHTSA has

decided to modify the requirement that results must remain visible for

two hours. It is not feasible, however, for the agency to eliminate the

requirement altogether. In part to facilitate the evaluation of these

devices, and also to be consistent with the DOT Alcohol Testing

Procedures (49 CFR Part 40), which provide that the waiting period

between screening and confirmation tests must be at least 15 minutes

but should be no longer than 20 minutes, NHTSA will modify its testing

methods so that the interpretation of results will be accomplished

within 20 minutes of dosing. Accordingly, the results of disposable

interpretive devices will need to remain visible for a period of only

20 minutes.

The notice explained that, to NHTSA's knowledge, no reusable

devices currently use interpretive readings and the agency believes it

is unlikely that manufacturers would begin to use such readings in

reusable devices. Accordingly, NHTSA proposed that the Model

Specifications would not include a methodology for testing reusable

interpretive devices. We requested comments on this aspect of the

proposed Model Specifications. The commenters that addressed this issue

agreed with the agency's proposal.

For disposable devices that use interpretive readings, NHTSA

proposed to combine Tests 1 and 2, and number the units and expose them

to the three BAC levels using a methodology that would not reveal to

the person interpreting the test the dosage received by any particular

unit. NHTSA requested comments on this proposed methodology. No

comments were received. The proposed methodology has been adopted

without change.

The February notice proposed to test devices to determine whether

acetone or, in the case of breath or saliva devices, cigarette smoke

affects the functioning of the instruments. The notice also requested

comments on whether devices should be tested for interference from

other substances.

With regard to the test for acetone interference, one commenter

agreed that there is a need for such a test. Another commenter strongly

recommended that the test be deleted from the Model Specifications. The

commenter argued that acetone is unlikely to interfere with the

measurement of breath alcohol and, if persons have levels of acetone

that are sufficiently high to cause interference, such persons should

not be performing safety sensitive functions. In addition, the

commenter stated that requiring devices to distinguish between alcohol

and acetone would greatly increase instrument cost and restrict

participation for certain instruments.

NHSTA has reconsidered its position on this issue, and decided that

alcohol screening devices should not be required to distinguish between

alcohol and acetone, particularly since the instruments used for

confirmation testing are capable of distinguishing between these

substances. Based on existing data,\2\ we do not expect a high

incidence of acetone interference and, in the unlikely event that a

device indicates a positive result due to the presence of acetone, this

will be detected in the confirmation test. The Model Specifications

have therefore been amended to eliminate the acetone test.

---------------------------------------------------------------------------

\2\Flores, A.L. and Frank, J.F., ``The Likelihood of Acetone

Interference in Breath Alcohol Measurement,'' Washington, DC, U.S.

Department of Transportation, National Highway Traffic Safety

Administration, Technical Report No. DOT HS 806 922, 1985.

Frank, J.F. and Flores, A.L., ``The Livelihood of Actone

Interference in Breath Alcohol Measurement,'' Alcohol, Drugs, and

Driving, 3 (2), 1-8, April-June 1987.

---------------------------------------------------------------------------

With regard to cigarette smoke and other interfering substances, we

received only one comment, which stated that non-interference from

smoking, eating and drinking should not be a conformance requirement

since these activities can be avoided before a test is performed. If

the evaluation of cigarette smoke is retained in the Model

Specifications, this commenter recommended that it be performed for

information purposes only.

NHTSA expects the likelihood of cigarette smoke interference will

be much greater than acetone interference, and has decided to retain

the cigarette smoke test. As provided in the Model Specifications, the

test will be performed in accordance with the manufacturer's

instructions. Any waiting period specified in the manufacturer's

instructions will be strictly observed. The test will be performed

within one minute after the person smokes the cigarette where no

waiting period is specified in the manufacturer's instructions. NHTSA

did not propose to conduct a test for interference from eating and

drinking, and we have not added any such test in the final Model

Specifications.

