In this appeal we are called upon to assess the constitutionality of the United States Department of the Army's practice of subjecting certain of its civilian employees to compulsory, random toxicological urine testing. Implemented in 1986, the testing program requires more than 9,000 of the Army's 450,000 civilian employees be tested in any of four circumstances, including "[pe]riodically ... on the basis of random criteria." Appellees, National Federation of Federal Employees and American Federation of Government Employees--union representatives of the covered employees--filed actions to enjoin the random aspect of the program. The District Court concluded that suspicionless urinalysis testing constituted an unreasonable search in violation of the Fourth Amendment, and accordingly entered a nationwide preliminary injunction against such testing. National Fed'n of Fed. Employees v. Carlucci, 680 F. Supp. 416 (D.D.C.1988).1 After this Court ordered the preliminary injunction stayed pending appeal, National Fed'n of Fed. Employees v. Carlucci, No. 88-5080 (D.C.Cir. Mar. 30, 1988), the District Court permanently enjoined the Army from implementing the random aspect of its testing program. National Fed'n of Fed. Employees v. Carlucci, 690 F. Supp. 46 (D.D.C.1988). The District Court stayed the permanent injunction pending resolution of this appeal. Id. at 55. We affirm in part, vacate in part, and remand for further proceedings.

In 1971, Congress directed the Secretary of Defense to "prescribe and implement procedures ... [to] identify, treat, and rehabilitate members of the Armed Forces who are drug or alcohol dependent persons." PUB.L. NO. 92-129, tit. V, Sec. 501(a), 85 Stat. 348, 361 (1971). Pursuant to this command, the Army initiated mandatory toxicological testing of its military forces, a program that by 1982 resulted in comprehensive military testing. See Joint Appendix ("J.A.") at 237. On April 8, 1985, the Department of Defense issued Directive 1010.9, authorizing urinalysis testing of civilians occupying or applying for "critical jobs,"2 and requiring each "DoD Component" to submit proposals for critical job designations. The Directive provided that covered civilians "may be required to participate in urinalysis testing" in four circumstances: (1) before appointment or selection; (2) "[p]eriodically ... on the basis of neutral criteria;" (3) upon probable cause to believe on-duty drug impairment; and (4) as part of a "mishap or safety investigation." Directive 1010.9 para. F.2.a. (1)-(4).

In order to implement the Directive, in 1986 the Army promulgated Army Regulation 600-85, Interim Change I11 ("AR 60085"). AR 600-85 provides for testing under the same four circumstances listed in the Department's Directive.3 Testing was authorized for any of fourteen classes of critical jobs, categorized as follows: aviation, guard and police, "personnel reliability program," and the Alcohol and Drug Abuse Prevention and Control Program. Id. Table 5-1.4 A brief description of classes of critical jobs accompanied the Regulation.

After Directive 1010.9 and Army Regulation 600-85 were issued, the Department of Health and Human Services, in accordance with Executive Order 12,564, "Drug-Free Federal Workplace," 3 C.F.R. Sec. 224 (1987), promulgated regulations to govern federal drug testing. See "Mandatory Guidelines for Federal Workplace Drug Testing Programs," 53 Fed.Reg. 11,970 (April 11, 1988) ("HHS Regulations" or "HHS Reg."). Congress ordered the Army to comply with the HHS Regulations. See Supplemental Appropriation Act of 1987, PUB. L. NO. 100-71, Sec. 503, 101 Stat. 391, 468-71 (codified at 5 U.S.C. Sec. 7301 note (Supp. V 1987)). The parties agree that the constitutionality of the Army's testing program is to be judged with reference to the procedures as they currently exist, rather than as originally announced. See Brief for Appellants at 6 n. 1; Brief for Appellees at 8 & n. 3; see generally California Bankers Ass'n v. Shultz, 416 U.S. 21, 53, 94 S. Ct. 1494, 1513, 39 L. Ed. 2d 812 (1974).

The HHS Regulations provide that, upon arriving at an assigned time and collection site, the subject employee is to prepare for the test by removing unnecessary outer garments and washing his hands. Thereafter the subject employee is to "remain in the presence of the collection site person." HHS Reg. Sec. 2.2(f)(6), 53 Fed.Reg. at 11,980. The sample is to be provided "in the privacy of a stall or otherwise partitioned area," id. Sec. 2.2(f)(7), 53 Fed.Reg. at 11,981, "unless there is reason to believe that a particular individual may alter or substitute the specimen to be provided," id. Sec. 2.2(e), 53 Fed.Reg. at 11,980.

