Oregon Occupational Safety & Health Division v. Moore Excavation, Inc.

SERCOMBE, J. The Oregon Occupational Safety and Health Division (OR-OSHA) seeks judicial review of an administrative law judge’s (ALJ) order vacating a citation issued by OR-OSHA to Moore Excavation, Inc. (Moore).1 Moore was cited for failure to “tag and withdraw from service a damaged portable ladder” in violation of 29 CFR section 1926.1053(b)(16),2 a federal regulation adopted in Oregon by reference pursuant to the Oregon Safe Employment Act (OSEA).3 OAR 437-003-0001(24)(d) (9/26/07); see also ORS 654.035 (authorizing the Director of the Department of Consumer and Business Services to adopt such regulations). Although Moore conceded that a violation of that regulation had indeed occurred, the ALJ vacated the citation after determining that OR-OSHA had failed to prove that, as a result of the violation, it was “reasonably predictable” that a worker would have become exposed to the hazard created by that violation. OR-OSHA assigns error to the ALJ’s vacation of the citation, arguing on judicial review that the ALJ erroneously imposed that standard of proof in assessing the employee “exposure” element of its prima facie case. It asserts that, “under the proper standard,” it was required to prove only “that the [regulation] was violated and that employees had access to the violative condition.” OR-OSHA concedes that evidence of employee exposure to the hazardous condition was required to prove the violation. However, OR-OSHA argues that it was required to prove only that employee exposure to the hazardous condition was possible — not, as required by the ALJ, that it was “reasonably predictable [.]” On review for errors of law, ORS 183.482(8)(a), we conclude that the ALJ correctly imposed a reasonable-predictability standard with respect to OR-OSHA’s burden to prove employee exposure — 569 — to the damaged ladder. On review for substantial evidence, ORS 183.482(8)(c), we conclude that the ALJ did not err in vacating the citation after determining that OR-OSHA had failed to meet its burden under that standard.4 In January 2008, Moore — an excavation company operating in the Pacific Northwest — was repairing a water line in Turner. The project was quite large, involving approximately “15,000 lineal feet of pipe and several discrete work areas.” On January 18, OR-OSHA Safety Compliance Officer Weaver opened an inspection at the site, contacting site superintendent Coate and conducting a “walk through” of the site. Importantly, at the time of Weaver’s inspection, the bulk of the work had been completed and the “active construction zone” within the three-mile worksite had been reduced to “roughly two to three thousand [lineal] feet” with approximately 18 employees working in that area; however, Weaver’s inspection encompassed both active and inactive areas. In the course of his inspection, Weaver “identified a 24-foot extension ladder that had been damaged” but had not been “tagged” (that is, marked with language such as “Do Not Use” as set forth in 29 CFR section 1926.1053(b)(16)). The ladder had been visibly damaged approximately one week prior to the inspection when it was run over by an excavator, and the top rungs were “severely bent.” At the time of the inspection, it was located inside of a large, “enclosed fenced area” that was “not a part of the active work zone at the time * * That “inactive” area was approximately 2,000 square feet in size, encompassed a large, excavated pit, and was surrounded by a “cyclone” fence topped with barbed wire. Weaver observed the ladder resting in the bottom of that pit, with the visibly damaged portion of it sticking out — 570 — and leaning up against the fencing.5 After Weaver pointed it out, a foreman was immediately called to remove the damaged ladder from the worksite. Ultimately, Weaver cited Moore for an “ [o] ther than serious violation” in connection with its failure to tag and remove the damaged ladder from the worksite and imposed a $0 penalty. Moore appealed to OR-OSHA, and that appeal was subsequently transferred for a hearing before an ALJ on August 5, 2010. Soon after, that ALJ issued an opinion and order vacating the citation and framing the issue as follows: “The question in this case is whether the mere possibility of exposure as opposed to a reasonable likelihood of exposure is enough to support the citation in this instance.” OR-OSHA then moved for reconsideration, and, on April 20, 2011, the ALJ issued an opinion and order on reconsideration affirming the citation — relying in large part on this court’s opinion in OR-OSHA v. David A. Mowat & ML Mowat Co., 237 Or App 576, 240 P3d 748 (2010) (Mowat), which we discuss below. Moore then moved for reconsideration, and, on July 28, 2011, a different ALJ6 issued a second opinion and order on reconsideration, which again vacated the