The commenter also suggested that, if the Model Specifications

continue to include a cigarette smoke test, that the method used for

conducting this test on saliva screening devices should be similar to

that used for breath screening devices. NHTSA concurs that this

comment, and has revised the Model Specifications to clarify its

application to both saliva and breath devices.

The agency also proposes to conduct high (40 deg.C) and low

(10 deg.C) ambient temperature and vibration tests for alcohol

screening devices to determine their ability to function under a range

of environmental conditions. NHTSA proposes that these tests would be

performed by an investigator at VNTSC. Five trials would be conducted

at .000 BAC under Test 3.2. Forty trials (including 20 at .008 and 20

at .032 BAC) would be conducted under each of these other tests.

One commenter, a manufacturer of a passive alcohol sensor, noted

that the proposed temperature range for testing is more severe than

that for EBT testers. This commenter is correct. The temperature range

is more severe because it is anticipated that screening tests may be

performed outside in widely varying temperature conditions. Tests

performed with EBTs are generally performed indoors where temperatures

are controlled. The proposed temperature range has been adopted without

change.

Another commenter, a manufacturer of a saliva test device,

suggested that the specimens for saliva testing should be held at body

temperature (37 deg.C) while performing the two ambient temperature

evaluations ``to stimulate real-life situations.'' NHTSA disagrees with

this comment. When saliva tests are being conducted in the field, the

temperature of the saliva will change soon after the sample is taken

from the person's mouth. NHTSA therefore believes the procedures

contained in its proposed Model Specifications more accurately simulate

the conditions under which actual testing will be conducted. This

portion of the Model Specifications has been adopted without change.

The manufacturer of an alcohol breath testing device commented that

disposable devices, which cannot be checked for calibration on a

periodic basis, should be evaluated throughout their useful life. This

manufacturer also recommended that devices which require that results

be checked through a visual inspection should be tested under a variety

of light conditions, such as fluorescent, mercury vapor, sodium vapor

and daylight.

NHTSA disagrees that the Model Specifications should provide for

the evaluation of disposal devices throughout their useful life. As

explained in the February 15 notice, manufacturers of alcohol screening

devices must meet the requirements contained in FDA's Good

Manufacturing Practices regulations for devices used for medical

purposes (21 CFR Part 820), and they must include labels on their

devices that meet the requirements contained in FDA's Labeling

regulations for devices used for medical purposes (21 CFR 809.10), even

if the devices are not to be used for medical purposes.

The Labeling Instructions for Alcohol Screening Devices included as

an Appendix to the February notice instructed, among other things, that

the label ``Provide the reagent's shelf life and opened expiration

dating, if applicable.'' In addition, manufacturers must determine

shelf life and expiration dating in accordance with FDA's regulations

on Good Manufacturing Practices.

NHTSA has asked users of alcohol screening devices to provide both

acceptance and field performance data to the agency's Office of Alcohol

and State Programs (OASP) when such data are available. As we explained

in the February notice, if information gathered indicates that a device

on the CPL is not performing in accordance with the Model

Specifications, that a manufacturer is not complying with FDA's Good

Manufacturing Practices, or that a device's label does not comply with

FDA's Labeling regulations, an investigation would be conducted and

appropriate measures would be taken. For these reasons, the Model

Specifications have not been amended to provide for the evaluation of

disposable devices throughout their useful life.

NHTSA accepts the recommendation that certain devices should be

tested under a variety of light conditions. The Model Specifications

have been amended to provide that interpretive devices which require

that results be checked through a visual inspection should be tested

under incandescent, mercury vapor, sodium vapor and daylight as well as

fluorescent conditions.

To conform with the Model Specifications, the notice proposed that

the device must perform with no positive results at each test performed

at .000 BAC, not more than one positive result at each test performed

at .008 BAC and not more than one non-positive result at each test

performed at .032 BAC. If the device is capable of providing a reading

of greater than 0.000 BAC and less than 0.020 BAC, the notice proposed

that the device must perform with not more than one such result at .000

BAC. No comments were received regarding this aspect of the proposal.