The sample is then sent to an HHS-approved laboratory, under procedures ensuring a secure chain of custody, id. Sec. 2.2(f)(17)-(26), 53 Fed.Reg. at 11,981, where it is tested for proscribed drugs.5 The sample is initially subject to a radioimmunoassay (RIA) test, and if the results of that test exceed certain cut-off levels, to a gas chromatography/mass spectrometry (GC/MS) test. See J.A. at 242. A sample is considered positive only if it is positive on both tests. HHS Reg. Sec. 2.4(g)(2), 53 Fed.Reg. at 11,983.

Any applicant who tests positive for proscribed drugs "will be denied further consideration" for employment, AR 600-85 para. 5-14c (3)(a), and employees who test positive "may be subject to adverse action," identified as reassignment or demotion to a noncritical job, or "if there is no job available for which the employee is qualified, separat[ion] from the service." Id. para. 5-14c (1)(b). Among the potential penalties for refusing to submit to a test or attempting to adulterate the test sample is "[r]emov[al] from Federal service." Id. para. 5-14c (4)(b) & c (5). The Regulation provides that any employee testing positive "shall, if eligible," be offered counseling or treatment. Id. para. 5-14e (4). Collateral use of the test results is proscribed by regulation and statute alike. See HHS Reg. Sec. 2.8, 53 Fed.Reg. at 11,986; Sec. 503(e), PUB. L. NO. 100-71, 101 Stat. 471; cf. AR 600-85 para. 5- 14e(5).

Appellees filed actions challenging the legality of the random aspect of the Army testing program.6 The Army argued that compulsory testing was justified by compelling governmental interests in the safety, security, and integrity of its workforce. The District Court, relying on our since vacated decision in Jones v. McKenzie, 833 F.2d 335, 340-41 (D.C.Cir.1987), vacated sub nom. Jenkins v. Jones, --- U.S. ----, 109 S. Ct. 1633, 104 L. Ed. 2d 149 (1989), replaced, 878 F.2d 1476 (D.C.Cir.1989), ruled only on appellees' Fourth Amendment challenge, concluding that mandatory, random urinalysis testing constituted an unreasonable search in violation of the Fourth Amendment. Although it conceded the Army's "compelling safety interest ... in maintaining a drug-free workplace," the Court reasoned that because the Army failed to show that drug testing measures on-the-job impairment, "urinalysis drug testing lacks the necessary nexus to the employer's safety concern to satisfy the Fourth Amendment." Carlucci, 680 F. Supp. at 418 (emphasis in original). The District Court dismissed the Army's nonsafety interests as insufficiently compelling to warrant "the substantial intrusion of mandatory, random urinalysis." Id. Accordingly, the Court enjoined such testing "unless the test is based upon a reasonable, articulable, and individualized suspicion that a specific employee is under the influence of drugs or alcohol while on duty." Id. at 436. On July 6, 1988, the District Court made final its injunction against such testing.

In decisions issued during the pendency of this appeal, the Supreme Court rejected Fourth Amendment challenges to two federally mandated drug testing programs. In the first, National Treasury Employees Union v. Von Raab, --- U.S. ----, 109 S. Ct. 1384, 103 L. Ed. 2d 685 (1989), the Court sustained as reasonable a United States Customs Service testing program that required pre-ascension testing of employees seeking positions directly involving the interdiction of illegal drugs or that require the incumbent to carry a firearm. Id. 109 S.Ct. at 1397. The Court held that, on balance, the "Government's compelling interests in safety and in the integrity of our borders" outweighed the privacy interests of these promotion-slated Customs agents, and thus warrantless, suspicionless testing, while a search, was not unreasonable. Id. at 1394, 1397. The Court did not pass on the constitutionality of the third category of subject employees--those whose new duties would require that they "handle classified material"--because the record did not reflect adequately whether the covered employees were "likely to gain access to sensitive information." Absent such a showing, the Court was unable to discern "whether the Service has defined this category of employees more broadly than necessary to meet the purposes of the Commissioner's directive." Id. at 1397. The Court therefore remanded the case to the Court of Appeals, so that it might "clarify the scope of this category of employees subject to testing." Id.

