This direct review proceeding presents a classic case of what Judge Friendly has aptly termed 'a new form of uneasy partnership' between agency and court that results whenever Congress delegates decision making of a legislative character to the one, subject to review by the other. Associated Industries v. United States Dept. of Labor, 487 F.2d 342, 354 (2nd Cir. 1973). The angularity of this relationship is only sharpened when, as here, Congress-- with no apparent awareness of anomaly-- has explicitly combined an informal agency procedure with a standard of review traditionally conceived of as suited to formal adjudication or rulemaking. The federal courts, hard pressed as they are by the flood of new tasks imposed upon them by Congress, surely have some claim to be spared additional burdens deriving from the illogic of legislative compromise. At the least, it would have been helpful if there had been some recognition by Congress that the quick answer it gave to a legislative stalemate posed serious problems for a reviewing court, and that there would inevitably have to be some latitude accorded it to surmount those problems consistently with the legislative purposes. The duty remains, in any event, to decide the case before us in accordance with our statutory mandate, however dimly the rationale, if any, underlying it can be perceived.

The petition before us seeks review of standards promulgated by the Secretary of Labor under the Occupational Safety and Health Act of 1970, 29 U.S.C. 651 et seq., (hereinafter OSHA). The standards in question regulate the atmospheric concentrations of asbestos dust in industrial workplaces. Petitioners are unions whose members are affected by the health hazards of asbestos dust. They challenge the timetable established by the standards for the achievement of permissible levels of concentration, and object to portions of the standards concerning methods of compliance, monitoring intervals and techniques, cautionary labels and notices, and medical examinations and records. We remand two of such issues to the Secretary for further consideration. In all other respects, the petition is denied.

Technological progress in industry appears not to have been accompanied uniformly by corresponding reductions in the health hazards of industrial working conditions. More than 2.2 million persons are disabled on the job each year, and in 1967 the Surgeon General estimated that approximately 400,000 new cases of occupational disease would occur in each succeeding year.1 The Chairman of the Committee on Labor and Public Welfare summarized the problem as follows:

Not only are occupational diseases which first came to light at the beginning of the Industrial Revolution still undermining the health of workers, but new substances, new processes, and new sources of energy are presenting health problems of ever-in-creasing complexity.

Foreword, Legislative History of the Occupational Safety and Health Act of 1970 (hereinafter Legis.Hist.).

OSHA, the first comprehensive attempt by Congress to deal with these problems,2 covers every employer whose business affects interstate commerce.3 Eschewing any attempt to establish substantive provisions to control all these various employers, the Act erects a general framework to govern the development of regulations, and delegates the task of formulating particular health and safety standards to the Secretary of Labor. Civil and criminal sanctions are provided to enforce compliance.

OSHA specifies the procedure to be followed in the promulgation of standards, and provides for the establishment of a research institute and the appointment of advisory committees to assist the Secretary.4 The substantive provisions of the Act impose a general obligation upon employers to provide safe working conditions. 29 U.S.C. 654(a)(1) (1970). The Secretary is required to promulgate standards to control particular health hazards that come to his attention. Certain types of controls, including monitoring, medical examinations, warnings, record keeping, and specific protective measures are specified by the statute itself, but the decision as to when and how they should be required with regard to particular health hazards is left to the Secretary.

Asbestos is a generic term applicable to a number of fibrous, inorganic, silicate minerals that are incombustible in air. Its commercial value is high, and its uses are many and varied. Asbestos can be woven into cloth, used in powder form, or incorporated into materials of various shapes and consistencies. Almost one million tons of asbestos are used in this country annually; and, for many purposes, it cannot easily be replaced with other substances.5

Unfortunately, asbestos is as hazardous to health as it is useful to industry. During its production and use, tiny asbestos fibers are released as a dust in the air, and, over the course of this century, thousands of workers have been killed or disabled by the effects of inhaling these fibers. There are no precise figures concerning the number of workers involved, but it is estimated that three to five million workers are exposed to some extent to asbestos fibers in the building construction and shipyard industries alone.6 While OSHA was under consideration in Congress, the health hazards of the asbestos industry were among the examples used to stress the need for legislation.7

Within a few months of the effective date of OSHA, petitioners requested the Secretary to establish an emergency standard to control concentrations of asbestos dust.8 The Secretary promptly issued a temporary standard and set in motion the procedure for establishment of a permanent standard. Notice of the proposed rulemaking was published, and interested persons were invited to submit their views. NIOSH submitted its recommendations, as did the Advisory Committee. These were made public, and the Secretary conducted a hearing at which various representatives and experts appeared on behalf of interested parties. On the basis of these recommendations and a formidable record of documents and oral testimony, including highly technical statements by expert witnesses, the Secretary established the standards in question.9 His statement of reasons covers some four and one-half pages of the Federal Register.10

Petitioners allege no procedural errors in the promulgation of these standards, but they characterize them as inadequate to protect the health of employees as required by the Act. They attack the Secretary's interpretation of OSHA in certain particulars, as well as the enforcement measures he has selected.

