OPINION AND ORDER

This case requires the Court to find a balance between conflicting and compelling interests. On one side of the scales is the interest of the New York City Transit Authority, and the public at large, in the safe and efficient operation of New York’s public transport. On the other side are the protected privacy interests of tens of thousands of the Transit Authority’s employees, represented here by the plaintiff unions.

The Americans with Disabilities Act of 1990 (“ADA”) entitles employees (of covered institutions) to maintain a special zone of privacy from their employers concerning their health and medical conditions. Employers may not make inquiries regarding an employee’s disabilities without a sufficient business justification. The Transit Authority has a longstanding practice of making certain general inquiries about its employees’ medical conditions before approving their applications for sick leave. In this respect, the Transit Authority acts no differently from many other employers; and until recently, the conflict between these routine inquiries and the ADA was not widely recognized. The Second Circuit’s recent decision in Conroy v. New York State Department of Correctional Services1 held that even such routine and general inquiries might reveal disabilities, and thereby invade the privacy interests protected by the ADA. Conroy set out a framework to guide the district courts in determining whether such inquiries are, in a particular case, justified by the employer’s business needs.. This case appears to be the first in which a trial court must apply Conroy’s analysis.

The plaintiff unions seek a declaratory judgment that the Transit Authority’s policies and practices relating to employee sick leave (“the Policy”) violate the prohibited .medical inquiries and examinations provision of the ADA, and request an injunction prohibiting the Policy's enforcement. The Transit Authority advances two justifications for the Policy: first, curbing sick leave abuse, and, second, maintaining workplace and public safety. Although I recognize the importance of the public interests behind the -Policy, I am bound to follow Conroy. Applying Con-roy’s analysis, I find that the first justification is sufficient to support the inquiries only with respect to those employees with egregiously poor attendance records, and the second is sufficient only with respect to employees with safety-sensitive jobs. Further factual development will be required to determine which employees are performing such safety-sensitive duties.

The Court has jurisdiction over this matter pursuant to 28 U.S.C. § 1331, as a controversy arising under federal law, and pursuant to 28 U.S.C. § 1343(4), as an action to secure equitable relief under a federal civil rights statute. A bench trial was held from September 7, 2004 to September 14, 2004. The following constitutes the Court’s findings of fact and conclusions of law.

*4381. FINDINGS OF FACT

A. The Parties2

Plaintiffs are three labor unions and their respective presidents: Transport Workers Union of America, Local 100 (“Local 100”), and its president Roger Toussaint; Local 100’s parent union, the Transport Workers Union of America (“T.W.U.”), and its president Sonny Hall; and Amalgamated Transit Union, Local 726 (“Local 726”), and its president Angelo Tanzi (collectively, “the Unions”). Defendants are the New York City Transit Authority (“Transit Authority”) and its subsidiary, the Manhattan and Bronx Surface Transit Operating Authority (“MAB-STOA”) (collectively, “the Authority”). The Authority employs more than 47,000 people. Local 100 represents over 32,000 of the Authority’s employees, including subway station cleaners, bus operators and other employees in Manhattan, the Bronx and Brooklyn. Local 726 represents over 1,300 of the Authority’s employees, including bus operators and maintenance employees in Staten Island. Some of the employees represented by the Unions are employed by the Transit Authority, and others by MABSTOA. The Unions and the Authority are, and at all relevant times have been, parties to collective bargaining agreements (“CBAs”). Local 100 and Local 726 each have separate bargaining agreements with the Authority. js

B. The Sick Leave Policy

The inquiries challenged by the Unions are part of the sick leave policy codified in the CBAs between the Authority and the Unions.3 The Policy involves three inquiries. First, any employee 4 who seeks sick leave must, before missing work, call the Authority to give notice at least one hour prior to the start of his or her scheduled tour of duty. This notice must include a brief statement of the nature of the illness or condition causing the absence.5

Second, on returning to work, the employee must submit a sick leave application form (known colloquially as a “sick form”).6 The employee must give the completed sick form to his or her supervisor. The sick form must be submitted by all employees after an absence of any length, regardless of whether the employee seeks paid or unpaid leave. The sick form requires the employee to state again the “nature of [the] disability” which caused him or her to be “unfit for work on account of illness during this period.” The form must be submitted within three days of the employee’s return from his or her absence.7 During that time, the employee returns to normal duty.8