In Skinner v. Railway Labor Executives' Association, --- U.S. ----, 109 S. Ct. 1402, 103 L. Ed. 2d 639 (1989), the Court sustained against Fourth Amendment challenge regulations promulgated by the Federal Railroad Administration ("FRA") authorizing certain carriers to conduct suspicionless toxicological tests. The regulations mandate blood and urine tests be administered to covered employees following major train accidents or incidents, and authorize, but do not require, railroads to administer breath and urine tests, or both, to employees who violate certain safety rules. See 49 C.F.R. Sec. 219.201(a)(1)-3; id. Sec. 219.301(b)(2)-(3). A variety of factors established the reasonableness of the program, among them, (1) the "limited" intrusions occasioned by the testing procedures, id. 109 S.Ct. at 1417; (2) the diminished expectation of privacy that attaches to employment in an "industry that is regulated pervasively to ensure safety," id. at 1418; and (3) the government's "compelling" or "surpassing" interest in railway safety, id. at 1419, 1422, an interest that could not adequately be protected by testing only upon individualized suspicion, id. at 1420-21.

Von Raab and Skinner make clear, consistent with our own precedent, see National Fed'n of Fed. Employees v. Weinberger, 818 F.2d 935, 942 (D.C.Cir.1987), that the Fourth Amendment is applicable to governmentally compelled urinalysis. See Skinner, 109 S. Ct. at 1412-13; Von Raab, 109 S. Ct. at 1390. The cases also establish that testing, if justified by needs beyond the normal need for law enforcement, need not necessarily be supported by a warrant, probable cause, or any level of particularized suspicion. In this case, like Von Raab and Skinner, the government's interests in testing are clearly other than the ordinary need for law enforcement. The testing program authorized by Directive 1010.9 has three stated purposes: (1) assisting in determining employee fitness, (2) identifying and treating drug abusers, and (3) maintaining national security and the internal security of the Defense Department. Directive 1010.9 para. D.1-3. See also AR 600-85 para. 5- 14a (1)-(3). Law enforcement appears nowhere among the program's stated goals, and for that matter, non-consensual disclosure of test results to police authorities is proscribed both by regulation and statute. See HHS Reg. Sec. 2.8, 53 Fed.Reg. at 11,986; Sec. 503(e), PUB. L. No. 100-71, 101 Stat. 471.

Even where a testing regime is motivated by needs other than law enforcement, the question remains whether the governmental interests asserted are so substantial as to justify the suspicionless privacy intrusions at issue. In Von Raab and Skinner the Supreme Court laid out a balancing test that, while not self-executing, focuses our attention on a single question: Does the government's need to conduct the suspicionless searches outweigh the privacy interests of the covered employees in such a fashion that it is "impractical to require a warrant or some level of individualized suspicion?" If so, the Army's testing scheme cannot be deemed unreasonable. See Von Raab, 109 S. Ct. at 1390, 1392. Accord Skinner, 109 S. Ct. at 1414.

Initially, appellees ask that we find Von Raab and Skinner of "little or no impact" on the present case, since the Army tests on a random, rather than post-accident or pre-ascension basis. Supplemental Brief for Appellees, American Federation of Government Employees at 2.7 To settle this contention, we need look no further than our recent decision in Harmon v. Thornburgh, 878 F.2d 484 (D.C.Cir.1989), where, while recognizing that a random plan might plausibly be considered different in kind from the programs approved by the Supreme Court, we concluded that "the random nature of the [subject] testing plan is a relevant consideration," but does not "require[ ] us to undertake a fundamentally different analysis from that pursued by the Supreme Court." At 489 (emphasis in original).

Throughout this litigation appellees have argued that testing is unreasonable because the Army has no interest in off-duty conduct that does not have some effect on on-duty performance. "At best," appellees correctly observe, "a positive result indicates only that at some time in the past, maybe days or weeks before the test, the tested employee ingested a drug in some unknown quantity." Brief for Appellees at 36 (footnote omitted). The District Court, relying on the Ninth Circuit's opinion in Railway Labor Executives' Association v. Burnley, 839 F.2d 575, 588 (9th Cir.1988), reversed sub nom. Skinner, 109 S. Ct. 1402, concluded that the inability of testing to differentiate between drug use that produces on-the-job impairment and that which does not was fatal to the program's constitutionality. Carlucci, 680 F. Supp. at 434.

