Chapter 5: Breath and Blood Testing

I.   Overview

§5.1   As part of nearly every drunk driving stop, law enforcement will request a chemical test. These tests can be to assist law enforcement in determining whether to arrest a citizen (the preliminary breath test) or to use as evidence in a trial (the Intoxilyzer 9000 and blood tests). It is important to understand each of these tests, their purposes, and their limitations.

II.   Preliminary Breath Tests

§5.2   Before an arrest, the police will usually request that a suspect take a preliminary breath test (PBT). MCL 257.625a(2) provides that “[a] peace officer who has reasonable cause to believe that a person” was operating while intoxicated may require the person to submit to a “preliminary chemical breath analysis.”

Before administering the PBT, a peace officer must “determine” that a person has not smoked, regurgitated, or placed anything in their mouth in the 15 minutes before the PBT. AC, R 325.2657a(1)(b). This is a different standard from the deprivation and observation period required by the Intoxilyzer 9000 (see §5.4). Although an officer need not observe a suspect for 15 minutes, the officer must at least inquire of the subject when they last smoked, regurgitated, or placed anything in their mouth. Failure to do so can result in the PBT not being considered when determining whether probable cause to arrest existed. In People v Robe, 337 Mich App 142, 972 NW2d 52 (2021), defendant’s PBT could not be considered in determining whether there was probable cause to arrest because defendant was observed for only 3 minutes before being tested instead of the 15 minutes required by the administrative rule. Of equal importance is that Robe was the first published opinion after People v Wujkowski, 230 Mich App 181, 583 NW2d 257 (1998), to hold that violation of the administrative rule, without more, was sufficient to call into question the reliability of the breath test. (See §5.9 for more on Wujkowski.)

A peace officer may arrest a person based in whole or in part on the PBT results, but the PBT has limited admissibility at trial. It is important to note that refusal to take the PBT will not result in any license sanctions, as it is not part of the implied consent law (discussed in chapter 12). Rather, refusal to take the PBT is a civil infraction punishable only by a fine for most drivers. MCL 257.625a(2)(d). Refusal of the PBT by an operator of a commercial motor vehicle is a misdemeanor, punishable by not more than 93 days’ imprisonment or a fine of up to $100 or both, and will result in a 24-hour out-of-service order. MCL 257.625a(4), (5). Drivers under the age of 21 who refuse the PBT will have two points placed on their driving record in addition to incurring the fine. MCL 257.320a(1)(v).

The results of a PBT are admissible in court only in the following circumstances:

1. To assist the court in determining a challenge to the validity of the arrest.
2. “As evidence of the defendant’s breath alcohol content, if offered by the defendant to rebut testimony elicited on cross-examination of a defense witness that the defendant’s breath alcohol content was higher at the time of the charged offense than when [an evidentiary] chemical test was administered.” This situation arises when the prosecutor attempts to use retrograde extrapolation. See §5.36.
3. “As evidence of the defendant’s breath alcohol content if offered by the prosecution to rebut testimony elicited on cross-examination of a prosecution witness that the defendant’s breath alcohol content was lower at the time of the charged offense than when [an evidentiary] chemical test was administered.” This situation arises in a case with a rising blood alcohol defense. See §5.37.

MCL 257.625a(2)(b)(i)–(iii).

These limitations on the use of PBTs in court do not impede a defendant’s constitutional right to present a complete defense. People v Parrott, 335 Mich App 648, 968 NW2d 548 (2021). In Parrott, the court also held that to the extent MCL 257.625a(2)(b) irreconcilably conflicts with MRE 702 and 703, the statute prevails. Consequently, a defendant does not have a right to present favorable PBT information unless the prosecutor opens the door by arguing that a defendant’s blood alcohol content (BAC) was higher at the time of driving that at the time of the evidentiary breath or blood test.