*439Third, in certain circumstances, employees are also required to have the “doctor’s certification” section of the sick form completed. In such cases, the employee’s doctor must certify that the employee’s illness “so incapacitated the employee that he/she was incapable of performing ( his/her duties” during a specific period of time. The doctor must also state briefly the employee’s “diagnosis/objective findings” and “treatmenf/prognosis and expected date of return.”9

The Authority’s employees are divided into several categories, established through collective bargaining, to determine when a doctor’s certificate is required. Local 726-represented 'employees must submit a doctor’s certificate after an absence of more than two days. Only thirty percent of Local 100-represented employ; ees, those with the heaviest record of sick leave usage, are required to submit a doctor’s certificate after an absence of more than two days. The remaining seventy percent are required to submit a doctor’s certificate after an absence of more than three days.10

Finally, all' émployees who have been placed on a “sick leave control list” are required to submit a doctor’s certificate after an absence of any length.11 Ah employee is placed on the list after taking six absences without a doctor’s certification in one year, or after a suspicious “pattern of usage.” 12 . A suspicious pattern of usage may mean consistent absence on the same day of the week or month, or adjacent to regular holidays or coincident with sports events, or similar suspicious conduct. It is not necessary that the employee have been proved to have abused sick-leave, nor does such proof by itself place an employee on the list.13

The Policy has existed in substantially its present form for decades. It was incorporated into .the CBAs between the Unions and the Authority in 1982, by order of a special independent arbitration panel. It has remained in the CBAs, which have been periodically renegotiated and renewed, since that time.14

C. The Purpose of the Policy’s Disclosure Requirements

The Authority has identified two main purposes served by the medical disclosures required by the Policy. First, the Authority asserts 'that the disclosures are used to detect and deter abuse of sick leave — that is; employees taking sick leave when they are not in fact too sick to work. Second, the disclosures may alert the Authority to dangerous medical conditions in safety-sensitive employees such as bus operators.15 I will first discuss the evidence offered in. support of these concerns — the high cost of absenteeism and the safety risks associated with impaired bus operators. I will then address whether the evidence establishes that the sick forms are reviewed by the Authority’s management in a way that effectively meets these concerns.

*4401. The Extent and Cost of Sick Leave Abuse

Employee absences impose substantial costs on the Authority and on the public. Aaron Stern, the Authority’s Deputy Budget Director for Budget Control, testified that paid sick leave costs the Authority over $74 million annually.16 This represents the total cost to the Authority of all paid sick leave, legitimate or illegitimate. This figure is based on “what would be the additional change in the number of positions that would be required to do the fixed amount of work if there were no paid sick leave ... [T]he answer is 657 positions fewer would be required if there were no paid sick leave. And that would save $74.4 million.”17

Stern’s figure of $74 million does not include the costs of unpaid sick leave. However, this cost is also considerable. Although it is not clear from the record precisely how much unpaid sick leave is taken, it appears that the Authority’s employees take nearly as many days of unpaid sick leave as they do paid sick leave.18 Although the Authority does not have to pay wages to employees when they take unpaid sick leave, such leave is not without its costs. The Authority must pay for substitute workers, which may require overtime payments,19 and must pay fringe benefits to both the replacement and the absent worker.20

Moreover, the Authority presented compelling evidence of substantial indirect costs of sick leave, whether paid or unpaid. For instance, Marie Bernoth, a manager responsible for subway station night cleaning crews, testified that employee absences caused significant disruption to subway cleaning services. Frequently, the Authority is unable to replace absent cleaning employees. This may result in stations going uncleaned, leading to unsanitary conditions.21 Even when the Authority is able to make replacements, employee efficiency and morale is decreased as they are placed in unfamiliar stations, and required to work in unfamiliar teams.22 Bus service may also be disrupted and delayed as a result of absenteeism.23 Sick leave (both paid and unpaid) thus affects important services, imposing costs on the public as well as on the Authority.

The percentage of sick leave that is illegitimate is not clear — and perhaps could never be measured with precision — but the Authority has presented evidence that it is significant. The clearest evidence of abuse, of course, is the fact that every year the Authority denies three to four thousand days of sick leave, on grounds of *441abuse.24 This is, however, a relatively small percentage of the total number of sick leave applications: in 2003, the Authority’s employees took a total of roughly 280,000 paid sick days.25 The number of sick leave applications that are denied amounts to 1-2% of the total of paid sick leave. Plainly, this quantity of claimed sick leave is abusive: given the difficulties of detection, the real amount is likely to be larger.