As a means of uncovering the information in which the Army has a legitimate interest, we agree that urine testing--unaided by blood or breath testing--is a blunt instrument. A single positive urine test result is silent as to when or how much of the drug was taken, the pattern of the employee's drug use, or whether the employee was intoxicated when the test was given. See J.A. 90-91, 102-04. While the question may have been a close one before the Supreme Court's recent pronouncements, particularly in view of our original opinion in Jones, see Jones v. McKenzie, 833 F.2d at 340-41, Von Raab and Skinner convince us that the District Court, like the Ninth Circuit in Skinner, adopted an overly restrictive view of the Fourth Amendment's command of reasonableness.

In Skinner, the Ninth Circuit, like the District Court in the present case, concluded that drug testing was constitutionally infirm because of its inability to differentiate drug use that results in on-duty impairment from that which does not. Burnley, 839 F.2d at 588. The Supreme Court concluded that this aspect of the Ninth Circuit's decision was flawed on several fronts. First, the analysis overlooked the principle that evidence " 'need not conclusively prove the ultimate fact in issue,' " so long as it makes a fact of consequence more or less probable. Skinner, 109 S. Ct. at 1421 (quoting New Jersey v. T.L.O., 469 U.S. 325, 345, 105 S. Ct. 733, 744, 83 L. Ed. 2d 720 (1985)). Second, the Skinner Court noted that under the Federal Railroad Administration regulations, urinalysis test results serve only as "a secondary source of information," as the regulations "plac[e] principal reliance on the results of blood tests, which unquestionably can identify very recent drug use." Id. Thus together, blood and urine tests provide a "highly effective means of ascertaining on-the-job impairment." Id. In the very same paragraph the Court noted a third flaw in the Ninth Circuit's analysis: It "failed to recognize that the FRA regulations are designed not only to discern impairment but also to deter it." Id.

Because Skinner can reasonably be read to depend for its conclusion on the co-administration of blood and urine tests, the Court's analysis in Von Raab--where the Customs Service administered only urine tests--is perhaps the more telling. In sustaining the Customs Service's practice of urine testing those whose new duties would require carrying a firearm, the Court identified the governmental interest as preventing on-the-job impairment--"the risk that employees who may suffer from impaired perception and judgment will be promoted to positions where they may need to employ deadly force." 109 S. Ct. at 1393. But the Court did not even advert to the inability of the testing procedures to differentiate on- and off-duty drug use or impairment. Rather, it referred only to the Service's broad responsibility to "ensur[e] against the creation of this dangerous risk," id., a risk that urinalysis testing reasonably, albeit imperfectly, helped detect and prevent. The same analysis obtains in the present case. To hold as did the District Court--that only a procedure able to detect on-duty impairment is constitutionally permissible--would be tantamount to holding urinalysis testing unreasonable per se8--a conclusion inconsistent with the Supreme Court's recent teachings.

It is readily apparent that the Army has a compelling safety interest in ensuring that the approximately 2,800 civilians who fly and service its airplanes and helicopters are not impaired by drugs. Employees in each of the covered positions--air traffic controllers, pilots, aviation mechanics and aircraft attendants--perform tasks that are frought with extraordinary peril: A single drug-related lapse by any covered employee could have irreversible and calamitous consequences.

The Supreme Court adverted to the government's compelling interest in maintaining air safety in Von Raab, quoting with apparent approval from United States v. Edwards, 498 F.2d 496 (2d Cir.1974), which sustained the requirement that all passenger carry-on luggage be searched before it could be stored aboard commercial airliners. See 109 S. Ct. at 1395 n. 3. In Edwards, Judge Friendly reasoned that " '[w]hen the risk is the jeopardy of hundreds of human lives and millions of dollars of property ... th[at] danger alone meets the test of reasonableness.' " Id. at 500 (citation omitted; emphasis in original). The Von Raab Court likewise concluded that where "the possible harm against which the Government seeks to guard is substantial, the need to prevent its occurrence furnishes an ample justification for reasonable searches calculated to advance the Government's goal." 109 S. Ct. at 1395 (footnote omitted). Compelling as the government's interest is in preventing the promotion of drug users to positions involving interdicting drugs and carrying firearms, it pales in comparison with the quintessential risk of destruction to life and property posed by aviation.