It should be noted that the term “preliminary chemical breath analysis” is broader than the formerly used “preliminary breath test” and allows law enforcement to request a preliminary breath test for not only alcohol but also controlled substances or other intoxicating substances. MCL 257.43a. However, as of March 2025, the author is unaware of any current breath testing technology that can discern the presence of a controlled substance or any other intoxicating substance other than alcohol. Although there are companies working to develop a breath test for marijuana, no device has yet been approved or placed on the market.

PBTs use either fuel cell technology or infrared spectroscopy to obtain their readings. MCL 257.625a(6)(g) requires the Michigan State Police (MSP) to set rules for the administration of PBTs and the approval of PBT equipment, which can be found at AC, R 325.2651 et seq. (appendix A). In addition to minor renumbering and rephrasing, the rules contain new testing procedures for PBTs, and the classifications for various certifications have changed. Attorneys are encouraged to read the new rules carefully. Those rules require all PBT instruments to be approved by the MSP and require any department wishing to use a PBT instrument to apply for approval. AC, R 325.2652(3), (4).

PBTs must undergo an accuracy test at least monthly, AC, R 325.2656a, and a class II, III, IVA, or IVB operator must perform the accuracy test, AC, R 325.2656a, .2658(7), table 3. The monthly testing requirement remains in place, but the testing procedure and classifications have changed. See generally AC, R 325.2657a–.2658. The agency performing the accuracy tests is supposed to maintain a record of them and forward those test records to the MSP. Defense attorneys should make a Freedom of Information Act (FOIA) request for the test result records from one or both agencies. In addition, the MSP has published its 2018 PBT training manuals online.

III.   Evidential Breath Tests

A. The Intoxilyzer 9000

§5.3   After an arrest, law enforcement will almost always request a citizen to submit to an evidential breath or blood test. In June 2023, Michigan began using the Intoxilyzer 9000, produced by CMI, Inc., based in Owensboro, Kentucky. The Intoxilyzer 9000 was developed in 2012 and is used in Texas, Georgia, Colorado, and other states. Although there are many similarities, the technology used in the Intoxilyzer 9000 differs in several ways from the DataMaster DMT, which was used in Michigan until the switch to the Intoxilyzer 9000. While the Intoxilyzer 9000 continues to use infrared spectroscopy, it employs pulsed spectroscopy to measure the alcohol content. It also uses different technology for its sample chamber and alcohol detection systems than those used by the DataMaster DMT. The MSP has published a truncated Intoxilyzer 9000 Operator Guide for Law Enforcement (Nov 2023) on its website. For comparison, the DataMaster DMT manual was 50 pages, not including the administrative rules, while the new Intoxilyzer 9000 manual is a brief 19 pages.

There are new procedures in the amended administrative rules, as discussed throughout this chapter, to reflect the use of this new instrument. Under the amendments, a weekly test is not required for the Intoxilyzer 9000. AC, R 325.2653(2). However, if an accuracy test is performed, the results must be retained within the instrument’s memory. Id. Furthermore, the instrument will no longer require an inspection for accuracy every 120 days but instead will be subject to inspection at least twice annually. AC, R 325.2653(4).

B. Operation of the Intoxilyzer 9000

1. 15-Minute Observation Period

§5.4   It has always been a requirement that the testing officer observe the subject for “not less than” 15 minutes before administering the evidential breath test to ensure that the subject has not smoked, regurgitated, or placed anything in their mouth. This remains the case with the Intoxilyzer 9000. Although the Intoxilyzer 9000 can be programmed to lock out the officers during the 15-minute waiting period, the MSP chose not to activate this feature purportedly because if there were multiple subjects in line to be tested, each officer could be performing the observation period while another subject is being tested. Locking them out automatically would require officers to wait at least 15 minutes from their turn to use the machine. The decision to not provide an automatic lockout is curious because the Intoxilyzer 9000 law enforcement training materials note that “[t]he 15-minute observation period is one of the most challenged procedures of an OWI arrest by defense attorneys. DO IT RIGHT!”

