United States of America


1120 20th Street, N.W., Ninth Floor

Washington, DC 20036-3457



v. OSHRC DOCKET NO. 17-0737





Kate O’Scannlain, Solicitor of Labor

Maria S. Fisher, Regional Solicitor

Robin Ackermann, Senior Trial Attorney

Ralph Minichiello, Senior Trial Attorney

U.S. Department of Labor, Office of the Solicitor, Boston, MA

For the Complainant

Melanie L. Paul

Raymond Perez, II

Jackson Lewis P.C., Atlanta, GA

For Respondents

Before: Keith E. Bell, Administrative Law Judge


Following its receipt of a complaint about workplace violence, OSHA commenced an

inspection of UHS of Westwood Pembroke, Inc.’s (“Pembroke’s”) facility located at 199 Oak

Street in Pembroke, MA (“Pembroke Hospital”). As a result of this inspection, Pembroke received

a Citation and Notification of Penalty (“Citation”) alleging it violated the Occupational Safety and

Health Act of 1970, 29 U.S.C. §§ 651-678 (the “Act”). The Citation alleges a violation of 29



U.S.C. § 654(a)(1), the provision commonly referred to as the general duty clause, for exposing

employees to the hazard of workplace violence in the form of patient on staff aggression.


Pembroke timely contested the Citation bringing the matter before the Commission. The

Secretary of Labor (“Secretary”) filed his initial Complaint on April 27, 2017. He moved to amend

his Complaint to add UHS of Delaware, Inc. (“UHS DE”) as a Respondent and to change the

classification of the Citation from “serious” to “repeat.” This request was granted in an October

30, 2017 Order, after which Pembroke and UHS DE together became the Respondents for this


Subsequently, the Secretary moved again to amend abatement in the Citation and

Complaint. This request was uncontested and was approved in a December 13, 2017 Order. A

hearing was held on July 17-20, 2018 and July 24-25, 2018.

For the reasons discussed, the Citation is affirmed as serious and a $12,675 penalty is


I. Jurisdiction


Pembroke filed a timely Notice of Contest bringing this matter before the Commission.

(Stip. 4-6.) Pembroke and its corporate parent, UHS DE, are both employers affecting commerce

1 The Joint Exhibit of the parties’ Stipulations indicates that: “Respondent timely filed its Notice of Contest on May

17, 2018.” (Ex. J-1, Sec’y Br. at 1; Tr. 194.) The undersigned accepts the acknowledgment that the Notice of Contest

was timely, but rejects the date listed in the Stipulation. The record reflects that the Secretary issued the Citation on

April 7, 2017, and Pembroke filed a Notice of Contest on April 12, 2017. Pembroke then filed an Answer to the

Secretary’s Complaint on May 17, 2017. (Resp’t Br. at 5.) The record does not reflect a document filed on May 17,


2 Stipulation 4 is: “The Occupational Safety and Health Review Commission has jurisdiction in this proceeding

pursuant to § 10(c) of the [Act].” Stipulation 5 is: “The Citation and Notification of Penalty underlying this proceeding

was issued on April 7, 2017.”





within the meaning of 29 U.S.C. § 659(c) and both are employers under the Act. (Stip. 1, 10-13.)

Based upon the record, including the parties’ admission to jurisdiction, the undersigned concludes

the Commission has jurisdiction over the parties and the subject matter of this case. (Stip. 1, 4-6.)

II. Background

A. Corporate Structure

Pembroke operated three facilities at the time of the inspection: Pembroke Hospital,

Westwood Lodge, and Lowell Treatment Center. (Tr. 697; Ex. R-38 at 2.) While each facility

was located at a separate address, collectively the sites shared a license and were a single legal

business entity. (Tr. 723-24.) Pembroke is a wholly-owned subsidiary of UHS DE. (Stip. 12.)

UHS DE, in turn, is a wholly-owned subsidiary of Universal Health Services, Inc. (“UHS”). (Stip.

13.) Thus, UHS owns UHS DE and indirectly owns Pembroke. (Stip. 13; Tr. 752-53.)

UHS DE is described as a “management company” for Pembroke. (Tr. 740; Ex. C-27.) At


the time of the inspection, Pembroke’s CEO and CFO were both employees of UHS DE. (Stip.

10-11; Tr. 598, 693, 757.) UHS DE employees provided oversight for the work of Pembroke’s

CEO and CFO. (Tr. 599, 696, 757-58, 1325, 1327-28.) In addition, Gina Gilmore, a Loss Control

Manager at UHS DE, handled Pembroke’s budget for workers’ compensation claims and visited

the facility every month. (Tr. 549-551.) During these visits, she would meet with Pembroke’s

CEO and attend the facility’s Aggression Reduction Team meetings. (Tr. 553-54.) The Director

3 Stipulation 1 states: “Respondents are employers engaged in a business affecting commerce within the meaning of

Section 3(5) of the Occupational Safety and Health Act of 1970, 29 U.S.C. § 6252(5).”

4 As explained by Stipulation 10, “Raymond Robinson, Gary Gilberti, Diane Airosus, and Gina Gilmore were

employed by UHS of Delaware, Inc. during the time of the inspection.” Thomas Hickey was the chief executive

officer (“CEO”) from the start of the inspection until January 21, 2017. (Tr. 692-93; Stip. 11 (“Thomas Hickey was

employed by UHS Delaware, Inc. during a portion of the time of the inspection.”)



of Nursing, Claire Kent, and the Risk Manager, Stacey Coke Burns, were Pembroke employees

but also reported to individuals employed by UHS DE. (Tr. 369-70, 457, 497-98, 502.)

B. Respondents’ Request for Judicial Notice

In its brief, Respondents assert that the Secretary has pursued general duty clause violations

at “several other [UHS] managed hospitals.” (Resp’t Br. at 3.) Respondents fail to discuss whether

these other citations were issued to UHS DE or entities affiliated with Pembroke. However, in a

footnote elsewhere in the brief, Respondents ask the undersigned to take judicial notice of “other

Citations that have been issue to behavioral hospitals.” (Resp’t Br. at 33.) It requests judicial

notice of three specific cases:

Secretary of Labor v. BHC Northwest Psychiatric Hospital, LLC d/b/a Brooke Glen

Behavioral Hospital, (OSHRC Docket 17-0063), Secretary of Labor v. UHS of

Westwood Pembroke, Inc., d/b/a Lowell Treatment Center (OSHRC Docket 17-

1302 and 17-1304) and UHS of Delaware, Inc. (“UHS DE”) and Premier

Behavioral Health Solutions of Florida, Inc., d/b/a Suncoast Behavioral Health

Center (OSHRC Docket 18-0657).

(Resp’t Br. at 33.) The first matter listed concerns BHC Northwest Psychiatric Hospital (“BHC”).

Respondents do not explain how or if that entity is affiliated with them. Id. The undersigned is

aware that a matter involving BHC has become a final order of the Commission. However, BHC

is appealing the matter and the matter is currently pending before the D.C. Circuit. BHC Nw.

Psychiatric Hosp. LLC d/b/a Brooke Glen Behavioral Hosp., 27 BNA OSHC 1862 (No. 17-0063,

2019) (ALJ), argued, 19-1087 (D.C. Cir. Jan. 9, 2020).

Federal Rule of Evidence 201(b) permits judicial notice of a “fact” that is not subject to

reasonable dispute because it: (1) is generally known within the court’s territorial jurisdiction; or

(2) can be accurately and readily determined from sources whose accuracy cannot reasonably be

questioned. Even though Respondents do not discuss their affiliation with BHC, the undersigned

can and has reviewed the ALJ’s decision, BHC Nw. Psychiatric Hosp. LLC d/b/a Brooke Glen




Behavioral Hosp., 27 BNA OSHC 1862 (No. 17-0063, 2019) (ALJ). However, taking judicial

notice of a non-binding decision that is still on appeal is not appropriate.

As for the other two matters highlighted by Respondents, both are active disputes without

final determinations. As such, there are no established facts for which judicial notice is

appropriate. At the hearing, Respondents did not attempt to offer any evidence regarding the

relevance of these matters or how Premier Behavioral Health Solutions of Florida, Inc. is affiliated

with them. As a result, the Secretary was deprived of an opportunity to respond to these claims.

Further, Respondents assert that UHS DE had only a minimal role at Pembroke Hospital and asks

for the entity to be dismissed from this pending matter. (Resp’t Br. at 40.) Asking the undersigned

to now undertake a search of other citations issued to behavioral health centers is not the purpose

of judicial notice and the late request is denied.

The undersigned notes that the Secretary offered evidence of another citation issued against

Pembroke, the same employer cited in the present matter. That citation, which became a final

order on May 27, 2016, is a part of the record before me and has been considered in reaching this

decision. (Exs. C-14 thru C-17.)

C. Nature of the Workplace

Pembroke Hospital is an inpatient psychiatric hospital with six patient care units, serving


up to 120 patients at a time. (Stip. 2-3; Tr. 290, 373; Ex. R-38 at 1, 8.) Four units focus on adults,

one is for adolescents, and the sixth is for geriatric patients. (Tr. 373.) Nurses and Mental Health

Associates (“MHAs”) work directly with patients in the units. (Tr. 669, 821-824.) Units always

have a registered nurse. (Tr. 380.)

5 Stipulation 2 states: “Pembroke Hospital is an inpatient psychiatric hospital.” Stipulation 3 states: “Pembroke

Hospital has six patient care units with a total of 120 beds.”




The MHAs at Pembroke Hospital have many responsibilities, including providing security

and fulfilling housekeeping responsibilities. (Tr. 50-52, 96.) The nurse on the unit could not see

patients from his or her workstation. (Tr. 58.) So, the MHAs functioned as the “eyes and ears of

the nursing staff” and were responsible for monitoring patient behavior, including for signs of

aggression or agitation. (Tr. 50.) MHAs searched patient belongings to identify prohibited items

that could hurt patients or staff. (Tr. 79, 81, 86.) MHAs conducted these searches alone in small

rooms on the unit. (Tr. 79, 81.) MHAs had to make sure certain doors remained locked. They

would open and close bathrooms when requested by patients, which might occur fifty times a shift.