Without quarrelling with the profound significance of these interests, appellees contend that testing is nevertheless unreasonable because the risk posed by a drug-impaired aviation employee is abated by the Army's extensive system of safeguards and supervision already in place. Brief for Appellees at 6 n. 2. This contention is a derivative of an argument appellees successfully pressed before the District Court: Drug testing is unreasonable because the Army's interests could be adequately achieved through the implementation of less intrusive alternatives such as supervisory observation and neurobehavioral testing. The District Court agreed, concluding that trained supervision and neurobehavioral testing would be both less intrusive and better able to detect those who are impaired at work because of drug use. Carlucci, 680 F. Supp. at 434.

While the Supreme Court has indicated that "the supervision to which the [ ] employees are already subject" is a relevant factor in determining the reasonableness vel non of testing, Von Raab, 109 S. Ct. at 1397, in Skinner the Court made clear that the reasonableness of a particular technique does not " 'necessarily or invariably turn' " on the existence of less intrusive alternatives. 109 S. Ct. at 1419 n. 9 (quoting Illinois v. Lafayette, 462 U.S. 640, 647, 103 S. Ct. 2605, 2610, 77 L. Ed. 2d 65 (1983)). Like appellees, respondents in Skinner maintained that the government's interests could be adequately addressed by enforcing proscriptions already in force and by training supervisory personnel in drug impairment detection. The Court rejected this claim in part because "insistence on less drastic alternatives would require us to second-guess the reasonable conclusions drawn by the FRA after years of investigation and study." Id. The same analysis governs the present case.

Initially, it does not appear from the record that neurobehavioral testing--which counsel for appellees described as "a computerized manual dexterity test," Transcript of Proceedings ("Tr.") at 26 (Oct. 24, 1989)--is either less intrusive or less degrading than urinalysis testing. Nor does the record reveal that it would be feasible to subject aviators or their work-product to the kind of day-to-day scrutiny that would appear necessary in order for supervisory observation to be effective. Cf. Von Raab, 109 S. Ct. at 1395. Moreover, the employees who are subject to testing in this category "can cause great human loss before any signs of impairment become noticeable to supervisors or others." See Skinner, 109 S. Ct. at 1419. While the Army, unlike the Customs Service, has not expressly considered the various alternatives to drug screening, it is, in light of its experience from fifteen years of testing its military personnel, better able than we to assess the efficacy of urinalysis testing.

The second category of tested civilians occupy "[c]hemical and nuclear surety positions" within the Personnel Reliability Program ("PRP"). AR 600-85, Table 5.1. This category includes nuclear reactor operators, nuclear weapons technicians, chemical ammunition maintenance specialists, quality assurance personnel, material handlers, laboratory workers, and "intrusion detection system maintenance personnel." J.A. at 204. Chemical and nuclear sureties are ostensibly "responsible for the custody, transportation, storage, maintenance, demilitarization, and security of surety materials." J.A. at 204.

We are, for reasons unrelated to the seriousness of the governmental interests invoked, unable to determine the reasonableness of this aspect of the testing program. On the current record it appears that among the employees tested as part of the PRP program are secretaries, engineering technicians, research biologists, and animal caretakers. See J.A. 533-65. It is not intuitive that a secretary would be "responsible for the custody, transportation, storage, maintenance, demilitarization, and security of surety materials." Nor is there any evidence in the record to assuage our doubt. Indeed, the only record material, an affidavit from a subject secretary, is to the effect that although she had been issued "protective equipment," she had not used it for anything but training. Nor does it appear that she has access to chemical or nuclear materials. See J.A. at 535-36.