Another curious change is that, while the Datamaster DMT required the officer to manually type in the starting time of their observation, the Intoxilyzer 9000 simply asks the yes/no question of whether the 15-minute rule has been complied with. It does not require the officer to enter a time for the start of the observation period. By not requiring the entry of the start time, the Intoxilyzer 9000 also eliminates the incentive to ensure that the observation start time is taken from the machine itself rather than being an arbitrary time before entering the breath-testing room. This is despite the following notation in the training materials: “Best practice: Use the clock on the breath instrument to track your 15-minute observation period.”

Defense counsel should review body camera footage or video from the breath-testing room, if available, to ensure this rule was complied with. The observation and deprivation requirement is discussed more fully in §5.9.

2. Officer’s Identifying Information

§5.5   The Intoxilyzer 9000 requires police officers to enter their identifying information, including their name and badge number. Check each of these carefully to determine whether the test was given by the arresting officer or another officer. The MSP does not issue law enforcement operator certificate numbers specifically for breath testing. Instead, an officer enters their Michigan Commission on Law Enforcement Standards (MCOLES) number to identify themselves, and MCOLES keeps track of officer certifications. This information provides the defense attorney with the opportunity to make a FOIA request for the officer’s training history from MCOLES to ensure that the officer is qualified to administer the Intoxilyzer 9000 and has an updated certification. This should be a routine part of every defense attorney’s practice before evidentiary hearings and trials. See form 5.1 for the MCOLES request form. Obtaining the training history will start to become even more important in the later half of 2025 because Intoxilyzer certifications are good for only two years, after which officers must renew them. The MSP has not stated whether officers will have to retake the entire Intoxilyzer training or take only a refresher course, but there will be some requirement to renew the training certification every two years.

3. Subject’s Identifying Information

§5.6   With Michigan licenses now having barcodes and magnetic stripes with the driver’s identifying information, officers do not have to enter the subject’s information into the machine. The Intoxilyzer 9000 has a barcode reader and a magnetic stripe reader that officers can use to simply scan the licenses.

4. Testing Sequence

§5.7   The Intoxilyzer 9000’s testing sequence includes a control test. This testing sequence is as follows:

• operation system test
• air blank
• subject sample
• air blank
• dry-gas calibration test at 0.08%
• air blank
• rest
• air blank
• subject sample
• air blank
• operation system test
• air blank

The Intoxilyzer 9000 runs internal diagnostic tests to make sure it is in working order. In addition, it runs air blanks to make sure it purges any contaminants in the breath test tube. The air blanks cause the breath tube to suck in the ambient air in the room. If it detects other sources of alcohol during that process, it is supposed to abort the test and notify the officer to start over.

In a major change from the DataMaster DMT, the Intoxilyzer 9000 does not run automatic weekly dry-gas calibration checks at 0.08 and keep a log of those tests. Instead, the Intoxilyzer 9000 runs a dry-gas calibration test targeting 0.08 between subject samples for each person who takes an Intoxilyzer 9000 test. Because the Intoxilyzer 9000 runs the dry-gas test between each subject sample, the MSP is not required to perform weekly dry-gas calibration tests as was required of the DataMaster DMT. AC, R 325.2653(2). Thus, as there is no weekly log of the results of calibration tests, defense attorneys who wish to see a “log” of calibration checks will have to ask for each subject test taken in the weeks and months before and after the subject test in question.

The Intoxilyzer 9000 also requires a full calibration check twice per calendar year, rather than every 120 days as was the case with the Datamaster DMT. AC, R 325.2653(4). There is no requirement that these two tests be 180 days apart. They can be on consecutive days or even 364 days apart. For example, if the MSP performs a calibration check on January 1 and there is a problem with the machine on January 15 that requires the MSP to fix the machine and then run additional calibration checks, the MSP will have fulfilled its twice-per-year requirement with the January calibration checks. Each subject test sequence (and each calibration check done by a law enforcement department) “must be retained either in log form by the agency where the instrument is installed or electronically within the instruments memory.” AC, R 325.2653(2).