OSHA is a self-contained statute in the sense that it does not depend upon reference to the Administrative Procedure Act for specification of the procedures to be followed. It prescribes that the process of promulgating a standard is to be initiated by the publication of a proposed rule. Interested persons are given a period of 30 days thereafter within which to submit written data or comments. Within this period any interested person may submit written objections, and may request a public hearing thereon. In such event, the Secretary shall publish a notice specifying the particular standard involved and stating the time and place of the hearing. Within 60 days after the completion of such hearing, the Secretary shall make his decision. Judicial review by the courts of appeals is provided.11

This procedure is characteristic of the informal rulemaking contemplated by Section 4 of the APA, 5 U.S.C. 553, and it was so understood by the Congress. inferences of policy drawn from those although describing it as 'legislative in type,' has provided that the oral hearing called for in the statute shall contain some elements normally associated with the adjudicatory or formal rule-making model. As indicated in the text of the regulations, set forth in the margin,12 the Secretary apparently concluded that this was necessary because of the necessity of having a record to which the statutorily mandated substantial evidence test could be meaningfully applied by a reviewing court. The only controversy we have in this case as to the procedural requirements of the statute is not with respect to the manner in which the rulemaking was done by the Secretary, but as to the reach of the substantial evidence test in the course of judicial review.

The substantial evidence test has customarily been directed to adjudicatory proceedings or formal rulemaking.13 The hybrid nature of OSHA in this respect can be explained historically, if not logically, as a legislative compromise. The Conference Report reflects that the Senate bill called for informal rulemaking, but the House version specified formal rulemaking and substantial evidence review. The House receded on the procedure for promulgating standards, but the substantial evidence standard of review was adopted.14

One question generated by this anomalous combination is whether the determinations in question here are of the kind to which substantial evidence review can appropriately be applied. The Government in its argument suggested that a proper accommodation could be effected by construing the statute to require substantial evidence review of factual determinations, while weighing the inferences of -olicy drawn from those facts in terms of their freedom from arbitrariness or irrationality. We do not believe this approach would affect the rigorousness of our review to the extent the Government seems to suppose, or that petitioners purport to fear.15 The analysis may, however, be useful for the purpose of clarifying the diverse nature of the judicial task imposed upon us by a statute like OSHA.16

Another problem arising from substantial evidence review of informal proceedings concerns the adequacy of the record to permit meaningful performance of the required review. Although this issue has not been directly raised in argument, it underlies much of the controversy concerning the sufficiency of the evidence to support various specific determinations of the Secretary. Thus some explication of the procedural implications of the prescribed substantial evidence standard of review should help to clarify our resolution of the particular substantive issues presented by this petition.

Faced with the fact that his determinations were commanded by Congress to be reviewed under a substantial evidence standard, the Secretary did voluntarily move his procedures significantly towards the formal model. He directed that (1) a qualified hearing examiner should preside over the oral hearing, (2) cross-examination should be permitted, and (3) a verbatim transcript made. The total record in this case was in part created under the conditions that obtain in a formal proceeding. In substantial remaining part, however, it consists of a melange of written statements, letters, reports, and similar materials received outside the bounds of the oral hearing and untested by anything approaching the adversary process.

Thus, in some degree the record approaches the form of one customarily conceived of as appropriate for substantial evidence review. In other respects, it does not. On a record of this mixed nature, when the facts underlying the Secretary's determinations are susceptible of being found in the usual sense, that must be done, and the reviewing court will weigh them by the substantial evidence standard. But, in a statute like OSHA where the decision making vested in the Secretary is legislative in character, there are areas where explicit factual findings are not possible, and the act of decision is essentially a prediction based upon pure legislative judgment, as when a Congressman decides to vote for or against a particular bill.

OSHA sets forth general policy objectives and establishes the basic procedural framework for the promulgation of standards, but the formulation of specific substantive provisions is left largely to the Secretary.17 The Secretary's task thus contains 'elements of both a legislative policy determination and an adjudicative resolution of disputed facts.' Mobil Oil Corp. v. FPC, 157 U.S.App.D.C. 235, 254, 483 F.2d 1238, 1257 (1973). Although in practice these elements may so intertwine as to be virtually inseparable, they are conceptually distinct and can only be regarded as such by a reviewing court.