We are likewise unable to discern the reasonableness of the inclusion of animal caretakers, of whom the record reveals nothing more than access to animals located in secured areas. See J.A. at 560. Although we might be more inclined to deem reasonable the testing of engineering technicians and research chemists who are exposed to an unspecified "chemical agent," J.A. at 540, 550, or of research biologists, who have access to facilities located in secured areas maintained under the Chemical Surety Program, J.A. at 545, 555, we see little reason to resolve these questions of the category's coverage when the District Court may on remand develop a more complete factual record.

It may very well be, as the record appears to reflect, that all of these individuals are required to report all prescription drug use, to update background information, and to be regularly medically examined. See J.A. at 292. If such is the case, the employees' privacy expectations may be considerably diminished. See supra p. 608 & infra pp. 612-613. It may also be that these employees have as part of their assigned duties access to highly dangerous chemical and nuclear material and sensitive information, such that compelling safety and security interests would be advanced by toxicological testing. Cf. Skinner, 109 S. Ct. at 1419 (adverting to governmental interest in preventing drug use by those "who have routine access to dangerous nuclear power facilities") (citations omitted); Harmon, 878 F.2d at 490 n. 8 (permitting testing of those " 'whose assigned position duties include maintaining, storing or safeguarding a controlled substance' ") (citation omitted). The simple fact is that while the Army has averred serious interests to be furthered by testing PRP employees, we are not able to assess the reasonableness of the category as drawn. Von Raab counsels that in such circumstances our responsibility is to remand. 109 S. Ct. at 1397.

Appellants maintain that compelling safety and security interests justifying testing the approximately 3,700 civilians employed in positions pertaining to law enforcement, most notably civilian police and guards. See J.A. at 203-04. The threat to public safety posed by drug-impaired, gun-toting civilian officers is manifest; the most apparent risk arising from the fact that these personnel are armed, "many with automatic weapons," J.A. at 362, and may have access to other munitions. At the Anniston Army Depot in Georgia, for example, guards have access not only to the .38 caliber revolvers that they carry, but also to shotguns, semi-automatic and automatic rifles, machine guns, grenades, and grenade launchers. J.A. at 275.

Appellees contend that the risk to public safety posed by a drug-impaired Army guard is more theoretical than real, as "the work of an Army guard is far less dangerous and far more mundane than that of their counterparts on the outside, who may have to use their firearms at any moment." Supplemental Brief for Appellees at 6. Our reading of Von Raab convinces us that this contention, while substantial, must fall. In Von Raab, the Supreme Court determined that the public interest "to prevent the promotion of drug users to positions that require the incumbent to carry a firearm" warranted pre-ascension testing of such personnel. 109 S. Ct. at 1393. While it might be argued that we are not in fact presented with " 'duties fraught with ... risks of injury,' " id., the Court's opinion is silent on the likelihood that the threatened harm would occur. Nor is it obvious to us that Customs agents who are uninvolved with drug interdiction more frequently discharge their weapons than do the Army guards in question. It certainly appears that here, as in Von Raab, "the public should not bear the risk that employees who may suffer from impaired perception and judgment will be promoted to positions where they may need to employ deadly force." Id. The very fact that these employees' duties entail the possession of deadly weapons belies appellants' argument that there is no real danger of their using them.

Furthermore, a second, and perhaps more grievous threat is posed by a drug-impaired guard in the present case; stationed as he is in an extraordinarily hazardous setting. Several of the installations where the security personnel are stationed test the newest and most technologically advanced weapons systems. For example, the Aberdeen Proving Ground and White Sands Missile Range house nuclear reactors. J.A. at 250. The Anniston Depot stores more than 2 million small arms and munitions ranging from .22 caliber to 1,000 pound bombs, toxic chemical agents, and thousands of rockets. J.A. at 272; Tr. at 9. In such settings, the careless discharge of a weapon or failure to prevent or detect fire, theft, or accidental or willful destruction to property carries the attendant risk of catastrophic consequences. While counsel for appellees argued that we ought not be influenced by "nightmare scenarios," Tr. at 23, as we have already explained, the reasonableness of the testing plan is determined in part by the potential gravity of the feared harm. In the context of this case, risk assessment weighs heavily, if not determinatively, in favor of reasonableness.

