Opinion

The Department of Motor Vehicles (DMV) suspended Joseph Freitas’s driver’s license after it found he violated Vehicle Code section 13353.2 by driving with a blood-alcohol concentration (BAC) of 0.08 percent or more. The suspension was upheld after an administrative hearing. Freitas appeals from the trial court’s denial of his petition for a writ of mandate to overturn the hearing officer’s decision.

The trial court’s ruling was based on its rejection of the unrebutted testimony of Freitas’s expert, who opined that the blood-testing procedure used to measure Freitas’s BAC was scientifically invalid. The court’s conclusions on this point are inconsistent with our holding in Najera v. Shiomoto (2015) 241 Cal.App.4th 173 [ 193 Cal.Rptr.3d 596 ] (Najera). We will reverse the judgment.

FACTS AND PROCEDURAL HISTORY

According to the police report, Officer Bright of the Department of the California Highway Patrol saw Freitas’s truck speeding at 10:10 p.m. on July 13, 2013. The truck was weaving between lanes and moving at 65 miles per hour on a street with a 45-mile-per-hour speed limit. It continued after Bright turned on his siren and flashing lights. Instead of stopping, the truck made three turns and then pulled into Freitas’s driveway. It rolled backward in the driveway. Finally it stopped and Freitas got out. He stood unsteadily, smelled strongly of alcohol, and had red eyes and slurred speech.

Bright administered four field sobriety tests. Freitas performed poorly. On the horizontal gaze nystagmus test, he had difficulty keeping his eyes on the stimulus. On the Romberg balance test, his 30-second time estimate was close to correct, at 31 seconds, but he swayed three to four inches side to side as well as forward and backward. During the one-leg-stand test, he was very unsteady. He swayed, dropped his foot twice, and hopped and stumbled twice. At that point, he admitted he had had too much to drink to be driving. During the walk-and-turn test, Freitas failed to keep his balance while standing heel to toe as Bright gave the instructions for the test. Then he counted 10 steps while taking nine, failed several times to touch his heel to his toe, did not keep his steps in line, and made the turn improperly. At some point during his encounter with Bright, Freitas admitted he had drunk four beers at a restaurant beginning at 4:00 p.m. and finishing 30 minutes before the stop.

Bright arrested Freitas for driving under the influence and advised him pursuant to Vehicle Code section 23612 that, by driving, Freitas had given his implied consent to a breath or blood test and could choose one or the other. Freitas chose the blood test and was transported to Kern Medical Center. A nurse drew two blood samples at 10:54 p.m.

At the time of the arrest, Bright served Freitas with a DMV form ordering the suspension or revocation of his license. The form advised him of his right to an administrative hearing.

Freitas’s blood samples were tested at the Kern Regional Crime Laboratory on July 17 and 18, 2013. They were found to have a BAC of 0.23 percent.

Freitas requested a hearing, which commenced on September 18, 2013, and was continued to July 28, 2014. The DMV submitted a sworn statement by the arresting officer, the police reports, the blood test report, and Freitas’s driving record.

Freitas’s expert, a chemist named Janine Arvizu, testified about the procedures used by the Kern Regional Crime Laboratory to test Freitas’s blood samples. She said the laboratory used a technique called gas chromatography to test for alcohol in a sample. The testing apparatus, a machine called a gas chromatograph, has a heated chamber containing two long, narrow, coiled columns. The inner surfaces of the columns are treated with chemical preparations, a different preparation in each column. A portion of the sample to be tested is introduced into each column in gaseous form. As a sample passes through a column, compounds in the sample react with the chemicals on the walls of the column. The reactivity of the compounds in the sample determines the length of time it takes for those compounds to reach the end of the column. Compounds that interact more take longer. When a hydrocarbon such as alcohol reaches the end of the column, it is combusted by a device called a flame ionization detector. The length of time from insertion of the sample to combustion at the detector, known as the retention time, is an indication of the identity of the compound.

