894 A.2d 690 (2006)
384 N.J. Super. 232
Kevin GEORGE, Petitioner-Respondent,
v.
CITY OF NEWARK, Respondent-Appellant.
Superior Court of New Jersey, Appellate Division.
Argued February 28, 2006.
Decided March 29, 2006.
Carolyn A. McIntosh, Assistant Corporation Counsel, argued the cause for appellant (Joanne Y. Watson, Corporation Counsel, attorney; Ms. McIntosh, on the brief).
William S. Greenberg, Newark, argued the cause for respondent (Paster & Greenberg, attorneys; Mr. Greenberg, on the brief).
Zulima V. Farber, Attorney General, attorney for The New Jersey Merit System Board (Todd A. Wigder, Deputy Attorney General, on the Statement In Lieu of Brief).
Before Judges KESTIN, LEFELT and R.B. COLEMAN.
The opinion of the court was delivered by
LEFELT, J.A.D.
After Police Sergeant Kevin George tested positive for marijuana, the City of Newark removed him from employment. Upon George's administrative appeal, the Merit System Board concluded that the City had unconstitutionally deprived George of his ability to contest the charge and reinstated the sergeant with back pay, benefits, and counsel fees. The City appeals and we vacate the Board's decision and remand for further proceedings.
Here is what happened. Pursuant to the Attorney General's Drug Screening Policy, the Newark Police Department, as part of its random drug testing program authorized by General Order 99-4, directed George to submit a urine specimen. The General Order required George to provide a minimum 60 ml of urine, with the specimen divided into a primary and split specimen, each containing at least 30 ml of urine. George complied, and his primary sample was delivered to the State Toxicology Laboratory.
The State Lab used a florescence polarization immunoassay (FPIA) test to screen George's primary sample preliminarily for the presence of eight illegal substances. The test's threshold level for cannabinoids, or marijuana, was 20 nanograms per milliliter (ng/ml). George's specimen tested positive for marijuana at 45.35 ng/ml.
Pursuant to the Drug Screening Policy, the positive test required the lab to perform a second confirmatory test of George's primary sample. This time the lab tested the sample by using gas chromatography/mass spectrophotometry (GC/MS) with the marijuana threshold at 10 ng/ml. The cut-off threshold level is set lower for GC/MS than for the FPIA, according to the State Lab, "because of the sensitivity of that particular methodology." George's sample again tested positive for marijuana at 23.0 ng/ml.[1]
Pursuant to the Attorney General's Law Enforcement Drug Testing Manual, before a positive result is reported, a State Lab medical officer reviews "the test results together with the medication information form." In this form the officer being tested discloses any medications taken within the last fourteen days. The medical officer seeks "to determine whether any of the substances listed on the form would explain the positive test result." If there is no explanation found in the form, the positive result is reported to the pertinent police department.
Based on the lab results, the City terminated George from his sergeant's position. After months of delay caused by some initial confusion as to whether a request had already been made, George requested by letter that his split sample be tested by LabCorp, a licensed independent laboratory that George had selected from a list the City had provided to him. Despite the delay in making the request, which exceeded the General Order's sixty-day time-frame, the Department sent the split sample directly to LabCorp.
According to testimony from a forensic toxicologist, who had been employed by the State Laboratory for twenty-seven years, the State had advised independent labs, such as LabCorp, to utilize the same testing threshold required by the Attorney General Drug Screening Policy for use in law enforcement drug tests. However, neither George nor anyone from the State Lab or the City verified that LabCorp had received this instruction or ensured in any fashion that LabCorp understood the threshold that had been utilized in the testing of George's primary sample.
