Appellant Officer Thomas B. Stueart was terminated from the Arkansas State Police after he tested positive for marijuana use during a random drug screening pursuant to the Department’s Drug Free "Workplace Policy. He appealed his termination to the Arkansas State Police Commission, which upheld it. The Pulaski County Circuit Court affirmed on appeal. Stueart argued to the Commission, as he does to this court, that certain required procedures set forth in the Drug Free Workplace Policy were omitted and that this prejudiced substantial rights. This point is well taken, and establishes reversible error. We hold that because the Commission ignored its own rules in affirming Stueart’s termination, its decision was based upon unlawful procedure. Ark. Code Ann. § 25-15-212(h)(3) (Repl. 1996). Accordingly, we reverse.
Pursuant to the Drug Free Workplace Act of 1988, the Arkansas State Police adopted a Drug Free Workplace Policy by General Order No. 104. The stated purpose of the order was to establish “the policies and procedures of the Arkansas State Police governing alcohol or drug testing of employees” and to prohibit “alcohol or drug abuse or drug misuse by employees, either on or off duty.”
The policy sets out specific steps to be taken in the chain of custody to ensure the reliability of the testing and to prevent tampering. A Drug Custody and Control form is used to properly document each step of the testing process. This required documentation is to show who received the samples, who opened them, and who tested them. The employee is required to sign the form at the time he is tested to confirm that he is the donor of the *49sample, that he has not altered it in any way, that the bottle was sealed in his presence with a tamper-proof seal, and that the information provided on the form and on the label affixed to the specimen is correct.
At the end of the testing, the laboratory is required to submit any positive results to the authorized Medical Review Officer for confirmation of results. The Medical Review Officer is defined in the policy as “[a] licensed physician or designated person who reviews all positive drug test results to determine whether or not such results were due to the tested employee’s proper use of a prescribed medication.” The policy states, “A positive test result shall only be reported when both the initial and confirmatory tests have been completed and the positive result is not adequately explained to the satisfaction of the Medical Review Officer by consultation with the employee or the employee’s physician.” [Emphasis in original.]
At the hearing, both sides agreed that the chain of custody was flawed because the forms were not filled out correctly by the persons in the chain. However, there was testimony at the hearing that established who received the specimen at each point, who opened it, that it remained sealed until received by the testing facility, and who tested it. The specimen was taken from Stueart at his home in Ashdown. The officer in charge of collecting the specimen took it to his home and put it in the refrigerator overnight. He then delivered it to another officer, who delivered it to yet another officer at a designated point on the highway, and that officer took it to Baptist Medical Center in Arkadelphia. The specimen was taken by courier the next day to Baptist Medical Center in Little Rock, the testing facility, where it was tested.
While Stueart did not sign the donor certification on the form, he testified that he gave a specimen to the officer as established by the testimony at the administrative hearing. However, the fatal flaw was that the final steps in the procedure were omitted entirely. These requirements are, as stated in the policy:
(1) that any positive results be submitted to the authorized Medical Review Officer for confirmation of results; and (2) that a positive test shall only be reported when both the initial and *50confirmatory tests have been completed and the positive result is not adequately explained to the satisfaction of the Medical Review Officer by consultation with the employee or the employee’s physician. [Emphasis in original.]
It was undisputed that these steps were omitted. Dr. Don Cashman, supervisor of the toxicology lab at Baptist Medical Center, testified on cross-examination as follows:
Q. What did you do when you saw that the results of this test were positive?
A. We sent those results to the Arkansas State Police.
Q. You have reviewed the policy. Now have you come to an opinion that you did not follow the policy?
A. It’s correct, yeah. Our process did not follow their policy.
Q. The last page of the drug policy, number 5, it says, “Submit any positive results to the authorized Medical Review Officer for confirmation of results.” Was that done?
A. No.
Q. And the bottom part, it says in 6a, “A positive test result shall only be reported when both initial and confirmatory results have been completed and the positive result is not adequately explained to the satisfaction of the Medical Review Officer by consultation with the employee or the employee’s physician.” Was that done?
A. There [sic] results were not submitted to a medical review officer.
Q. So 6a was not done, either; is that correct?
A. That’s correct.
By affirming Stueart’s termination in the face of an admitted failure to follow the Department’s stated policy, the Commission failed to follow its own rules. This failure distinguishes this appeal from a typical appeal from an exercise of judgment by an administrative agency in which our standard of review is limited to a determination of whether the agency’s action is arbitrary and capricious, or whether its findings are unsupported by the record. See, e.g., Arkansas Dep’t of Human Servs. v. Kistler, 320 Ark. 501, 898 S.W.2d 32 (1995). Rather, we are concerned with whether the Commission’s decision is based upon unlawful procedure. Regional Health Care Facilities, Inc. v. Rose Care, 322 Ark. 767, 912 *51S.W.2d 409 (1995). We have held that a procedure is “unlawful” when an agency fails to follow that which it has prescribed. Id. As the District of Columbia Circuit Court of Appeals stated in Panhandle Eastern Pipeline Co. V. F.E.R.C., 613 F.2d 1120 (D.C. Cir. 1979), which we quoted in part in Rose Care, “It has become axiomatic that an agency is bound by its own regulations. The fact that a regulation as written does not provide [the agency] a quick way to reach a desired result does not authorize it to ignore the regulation . . . . ” Id. at 1135. The decision of an administrative agency may be reversed “if the substantial rights of the petitioner have been prejudiced because the administrative findings . . . are . . . made upon unlawful procedure.” Rose Care, 322 Ark. at 771, 912 S.W.2d at 411 (quoting Ark. Code Ann. § 25-15-212(h)(3)).