From extensive and often conflicting evidence, the Secretary in this case made numerous factual determinations. With respect to some of those questions, the evidence was such that the task consisted primarily of evaluating the data and drawing conclusions from it. The court can review that data in the record and determine whether it reflects substantial support for the Secretary's findings. But some of the questions involved in the promulgation of these standards are on the frontiers of scientific knowledge, and consequently as to them insufficient data is presently available to make a fully informed factual determination. Decision making must in that circumstance depend to a greater extent upon policy judgments and less upon purely factual analysis.18Thus, in addition to currently unresolved factual issues, the formulation of standards involves choices that by their nature require basic policy determinations rather than resolution of factual controversies. Judicial review of inherently legislative decisions of this sort is obviously an undertaking of different dimensions.19

For example, in this case the evidence indicated that reliable data is not currently available with respect to the precisely predictable health effects of various levels of exposure to asbestos dust; nevertheless, the Secretary was obligated to establish some specific level as the maximum permissible exposure. After considering all the conflicting evidence, the Secretary explained his decision to adopt, over strong employer objection, a relatively low limit in terms of the severe health consequences which could result from over-exposure. Inasmuch as the protection of the health of employees is the overriding concern of OSHA, this choice is doubtless sound, but it rests in the final analysis on an essentially legislative policy judgment, rather than a factual determination, concerning the relative risks of underprotection as compared to overprotection.

Regardless of the manner in which the task of judicial review is articulated, policy choices of this sort are not susceptible to the same type of verification or refutation by reference to the record as are some factual questions. Consequently, the court's approach must necessarily be different no matter how the standards of review are labeled. That does not mean that such decisions escape exacting scrutiny, for, as this court has stated in a similar context:

'This exercise need be no less searching and strict in its weighing of whether the agency has performed in accordance with the Congressional purposes, but, because it is addressed to different materials, it inevitably varies from the adjudicatory model. The paramount objective is to see whether the agency, given an essentially legislative task to perform, has carried it out in a manner calculated to negate the dangers of arbitrariness and irrationality in the formulation of rules for general application in the future.

Automotive Parts & Accessories Ass'n v. Boyd, 132 U.S.App.D.C. 200, 407 F.2d 330, 338 (1968).

We do not understand Congress to have in this instance nullified this approach for all purposes by directing substantial evidence review. As noted above, that provision is important as an indication of how we should approach certain kinds of questions and what kind of record we should demand of the Secretary. But it is surely not to be taken as a direction by Congress that we treat the Secretary's decision making under OSHA as something different from what it is, namely, the exercise of delegated power to make within certain limits decisions that Congress normally makes itself, and by processes, as the courts have long recognized and accepted, peculiar to itself. A due respect for the boundaries between the legislative and the judicial function dictates that we approach our reviewing task with a flexibility informed and shaped by sensitivity to the diverse origins of the determinations that enter into a legislative judgment.

What we are entitled to at all events is a careful identification by the Secretary, when his proposed standards are challenged, of the reasons why he chooses to follow one course rather than another. Where that choice purports to be based on the existence of certain determinable facts, the Secretary must, in form as well as substance, find those facts from evidence in the record. By the same token, when the Secretary is obliged to make policy judgments where no factual certainties exist or where facts alone do not provide the answer, he should so state and go on to identify the considerations he found persuasive.

Judge Friendly concluded his ruminations in Associated Industries with an expression of doubt as to 'whether judicial review of legislative standards resulting from informal rule-making will ultimately prove to be feasible.'20 That is certainly a serious and substantial question. Whether it can eventually be answered affirmatively must depend in large measure upon the care and good sense with which both the delegatee of what is essentially legislative power and the reviewing court go about their respective duties. In the case of OSHA, the Secretary has wisely acted by regulation to go beyond the minimum requirements of the statute and to expand his capacity to find facts by providing an evidentiary hearing in which cross-examination is available. We think it equally the part of wisdom and restraint on our part to show a comparable flexibility, and to be always mindful that at least some legislative judgments cannot be anchored securely and solely in demonstrable fact. Such a principle, far from being destructive of the Congressional purpose to provide judicial review, seems to us within the Congressional contemplation as essential to its preservation.

Before addressing the specific challenges made to the Secretary's action, we examine two further problems raised by petitioners involving statutory construction. One has to do with the weight to be accorded by the Secretary to the NIOSH recommendations. The other relates to whether the Secretary may take economic considerations into account.

With respect to the former, the statute directs NIOSH to develop criteria documents that describe safe levels of exposure, and the Secretary is to promulgate standards that insure that employees are protected. The language employed by Congress in these two mandates is essentially identical except that the Secretary must consider elements of feasibility.21 From this similarity petitioners argue that the determinations of NIOSH are meant to be conclusive on the question of what exposure levels adequately protect health, and that the Secretary may deviate from the NIOSH document only to the extent dictated by feasibility.

The Act merely says that the Director of NIOSH shall immediately forward recommended standards to the Secretary without specifying how the Secretary is to use them, but the procedure prescribed for the formulation of standards militates against petitioners' position. It is the Secretary rather than NIOSH who conducts hearings and receives the comments of interested persons. The Secretary may also appoint a special advisory committee to assist him in his standard-setting functions, and receive recommendations from it, as he did here.