Shortly thereafter, LabCorp tested the split sample using a threshold level of 50 ng/ml, which is the threshold commonly used in private industry, and reported that the sample was negative for the presence of marijuana. The City apparently reported the result to George some ten months after the test had been performed. The administrative record does not reflect whether the split sample sent to LabCorp had been destroyed by the time George was notified of the result; whether the sample had likely degraded over time so that it could not be accurately tested; or whether LabCorp could, at the time of the hearing, have produced the actual testing result.[2]
George denied knowingly ingesting marijuana and maintained that any exposure to marijuana came as a result of his job responsibilities as a police officer. George claimed that he had to handle marijuana evidence, which was often contained in plastic bags, and occasionally in the form of cigarettes.
George appealed his termination to the Merit System Board, and the matter was transmitted to the Office of Administrative Law. An administrative law judge conducted a hearing at which the parties stipulated to the chain of custody and the voiding and collection procedures employed by the City. The ALJ found George's testimony that he did not knowingly ingest marijuana to be unbelievable and issued an initial decision affirming the City's termination of George.
The Board concluded that it was the City's obligation to advise the independent lab of the proper threshold and reversed the initial decision. The Board determined LabCorp was "apparently" not advised of the threshold, and that the City's failure to so advise the lab rendered the testing process "fundamentally flawed" and deprived George of his "due process" opportunity to establish that the positive reading was false. The Board found the flaw "especially egregious in this case, since the initial test produced a positive result of 23 ng/ml, barely over the threshold of 20 ng/ml."
Consequently, the Board directed that George be reinstated and awarded back pay of $111,324.30; $631.71 to reimburse George for payments he made to maintain health insurance coverage within 30 days of receipt of the Board's decision; counsel fees of $10,255; and $1,000 for his expert witness costs. The City reinstated George on March 3, 2004,[3] and thereafter appealed to this court.[4]
Ordinarily, administrative decisions are accorded substantial deference. In re Musick, 143 N.J. 206, 216-17, 670 A.2d 11 (1996). We generally will overturn only those administrative determinations that are arbitrary, capricious, unreasonable, or violative of expressed or implicit legislative policies. Campbell v. Dep't of Civil Serv., 39 N.J. 556, 562, 189 A.2d 712 (1963). Any administrative decision that is unsupported by substantial, or sufficient, credible evidence in the record will be reversed as arbitrary, capricious, or unreasonable. In Re Taylor, 158 N.J. 644, 656-57, 731 A.2d 35 (1999); Henry v. Rahway State Prison, 81 N.J. 571, 579-80, 410 A.2d 686 (1980).
Preliminarily, we agree with the Board that George had a protected interest in his continued employment with the City. See In re Carberry, 114 N.J. 574, 583-84, 556 A.2d 314 (1989) (state trooper had a property interest in continued employment during good behavior, subject to termination only for cause). We also agree that George had a right to have the split sample tested at the same threshold as the primary sample. We further agree it is within the Board's authority to determine whether George, the City, or some other entity was obliged to inform LabCorp of the proper testing parameters. See DelRossi v. Dept. of Human Servs., 256 N.J.Super. 286, 292, 606 A.2d 1128 (App.Div.1992) (holding that "in disciplinary proceedings the Board's role is clearly adjudicatory and the Board acts on a case by case basis"). Upon review of decisions by the Board, we are bound to give weight to its presumed expertise in personnel matters. See Shahmoon Indus., Inc. v. Dept. of Health, 93 N.J.Super. 272, 282-83, 225 A.2d 699 (App.Div.1966), certif. denied, 49 N.J. 358, 230 A.2d 392 (1967). However, on questions of law, such as whether George in this case was deprived of due process, we are not bound by the agency decision and will review the agency's determination de novo. Taylor, supra, 158 N.J. at 658, 731 A.2d 35.
In this matter, we question whether the record contains sufficient evidence supporting the Board's decision that a due process violation occurred and that the entire testing process was "fundamentally flawed." Although the Board believed that the split sample, if properly tested, would have constituted persuasive evidence that the State Laboratory's results were "somehow invalid" or a "false positive," we find little to no support for this in the record.