(Tr. 55, 827.) MHAs also accompanied patients off units for certain meetings, fresh air breaks,


and meals. (Tr. 55-56.) If a patient was unable to leave the unit for meals, an MHA would leave

the unit instead and get the patient’s meal. (Tr. 52.) MHAs also left the units for laundry and

other housekeeping tasks. (Tr. 50-54, 96.) Consequently, while more than one MHA was assigned

to each unit, the staff was often divided between those on the unit and those off the unit. (Tr. 822,


D. Prior OSHA Inspections of Pembroke

About a year before the inspection leading to the Citation before the undersigned, OSHA

inspected Pembroke Hospital. (Ex. R-38.) That inspection led to the issuance of a letter on

October 2, 2015, which indicates that MHAs were exposed to workplace violence hazards but that

“it is not considered appropriate at this time to invoke the Section 5(a)(1), the general duty clause


of the [Act].” (Ex. R-37 at 2 (emphasis added).) The letter does not state that OSHA considered

6 Mealtimes themselves were a “very high time for assault” at Pembroke. (Tr. 960; Exs. C-55 at 2, C-58 at 2.)

7 Respondents neglect to note the critical phrase “at this time” when discussing this document. (Resp’t Br. at 21.)




workplace violence hazards to be outside of the scope of the general duty clause. (Resp’t Br. at


21.) Rather, OSHA exercised its discretion to issue a letter at the end of the investigation. The

letter goes on to outline several methods of feasible abatement for workplace violence hazards


Pembroke could adopt, but does not demand any specific action. (Ex. R-37.)

After the letter was issued, OSHA commenced an inspection of Pembroke’s facility in

Lowell, MA (the Lowell Treatment Center). After the inspection, the Secretary issued Pembroke

a citation for a violation of the general duty clause because it exposed employees to workplace

violence in the form of, among other things, patients assaulting staff through punches, kicks, bites,

scratches, and hair pulling. (Exs. C-14, C-15 at 2, R-38 at 2.) While Pembroke initially contested

the citation, it subsequently agreed to accept it as issued and agreed to take several abatement

actions to address the workplace violence hazards at the Lowell Treatment Center. (Ex. C-15.)

E. Current Inspection

In October 2016, OSHA received another complaint about worker safety at Pembroke

Hospital. (Tr. 277; Ex. C-8.) The complaint cites a worker injury following an instance of patient

aggression. (Ex. C-8.) This complaint led OSHA to commence another inspection of Pembroke

Hospital. (Tr. 277.) OSHA Compliance Safety and Health Officer Salvatore Insogna (“CO”) led

the investigation. Id. He first visited the site on October 11, 2016. (Tr. 342.) In addition to the

initial complaint triggering the investigation, OSHA received a second complaint on March 27,

8 There is no evidence that the conditions, particularly those related to the cited hazard, were the same during the two

inspections. For example, the number of staff injuries increased in the year after this letter was sent. (Exs. C-60; C-

68, C-69; R-38 at 3.)

9 After the 2015 inspection, but before the one which led to the present matter, OSHA conducted another inspection

of Pembroke Hospital. That investigation commenced after OSHA received two employee complaints alleging several

hazards related to workplace violence. (Ex. R-38 at 1-2.) Pembroke Hospital participated in the investigation,

including by answering questions and providing responses to various document requests. Id. at 3-5. Ultimately,

OSHA closed this inspection on August 4, 2016, without issuing any citations. Id. at 5.



2017. (Ex. C-11.) Like the prior complaint, the second one cites employee injuries from

workplace violence in the form of assaults by patients against staff. Id. The CO’s investigation

included a review of documents related to employee injuries and illnesses, and interviews with ten

or more non-management employees. (Tr. 290-92, 297; Ex. C-68.) When the investigation

concluded, the Secretary issued the Citation that is the subject of this litigation.

III. Legal Standard

A. Due Process Claims

Respondents claim that the Citation’s issuance violated due process. (Resp’t Br. at 19.)

The Commission ensures that OSHA provides due process by upholding violations of the general

duty clause only where the employer or its industry recognized the hazard before the citation was

issued. The inquiry is whether the employer or its industry could reasonably foresee an accident

capable of causing death or serious injury. Armstrong Cork Co., 8 BNA OSHC 1070, 1074 (No.

76-2777, 1980), aff'd, 636 F.2d 1207 (3d Cir. 1980) (unpublished). Knowledge of the hazard

provides adequate notice to satisfy the requirement of due process. See e.g., Cape & Vineyard

Div. of New Bedford Gas & Edison Light v. OSHRC, 512 F.2d 1148 (1st Cir. 1975) (finding that

actual knowledge of the hazard provides fair notice); Bethlehem Steel Corp. v. OSHRC, 607 F.2d

871, 875 (3d Cir. 1979) (finding that fair notice is addressed by the requirement that the hazard is

recognized); Babcock & Wilcox Co. v. OSHRC, 622 F.2d 1160, 1164 (3d Cir. 1980) (concluding

that either the employer or its industry must be aware of the hazard).

Respondents raise no challenges to how the Secretary defined the hazard here or

Pembroke’s recognition of both the hazard and its presence at Pembroke Hospital. Indeed, they

stipulated that the “hazard of workplace violence, specifically defined in this case as violence




and/or assault by patients against staff, was recognized by [Pembroke].” (Stip. 8.) And they

stipulated that the “hazard of workplace violence, specifically defined in this case as violence

and/or assault by patients against staff is recognized in the industry.” (Stip. 9.)

Even if Pembroke had not stipulated to the definition and recognition of the hazard, the

record still provides ample support for such findings. Patient attacks on employees routinely

occurred at Pembroke Hospital. (Exs. C-8, C-11, C-68.) One employee indicated he was hit or

kicked approximately fifty different times by various patients in his four years working for

Pembroke. (Tr. 56-57.) See Gen. Dynamics Land Sys. Div., Inc., 15 BNA OSHC 1275, 1285 (No.

83-1293, 1991) (accidents put the employer on notice of the hazard), aff’d, 985 F.2d 560 (6th Cir.

1993) (unpublished). Pembroke was aware of these events through accident reports employees

could choose to complete after incidents and through the direct knowledge of supervisors, some

of whom were exposed to the hazard themselves. (Tr. 794, 802; Exs. C-55, C-56.)

Moreover, Pembroke had specific and direct knowledge of the Secretary’s view that the

general duty clause applied to the hazard of workplace violence. On May 19, 2015, Pembroke was

cited for violating the general duty clause because its employees “were exposed to acts of

workplace violence.” (Ex. C-14.) This citation was issued in 2015 and became a final order on

May 27, 2016, months before the commencement of the inspection and issuance of the present

citation. (Exs. C-15 thru C-17.) So, Pembroke accepted responsibility for a violation of the general

duty clause based upon the presence of the hazard of workplace violence before the investigation


leading to the instant Citation commenced. This establishes direct notice of the applicability of

10 This citation related to the Lowell Treatment Center. After this Pembroke’s acceptance of this citation, another

inspection of Pembroke Hospital concluded without the issuance of any citations. (Ex. R-38 at 5.) However, OSHA

never indicated there were no hazards at Pembroke Hospital or that Pembroke did not have a duty to provide its

employees with a workplace “free from recognized hazards that are causing or are likely to cause death or serious

physical harm.” 29 U.S.C. § 654(a)(1).



the general duty clause to workplace violence hazards at Pembroke’s facilities. See Corbesco Inc.

v. Sec'y of Labor, 926 F.2d 422, 428 (5th Cir. 1991) (holding notice is provided through

Commission decisions); Armstrong, 8 BNA OSHC at 1073 (finding that the employer itself

recognized the hazard because it took an abatement action).

Besides Pembroke’s direct knowledge, the Secretary provided other evidence that the

industry was on notice of the hazard. First, there is no dispute that experts familiar with the

industry would take the hazard of workplace violence into account in prescribing a safety program

for a behavioral health hospital. See Nat’l Realty & Constr. Co. v. OSHRC, 489 F.2d 1257, 1266

(D.C. Cir. 1973). Experts for both parties recognized that any risk assessment of a facility like

Pembroke Hospital would include assessing workplace violence and the hazard would be relevant

to any safety program. (Tr. 903, 915-17, 1455.) Second, OSHA published guidance documents

to inform the industry of the hazard and ways to address it. In 2015, it published Preventing

Workplace Violence: A Road Map for Healthcare Facilities (“Road Map”). (Ex. C-102.) The

Road Map describes the problem of workplace violence in healthcare facilities and provides

information on how to identify hazards and respond to risks identified. Id. The Road Map is

related to an earlier OSHA publication, Guidelines for Prevention of Violence in Healthcare

(“Workplace Violence Guidelines”). (Exs. C-97, C-102 at 4.) Neither the Workplace Violence

Guidelines nor the Road Map constitute specific standards. At the same time, the documents do

not suggest that there is some type of immunity from citation for failing to free a workplace of

recognized hazards, as the general duty clause requires. Id. These OSHA publications bolster the

uncontested point that Pembroke and its industry are aware of the hazard.

Perhaps in recognition of its actual awareness of both the hazard and its presence at

Pembroke Hospital, Respondents suggest another due process test. (Resp’t Br. at 20.) They argue



the Citation failed to set forth the “specific actions” to be taken so that Respondents can avoid

future citations. Id. Respondents argue they must have notice of the Secretary’s views regarding

appropriate abatement before a citation can be issued. Id. at 20-21, 23.

Respondents attempt to conflate the due process fair notice requirement, which is driven

by Constitutional constraints, with the abatement requirement, which is a statutory interpretation

created by binding precedent. The Constitution precludes depriving “any person of life, liberty,

or property, without due process of law.” U.S. Const. amend V. The Commission sets forth

requirements to ensure that proceedings before the Commission provide due process. For example,

the citation must sufficiently define the hazard and the abatement so that the employer can

reasonably defend itself. The Secretary must make the employer aware of what he alleges are

feasible means of abatement prior to the hearing, but the Commission has not required such

information before the issuance of a citation. See Erickson Air-Crane Inc., No. 07-0645, 2012 WL

762001 (O.S.H.R.C, Mar. 2, 2012). In Erickson, the Secretary proposed various abatement

methods but the ALJ concluded that none of the proposed methods were feasible. Id. at *2-3.

Nonetheless, the ALJ went on to determine that the hazard could be feasibly abated through some

other action. Id. The parties had neither tried nor consented to try whether that action was feasible.

Id. The Commission considered this a due process violation because the pleadings did not put the

employer on notice that the alternative abatement method was at issue. Id. Notice must be given

before the record closes as to what is being considered feasible means of abatement. Id. The

Commission did not hold that the employer must know before the citation’s issuance what





constitutes feasible means of abatement. Id.

Respondents cite cases where the parties disputed the presence of a hazard, the employer’s

recognition of it, or are otherwise not relevant to the due process issues before me. (Resp’t Br. at

20-21.) Respondents rely on Asamera Oil (U.S.), 9 BNA OSHC 1426 (No. 1426, 1980)

(consolidated), a non-binding ALJ opinion addressing a specific standard, not the general duty

clause, and a non-binding Ninth Circuit decision, Donovan v. Royal Logging Co, 645 F.2d 822


(9th Cir. 1981). See Asamera, 9 BNA OSHC at 1427 (affirming the ALJ’s decision but according

it only “the precedential value of an unreviewed judge’s decision”); Integra Health Mgmt., Inc.,

27 BNA OSHC 1838, n.15 (No. 13-1124, 2019) (noting that Royal Logging is “not relevant

precedent” for cases, like the present matter, that cannot be appealed to the Ninth Circuit).

