April 15, 1986

Peter G. Nash, Esq.
Ogletree, Deakins, Nash, Smoak and Stewart
1200 New Hampshire Avenue, N.W., Suite 230
Washington, D.C. 20036

Dear Mr. Nash:

This is in response to your February 10 letter concerning the Cotton Dust Standard. You requested on behalf of your client, the Minnesota Mining and Manufacturing Company (3M), a respirator manufacturer, that the Occupational Safety and Health Administration (OSHA) grant an administrative stay of enforcement of the respirator selection table of the Cotton Dust Standard. Specifically, you have objected to OSHA's decision to limit the use of disposable respirators to situations where cotton dust concentration is no greater than five times the permissible exposure limit (PEL).

The new standard defines "disposable respirators" to mean that the filter element is an inseparable part of the respirator. 29 CFR 1910.1043(f)(2)(i) Table 1 note 1 (50 Fed. Reg. 51175)(Dec. 13, 1985). As a manufacturer of one such [disposable] respirator, the "8710", your client has requested the agency to reconsider its decision to assign this class of respirators a protection factor (PF) of five, and to stay the implementation of the respirator selection table pending reconsideration.

[This document was edited on 8/18/99 to strike information that no longer reflects OSHA policy.]

The assigned protection factor reflects the ratio of contaminant outside the respirator to contaminant inside the respirator. It is set by taking into account two measurable elements -- filter efficiency and fit factor. In addition, the Agency must take into consideration that respirators do not perform as well in the workplace as they do under laboratory conditions.

"Filter efficiency" tests the amount of the contaminant which is captured by the filter itself. The "fit factor" is a measurement of the leakage that occurs between the respirator face seal and the face. Face seal leakage can be determined by use of either a quantitative or qualitative fit test. A "quantitative" fit test (QNFT) uses laboratory equipment to numerically measure leakage of the seal between the respirator facepiece and the wearer's face. A "qualitative" fit test (QLFT) assesses the adequacy of respirator fit by determining whether or not an individual wearing the respirator can detect the odor, taste, or irritation of a chemical introduced in the vicinity of the wearer's head.

A simple method of determining prior to each use whether a respirator has any gross leaks is the use of a "positive and negative pressure fit check." To perform this test, an employee places his hand over the respirator's inhalation or exhalation valves and blows or inhales to determine whether air is escaping from the face seal. This method is not adequate for assessing fit factor, however.

In the cotton dust standard, OSHA developed a table which assigns protection factors to respirators for protection against exposure to cotton dust (50 Fed. Reg. 51175, Table 1). For each different category listed in the table, the specified respirators can be used for protection up to the assigned multiple of cotton dust concentration. In developing this table, the Agency accepted the respirators' filtering capacity and looked to fit to determine the different protection factors. Accordingly, although there are various types of filter media, we uniformly accepted the filtering capacity of all particulate filters. Thus, all disposable respirators with any type of particulate filters are grouped together and given an assigned protection factor of five. This includes single use, dust, dust and mist, or dust, mist and fume.

The Agency's decision to assign disposable respirators a protection factor of five is explained in the preamble to the standard at 50 Fed. Reg. 51153-54 (Dec. 13, 1985). OSHA noted that it was now using the term "disposable respirators" to refer to the same respirators which had been referred to as "single-use" respirators in the 1978 cotton dust standard. This definitional change was made for consistency purposes. The National Institute of Occupational Safety and Health (NIOSH), while recognizing that there have been no significant changes in the construction and performance of these respirators' had tested some "single-use" respirators and approved them as respirators with "replaceable" filters. 30 CFR Part 11 MSHA/NIOSH Respirator Testing and Certification Requirements. Regardless of this certification action, NIOSH, in a report submitted to OSHA's new cotton dust standard record, supported the respirator provisions in the proposed standard, which were adopted in the final standard. Likewise, Carrol Bailey, a certified industrial hygienist, testified in favor of these provisions.

[This document was edited on 8/18/99 to strike information that no longer reflects current OSHA policy.]

In the preamble OSHA also responded to the contentions made by 3M representatives during the hearing and in a short written comment (Ex. 187-12). In summary terms, you objected to the assignment of a protection factor of five because: (1) NIOSH and the American National Standards Institute (ANSI) have recognized that disposable respirators are equivalent to non-disposable respirators with replaceable filters; (2) OSHA allows use of disposable respirators as protection against lead for ten times the PEL; (3) there are available and accepted fit tests to assess the facefit of disposable respirators; and (4) limiting the use of disposable respirators to a protection factor of five would force the cotton dust industry to purchase more expensive and less comfortable respirators.

