Allan Wagner appeals from a district court judgment affirming the administrative suspension of his driving privileges. We reverse.
Wagner was arrested for driving under the influence of alcohol. He submitted to an Intoxilyzer test administered by Damon Busch, the arresting officer. The test indicated a blood-alcohol concentration of 0.10 percent by weight. Busch took possession of Wagner’s driver’s license and Wagner requested and received an administrative hearing pursuant to NDCC § 39-20-05.
At the hearing, Busch testified that he had followed the state toxicologist’s approved method for administering the Intox-ilyzer test. During cross-examination, however, Busch admitted that he did not insert into the machine, at the time the approved method required, the card on which the test results are recorded. Wagner argued that this lapse was a failure on the part of Busch to follow the approved method. At the close of the hearing, the hearing officer found, in part, that Busch properly performed the Intoxilyzer test and concluded that Wagner was properly tested. Accordingly, Wagner’s driver’s license was suspended for ninety-one days. The district court affirmed the suspension and Wagner appealed. The only issue on appeal is whether the hearing officer properly admitted the Intoxilyzer test results into evidence.
Under section 39-20-05(2), NDCC, one issue at Wagner's administrative hearing was whether his Intoxilyzer test results demonstrated that he had a blood-alcohol concentration of at least 0.10 percent by weight. Section 39-20-07(5), NDCC, governs the admissibility of results of a chemical analysis of a driver’s breath to determine blood-alcohol content:
“The results of the chemical analysis must be received in evidence when it is shown that the sample was properly obtained and the test was fairly administered, and if the test is shown to have been performed according to methods and with devices approved by the state toxicologist, and by an individual possessing a certificate of qualification to administer the test issued by the state toxicologist....”
The State Toxicologist’s Approved Method To Conduct Breath Test With Intoxilyzer provides in part:
“To initiate a test depress the ‘Start Test’ switch. The display will flash ‘Insert Test Record’. Before the display time expires, insert a Form 106-1 into the instrument. Once the Form 106-1 is inserted, the Intoxilyzer will perform a series of diagnostic checks.”
Busch stated during cross-examination that he had not inserted Form 106-1, the Intoxi-lyzer Test Record and Checklist, before the display time had expired. He did not contradict this admission during redirect examination, and neither Busch, nor any other witness, testified as to the significance of inserting the Form at a time other than that directed by the approved method.
Wagner contends that the test was not performed in accordance with the approved method and, consequently, the hearing officer erred in concluding that Wagner had been fairly tested. We agree.
When the director fails to establish compliance with the State Toxicolo*600gist’s directions which go to the scientific accuracy of the test, the director must prove fair administration of the test through expert testimony. Schwind v. Director, 462 N.W.2d 147 (N.D.1990); see also Glaspey v. Baches, 462 N.W.2d 635 (N.D.1990); Schirado v. North Dakota St. Hwy Com’r, 382 N.W.2d 391 (N.D.1986); Moser v. North Dakota State Hwy Com’r, 369 N.W.2d 650 (N.D.1985). When there is a deviation from the State Toxicologist’s approved method which goes to the scientific accuracy of a test, a hearing officer cannot assume that the test was accurate. Price v. North Dakota DOT Director, 469 N.W.2d 560 (N.D.1991). To establish fair administration, the director can prove that those portions of the approved method which go to the scientific accuracy and reliability of the test have been “scrupulously” complied with, or he can offer expert testimony. State v. Schwalk, 430 N.W.2d 317, 323-24 (N.D.1988).
The director contends that Wagner did not produce any evidence that the untimely insertion of the Form 106-1 compromises the accuracy of the test results. The director also contends that expert testimony is not necessary to address the scientific accuracy of the Intoxilyzer test in this case. Relying on a Minnesota case, State v. Habisch, 313 N.W.2d 13 (Minn.1981), the director argues that the results of the self-diagnostic exams run by the Intoxilyzer after Busch inserted the test card proves that the machine was operating properly. In Habisch, the Minnesota Supreme Court considered the significance of using a 62-day-old standard solution when the normal procedure was to change the solution every thirty days. We considered a similar question in Pladson v. Hjelle, 368 N.W.2d 508 (N.D.1985).
In Habisch and Pladson, the responsible administrative agencies had prescribed a method for conducting the test and the tests had been conducted according to that method. In both cases, the approved method did not say when the standard solution sample had to be changed and the drivers, therefore, raised a question that was not answered by the approved method. Habisch, 313 N.W.2d at 16; Pladson, 368 N.W.2d at 512-13. Therefore, the burden fell on the driver to produce some evidence discrediting the test results when the director had made a prima facie showing of fair administration by establishing first that the approved method was followed. Pladson, 368 N.W.2d at 513. See also Fargo v. Komulainen, 466 N.W.2d 610, 612 (N.D.1991). In this case, the director did not make a prima facie showing that the approved method was followed. Therefore, Wagner was not required to offer further evidence showing the significance of Busch’s deviation from the approved method.
The director argues that the deviation from the approved method could not have affected the test results. It is true that not all deviations from an approved method are fatal. We have required expert testimony to meet the foundational requirements for the admission of chemical analysis results when the director has not shown scrupulous compliance with those portions of an approved method which prescribe the procedures for collecting and testing the samples used in the chemical analysis and which, therefore, go to the scientific accuracy and reliability of the test. E.g., Schwalk, 430 N.W.2d at 323-24; Moser, 369 N.W.2d at 653. When, however, we have been able to say that the deviation involved some clerical or ministerial aspect of an approved method and, therefore, could not have affected the test results, we have upheld a license suspension. Schwind, 462 N.W.2d at 152 [deputy, not nurse, inspected and certified seal on blood sample]; e.g., Heinrich v. North Dakota St. Hwy. Com’r, 449 N.W.2d 587, 589 (N.D.1989) [post-test correction of standard solution number]; Schense v. Hjelle, 386 N.W.2d 888, 891 (N.D.1986) [discrepancy in serial number of simulator used in Intoxy-lizer test]. We cannot say, without expert advice, that the failure to follow the portion of the approved method governing the timely insertion of the test record could not affect the test results.
Because the testing officer deviated from a portion of the State Toxicologist’s Ap*601proved Method that may affect scientific accuracy and reliability of the test, and there was no expert testimony on the effect, if any, of the deviation on the accuracy of the test results, the director failed to show that the test was “fairly administered.” We conclude that the hearing officer should not have admitted the Intoxilyzer test result into evidence.
Wagner’s license was suspended for his operation of a vehicle while having a blood-alcohol content of at least 0.10 percent by weight. The Intoxilyzer results were the only evidence presented to demonstrate that Wagner’s blood-alcohol content was 0.10 percent or greater. Consequently, the license suspension must be reversed.
The judgment of the district court affirming the administrative suspension of Wagner’s license is reversed.
GIERKE and VANDE WALLE, JJ„ concur.