Respondents’ reliance on Davey Tree Expert Co., 11 BNA OSHC 1898 (No. 77-2350, 1984) is

also misplaced. (Resp’t Br. at 24.) Unlike that case, in this matter, Respondents explicitly chose


not to contest the Secretary’s definition of the hazard or Pembroke’s recognition of it. (Stips. 7-

9.) Here, Pembroke recognized the hazard of patient on staff violence and knew this hazard was

present in its facility. (Stips. 8-9.)

The recognition requirement relates to “knowledge of the hazard, not recognition of the

11 A.H. Sturgill Roofing, Inc., 27 BNA OSHC 1809 (No. 13-0224, 2019) is also instructive. There, the citation was

somewhat ambiguous about the scope of the proposed abatement. 27 BNA OSHC 1818-19. However, before the

record closed, the Secretary subsequently clarified it. Id. The Commission did not require the Secretary to show that

the employer understood the proposed methods of abatement before the citation was issued. Id. The Commission

was satisfied that during the hearing the Secretary made clear that the measures in the citation were alternative means

of abatement. Id.

12 This case can be appealed to the First and D.C. Circuits. See 29 U.S.C. § 660(a) (“Any person adversely affected

or aggrieved by an order of the Commission ... may obtain ... review ... in any United States court of appeals for the

circuit in which the violation is alleged to have occurred or where the employer has its principal office, or in the Court

of Appeals for the District of Columbia Circuit ....”); Kerns Bros. Tree Serv., 18 BNA OSHC 2064, 2067 (No. 96-

1719, 2000) (Commission generally applies law of circuit where it is probable case will be appealed).

13 Respondents also quote the discussion in Missouri Basin Well Service, 26 BNA OSHC 2314 (No. 13-1817, 2018)

about how the hazard must be defined and attempt to apply the statements more broadly. (Resp’t Br. at 20, 28.) In

this matter, there is no dispute as to how the hazard is defined or Pembroke’s recognition of it.



means of abatement.” Kansas City Power & Light Co., 10 BNA OSHC 1417, 1422 (No. 76-5255,

1982), citing Gen. Dynamics Corp., Quincy Shipbuilding Div. v. OSHRC, 599 F.2d 453 (1st Cir.

1979). See also Brock v. L.R. Willson & Sons, Inc., 773 F.2d 1377 (D.C. Cir. 1985) (noting in the

context of a violation of a specific standard that “The constitution does not require that employers

be actually aware that the regulation is applicable to their conduct”). Fair notice challenges to a

general duty clause citation fail when the abatement measures are available to and readily

knowable by the industry. Integra, 27 BNA OSHC 1851 at n.15. Like the abatement measures in

Integra, the abatement measures proposed for Pembroke track the Workplace Violence Guidelines

and Road Map as well as those specified in a letter issued to Pembroke after a previous inspection.

Id. The Secretary set forth specific abatement measures and Respondents were informed of them

in advance of the hearing. The Secretary’s Amended Complaint sets forth his proposed abatement

and there is no allegation that Respondents were not aware of his proposal well before the hearing.

The feasibility of the specific measures discussed herein was tried with the full consent of the

parties. (Jt. Pre-Hr’g Stmt. at 16.) See Beverly Enters., 19 BNA OSHC 1161, 1168-69 (No. 91-

144, 2000) (consolidated) (excusing a lack of precision in the complaint because, “at the time of

the hearing” the Secretary sufficiently specified the alleged hazard).

Respondents then take another tack and argue that because they were not cited after prior

inspections, they lacked fair notice they would be cited in the future. (Resp’t Br. at 21-22.)

Respondents make this argument without citation to precedent. Id. An employer cannot deny the

existence of a hazard or its knowledge of it “by relying on the Secretary's earlier failure to cite the

condition.” Seibel Modern Mfg. & Welding Corp., 15 BNA OSHC 1218, 1244 (No. 88-821, 1991).

Employers must comply with the Act regardless of whether they have been previously informed

that a violation exists. Id.




Respondents claim that the Citation, including the Secretary’s proposed abatement,


violates the Constitution is rejected.

B. Test to Establish a General Duty Clause Violation

The general duty clause requires every employer to provide its employees with a workplace

“free from recognized hazards that are causing or are likely to cause death or serious physical


harm.” 29 U.S.C. § 654(a)(1). As interpreted by the Commission, to establish a violation of this

clause, the Secretary must show: (1) there was an activity or condition in the employer’s workplace

that constituted a hazard to employees; (2) either the cited employer or its industry recognized that

the condition or activity was hazardous; (3) the hazard was causing or was likely to cause death or

serious physical harm; and (4) there were feasible means to eliminate the hazard or materially

reduce it. Waldon Health Care Ctr., 16 BNA OSHC 1052, 1058 (No. 89-3097, 1993). The

evidence must also show the employer knew, or with the exercise of reasonable diligence, could

have known of the hazardous condition. Otis Elevator Co., 21 BNA OSHC 2204, 2207 (No. 03-

1344, 2007).

Despite agreeing that Waldon sets forth the test for finding a violation of the general duty

clause, Respondents, citing National Realty, argue that the Secretary must meet various other

14 Separate from these Constitutional considerations is the requirement for the Secretary to establish feasible means

of abatement. The requirement is not set forth in the text of the general duty clause itself. 29 U.S.C. § 654(a)(1). See

e.g., SeaWorld of Fla. v. Perez, 748 F.3d 1202, 1207 (D.C. Cir. 2014); A.C. Castle Constr. Co., Inc. v. Acosta, 882

F.3d 34, 44 (1st Cir. 2018) (limiting the scope of the fair notice requirement); Mark A. Rothstein, Occupational Safety

and Health Law §§ 6:1, 6:9 (2019 ed.). It arose as a means to interpret the Act so as to avoid the application of strict

liability. See 29 U.S.C. § 651(b) (requiring employers “to assure so far as possible … safe and healthful working

conditions”). Respondents other contentions that the Secretary failed to meet the abatement prong of the test for

establishing a violation of the genera duty clause as developed by the Commission and relevant Circuits are discussed


15 The parties stipulated that this is an applicable principle of law. (Am. Pre-Hr’g Stmt. at 16; Resp’t Br. at 6.)





requirements. (Am. Pre-Hr’g Stmt. at 16; Resp’t Br. at 6, 28-30.) In National Realty, the D.C.

Circuit distinguishes between an unrecognized hazard and one, like the situation at hand, that is

recognized. 489 F.2d at 1266. Hazard recognition does not automatically trigger liability. Id.

The Secretary still must demonstrate “at the hearing” what the employer should have done. Id. at

1266-68. When discussing whether feasible abatement can be found after the hearing, as opposed

to beforehand, the D.C. Circuit notes the concept of preventability. Id. If the conduct resulting in

the hazard is something “conscientious experts, familiar with the industry” would take it into

account when “prescribing a safety program,” then the Secretary can establish a violation. Id. at

1266. In the present matter, the testifying experts agree that workplace violence is something

facilities like Pembroke Hospital must consider when prescribing a safety program. While the

hazard is not entirely preventable, there are steps employers can take to reduce the likelihood of

such incidents and the severity of the ones which occur. See Integra, 27 BNA OSHC 1841 at n.3

(discussing National Reality in the context of a general duty clause violation based on workplace

violence hazards); Armstrong, 8 BNA OSHC at 1074 (noting that the inquiry into foreseeability is

limited to whether the hazard will result in serious injuries or death).

1. Presence of a Hazard

While National Realty analyzes what it means for a workplace to be free of a recognized

hazard, Waldon and more recent cases apply a four-part test to assess general duty clause

violations. Under this framework, the first element the Secretary must prove is that “a condition

or activity in the workplace presented a hazard.” 16 BNA OSHC at 1058. The parties stipulated:

Employees at the worksite were exposed to the hazard of workplace violence, specifically defined

16 National Realty predates Waldon and the development of the Commission’s four-part test for sustaining violations

of the general duty clause. See Integra, 27 BNA OSHC 1841 at n.3.




in this case as violence and/or assault by patients against staff, during the six months prior to the

issuance of the citation (October 11, 2016 to April 7, 2017). (Stip. 7; Tr. 203, 923.) At the hearing,

Respondents’ counsel also made plain that they conceded the presence of a hazard to which

employees were exposed. (Tr. 194-95.) The Secretary established that employees at Pembroke

Hospital were exposed to the hazard of violence and/or assault by patients against staff.

2. Recognition and Knowledge of the Hazard

The second element looks at whether the employer or its industry recognized the condition

as a hazard. Again, there is no dispute that Pembroke and the behavioral health industry both


recognized the hazard of workplace violence in the context of patient on staff violence. (Stips.

8-9; Resp’t Br. at 1.) As for knowledge of the hazard, establishing this does not require a showing

that the employer was actually aware that it was in violation of the Act. See e.g., Peterson Bros.

Steel Erection Co., 16 BNA OSHC 1196, 1199 (No. 90-2304, 1993), aff’d, 26 F.3d 573 (5th Cir.

1994). Knowledge is established if the record shows the employer knew or should have known of

the conditions constituting a violation. Peacock Eng’g Inc., 26 BNA OSHC 1588, 1592 (No. 11-

2780, 2017).

Here, Pembroke had actual knowledge of the hazard’s on-going presence at its workplace.

It also knew its employees were exposed to both actual and potential incidents of workplace

violence. Pembroke knew patients had the potential to become aggressive towards staff. (Resp’t

Br. at 1.) It trained employees about the fact that patients may be assaultive or even homicidal.

(Exs. C-30, R-32.) And it knew that injuries from violence had occurred. Written employee

accident reports detailing injuries from violence were provided to management. (Ex. C-3.)

17 In addition to the stipulation, Respondents’ counsel also acknowledged at the hearing that both Pembroke Hospital

and the industry as a whole recognized the hazard. (Tr. 924.)




Employee injuries were discussed by management at daily meetings. (Tr. 1441-42.) An MHA

explained that she repeatedly communicated safety concerns to multiple supervisors, including


discussing a violent incident directly with the Director of Nursing, Claire Kent. (Tr. 853.) Loss

control reports provided to Pembroke’s CEO routinely documented many employee injuries and

noted the link between the injuries and patient aggression. (Exs. C-55 thru C-60.) Supervisors

were also made aware of occurrences of the hazard through verbal reports and by witnessing

assaultive behavior directly. (Tr. 559, 564, 1263, 1268, 1302-3; Exs. C-3, C-55 thru C-60, R-38,

R-39, R-40.)

The Secretary established recognition and knowledge of the hazard of workplace violence

present at Pembroke Hospital.