It has been adopted without change, except that the final Model

Specifications clarify that there can be no more than one ``can't

tell'' result for disposable interpretive devices.

4. FDA Involvement

When alcohol screening devices are used for medical purposes, the

manufacturers of the devices are required to obtain marketing clearance

from the Food and Drug Administration (FDA), in accordance with FDA

regulations that address issues such as quality assurance in

manufacturing, shelf-life and labeling. Currently, FDA does not assert

jurisdiction (provide marketing clearance) for alcohol screening

devices used for law enforcement purposes and workplace testing.

However, because of the nature of alcohol screening devices and the

conditions under which they are to be used, NHTSA stated in its

February 15 notice that it is important for manufacturers of these

devices to conform with certain requirements, imposed by FDA on devices

used for medical purposes, prior to the inclusion of the devices on

NHTSA's CPL.

Accordingly, NHTSA proposed to require that each device submitted

for testing under the Model Specifications be accompanied by a self-

certification from the manufacturer, certifying that it meets the

requirements contained in FDA's Good Manufacturing Practices

regulations for devices used for medical purposes (21 CFR Part 820),

and that the device's label meets the requirements contained in FDA's

Labeling regulations for devices used for medical purposes (21 CFR Part

809.10), even if the devices are not to be used for medical purposes.

NHTSA received a number of comments regarding this aspect of its

proposal. One commenter favored direct FDA regulation of all workplace

alcohol testing products and, if necessary, FDA enforcement. This

commenter encouraged DOT and NHTSA to continue their discussions with

FDA. Another commenter agreed that the guidelines written in FDA's Good

Manufacturing Practices regulation could be useful as a basis for

labeling and manufacturing requirements, but this and other commenters

recommended that FDA not get involved. According to one commenter,

``FDA is already overloaded, and long delays could result from their

involvement in this project.'' Another commenter recommended that, ``if

an instrument is not to be used in the medical field . . . FDA [should]

not assert jurisdiction.''

By requiring a self-certification, NHTSA was not proposing to

require that manufacturers obtain FDA marketing clearance, but simply

that the manufacturers self-certify that they meet the above-referenced

requirements. NHTSA stands by this aspect of its proposal.

For technical assistance or a copy of the Device Good Manufacturing

Practices Manual for Medical Devices, manufacturers should contact

FDA's Division of Small Manufacturers by calling toll free at 1-800-

638-2041.

NHTSA's February notice included, as an Appendix, a proposed set of

Labeling Instructions for Alcohol Screening Devices that had been

prepared in consultation with FDA to assist manufacturers of alcohol

screening devices in developing a label that conforms to 21 CFR Part

809.10. The labeling instructions addressed issues such as restrictions

that may apply to operators of the device and conditions under which

the device should or should not be operated.

One respondent commented on certain aspects of the labeling

instructions. The commenter supported the inclusion of details on

calibration, calibration frequency, and the manufacturer's name,

address, and telephone and fax numbers, but disagreed that an ``800''

number is necessary. In addition, the commenter stated that frequency

is subject to use, and some users will prefer to return a unit to the

manufacturer rather than engage in its calibration.

For the convenience of users, many of whom will be conducting

alcohol screening tests in the field, the Labeling Instructions for

Alcohol Screening Devices, which are included as an Appendix to today's

notice, continue to provide that manufacturers list an 800 number the

user may contact for further information or technical assistance. With

regard to the calibration of devices, the Labeling Instructions

continue to provide that disposable devices are pre-calibrated, and

need no additional calibration. They also continue to provide that

reusable devices require calibration, and instruct that the labels on

such devices provide information regarding how calibrations are to be

conducted, instructions for calibration and recalibration and the

criteria for acceptability of calibration.