The rights that were prejudiced in this case were indeed substantial. Although the United States Supreme Court has definitively approved drug testing, even when conducted without a reasonable suspicion that the subject is “using,” the Court has said that the collection and testing procedure must satisfy reasonableness requirements in order to protect the employee’s constitutional rights. National Treasury Employees Union v. Von Raab, 489 U.S. 656 (1989); Skinner v. Railway Labor Executives Ass’n, 489 U.S. 602 (1989).
For such a test to be “reasonable,” it must be reasonably related to the objectives of the test “as actually conducted.” National Federation of Fed. Employees v. Weinberger, 818 F.2d 935 (D.C. Cir. 1987). We find the following guidelines instructive in determining whether a methodology of drug testing meets the reasonableness requirement:
In determining whether the searches “as actually conducted” were reasonable, the methodology of the searches must be examined in relation to the goals of the drug testing methods. If the, testing is not conducted in ways in which there is absolute confidence in the reliability of the results, then the purposes of testing are not secured. Treating someone for drug abuse who is not a drug user is counter-productive, as is disciplining someone who is not a drug user. ... It is critical that testing procedures be accurate and rehable.
*52Kevin B. Zeese, Drug Testing Legal Manual, § 5.01 [3] [d] at 5-18.39 (1995).
Stueart clearly had substantial rights that were placed at risk by the screening for drugs. Not only did he have a right to standards of reasonableness under the Fourth Amendment and under the Due Process Clause, but he also had a substantial interest in continued employment. Adherence to the procedure adopted by the agency to ensure absolute confidence in the reliability of the results was essential to protect these substantial interests. However, in affirming Stueart’s dismissal, the Commission ignored its own requirement that the results be confirmed by a medical review officer. This failure to follow its own rules deprived Stueart of the rights that the procedure was designed to protect.
The testimony of Dr. Henry F. Simmons at the hearing illustrates the importance of the responsibilities of the Medical Review Officer. Dr. Simmons, who serves as a medical review officer for the Little Rock Police Department, testified as follows:
[Assuming [the chain of custody is] intact, then I have to conduct, as a licensed physician ... an interview with the individual, unless he or she expressly declines the opportunity, to look for a medical basis for the result.
If I can find a medical basis, then I can make a negative report to the company. If I can’t, I have to make a positive report.
It is readily apparent from this description of this medical review officer’s duties, which are identical to those of the Medical Review Officer in the Arkansas State Police Program, that Stueart was deprived of the substantial rights protected by this procedure. He was deprived of the opportunity for an expert to determine whether there might be another explanation, besides marijuana use, for the positive result. Unlike the flaws in the chain of custody, this breach cannot be cured with affidavits and testimony.
Stueart properly raised this issue during the hearings before the Commission, and it ruled that the omission of the Medical Review Officer’s report was irrelevant. His argument and the Commission’s response are quoted in pertinent part as follows:
*53MR. YEARGAN: What I’m saying, sir, is the State Police has used the policy to terminate an individual, but it did not follow the procedures set forth in the policy to terminate him.
CHAIRMAN EILBOTT: I understand, but my question is, what is the relevance of a medical review officer where the substance is not a substance that is due to the tested employee’s proper use of a prescribed medication?
MR. YEARGAN: It wasn’t followed through. No one ever checked with him. It wasn’t done.
The Commission was clearly presented with Stueart’s argument that the failure to follow procedure, the required final step of a medical review officer’s evaluation, deprived him of substantial rights. The Commission determined — notwithstanding its own requirement for a final evaluation by a medical review officer — that the failure to follow its own rules was irrelevant.
We hold that the Commission’s failure to follow its own rules prejudiced the same substantial rights that the rules were promulgated to protect, and that this failure requires that we reverse the decision of the Commission to terminate Stueart. We direct that Stueart be reinstated and restored to the benefits of his employment, with appropriate consideration given to set-offs from earnings and benefits that he may have received. We reverse and remand to the trial court with instructions to remand to the Commission for further action consistent with the holdings of this opinion.
Glaze and Brown, JJ., dissent.
Tom Glaze, Justice, dissenting. I respectfully dissent. To the Arkansas State Police Commission’s credit, it has adopted a Drug Free Workplace policy, and in being called on to enforce that Policy, the Commission upheld Officer Thomas B. Stueart’s test, showing he was positive for marijuana use. While minor deviations occurred in following the State Police Department’s drug-testing rules when testing Stueart, the Commission upheld Stueart’s test because he failed to show how those deviations in any way tainted his test results.