citation. Addressing our opinion in Mowat, the ALJ ultimately reasoned that “[a] 11 Mowat really tells us here is that OR-OSHA [did] not have to prove that a broken ladder is hazardous to use, a proposition that employer does not challenge.” Addressing OR-OSHA’s burden to prove employee “exposure” (still in existence post -Mowat), the ALJ reasoned: — 571 — “First, it is obvious that the ‘zone of danger’ here is restricted to use of the damaged ladder itself. *** [H]ere, a violation would be established if it were reasonably predictable that, as a result of employer’s failure to tag and/ or withdraw from service the defective ladder, an employee would attempt to use the damaged ladder. Although there was conflicting testimony regarding whether the gate to the fenced work site here was locked, there is no persuasive evidence that the work site itself was an active site during the time the defective ladder was stored there” (Emphases added.) OR-OSHA now seeks judicial review. As noted, OR-OSHA contends on review that the ALJ erred in requiring that it “prove that there was [reasonably predictable] employee exposure to a hazard because of Moore’s failure to comply with the regulation” and argues that the citation should have been affirmed in this instance. That is, it argues that it was required to prove only that the regulation was violated and that “employees had access to the violative condition” — eschewing the ALJ’s requirement that OR-OSHA prove that it was “reasonably predictable that, as a result of the violative condition ***, a worker would become exposed to the actual risk of injury by attempting to use the defective ladder.” (Emphasis added.) In addition, relying on Mowat, 237 Or App at 580-83, OR-OSHA argues that “[n]othing in the terms of the [regulation] at issue here requires OR-OSHA to prove the existence of a hazard.” For its part, Moore “accepts the existence of a hazard and asserts only that there was no reasonably predictable employee endangerment relative to that hazard.” That is, Moore concedes that its violation of 29 CFR section 1926.1053(b)(16) “created a potentially hazardous condition” under the principle set forth in Mowat and discussed below; however, it argues that the ALJ correctly concluded that OR-OSHA failed to carry its burden of proving reasonably predictable employee exposure to that hazardous condition. Ultimately, Moore (1) asserts that “reasonable predictability” is the appropriate standard for proving employee exposure to a violative condition under the OSEA and (2) argues that the ALJ correctly determined that OR-OSHA failed to meet its burden of proof under that standard. We agree with Moore. — 572 — First, in order to clarify the parties’ arguments and the ALJ’s reasoning, we address our 2010 decision in Mowat and the distinction between proof of “hazardousness” and proof of “exposure.” In Mowat, we reviewed an ALJ’s order vacating a citation for violation of a federal regulation adopted under the OSEA that prescribed specific standards for “sloping and benching system[s]” at excavation sites. 237 Or App at 578. There, the ALJ had vacated the citation “on the ground that OR-OSHA had failed to prove that [the employer’s] violation of [the] safety [regulation] had created a hazard for [its] employees.” Id. (emphasis added). We reversed, reasoning that established federal law “draws a distinction between two types of rules” — those that “explicitly or implicitly make hazardousness an element of the rules” and those that presume the existence of a hazard. Id. at 580-82. We held that the regulation at issue fell into the latter category, stating that the regulation “presumes that an employer’s failure to conduct an excavation in accordance with one of the four options in the [regulation] creates a cave-in hazard for employees who could be affected by a cave-in. Hence, OR-OSHA was not required to present evidence as part of its prima facie case that the excavation at the [employer’s] work site had, in fact, created a hazardous condition for [the employer’s] employees.” Id. at 582-83. Like the regulation in Mowat, 29 CFR section 1926.1053(b)(16) falls into the latter category.7 A regulation mandating that employers mark damaged and unsafe equipment (e.g., a broken ladder) and remove it from service implicitly presumes that, if left unmarked and accessible, that damaged and unsafe equipment will present the hazard that a worker may use it. OR-OSHA was therefore not required to prove that Moore’s failure to tag and remove the ladder from the worksite created a hazardous condition for — 573 — employees. However, the fact that OR-OSHA did not need to prove that a violation of 29 CFR section 1926.1053(b)(16) created a hazardous condition did not relieve it of its burden to prove employee exposure to that hazardous condition. In Mowat, the employer’s workers were working in and around an excavation site with a “sloping and benching system” that did not meet the safety standards set forth in the applicable regulation (i.e., an excavation site that was itself a hazard). 