C. Second and Subsequent Tests

§5.8   The administrative rules require the officer to offer a second test to verify the results of the first test sequence. However, after an individual submits to the first test sequence, it is not considered a violation of the implied consent law to refuse the second test sequence. AC, R 325.2655(f). In fact, even if the citizen submits to the second test sequence, if the test is not completed for some reason, there is no requirement that the officer offer a third test sequence. People v Fosnaugh, 248 Mich App 444, 639 NW2d 587 (2001). In Fosnaugh, defendant challenged the admissibility of the breath test results (from the previously used DataMaster) because the second test resulted in a reading of “invalid sample” and a third test was not administered. The court noted that the administrative rules required a third test only if the second test varied from the first test, not if mouth alcohol was present in the second test. Thus, the lack of a confirming test result went to the weight, not the admissibility, of the first test result. 248 Mich App at 452. All of the parties assumed that “invalid sample” meant that mouth alcohol was present in the second sample, although there is nothing in the record as reported by the court of appeals that substantiates this. An invalid sample may result from mouth alcohol, but it could also be caused by a poor flow rate. The Intoxilyzer 9000 Operator Guide for Law Enforcement, at 9 (Nov 2023), states that, when a breath sample is rejected, “the correct procedure is to follow the on-screen instructions which will instruct the operator to conduct another test, or in other instances, to seek a blood sample.”

If the first two test results are not within the preset tolerances of each other, the Intoxilyzer 9000 will automatically request a third sample, and the sequence will be the same as the first two sequences. Although there is nothing in the administrative rules, the Intoxilyzer 9000 manual, or the written training materials requiring a new mouthpiece for each test, law enforcement is taught to use a new mouthpiece for each test sequence.

The Intoxilyzer 9000 reports the results to the third digit (i.e., 0.080). The administrative rules set the tolerance requirements as follows:

AC, R 325.2655(g).

Thus, the Intoxilyzer 9000 is allowed to vary by as much as 12.5 percent above or below a reading of 0.08 (readings of 0.07 to 0.09 are acceptable to confirm a 0.08) and as much as 13.33 percent above or below a reading of 0.15 (readings of 0.13 to 0.17 are acceptable to confirm a 0.15). This means that the administrative rules allow a total tolerance of 25 percent for a 0.08 and 26.67 percent for a 0.15, numbers that would never be acceptable in the scientific community.

D. Sources of Operator Error

1. In General; 15-Minute Deprivation and Observation Period

§5.9   The Intoxilyzer 9000 has a fairly automated process designed to keep the chances of operator error to a minimum. However, there are still many ways an officer can improperly administer the test sequence.

The admissibility of evidential breath test results at trial is governed by MCL 257.625a(6)(g), which requires the MSP to promulgate rules for the administration of breath tests, which can be found at AC, R 325.2651–.2659 (appendix A). AC, R 325.2655(e) states:

An individual shall be administered an evidential breath alcohol analysis on an evidential breath alcohol test instrument only after being observed for not less than 15 minutes before collection of the breath sample by at least 1 appropriate class operator that is certified in accordance with R 325.2658. The observation period may be conducted by more than 1 operator working in concert. During the observation period, the individual shall not have smoked, regurgitated, or placed anything in his or her mouth, except for the mouthpiece associated with the performance of the evidential breath alcohol analysis. The operator need not stare continuously at the individual, but shall be close enough to be aware of the individual’s actions and conditions. The operator may complete paperwork, enter data into the evidential breath alcohol test instrument, or conduct other reasonable tasks during the observation period, if the individual is within the operator’s field of vision. Breaks in the observation period lasting only a few seconds do not invalidate the observation if the operator can reasonably determine that the individual did not smoke, regurgitate, or place anything in his or her mouth during the breaks in the observation.

Dr. Kurt M. Dubowski, a pioneer and recognized authority in breath alcohol testing, notes that the scientific safeguard of a minimum 15-minute pretest deprivation-observation period is particularly important. The 15-minute deprivation-observation period is considered indispensable in forensic breath alcohol measurement and is endorsed and recommended by the National Safety Council Committee on Alcohol and Other Drugs. Kurt M. Dubowski, Quality Assurance in Breath-Alcohol Analysis, 18 J Analytical Toxicology 306 (Oct 1994). Every OWI defense attorney should obtain and read a copy of Dr. Dubowski’s article.