3. Serious Physical Harm

The third element of the test is also met: the cited hazard was causing or was likely to cause

death or serious physical harm. A hazard is likely to cause death or serious physical harm if the

likely consequence of employee exposure to the hazard would be serious physical harm.

Morrison-Knudsen Co./Yonkers Contracting Co., 16 BNA OSHC 1105, 1122 (No. 88-572, 1993).


The Secretary offered expert testimony on this point. (Tr. 919; Exs. C-66, C-99.) Dr. Robert

Welch opined that at the time of the inspection the hazard of workplace violence was causing

serious injury at Pembroke Hospital. (Tr. 923-24, 926.) Before reaching his conclusion, Dr. Welch

18 This MHA (KS) was employed by Pembroke for over two years. (Tr. 792.) She left her position approximately

two weeks before the hearing to work in another type of psychiatric care facility. (Tr. 792-93.) Her demeanor lent

her testimony credibility. Her testimony about safety concerns related to the hazard of workplace violence at

Pembroke was supported by the documentary evidence.

19 Respondents’ expert, Monica Cooke, was offered only as an expert who evaluated the suggested feasible abatement

methods offered by the Secretary. (Tr. 1463.) She was not offered (or accepted) as an expert who assessed whether

the hazard was causing serious physical harm.




reviewed the employee accident reports and records of OSHA recordable injuries for the two years

before the Citation. (Tr. 926.) He identified thirty-two injuries attributable to patient violence

over that time. Id. The incidents included episodes of staff being bit, concussed, head butted, spit

at, and suffering hand injuries. Id.

Other record evidence supports Dr. Welch’s conclusions. Pembroke provided training to

new employees about risk management. (Tr. 519; Exs. C-3, C-30.) The training explains that


Pembroke’s patients may have the potential for assaultive or even homicidal behavior. (Ex. C-

30 at 19, 22.) Pembroke employees reported suffering physical harm as a result of patient on staff

violence. (Exs. C-3, C-56, C-57, C-58; Tr. 648, 794, 810-11.) The CO reviewed Pembroke

Hospital’s records of OSHA recordable injuries and found that a “very substantial majority were

workplace violence related.” (Tr. 292.) Injuries to employees constitute at least prima facie

evidence that the hazard was likely to cause death or serious injury. See e.g., Usery v. Marquette

Cement Mfg. Co., 568 F.2d 902, 910 (2d Cir. 1977).

Pembroke’s records detail five different employee injuries from violence in one month


alone of the inspection period. (Exs. C-3, C-56.) KS worked as an MHA on the adolescent unit

on both the day and evening shifts. (Tr. 793-94.) She described multiple assaults over the nearly

two-year period when she worked for Pembroke Hospital. (Tr. 792-94.) During the inspection

period, on November 12, 2016, KS was injured by a patient. (Exs. C-3, C-74; Tr. 794, 803, 857.)

The incident occurred on the adolescent unit on a Saturday during the evening shift. (Tr. 794,

853.) KS and another staff member tried to calm an agitated patient verbally, consistent with their

20 Respondents risk management training does not appear to label any particular behavior as “workplace violence,”

but discusses “issues inherent” in the patient population and indicates that patients are to be monitored for increases

in “assaultive/homicidal potential ….” (Ex. C-30 at 8-9.)

21 Five employee injury reports all dated in November 2016 refer to injuries from patient aggression. (Exs. C-3, C-60

at 2-3.) In each of these circumstances, the employee required medical attention. (Tr. 563; Ex. C-60 at 3.)





training. (Tr. 795-96.) Their efforts were not successful, and the patient began throwing

shampoo bottles and other objects. (Tr. 796.) The patient then proceeded to physically assault

KS. Id. KS attempted to restrain the patient, but before she could, the patient was able to grab

KS’s hair. Id. The patient pulled with such force as to lift KS’s head and remove chunks of hair.

(Tr. 796, 1266.) Another employee, JQ, was also punched in the head as he tried to assist with the

restraint. (Tr. 1264-65.)

Following the attack, the patient was put on 1:1 observation. (Tr. 801.) 1:1 observation is

a staffing arrangement to assign a staff member to stay with a single patient. (Tr. 77-78.) This is

typically done as a result of violent or self-injurious behavior and must be approved by a doctor.

Id. However, not long after this arrangement was put into place, the order for 1:1 observation was

removed. KS was concerned by this, particularly because of the severity of the incident and the


fact that it was not the first time the patient exhibited violent behavior. (Tr. 797-98.) She

informed the nursing supervisor, Kristin Devane, of her concerns. (Tr. 814-15; Ex. C-3.) The

supervisor did not dispute the concern for violent behavior but indicated there was not enough staff

at the time to have the patient on 1:1 observation. (Tr. 815.)

KS completed her shift and returned to work the next day, which was a Sunday. (Tr. 803-

4.) Near the end of her shift, a physical conflict developed between two groups of patients. The

same patient involved with the attack against KS on the prior day acted as the “ringleader” for one

of the factions. (Tr. 804.) When the staff attempted to intervene to protect the patients, the patients

began to assault staff. (Tr. 804-5.) Patients kicked staff in their heads, their backs, and their legs.

22 AS, a former MHA at Pembroke, explained that employees were trained to verbally de-escalate patients by using

conversation to engage a patient and then attempt to calm them down. (Tr. 89.)

23 For example, about one week before her assault, KS reported to her supervisor that the same patient tried to hit her.

(Tr. 798.) The patient was not placed on 1:1 observation after that incident. (Tr. 799.)




Staff members were punched, scratched, and bitten during the incident. (Tr. 805.) The incident

lasted for about two hours before it was appropriately contained. (Tr. 807.)

Besides the assaults on November 12 and 13, 2016, KS described another incident which

occurred on November 3, 2016. The incident involved a dispute over a music player. Many

patients rely on music as a method to calm themselves. Two patients began fighting over an MP3

music player. (Tr. 809.) KS attempted to intervene to restrain the primary aggressor. Id. KS

sustained a tear to the triangular fibrocartilage complex in her wrist. (Tr. 809, 811.) The injury

required emergency treatment and eventual surgery. (Tr. 809-11.) KS still has ongoing numbness

in two fingers. (Tr. 811.) Two additional employees assisted with the restraint and were also

injured. (Tr. 810.) One of these employees was kicked with enough strength to knock her into the

opposite wall. Id. The assault caused her to suffer cranial nerve damage. (Tr. 810-811.)

KS’s experiences at Pembroke Hospital were not unique. Other employees also described

injuries from workplace violence. AS, an MHA, indicated he was hit or kicked approximately 50


times during his four-year employment with Pembroke. (Tr. 47, 56-57.) He described one

instance that occurred during the inspection period. He was about to commence the required search

of a newly arrived patient’s belongings. (Tr. 86.) The patient became angry and pushed AS into

a wall. Id. Another nurse recalled multiple assaults, including an incident where a patient nearly


tore an employee’s ear off. (Tr. 217-18; Ex. C-60.)

The assaults against staff resulted in multiple employee injuries. (Tr. 805; Exs. C-3, C-68,

24 AS was employed by Pembroke during the inspection and left after the Citation was issued. (Tr. 47.)

25 In addition, a former nurse at the facility, TL, indicated that in her three years at the facility serious staff injuries

from assaults regularly occurred. (Tr. 651-52.) She left before the latest investigation commenced. However, her

testimony is relevant to whether the recognized hazard was capable of causing serious physical harm. Although TL

was terminated, her testimony had multiple indices of credibility. She explained that she had wanted to leave the

position but remained on for financial and family reasons. She was forthright in her testimony, directly answering

questions without a suggestion of malice or ill will toward her former employer.




C-69.) Employee injuries included concussions, bruises, scratches, and bites. (Tr. 364, 446, 448,

466-67, 805; Exs. C-3, C-60, C-68, C-69.) Some of the injuries required employees to seek care

in an emergency room and some staff injuries persisted for a long time. (Tr. 805, 810-11; Exs. C-

3, C-60, C-68, C-69.) While some injuries are documented in accident reports, the decision to

complete an accident report was left to the employees themselves. (Tr. 1273-74.) Further, besides

incidents requiring medical attention, there were many more incidents of aggression against


employees. (Tr. 563; Exs. C-57, C-58, C-60 at 3, C-61 at 1.) Respondents do not dispute these

accounts of routine assault and injury. The Secretary established that the recognized hazard was

causing serious physical harm at Pembroke.

4. Pembroke’s Abatement Measures Were Inadequate

As to the fourth element, abatement, the parties agree that under Waldon the Secretary must

show there are feasible means to eliminate or materially reduce the hazard. (Resp’t Br. at 6.)

However, they disagree as to what constitutes such a showing and whether the Secretary satisfied

the requirements in this case.

Although the parties cobble together snippets of various cases in an attempt to create

abatement tests suited to their respective positions, the Commission’s decision in Integra sets forth

a straightforward three prong test to determine whether the Secretary established that there are

feasible means to materially reduce a hazard. 27 BNA OSHC at 1849-50. In Integra, which also

involved a violation of the general duty clause due to the hazard of workplace violence, the

Commission explains that the threshold question is whether the abatement actions the employer

26 In October and November 2016, there were seven incidents of aggression for which medical attention was sought

and an additional six incidents of aggression, such as hits or kicks, for which the affected employees declined to seek

medical attention. (Tr. 563; Ex. C-60.) Overall, UHS DE tracked 151 incidents of aggressive behavior for the fourth

quarter of 2016. (Ex. C-61.)



took are inadequate. Id. at n.14, quoting U.S. Postal Serv., 21 BNA OSHC 1767, 1774 (No. 04-

0316, 2006). If the Secretary shows that the employer’s abatement is inadequate, then he must

propose abatement measures which can be put into effect, i.e., show that the measures are

“feasible.” Id. at 1849-50. If the measures are feasible, then the Secretary must show the measures

will materially reduce the incidence of the hazard, i.e., the Secretary must show the identified

measures will be effective. Id.

Turning first to the abatement measures in place at the time of the inspection, Pembroke’s

approach to minimizing the hazard centers around training, policies, and procedures. (Resp’t Br.

at 24-27.) Pembroke describes its abatement methods for the hazard to include: (1) staffing, (2)

written policies, (3) management commitment, (4) employee participation, (5) worksite analysis

and hazard identification, (6) hazard prevention and control, (7) training, (8) program evaluation,

(9) medical orders, and (10) procedures and equipment for responding to staff calls for assistance.

(Resp’t Br. at 2, 24-27.) Collectively, these things constitute Pembroke’s Workplace Violence

Prevention Program, which Respondents’ argue appropriately protected staff from patient

aggression. (Resp’t Br. at 27.) Respondents concede that their abatement efforts did not eliminate

the hazard of workplace violence. Id. Still, they argue the measures adequately addressed the

hazard and the Secretary failed to establish otherwise. (Resp’t Br. at 23.)

