This is a Constitutional challenge to random urine tests for bus dispatchers and instructors.

Facts.

This case was dismissed with prejudice on a Rule 12(b)(6) motion, so we evaluate the complaint de novo to decide whether it states a claim upon which relief could be granted, if the facts alleged were proved. Idaho Sporting Congress, Inc. v. United States Forest Service, 92 F.3d 922 (9th Cir.1996). The constitutionality of agency drug testing regulations is reviewed de novo. International Bhd. of Teamsters v. Department of Transp., 932 F.2d 1292, 1298 (9th Cir.1991).

According to the complaint, Mr. and Mrs. Gonzalez both work for the municipal bus service in Los Angeles. Mrs. Gonzalez is a radio dispatcher, Mr. Gonzalez an instructor. Their employer, the Los Angeles County Metropolitan Transportation Authority, conducts random unannounced testing for drugs and alcohol pursuant to the federal Omnibus Transportation Employee Testing Act of 1991 and regulations thereunder. 49 U.S.C. § 5331; 49 C.F.R. Part 653-54.

Mrs. Gonzalez was subjected to a urine test at a medical facility pursuant to the policy. The test came out negative, but the process was so upsetting to her, according to the complaint, that she needed medical treatment and was unable to return to work afterward. She was a victim of child abuse, which made her unusually sensitive to interference with her bodily privacy. Mr. Gonzalez has not been tested, but is upset both by what happened to his wife and by the prospect of being tested himself.

According to the complaint, the tests are based on the classification of both their jobs as "safety-sensitive," but that classification cannot be justified. Also, the tests are not accurate enough to be justifiable to serve any public purpose, nor are there any screening procedures to protect unusually sensitive people like Mrs. Gonzalez.

The complaint claims that the tests amount to an unconstitutional search, violative of the Fourth Amendment, and also Article I, sections 1 and 13, of the Constitution of the State of California. Suit is brought under 42 U.S.C. § 1983 against the Metropolitan Transportation Authority, its past and present general managers, and its past and present officials in charge of implementing the testing program. The prayer is for a declaratory judgment that the testing program is unconstitutional facially and as applied, damages for lost earnings and emotional distress, and an injunction to prohibit testing Mr. and Mrs. Gonzalez pursuant to the program.

The district court ruled that (1) federal regulations requiring urine testing of employees implicate the Fourth Amendment under Skinner v. Railway Labor Executives' Ass'n., 489 U.S. 602, 109 S. Ct. 1402, 103 L. Ed. 2d 639 (1989), because they invade reasonable expectations of privacy, as plaintiffs argued; (2) reasonableness under the Fourth Amendment is evaluated by balancing the intrusiveness of the test against the government interest satisfied by testing, under National Treasury Employees Union v. Von Raab, 489 U.S. 656, 109 S. Ct. 1384, 103 L. Ed. 2d 685 (1989); (3) but the federal regulations requiring random testing of employees such as plaintiffs satisfy the balancing test. Defendants' motion to dismiss was accordingly granted with prejudice.

Analysis.

I. The regulations.

The Transit Authority argues that the urine tests can properly be required of Mr. and Mrs. Gonzalez under duly authorized federal regulations. The Gonzalezes argue that their jobs do not properly fit under the "safety sensitive personnel" definition under the regulation, or if they do, the definition is broader than the statute permits.