In the cotton dust preamble, OSHA responded to each of these contentions. First, OSHA acknowledged that NIOSH does judge "disposable" respirators equal to respirators with disposable filters. But, as explained in the preamble, NIOSH's certification system does not quantitatively account for leakage which may result from improper fit. Rather, NIOSH tests only filter efficiency and breathing resistance; and these tests do not result in the assignment of a protection factor. Consequently, the NIOSH certification does not indicate whether the disposable respirator provides as much protection as the replaceable-filter half mask respirators when fit as well as filter efficiency are taken into account.

In response to 3M's assertion that ANSI likewise recognizes the equivalency of disposable and nondisposable respirators, the preamble explained that, while ANSI did perform qualitative fit tests on high efficiency respirators, it did not perform these tests on disposable respirators. Hence, the ANSI respirator protection factor table was developed without any fit testing for disposable respirators.

Regarding 3M's general statement about the availability and acceptability of facefit tests, the preamble emphasized that the simplest such test--the fit check--usually can be performed effectively only on rigid elastomeric half mask respirators, which have inhalation and exhalation valves. These valves can be blocked off easily by the employee's hand. The disposable respirators permitted for use under the cotton dust standard do not, as the preamble explains, have either inhalation or exhalation valves. Therefore, a simple effective fit check cannot be performed on these respirators.

The preamble further explained that QNFT cannot be performed on disposable respirators, because their construction precludes use of the probe necessary to perform these tests. Moreover, the Agency indicated it would be inappropriate to require employers to conduct QLFT regularly, since these tests are time consuming and inappropriate in light of the nature of the hazard associated with cotton dust exposure. In this regard, OSHA emphasized that, since virtually no employees in the textile industry are exposed at levels greater than five times the PEL for an eight hour period, the requirement to perform QLFTs was unwarranted.

The preamble also explained that, although OSHA does allow the use of QLFT to establish a protection factor of ten for disposable respirators under the lead standard, the adoption of similar provisions for the cotton dust standard is inappropriate. OSHA explained that the blood lead tests required under the lead standard provide a reasonably direct indication of whether lead is getting into the employee's breathing zone. In contrast, the pulmonary function testing required under the cotton dust standard cannot indicate how much cotton dust enters the breathing zone.

Finally, the preamble emphasized that assigning disposable respirators a protection factor of five would not force employers to purchase more expensive and less comfortable respirators. This is so because virtually no employees are actually exposed at levels above five times the PEL for an eight hour period.

In your request for reconsideration and for an administrative stay, you have raised several of the same arguments addressed in the preamble, and explained why you disagree with OSHA's conclusions about these matters. In addition, you claim that by redefining the "single use" respirators as "disposable" respirators and still according them a protection factor of five, the Agency has unjustifiably changed the respirator selection table established in the 1978 cotton dust standard.

In support of many of your arguments, you have included with your stay request several attachments which were not submitted during the comment period or at the public hearing. Although a 3M representative testified at the hearing that you would submit information about tests that were performed with the 8710 respirator at some textile mills, you failed to submit these data to the record (see Docket H-052E, 92383 Transcript 904).

In considering your pending request for reconsideration, OSHA nonetheless considered your belatedly-submitted reports. After consideration of all your objections and the information attached to your request, the Agency adheres to its prior view, and accordingly must deny your request.

First, it deserves emphasis that virtually all employers in the cotton industry can meet the standard's respirator requirements by providing respirators with a protection factor of five, such as 3M's disposable respirators. This is so because, to our knowledge, no worker in cotton textile manufacturing operations and few, if any, in waste processing operations are exposed to cotton dust at concentrations that exceed five times the PEL (see Centaur Report, Docket H-052E, Exhibit 105). Consequently, little weight can be given to 3M's speculative claims regarding lost sales and profits.