The Act, or so it seems to us, must be taken as contemplating that the Secretary may consider all of this information as well as that received from restrict the advisory committees and interested restirct the advisory committees and interested parties to comments relating solely to feasibility, a role petitioners themselves clearly-- and, we think, legitimately-- exceeded at the hearing and in their arguments before this court. The NIOSH recommendation was undoubtedly important in the eyes of Congress as an aid to the Secretary, but we cannot see that it was intended as more than that.

In connection with the second issue, we note that the statutory authority for the promulgation of standards reads in relevant part:

The Secretary . . . shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity . . ..

29 U.S.C. 655(b)(5).

The standards as promulgated retain the concentration level specified by the temporary emergency standard until 1976 when a lower permanent standard becomes effective. The Secretary explained his decision to delay, two years longer than the period suggested by NIOSH, implementation of the tougher standard as 'necessary to allow employers to make the needed changes for coming into compliance,' and petitioners argue that the Secretary improperly considered economic factors in reaching this conclusion. We conclude that the factors entering into the Secretary's conclusion could properly include problems of economic feasibility.

There can be no question that OSHA represents a decision to require safeguards for the health of employees even if such measures substantially increase production costs. This is not, however, the same thing as saying that Congress intended to require immediate implementation of all protective measures technologically achievable without regard for their economic impact. To the contrary, it would comport with common usage to say that a standard that is prohibitively expensive is not 'feasible.'22 Senator Javits, author of the amendment that added the phrase in question to the Act, explained it in these terms:

As a result of this amendment the Secretary, in setting standards, is expressly required to consider feasibility of proposed standards. This is an improvement over the Daniels bill, which might be interpreted to require absolute health and safety in all cases, regardless of feasibility, and the Administration bill, which contains no criteria for standards at all.

S.Rep.No.91-1282, 91st Cong., 2d Sess., at 58; Legis.Hist. at 197.

The thrust of these remarks would seem to be that practical considerations can temper protective requirements. Congress does not appear to have intended to protect employees by putting their employers out of business-- either by requiring protective devices unavailable under existing technology or by making financial viability generally impossible.

This qualification is not intended to provide a route by which recalcitrant employers or industries may avoid the reforms contemplated by the Act. Standards may be economically feasible even though, from the standpoint of employers, they are financially burdensome and affect profit margins adversely. Nor does the concept of economic feasibility necessarily guarantee the continued existence of individual employers. It would appear to be consistent with the purposes of the Act to envisage the economic demise of an employer who has lagged behind the rest of the industry in protecting the health and safety of employees and is consequently financially unable to comply with new standards as quickly as other employers.23 As the effect becomes more widespread within an industry, the problem of economic feasibility becomes more pressing. For example, if the standard requires changes that only a few leading firms could quickly achieve, delay might be necessary to avoid increasing the concentration of that industry. Similarly, if the competitive structure or posture of the industry would be otherwise adversely affected-- perhaps rendered unable to compete with imports or with substitute products24 -- the Secretary could properly consider that factor. These tentative examples are offered not to illustrate concrete instances of economic unfeasibility but rather to suggest the complex elements that may be relevant to such a determination.25

With the aid of the foregoing analytic background of the procedural and substantive provisions of the Act, we turn to the specific objections raised by petitioners to the standards.

1. Effective Date for the Two Fiber Standard

The most important aspect of setting the standards was the determination of an acceptable dust concentration level. Under the emergency standards, the eight hour time-weighted average airborne concentration of asbestos dust had been limited to five fibers greater than five microns in length per milliliter of air (hereinafter 'the five fiber standard'). 36 F.R. 23207, 23208.26 A principal issue at the hearings on the permanent standards was whether the standard should remain at five fibers or be lowered to two. Proponents of standards ranging from zero to twelve fibers appeared, and it is fair to say that the evidence did not establish any one position as clearly correct. The Secretary decided to resolve this doubt in favor of greater protection of the health of employees, and established the two fiber standard recommended by NIOSH and his Advisory Committee as the level ultimately to be achieved.

Industry representatives testified that they simply could not reduce concentrations to the two fiber level in the foreseeable future. In the course of formulating its proposal, NIOSH had undertaken a limited analysis of industry's capacity to comply and had recommended delaying the effective date of the two fiber standard for two years, i.e., July 1, 1974. The Secretary decided to retain the five fiber standard for approximately four years (July 1, 1976) before requiring the reduction to two fibers, in order to give employers time to prepare for the lower limit. Petitioners assert that the four year delay permitted by the Secretary is too long because (1) the health of employees is endangered thereby, and (2) employers do not need that much time.

a. Health Hazards Occasioned by the Delay

The Secretary solicited the views of several experts on the question of the predictable health effects of maintaining a five fiber standard until 1976. The experts differed sharply in some of their opinions, but their responses are generally cautious and reflect deficiencies in available data concerning the relationship between exposure to absestos dust and the likelihood of disease. The record indicates that no precise prediction of increased harm can be made at this time.27