George's expert directly attacked only the reliability of the FPIA screening test. The FPIA's manufacturer's statistics revealed that, at a threshold of 25 ng/ml, the GC/MS test confirmed only twenty-seven of fifty-five tests found positive for cannabinoids through FPIA testing. Thus, according to George, the FPIA had twenty-eight false positives. However, the record contains no statistics pertaining to false positives for the GC/MS test. In fact, there was no evidence challenging the reliability of the GC/MS results.
George's own expert was asked "if that same positive test for marijuana was confirmed on the GC/MS with a simultaneous quality control test, would that then change your mind" regarding the need for a "control run on the same day that the testing of a sample occurs?" The expert replied, "I have great faith in GC/MS confirmation. I don't have much trouble after reviewing the documents of the GC/MS work that was done on supposedly Mr. George's samples."
Later, in his examination, the expert explained he thought the GC/MS testing was done properly and that he had "absolutely no problem" with the GC/MS. His only problem was with the FPIA result, and he refuted and found unacceptable the City's expert's contention that of 80,000 tests there were no false positives. The ALJ then asked, if there were an FPIA positive reading that was confirmed by GC/MS testing of the same sample, "would that confirmation, in your mind, show the absence of a false positive?" George's own expert answered that such a result would indicate "it was a positive sample."
On redirect examination, George's counsel asked "[a]s you go down to the lower levels [for the threshold], closer to the limit of protection [with GC/MS testing] does it also make it less reliable than at the higher levels?" George's expert replied, "I think, by and large, GC confirmation is recognized as a gold standard." Needless to say, both City experts from the State Lab agreed with the conclusion that an FPIA positive reading confirmed by GC/MS indicates a correct result and that the GC/MS was the drug testing "gold standard."
Besides overestimating the potential invalidity of the State Lab's positive test, the Board also mistakenly underestimated the weight of the test results in reaching its conclusion that the entire testing process was flawed. The Board said the City's "flaw" in failing to advise LabCorp "of the threshold level at which to perform the second test . . . is especially egregious in this case, since the initial test produced a positive result of 23 ng/ml, barely over the threshold of 20 ng/ml."
The Board mistakenly considered the GC/MS result in light of the FPIA threshold. The FPIA screening threshold was set at 20 ng/ml and the test found George's sample positive at 45.35 ng/ml, but the confirmatory GC/MS result had a threshold of 10 ng/ml and George's sample tested positive at 23 ng/ml. Thus the confirmatory result may have been well under private industry's threshold of 50 ng/ml, but it was more than double the law enforcement threshold.
In addition, the Board failed to consider that there were neither rules nor guidelines that placed the notification burden on the City at the time of George's test. The process followed by the City required George to notify the City of his lab selection and the City thereafter to forward the split sample to the private lab. It is true that once instructed, the City acted without notifying George, and George had no contact with any of the transmitting documents. This process would, for practical reasons, make the City the logical entity to notify the private lab, but George also knew which lab he had selected and when he submitted his request for additional testing. There is no reason why some follow-up by George could not have also been expected.
Furthermore, George was able to challenge several aspects of the State Lab's test. George argued that the primary sample might not have been frozen promptly prior to transmission to the lab, that aged reagents and insufficient calibration and controls may have skewed the test, that there would have been high false positives in the FPIA test, that the results were "borderline," that there were numerous steps that "are subject to potential error," and that it was reasonably possible to test positive from handling marijuana in powder form. Thus, the "flaw" with regard to the split sample did not deprive George of his entire defense.
Moreover, in its rejection of interim relief pending appeal, the Board indicated that "[w]here as here, George denied ingesting marijuana, and the only evidence presented to refute that denial was the initial test result, the proper testing of the split sample was especially important." Here, the Board ignored the findings of the ALJ, which were at least in part based on credibility, that George "was well aware that he had ingested marijuana prior to the administration of the random test," and that the State Lab did retest the primary sample after the initial positive result.