The Secretary finds fault with Pembroke’s existing measures because of the number of

workplace violence incidents that continued to occur at the facility. (Sec’y Br. at 52.) While

Pembroke cites various policies as part of its Workplace Violence Prevention Program, most are

brief, generic, and focus on patient care rather than protecting staff. For example, the one-page

Management of Assaultive Behavior policy refers to patient arguments with other patients and

visitors, but not arguments with staff. (Ex. R-10.) It calls for patients to be “monitored” for violent




behavior and changes related to “assaultive potential” are to be reported to the supervising nurse.

Id. It does not reference to precautions staff should take. Id. Similarly, the one-page Patient

Observation Rounds checklist has a space for nurses to check a box indicating the patient has the

potential for “assaultive or destructive behavior,” but does not specifically refer to staff safety risks


or how to address them. (Ex. R-17.)

Further, some of the measures called for by the policies were not consistently or thoroughly

implemented. There was significant disconnect between the stated policies and what occurred in

practice. For example, Pembroke’s Seclusion and/or Restraint Use policy does not require

debriefings if a workplace violence incident did not result in a patient being restrained or


secluded. (Ex. R-12 at 2.) Also, the assessments Pembroke did conduct after workplace violence

incidents were often incomplete. (Ex. C-74.) Staffing levels were not considered relevant to the

assessment of injuries. (Tr. 513-14.) Even if an employee cited insufficient staffing as a cause of

an injury, such information was not included in the analysis. (Tr. 559.) Additionally, although

called for by the policies, patient treatment plans were not always updated after a patient had been

restrained or secluded. (Ex. R-13.) The December 2016 Massachusetts Department of Mental

Health (“DMH”) survey, which overlapped the OSHA inspection period, made recommendations

to address shortfalls with Pembroke’s debriefing practices, such as failing to adequately note


triggers for violent behavior. Id. at 10-11. So, while Pembroke indicated that it conducted root

cause analyses of injuries, its program was incomplete. (Tr. 852, 854-58; Exs. C-26 at 12, R-12,

27 The form specifies that only nurses should complete this section of the form; not the MHAs who typically perform

the Observation Rounds. (Ex. R-17.)

28 Although Pembroke cites this document as its policy on debriefings, it indicates that Pembroke “shall develop

procedures to ensure debriefing activities occur.” (Ex. R-12 at 2 (emphasis added).)

29 After the December 2016 regulatory visit, Pembroke indicated to the DMH that it corrected, or planned to correct,

these deficiencies. (Exs. C-26, R-13.)



R-13.) See Chevron Oil Co., 11 BNA OSHC 1329, 1333 (No. 10799, 1983) (finding employer’s

abatement approach inadequate in part because of the lack of accurate measuring and monitoring

of data collected through inspections).

Dr. Welch testified in support of the Secretary’s view that Respondents failed to adequately

abate the hazard. Dr. Welch practiced medicine as an attending psychiatrist at multiple facilities

during his career and his work experience includes training staff about the management of violence

in psychiatric care facilities and handling behavioral emergencies. (Ex. C-66; Tr. 915.) Also, he

previously served on the Workplace Violence Committee for a healthcare system. (Ex. C-66.)

Before offering his opinion at the hearing, Dr. Welch reviewed and analyzed: employee

statements, an employee survey focusing on safety, notes from employee interviews, OSHA 300

logs, UHS DE loss control reports, staffing grids and 1:1 staffing information, employee accident

reports, photographs of the physical layout of space at Pembroke Hospital, and DMH inspection

reports and corrective action plans for the facility. (Tr. 919-20, 922-23, 1143-44.) Dr. Welch also

reviewed approximately 60 articles from medical, nursing, and hospital security literature. (Tr.

921-922.) Finally, he surveyed other psychiatric facilities in the same geographic area as

Pembroke Hospital to assist with the evaluation of the Secretary’s proposed methods of abatement.

(Tr. 920-21.)

The abatement measures in place at Pembroke Hospital were inadequate because the

Secretary showed: (1) the hazard remains at Pembroke Hospital; (2) staffing was inadequate; (3)

employees lacked sufficient means to summon assistance; and (4) Pembroke failed to properly

implement patient de-escalation plans.

a) Hazard Remains at Pembroke Hospital

Despite clinical attempts to address patient aggression, the hazard of workplace violence




remained prevalent at Pembroke Hospital during the inspection period. All the non-management

employees described multiple incidents of workplace violence and Respondents’ records detail

additional injuries. Dr. Welch identified “many, many reports” of Pembroke employees

documenting their injuries resulting from workplace violence. (Tr. 927.) OSHA 300 and 300A

forms for 2016 identified 23 recordable injuries. (Tr. 290-92, Ex. C-68.) From this self-reported

information, the CO compared rates of injury at Pembroke Hospital with national averages for


psychiatric hospitals. (Tr. 293.) In comparing injury rates, the CO only considered injuries that

resulted in lost work time, restricted duties or job transfers. (Tr. 291, 293-94.) Pembroke

Hospital’s injury rate was approximately two and a half times the national average for psychiatric

facilities. (Tr. 293-94, 355, 927.)

Respondents point out that the existence of injuries does not establish that its abatement

efforts were inadequate. (Resp’t Br. at 27.) While accurate, the presence of a higher than average

injury rate is relevant to assessing the sufficiency of Pembroke’s program. SeaWorld, 748 F.3d at

1215 (existing safety procedures held inadequate where evidence showed employer’s training and

protocols did not prevent continued injuries). Injuries and incidents are not dispositive, but they

do lend support to the Secretary’s claim that the abatement methods, as implemented, were


b) Staffing was Inadequate

One of Pembroke’s abatement methods was staffing. Both experts agreed that staffing is

a way to reduce the hazard of workplace violence in a behavioral health setting. (Tr. 937-40, 942-

952, 959-60, 966, 980-81, 1510.) Dr. Welch explained that higher staffing levels mitigate violence

30 Because it is based on days away, restricted duties and transfers, the rate of injury is often referred to as the DART

rate. (Sec’y Br. at 7.)




and provide a safer working environment. (Tr. 946-47, 952.) He discussed peer reviewed studies


that supported his view that adequate staffing reduced the hazard. (Tr. 950-52, 962, 989; Exs.

C-94, C-96, C-98, C-100, C-101.) Understaffing is linked to more assaults while higher staff to

patient ratio resulted in fewer patients escalating to a point where they became violent and fewer

patient emergencies. (Exs. C-96, C-97, C-98, C-100, C-101; Tr. 980-81.) Pembroke employees

working at the site during the inspection period confirmed that, in their experience, insufficient

staffing harmed worker safety. (Tr. 90, 329-31, 325-26, 666, 793.)

Dr. Welch concluded Pembroke’s staffing was inadequate to manage the risk of workplace

violence at Pembroke. (Tr. 939, 1067.) Most shifts at Pembroke had less staff per patient than the

national averages for similar facilities. (Tr. 939, 1046-47.) Pembroke does not dispute this, but

argues it follows the minimum staffing requirements set by the DMH for patient care. (Resp’t Br.

at 13, 26.) The DMH requirements vary with the number of patients (i.e. the census) at the facility.

(Ex. C-20.) Pembroke uses a document referred to as the “grid” to determine when the census


reaches a point that additional hours are needed under the DMH requirements. (Ex. C-20; Tr.

371-72, 387-88, 930-33, 1316.) At a minimum, there is one nurse and one MHA for each unit.

(Ex. C-20; Tr. 380.) As the census increases, additional staffing hours may be added. (Tr. 371-

72, 387-88; Ex. C-20.) Each additional patient does not automatically trigger an increase in

staffing. (Ex. C-20.) The grid is focused on minimum staffing levels per patient for their medical

31 Dr. Welch relied on several studies, including, Paul Morrison & Michael Lehane, Staffing Levels and Seclusion

Use, 22 J. of Advanced Nursing 1193-1202 (1995) (hereafter, Lehane Study). The Lehane Study found “a highly

significant difference between the levels of staffing” and the number of violent incidents requiring patient seclusion.

Id. at 1193. Similarly, Vanya Hamrin, et al., A Review of Ecological Factors Affecting Inpatient Psychiatric Unit

Violence: Implications for Relational and Unit Cultural Improvements, Issues in Mental Health Nursing, 30:214-226

(2009), also found that violence is prevented by having adequate staff to patient ratios. (Ex. C-101 at 9.)

32 Because admissions and discharges alter the number of patients, the census is based on the average number of

patients on the unit as counted at least once during a shift.




care. (Tr. 931.) In setting the minimum staffing levels, the DMH does not focus on staff safety or

workplace violence. (Tr. 931-33.)

Besides the individuals “on the grid,” other personnel can assist with workplace violence

incidents. (Tr. 475-76.) During the day shift on weekdays, additional support staff and medical

personnel are available to prevent and respond to workplace violence situations. (Tr. 475, 819,

1299.) However, not all personnel can respond to a code or request for assistance. (Tr. 1299.) At

least one person must stay on each unit at all times. (Tr. 375-77, 871; Ex. R-17.) The employee

assigned to conduct patient checks is limited to that task only. Id. Further, the number of

additional potential responders is not the same on all three shifts. On the evening shift, there is a

single “crisis intervention specialist” who can assist units as needed. (Tr. 871, 1233.) The Director

of Nursing explained that “a lot of crises” happen on the evening shift. (Tr. 1437.) The crisis

intervention specialist is not involved in direct care and tries to be “proactive” to identify problems

before they become “huge issues.” (Tr. 473.)

On the night shift, each staff member was responsible for the greatest number of patients.

(Tr. 388; Exs. C-20, C-78.) There is no crisis intervention specialist and fewer other employees

as well. (Tr. 374; Ex. C-78.) For example, by the grid, there only needs to be one nurse per adult


unit unless the census for that unit exceeds twenty-three patients. (Ex. C-20.) And, even when

the census reaches a point that more staffing hours are required, the grid typically only requires

four more hours of care, i.e., an additional employee for only half of the eight-hour shift. Id. After

the four hours of care are provided, the employee could be sent home, even if additional admissions

33 The adolescent unit also only has one nurse on the night shift, but its census is capped at eighteen patients. On the

night shift, Pembroke added one “float” position on for the entire hospital. (Tr. 375, 730.) The DMH concluded that

this was insufficient. (Ex. C-47 at 6-7.) During its four-day survey in the spring of 2016 (before the OSHA inspection

commenced, the DMH noted one unit operating with 22 patients but only two employees (one nurse and one MHA).

Id. While there was a float, that employee was also needed in another unit. Id.




are anticipated. (Tr. 107-8, 1316.)

As with the night shift, on weekends there were fewer employees at the hospital who could

respond in the event of a workplace violence incident. There was only one doctor for all 90 patients

on weekends. (Tr. 961-62.) There were fewer social workers and case managers as well. (Tr.