The Secretary must establish those standards that most adequately insure that no employee will suffer material impairment of health. We cannot say, on the basis of the conflicting testimony in the record, that the Secretary erred in his prediction of the health effect of the four year delay,28 but neither can we say that employees are not exposed to some additional risk of disease because of greater exposure. In view of the Act's express allowance for problems of feasibility, the Secretary's decision to allow a four year delay is not irrational with regard to those industries that require that long to meet the standard. It is appropriate to allow sufficient time to permit an orderly industry-wide transition since, in those cases, the indeterminate degree of risk involved is counterbalanced by considerations of feasibility; it is not, however, a risk to which employees should be needlessly exposed.

b. Industrial Compliance Capability

The evidence indicates that significant inter-industry, as well as intra-industry, differences exist concerning the time needed by employers to meet a two fiber standard. Within particular industries the concentration levels at some plants, usually newer ones, are much lower than at others. Supp.App. 12b. More importantly, some industries could implement a two fiber standard more quickly than others.29 The Director of NIOSH recommended that the Secretary require compliance sooner than 1976 where possible, and that he prohibit degradation of workplaces with concentrations currently below the limits.

Despite this recommendation and the evidence of these differences, the Secretary issued a single uniform effective date for all employers in all industries. He explained this decision as follows:

It is concluded there should be one minimum standard of exposure to asbestos applicable to all workplaces exposed to any kind of mixture of kinds of asbestos. Reasons of practical administration preclude a variety of standards for different kinds of asbestos and of workplaces.

We cannot say on this record that an attempt to assign differing effective dates to employers within an industry based on the time needed by each employer to alter his plant would be practicable. However, insofar as inter-industry differences are concerned, those reasons of practical administration are neither explained nor readily apparent.30

Government counsel suggested at oral argument that the Secretary possessed insufficient information at the time of formulation of the standards to differentiate among industries. That may be true, but it is unclear whether the lack of information is a cause or result of the Secretary's approach. When the Secretary sought opinions concerning compliance capabilities, the responses suggested that some industries could comply almost immediately, some could comply within two years, and some might require longer. The lack of more specific information may be partially attributable to the Secretary's failure to seek it. The record reflects little, if any, effort to cross-examine industry witnesses on this point, and, notwithstanding their insistence that a two fiber standard was not practicable, such questioning might have elicited some information pertinent to inter-industry distinctions.31

Separate standards for different industries would not appear to create opportunities for employers in one industry to challenge their standards on the grounds that standards for another industry were less demanding. The only relevant question would be whether the time schedule established for each industry was feasible for that industry; therefore, comparisons with the standards established for a different industry with different technological problems would be pointless unless the two industries were in competition with one another.32 If one industry could gain a competitive advantage over another by virtue of differing standards, employers in the disadvantaged industry might challenge the standards on a comparative basis, but the threat of problems of that type does not appear on the basis of the record before us to justify a uniform standard for all industries.

It may be that the task of devising categories and classifying employers by industry would be unmanageable in view of the many diverse uses of asbestos. However, there is no evidence to that effect in the record, and it is not for the court to guess at the Secretary's reasoning or to supply justifications for his action. We have noted his cryptic reference to 'reasons of practical administration,' but, insofar as inter-industry distinctions are concerned, those reasons are not self-evident. Therefore, we remand this aspect of the standards to the Secretary for clarification or reconsideration.

Within six months of publication of the standards, all employers are required to monitor workplaces to determine whether the concentrations of asbestos dust are within the allowable limits; thereafter, monitoring must in all cases be of such pattern and frequency as to identify accurately the levels of exposure. Monitoring must occur no less frequently than once every six months where the concentrations may reasonably be foreseen to exceed the standards. Petitioners object that these provisions are inadequate to protect the health of employees.33

The monitoring provisions are especially important because the results of that process often determine when and what protective measures are required.34 Petitioners argue that the reasonable foreseeability qualification vests control of this key provision in the discretion of the employer. They suggest that an employer may evade the controls imposed by the standards simply by deciding that no violations can reasonably be foreseen at his plant, thereby exempting his business from the requirement that monitoring be conducted once every six months. Although it is true that the standards require the employer to exercise some judgment concerning the likelihood of a violation, we do not give the foreseeability exception the broad construction feared by petitioners.

Periodic monitoring is important to insure that concentration levels have not been allowed to increase since the initial monitoring. Further, monitoring techniques may not be entirely uniform,35 and, in any event, concentrations of asbestos in the air may not remain stable throughout a working day or from one day to the next. Thus whether or not the initial monitoring reflects a violation will be the product of several variables other than the actual average concentration of asbestos dust in that plant. In light of these factors, satisfactory results in the initial monitoring would not necessarily justify a conclusion that concentrations in excess of the limits could not be foreseen. It might well require several samples demonstrating consistently low exposure levels before such a prediction could be made with any confidence. Thus we consider the qualification based on reasonable foreseeability to be a narrow one applicable only when considerable information regarding the workplace is available. Even then occasional monitoring would be required in order to insure that no changes had occurred, but the six months schedule would not apply.