It is important to note that officers can tag team the deprivation-observation time period as long as each officer is certified as an operator in the class required for the particular breath machine. And, although Dr. Dubowski believes otherwise, minor lapses in the observation period will not necessarily require suppression of the breath test. There are two cases that set the outer boundaries for lapses in the observation period.

At the low end, not requiring suppression of the breath test results, is People v Wujkowski, 230 Mich App 181, 583 NW2d 257 (1998). In that case, the administrator failed to observe defendant for six seconds. The court found that lapse to be harmless. Indeed, defendant in Wujkowski did not even argue that there was a problem with the DataMaster DMT (predecessor to the Intoxilyzer 9000) test. Instead, he simply argued that the police did not follow the administrative rules, so the test should be suppressed. The Wujkowski court reminded defense counsel that statutory violations do not require suppression unless the violation rises to the level of a constitutional error. However, the Wujkowski court acknowledged that the more severe the violation, the less reliable the test results are, and the more likely suppression is the appropriate remedy. 230 Mich App at 188.

Indeed, Wujkowski compared its six-second lapse, which it viewed as harmless, to the eight-minute observation in People v Boughner, 209 Mich App 397, 398, 531 NW2d 746 (1995) (citing People v Tipolt, 198 Mich App 44, 46, 497 NW2d 198 (1993)), in which the court quashed the test results, noting that the purpose of the administrative rules is to ensure the accuracy of those tests. Wujkowski acknowledged that the test in Boughner was “sufficiently questionable so as to preclude the test results from being admitted.” Wujkowski, 230 Mich App at 188. In addition, although it involved the 15-minute deprivation rule for PBTs, People v Robe, 337 Mich App 142, 972 NW2d 52 (2021), ruled that a 3-minute observation was not sufficient for the PBT and quashed defendant’s arrest, which had been based largely on the PBT results.

2. Failure to Replace the Mouthpiece Between Tests

§5.10   The administrative rules, the Intoxilyzer 9000 operator guide, and the written law enforcement training manual are all silent on whether a new mouthpiece is required for each test, but law enforcement officers are told during their training on the Intoxilyzer 9000 to use a new mouthpiece for each test sequence.

E. Other Sources of Error

1. In General; Failures in Accuracy Checks

§5.11   Most other sources of error result from issues inherent in infrared spectroscopy breath testing itself, which is discussed in §§5.13–5.17. However, there are several other areas of human error to explore.

The previously used Datamaster DMT required a weekly dry-gas test to check the accuracy of the machine at 0.08. However, the Intoxilyzer 9000 instead runs a dry-gas check within each subject sequence. Because of this, defense counsel must request copies of tests run before and after the client’s test to ensure they were all within tolerance. This is known as the COBRA data. The personal information of the test takers is replaced with “Subject Test,” but the results of each individual test that was run, including the dry-gas test run in between sequences and any exception messages, are displayed. A sample of the COBRA data is shown in exhibit 5.1. This is a more difficult way to review the data because it requires defense counsel to sift through many more subject tests. However, it also allows defense counsel to spot trends among the data provided by the actual test sequences. For example, in the sample COBRA data in exhibit 5.1, there are exception messages from August 7, 2024, where it appears the test subjects were having difficulty providing samples, and on two occasions that day, the officer stopped the test. Eventually, defense counsel should be given remote access to the COBRA database without having to resort to a FOIA request. Unfortunately, it will likely be at least a couple of years before that access is granted, as the MSP is still working out the logistics of tying all the machines together in one network so that the data can be automatically centrally stored. Because of this, defense counsel will have to obtain information from both the MSP and the local police agency because there is not currently a central repository for Intoxilyzer 9000 tests.