These Model Specifications are not regulations. Organizations and

agencies may adopt these Model Specifications and rely on NHTSA's test

results or may conduct their own tests according to their own

procedures and specifications. It should be noted, however, that

transportation employers covered by 49 CFR Part 40, Procedures for

Transportation Workplace Drug and Alcohol Testing Programs, are

required to use only alcohol testing devices that meet the criteria

established by that regulation.

NHTSA intends to begin testing of alcohol screening devices

immediately, and hopes to publish a CPL of devices that have been

tested to date and conform to these Model Specifications within 30 days

from today's date. The CPL will be updated and published periodically,

as further testing is completed. Once the first CPL is published, DOT

will develop and issue procedural rules for using approved alcohol

screening devices in transportation workplaces, including provisions

for how and where such devices can be used and the steps that must be

taken to collect bodily fluids. Employers are reminded that these

screening devices are not authorized for use under 49 CFR Part 40 until

that regulation is amended.

Procedures

The procedures proposed in the February 15 notice have been adopted

without change. Testing of products submitted by manufacturers to these

Model Specifications will be conducted by the DOT Volpe National

Transportation Systems Center (VNTSC), DTS-75, Kendall Square,

Cambridge, MA 02142. Tests will be conducted semiannually, or as

necessary. Manufacturers are required to apply to NHTSA for a test date

by writing to the Office of Alcohol and State Programs (OASP), NTS-21,

NHTSA, 400 Seventh Street, S.W., Washington, D.C. 20590. Normally, at

least 30 days will be required from the date of notification until the

test can be scheduled.

One week prior to the scheduled initiation of the test program,

manufacturers will be required to deliver their devices to VNTSC. If

the devices are disposable, the manufacturer must deliver 300 such

devices; if the devices are disposable, interpretive and require that

results be checked through a visual inspection (and therefore must be

tested under various light conditions), the manufacturer must deliver

600 such devices; if the devices are reusable, the manufacturer must

submit only a single device. If a manufacturer of a reusable device

wishes to submit a duplicate, backup instrument, it may do so. The

manufacturer shall be responsible for ensuring that the devices operate

properly and are packaged correctly. The manufacturer must also deliver

the operator's manual (or instructions) and the maintenance manual (if

any) normally supplied with the purchase of the device, as well as

specifications and drawings which fully describe these devices.

Proprietary information will be respected. (See 49 CFR Part 512,

regarding the procedure by which NHTSA will consider claims of

confidentiality.)

In addition, the manufacturer must submit a self-certification,

certifying that the manufacturer meets the requirements in FDA's Good

Manufacturing Practices regulations for devices used for medical

purposes (21 CFR Part 820), and that the device's label meets the

requirements in FDA's Labeling regulations for devices used for medical

purposes (21 CFR Part 809.10), even if the devices are not to be used

for medical purposes. See the Appendix to this notice.

The manufacturer has the right to check its devices between the

time of their arrival at VNTSC and the start of the tests, but will

have no access to the devices during the tests. Any malfunction of a

device which results in failure to complete any of the tests

satisfactorily will result in a determination that the device does not

conform to the Model Specifications. If a device is found not to

conform, it may be resubmitted for the next testing series after

appropriate corrections have been made.

NHTSA plans to begin testing of alcohol screening devices

immediately to determine whether they comply with the performance

criteria included in the Model Specifications.

A Conforming Products List (CPL) will be updated and published

periodically. It will include a list of alcohol screening devices that

were submitted with the proper certifications and found to meet or

exceed the Model Specifications.

One commenter requested that manufacturers should be permitted to

commercialize their products as soon as they receive notification from

NHTSA that their product has been found to meet or exceed the Model

Specifications, rather than wait until the CPL listing their device is

published. NTSHA intends to notify manufacturers that their devices

meet the Model Specifications, and manufacturers may receive such

notices and an evaluation report prior to the publication of a CPL

listing their instrument. A decision about the point at which it would

be appropriate for manufacturers to commercialize their instruments,

however, is outside the scope of this notice.