The Authority has presented anecdotal evidence intended to show that the real amount of sick leave abuse is much more substantial than the amount that is actually denied. The Authority’s witnesses testified that sick leave usage rises on certain days. For example, sick leave usage is said to be higher than usual when the weather is especially bad, or especially good, and on the days of major sporting events. Sick leave also rises around regular holidays, suggesting that employees are using sick leave to extend their vacations.26 I find this testimony, coming from witnesses with long experience both as employees of the Authority and as managers, to be credible. The evidence shows that there is a significant level of sick leave abuse.

However, the Authority has not proved that the abuse is so widespread that overwhelming numbers of employees are abusing the leave policy. The evidence of suspicious patterns of usage is anecdotal and impressionistic: while it shows significant and troubling levels of abuse, it is not clear how common such abuse is. The Authority has also offered a great deal of statistical evidence intended to show that sick leave abuse has reached such epidemic levels that it may be considered the norm among the Authority’s employees. This evidence is not convincing.

For instance, the Authority offered evidence that usage of paid sick leave by MABSTOA employees was greater in 2004 than in 2003, following an increase in the amount of authorized paid sick leave.27 The Authority asks the Court to infer that the explanation for this increase is that employees took advantage of the increase in paid sick leave to take leave when they could have worked. For several reasons, this inference is unwarranted.

First, while it is true that the usage of paid sick leave increased in 2004, usage of unpaid sick leave declined at the same time, so that the increase in total sick leave, while real, is less striking than the increase in paid sick leave taken on its own.28 Second, this increase in sick leave usage could also support an inference that, before the increase in paid sick leave, employees were prone to work when they were in fact too sick to do so, to avoid losing income. Third, the Authority’s own figures indicate that employees who did not receive an increase in paid sick leave also increased their sick leave during the *442same period by a similar amount.29 Fourth, the Authority has not presented any evidence that, the increase from 2002 to 2003 is more than the normal fluctuation between years, which could be caused by any number of factors (e.g., a bad flu year).30

The Authority also points to the fact that sick leave usage is greater in May, the month in which employees’ sick leave “banks” are replenished, than in other months.31 The Authority argues that this demonstrates that employees are taking advantage of the replenished stock of sick leave by taking leave when they are not ill. However, it is equally likely that sick employees aré taking paid sick leave in May, while in the later months of this cycle, they either take unpaid leave or worked despite their .illness, having depleted their banks of paid sick leave.

For the foregoing reasons, I find that there is a significant level of sick leave abuse at the Authority. However, the evidence 'is not sufficient to show that such abuse is so widespread as to be the norm among the Authority’s employees.32

2. Workplace and Public Safety Concerns

At trial, the Authority focused on bus operators as an example of a safety-sensitive class of employees. The danger that may be posed by a bus driver who is unfit for duty due to illness or medication is obvious — a -single accident could injure many people.33 The law recognizes the importance of these concerns.34 ‘The Au*443thority is under a legal obligation not to “permit a driver to operate a bus while the driver’s ability or alertness is so impaired, or so likely to become impaired, through fatigue, illness or any other cause, as to make it unsafe for him to begin or continue to operate the bus.”35 The Authority is also required to prohibit drivers from operating a bus within six hours of having consumed a drug or intoxicating liquor, including certain common medications such as Nyquil.36

Pursuant to a pre-trial agreement, no evidence was presented as to the safety-sensitive nature of other jobs. It is likely that some other classes of employees— such as tower operators, or certain maintenance employees — raise similar safety concerns, while others do not.