Some jobs pose particularly difficult monitoring problems, and petitioners argue that monitoring once every six months is inadequate to control such situations effectively.36 That may be true, but the problem should be adequately handled by the general requirement that monitoring be tailored to reflect exposure levels accurately. Where excessive exposure may reasonably be foreseen, monitoring is required at least once every six months and more frequently where appropriate.

Petitioners seem to fear that, because of the imprecise language of this requirement, the six months maximum interval between monitoring samples will in practice become the minimum as well, and they consider that inappropriate. The imprecision is perhaps unavoidable, however, in view of the multitude of diverse industrial situations involved. The most effective manner in which to deal with problems of this sort would appear to be to invoke the procedures for detecting and correcting violations in particular workplaces.37 In the context of a claim concerning a specific situation, it would be possible to develop a record on the question of the adequacy of the monitoring interval for that particular workplace-- a record necessarily absent in the context of these proceedings to establish a general guideline.

3. Methods of Compliance-- Scope of Applicability

The Act specifies particular methods of compliance that must be employed in order to meet concentration levels, but there is some ambiguity as to the situations in which those provisions are operative. Part of this uncertainty stems from the following sentence quoted from the initial monitoring requirement:

If the limits are exceeded, the employer shall immediately undertake a compliance program in accordance with paragraph (c) of this section.

29 C.F.R. 1910.93a(f)(1).

Petitioners fear, and the Government brief seems to suggest, that this language restricts applicability of the provisions concerning methods of compliance to those situations in which the initial monitoring reflected a violation. Those provisions are not by their terms so restricted, nor does an examination of this language suggest that they were meant to be. They state broadly, for example, that engineering controls 'shall be used to meet the exposure limits . . ..' The provision dealing with particular tools states:

All hand-operated and power-operated tools which may produce or release asbestos fibers in excess of the exposure limits prescribed in paragraph (b) of this section . . . shall be provided with local exhaust ventilation systems . . ..

29 C.F.R. 1910.93a(c)(1)(iii).

The modifying clause italicized above would hardly seem necessary if this provision applied only to those workplaces where a violation has already been detected.38 Finally, the work practices provisions are concerned with activities that may create high concentrations of asbestos for short periods of time-- situations that can be identified as dangerous without resort to monitoring and as to which monitoring is least effective.39

Specific control measures can be used not only to correct violations that have been discovered but also as a supplement to the monitoring procedure. Whether monitoring detects impermissible concentrations of asbestos dust in a particular workplace may depend upon when and how the sample is taken. The danger that hazardous conditions will consequently exist undetected and unremedied can be reduced by requiring that certain objective control measures be employed as well. Whereas monitoring requires skilled technicians, anyone, including a labor union representative, could determine whether exhaust fans or wet handling methods were being employed.40 Since the record indicates a shortage of monitoring personnel, this may prove to be one of the most important benefits of these provisions.

For these reasons, we do not construe the methods of compliance provisions as being applicable only in those instances in which a violation has already been detected. Although the terms of those provisions are often sufficiently broad to allow considerable flexibility in the choice of methods, they are generally applicable to all workplaces covered by the asbestos standards.

4. Use of Moisturization as a Control Method

The standards require that asbestos be moisturized before being handled and worked to avoid creating dust 'unless the usefulness of the product would be diminished thereby.'41 Petitioners argue that this requirement should not be modified by considerations of utility. In considering this issue it is important to recall that this is a supplemental regulation. The employer must comply with the concentration limits in any event; the issue is whether he must use this particular method.

Moisturization is effective and is subject neither to the uncertainties of monitoring nor to the technical problems of malfunction and repairs incident to dust collection and exhaust systems. The Secretary recognized these advantages and required that this method be employed where practicable, but left to the employer the opportunity to seek other methods in some circumstances. This solution seems to accommodate as nearly as possible the interests of safety and efficiency. Petitioners' objection is that flexibility is again achieved by relying on a judgment by the employer. Although we can understand the concern on the part of labor representatives in a matter of this sort, it is too early in the history of this statute to presume that employers will not make a good faith effort to comply.