The reason for using two differently prepared columns is that for any given single column, the retention time for alcohol is the same as the retention time for numerous other volatile organic compounds that can be present in a blood sample. Data from a single column consistent with the presence of alcohol would also be consistent with the presence of a different compound or alcohol plus another compound. A sample yielding a positive result from a single column thus might contain no alcohol or might contain less alcohol than the result indicates. Results from the second column, which are based on a different chemical principle, are necessary to confirm the presence and quantity of alcohol.

Arvizu testified that the necessity of confirmation of a BAC result by the second column was widely recognized in the scientific community. All literature she had reviewed on the subject agreed that gas chromatography cannot determine the presence and quantity of alcohol in a blood sample without the use of two columns. Instructions from the manufacturers of the gas chromatograph and the columns used by the Kern Regional Crime Laboratory also stated that results from two columns must be used for measuring BAC. One column can “tentatively identify” alcohol, but “simply cannot confirm its identity” or “how much might be present.”

Arvizu testified that, although the Kern Regional Crime Laboratory used a dual-column gas chromatograph with two columns installed, the test results issued by the lab reflected data from a single column only. This meant the results were not reliable. Arvizu’s testimony on this point included the following exchange with Freitas’s counsel:

The DMV presented no evidence in rebuttal to Arvizu’s testimony.

The hearing officer’s responsibility was to determine whether a preponderance of the evidence showed three factors: (1) a police officer had reasonable cause to believe Freitas was violating Vehicle Code section 23152 or 23153; (2) Freitas was lawfully arrested; and (3) Freitas was driving with a BAC of 0.08 percent or more. (Gananian v. Zolin (1995) 33 Cal.App.4th 634, 638 [ 39 Cal.Rptr.2d 384 ]; Veh. Code, § 13557, subd. (b)(3).) He issued his decision on August 9, 2014. He found the arrest was lawful and the officer had reasonable cause to believe Freitas was driving while intoxicated. The hearing officer briefly summarized and rejected Arvizu’s testimony, stating that the evidence of a recommendation by the testing equipment manufacturers to use the dual-column method did not establish a violation of the applicable regulations. (Cal. Code Regs., tit. 17, § 1215 et seq.) The hearing officer concluded that the expert testimony did not suffice to show the test results were unreliable. The suspension was upheld.

Freitas filed his petition in the trial court on August 18, 2014. The matter was heard on January 26, 2015. In a minute order issued February 2, 2015, the court found in favor of the DMV. Observing that the DMV was entitled to a presumption that it acted in conformity with state regulations, the court held that it was required to presume that “single[-]column [gas chromatograph] analysis was capable of identifying and quantifying ethyl alcohol concentrations in [Freitas’s] blood sample with precision, and that it did so.” Then the court found that Arvizu’s testimony failed to rebut this presumption. Arvizu did not rely on “calibration or similar maintenance records” to show that the laboratory’s gas chromatograph “was producing inaccurate results at the time” Freitas’s sample was tested. Further, Arvizu did not ‘“testify regarding [Freitas’s] actual BAC, or state that [his] BAC was lower than .23 [percent]. No evidence was presented that [Freitas’s] reported .23 [percent] BAC was, in fact, wrong.” “The evidence that the test result might be wrong is not sufficient to overcome the presumption,” the court concluded.

In addition to finding the blood test results to be valid, the trial court relied on corroborating evidence of Freitas’s alcohol intoxication. It pointed to evidence of his speeding and erratic driving, his failure to pull over until after driving home, his unsteadiness on his feet, smell of alcohol, red eyes, slurred speech, and admission he had been drinking beer.

The court entered judgment against Freitas on February 19, 2015. A stay on the suspension of his license was lifted.

