{¶ 1} The issue presented in this case is whether the state substantially complies with the alcohol-testing regulations set forth in the Ohio Administrative Code when it fails to use a solid anticoagulant in a blood test.
I
{¶ 2} On July 28, 2001, Trooper Donald A. Ward arrested defendant-appellee, Chadd A. Burnside, for driving under the influence of alcohol. In the early hours of that morning, Trooper Ward stopped Burnside’s vehicle for a speeding violation. During the stop, Trooper Ward observed that Burnside had glassy eyes and smelled of alcohol. Burnside admitted that he had consumed “a few” alcoholic beverages and consented to a standard field sobriety test. Upon the completion of the field sobriety test, Trooper Ward arrested Burnside and charged him with driving under the influence of alcohol in violation of R.C. 4511.19(A)(1).
{¶ 3} Trooper Ward transported Burnside to the State Highway Patrol Post, where he asked Burnside to submit to a chemical breath test. Burnside refused and requested a blood test. Trooper Ward then transported Burnside to the Fairfield Medical Center, where a phlebotomist drew a blood sample. After properly sealing the blood sample, Trooper Ward mailed the sample to the State Highway Patrol Laboratory in Columbus, Ohio. A criminalist at the laboratory tested the sample and determined that Burnside’s blood-alcohol content was 0.169 grams per 100 milliliters. As a result, Trooper Ward additionally charged Burnside with operating a motor vehicle with a prohibited blood-alcohol level in violation of R.C. 4511.19(A)(2).
*154{¶ 4} On August 1, 2001, Burnside entered pleas of not guilty to charges of speeding, driving under the influence of alcohol, and driving with a prohibited level of alcohol. Burnside thereafter filed a motion to suppress the blood-test results, asserting that the state had failed to substantially comply with the blood-testing procedures in Ohio Adm.Code 3701-53-05. At the suppression hearing, the state presented testimony from the arresting officer, the phlebotomist who drew the blood sample, and the criminalist who tested the blood sample. Burnside presented no evidence. During closing argument, Burnside asserted that the blood-test results should be suppressed because the state had failed to establish that it (1) used a nonvolatile antiseptic on his skin prior to collecting the blood sample and (2) drew the blood into a vacuum container with a solid anticoagulant. The trial court overruled Burnside’s motion to suppress, holding that the state had substantially complied with the blood-testing procedures in Ohio Adm.Code 3701-53-05.
{¶ 5} On December 4, 2001, Burnside entered a plea of no contest to operating a motor vehicle with a prohibited level of alcohol in violation of R.C. 4511.19(A)(2). In exchange, the state dismissed the remaining charges. Burnside appealed his conviction to the Fifth District Court of Appeals, alleging that the trial court had erred in denying his motion to suppress the blood-test results. The court of appeals reversed the judgment of the trial court on the basis that the state failed to substantially comply with Ohio Adm.Code 3701-53-05 and certified its judgment to be in conflict with that of the Seventh District Court of Appeals in State v. Zuzga (2001), 141 Ohio App.3d 696, 753 N.E.2d 229.
{¶ 6} This cause is now before this court upon our determination that a conflict exists in case No. 2002-1524 and pursuant to the allowance of a discretionary appeal in case No. 2002-1440.
II
{¶ 7} The court of appeals certified the following issue for our determination: “When there is no evidence that a solid anticoagulant is used in a blood test to determine alcohol content as required by O.A.C. 3701-53-05(C), dos [sic] the State still meet its burden of substantial compliance with Department of Health regulations?” Our analysis of the certified issue begins with the appropriate standard of review.
A. Standard of Review
{¶ 8} Appellate review of a motion to suppress presents a mixed question of law and fact. When considering a motion to suppress, the trial court assumes the role of trier of fact and is therefore in the best position to resolve factual *155questions and evaluate the credibility of witnesses. State v. Mills (1992), 62 Ohio St.3d 357, 366, 582 N.E.2d 972. Consequently, an appellate court must accept the trial court’s findings of fact if they are supported by competent, credible evidence. State v. Fanning (1982), 1 Ohio St.3d 19, 1 OBR 57, 437 N.E.2d 583. Accepting these facts as true, the appellate court must then independently determine, without deference to the conclusion of the trial court, whether the facts satisfy the applicable legal standard. State v. McNamara (1997), 124 Ohio App.3d 706, 707 N.E.2d 539. We therefore consider whether the facts in the instant case demonstrate substantial compliance with the Department of Health regulations under a de novo standard of review. Resolution of this issue requires an examination of the blood-testing procedures in Ohio Adm.Code 3701-53-05.