3. The Procedures for Review of Sick Forms and Medical Diagnoses

a. Review for Purposes of Detecting Sick Leave Abuse

An employee’s sick form is reviewed, in the first instance, by his or her supervisor.37 The Authority’s Sick Leave Manual instructs supervisors to look for irregularities in the form, such as signs of forgery of the doctor’s certificate, or discrepancies between the dates on which the doctor certifies that the employee was ill and the period of absence actually taken.38 Supervisors are also instructed to determine whether the length of absence appears reasonable in light of the diagnosis and treatment described.39 They may also examine whether the employee’s self-diagnosis matches that of the doctor.40 Suspicious forms are submitted to the Labor Cost Control Unit in the Labor Relations Department for further investigation, which may involve review of the stated diagnosis by the Authority’s medical professionals.41 Only a “small percentage” of sick forms are referred for investigation on suspicion of abuse.42 Nevertheless, the Authority denies three to four thousand days of sick leave a year on grounds of abuse.43 Every year, the Authority initiates thousands of disciplinary actions for sick leave abuse.44 In addition, the Authority regularly pursues doctors who have provided fraudulent certificates, recouping money from them and in some cases causing them to lose their licenses.45

Although thousands of sick forms are referred every year for further review, denial or disciplinary action, only a small number of these are questioned because of irregularities in the medical disclosures challenged by the Unions. In the majority of cases, if a form is questioned, it is because of other irregularities. Nonetheless, a significant quantum of abuse is detected as a result of the medical disclosures: one witness estimated that 8-10% of those forms that are sent for further review are questioned because of the medical disclosures, rather than other irregularities.46

*444While the Authority has adduced evidence that the Policy results in the denial of some sick leave and related disciplinary actions, the statistics cannot show that the Policy deters abuse.47 This absence of evidence is not surprising. The Policy has been in place essentially unchanged for decades, and the Authority therefore has had no opportunity to observe levels of sick leave abuse in the Policy’s absence. Despite this understandable gap in the Authority’s statistical presentation, it strains credulity to believe that the Authority’s investigations and disciplinary actions, including suspensions and dismissals, have no deterrent effect. I therefore find that the Policy has some deterrent effect, albeit one that is unquantified and possibly unquantifiable.

b. Review for the Purpose of Detecting Dangerous Medical Conditions

The sick forms are also reviewed for evidence of medical conditions affecting fitness to work. When reviewing bus operators’ sick forms, supervisors are instructed to contact the Authority’s medical professionals, at a Medical Assessment Center (“M.A.C.”), if a form raises questions about the employee’s fitness.48 Employees whose medical conditions render them unfit to work may be kept off duty.

The procedure for review for fitness is somewhat informal and haphazard. Supervisors are given no formal training in detecting dangerous conditions, nor a list of particular conditions to look for. Caten-zaro described the process as a matter of informal exchange of experience: “[w]ork-ing the depot, employees learn through their supervisors. We teach them ... It is informal-type training. It occurs within the depot.”49 Inexperienced supervisors are instructed to “error [sic] on the side of safety” and refer all questionable medications or conditions to experienced managers or to the M.A.C.50

Nonetheless, it is clear that sick forms are reviewed for safety purposes with some regularity. Catenzaro testified that “forms get brought to us on a fairly regular basis if there is a question.”51 Caten-zaro further testified that around 50-100 forms per year, of the hundreds or thousands he sees at his depot each year, *445“raise a flag” and are sent on for further investigation.52 Dr. Michelle Alexander, the Authority’s assistant vice president of occupational health services, testified that supervisors at the depot call doctors at the M.A.C. with questions about medication “daily.”53 I find both Catenzaro’s and Alexander’s testimony credible.

Alexander also testified as to why the Authority carries out its own review of the diagnoses, rather than rely on a doctor’s certification of an employee as fit to work:

Most private doctors are totally unaware of the standards and the actual requirements that most of our employees do.... I have been working for the Transit Authority for many years and I am constantly amazed at what the private doctor will say the employee can do and return to work doing because they don’t understand what the employee really does.54

This testimony is undoubtedly credible.

D. The Authority’s Other Inquiries

The Authority makes a number of other health-related inquiries, mostly for the purpose of detecting dangerous medical conditions, which the Unions do not challenge. Bus operators who have been out sick for more than twenty-one consecutive days are required to take a medical exam to determine their fitness before returning to work.55 In addition, bus operators are subject to random drug testing56 and are required to report their drug and medication usage to the Authority.57 Bus operators are also required to take a physical exam at least biannually.58 Employees who work in confined spaces or with hazardous chemicals are required to have more frequent physical examinations.59 On a more informal basis, dispatchers are required to give bus operators a “non-medical look-over” before allowing them to go on duty.60 Finally, “employees may be ordered to submit to a medical examination at the direction of the medical director, department head or a member of management at any time if, in his/her opinion, one is necessary.”61 In addition, the Authority investigates sick leave abuse by means other than review of the sick form: for example, the Authority may call the absent employee’s house, or conduct surprise home investigations.62