In this vein, it should also be noted that the respirator selection table under scrutiny here pertains only to cotton dust. To the extent 3M wishes to participate in establishing broad general principles regarding the acceptability of its disposable respirator, we invite you to submit evidence during the comment period and public hearings that will follow publication of the proposed amendment to OSHA's Respiratory Protection Standard (29 CFR 1910.134). In view of the reality of current conditions in the cotton industry and given that OSHA has scheduled a full re-evaluation of the Respiratory Protection Standard, a reconsideration of or stay of the provisions of the cotton dust respirator selection table would serve no useful purpose.

[This document was edited on 8/18/99 to strike information that no longer reflects current OSHA policy.]

Your stay request focuses largely on your assertion that OSHA has changed the respirator selection table with regard to 3M's 8710 respirator. This premise is incorrect. When OSHA originally proposed a cotton dust standard in 1976, single use respirators, such as 3M's 8710, were not permitted at all. During the rulemaking that followed that proposal, the pertinent issue was whether these respirators should be permitted for any purpose to protect against cotton dust exposure. In the final 1978 cotton dust standard, OSHA determined that these respirators would be permitted, and assigned them a protection factor of five.

When OSHA published its proposal in 1983 for an amended cotton dust standard, the Agency again proposed that these-respirators be assigned a protection factor of five. The Agency included these respirators under a different nomenclature, i.e., "disposable" respirators, to reflect current practice. Despite this definitional change, the Agency made it clear in this proposal that it was referring to the same respirators which had previously been labeled "single-use," and that it intended to continue an assigned protection factor of five for those respirators. In promulgating the final standard, the Agency, well aware of exactly which respirators it was addressing, determined that it was appropriate to maintain the protection factor of five for these respirators. Accordingly, the final standard reflects the maintenance of the status quo with regard to these respirators; it does not, as you suggest, reflect any change in the protection factor accorded them.

This point is not altered by your statements that the filter efficiency of the 8710 has been improved⁽¹⁾. Even if your statement is true, you have admitted that "the construction of the 3M 8710 has not basically changed" (Request for Administrative Stay and Reconsideration at p. 12 n.11). In the Agency's view, it is the construction of these respirators that most saliently distinguishes them from the elastomeric half-mask respirators for the purpose of assigning the protection factor. As explained in the preamble, the fact that the construction of these respirators makes assurance of a close fit less certain justifies assigning them a protection factor of five. For this reason, your reliance on NIOSH's certification is misplaced, since that certification allows only-for filter efficiency, and does not account quantitatively for leakage due to fit.

Likewise, your reliance on ANSI's table does not detract from the informed judgment and expertise of OSHA's scientists that disposable respirators should be assigned a protection factor of five for cotton dust. To the extent that you point to OSHA's nonbinding internal use of ANSI's table for enforcement of OSHA's general respirator provisions at 29 CFR 1910.134, it deserves emphasis that the respirator standard broadly covers over 400 substances, whereas the decision here was based specifically upon OSHA's evaluation of cotton dust hazards. The cotton dust standard and its respirator selection table, promulgated under §6(b) of the OSHA Act, takes precedence over the general respirator standard with respect to the selection of respirators to protect against cotton dust exposure.

In response to OSHA's suggestion that the difficulty of assuring proper fit distinguishes disposable respirators from half-mask respirators, you refer to documents included in the docket for OSHA's lead standard and to Attachment F submitted with your Stay Request. Regarding the lead standard, you rely on a rulemaking which allows the use of QLFT as well as QNFT to establish a protection factor of ten in protecting against lead exposure. As explained in the preamble to the cotton dust standard and summarized earlier in this letter, the Agency does not believe similar provisions are appropriate for the cotton dust standard for several reasons. To briefly reiterate, the lead standard requires frequent blood lead monitoring, which would most directly indicate if lead is present in the employee's breathing zone; the pulmonary function tests required by the cotton dust standard do not provide any equivalent indication of whether or not the contaminant is present in the employee's breathing zone. Second, in view of the Agency's finding that few if any workers are exposed to cotton dust in excess of five times the permissible exposure level, the Agency determined that the inclusion of detailed QLFT provisions such as those included in the lead standards were unwarranted. If detailed QLFT requirements were included in the cotton dust standard, the Agency would be imposing an unnecessary burden on employers and on its own enforcement personnel, who would have to use Agency time and resources to assure that employers were properly performing QLFTS. Finally, OSHA made it clear in the QLFT lead rulemaking that it was drawing conclusions only for the lead standard.