237 Or App at 578-79. As a result of those facts and the parties’ attendant arguments, we did not have occasion in Mowat to address the question of employee exposure. There, the hazard was effectively the entire work-site, and employees were indisputably present within that hazardous area. Id. at 580-83. Here, on the other hand, where the hazard at issue was a damaged and unsafe ladder that was required to be tagged and removed from service, employee exposure to the hazard was not, in contrast with Mowat, an implicit certainty. In short, we agree with the ALJ’s reasoning that “[a] 11 Mowat really tells us here is that OR-OSHA d[id] not have to prove that a broken ladder is hazardous to use, a proposition that employer does not challenge.” Although this is the first time that the issue of employee exposure has been squarely before this court, we have previously discussed it, albeit largely in dictum. See Oregon Occupational Safety v. Mad Creek Logging, 123 Or App 453, 455-56, 861 P2d 365 (1993), rev den, 318 Or 381 (1994); Accident Prev. Div. v. Stadeli Pump, 18 Or App 357, 367, 525 P2d 170 (1974). First, in Stadeli Pump, we briefly acknowledged the employee-exposure element of OR-OSHA’s prima facie case and implicitly distinguished it from the element of hazardousness: “We agree with the hearing officer that it is not necessary that a state inspector ‘catch’ a workman in an unshored trench in order to prove that ‘the hazard is one to which employees are exposed’ and that the code has been violated. Other evidence may permit an inference that an unshored trench is a place where employees are currently working, although not present at the moment of inspection.” 18 Or App at 367 (emphases added). — 574 — In 1993, in Mad Creek Logging — a case that involved a rule requiring that wire rope be wound in a certain manner — we addressed the element of employee exposure and again implicitly distinguished it from the element of hazardousness while disavowing the need for proof of “actual” exposure: “The parties agree that, in order to establish a violation, OR-OSHA must show that an employee was exposed to an unsafe condition as a result of a failure to comply with the standard. Employer agrees that the spool was improperly wound, but contends, and the referee found, that OR-OSHA has failed to establish a violation of the rule, because it has not presented evidence of employee endangerment. See, e.g., Secretary v. Bechtel Corp., 2 OSH 1336, 1337 (1974). The referee found: “‘Here, neither the lay nor documentary evidence established any actual exposure to [employer’s] employees, as a result of the improperly spooled wire. At most, defendant presented evidence that the improperly spooled wire wire [sic] could result in a hazzard [sic]. According to the testimony of Mr. David Wooley, however, even properly spooled wire could result in a hazzard [sic], Wooley further testified that without performing a “pull test,” he could not estimate the degree of damage to the wire.’ “OR-OSHA contends that the referee’s conclusion is ‘legally flawed,’ because his analysis requires a finding of an ‘actual hazard,’ i.e., direct physical endangerment, as opposed to a potential hazard, before a citation will issue. That is not a correct reading of the referee’s order. Certainly, a failure to comply with a safety rule that gives rise to a potential risk of injury to an employee is a violation, even in the absence of evidence that the employee was actually in the zone of danger. The point of the referee’s decision is that the evidence fails to show damage to the cable as a result of the violation that could create a risk to employees, either actual or potential. There is substantial evidence to support the referee’s findings, and they require that the citation * * * be dismissed.” 123 Or App at 455-56 (brackets in original; emphases added). — 575 — As illustrated above, Mowat did not address the employee-exposure element of OR-OSHA’s prima facie case. Moreover, it expressly disavowed the notion that Mad Creek Logging established a principle applicable in all enforcement proceedings under the OSEA (given, in part, the stipulated burden of proof in that case). Mowat, 237 Or App at 583. Thus, we agree with the ALJ that our case law does not provide applicable guidance regarding the requisite standard of proof for employee exposure — establishing only the principle, which neither party disputes, that OR-OSHA need not prove actual employee exposure to a hazardous condition. Given that the regulation at issue here is, as noted, a federal regulation adopted in Oregon by reference, see OAR 437-003-0001(24)(d) (9/26/07), we turn to federal authority. The Supreme