NHTSA intends to modify and improve these Model Specifications as

new data and test procedures become available and to alter the test

procedures, if necessary, to meet unique design features of a specific

device. For each such modification, NHTSA would provide notification in

the Federal Register and would retest devices when necessary.

OASP is the point of contact for information about acceptance

testing and field performance of devices. NHTSA requests that users of

these devices provide both acceptance and field performance data to

OASP when such data are available. Information from users will help

NHTSA monitor whether alcohol screening devices are performing

according to the NHTSA Model Specifications.

If information gathered indicates that a device on the CPL is not

performing in accordance with the Model Specifications, NHTSA will

direct VNTSC to conduct a special investigation. An investigation may

include visits to users and additional tests of the device obtained

from the open market. If the investigation indicates that the devices

actually sold on the market are not meeting the Model Specifications,

the manufacturer will be notified that the device may be removed from

the list. In this event, the manufacturer will have 30 days from the

date of notification to reply. Based on the VNTSC investigation and any

data provided by the manufacturer, NHTSA will decide whether the device

should remain on the list. If the device is removed from the list, the

manufacturer will be permitted to resubmit an improved device to VNTSC

for testing when it believes the problems causing its failure have been

resolved. Upon resubmission, the manufacturer must submit a statement

describing what has been done to overcome the problems which led to

failure of the device.

If information gathered indicates that the manufacturer of a device

on the CPL does not comply with the requirements in FDA's Good

Manufacturing Practices regulations for devices used for medical

purposes or that the device's label does not comply with the

requirements in FDA's Labeling regulations for devices used for medical

purposes, NHTSA will investigate the matter in consultation with FDA

and will notify the manufacturer that the device may be removed from

the list. The manufacturer will have 30 days from the date of

notification to reply. Based on any data provided by the manufacturer

and investigative findings, NHTSA will decide whether the device should

remain on the list. If the device is removed from the list, the

manufacturer will be permitted to resubmit a self-certification,

certifying that the manufacturer complies with these FDA requirements

when it believes the problems causing its non-compliance have been

resolved. Upon resubmission, the manufacturer must submit a statement

describing what has been done to overcome the problems which led to

non-compliance.

This action has been analyzed in accordance with the principles and

criteria contained in Executive Order 12612 and it has been determined

that it has no federalism implication that warrants the preparation of

a federalism assessment.

In accordance with the foregoing, the Model Specifications for

performance testing of alcohol screening devices are set forth below.

Authority: 23 U.S.C. 402; delegations of authority at 49 CFR

1.50 and 501.

Michael B. Brownlee,

Associate Administrator for Traffic Safety Programs.

Model Specifications for Alcohol Screening Devices

1. Purpose and Scope

These specifications establish performance criteria and methods for

testing of alcohol screening devices. Alcohol screening devices use

bodily fluids to detect the presence of 0.020 or more BAC with

sufficient accuracy for screening purposes. These specifications are

intended primarily for use in the conformance testing of alcohol

screening devices.

2. Classification

2.1 Disposable Alcohol Screening Devices

Alcohol screening devices designed for a single use.

2.2 Reusable Alcohol Screening Devices

Alcohol screening devices designed to be reused.

3. Definitions.

3.1 Alcohol

The intoxicating agent in beverage alcohol, ethyl alcohol or other

low molecular weight alcohols including methyl or isopropyl alcohol.

3.2 Alcohol Screening Device

A device that is used to detect the presence of 0.020 or more BAC.

The device may measure any bodily fluid for this purpose, but shall

provide output in BAC units. Test results may be indicated by numerical

read-out or by other means, such as by the use of lights or color

changes.

3.3 Blood alcohol concentration (BAC)

Grams alcohol per 100 milliliters of blood or grams alcohol per 210

liters of breath in accordance with the Uniform Vehicle Code, Section

11-903(a)(5)\3\ (BrAC is often used to indicate that the measurement is

a breath measurement); or grams alcohol per 100 milliliters of saliva.