The statute requires the Secretary of Transportation to issue regulations for random testing and to make the decision which categories of employees are to be treated as "safety-sensitive" and tested. Congress required the Secretary to issue regulations "to conduct ... random testing of mass transportation employees responsible for safety-sensitive functions (as decided by the Secretary) for the use of a controlled substance ... and to conduct ... random ... testing of such employees for the use of alcohol...." 49 U.S.C. § 5331(b)(1)(A) (emphasis added).2

The regulations define "safety-sensitive function" to include anyone who operates a "revenue service vehicle" such as a city bus, "including when not in revenue service." 49 C.F.R. § 653.7; 49 C.F.R. § 654.7. That may cover Mr. Gonzalez, described by the complaint as a "Transit Operations Supervisor-Instructor." We understand from the briefs and oral argument that he supervises trainees who drive buses without passengers on board. The definition also includes "[c]ontrolling dispatch or movement of a revenue service vehicle." Id. That covers Mrs. Gonzalez, a bus dispatcher. As an instructor, Mr. Gonzalez is "controlling ... movement" of a bus when he is monitoring a trainee who is driving an empty bus.3 Thus the federal regulations required the municipal bus system to test Mr. and Mrs. Gonzalez, and the Department of Transportation was authorized by the statute to issue the regulations. The only serious question is whether the regulations are constitutionally permissible as applied to Mr. and Mrs. Gonzalez.

II. Jurisdiction.

The Transit Authority argues that the court lacked jurisdiction over the Gonzalezes' challenge because they did not join the Federal Transit Administration and the Department of Transportation, and these entities are indispensable parties under Federal Rule of Civil Procedure 19. This contention is incorrect. Whether a party is necessary and indispensable is a pragmatic and equitable judgment, not a jurisdictional one. Simpson v. Alaska State Comm'n for Human Rights, 608 F.2d 1171, 1174-75, 1175 n. 5 (9th Cir.1979); 7 Charles A. Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice & Procedure §§ 1601, 1603 (1986). If the complaint otherwise states a claim, there is no apparent reason why these parties could not be joined, and it is for the district court initially to decide whether joinder is appropriate and feasible.

III. Qualified immunity.

Appellees argue that all the individual defendants should be dismissed because they are entitled to qualified immunity, and would have had no reason to think that administering the urine tests was unconstitutional. The district court did not reach the issue of what claims for relief if any would be subject to qualified immunity, because it dismissed the case for failure to state a claim. Accordingly, we need not decide whether, for example, the damages claims would fail because of qualified immunity. Of course good faith that would shield individuals from damages judgments would not avoid prospective relief based on unconstitutionality of the regulation as applied, were that to be the correct result.

IV. Constitutionality.

The Gonzalezes argue that the regulations, insofar as their positions are included as "safety-sensitive," reach further than the Fourth Amendment permits. They claim, basically, that what they do affects safety so little, and the tests are of so little utility to protecting public safety, that the interference with their privacy cannot be justified. Thus for municipal bus service to subject them to urine tests is, they argue, an unreasonable search and seizure.

Collection and testing of urine pursuant to government directive is a search under the Fourth Amendment. Skinner v. Railway Labor Executives Ass'n, 489 U.S. 602, 617, 109 S. Ct. 1402, 103 L. Ed. 2d 639 (1989). The reasons are that urine tests can reveal private medical facts, and that "the process of collecting the sample to be tested, which may in some cases involve the visual or aural monitoring of the act of urination, itself implicates privacy interests." Id.; see also National Treasury Employees Union v. Von Raab, 489 U.S. 656, 665, 109 S. Ct. 1384, 103 L. Ed. 2d 685 (1989). For a warrantless search of this sort to be reasonable, usually some quantum of individualized suspicion has been required, though that is not necessary to reasonableness. Skinner, 489 U.S. at 624, 109 S. Ct. 1402. In the absence of individualized suspicion, the reasonableness of such a search depends on balancing the "special needs" of the government against the extent of the intrusiveness of the testing procedure. Chandler v. Miller, 520 U.S. 305, 318, 117 S. Ct. 1295, 137 L. Ed. 2d 513 (1997).