B. The Blood-Testing Procedures in Ohio Adm.Code 3701-53-05.
{¶ 9} The General Assembly established the threshold criteria for the admissibility of alcohol-test results in prosecutions for driving under the influence and driving with a prohibited concentration of alcohol in R.C. 4511.19(D). That section, which governs the admissibility of alcohol-test results, provides that a defendant’s blood, breath, or urine “shall be analyzed in accordance with methods approved by the director of health by an individual possessing a valid permit issued by the director of health pursuant to section 3701.143 of the Revised Code.” R.C. 3701.143 requires the director of health to “determine, or cause to be determined, techniques or methods for chemically analyzing a person’s blood.”
{¶ 10} In accordance with this statutory mandate, the Director of Health promulgated the following alcohol-testing regulations in Ohio Adm.Code 3701-53-05:
{¶ 11} “(A) All samples shall be collected in accordance with section 4511.19, or section 1547.11 of the Revised Code, as applicable.
{¶ 12} “(B) When collecting a blood sample, an aqueous solution of a nonvolatile antiseptic shall be used on the skin. No alcohols shall be used as a skin antiseptic.
{¶ 13} “(C) Blood shall be drawn ivith a sterile dry needle into a vacuum container with a solid anticoagulant, or according to the laboratory protocol as written in the laboratory procedure manual based on the type of specimen being tested.
{¶ 14} “(D) The collection of a urine specimen must be witnessed to assure that the sample can be authenticated. Urine shall be deposited into a clean glass or plastic screw top container which shall be capped, or collected according to the laboratory protocol as written in the laboratory procedure manual.
*156{¶ 15} “(E) Blood and urine containers shall be sealed in a manner such that tampering can be detected and have a label which contains at least the following information:
{¶ 16} “(1) Name of suspect;
{¶ 17} “(2) Date and time of collection;
{¶ 18} “(3) Name or initials of person collecting the sample; and
{¶ 19} “(4) Name or initials of person sealing the sample.
{¶ 20} “(F) While not in transit or under examination, all blood and urine specimens shall be refrigerated.” (Emphasis added.)
{¶ 21} The blood-testing procedure in Ohio Adm.Code 3701-53-05 thus requires the state to (1) use an aqueous solution of a nonvolatile antiseptic on the skin, (2) use a sterile dry needle to draw blood into a vacuum container with a solid anticoagulant, (3) seal the blood container in accordance with the appropriate procedure, and (4) refrigerate the blood specimen when it is not in transit or under examination. The purpose of these regulations is to ensure the accuracy of the alcohol-test results. State v. Dickerson (1986), 25 Ohio St.3d 64, 65-66, 25 OBR 86, 495 N.E.2d 6.
C. Judicial Interpretation of Ohio Adm.Code 3701-53-05
{¶ 22} We first addressed the application of the Department of Health regulations that govern alcohol testing in State v. Steele (1977), 52 Ohio St.2d 187, 6 O.O.3d 418, 370 N.E.2d 740. Steele involved the admissibility of a breath-alcohol test administered by a police officer who failed to continuously observe a subject for the requisite 20-minute observation period prior to administering the test. We concluded that the failure to observe the subject for a “few seconds” while the officer exited and walked around his patrol car did not render the test results inadmissible. Steele thus established that rigid compliance with the alcohol-testing procedures in the Ohio Administrative Code is not a prerequisite to the admissibility of alcohol-test results.
{¶ 23} Nearly a decade later, we again addressed whether the state had complied with the Department of Health regulations relating to alcohol testing in State v. Plummer (1986), 22 Ohio St.3d 292, 22 OBR 461, 490 N.E.2d 902. In Plummer, we considered the consequences of the state’s failure to comply with a regulation that required it to refrigerate urine samples that were not in transit or under examination. Concluding that a three-to-four-hour interval without refrigeration did not render the test results inadmissible, we held that “[ajbsent a showing of prejudice to a defendant, the results of a urine-alcohol test administered in substantial compliance with Ohio Adm.Code 3701-53-05 are admissible in a prosecution under R.C. 4511.19.” Id. at syllabus.