960-62.) Patients often were not seen by a psychiatrist, social worker, or case manager until

Monday if they were admitted on the weekend. Id. In fact, 40% of assaults took place on weekends


when there was significantly less staff on site. (Tr. 960.)

The Secretary contends that Respondents failed to have sufficient staff to abate the hazard


of workplace violence. The evidence reveals that staffing was the overwhelming safety issue for

employees and Dr. Welch agreed with employee concerns. (Tr. 306.) According to Dr. Welch,

the baseline level of staffing at Pembroke Hospital was both inadequate and significantly below

other hospitals he surveyed or worked in. (Tr. 939, 1067-68.) The staff to patient ratios were

inadequate to successfully manage the hazard. (Tr. 1067.) Dr. Welch cited multiple occasions

during the inspection period when Pembroke fell below its own minimum staffing requirements

and employees reported injuries. (Tr. 942.) Studies have found that there is a higher rate of

violence and aggression from patients when each staff member is responsible for a higher number

of patients. (Tr. 943, 946-47; Ex. C-101.) As discussed in more detail below, the Secretary showed

that Respondents’ approach to staffing resulted in an insufficient number of people to

34 Ms. Gilmore’s letters regarding her loss control visits to Pembroke Hospital also note increased episodes of

aggression on weekends. (Ex. C-59.)

35 The Secretary does not specifically allege that staffing was inadequate from a medical perspective. However, he

does argue that the “unmet needs” of patients increases the likelihood of a workplace violence incident. (Sec’y Br.

37-38.) As discussed above, patients were dependent upon staff for meals, bathroom access, medication and other

needs. If there was inadequate staff to respond to such needs, certain patients can become aggressive and angry. (Tr.

946-47; Ex. C-101 at 7-9.) “A recurring theme in the research was that unmet needs trigger aggression or violence.”

(Ex. C-101 at 9.) Pembroke’s training materials on de-escalation also acknowledged this can occur. (Ex. R-32.)




accommodate medical orders for 1:1 supervision, to safely address changes in patient acuity, and

to handle new patient admissions. (Sec’y Br. at 51, 54, 61.) Pembroke also failed to maintain

adequate staffing levels for appropriate de-escalation and for therapeutic activities.

(1) Inadequate Staffing for 1:1 Medical Orders

Pembroke argues that one of its methods for addressing the hazard was to allow staff to

seek a higher level of observation for certain patients. (Resp’t Br. at 14, 27.) MHAs and nurses

could request a doctor to order an MHA to be assigned directly to a particular patient. Doctors

could order an MHA to stay within arm’s reach of a single patient. (Tr. 78, 683; Ex. R-11 at 2.)

This level of observation is referred to as a 1:1, meaning one employee for one patient. The

arrangement permits focused de-escalation and prompt notification of co-workers if a patient starts

to become violent. (Tr. 1040-41, 1045, 1125.) While assaultive behavior could support a 1:1

assignment, employee safety was not a specific reason for a patient to be placed on 1:1


observation. (Tr. 1191; Ex. R-11 at 2.) Typically, such arrangements were called for because

the patient was injuring themselves. (Tr. 479-80, 1191; Ex. R-11.)

The Secretary agrees with the importance of 1:1 assignments as a method for both reducing

workplace violence incidents and reducing the likelihood such incidents cause serious injuries.

(Tr. 1045-46.) He does not call for more 1:1 assignments or cite issues with Pembroke’s policies

addressing 1:1s. Rather, the Secretary’s concern lies with how Pembroke addressed, or failed to

address, the need for staff to allow the implementation of orders for 1:1s. In particular, there was

not always enough staff to fill a recommendation for 1:1 staffing. (Tr. 233-36, 453-55, 657, 666,

814-15; Ex. C-78.) Nor did placing a patient on 1:1 supervision typically result in additional staff

36 Pembroke’s policy permitted “special precautions” for “Assaultive/Destructive Behavior” as well “other reasons

deemed appropriate by the physician or nursing staff.” (R-11 at 2.)



being assigned to the unit. (Tr. 332, 383, 453, 945-46; Ex. C-38.) Often an MHA was re-assigned

from helping a group of patients on the unit to instead focus on a single patient. (Tr. 383, 388-89,

946.) This was particularly the case if more than one patient required 1:1 staffing. (Tr. 332, 383.)

Ms. Kent acknowledged it was not a policy to bring an additional staff member in to cover the first

1:1 ordered on a unit. Id. Typically, it would be covered on the next shift, regardless of when the

doctor first ordered it. Id. If there was a need to place a second patient on 1:1 observation due to

violent behavior, Pembroke generally did not add additional staff. (Tr. 332, 383, 388-89; Ex. C-

47.) Staff reported “tension” between management’s desire to remove a patient from 1:1

observation to improve the staff to patient ratio and the belief of front-line workers that doing so

was premature and could result in assaultive behavior. (Tr. 235-36, 664, 657-58.)

Even with a history of violence, 1:1 observation was not assigned until after a patient was

violent towards others or engaged in serious self-injurious behavior at Pembroke Hospital. The

potential for violence to employees did not result in additional staff. (Tr. 77-78, 799.) The patient

had to demonstrate violence or self-injurious behavior to a sufficient degree before a doctor would

order 1:1 observation. (Tr. 77-78.) If a patient previously assaulted medical professionals or others

before arriving at Pembroke, that was not sufficient. Id. They had to exhibit violent behavior at

Pembroke. Id.

For example, during the inspection period, a patient attempted to punch an MHA in the

face. (Tr. 798.) The incident was reported to supervisors. Id. A week later, the same patient

succeeded in assaulting the MHA. (Tr. 798-99.) Even at that point, the patient was only briefly

placed on 1:1 observation for part of a shift. (Tr. 800-1.) According to the MHA, the 1:1

observation was discontinued not because anyone thought the patient no longer needed it, but





because there was no female MHA available to fill the role. (Tr. 801-2.) The same patient was

then involved in another violent attack against staff the following day. (Tr. 804-5.)

Pembroke understood the need for additional staff to implement1:1 assignments. In 2016,

before the latest OSHA inspection commenced, the DMH identified several deficiencies in

Pembroke’s approach to staffing. (Ex. C-47.) The DMH found that Pembroke lacked enough staff


to allow for emergency coverage. Id. In response to the DMH’s investigation, in December

2016, Pembroke submitted a Plan of Correction to the DMH that called for 1:1 assignments to be

handled by additional staff, rather than by reassigning workers from providing care for all patients

to caring for a single patient. (Tr. 396, 703-4; Exs. C-26, C-38, C-47.) In practice, Pembroke

admits that it failed to adhere to its Plan of Correction. (Tr. 332, 736; Ex. C-38.) On multiple

shifts during the inspection period, placing a second patient on 1:1 observation did not increase

the number of staff members on the unit, despite Pembroke’s commitment in the Plan of


Correction. (Exs. C-26, C-38.)

The Secretary argues that Respondents tried to conceal the number of patients on 1:1


observation to make their staff to patient ratio appear better than it was in reality. (Sec’y Br. at

31.) Melissa Heron prepared staffing records for production to OSHA. (Tr. 417-19.) The staffing

records initially contained the number of patients, the total number of staff, and how many

employees were assigned a single patient (i.e., how many were assigned to 1:1 observations). (Tr.

37 Per Pembroke’s “Special Precautions Guidelines” there was a preference for the assigned staff member to be of the

same gender, but it was not required other than during certain time, such as for personal hygiene and toileting. (Ex.


38 While DMH findings relate to patient care, not employee safety, they still had the impact of notifying Respondents

that its staffing levels were low.

39 The Secretary does not allege that Pembroke’s approach to 1:1 staffing would still be deficient if it had followed

what was called for by the December 2016 Plan of Correction consistently during the inspection period.

40 The undersigned notes that the DMH review of Pembroke over three days in December 2016 found that

documentation about what level of precaution patients were on was not consistent. (Ex. R-13.)




419, 423-24, 887-89; Ex. C-78.) Before turning over the documents, Pembroke removed the

information related to 1:1 assignments. (Tr. 417-19, 889.) By removing this information, it could

appear that there was adequate staff for all patients and that the units were less acute. (Tr. 887-88,

1054-55.) This perception was corrected when before the hearing, Pembroke provided the

removed information to OSHA. (Tr. 417-19, 421-22, 883-84, 887-88; Ex. C-78.) Considering the


record as a whole, the Secretary did not establish the editing warranted adverse inferences.

Dr. Welch concluded that Pembroke’s approach to staffing was inadequate and left too few

employees responsible for too many patients and tasks. The approach of not consistently adding

staff when there was a doctor ordered 1:1 observation reduced the effectiveness of Pembroke’s

Special Precaution Guidelines, and left staff more vulnerable to the hazard.

(2) Inadequate Staffing for Acuity

Pembroke argues that it altered staffing based on a unit’s acuity to address the hazard. In

the context of a behavioral health hospital, the “acuity” of a unit refers to the level of activity and

risk. (Tr. 1001-2.) Factors influencing the level of acuity include how many patients, staff, and

visitors are present, and the condition of the patients. (Tr. 1002-3.) The higher the level of acuity

within a unit, the more staff that are needed. (Tr. 576.)

In practice, employees explained that increasing staff for acuity did not consistently

41 The Secretary also cites Pembroke’s coaching of employees only to answer the questions investigators asked and

to not offer information. (Tr. 113, 119, 125; Ex. C-1.) After management provided information about how it expected

employees to act during investigations in writing, employees were specifically directed to destroy the instructions.

(Tr. 125.) Respondents argue that the instructions were perfectly appropriate. The written instructions do not direct

employees to lie or provide misinformation. (Ex. C-1.) AS acknowledged that he did not fear reprisal after speaking

with OSHA. (Tr. 184.) Although the Secretary’s curiosity about why Pembroke would direct employees to destroy

instructions about regulatory investigations is understandable, he did not establish that the coaching practices warrant

adverse inferences.





occur. (Tr. 132, 307; Ex. C-22.) Dr. Welch evaluated staffing at the facility and agreed with the


employees’ view that Pembroke did not sufficiently staff for acuity during the inspection period.

(Tr. 1053-54, 1056, 1058.) Pembroke’s claims about increasing staff for acuity were not supported

by its records. (Tr. 1055-56.) Although Pembroke took steps to ensure that the minimum level of

staffing called for by the grid was met before employees could leave, this requirement did not

extend to situations where a unit’s level of acuity required a higher ratio of staff per patient. (Tr.

454.) Pembroke’s goal was to ensure that the base level of staffing for patient care was met, not

to have enough staff for purposes of abating the hazard of workplace violence.