The urine testing cases decided by the Supreme Court over the last decade have gone both ways depending on their quite particularized facts. Skinner, 489 U.S. 602, 109 S. Ct. 1402, 103 L. Ed. 2d 639 (1989), the first of this line of authority, dealt with blood and urine tests of railway employees who had been involved in train crashes or safety rule violations. The Court discussed a century of alcohol problems among railway employees, and a study indicating that "23% of the operating personnel were 'problem drinkers.' " Id. at 607 n. 1, 109 S. Ct. 1402. The urine tests were held to be constitutional, because the urine collection procedure was appropriate, employees' reasonable expectations of privacy were diminished by their participation in an industry pervasively regulated for safety, and the persons tested "discharge duties fraught with such risks of injury to others that even a momentary lapse of attention can have disastrous consequences" and can "cause great human loss before any signs of impairment become noticeable to supervisors or others." Id. at 628, 109 S. Ct. 1402.

Treasury Employees, which came down the same day, held that urine testing of some Customs Service employees was constitutionally permissible, but the testing of others might not be, depending on further development of the facts. Those who sought promotion to positions involving drug interdiction or carrying a firearm could be tested. But whether the Customs Service could also test employees required to "handle classified material" was remanded, because inclusion of such positions as "animal caretaker" and "accounting technician" raised "the question whether the Service had defined this category of employees more broadly than is necessary." Id. at 678, 109 S. Ct. 1384.

In Vernonia School District v. Acton, 515 U.S. 646, 115 S. Ct. 2386, 132 L. Ed. 2d 564 (1995), drug testing of student athletes was held to be constitutionally permissible. The Court reasoned that "children ... committed to the temporary custody of the State as schoolmaster" lacked some of the most fundamental rights of self-determination and privacy, and that athletes who ordinarily showered together had even less of an expectation of privacy. Further, the urine was collected in circumstances intruding no more on privacy than urinating in a public restroom ordinarily did, and information obtained from the urine and disclosure of that information were sharply limited.

Chandler v. Miller, 520 U.S. 305, 117 S. Ct. 1295, 137 L. Ed. 2d 513 (1997) held that it was unconstitutional to require candidates for state public office to subject themselves to urine tests. Though the tests and results were conducted with full privacy and the candidates could control dissemination of the results, there was no evidence that the state was afflicted with drug-abusing officeholders, and the testing procedure was not likely to be effective for screening out drug abusing candidates. The testing protected a state interest in symbolism and image rather than safety, an insufficient basis for the warrantless searches without individualized suspicion.

We have also decided a few urine testing cases. Bluestein v. Skinner, 908 F.2d 451 (9th Cir.1990), upholds random drug testing of flight crew members. International Bhd. of Teamsters v. Department of Transportation, 932 F.2d 1292 (9th Cir.1991), upholds the constitutionality of drug testing for bus and commercial truck drivers. In Railway Labor Executives v. Skinner, 934 F.2d 1096 (9th Cir.1991), we upheld random testing of railroad workers, even without a crash or safety violation.

We are unable, on the record before us, to determine whether our precedents upholding testing are controlling or distinguishable. This case went to judgment on a pleading which did not exclude the possibility that facts might be proved under it, pursuant to which the urine testing would be unconstitutional. Each of our own decisions may be distinguishable, depending on the facts as they are further developed in the case at bar. In Bluestein, testing was limited to employees whose impairment would cause physical risk to passengers, and the test procedure was designed to assure reasonable balancing of accuracy, privacy, and dignity. But here we do not yet know whether dispatchers and instructors would cause physical risk to passengers if impaired, because we do not know exactly what they do, and we know little about the testing procedure. In Teamsters, we did not address the issue of testing dispatchers and instructors, we were able to ascertain how the tests were done and how much they impinged on privacy, and we noted that truck drivers were already subjected to extremely invasive physical examinations including urinalysis even without the challenged procedure. What was critical in Teamsters was that the persons tested could be impaired "behind the wheel." Teamsters, 932 F.2d at 1304. The employees at issue in the case at bar apparently do not get behind the wheel. In Railway Executives, unlike the case at bar, we had a record establishing how the tests were done, so that we could evaluate accuracy, privacy and dignity.