If a criminal defendant had been arrested and tested for drugs, and procedural deviations occurred in the handling of the *54defendant’s specimen, our case law would require the defendant to show the evidence had not been tampered with before the test results would be excluded at trial. Crisco v. State, 328 Ark. 388, 943 S.W.2d 582 (1997) (The trial court must be satisfied within a reasonable probability that the evidence has not been tampered with, but it is not necessary for the State to eliminate every possibility of tampering). Here, the majority opinion declines to follow that accepted case law. Instead, the majority adopts a rule whereby Officer Stueart can have his test excluded from the Commission’s consideration even though he has failed to show his test had been tampered with or that he had been prejudiced by the manner in which the test had been processed. I submit that Officer Stueart’s rights were fairly protected, and the Commission’s right to terminate him was supported by substantial evidence — which, from the record, is the only issue addressed and decided by the Commission and circuit court.
Even the majority opinion expresses agreement that the evidence reflects that Officer Stueart’s voluntarily given specimen was properly taken from Stueart and sealed, and that the specimen remained sealed until it was tested at the Baptist Medical Center in Litde Rock. And while the majority seems to agree that no tampering appears to have occurred with Stueart’s specimen, it complains only that (1) Stueart failed to sign the urine-donor form, and (2) no medical review officer was asked to confirm Stueart’s positive test results or to consult with him to determine if there might be another explanation for the results.
Two observations need to be made. One, Stueart’s failure to sign the donor form was insignificant, and he made no assertion that anything occurred to the specimen as a result of his having failed to sign the form. In fact, the record reflects Stueart’s urine specimen had been taken and sealed in his presence, and the sample remained sealed until tested. Two, although no medical review officer was assigned to review Stueart’s drug test to determine another explanation for the positive results, Stueart had every opportunity to present an expert at his hearing to offer explanations that might have accounted for such results. Instead, Stueart merely relied on the testimony of Dr. Henry Floyd Simmons, who gave no alternative explanation except to say, categor*55ically, “I can tell you with utter confidence that as soon as I realized that the employee applicant had not signed the chain of custody, it (the test) would have been over.” Simmons stated that, in federal (not state) testing procedures, federal law required him to end his review in such circumstances. Under the State Police Department’s policy, a medical review officer’s function is to review all positive test results to determine whether such results were due to the employee’s proper use of a prescribed medication. Significantly, Dr. Simmons and Dr. Cashman testified that there is only one FDA drug that would actually make one positive for marijuana and that was Marinol — which Stueart was not taking. Dr. Simmons’s opinion concerning the law or legal validity of the chain of custody matter in this state case is irrelevant. State law places the Commission in the role to weigh the evidence to determine if discrepancies in the chain of custody in this case tainted Stueart’s test.
At this stage, I would point out that there was more than substantial evidence to support the Commission’s findings and termination decision, and that evidence, bearing on Stueart’s test results, was largely made by Dr. Dan Cashman, supervisor of the toxicology lab at Baptist Medical Center.
Again, the evidence reveals that the specimen was taken and sealed in Stueart’s presence and remained sealed until received for testing. Cashman testified that when he received the subject specimen, the external chain of custody information contained Stueart’s name which was entered into the computer. The specimen was assigned lab number 62149 which followed the specimen throughout the testing procedure. This numbered, urine specimen received a value number of .2872, meaning it tested positive for THC. The specimen was then retested and THC was confirmed. Cashman further explained that he had satisfied himself that the chain of custody was proper and that he had a clean specimen.
Finally, the majority opinion cites Regional Health Care Facilities, Inc. v. Rose Care, Inc., 322 Ark. 767, 912 S.W.2d 409 (1995), for the rule than an administrative agency may be reversed if the substantial rights of the petitioner have been prejudiced because the *56administrative findings are made upon unlawful procedure. That rule is good law, but, as discussed above, Stueart simply offered no evidence concerning how he was prejudiced by the minor deviations found in the State Police Department’s chain of custody in handling Stueart’s specimen. While I agree the Department failed to comply with its own drug-testing procedure and Stueart’s right was violated in this respect, Stueart was not shown to have been prejudiced. Obviously, his job termination is not the prejudice in issue. If that were true, a defendant charged with a drug offense could always have his test results suppressed when minor discrepancies occurred in the State’s chain of custody merely because he had been arrested, tried, and convicted.
In conclusion, the record here reflects Stueart’s urine sample had been taken and sealed in his presence, and the sample remained sealed until tested. Furthermore, substantial evidence reflects that the specimen tested was Stueart’s and the test showed positive for marijuana. The majority opinion fails to mention any prejudice which affected Stueart as a result of the procedural deviations that occurred in this matter. The only prejudice I can conceive is that Stueart was prevented from giving his explanation as to why his tests returned positive, but Stueart was given an extensive hearing where he could have offered such an explanation and did not.1
Because I believe substantial evidence exists that supports the Commission’s decision, I would affirm.
Brown, J., joins this dissent.