Nor can we agree with appellees that a different balance should be struck in the present case than in Von Raab and Skinner because "unlike Customs agents and railroad employees, the Army's civilian guards ... possess undiminished privacy expectations." Supplemental Brief for Appellees at 6-7. We are, to say the least, baffled by this contention. Civilian guards and police undergo a variety of privacy diminishing tests and investigations as a condition of employment in a high-security, military context. Not only are security guards subject to a series of pre-employment screenings, including investigations into criminal, medical, employment, and educational histories, once hired, security personnel are subject to quarterly blood tests, annual medical examinations, annual criminal records checks, and periodic security investigations. J.A. at 275. Von Raab explicitly teaches that such intrusions may be expected to reduce one's expectation of privacy, and, hence, may enhance the reasonableness of testing. Von Raab, 109 S. Ct. at 1397 ("background investigations, medical examinations, or other intrusions ... may be expected to diminish ... expectations of privacy in respect of a urinalysis test"). See also Skinner, 109 S. Ct. at 1418 (noting that expectation of privacy reduced by requirement of periodic physical examinations). These employees' expectations of privacy are also reduced by the requirement that they report any illness or incident that results in use of prescription drugs. J.A. at 275.

In short, "[t]he operational realities of the workplace," O'Connor v. Ortega, 480 U.S. 709, 717, 107 S. Ct. 1492, 1502, 94 L. Ed. 2d 714 (1987) (plurality opinion), are such that "a diminished expectation of privacy attaches to information relating to the physical condition of covered employees," Skinner, 109 S. Ct. at 1419. Because the civilian guards' expectations of privacy are severely reduced as a condition of employment in a high-security, military context, mandatory, random urinalysis testing constitutes a modest additional privacy intrusion, and cannot, on this basis, be deemed unreasonable.

The fourth category of tested positions includes those "[c]ivilian treatment staff of the Alcohol and Drug Abuse Prevention and Control Program (ADAPCP) whose duties involve direct contact with clients, civilian employees of the Army's Drug Testing Laboratories (DTL), and all employees involved in the chain of custody process for biochemical testing." J.A. at 203. Table 5.1 breaks down the employees into two classes: "direct service staff"--who, as best we can determine, are primarily drug counsellors--and employees at Army forensic drug testing laboratories. The Army initially justified testing these drug counsellors and laboratory workers as necessary to maintain the testing program's "continued credibility" and to ensure proper identification and treatment among other personnel. J.A. at 203. In this litigation the Army has similarly argued that testing is justified on the basis of its "strong integrity interest in insuring an absence of all drug use" within the Army. Brief for Appellants at 34. In short, for these civilians "drug use is inconsistent with the job." Id. at 34-35.

The Army's assertion of "credibility" and "integrity" interests to support testing these civilians is akin to the rationale employed by Von Raab to approve the Customs Service's program of testing its front-line interdiction personnel. Although in Harmon we characterized the interest sanctioned by the Supreme Court as the "government's interest in maintaining the integrity of its workforce," Harmon, 878 F.2d at 488, we noted that, taken to its logical end, the integrity rationale would justify the random testing of every federal employee--a result that would be inconsistent with Von Raab's essential teachings. Id. at 11 & n. 7. Thus, in our view Von Raab rested not so much on the Service's undifferentiated interest in ensuring its reputation and credibility as on the myriad dangers posed by drug-using interdiction agents. The Court identified threats to the "national interest in self-protection," 109 S. Ct. at 1393, to "our Nation's first line of defense" against the scourge of illegal drugs, id. at 1392, and to the health and safety of the agents, who daily encounter a criminal element that "do[es] not hesitate to use violence to protect [its] lucrative trade and avoid apprehension," id.

In Harmon we concluded that the Justice Department's generic interest in the integrity of its workforce did not justify the testing of any of the three categories there under review: prosecutors in criminal cases, employees with access to grand jury proceedings, and personnel holding top secret national security clearances. Harmon, 878 F.2d at 491. Although we conceded, arguendo, the government's interest in preventing drug use by prosecutors "having substantial responsibility for the prosecution of federal drug offenders," id. (footnote omitted), we concluded that the responsibility rested with the Department of Justice to propose such a classification, id. at 493-95.