Vernonia School District v. Acton, 515 U.S. 646, 653, 115 S. Ct. 2386, 132 L. Ed. 2d 564 (1995), holds that reasonableness is judged by balancing the search's intrusion on the individual's Fourth Amendment interests against its promotion of legitimate governmental interests. Vernonia holds that the factors to be considered are: (1) the nature of the privacy interest upon which the search intrudes; (2) the character of the intrusion; (3) the immediacy of the government concern and the efficacy of the search for meeting it. Id. at 654-64, 115 S. Ct. 2386. We lack the factual record to perform the Vernonia evaluation. As to "nature of the privacy interest," the municipal bus authority employees obviously are not children with a reduced privacy interest, as in Vernonia, and the record does not establish whether bus dispatchers and other transit employees are subject even without the urine tests to the kind of comprehensive physical examinations that were a factor in Teamsters, 932 F.2d at 1300. The record does not establish whether, as in Skinner, bus dispatchers and instructors have traditionally been subject to extensive regulation for safety.

Vernonia holds that we are next to consider the "character of the intrusion," because urine testing "intrudes upon 'an excretory function traditionally shielded by great privacy.' " Vernonia, 515 U.S. at 658, 115 S. Ct. 2386. We do not know whether, as in Vernonia, the bus employees are tested in "conditions nearly identical to those typically encountered in public restrooms," Id. at 658, 115 S. Ct. 2386 or what other procedures may be used.

The next Vernonia factor is "the nature and immediacy of the governmental concern and the efficacy of this means for meeting it." Vernonia, 515 U.S. at 660, 115 S. Ct. 2386. Obviously the government concern with preventing accidents involving municipal buses is very great. In a Third Circuit case on bus drivers, the transit authority "presented extensive evidence of a severe drug abuse problem among its operating employees," Transport Workers Union v. Southeastern Pennsylvania Transp. Auth., 884 F.2d 709, 711 (3d Cir.1989), but no evidence has been presented in the case at bar. But with a record limited to the complaint in this case, it would be possible for the evidence to establish efficacy or inefficacy.

One issue of efficacy is that the complaint does not establish how testing employees in the positions at issue, instructor and dispatcher, affects safety. It does not show exactly what they do, and how being under the influence of alcohol or drugs when they do it might affect safety. So far as we can tell, neither appellant drives a bus. It may be that the risk to public safety is great, because even though they are not behind the wheel, these employees can greatly increase the risk of an accident by those who are, as for example a drunk mechanic might. But it is also possible that plaintiffs can prove that dispatchers and instructors are no more of a threat to public safety than the animal handlers in Treasury Employees or the holders of high public office in Chandler. We understand from the briefs and argument that a dispatcher sits in a booth and talks on a microphone, telling drivers where more buses are needed and gives directions so a bus driver can get there. Possibly their duties make them as dangerous if drunk on duty as drunk air traffic controllers would be. On the other hand, buses go a lot slower than airplanes, so the air traffic controller comparison might not be apt. It is also possible that a drunk dispatcher is no more dangerous to public safety than a gas station attendant who gives an inquiring motorist mistaken directions, or a drunk shipping clerk who sends out ten of something when only one was ordered. It is conceivable that the one time of the day that a drunk dispatcher does not threaten public safety is when on the job, because he or she is then assuredly off the road. The record does not compel either inference at this stage.

The other aspect of efficacy is whether the tests are reasonably designed to accomplish their objective, and that too is subject, on the limited record before us, to proof either way. It does not show whether the tests are adequately designed to catch those who should fail them, by such means as surprise and monitoring to avoid deception. The inefficacy of the drug tests on candidates, performed by appointment with no monitoring, was among the reasons why they were unconstitutional in Chandler, 520 U.S. at 319-20, 117 S. Ct. 1295. Without some sort of monitoring and surprise, workers who do drink and use drugs regularly can substitute clean urine for their own, rendering the indignity on all the clean workers pointless.