Court has intimated that, in deciding cases under the OSEA, we may look to federal case law for guidance where the rule at issue has a “counterpart in the federal [Occupational Safety and Health Act (OSHA)].” OR-OSHA v. Don Whitaker Logging, Inc., 329 Or 256, 263, 985 P2d 1272 (1999). More generally, given our express recognition that the OSEA is based on the federal OSHA, Skirvin v. Accident Prevention Division, 32 Or App 109, 111 n 1, 573 P2d 747, rev den, 282 Or 385 (1978), we have repeatedly looked to federal case law for guidance in deciding cases under the OSEA. See, e.g., OR-OSHA v. CBI Services, Inc., 254 Or App 466, 474-82, 295 P3d 660, rev allowed, 353 Or 747 (2013) (relying on federal case law in assessing element of OR-OSHA’s prima facie case); Mowat, 237 Or App at 580-83 (relying on federal circuits’ approach to interpretation of workplace safety rules); OR-OSHA v. Tom O’Brien Construction Co., Inc., 148 Or App 453, 456, 461, 941 P2d 550 (1997), aff’d, 329 Or 348, 986 P2d 1171 (1999) (citing federal case law in assessing potential affirmative defense to element of OR-OSHA’s prima facie case); Skirvin, 32 Or App at 113-15 (relying on Ninth Circuit’s reasoning in assessing element of OR-OSHA’s prima facie case); cf. Marks v. McKenzie High School Fact-Finding Team, 319 Or 451, 457-60, 878 P2d 417 (1994) (federal case law relevant in interpreting Oregon law modeled in part on federal law). Further, as particularly relevant here, we stated in Mowat that “[t]he federal courts — 576 — have developed a body of law that has affected the adoption and interpretation of federal workplace safety rules” and explicitly relied on that body of law in fashioning the above-mentioned standard concerning whether a rule carries a presumption of hazardousness. 237 Or App at 580-82. The OSEA was signed into law on July 22, 1973— approximately three years after the enactment of the federal OSHA. Or Laws 1973, ch 833. As noted, in 1974, we recognized employee exposure as an element of OR-OSHA’s prima facie case. Stadeli Pump, 18 Or App at 367. During that same time period, the federal Occupational Safety and Health Review Commission (OSHRC) and federal courts were grappling with the appropriate standard of proof for employee exposure. At the behest of the Fourth Circuit, the OSHRC expressly disavowed proof of “actual” exposure as a requirement, given OSHA’s preventative purpose, and developed the “rule of access” standard cited extensively by Moore in arguing that the ALJ correctly applied a “reasonable predictability” standard. See Brennan v. Gilles & Cotting, Inc., 504 F2d 1255, 1263-64, 1267 (4th Cir 1974) (remanding case to OSHRC for determination as to whether “access” or “actual employee exposure” was appropriate standard); Secretary of Labor v. Gilles & Cotting, Inc., 3 OSH Cas (BNA) 2002, 1976 WL 5933 at *4 (OSHRC, Feb 20, 1976) (“On balance we conclude that a rule of access based on reasonable predictability is more likely to further the purposes of the Act than is a rule requiring proof of actual exposure.” (Emphasis added.)). That “rule of access” standard based on “reasonable predictability” of employee exposure has subsequently been applied with relative consistency by the OSHRC. See, e.g., Secretary of Labor v. Field & Associates, Inc., 19 OSH Cas (BNA) 1379, 1383 (2001) (“The Secretary could establish exposure by showing that employees were actually exposed to the hazard, or that it was reasonably predictable that during the course of their normal work duties, employees might be in the ‘zone of danger’ posed by the [violative] condition.”); see generally Mark A. Rothstein, Occupational Safety and Health Law § 5:13 (2013 ed). And, although phrased differently by some courts, the standard — 577 — derived from Gilles & Cotting, Inc. — which ultimately requires, simply, that the agency prove that it was reasonably predictable that one or more employees had been, were, or would be exposed to the hazard presented by the violative condition at issue — has been endorsed by a majority of the federal appellate courts that have considered the issue and remains the prevailing standard of proof with respect to employee exposure under the federal OSHA. See, e.g., R. Williams v. Occupational Safety & Health, 464 F3d 1060, 1064 (9th Cir 2006); George J. Igel & Co., Inc. v. OSHRC, 50 Fed Appx 707, 713-14 (6th Cir 2002); N & N Contractors v. Occupational Safety & Health, 255 F3d 122, 127 (4th Cir 2001); Daniel Intern. Corp. v. Donovan, 705 F2d 382, 387-88 (10th Cir 1983); Mineral Industries, Etc. v. Occupational Safety, 639 F2d 1289, 1294 (5th Cir 1981); Secretary of Labor v. Fabricated Metal Products, Inc., 18 OSH Cas (BNA) 1072, 1074-75, 1074 n 5 (1997) (citing “seminal” OSHRC decisions and “emphasiz[ing] that *** the inquiry is not simply into whether exposure is theoretically possible. Rather, the question is whether