Even by the grid, the facility was understaffed for parts of shifts on at least three dates

during the inspection period. (Tr. 942; 1443-44.) Injuries from workplace violence incidents were

documented on understaffed shifts. (Tr. 942, 960; Ex. C-78.) In one instance, on November 14,

2016, one adult unit was operating for the evening shift with only one nurse and one MHA. There

had been a series of assaults in the same unit during prior shifts. Just at the start of the night shift,

a patient bit an employee during a restraint. As was always the case on the night shift, there was

no crisis intervention specialist who could move onto the unit if the violent behavior continued

and only one float employee for all the units.

Inadequate staffing harmed preventative actions and responses to violence. Both experts

agreed that when staff is available to attend to patients’ initial signs of agitation, verbal intervention

alone can be effective at preventing physical violence. (Tr. 947, 1512.) The experts also agreed

42 In 2016, UHS DE conducted a staff survey of “safety culture.” Several of the comments cite a need to staff

appropriately for acuity. (Ex. C-22.) Although admissible, less weight is given to these survey results than the

testimony of the current and former employees at the hearing. There is no information about the credibility of the

persons providing the information in the survey and it is not clear what time period the survey responses relate to. Id.

43 The undersigned notes that in contrast to the scope of Dr. Welch’s review of staffing at Pembroke, Ms. Cooke only

reviewed two days of staffing records in depth.




that having numerous potential responders in the event of behavioral health emergencies aids

safety. (Tr. 947-48, 1512.) It facilitates addressing the individual whose behavior started the

situation and limits the ability of other patients to take advantage of an employee’s isolation or

distraction during a workplace violence incident. Id. Dr. Welch explained that “at least five


people” are required for each behavioral health emergency. (Tr. 947-48, 1000.) He cited an

employee accident report involving a patient who weighed 388 pounds. (Tr. 948.) A situation

involving such a large patient could require as many as eight employees to ensure safety for all.

Id. At times during the inspection period, there were only twelve MHAs for the whole facility.

(Ex. C-78.)

KS, a former MHA at Pembroke, explained that she did not feel safe working at Pembroke

because routinely there was not enough staff to address the unpredictability of patient behavior.

(Tr. 793.) She described one situation during the inspection period that spiraled into a melee. (Tr.

794.) Even after calls for all available staff to respond, there was still not enough support. (Tr.

806, 1482-83.) The situation progressed to a point where some patients started to restrain other

patients for the safety of everyone in the unit. (Tr. 794, 807, 856, 948.) KS noted that the situation

would have been worse had it not occurred near the end of one shift and the start of the next. (Tr.

806.) Because of this coincidence, some people who had arrived early for their shift could assist

with bringing the situation under control. Id. Had the incident occurred mid-shift such resources

would not have been available. Id.

44 Pembroke’s policy Code Greens also contemplates the involvement of multiple employees. (Ex. R-9 at 3.)




(3) Inadequate Staffing for Admissions


In addition to prior violent incidents, new admissions also impact the acuity of the units.

(Tr. 1057-58; Exs. C-55, C-101 at 9.) When a patient arrived at Pembroke, a nurse would briefly

take vital signs just inside the door to the unit. (Tr. 75.) An MHA would then bring the patient to


“a very small room” on the unit. Id. The MHA would continue the admissions process alone in

the room with the patient. (Tr. 75, 81.) There were no cameras in the room and only a very small,


high window. (Tr. 88, 512.) The MHA’s assessment included searching the patient and their

belongings for contraband, i.e. items that could harm staff or patients. (Tr. 81, 837, 1181-82; Exs.

R-1, R-8.) AS explained that he did not feel safe to be alone with patients at this point in the

admissions process. (Tr. 81-82.) First, the patient’s condition was not yet well understood. Id.


Second, the patient and MHA are in the room alone before the contraband search is completed.

Id. Pembroke’s Contraband and Restricted Articles policy notes numerous types of objects that

could pose a danger to staff. (Ex. R-1.) AS explained that some patients became aggressive during


the searches. (Tr. 81-82.) In addition to contraband a patient may come in with, the room used

for the searches also had multiple objects that could be used as projectiles. Id. The location and

size of the room made it difficult to call for help even with a radio. (Tr. 82.) Thus, admissions

45 A UHS DE conducted regular “loss control” visits of Pembroke Hospital and prepared summary letters after the

visits. (Ex. C-55.) The December 12, 2016 letter notes an increase of “aggression episodes on weekends due to large

number of admissions ….” Id. at 2.

46 AS indicated that the room was eight feet by eight feet in one of the units. (Tr. 82.) Some of the other units might

have had slightly bigger rooms. Id.

47 The window is described as about the size of “loaf of French bread.” (Tr. 88.) There were no security cameras on

the units, but a few offices have cameras. (Tr. 512.)

48 Searches were also required after patients returned to the units from appointments off-site. (Ex. R-8.)

49 Per Pembroke’s policy, patients could decline the searches. (Ex. R-8.) According to the policy, the patient was to

“remain in open areas unit assessed” if they declined. Id. The policy does not call for them to be secluded from staff.





were one of the “scarier” parts of the job. Id.

After the MHA completed his or her assessment, the new patient moved to the common


room where the existing patients were until it was possible to meet with a physician. (Tr. 91-92,

861-62.) The wait to meet with a physician could take several hours, particularly on weekends,

when only one doctor attended the hospital. (Tr. 861-62.) New patients would often become

agitated or aggressive after being moved into the common room, particularly if in need of


medication,. (Tr. 79.)

Admissions impacted acuity in multiple respects. First, arriving patients were typically

distressed and at their sickest. (Tr. 1057.) As they were not fully assessed yet, the practice of

having the new patient enter the common room directly increased risk and acuity. (Tr. 1014.)

Second, Pembroke did not set separate limits for the number of violent patients being admitted.

(Tr. 77.) Often there were multiple admissions of patients with a history of violent behavior during

a single shift. Id. Third, the admission process itself was very time consuming and could occupy

both a nurse and an MHA for long periods. (Tr. 1058.) Admitting a patient typically required

twenty to thirty minutes of time from an MHA and as much as 90 minutes for the nursing tasks.

(Tr. 653-54, 837-38, 1006-7.) In general, according to the Director of Nursing, there were 6.9

admissions per day. (Tr. 1346.) But admissions, both in terms of their number and their timing,

varied significantly. (Tr. 863, 1008, 1382, 1436, 1516). Pembroke’s Director of Admissions

explained that “every day is very different.” (Tr. 1382.) Sometimes there would be as many as

50 It can take up to twenty-four hours before a physician completes his or her complete assessment of the patient. (Tr.

1291; Ex. R-7.) Typically, at Pembroke, a doctor sees a patient within two hours to assess their competency and

conduct a brief assessment. Id. Patients are also seen by a social worker within three days (72 hours). (Tr. 1423; Ex.


51 Ms. Gilmore, the UHS DE Loss Control Manager, also noted that the lack of clinical case management for patients

admitted over the weekends can increase these patients becoming upset. (Ex. C-58 at 2.)



seven admissions to a single unit during a shift. (Tr. 863.) Dr. Welch pointed to another shift

during the inspection period with fourteen admissions. (Tr. 1014.) While admitting a patient, the

staff member could not focus on de-escalating potentially violent patients and timely intervening

to address assaultive behavior towards staff. Fourth, admissions could result in staffing levels that

fell below the minimum “floor” required by the grid. (Tr. 455.) Pembroke Hospital had no staff

dedicated to handling the admissions tasks. Moreover, it did not consistently staff in anticipation

of admissions. (Tr. 242-43, 381-82, 455, 878-79; Ex. C-78.) Rather, the shift would often continue

with the same number of employees working despite the number of admissions. (Tr. 878-79.)

Additional staff would only be added for the next shift. Id. Fifth, when there was a high influx of

patients, admission tasks had to be performed hastily and sometimes potential weapons were

overlooked. (Tr. 837.) This could increase the risk of injury from workplace violence. For these

reasons, Dr. Welch concluded that Respondents failed to assign enough staff for admissions and

that this increased both the likelihood and potential severity of workplace violence. (Tr. 962, 1056-

58.) Admissions increased the number of patients for each staff member and created additional

work to assess the newly admitted patients. (Tr. 838-39, 1057-58; Ex. C-55.)

Respondents note that discharges reduce the number of patients. Discharges only occurred

on the day shift when there already was additional staff who could respond to workplace violence

events. In contrast, admissions occurred throughout the twenty-fours and were not evenly

distributed. (Tr. 1382, 1014.) Admissions also did not automatically result in increased staff until

the next shift, while discharges could result in staff being sent home mid-shift. (Tr. 107-8, 1316.)

Staff reductions occurred even when new patients were anticipated but had not yet arrived. (Tr.

107-8, 838-39.) Discharges also impacted the acuity of units. The patients likely to be discharged

are, by definition, the healthiest and most stable. (Tr. 1056-57.) The most stable patients were




then replaced with the newly admitted, sicker patients. (Tr. 1005, 1056-57.) The patients arriving

for admission are “acutely decompensated.” (Tr. 1012.) They may not yet be medicated, and Dr.

Welch described it as a “very high-risk time.” Id.

The Secretary established that Pembroke’s approach to admissions was inadequate.

(4) Inadequate Staff to Verbally De-escalate Patients

Pembroke argues that its use of verbal de-escalation, which included a written policy and

training, was adequate. Pembroke’s Management of Aggression policy was a one-page document

described by the risk manager as a review for staff about how to manage assaultive behavior. (Tr.

1186-87; Ex. R-10.) But the document does not provide guidance on how to accomplish this. (Ex.

R-10.) Instead, it notes that patients will be assessed for assault risk and monitored for behavior.

Id. Changes in behavior or events related to the potential for assaultive behavior are to be reported

to the charge nurse, and such reports may alter treatment. Id. Staff did receive other training,

including being taught a technique called Handle with Care, which the Risk Manager indicated


was a method to safely intervene with patients at risk of harming themselves or others. (Tr. 747,

1188, 1419-20; Ex. R-32.)

In addition to the Handle with Care training, Respondents allege that Pembroke also

presents a “Workplace Violence Prevention PowerPoint” to employees and indicates that Exhibit

R-32 is this presentation. (Resp’t Br. at 7.) At the hearing, Mr. Quinn described Exhibit R-32, not

as a Workplace Violence Prevention PowerPoint, but as the “Handle With Care with the Verbal

De-escalation PowerPoint.” (Tr. 1235.) He did not refer to it as workplace violence training and

52 Pembroke’s “Seclusion and/or Restraint Use” policy indicates that staff was also to be trained annually in “Handle

with Care” and “Trauma Informed Care.” (Ex. R-12 at 1.) Ms. Kent cited trauma informed care as being part of the

initial training, but it’s unclear if that was also part of the annual re-training. (Tr. 1420-21.)





his description matches what is listed on the first page of the document. Id. Ms. Burns,

Pembroke’s risk manager, was not asked whether Exhibit 32 constituted Pembroke’s workplace

violence training. (Tr. 497.) She testified that, in general, employees were trained about workplace

violence, but she did not provide such information to staff. (Tr. 1219-20.) When shown Exhibit

C-31, a different collection of PowerPoint slides that uses the phrase “workplace violence,” she

could not confirm whether it was used at Pembroke to train employees. Id. So, the record does


not show that employees were trained on, given, or reviewed Exhibit C-31.