II. APPLICABLE LAW

The Unions argue that the Policy violates the ADA, which prohibits covered employers from making inquiries of their employees that may tend to reveal a disability.63 The relevant section of the ADA provides that:

*446A covered entity shall not require a medical examination and shall not make inquiries of an employee as to whether such employee is an individual with a disability or as to the nature or severity of the disability, unless such examination or inquiry is shown to be job-related and consistent with business necessity.64

My analysis of this law must follow the framework set out in the Second Circuit’s recent opinion in Conroy, in which the court addressed a policy very similar to the one at issue here. The .employer iñ Conroy required employees returning from sick leave to produce a “general diagnosis” and medical certification.65 The Conroy court held that the general diagnosis requirement was an inquiry within the meaning of the ADA prohibition, because it “may tend to reveal a disability.”66

Such an inquiry is lawful only if it is shown to be “job-related and consistent with business necessity.”67 This is an affirmative defense as to which the employer has the burden of proof. The employer is required to make a two-part showing.

First, the employer must “show that the asserted business necessity is vital to the business.”68 The employer must “show more than that its inquiry is consistent with mere expediency [or that] an inquiry is convenient or beneficial to its business.”69 Conroy recognized that “business necessities may include ensuring that the workplace is safe and cutting down on egregious absenteeism.”70 However, an employer “cannot merely rely on reasons that have been found valid in other cases,” but must show that the necessity is vital on the facts of the particular case.71 What constitutes a business necessity may vary from workplace to workplace.72

Second,

[t]he employer must also show that the examination or inquiry genuinely serves the asserted business necessity and that the request is no broader or more intrusive than necessary. The employer need not show that the examination or inquiry is the only way of achieving a business necessity, but the examination or inquiry must be a reasonably effective method of achieving the employer’s goal.73

Moreover, if the employer’s inquiries are part of a general policy affect*447ing broad classes of employees, then “[i]n defining a class subject to a general policy, the employer must show that it has reasons consistent with business necessity for defining the class in the way that it has.”74 The employer need not make a particularized showing that the inquiry is justified with respect to each individual employee. However, the employer must show that the policy is generally justified with respect to the class affected.75 The court should grant some deference to the employer in determining how to define a class subject to a general policy.76

For example, Conroy discussed a policy applicable to employees who had been identified as probable sick leave abusers. Conroy held that this policy was probably justifiable, if the class affected was genuinely focused on a small group of consistent attendance abusers.77 “Nonetheless, if the policy ultimately affects a class of so-called attendance abusers that is much larger than that small group of employees with truly egregious attendance records ... [the employer] will find it more difficult to prove business necessity.”78

III. CONCLUSIONS OF LAW

A. Are the Policy’s Disclosures Within the Scope of the ADA Prohibition?

As a threshold matter, the Authority argues that the Policy is not an “inquiry” within the meaning of the ADA prohibition. Conroy clearly forecloses this argument. The inquiries here are almost identical to those in Conroy. Like those inquiries, the inquiries made by the Authority “may tend” to reveal disabilities or perceived disabilities.79 They are therefore covered by the ADA prohibition, and the Authority must show that they are job-related and consistent with business necessity.

B. Is the Policy Job-Related and Consistent With Business Necessity?

The Authority claims that the Policy is justified by two business necessities: curbing sick leave abuse, and ensuring safety. With respect to each justification, the Authority must show (1) that the claimed purpose is in fact a business necessity, (2) that the Policy genuinely contributes to it, and (3) that the class affected is reasonably defined. For the following reasons, I find that the asserted business necessity of curbing sick leave abuse is not sufficient to justify the Policy other than with respect to the class of employees on the sick leave control list. The asserted business necessity of ensuring safety is sufficient to justify the Policy with respect to bus operators, and any other group of employees who perform safety-sensitive work.80

1. Sick Leave Abuse

a. Is Curbing Sick Leave Abuse a Business Necessity?

Conroy recognized that curbing absenteeism may be a business necessity, *448in some workplaces, but held that the employer has the burden of showing that this goal is a business necessity on the facts of each particular case. The Authority has met this burden: it is vital to the Authority’s business that it control sick leave abuse. Sick leave imposes enormous costs on the Authority, and on the public at large, draining public funds and disrupting vital services. The Authority has demonstrated that some significant percentage of these costs results from abuse of sick leave. Though the precise quantum of abuse is unclear, the cost to the Authority and the public of any degree of abuse is significant.