DISCUSSION

In ruling on a petition for a writ of mandate seeking reversal of the suspension of a driver’s license, a trial court must apply its independent judgment to determine whether the weight of the evidence supports the administrative decision. (Lake v. Reed (1997) 16 Cal.4th 448, 456 [ 65 Cal.Rptr.2d 860 , 940 P.2d 311 ].) We must uphold the trial court’s factual findings if they are supported by substantial evidence in the record. In deciding whether there is substantial evidence, we resolve all evidentiary conflicts and draw all reasonable inferences in favor of the trial court’s decision; we cannot reverse that decision merely because a different decision could also reasonably have been reached. (Id. at p. 457.) To the extent the appeal involves pure questions of law, including the interpretation of statutes and regulations, we review those questions de novo. (Borger v. Department of Motor Vehicles (2011) 192 Cal.App.4th 1118, 1121 [ 121 Cal.Rptr.3d 816 ]; Manriquez v. Gourley (2003) 105 Cal.App.4th 1227, 1233 [ 130 Cal.Rptr.2d 209 ] (Manriquez).)

At the administrative hearing, the DMV had the burden of proving by a preponderance of the evidence that Freitas had a BAC of 0.08 percent or more. (Manriquez, supra, 105 Cal.App.4th at p. 1232 .) The DMV can do this, however, by merely submitting blood-alcohol test results recorded on official forms. (Shannon v. Gourley (2002) 103 Cal.App.4th 60, 64 [ 126 Cal.Rptr.2d 327 ].) This is because (1) provisions of title 17 of the California Code of Regulations (specifically, Cal. Code Regs., tit. 17, § 1215 et seq.) regulate the collection and testing of blood samples for determination of alcohol concentration; (2) Evidence Code section 664 creates a rebuttable presumption that official duties (such as the duty to follow regulations) have been carried out; and (3) Evidence Code section 1280 establishes a hearsay exception for records made by public employees. (Shannon, supra, at p. 65 .) Consequently, ‘“[t]he recorded test results are presumptively valid and the DMV is not required to present additional foundational evidence.” (Ibid.)

After the DMV has made its initial showing by means of these official test result records, the burden shifts to the driver “to demonstrate that the test was not properly performed.” (Imachi v. Department of Motor Vehicles (1992) 2 Cal.App.4th 809, 817 [ 3 Cal.Rptr.2d 478 ].) Among the ways the driver can do this is by showing that the particular machine used to test the sample malfunctioned or was improperly calibrated or employed. (People v. Vangelder (2013) 58 Cal.4th 1, 34 [ 164 Cal.Rptr.3d 522 , 312 P.3d 1045 ].) If the driver does this, the burden of proof shifts back to the DMV to show that the results are reliable despite the facts presented by the driver. (Manriquez, supra, 105 Cal.App.4th at p. 1233 .)

In this case, it is undisputed that Freitas was lawfully arrested and the officer had reasonable cause to believe he was driving under the influence. It also is undisputed that the DMV submitted official records adequate to make an initial showing that Freitas was driving with a BAC of 0.08 percent or more. The burden of producing evidence to the contrary thus was shifted to Freitas. The trial court concluded that Freitas failed to shift the burden back to the DMV. This conclusion was based primarily on the view that Arvizu’s testimony was inadequate to undermine the blood test results.

The court’s view was based on an error of law, for the court misapplied the presumption of regulatory compliance. The applicable regulation requires blood to be tested by a method “capable of the analysis of ethyl alcohol with a specificity which is adequate and appropriate for traffic law enforcement.” (Cal. Code Regs., tit. 17, § 1220.1, subd. (a)(2).) It is rebutta-bly presumed that the method employed had this capability. Contrary to the trial court’s opinion, however, rebutting this presumption did not require Freitas to prove the gas chromatograph was improperly calibrated or maintained. It also did not require him to present proof of his actual BAC. A driver can rebut the presumption by showing that a testing apparatus was improperly employed when the driver’s sample was tested. (People v. Vangelder, supra, 58 Cal.4th at p. 34 .)

The proper employment of the apparatus includes using the data it produces in a scientifically valid manner. Arvizu’s uncontroverted testimony showed the laboratory was not using a valid methodology in the process of getting from the raw data produced by the gas chromatograph to the conclusion that the samples had a BAC of 0.08 or more. The point of her testimony was that, even assuming the machine was working correctly with both columns installed, the reported conclusion that Freitas’s samples had a BAC of more than 0.08 percent was based on only one column’s data, and, as a matter of well-established scientific principle, one column’s data are incapable of supporting this conclusion. The trial court stressed that Arvizu did not show the results were incorrect, but the possibility that the results of a scientifically invalid procedure might happen to be correct by chance hardly shows the presumption of validity to be unrebutted.