In Capua v. City of Plainfield, 643 F.Supp. 1507, 1521 (D.N.J.1986), the court held that the inability of fire fighters to evaluate and review their test results or to elect to have their urine specimens retested by an independent facility offended fundamental fairness and due process. The City, in this case, argues that Capua does not apply because in that case the process precluded re-testing, where here George had an opportunity under the pertinent drug testing process to have his split sample retested. Instead, the City claims Forte v. South Woods State Prison, 2002 WL 341252, *1-*5, 2002 N.J. AGEN LEXIS 120, *1-*16 (OAL Jan. 22, 2002)[5] is of more significant import.
In Forte, a senior corrections officer was subjected to a random urine test. Id. at *1, 2002 N.J. AGEN LEXIS at *2. After the urine tested positive for cocaine, the officer requested that his split sample be tested by an independent laboratory. Id. at *1-*2, 2002 N.J. AGEN LEXIS at *3. Unfortunately, however, the independent laboratory, which was also LabCorp, tested the sample for opiates, not cocaine. Ibid. The Administrative Law Judge found that because the officer signed a request form acknowledging the accuracy of the information provided to the lab, he was at least partially responsible for the lab's error. Id. at *2-*3, 2002 N.J. AGEN LEXIS at *6. In addition, the officer never challenged the accuracy of the State Lab test and never denied using cocaine. Ibid. Consequently, the judge sustained the disciplinary action taken by the appointing authority for Forte's use and possession of cocaine. Id. at *3, 2002 N.J. AGEN LEXIS at *7.
Needless to say, George asserts that Forte is inapposite. Unlike the officer in Forte, George's only involvement with his split sample was to instruct the City to release it to LabCorp. Furthermore, George denied using marijuana and attacked the accuracy of the State Lab test. Although we agree that these distinctions differentiate George's case from Forte, we find other precedents more relevant and persuasive.
The failure of LabCorp to test the split specimen at the same threshold as the primary specimen, according to the Board and George, rendered the independent test useless. Without the precise numerical result of the independent test, this observation is accurate. George's situation, therefore, is analogous to the plight faced by criminal defendants when the State loses or otherwise damages or suppresses favorable evidence.
Without bad faith on the part of the State, "failure to preserve potentially useful evidence does not constitute a denial of due process of law." Arizona v. Youngblood, 488 U.S. 51, 57, 109 S.Ct. 333, 337, 102 L.Ed.2d 281, 289 (1988). It is only the suppression of exculpatory evidence that violates due process "irrespective of the good faith or bad faith of the prosecution." Brady v. Maryland, 373 U.S. 83, 87, 83 S.Ct. 1194, 1196-97, 10 L.Ed.2d 215, 218 (1963).
In this case, a correct independent test result cannot be considered exculpatory at the time the split specimen was sent to LabCorp. At best it was potentially exculpatory. Even a negative result at zero ng/ml or well below the correct threshold for marijuana would still only be evidence of a mistake in the two positive State Lab results and not clearly exculpatory.
When there has been "suppression, loss, or destruction of physical evidence," our courts focus on three factors to determine whether a due process violation has occurred. State v. Dreher, 302 N.J.Super. 408, 483, 695 A.2d 672 (App.Div.), certif. denied, 152 N.J. 10, 702 A.2d 349 (1997), cert. denied, 524 U.S. 943, 118 S.Ct. 2353, 141 L.Ed.2d 723 (1998). These factors are: "(1) the bad faith or connivance by the government; (2) whether the evidence was sufficiently material to the defense; and (3) whether the defendant was prejudiced." Ibid. (citing State v. Hollander, 201 N.J.Super. 453, 479, 493 A.2d 563 (App.Div.), certif. denied, 101 N.J. 335, 501 A.2d 983 (1985)); see also State v. Serret, 198 N.J.Super. 21, 26, 486 A.2d 345 (App.Div.1984). Where the lost evidence is merely potentially exculpatory, the "court's finding of bad faith is crucial." See State v. Ruffin, 371 N.J.Super. 371, 392, 853 A.2d 311 (App.Div.2004); see also State v. Brent, 265 N.J.Super. 577, 586-87, 628 A.2d 372 (App.Div.1993), rev'd on other grounds, 137 N.J. 107, 644 A.2d 583 (1994) "In the absence of bad faith, relief should be granted only where there is a `showing of manifest prejudice or harm' arising from the failure to preserve evidence." Dreher, supra, 302 N.J.Super. at 489, 695 A.2d 672 (quoting DeVitis v. N.J. Racing Comm'n, 202 N.J.Super. 484, 494, 495 A.2d 457 (App.Div.), certif. denied, 102 N.J. 337, 508 A.2d 213 (1985)).