The Authority is required by New York law to operate for the “convenience and safety of the public on a basis which will enable the operations thereof, exclusive of capital costs, to be self-sustaining.”81 Sick leave abuse frustrates the Authority’s obligations, under the law and to the public, to provide convenient and safe service on a self-sustaining basis. I therefore find that curbing sick leave abuse is a legitimate business necessity for the Authority.

b. Does the Policy Genuinely Serve the Stated Purpose?

The evidence shows that the medical diagnoses contained within the Policy serve the purpose of curbing sick leave abuse. Most obviously, the medical diagnoses result in the denial of a significant number of applications for sick leave. Every year, the Authority denies thousands of applications for sick leave,82 and issues thousands of disciplinary action notices against employees for abuse of the sick leave policy.83 Many of these denials of sick leave and disciplinary actions arise out of irregularities in the medical diagnoses. If the Authority could not obtain a general diagnosis, its ability to detect abuse would be significantly reduced. In addition, the Authority has used the information contained in the medical diagnoses to detect doctors selling fraudulent sick leave excuses, putting some of them out of business and recouping tens of thousands of dollars. Even if the Court were to assume—quite implausibly—that the Authority’s disciplinary actions and investigations do nothing to deter additional fraudulent claims by employees, or fraudulent diagnoses by doctors, it would still be clear that the medical diagnosis requirement serves the Authority’s legitimate business purposes. The diagnosis requirement reduces the costs associated with sick leave abuse by making it possible for the Authority to deny fraudulent sick leave, saving significant sums of money. Because I also find that the diagnosis requirement has some, albeit unquantifiable, deterrent effect, there can be no question that the diagnosis requirement genuinely serves the purpose of curbing sick leave abuse.

It may be true, as the Unions argue, that the contribution of the diagnosis requirement to this purpose is small, when compared to those aspects of the Policy that they are not challenging.84 However, Conroy does not require that the challenged inquiry be “the only way of achieving a business necessity,” only that it “genuinely serves” the purpose in a “reasonably effective” manner.85 The Authori*449ty has met this test.

c. Is the Class Affected Reasonably Defined?

Different aspects of the Policy affect different classes of employees. The first inquiry, the requirement that employees call in to give notice, and the second inquiry, the requirement that employees give a self-diagnosis on their sick forms, affect all employees taking any amount of sick leave. The third inquiry, the requirement of a doctor’s certificate, affects various classes differently, but ultimately affects all employees taking three or more days of sick leave. Employees on the sick leave control list must submit a doctor’s certificate after an absence of any length. The class of employees affected by at least some aspect of the Policy is thus very broad. Conroy held that an employer seeking to justify a general policy aimed at such a broad group would have a “more difficult” burden than an employer whose policy affected a narrower, more carefully drawn group.86

Conroy requires the Court to grant the Authority some deference in defining the class subject to the Policy.87 But even when all due deference is accorded to the Authority, it has not met its burden of showing a reasonable basis for making inquiries of all employees taking any amount of sick leave, or even of all employees taking three or more days of sick leave. The Authority has not shown that it has any reasonable basis to suspect that such broad swathes of its employees are sick leave abusers. Only a small percentage of sick leave has been clearly shown to be abusive. While that quantity of abuse imposes significant costs, and the Authority is justified in enacting a narrowly drawn policy to combat it, there has been no showing that abuse is so widespread as to justify the virtually universal sweep of the current Policy. Moreover, the Authority is fully capable of identifying those employees who do have truly egregious attendance records, and in fact already maintains a control list of such employees. The asserted business necessity of curbing sick leave abuse therefore cannot justify the Policy as it currently stands.

As to the class of employees on the sick leave control list who take any amount of sick leave, I find that the Policy is justified. The sick leave control list is a narrowly drawn class. Although the list undoubtedly includes some innocent employees, the Authority has reasonable grounds for suspecting that employees on the list are potential attendance abusers.88

2. Workplace and Public Safety Concerns

a. Is Safety a Genuine Business Necessity?