The nature of the trial court’s error is well illustrated by the remark in its minute order that Arvizu must be wrong because “the consequence of accepting [Arvizu’s] conclusion would be that single[-]column [gas chromatography] analysis would never be acceptable, even though it strictly complies with [the applicable regulations in Cal. Code Regs., tit. 17].” This remark reflects misunderstanding of both the regulations and the testimony. The regulations state that the blood-te sting method must be capable of alcohol analysis adequate for enforcing the law. (Cal. Code Regs., tit. 17, § 1220.1, subd. (a)(2).) Arvizu’s testimony, if correct, established that single-column gas chromatography is not capable of determining either the presence or the concentration of blood alcohol. This means that single-column gas chromatography is indeed never valid and therefore cannot comply, strictly or otherwise, with the regulations. The DMV did not attempt to rebut Arvizu’s testimony and there is nothing in the record on the basis of which the trial court could reasonably have rejected it.

In so holding, we follow this court’s recent decision in Najera, supra, 241 Cal.App.4th 173 . That case involved the same expert, Arvizu, giving the same uncontroverted testimony that single-column gas chromatography was incapable of valid measurement of BAC. (Id. at pp. 177-178.) We held (id. at p. 182) that this testimony rebutted the presumption that the laboratory was using methodology “capable of the analysis of ethyl alcohol with a specificity which is adequate and appropriate for traffic law enforcement.” (Cal. Code Regs., tit. 17, § 1220.1, subd. (a)(2).)

The above discussion disposes of the DMV’s main argument, that Arvizu’s testimony failed to rebut the presumption that the blood test results were valid because the regulations are presumed to have been followed. Several additional arguments remain to be addressed.

First, the DMV maintains Najera does not apply here because the procedural posture is reversed. There, because the driver won in the trial court, the question was only whether Arvizu’s testimony was sufficient to support the trial court’s finding that it rebutted the DMV’s prima facie case. Here, the driver lost, and the trial court found the expert testimony failed to rebut the DMV’s prima facie case, so the question is whether a reasonable finder of fact could so conclude. The DMV argues that a reasonable finder of fact could do so even if he or she also could find as the trial court did in Najera.

This attempt to distinguish Najera fails for two reasons. First, the trial court here committed legal error in holding that the presumption of regulatory compliance could not be rebutted without evidence that the gas chromato-graph was improperly maintained and calibrated or that Freitas’s BAC was actually below 0.08 percent. In reality, testimony like Arvizu’s, attacking the scientific validity of using only one column’s data to determine the sample’s BAC, also can rebut the presumption. Second, the DMV presented no rebuttal evidence. There is nothing in Arvizu’s testimony or elsewhere in the record on the basis of which a reasonable finder of fact could conclude that her testimony failed to carry Freitas’s burden and thereby shift the burden back to the DMV.

Next, the DMV contends “there was no evidence that the crime lab’s failure to ‘report’ the results from the second column meant that it did not consider the data from both columns.” The record does not support this view. Arvizu testified in effect that, according to her review of the materials provided by the DMV, the reported BAC results from the Kern Regional Crime Laboratory were based on data from only one column of the gas chromatograph. There is no other reasonable interpretation of her testimony. The trial court interpreted the evidence the same way. It ruled not that the evidence failed to show the reported BAC was based on data from only one column, but that it failed to show the lab’s use of a single-column method undermined the presumption of regulatory compliance. It would be sheer speculation to say the lab must have employed the second-column data even though the materials it provided reflected only one column’s results.

As we held in response to a similar argument in Najera, Arvizu’s testimony shifted the burden of proof back to the DMV. The DMV was then obligated to present evidence that it used data from both columns, if any such evidence existed. (Najera, supra, 241 Cal.App.4th at p. 184 .) The DMV presented no such evidence.