*157{¶ 24} In the wake of Plummer, courts have applied a burden-shifting procedure to govern the admissibility of alcohol-test results. E.g., State v. Zuzga, 141 Ohio App.3d at 698-699, 753 N.E.2d 229. The defendant must first challenge the validity of the alcohol test by way of a pretrial motion to suppress; failure to file such a motion “waives the requirement on the state to lay a foundation for the admissibility of the test results.” State v. French (1995), 72 Ohio St.3d 446, 451, 650 N.E.2d 887. After a defendant challenges the validity of test results in a pretrial motion, the state has the burden to show that the test was administered in substantial compliance with the regulations prescribed by the Director of Health. Once the state has satisfied this burden and created a presumption of admissibility, the burden then shifts to the defendant to rebut that presumption by demonstrating that he was prejudiced by anything less than strict compliance. State v. Brown (1996), 109 Ohio App.3d 629, 632, 672 N.E.2d 1050. Hence, evidence of prejudice is relevant only after the state demonstrates substantial compliance with the applicable regulation.
{¶ 25} Against this backdrop, we turn to the instant case.
Ill
{¶ 26} The state does not dispute that it had the burden to establish compliance with Ohio Adm.Code 3701-53-05, that this provision requires the use of a solid anticoagulant, and that it failed to establish the use of an anticoagulant. The state’s argument, rather, is that it need not use a solid anticoagulant to substantially comply with Ohio Adm.Code 3701-53-05. We begin our analysis by examining the parameters of the “substantial compliance” standard.
A
{¶27} In determining the admissibility of alcohol-test results regulated by Ohio Adm.Code 3701-53-05, we have observed that “there is leeway for substantial, though not literal, compliance with such regulations.” Plummer, 22 Ohio St.3d at 294, 22 OBR 461, 490 N.E.2d 902. The state must therefore establish that it substantially complied with the alcohol-testing regulations to trigger the presumption of admissibility. Our conclusion that the state must establish substantial compliance rather than strict compliance, however, does not relieve the state of its burden to prove compliance with the alcohol-testing regulations, but rather defines what compliance is.
{¶ 28} Although we have not had occasion to expound upon the substantial-compliance standard, appellate courts have developed two approaches to deter*158mine whether the state has substantially complied with Ohio Adm.Code 3701-53-05. One approach is to consider whether the noncompliance rendered the test results unreliable. See, e.g., State v. Gray (1980), 4 Ohio App.3d 47, 50, 51, 4 OBR 96, 446 N.E.2d 469. Under this approach, a court will conclude that the state has substantially complied with the Department of Health regulations if the alleged deviation did not affect the reliability of the test results. Id. The other approach for determining substantial compliance is to consider whether the alleged deviation prejudiced the defendant. See, e.g., State v. Zuzga (2001), 141 Ohio App.3d 696, 701, 753 N.E.2d 229. Under this approach, a court will conclude that the state has substantially complied with the Department of Health regulations so long as the alleged deviation did not cause an erroneously higher test result. Id.
{¶ 29} The import in denominating between these two approaches lies not in understanding the difference between them, but rather in recognizing the similarity: both require a judicial determination of what effect, if any, noncompliance had on the alcohol-test results. This determination, however, often requires judges to speculate why the Director of Health adopted a given regulation. One judge, charged with determining whether the failure to strictly comply with a regulation rendered alcohol-test results unreliable, deplored the fact that “most judges, myself included, do not know enough about chemistry, physics, or scientific testing so as to be able to know why the Department of Health adopted some of the required procedures.
{¶ 30} “* * *
{¶ 31} “* * * Thus, since I cannot know whether there was substantial compliance in this case, I am left with having to guess.” State v. Mitchell (Mar. 31, 1995), 6th Dist. No. L-92-227, 1995 WL 136820 (Grey, J., dissenting).
{¶ 32} This sentiment is not surprising when one considers the more fundamental problem with such a method of determining admissibility: a judicial determination that an alcohol test, although not administered in strict compliance with the alcohol-testing regulations, is reliable and therefore admissible may subvert the rule-making authority and the statutory mandate of the Director of Health. Indeed, the General Assembly instructed the Director of Health — and not the judiciary — to ensure the reliability of alcohol-test results by promulgating regulations precisely because the former possesses the scientific expertise that the latter does not. See R.C. 4511.19(D)(1). Notwithstanding this statutory mandate, however, courts have concluded that the state need not show strict compliance with the regulations prescribed by the Director of Health if a judge deems the test results reliable. The problem, of course, is that such an approach is inconsistent with R.C. 4511.19, which provides that compliance with the regulations, rather than a judicial determination as to reliability, is the criterion *159for admissibility. See Cincinnati v. Sand (1975), 43 Ohio St.2d 79, 72 O.O.2d 44, 330 N.E.2d 908.