A more complete record can also illuminate another aspect of efficacy, the Bayes' theorem problem that affects any random test given to a low incidence population. Nothing in this world is perfect. Suppose the combination of errors in the tests, including containers marked with someone else's name or number than the person who urinated into them, typographical errors in the reports of test results and identifications of which employees produced which results, anomalous chemical reactions with other substances in people's bodies such as medications and foods, and other random errors, cause an error rate such that one person out of 500 gets a report of "dirty" urine when it was actually "clean." Suppose that there is a high rate of alcohol drug use among the employees (in Skinner the Court found it relevant to note that the employees had a 23% "problem drinkers" rate), and on any particular day one worker in 10 has alcohol or drugs in his blood. Then with a 1/500 false positive rate, out of 1,000 tests, 2 will be positive even though the employee's urine was clean, and 100 will be positive correctly. Only one of the positives out of every 51 is false. Fifty out of 51 are accurate. That is a fairly effective test, in terms of reliability.

But if the workers are generally "clean," the reliability of the test goes way down. Suppose on a particular day only one worker in 500 has ingested drugs or alcohol. Then with a 1/500 false positive rate, out of 1,000 tests, 2 will be correct positives and 2 will be false positives. Half the employees who get a "dirty" urinalysis report are unjustly categorized. A positive result is as likely to be false as true on so clean a population, even though the test is identical to the one that was quite effective for a population with a higher incidence of drug and alcohol usage.

Thus the data on the extent of the drug and alcohol problem is useful not only for evaluating whether "the nature and immediacy of the governmental concern," Vernonia, 515 U.S. at 660, 115 S. Ct. 2386 justifies the testing, but also for evaluating whether the tests achieve reasonable efficacy in distinguishing those who have used alcohol or drugs from those who have not. This is not to say that a low rate of alcoholism or drug use necessarily bars testing. The low rate may result from the success of the test in deterring alcohol and drug abuse, and the false positives may be filtered out more or less adequately by a second test. See Treasury Employees, 489 U.S. at 656 n. 3, 109 S. Ct. 1384. But we need a record to determine whether the burden of false positives on those so classified, and the indignity of the testing on all those subjected to it, so outweighs the value of the testing as to make the testing an unreasonable search in the absence of any individualized suspicion.

Congress, in providing for mass transportation funding, unquestionably can constitutionally limit use of its funds in such a way as to keep drunk bus drivers off the street. But the use of random urine testing involves a search which must be "reasonable" in order to satisfy the Fourth Amendment. It is not enough to say that the test is performed in a doctor's office. We also take off all our clothes, wear gowns open in the back, and allow strangers to touch and probe our bodies in a doctor's office; that does not mean that the government can compel us to do these things without individualized suspicion.

We cannot simply defer to the administrative agency's determination that dispatchers perform a "safety-sensitive function." If that were the proper way to resolve the constitutional question, then the Supreme Court would have deferred Treasury Employees to the Treasury Department determination that employees who handled classified materials performed safety-sensitive functions. Instead, the Court remanded that case "to assess the reasonableness of the Government's testing program insofar as it covers employees who are required 'to handle classified material' " because the record "raises in our minds the question whether the Service has defined this category of employees more broadly than is necessary." Treasury Employees, 489 U.S. at 677-78, 109 S. Ct. 1384.

Conclusion.

The tests challenged by Mr. and Mrs. Gonzalez and the regulations pursuant to which they were performed address a serious concern. But the record is inadequate to determine whether, as applied in this case, they address it in a constitutional manner. We do not know, from the record we have, whether the employees at issue would pose a substantial immediate threat to public safety if impaired by drugs or alcohol, or whether the procedure for testing them would be reasonably effective for finding out if they are impaired, or whether the tests as performed were an undue invasion of their privacy. Facts might be proved under the complaint which would entitle plaintiffs to relief. Accordingly we REVERSE and REMAND.