Pursuant to a pre-trial agreement, the Authority made its case that maintaining public safety was a business necessity with reference only to bus operators. The record shows clearly that it is a business *450necessity for the Authority to ensure that bus operators are fit to perform their duties. The danger that may be posed by an unfit bus operator is obvious and undisputed.

b. Does the Policy Genuinely Serve the Stated Purpose?

The Unions argue that the Policy does not in fact serve the purpose of maintaining safety. The Unions point first to the fact that employees are given three days from .their return to work to submit then-sick forms, during which time they .are permitted to resume their work duties. This clearly limits the ability of the Policy to detect dangerous- conditions in a timely manner. This limitation is mitigated to some extent, however, by the requirement that employees call in with a statement of their condition before taking leave, and by the fact that many employees choose to submit their sick forms before the three-day-period elapses.

The Unions also make much of the fact that the diagnoses are not reviewed in the first instance by personnel with any training in detecting dangerous conditions or medications. The informal, rule-of-thumb character of the first stage of review of sick forms is certainly troubling. This concern may be partly alleviated by the fact that supervisors are instructed to err on the side of caution and to refer any medication or condition they do not recognize to medical reviewers or to their experienced superiors.

Finally, the Unions contrast the Policy with other inquiries made by the Authority — for example, the requirement that bus operators out sick for more than twenty-one days take a physical exam, or the requirement that all bus operators take a biannual physical' exam — which the Unions concede to be reasonably effective methods of ensuring that employees are fit to work. The Unions argue that in light of these other, more effective inquiries, the challenged Policy is neither necessary nor productive. The Unions also argue that the Authority could achieve its purposes by requiring an employee’s doctor to confirm his fitness to work, without disclosing his condition.

Despite the limitations of the Policy, and the possibility of alternative approaches to the problem, I find that the Authority has met its burden of showing that the Policy is reasonably effective in maintaining safety. The Policy is clearly not optimally effective, but Conroy requires' no more than that it “genuinely serves” its stated purpose in a “reasonably effective” manner.89 The criticisms that the Unions level at the Policy might also be made of the inquiries they do not challenge, or indeed of any reasonably cost-effective inquiry that could be imposed. The bus operators’ biannual physical, for example, could be conducted too infrequently to effectively detect dangerous conditions. Each of these inquiries, however, contributes in some way to the purpose of detecting dangerous conditions, with each inquiry acting as a backstop to the others.

Nor am I persuaded that the Authority could reliably achieve its purposes by asking an employee’s doctor to certify him fit for work. Doctors may underestimate the requirements of a bus operator’s job. A few, the record shows, are negligent, or more interested in their patient than in the Authority; some are simply for sale. The Authority is under a heavy obligation to keep unfit drivers off the city’s streets: it should not have to trust the judgment or scruples of its employees’ doctors to achieve this purpose. The potential risks in this case are severe indeed. If the Policy makes even a small contribution to *451reducing the risks posed by unfit drivers, it is amply justified.

c. Is the Affected Class Reasonably Defined?

The Authority has presented evidence that it has a reasonable basis for concluding that bus operators who require sick leave pose a genuine safety risk. This is true even of bus operators who take only one day of sick leave; even a brief, minor illness may render a driver unsafe behind the wheel (e.g., a driver who has taken certain ordinary, nonprescription cold remedies).

For the foregoing reasons, the Policy is justified with respect to the class of bus operators. The Authority may continue to enforce the Policy as it stands with respect to that class. Pursuant to the pre-trial agreement, the Authority has presented no evidence on the safety risks associated with employees other than bus operators. As a result, I cannot make a decision as to whether safety considerations might justify the application of the Policy to other groups.90 A second trial will be needed to determine which other groups might also pose safety risks.

C. The Scope of Permissible Inquiries 91

The Authority may therefore apply the Policy as it stands to those employees who meet the criteria of the sick leave control list, and to employees with safety-sensitive roles, including bus operators. With respect to employees not on the control list and not in safety-sensitive positions, the Authority may make the following inquiries, which do not have the potential to reveal disabilities and so do not run afoul of the ADA. The Authority may require an employee to call in advance of an absence, but may not require the employee to describe the nature of his illness. The Authority may require an employee to submit a sick form on his return, in which the employee must state that he was unfit to work due to illness during the period of absence, but may not ask the employee to state the nature of his disability. The Authority may also require an employee to submit a doctor’s certificate for absences of certain lengths, as determined through collective bargaining. The Authority may require that the doctor certify that the employee was incapable, due to illness, of performing his duties during a specific period, and that the employee is now fit to resume his duties, but may not require the *452doctor to describe the nature of the illness or treatment.