In addition to the dry-gas test being run in between subject tests, the MSP must also perform twice-yearly accuracy tests. Those tests are to be performed by an appropriate class officer, and the results must be signed by the person performing the accuracy test and forwarded to the MSP, which keeps records of them. These test results should be requested from the MSP and reviewed carefully to ensure that the tests were properly conducted, that they used approved solutions, and that the machines were performing correctly. The date of the most recent maintenance test is listed on the evidence ticket. Those test are part of the foundation the prosecutor must lay to ensure that the Intoxilyzer 9000 test results are admissible.

In September 2020, the court of appeals held that the testing logs (which were kept when the Datamaster DMT was the machine used) were not testimonial and not subject to the Confrontation Clause. People v Fontenot, 333 Mich App 528, 963 NW2d 397 (2020), vacated in part, 509 Mich 1073, 975 NW2d 919 (2022). The court also ruled that the testing logs were admissible under the business records exception to the hearsay rule, MRE 803(6). In 2022, the Michigan Supreme Court held oral arguments on the application for leave to appeal and, rather than granting leave, vacated part II(C) of the court of appeals opinion. This means that it did not address the Confrontation Clause issue, so the current state of the law in Michigan is that breath test calibration and control logs are not testimonial and can be introduced as business records if they satisfy the court of the trustworthiness of the records. The supreme court vacated the portion of the court of appeals opinion that held that the trustworthiness of the records goes to the weight of the evidence, not the admissibility. The court made it clear that if the courts have some evidence that the records are not trustworthy, it can consider that evidence in determining whether the records should be admitted.

Because weekly logs are no longer being kept, it is an open issue whether all the accuracy tests will likewise be admissible as a business record, given that the documents showing the results of each individual test included in the COBRA data will have been prepared specifically for going to court in that case. This is an area of the law that will need to be developed.

One case that was decided after four days of evidentiary hearings is People v Aubrey, No 2019-004326-OD (52-4 Dist Ct Dec 23, 2020). There, the judge acknowledged that she was bound to follow Fontenot. However, after determining that the data from the evidential breath test in question were not credible due to fraud in the testing procedure, she struck a middle ground by allowing the testing logs to be admitted without a witness but suppressing the evidential breath test as unreliable. The full opinion is attached as exhibit 5.2.

As noted, the Intoxilyzer 9000 is subject to accuracy testing procedures, albeit only twice per calendar year. The results of these tests are contained in a document titled “FS-165 Intoxilyzer 9000 Inspection Worksheet.” Each worksheet contains the instrument number, along with other information such as the barometric pressure on the date of the test, information on the dry-gas cylinder, and a checklist of tests performed on the machine to ensure its accuracy. Those test results are accompanied by the actual wet-bath test result printouts called the “Intoxilyzer—Alcohol Analyzer,” which look similar to subject sample printouts, and a “Michigan Forensic Breath Alcohol Analytical Report,” which concerns the dry-gas test that is run in between two “air blank” tests. All these documents should be scrutinized to ensure the accuracy of each instrument.

2. Radio Frequency Interference

§5.12   One of the purposes of the twice-annual inspections is to verify that the radio frequency interference (RFI) detector is working. Establishing that there was not any RFI is a problem for the prosecution because, if the RFI detector is not working, the machine will report no RFI detected even if there is RFI in the room. The only way to truly know whether there is RFI in the breath testing room is to have an electrical engineer take independent measurements at the time of testing. Of course, replicating all the sources of RFI at the time and date the breath test is being administered is nearly impossible, given all the things that can cause RFI, like cell phones and police radios.

Although the Intoxilyzer 9000 is enclosed in a metal box to, in theory, protect against the penetration of RFI, it is not without potential vulnerability. For instance, it has a number of USB ports on its back and side that may act like antennas. Although there has not been a study performed, based on this author’s experience, it appears that RFI is a big problem at many law enforcement locations, as there have been many instances of a machine reporting RFI. Many times, the printouts report RFI for one of the subject tests, but not the other. So even though the twice-annual tests include an RFI check, that may not be sufficient to ensure that no RFI is present on the date of a subject test.