employee entry into the danger zone is reasonably predictable.”). As set forth by a leading treatise, federal law dictates that the agency must show, in the absence of proof of actual exposure, that “it is reasonably predictable that employees, by ‘operational necessity’ or otherwise (including inadvertence) in the course of their work or associated activities (e.g., going to rest rooms) will be in the zone of danger created by the cited condition.” Randy S. Rabinowitz ed., Occupational Safety and Health Law 83 (2d ed 2002) (citations omitted). The federal cases cited by OR-OSHA are inapposite at best. They either directly contradict OR-OSHA’s position as to the appropriate standard, see, e.g., N & N Contractors, 255 F3d at 127 (explicitly stating that, “[t]o establish employee exposure to a risk of injury, the [agency] must demonstrate that it is reasonably predictable either by operational necessity or otherwise (including inadvertence), that employees have been, are, or will be in the zone of danger” (citation and internal quotation marks omitted)), or, alternatively, do not address the appropriate standard but rather state that the agency was required to prove “employee access” in the course of either (1) simply setting forth the — 578 — agency’s prima facie case without discussing the accompanying standard of proof or (2) rejecting a party’s argument that proof of actual access was required. See Modern Continental v. Occupational Safety, 305 F3d 43, 49 (1st Cir 2002) (former scenario); R. Williams, 464 F3d at 1064 (latter). As noted above, “access” in the context presented here is no different from “exposure”; rather, the question before us is whether such access or exposure must have been reasonably predictable or merely possible. Federal authority points us toward the former standard.8 Despite its arguments to the contrary, OR-OSHA’s proposed standard is neither incompatible with, nor even markedly different from, the prevailing federal standard that we now adopt. Indeed, OR-OSHA acknowledged during oral arguments before this court that the appropriate standard must include “some sort of reasonableness component” and expressly suggested that we adopt a standard mandating that access to the violative condition be proved “reasonably possible.” Moreover, in addressing the concept of access, OR-OSHA acknowledged that “anything is theoretically possible” and that, under the circumstances presented by this case, an employee could have “come in with bolt cutters” to break into the fenced enclosure and access the ladder. That hypothetical scenario illustrates precisely why reasonable predictability is the prevailing standard nationwide. In short, imposition of a pure, theoretical “access” standard would effectively relieve OR-OSHA of its burden of proof as to employee exposure and create a de facto strictliability test for that element of its prima facie case. Cf. CBI Services, Inc., — 579 — 254 Or App at 476-77, 476 n 10 (noting that the “Supreme Court has explicitly stated that the OSEA ‘is a fault-based system’” and relying on both Oregon and federal authority in rejecting the “imposition of a strict-liability approach to the knowledge requirement” of OR-OSHA’s prima facie case). Finally, having concluded that the ALJ did not err with respect to the standard of proof demanded of OR-OSHA, we turn to the disposition of this case. As noted, we agree with the ALJ that OR-OSHA failed to satisfy its burden with respect to employee exposure. Put another way, a reasonable factfinder could conclude from this evidentiary record, taken as a whole, that it was not reasonably predictable that an employee would be exposed to the hazardous ladder. See ORS 183.482(8)(c) (substantial evidence standard of review). The area where the ladder was stored was not an active work zone and was expressly “off limits” to nonsupervisory personnel, the ladder was at the bottom of a pit within that area, and that fenced, inactive area was situated within a large, three-mile worksite with only approximately 18 employees present. Weaver’s thoughts regarding the possibility of the work zone becoming active “at any time” and employees’ potential use of the ladder in an “emergency situation” do not amount to proof of reasonably predictable exposure. They constitute mere speculation, and were correctly treated as such by the ALJ. As to the evidence that a few supervisory employees had keys to the purportedly locked area, as well as the footprints observed by Weaver, OR-OSHA again failed to carry its burden as to that speculative evidence. It did not adduce proof that any supervisory employee was unaware that the ladder was damaged such that he or she might have accessed the enclosure and taken the ladder out for use at the active worksite. Nor did OR-OSHA adduce proof that the footprints were not simply made while the work area was active — or even by the very worker or workers who moved the ladder to that location in the first place. In sum, OR-OSHA failed to adduce any evidence that it was reasonably predictable that a worker at the three-mile worksite would enter the inactive, expressly restricted — 580 — area to use, or transport for use, the ladder — thereby exposing him- or herself (or others) to the hazard presented by use of that ladder. OR-OSHA thus failed to carry its burden as to the element of employee exposure under the standard set forth above, and we conclude that the ALJ neither erred in imposing that standard nor in applying it. Affirmed. Pursuant to ORS 654.290(2)(b), the ALJ’s order is deemed to be a final order of the Workers’ Compensation Board for purposes of judicial review. 29 CFR section 1926.1053(b)(16) provides: “Portable ladders with structural defects, such as, but not limited to, broken or missing rungs, cleats, or steps, broken or split rails, corroded components, or other faulty or defective components, shall either be immediately marked in a manner that readily identifies them as defective, or be tagged with ‘Do Not Use’ or similar language, and shall be withdrawn from service until repaired.” The OSEA is codified at ORS 654.001 to 654.295, ORS 654.412 to 654.423, ORS 654.750 to 654.780, and ORS 654.991. For ease of reading, we use the term employee “exposure” where possible throughout this opinion; however, the same concept is encompassed by terms such as employee “endangerment,” “access,” and others quoted herein in connection with judicial and administrative proceedings. As relevant here, those words all ultimately describe the concept that employees may be subjected to the dangers posed by a hazardous condition at the workplace. We simply hold that, no matter the language used, OR-OSHA was required to prove that such an occurrence was not merely theoretically possible, but rather reasonably predictable under the circumstances in this case. Coate testified that “the enclosed area where the ladder had been placed was locked” and that only a few supervisory employees had keys to the gate. He further testified that “enclosed fences [were] required to be locked on all inactive work zones” and that, “to the best of his knowledge [,]” there was signage on the fence “with ‘Do Not Enter’ or other similar verbiage” indicating that the area was “off-limits[.]” However, Weaver testified that he did not “remember anyone unlocking a gate” as he entered the area and noted the presence of footprints on the ground despite the fact that there were no workers present at the time of the inspection. No subsequent investigation was conducted with respect to the origin of the footprints. Weaver testified “that he felt that the ladder had not been ‘removed from service,’ [in accordance with 29 CFR section 1926.1053(b)(16)] as it could have been grabbed and used by an employee in an emergency situation!.]” He further testified that he felt (although no supporting evidence was adduced) that the area could have become an active worksite “at any time[.]” The initial ALJ died soon after the first opinion and order on reconsideration affirming the citation issued. The matter was then transferred to another ALJ, who issued the final order now under review. Again, that rule provides: “Portable ladders with structural defects, such as, but not limited to, broken or missing rungs, cleats, or steps, broken or split rails, corroded components, or other faulty or defective components, shall either be immediately marked in a manner that readily identifies them as defective, or be tagged with ‘Do Not Use’ or similar language, and shall be withdrawn from service until repaired.” Likewise, our sister states have predominately adopted an approach based on reasonable predictability — most often modeled on the federal standard. See, e.g., Department of Labor v. Morel Construction Co., Inc., 359 SW3d 438, 448-49 (Ky Ct App 2011) (citing decisions under the federal OSHA and further elaborating that “Reasonable predictability requires more than a hypothetical possibility of exposure, though less than a certainty”); National College of Business & Technology, Inc. v. Davenport, 57 Va App 677, 686-87, 705 SE2d 519, 524-25 (2011) (“[T]o resolve whether credible evidence supports the circuit court’s factual finding that employees were ‘exposed’ to the asbestos, we must * * * determine whether it is possible that an employee could be exposed to asbestos, i.e., whether it is reasonably predictable that employees have been or will be in the zone of danger.”); Adkins v. Aluminum Co. of America, 110 Wash 2d 128, 146-47, 750 P2d 1257, 1268-69 (1988) (“To establish employee access, the [agency] must demonstrate a reasonable predictability that, in the course of their duties, employees will be, are, or have been in the zone of danger.” (Emphasis in original.)).