---------------------------------------------------------------------------

\3\Available from the National Committee on Traffic Laws and

Ordinances, 405 Church Street, Evanston IL 60201.

---------------------------------------------------------------------------

3.4 Calibrating Unit

A device that produces an alcohol-in-air test sample of known

concentration that meets the NHTSA Model Specifications for Calibrating

Units (49 FR 48865).

3.5 Breath Alcohol Sample Simulator (BASS)

A device that provides an alcohol-in-air test sample with known and

adjustable alcohol concentration profile, flow rate, and air

composition at 34 deg. centigrade. (See NBS Special Publication 480-41,

July 1981\4\ for a description of a BASS unit suitable for use in the

required testing.)

---------------------------------------------------------------------------

\4\Available from Superintendent of Documents, U.S. Government

Printing Office, Washington, D.C. 20402.

---------------------------------------------------------------------------

3.6 Bodily Fluid

Any bodily fluid capable of being used to estimate alcohol

concentration, provided the relationship between such bodily fluid and

BAC has been established according to scientifically acceptable

standards. Such fluids include but are not limited to blood, exhaled

deep lung breath and saliva.

3.7 Scientifically Acceptable Substitutes

Fluids that have been scientifically accepted as equivalent to

bodily fluids for testing purposes, such as aqueous alcohol test

solutions on a one-to-one basis for blood or saliva.

4. Test Methods and Requirements

Testing will be performed according to the instructions which

normally accompany the submitted device and under the conditions

specified in the tests below.

4.1 Test 1. Precision and Accuracy

Perform 40 trials under normal laboratory conditions using

fluorescent light, including 20 trials at 0.008 BAC and 20 trials at

0.032 BAC. Use the BASS device for breath devices and preparations of

bodily fluids or scientifically acceptable substitutes for non-breath

devices.

For disposable alcohol screening devices that indicate the presence

of alcohol in a manner that requires interpretation, combine Tests 1

and 2, in accordance with 4.3 below.

For alcohol screening devices that indicate the presence of alcohol

in a manner that does not require interpretation, perform the test

using a VNTSC investigator. To conform at 0.008 BAC, not more than one

positive result. To conform at 0.032 BAC, not more than one non-

positive result.

4.2 Test 2. Blank Reading

Perform 20 trials under normal laboratory conditions using

fluorescent light at 0.000 BAC. Use non-alcoholic human breath for

breath devices and preparations of non-alcoholic bodily fluids or

scientifically acceptable substitutes for non-breath devices.

For disposable alcohol screening devices that indicate the presence

of alcohol in a manner that requires interpretation, combine Tests 1

and 2, in accordance with 4.3 below.

For alcohol screening devices that indicate the presence of alcohol

in a manner that does not require interpretation, perform the test

using a VNTSC investigator. To conform, no positive results. If the

device is capable of providing a reading of greater than 0.000 BAC and

less than 0.020 BAC, not more than one such result.

4.3 Methodology for Combining Tests 1 and 2 for Disposable

Interpretive Devices

Perform the test under normal laboratory conditions using

fluorescent light using ten individuals who qualify as test

interpreters (according to the manufacturer's restrictions, if any) and

who have no knowledge of test BACs. Ask each individual to read the

manufacturer's instructions for interpretation of the device's read-

out.

Label sixty devices from 1 to 60 and randomly separate them into

three groups of twenty. Record the numbers in each group. Use two of

the groups of devices for Test 1 and the remaining group for Test 2.

Dose each group at the BAC levels specified in Tests 1 and 2. Order the

sixty devices into a single set from 1 to 60 and ask each individual to

independently interpret the results of these trials.

Ask each individual to record each result as being one of the

following: ``at .00 BAC''; ``above .00 and below. 02 BAC''; ``at or

above .02 BAC''; or ``can't tell''. Dosing of devices and

interpretation of results will be accomplished within a twenty minute

period.