D. The Authority’s Unclean Hands/Laches Defenses

The Authority also argues that the Unions’ claims are barred on grounds of unclean hands or laches. The Authority argues that the Unions could have challenged the Policy at any time after the enactment of the ADA in 1992. During that time, the CBAs have been renegotiated numerous times, and the Unions have not negotiated for the removal of the Policy, or gone to impasse over it. The Authority argues that the Unions have therefore had the benefit of bargains negotiated on the understanding that the Policy would remain in place, and are estopped from challenging it now.

I have already rejected the Authority’s argument that the Unions had waived their standing to challenge the Policy by this conduct:

The nature of collective bargaining is that it is a negotiation- — one party relinquishes something (either by lowering its demands or agreeing to a provision that the other party values) in order to receive a concession that it deems more worthwhile. Therefore, in order to eliminate the sick leave policy from the collective bargaining agreement, the Unions would have been required to sacrifice some other benefit for their members. But if the policy is not legal, the Unions should not have to make that sacrifice. This is especially true where, as here, the challenged policy was inserted unilaterally by the Authority into the original collective bargaining agreement. Thus, the Unions would have been required to bargain to remove a pre-existing policy that the Authority valued; the cost of winning that concession would likely have been considerable.92

This reasoning still applies, whether the Authority characterizes this argument as a challenge to standing or a defense of unclean hands. The Unions were perfectly entitled to act as they did: they were not required to make sacrifices, whether through negotiation or impasse, in order to remove an illegal provision from the bargaining contract. Because there has been no showing of any bad faith, fraud or deceit on the part of the Unions, the defense of unclean hands cannot succeed.93

To prevail on its defense of laches, the Authority must show that “(1) the [Unions] knew of the [Authority’s] misconduct; (2) the [Unions] inexcusably delayed in taking action; (3) the defendant was prejudiced by the delay.”94 Typically, prejudice ensues where the passage of time has resulted in “unavailability of witnesses, changed personnel, and the loss of pertinent records.”95 In an appropriate case, however, a defendant may show it has been prejudiced when it “has changed his position in a way that would not have occurred if the plaintiff had not delayed.”96

*453The Authority argues that it has suffered prejudice as a result of the Unions’ long delay in bringing suit, having changed its position in negotiations in reliance on the Unions’ failure to make a legal challenge to the Policy. Although the Authority has arguably suffered some prejudice as a result of the delay, the laches defense is unavailing, for the following reasons. First, the Authority has not met its burden of showing that the Unions’ delay was inexcusable. The record shows that the Unions brought this suit promptly on learning of its possibility, following a decision from the Northern District of New York in Fountain v. New York State of Correctional Services, holding that the ADA prohibition is applicable to policies similar to those at issue here.97 It appears that neither the Authority nor the Unions were aware of the illegality of the Policy before Fountain. The Unions’ members should not be penalized for the understandable failure of their lawyers to anticipate this novel and imaginative application of the ADA.98 Second, because the Unions brought this suit within the statute of limitations provided by the ADA, the Court may not apply laches to their claim.99 “Separation of powers principles dictate that federal courts not apply laches to bar a federal statutory claim that is timely filed under an express federal statute of limitations.”100 Third, the Unions seek only to enjoin the Authority’s future conduct. Laches is generally not applicable where a plaintiff seeks only to enjoin continuing future unlawful conduct.101 Where there is a continuing violation, “even in equity ... laches does not bar a claim for prospective injunctive relief.”102

IV. CONCLUSION

For the foregoing reasons, the Unions are entitled to a declaratory judgment to the following effect: the Policy’s inquiries are within the scope of the ADA Prohibition, and the asserted business necessity of curbing sick leave abuse is adequate to justify the Policy as it stands only with respect to those employees who meet the criteria of the Authority’s sick leave control list. However, the asserted business necessity of maintaining safety is sufficient to justify the Policy with respect to safety-sensitive employees, including bus operators. A further trial will be required to determine whether safety concerns may justify the Policy (as it stands) with respect to other groups of employees, or to *454all employees. Because it is not yet clear which groups may pose safety risks, the Policy may continue to be enforced until such time as that issue is determined.

SO ORDERED.