Further, this aspect of the Secretary's standards is consistent with the general Congressional scheme. OSHA applies to an estimated 4 million employers and 57 million employees, but the Secretary has only about 600 inspectors and 100 industrial hygienists. In this situation some reliance on self-policing by employers is clearly necessary. Although this self-regulatory aspect of the statute would make specific, objective standards desirable, devising standards of this sort that will be generally applicable to thousands of different situations involves enormous practical difficulties. Setting forth inflexible requirements of particular methods would be effective only if the Secretary could anticipate all possible problems and devise a uniform approach appropriate to each. Rather than attempt to do so, the Secretary has in some instances employed a somewhat open-ended scheme to allow development of different methods for different workplaces. As a safeguard against abuses by the employer, Congress has provided employees or their representatives with numerous opportunities to participate in the policy process or to instigate enforcement by the Secretary.42

In order to insure that employees handle materials likely to produce asbestos dust carefully and to prevent persons from entering areas where they will be exposed to such dust needlessly, cautionary labels and warning signs are required by the Secretary's standards. Petitioners assert that the language selected by the Secretary is not strong enough for these purposes, and they have suggested alternatives that they consider better suited to the task.43 Although the language favored by petitioners was recommended by NIOSH and was included in the standards as originally proposed, we cannot say it is required by OSHA. After examination of the standards promulgated in light of the terms of the statute, we conclude that they are within the range of the Secretary's discretion.

The standards require that all employers provide a medical examination (1) when an employee is first assigned to a job involving exposure to asbestos, (2) when he leaves that job, and (3) annually during his employment in the position. Petitioners do not attack this timetable or the substantive requirements governing the nature of the examination, but they do assert that the examinations should be given by a physician of the employee's choice, and that the results should be given to the employer only if the employee decides to do so.44

Petitioners argue that the standards violate the principle of physician/patient confidentiality because they would allow the physical examinations to be conducted by company doctors and would make the results available to the employer. Confidentiality is necessary, they argue, to avert the possibility that, in hiring and discharging employees, employers will discriminate against those with symptoms of asbestos-related disease or prior histories of exposure to asbestos dust. The Secretary recognized this potential problem, and stated that uses of the records would be scrutinized carefully. However, he did not consider the possibility of such abuse sufficient to outweigh the opposing considerations.

The standards require the employer to take into consideration the result of an employee's most recent physical examination in making assignments to jobs requiring the use of respirators. An employee who cannot safely perform such a job is to be reassigned without loss of seniority or wages. The Secretary reasoned that the salutary purposes of this provision could not be fulfilled if employers were denied access to the medical records.45

Since the results are to be made available to the employer, allowing the employer to select the physician has the advantages of both convenience and efficiency. Some employers already maintain an industrial medical staff skilled in dealing with asbestos-related medical problems,46 and the requirements of the standards as promulgated appear likely to encourage the development of additional staffs of this sort. The employee would then benefit from the expertise associated with specialization, and the state of medical knowledge concerning these diseases should likewise be improved. All of these factors, identified in the record, operate to make the Secretary's decision on this point a nonarbitrary one.

Many of the problems the Secretary faced in establishing standards regarding asbestos dust were directly attributable to the lack of reliable information concerning asbestos-related diseases. The Act attempts to correct this deficiency by requiring that:

Each employer shall make, keep and preserve, and make available to the Secretary or the Secretary of Health, Education, and Welfare, such records regarding his activities relating to this chapter as the Secretary . . . may prescribe by regulation as necessary or appropriate for the enforcement of this chapter or for developing information regarding the causes and prevention of occupational accidents and illnesses.

29 U.S.C. 657(c)(1).

Under the standards promulgated by the Secretary to implement this provision of the statute, employers must retain for twenty years records of each employee's required medical examinations. The standards also require that employers keep records of personal and environmental monitoring, but these records need only be kept for the most recent three years.

Petitioners challenge both requirements as inadequate to protect the health of employees and to advance the present state of knowledge in this area. They argue that, in view of the long latency period associated with many asbestos-related diseases, all records, beginning with initial exposure and continuing throughout the life of each employee, should be retained. Since the recordkeeping requirements for medical examinations and monitoring differ not only in retention period but also in purpose, we treat each separately and reach different conclusions with regard to the two standards.

a. Medical Records

Records of an employee's prior physical examinations may assist a physician in the detection, diagnosis and treatment of that employee's illnesses. Although completeness would be desirable, it may be that the most critical medical records are the most recent ones, perhaps reflecting the examinations of the past two to five years.47 Since Congress expressed the desire to avoid unnecessary recordkeeping and to impose the 'minimum burden upon employers,' 29 U.S.C. 657(d),48 the Secretary's decision to require retention of only records from the most recent twenty years was permitted him under the Act.

Records of past medical examinations may also be useful in research; for example, they may facilitate the plotting of the longitudinal course of disease or the evaluation of the long-term effects of various levels of exposure to asbestos fibers. These purposes might require a longer retention period than would often be critical for treatment of particular patients; however, even for research, records of some effects of exposure are detectable in the examinations.49 Thus, even though that often will not occur until several years after initial exposure, the Secretary could reasonably conclude that the records of twenty years would ordinarily encompass the period during which disease could be detected and studied.