The DMV further argues that the data from the second column were provided to Freitas in discovery and Arvizu could have analyzed those data, determined whether they supported the reported BAC result, and testified on that point. Because she failed to do so, the DMV says, her opinion was unsupported.

Freitas never claimed the second column was not operating or the gas chromatograph did not generate data from the second column. The mere facts that the data existed and were produced in discovery have no bearing on the validity of test results derived without reference to those data. As for the DMV’s argument that Arvizu should have analyzed the data and determined whether they supported or undermined the first-column results on which the DMV relied when it suspended Freitas’s license, this was the DMV’s burden, as we held in response to a similar argument in Najera, supra, 241 Cal.App.4th at page 183 : “If Najera’s blood test results really were supported by the data from the second column despite the DMV’s failure to use those data when it suspended Najera’s license, it was up to the DMV, not Najera, to present evidence to that effect at the hearing.”

The DMV also avers that Arvizu could have tested, but did not test, Freitas’s blood samples independently, as permitted by regulation. (Cal. Code Regs., tit. 17, § 1219.1, subd. (g).) As we held in Najera, this argument again overlooks the fact that the expert testimony shifted the burden back to the DMV. (Najera, supra, 241 Cal.App.4th at p. 184 .)

Next, the DMV contends there is another aspect of the Evidence Code section 664 presumption of regulatory compliance that Arvizu’s testimony failed to rebut. The State Department of Health Care Services is required to evaluate a blood laboratory’s methods to make sure they satisfy regulatory standards. (Cal. Code Regs., tit. 17, § 1220.1, subd. (b).) The DMV says the trial court was required to presume the State Department of Health Care Services did this evaluation and found the Kern Regional Crime Laboratory was in compliance. Therefore, according to the DMV, the court was compelled to presume the laboratory was found to be using a method “capable of the analysis of ethyl alcohol with a specificity which is adequate and appropriate for traffic law enforcement.” (Cal. Code Regs., tit. 17, § 1220.1, subd. (a)(2).) In the DMV’s view, this means the court could not rely on expert testimony to find that the method used was not, in reality, capable of valid blood-alcohol analysis.

As we stated in Najera, this would be “a presumption too far,” because it would contravene the principle that a driver can “ ‘demonstrate that the test was not properly performed.’ ” (Najera, supra, 241 Cal.App.4th at p. 183 .) It is incorrect because it “would amount to a presumption that anything the lab does has been approved and thus is unchallengeable.” (Ibid.) A driver can rebut the presumption that the State Department of Health Care Services found a lab’s methods capable of adequate analysis by demonstrating that its methods are not, in fact, capable of adequate analysis. That is what happened here.

Finally, the DMV asserts that the data from one column of the gas chromatograph should count as at least some evidence to be corroborated by other indications of Freitas’s intoxication. The DMV’s thinking here is that if a positive result from one column means that some volatile organic compound is in the sample at a concentration of 0.08 percent or more, then other evidence, such as the driver’s admission to having been drinking, could confirm that the compound is alcohol. This reasoning may be superficially attractive, but it does not appear to us to be correct. Arvizu’s testimony indicated that a positive result from one column is consistent with alcohol, some other compound, or both alcohol and another compound. If a driver has consumed alcohol, but not enough alcohol to have a BAC of 0.08 percent, a test result based on data from one column could falsely show a BAC of 0.08 percent or more because the sample contained both alcohol and another volatile organic compound. For this reason, the lab’s failure to use the data from the second column means the test results cannot be used to support a finding of a BAC of 0.08 percent or more even in conjunction with behavioral evidence of impairment. If this other evidence did not show a BAC of 0.08 percent on its own, the single-column result would not be able to fill the gap, because a single-column positive result is always consistent with the presence of alcohol and another compound in unknown proportions.