The most basic rationale to support testing ADAPCP personnel is that successful performance of their assigned duties may reasonably be viewed as depending on their abstinence from illicit drug use. It is apparent that drug counselors who themselves use illicit drugs, like drug-using interdiction agents, may "because of their own drug use, [be] unsympathetic to their mission." See Von Raab, 109 S. Ct. at 1393. This concern is all the more pressing because of a drug counsellor's full-time, largely unstructured contact with drug users. While the consequences of a drug counsellor's misplaced sympathies may not be as extreme as those attributable to drug use by drug interdiction personnel--risks to "our Nation's first line of defense" against the mass importation of illegal drugs and to the health and safety of the agents--the Army maintains a legitimate interest in ensuring that its employees are allegiant to their essential mission.

As in the case of the "drug prosecutor" alluded to in Harmon, illicit drug use by an employee whose assigned duty is to counsel against the use of drugs is so dissonant with his responsibilities, he should reasonably expect to provide extraordinary assurances of trustworthiness and probity. While "most ... government employees in general" have no reason to expect inquiry into their fitness and probity, id. 109 S.Ct. at 1394, drug counsellors, like drug prosecutors, reasonably should expect heightened scrutiny into activities which evidence such a basic infidelity to their mission.

The Army also possesses a legitimate interest in ensuring effective treatment and rehabilitation of known drug users. ADAPCP personnel service both military and civilian personnel, J.A. at 365, who, upon successful completion of a treatment and rehabilitation program are eligible to return to their former jobs, J.A. at 235-36. The Army's interest in rehabilitation is even more urgent when the client is a critical employee, who, it appears from the regulations, may remain in his critical position after testing positive for proscribed drugs. See AR 600-85 para. 5-14c(4). Thus, ADAPCP personnel represent a vital phase in the Army's campaign to detect and eradicate ongoing drug use. Their own drug use portends the possible frustration of these efforts.

The Army's compelling interest in preventing drug use among the other categories of critical personnel carries a collateral interest in ensuring effective detection. To this extent, serious governmental interests may be furthered by testing those in the laboratory and in the biochemical chain of custody, upon whom the legitimacy of the entire program depends. However, a drug-related lapse by such an employee does not portend either direct or irreparable harm, as would, for example, a lapse by an air traffic controller, pilot, or guard. Absent either a "clear, direct nexus" between the duties of a lab technician or other employee in the chain of custody and the nature of the feared harm, Cf. Harmon, 878 F.2d at 489-90 and absent any compelling reason to expect that drug use will result in misplaced sympathies for their responsibilities, testing these employees lacks the necessary causal connection between the employees' duties and the feared harm.

While it is certainly true that a drug-using lab technician might be susceptible to bribery, so might a drug-using antitrust prosecutor, who we concluded in Harmon, cannot be made subject to testing. We note that the likelihood of complicity is minimized by the absence of the tested individual's name on the specimen label. See HHS Reg. Sec. 2.2(f)(19)-(20), 53 Fed.Reg. at 11,981 (label to include individual's specimen number and initials). Significantly, these personnel work in a more "traditional office environment[ ]" than the other employees we have considered, such that drug use might more easily be detected.9 See Von Raab, 109 S. Ct. at 1395; Harmon, 878 F.2d at 489. Further, there is nothing in the record to indicate that the laboratory personnel possess diminished expectations of privacy.10 Because we have not been shown that the governmental interests in testing either laboratory workers or those in the specimen chain of custody outweigh the privacy expectations of those employees, we hold that the testing of these employees is not reasonable within the meaning of the Fourth Amendment.

The District Court correctly determined that the governmental interests in testing laboratory workers and those involved only in the chain of sample custody do not outweigh the individuals' expectations of privacy, and to that extent only we affirm its decision. We hold that random, mandatory urinalysis testing of those employees who occupy positions in the aviation, police/guard, and direct service staff (ADAPCP) program is reasonable. We acknowledge, of course, that the integrity of bodily freedom is a cherished value in our society, cf. Schmerber v. California, 384 U.S. 757, 772, 86 S. Ct. 1826, 1852, 16 L. Ed. 2d 908 (1966), and that drug testing limits that freedom. But because of the surpassing governmental interests supporting the testing and because of the employees' already diminished expectations of privacy, we believe that, on balance, testing these employees is reasonable. Because we are on the current record unable to assess the reasonableness of testing PRP employees, the case is remanded for further proceedings consistent with this opinion.

So ordered.