To conform, with each interpreter, no positive results at .000 BAC,

not more than one positive result at .008 BAC, not more than one non-

positive result at .032 BAC and not more than one ``can't tell''

result. If the device is capable of providing a reading of greater than

0.000 BAC and less than 0.020 BAC, with each interpreter, not more than

one such result at .000 BAC.

4.4 Test 3. Light Conditions (only interpretive devices which require

that results be checked through a visual inspection)

Perform Tests 1 and 2, in accordance with 4.3, under each of the

following light conditions: incandescent light; mercury vapor light;

sodium vapor light; and daylight.

Under each light condition, the device must meet the criteria

established in 4.3: To conform, with each interpreter, no positive

results at .000 BAC, not more than one positive result at .008 BAC, not

more than one non-positive result at .032 BAC and not more than one

``can't tell'' result. If the device is capable of providing a reading

of greater than 0.000 BAC and less than 0.020 BAC, with each

interpreter, not more than one such result at .000 BAC.

4.5 Test 4. Cigarette smoke interference (only breath and saliva test

devices)

Perform five trials at 0.000 BAC. Select an alcohol-free person who

smokes cigarettes for this test. Ask the person selected to smoke

approximately one half of a cigarette. Within one minute after smoking,

or after a waiting period specified in the manufacturer's instructions,

administer the alcohol screening device test according to the

manufacturer's instructions. Then ask the person to smoke another

inhalation and repeat the test to produce a total of five trials.

To conform, no positive results.

4.6 Temperature

Test at low and high ambient temperature.

4.6.1 Test 5.1 Low Ambient Temperature

Perform 40 trials at 10 deg.C, including 20 trials at 0.008 BAC

and 20 trials at 0.032 BAC. Use a calibrating unit for this test for

breath devices and preparations of bodily fluids or scientifically

acceptable substitutes for non-breath devices.

To conform at 0.008 BAC, not more than one positive result. To

conform at 0.032 BAC, not more than one non-positive result.

4.6.2 Test 5.2 High Ambient Temperature

Perform trials of 40 devices at 40 deg.C, including 20 trials at

0.008 BAC and 20 trials at 0.032 BAC. Use a calibrating unit for this

test for breath devices and preparations of bodily fluids or

scientifically acceptable substitutes for non-breath devices.

To conform at 0.008 BAC, not more than one positive result. To

conform at 0.032 BAC, not more than one non-positive result.

4.7. Test 6. Vibration

Perform 40 trials, including 20 trials at 0.008 BAC and 20 trials

at 0.032 BAC. Use a calibrating unit for this test for breath devices

and preparations of bodily fluids or scientifically acceptable

substitutes for non-breath devices.

Mount the screening device on a shake table and vibrate the table

in simple harmonic motion through each of its three major axes, as

specified below. Sweep through each frequency range in 2.5 minutes,

then reverse the sweep to the starting frequency in 2.5 minutes. The 40

disposable testers may be placed in a suitable box mounted on the shake

table. Test after vibration.

------------------------------------------------------------------------

Amplitude

(inches,

Frequency (hertz) peak to

peak)

------------------------------------------------------------------------

10 to 30................................................... 0.30

30 to 60................................................... 0.15

------------------------------------------------------------------------

To conform at 0.008 BAC, not more than one positive result. To

conform at 0.032 BAC, not more than one non-positive result.

Appendix

Labeling Instructions for Alcohol Screening Devices Intended Use

Provide the intended use including the specimen matrix (e.g.

saliva, breath), the assay type (quantitative, semi-quantitative)

the purpose of performing the assay and the individual designated to

perform the assay.

e.g. This product is intended for the (quantitative,

semiquantitative) determination of alcohol in--define matrix (for

e.g., saliva, breath, sweat) to perform screening alcohol assays.

This product is recommended for use by individuals who have been

trained in the administration of screening devices.