Applying these three factors to George's situation leads us to conclude that a remand is in order. First, there is no evidence the City acted in bad faith when it did not advise LabCorp of the proper testing threshold. In fact, the failure seems completely understandable given the absence of preexisting direction or instruction regarding such a requirement.
However, the record indicates that the City waited over ten months before revealing the LabCorp test result to George. We do not know the reason for this delay or what LabCorp did with the split specimen or what the lab did during the delay with the specific numerical result of the test. In addition, we do not know precisely why LabCorp tested the specimen at the higher threshold or when, if ever, the State advised the private laboratories of the protocol governing law enforcement drug testing.
Second, the evidence was obviously material to the defense. The split specimen, when correctly tested, possessed an apparent potential exculpatory value. This evidence was unique in that, without it, George would be unable to obtain comparable evidence of the primary urine sample that was tested by the State Lab. See California v. Trombetta, 467 U.S. 479, 489, 104 S.Ct. 2528, 2534, 81 L.Ed.2d 413, 422 (1984); Hollander, supra, 201 N.J.Super. at 479-80, 493 A.2d 563.
Third, the gravity of any prejudice allegedly suffered by George seems to us to be factually disputed. The record contains no evidence of any false positive rate associated with the GC/MS test. If the false positive rate or any potential error rate is low for this test, there is a serious question regarding whether defendant was prejudiced by not having a proper split specimen test result. In addition, George had the opportunity to challenge the State Lab test results and conducted a vigorous defense. Furthermore, the ALJ found George's testimony denying any voluntary marijuana ingestion to be incredible, though this finding was based only in part on credibility, with some of the judge's assessment obviously dependent on the testing evidence.
Consequently, we vacate the Board's decision and remand for further proceedings so the Board, after the necessary factual exploration and evaluation by an ALJ, may reconsider its conclusions that the City's failure constituted a due process violation and rendered the testing process "fundamentally flawed." On the remand, the parties shall be accorded an opportunity to address the open issues discussed in this opinion, including the statistical reliability of the GC/MS test, the City's reasons for delaying disclosure of LabCorp's test results, any information supplied by the State to licensed independent labs, and LabCorp's actions and practices from receipt of the split specimen until the date of the hearing before the ALJ.
Vacated and remanded for further proceedings consistent with this decision.[6]
NOTES
[1] According to the State's forensic toxicologist, the FPIA measures five different marijuana metabolites, while the GC/MS isolates only the major metabolites, so the GC/MS finding is "always less than what you [find on the] FPIA."
[2] George's appellate brief argues that by the time the City notified George "the sample could not be retested at the lower cut-off levels," but there is no direct support in the record for that contention. In addition, the Board's brief argues, without support in the record, that "the negative result from his split sample is useless to George since LabCorp would not have preserved the actual reading that may have exonerated George."
[3] After reinstatement, the City removed George's name from an eligible list for promotion to police lieutenant based upon his unsatisfactory employment record. Upon George's appeal, the Board affirmed this removal.
[4] We stayed the monetary award pending the instant appeal.
[5] This initial decision was affirmed by the Board on September 10, 2002 In re Forte, DOP XXXX-XXXX, 2002 WL 32539188, final decision, (February 27, 2002), http://lawlibrary. Rutgers.edu/oal/ final/csv 00659-01.pdf.
[6] We add for the parties' assistance on the remand that we find no merit in the City's arguments objecting to the Board's back pay award.