To the extent that your argument relies on the existing allowance of a ten protection factor for disposable respirators for lead, we note that this allowance is only pursuant to an administrative stay rather than a final regulatory determination. Prior to our final determination, NIOSH must finalize the study referred to in footnote 1 of this letter.

In a final effort to show that 3M's disposable respirators can be effectively fit checked to allow for a protection factor of ten, you have attached results of tests you performed on the 8710 respirator. You explain your belated submission of these results by suggesting that you had no notice that OSHA was concerned about fit checking of disposable respirators. We believe that the record plainly evidences the Agency's concern about respirator fit, which of course encompasses any relevant fit check data. Accordingly, we are unpersuaded by your suggestion that you had no previous opportunity to submit these test results.

In any event, our review of this data has not convinced us that reconsideration is appropriate. Given the small size of the study, it fails to provide adequate statistical evidence that your suggested fit check procedure is effective to warrant a protection factor of ten for the 8710 respirator.

Specifically, you tested a total of 23 subjects in several trials to see how many of those who failed a QNFT would also fail a positive pressure fit check procedure (PPFC) performed by 3M. Using a fit factor value of 100⁽²⁾, 16 workers failed the QNFT one time, and one worker failed it in two separate instances. Of these workers, two were erroneously passed by 3M's PPFC. Thus, the estimated probability of this type of error⁽³⁾ under conditions similar to those of 3M's study is 2/17= 0.12 = 12%, with a 2 sided 95% confidence interval of 1.5% to 36.5%. (This uncertainty interval for the error rate of the PPFC is the best interpretation that can be obtained from the available data.) Accordingly, as many as 36 per 100 improperly-fitted wearers of 3M's 8710 respirator could be erroneously passed by 3M's PPFC procedures. This is not strong laboratory data to justify a protection factor of ten.

When a similar analysis is performed using a screening level of ten, a similar result is seen. Of seven wearers who had fit factors less than ten, none of them was passed by the test. The estimated probability beta of a beta error under conditions similar to the 3M study is 0/7 = 0.0 = 0%, with a 2 sided 95% confidence interval of 0% to 41%. This means that as many as 41 per 100 improperly-fitted wearers of 3M's 8710 respirator could be erroneously passed by 3M's PPFC procedure. As with the previous analysis, this data does not provide strong laboratory evidence that 3M's PPFC is effective to allow a protection factor of ten for the 8710 respirator.

For the reasons stated in the cotton dust preamble, and based upon the foregoing analysis of additional information you have brought to our attention, we do not believe that reconsideration of or a stay of the cotton dust respirator selection table is appropriate. Accordingly, we must deny your request.

Sincerely,

Patrick R. Tyson
Acting Assistant Secretary
for Occupational Safety and Health

Footnote(1) To the extent you rely on an incomplete draft of a NIOSH study submitted as Attachment B to your Request, we note first that this document is part of OSHA's asbestos docket; it is not part of the cotton dust docket. In any event, the report is only a preliminary and incomplete draft of NIOSH's study, which NIOSH does not regard as representing its final conclusion. Finally, the completed draft (which, we again emphasize is just a preliminary draft) shows that the 8710 has a filter efficiency which is significantly below that for most other respirators included in the draft of the study attached to your request (see Attachment "A" to this letter). (Back to text)

Footnote(2) The fit factor value is used as a screening level for an individual to determine what is an acceptable fit for a half-mask respirator where a protection factor of ten is desired. This value should be chosen in order to achieve a high probability of attaining the desired protection under actual working conditions (working protection factor or WPF). As the projected WPF is increased, the fit value must be higher. You have previously acknowledged that a fit value of 100 is appropriate to test for a WPF of ten. (OSHA Docket H049A, Ex. 39 p. 12) NIOSH has supported this view (OSHA Docket H049A, Ex. 65, pp. 2-4). On pages 1 and 4 of Attachment F to your present stay request, however, you seem to suggest that your fit test and fit check must only reject those wearers with fit factors less than ten and not 100 as you previously stated. As demonstrated above, the use of either screening value does not support the assignment of a ten PF for the 8710. (Back to text)

Footnote(3) This type of error is referred to as a "beta error." NIOSH has defined the beta error as the occurrence of a false negative response from the fit test (or fit check) among those wearers with inadequate fit factors. For the data in Attachment F, it would be the occurrence of a "pass" for a wearer with a fit factor less than the assigned fit factor value. (Back to text)