Further, we note that twenty years simply marks the minimum time which records must be maintained by employers. If retention is required for a longer period of time for purposes of research, the research organization can obtain copies of the records from the employer and keep them. The standards require an employer to supply these records to the Government upon request, and the twenty year period appears reasonably adequate for these purposes. This arrangement seems most practical since the records of every employee may not be needed for research. Ordinarily, sampling techniques would be employed to select limited numbers of representative records. After research groups have had a reasonable opportunity to obtain the records, and once the other records have been digested in statistical summaries, the remaining records arguably are of little utility.

b. Records of Exposure Levels Detected by Monitoring

The three year retention period for the monitoring records seems surprisingly short in comparison to the twenty years requirement for medical records-- especially in view of their respective functions. Whereas the medical records of primary importance may be those beginning with the first manifestations of a disorder, a complete record of an employee's history of exposure to asbestos prior to the development of disease would be important to research. At this point in time, relatively little is known about the causal relationship between exposure to asbestors dust and various diseases. Persons manifesting disorders are being studied, but that research is hindered by the lack of information concerning the exposure levels at industrial workplaces in the past. As the Secretary observed:

We have now evidence of the consequences of exposure, but we do not have, in general, accurate measures of the levels of exposure occurring 20 or 30 years ago, which have given rise to these consequences.

For this reason, in the proceedings before the Advisory Committee two members of the committee suggested that a twenty year period for exposure records and a five year requirement for medical records would be appropriate.51

The Secretary did not explain his decision to require only a three year retention period for exposure records, but the Government has suggested several justifications in its brief. Passing over the problem of considering a rationale advanced by counsel rather than by the Secretary, we have examined those arguments and conclude they do not adequately clarify the Secretary's action.

The Government notes that the Act specifies that citations for violations must be issued within six months of the occurrence of the violation. Thus it would appear that a period of three years is more than adequate for enforcement purposes. This argument is accurate, but it does not speak to the Secretary's obligation under OSHA to require retention of those records necessary to develop information concerning the causes of disease.

The second justification offered is that it is the responsibility of NIOSH, the Secretary, and HEW to collect the exposure data necessary to re-evaluate the standards. 29 U.S.C. 671, 673. Employers must supply records on request,52 and the Government asserts that three years is an adequate period of time to allow for the agencies to make such requests.

A single permanent storage center for all these records might well be the best solution. Since employees may change jobs and employers may cease operations, continuity may be best achieved if the agency collects and retains the records.53 If such a program is in fact implemented, the three year period may be acceptable, but there is no indication that current agency procedure includes compilation of such records. Further, the following footnote from the Government's brief raises some question concerning the interpretation of the purposes of recordkeeping:

These records are not relevant concerning individual records of exposure. Monitoring of employee exposure is not continuously performed, therefore, worklife exposure is never truly obtainable. Unless specific employees are linked with exposure data these records are not relevant for purposes of determining cumulative employee exposure. Because these records are not relevant to total employee exposure there is no need to transfer monitoring records if a company goes out of business.

Brief for Respondents at 50 n. 35.

As noted above, data concerning prior exposure is considered very important in determining the causal relationship between exposure to asbestos dust and disease. To the extent that the note quoted above suggests that collection of such data is unnecessary, it reflects a misinterpretation of the Act. It may be that an adequate basis for research may be established on the basis of exposure levels generally prevalent in the industry without linking each employee to the particular samples taken from his workplace, but the purposes of the recordkeeping requirements cannot be fulfilled without providing some adequate means of relating health records to exposure levels.

The Secretary has provided no explanation of the relatively short retention period for monitoring results, and we find no adequate assurance in the record that the requirement as promulgated will provide the date needed for research into the causes and prevention of asbestos related disease. Consequently, we remand the recordkeeping requirements to the Secretary for such modification or clarification as may be necessary to insure that the statutory objectives will be fulfilled.

Except for the remand we order for reexamination of (1) the uniform application of the 1976 effective date for the two fiber standard and (2) the three year retention period for monitoring records, the petition for review is denied. All of the challenged features of the standards appear to partake of an essentially legislative type of decision making by the Secretary in the performance of the broad delegation made to him by Congress. Had any one of these decisions been made in the first instance by Congress itself and embodied in the statute, its vulnerability to judicial scrutiny would have been dubious indeed. In this context, therefore, judicial review inevitably runs the risk of becoming arbitrary supervision and revision of the Secretary's efforts to effectuate the legislative purposes in an area where variant responses might each be legitimate in the sight of Congress.

What, in our view, differentiates the two provisions we have remanded from those we have left untouched is that the record, examined closely in relation to the relevant concerns of the Act, leaves nagging questions-- even for the inexpert observer-- as to the reason and rationale for the Secretary's particular choices. However the statutory standard for our review may be characterized, we consider that our dispositions fall within it.