The DMV cites Coffey v. Shiomoto (2015) 60 Cal.4th 1198 [ 185 Cal.Rptr.3d 538 , 345 P.3d 896 ] (Coffey) in support of its view that the single-column result can be considered in conjunction with behavioral evidence even if Arvizu’s testimony rebutted the presumption that the blood test results were reliable. Coffey was pulled over for driving erratically. Fifty-six minutes after the stop, she took a breath test which showed a BAC of 0.08 percent. Three minutes later, a second breath test showed a BAC of 0.09 percent. A blood sample taken 24 minutes after that was found to have a BAC of 0.095 or 0.096 percent. Coffey also had red eyes, failed field sobriety tests, and smelled of alcohol. {Id. at pp. 1203-1205.) At the DMV hearing, her expert testified that the breath and blood test results showed Coffey’s BAC was rising from 0.08 percent when she was stopped, so he would expect it to have been below 0.08 percent while she was driving. {Id. at p. 1205.) The Supreme Court held that this testimony rebutted the presumption that the test results showed Coffey was driving with a BAC of 0.08 percent or more. {Icl. at pp. 1209-1211.) This shifted the burden back to the DMV, but the test results could still be considered, in conjunction with the other evidence, in determining whether Coffey’s BAC was 0.08 percent while she was driving. {Id. at pp. 1211-1217.) In light of this, the Supreme Court concluded that substantial evidence supported the trial court’s finding against Coffey. {Id. at pp. 1217-1218.)

Coffey is distinguishable from this case. In Coffey , the expert’s testimony did not indicate anything intrinsically wrong with the test results. He only said the results were consistent with Coffey’s BAC being lower before the breath and blood samples were taken. There was no evidence that the testing methodology failed to determine validly the BAC in the samples. Here, by contrast, the uncontroverted expert testimony showed that the blood lab’s methodology was incapable of accurately measuring BAC.

It might be supposed that the behavioral, nonquantitative evidence of Freitas’s impairment could, in principle, support the trial court’s ruling by itself (i.e., even when the blood test results are disregarded) and that we should remand for consideration of this. We do not think so, however. In Coffey , the court assumed “that non-chemical-test evidence cannot by itself prove a driver’s exact BAC at the moment the driver is stopped by a police officer,” but refrained from making any definitive statement on that question. (Coffey, supra, 60 Cal.4th at p. 1216 .) Justice Liu, however, in a concurring opinion, made a persuasive argument that such evidence alone cannot establish that a driver drove with BAC of 0.08 percent or more: “Evidence that a driver’s behavior is not impaired tends to prove that her BAC was below 0.08 percent because we can rationally surmise, given the Legislature’s choice of the 0.08 percent BAC threshold, that a BAC of 0.08 percent is associated with an unsafe degree of impairment. But the converse is not true. Absent foundational evidence, a driver’s impairment does not generally tend to show that her BAC was 0.08 percent or higher because we have no way of correlating a specific type or degree of impairment with a particular BAC in a close case. The fact that 0.08 percent BAC is a threshold associated with an unsafe degree of impairment does not imply that no impairment occurs below that threshold. Without evidence that correlates particular behavioral impairments with particular BAC levels, signs of impairment do not generally establish that it is more likely a driver’s BAC is 0.08 percent as opposed to 0.06 or 0.07 percent. In other words, signs of impairment do not generally have a tendency in reason to prove a BAC of 0.08 percent or greater.” (Coffey, supra, 60 Cal.4th at p. 1218 (conc. opn. of Liu, J.).)

Justice Liu concluded that the evidence of behavioral impairment could not by itself have supported the judgment against the driver, but it could properly be used to support the DMV’s argument that if Coffey’s BAC was at 0.08 percent when she was pulled over, she was probably also over the threshold at the time when she was driving, despite the expert testimony. ( Coffey, supra, 60 Cal.4th at pp. 1219-1220 (conc. opn. of Liu, J.).) Here, there are no valid chemical test results at all, so there is nothing for the nonquantitative evidence of impairment to bolster.

DISPOSITION

The judgment is reversed. The trial court is directed to issue a writ of mandate ordering the DMV to set aside its order issued August 9, 2014, suspending Freitas’s driving privilege. Costs on appeal are awarded to Freitas.