{¶ 33} This problem is particularly acute where, as here, the state has failed to proffer evidence that it complied with a particular regulation directly related to blood-alcohol testing. To state it succinctly: A court infringes upon the authority of the Director of Health when it holds that the state need not do that which the director has required. Such an infringement places the court in the position of the Director of Health for the precise purpose of second-guessing whether the regulation with which the state has not complied is necessary to ensure the reliability of the alcohol-test results. This approach further precipitates conflicting decisions from lower courts and impedes the public policy of achieving uniformity and stability in the law. Painter, Ohio Driving Under the Influence Law (2003), Section 9.3,116.
{¶ 34} Nevertheless, we are cognizant that if “we were to agree * * * that any deviation whatsoever from th[e] regulation rendered the results of a [test] inadmissible, we would be ignoring the fact that strict compliance is not always realistically or humanly possible.” Plummer, 22 Ohio St.3d at 294, 22 OBR 461, 490 N.E.2d 902. Precisely for this reason, we concluded in Steele that rigid compliance with the Department of Health regulations is not necessary for test results to be admissible. Steele, 52 Ohio St.2d at 187, 6 O.O.3d 418, 370 N.E.2d 740 (holding that the failure to observe a driver for a “few seconds” during the 20-minute observation period did not render the test results inadmissible). To avoid usurping a function that the General Assembly has assigned to the Director of Health, however, we must limit the substantial-compliance standard set forth in Plummer to excusing only errors that are clearly de minimis. Consistent with this limitation, we have characterized those errors that are excusable under the substantial-compliance standard as “minor procedural deviations.” State v. Homan (2000), 89 Ohio St.3d 421, 426, 732 N.E.2d 952.
{¶ 35} With these principles in mind, we consider whether the state substantially complied with the alcohol-testing regulations in the instant case.
B
{¶ 36} The state asserts that it substantially complied with the alcohol-testing regulations notwithstanding its failure to establish the use of a solid anticoagulant. We disagree. Ohio Adm.Code 3701-53-05(C) declares in no uncertain terms that “[b]lood shall be drawn * * * into a vacuum container with a solid anticoagulant.” (Emphasis added.) This language does not advise the use of a solid anticoagulant when drawing a blood sample; it demands it. Indeed, the state failed to produce any evidence that it complied with Ohio Adm.Code 3701-*16053-05(C). As a result, we cannot conclude that such an error is de minimis and therefore permissible under the substantial-compliance standard. Given that the state failed to establish substantial compliance, its insistence that the defendant did not show prejudice is immaterial; rather, any evidence of prejudice would have been relevant only after the state demonstrated substantial compliance with the alcohol-testing regulations in Ohio Adni.Code 3701-53-05(C).
Jason M. Griggs, Fairfield County Assistant Prosecuting Attorney, and Terre Lynne Vandervoort, Lancaster Law Director, for appellant. Dagger, Johnston, Miller, Ogilvie & Hampson and Scott P. Wood, for appellee.{¶ 37} Finally, we address the state’s argument that the alcohol-test results in the instant case should be admissible because the use of a solid anticoagulant was not necessary to ensure the reliability of the alcohol testing. This argument is properly directed not to us but to the Director of Health, whose charge it is to promulgate regulations that will ensure the reliability of alcohol-test results. To hold otherwise would be to speculate, with neither the requisite expertise nor the statutory authority, whether the failure to use a solid anticoagulant affected the reliability of the alcohol-test results in the instant case.
{¶ 38} Accordingly, we affirm the judgment of the court of appeals.
Judgment affirmed.
Hildebrandt, F.E. Sweeney, Pfeifer, Klatt, Lundberg Stratton and Brown, JJ., concur. Lee H. Hildebrandt Jr., J., of the First Appellate District, sitting for Resnick, J. William A. Klatt, J., of the Tenth Appellate District, sitting for Cook, J. Susan Brown, J., of the Tenth Appellate District, sitting for O’Connor, J.