Description of Testing System

Provide the principles of the procedure for performing the

alcohol screening assay.

e.g. This product uses alcohol dehydrogenase, infrared

technology, etc. to perform the test.

Chemical Reaction Sequence

Describe the chemical reaction sequence, if applicable.

Reagents

List the concentration, strength, composition of the reactive

ingredients.

List the non-reactive ingredients.

Reagent Preparation and Storage

Provide instructions for preparing the reagents, if appicable.

Provide instructions for storing the reagents, if applicable.

Provide any signs of deterioration of the reagents, if

applicable.

Provide the reagent's shelf life and opened expiration dating,

if applicable.

e.g. Unopened tests are stable until the date printed on the

product container when stored at 22-28 deg.C. Opened test must be

used at once.

Provide a caution not to use the reagents beyond the expiration

dating.

Precautions:

1. List any reagents that may be hazardous such as caustic

compounds, sodium azide or other hazardous reagents and instructions

for disposal, if applicable.

2. If visually read, warn the user the result should not be

interpreted by readers who are color-blind or visually impaired.

3. Provide warning to user to treat all samples as potentially

infective. Include instructions for handling and disposal of the

sample.

Speciment Collection

Provide instructions for collecting and handling the sample.

Provide criteria for specimen rejection, if applicable.

Calibration

Disposable tests are pre-calibrated. No additional calibration

is required.

Reusable (Instrumented) tests require calibration.

Provide information regarding how calibrations are to be

conducted, if applicable, including the number and concentration of

calibrators, and the frequency of calibration.

Provide instructions for calibration and recalibration.

Provide the criteria for acceptabiity of calibration.

Test Procedure (Disposable)

Provide adequate step-by-step instructions for performing the

test.

If the test is disposable (non-instrumented) and involves a

color reaction, include the time frame for which the test must be

read and recorded.

e.g. Read within 15 minutes.

Test Procedure (Reusable/Instrumented)

Provide adequate step-by-step instruction for performing the

test.

Provide the installation procedures and, if applicable, any

special requirements.

Provide the space and ventilation requirements.

Provide the description of the required frequency of equipment

maintenance and function checks.

Provide the instructions for any remedial action to be taken

when the equipment performs outside of operating range.

Provide any operational precautions and limitations.

Provide instructions for the protection of equipment and

instrumentation from fluctuations or interruptions in electrical

current that could adversely affect test results and reports, if

applicable.

Quality Control (QC)

Disposable Tests

If applicable, the function and stability of the test can be

determined by examination of the procedural ``built in'' controls

contained in the product. If these controls are not working, the

test is invalid and must be repeated.

Disposable/Instrumented Devices

If external quality control materials are used, provide number,

type, matrix and concentration of the QC materials.

Provide directions for performing quality control procedures.

Provide an adequate description of the remedial action to be

taken when the QC results fail to meet the criteria for

acceptability.

Provide directions for interpretation of the results of quality

control samples.

Results

Describe how the user obtains the test results, from a colored

bar, instrument read-out, printout, etc.

Describe the results in terms of blood alcohol concentration.

Describe what concentration indicates a positive result and what

concentration indicates a nagative result.

Limitations

List the substances or factors that may interfere with the test

and cause false results including technical or procedural errors.

Dynamic Range

Provide the operating range of the product.

Precision and Accuracy

Precision and Accuracy specifications are included in the

National Highway Traffic Safety Administration's (NHTSA's) Model

Specifications for Alochol Screening devices. Only devices that meet

these model specificiations will be included on NHTSA's Conforming

Products List for alcohol screening devices.

Specificity

List the substances that have been evaluated with your product

that do or do not interfere at the concentration indicated.

References

Provide pertinent bibliography

Technical Assistance

List an 800 number the user may contact for further information

or technical assistance.

[FR Doc. 94-18750 Filed 7-28-94 4:30 pm]

BILLING CODE 4910-59-M