Dr. Welch acknowledged that almost all behavioral health hospitals use verbal de-

escalation and that it is an appropriate, albeit incomplete, intervention for certain patients. (Tr.

1010.) He did not raise specific issues with Pembroke’s training but explained that Pembroke’s

approach to handling workplace violence hazards requires enough staff so employees can promptly


identify who is escalating and intervene quickly. Id. Pembroke failed to consistently have

enough staff to implement the techniques called for by its Handle With Care Program. Id.

So, as implemented, Pembroke’s approach to verbal de-escalation was not sufficiently

effective at abating the hazard. (Tr. 1010.) Training could not make up for the lack of adequate

53 Exhibit R-32 does not include the term “workplace violence” but does refer to violence and “dangerous behaviors.”

(Ex. R-32.) It states that one of the goals of the training is to “Reduce staff and patient injuries.” Id.

54 Respondents did not introduce Exhibit C-31 (a PowerPoint presentation discussing workplace violence) or raise it

with any witness. The Secretary’s counsel showed Ms. Burns Exhibit C-31 on cross examination. (Tr. 1219-20.) Ms.

Burns did not use the materials herself and was unsure if the other employees who conducted training used them. Id.

55 Dr. Welch also noted that verbal de-escalation can be ineffective. (Tr. 1010-1011.) For example, if the patient is

already violent or if their untreated psychosis precludes rational conversation, verbal de-escalation is ineffective. Id.

Ms. Gilmore, the Loss Control Manager for UHS DE, noted in both her October 25, 2016, and December 12, 2016

letters that while non-clinical staff can recognize escalating situations and had training on verbal de-escalation, at the

time of her visit certain employees “DO NOT know how to defend themselves if attacked.” (Ex. C-55 at 2

(capitalization in original); Ex. C-59 at 2 (same capitalization).)





staff to handle psychiatric emergencies. (Ex. C-96 at 15.)

(5) Inadequate Staff for Therapeutic Activities

Dr. Welch explained that therapeutic group activities are effective at reducing incidents of

workplace violence. (Tr. 949, 968-69; Ex. C-101.) Regularly scheduled and predictable activities

promote engagement and decrease levels of acuity on the units. (Tr. 820-21, 949-50, 1061-62.)


Pembroke recognized the importance of regularly scheduled therapeutic groups. (Exs. C-55, C-

57, C-59, R-32.) Its de-escalation training called for increasing structure as a means to control

crises. (Ex. R-32 at 13.) The Handle With Care training cites schedules and programs as helpful

in avoiding power struggles that can lead to violence. (Ex. R-32 at 22.) Before regulatory visits,

Pembroke instructed employees to check to make sure details about groups were written, that


groups occurred, and that schedule changes were noted. (Ex. C-1 at 1-2, 4-5, 7-8.) Pembroke

also agreed with the DMH’s audit findings that the posted information about groups failed to

adequately describe what groups were actually going to occur. (Exs. C-26, R-13, R-29.)

Despite this recognition of the importance of making sure groups occurred, groups were

regularly canceled due to insufficient staff. (Tr. 221.) KS indicated that at least three times a week

a group would be canceled. (Tr. 819.) This most frequently occurred on the evening shift, which

had fewer available employees. Id. Still, even on the day shift, groups were routinely canceled.

56 The Handle with Care program also called for structure and having employees be able to provide calming modalities.

(Ex. R-32.) As discussed below in addition to a lack of adequate staffing, Pembroke also fell short in implementing

these aspects of the program. Id.

57 Ms. Gilmore assessed whether groups were occurring and how they were functioning as part of her routine reviews

of Pembroke Hospital. (Tr. 549; Exs. C-55, C-57, C-59.) Her reports discuss how groups can help patients and how

not having groups as scheduled can trigger aggression. (Ex. C-59 at 2; C-57 at 2.)

58 One witness said that she was not aware of any policies related to when groups would be canceled. (Tr. 228.) The

risk manager indicated that Pembroke had an “activities therapy policy,” as well as policies related “running groups”

and the “staff to patient ratio” for groups. (Tr. 535.) Pembroke did not introduce any of these policies into the record.




(Tr. 50, 137.) Groups were canceled because staff was too busy with admissions or because of the

need to respond to emergencies. (Tr. 138, 221, 677-78, 820.) There was not enough staff to

consistently run the groups and handle the other tasks. Id. Groups were an “afterthought,” rather

than an effective means of abatement. (Tr. 677-78; Exs. C-26 at 3, R-13, R-28, R-29.)

c) Inadequate Means to Summon Assistance

Pembroke argues that staff could adequately summon additional assistance when needed

to prevent or end patient on staff violence. Pembroke’s units often only had three employees. (Tr.

380, 1047; Ex. C-20.) There is no debate that this was insufficient to address all workplace

incidents. Staff could verbally announce a “Code Green” to seek assistance in the event of violence

or if they anticipated the need for additional support. (Tr. 57; Ex. R-9.) Typically, at least one or

two times a day, staff would seek assistance through the Code Green process. (Tr. 827.)

Code Greens could also be called over two-way radios (i.e., walkie talkies) if available or


by using a phone that could connect to an overhead public address system. (Tr. 57, 59-60.)

Employees were instructed to stay with the patient even when summoning assistance. (Ex. R-9.)

However, if the employee did not have a radio, they would need to access the one phone located


in the middle of the unit. (Tr. 59.) There was no phone in the area used for admissions. Id. The


number of two-way radios on each unit varied but there were never more than four. (Tr. 60-61,

59 Per Pembroke’s Code Green policy, psychiatric emergencies also “should be announced using walkie talkies and

the overhead paging system.” (Ex. R-9.) Employees described using either a walkie or the phone system. (Tr. 57,

59-60.) Some incidents would require multiple calls for assistance.

60 It is unclear if patients also used these phones. (Ex. R-28.) An audit by the DMH from May 2016 specifically

called for more phones in at least two units. Id.

61 More recently, Pembroke increased the number of radios to six per unit and implemented steps to make sure each

unit had a sufficient number of working radios. (Tr. 456.) The Secretary does not argue that this revised approach

was still deficient. However, the suitability of Pembroke’s abatement program is assessed as of the time in the




456-57.) Staff who could not access the phone or a radio could verbally request that someone else

commence the Code Green process. (Tr. 59.) There was no equipment or procedures to silently

alert others of the need for assistance.

Employees discussed various issues with the radios. First, there were not enough to permit

every staff member to carry one while working on the units. (Tr. 60-61, 231; Ex. C-78 at 3.)

Second, the radios had reliability issues. (Tr. 60.) Problems included static, faulty batteries, and

radio airwaves shared with maintenance workers. (Tr. 60, 815.) During fresh air breaks for

patients, a single MHA is responsible for seven patients. (Tr. 56.) At such times, staff must rely

on the radio to summon additional assistance. Id. A nurse described multiple occasions when she

attempted to use the radio to call for a Code Green and just heard static or was not able to reach

enough employees. (Tr. 221.) Sometimes staff would have to make multiple calls for help with

both the radios and through the one phone on the unit in order to get enough people to respond to

the situation. (Tr. 806-7.) One nurse described a situation during the inspection period that

continued to escalate as attempts to get assistance were repeatedly made. (Tr. 806, 1482-83.) In

short, there was both difficulty in communicating the need for help and then a lack of people who

could respond. (Tr. 221, 815.)

Dr. Welch concluded that Pembroke lacked an adequate system for employees to summon

assistance. (Tr. 1000.) In addition to the problems with the radios and the limited number of

phones, there was also no way to silently seek assistance. Staff communications over the radios,

verbal calls for assistance, and use of the unit phone could be audible to potential assailants. In

contrast, panic alarms allow for discreet requests, which may help to avoid escalating a patient

further. (Tr. 1049.) The Secretary established that Pembroke’s approach was inadequate.




d) Failure to Properly Implement De-Escalation Plans

To help manage patient aggression and thereby reduce incidents of workplace violence,

Pembroke sought information from patients about activities or interventions that successfully

calmed them in the past. (Tr. 512, 1255, 1293; Exs. R-12, R-18, R-32 at 9.) This information was

then used to develop crisis intervention plans for staff to follow in order to limit patient aggression.

(Tr. 512-13.) Staff could augment the information from the patient based on what they observed

during treatment. Id.

The Secretary does not dispute that Pembroke sought information for de-escalation plans

and communicated the findings to staff. His contention is that the facility lacked sufficient

equipment to implement one aspect of the plans. Staff were instructed as part of their verbal de-

escalation training to employ the calming techniques patients identified. (Ex. R-32.) “Listening

to music” is the first option listed in the “helping or comforting strategies” section of the form used

for the de-escalation plans and many of the plans called for its use. (Tr. 830, 1434; Ex. R-18.) Dr.

Welch and employees reported that music was effective for de-escalating aggressive patients. (Tr.

73-74, 666, 830-31, 962-63, 1064, 1245, 1293, 1306.) However, employees discussed situations

where they were unable to allow patients to listen to music because there were too few devices.

(Tr. 74, 667.) The insufficient number of music playing devices also led to conflicts among


patients. (Tr. 74-75, 809-10.) The failure to have the equipment necessary to implement the

plans rendered the policy ineffective at abating the hazard. See Chevron, 11 BNA OSHC at 1332-

33 (finding inspection program to be an inadequate method of abatement because of how it was


62 About six months before the hearing, Pembroke had implemented a new approach that improved patient access to

music for de-escalation. (Tr. 1255-56.) The Secretary’s arguments are limited to the conditions observed at the time

of the inspection.




e) Pembroke’s Abatement was Inadequate

Pembroke’s approach to staffing, summoning assistance, and implementing de-escalation

plans was inadequate. While Pembroke’s training program was not shown to be deficient, staff


training alone is insufficient to abate the hazard. Dr. Welch, supported by research, concluded

that to address the hazard, workplace violence programs must include adequate staffing and the

use of engineering controls such as maintaining systems to quickly summon assistance. (Tr. 986,

1000.) Despite the abatement measures implemented by Pembroke, the facility continued to

experience higher than average injuries. Pembroke failed to appropriately implement the safety

measures it had identified to address the hazard. See SeaWorld, 748 F.3d at 1206, 1215 (finding

existing procedures inadequate). The Secretary established that Pembroke’s program was

inadequate. Cf. U.S. Postal, 21 BNA OSHC at 1773-74.

The general duty clause requires employers to “take all feasible steps” to protect against

recognized hazards and implement every abatement measure “whenever it is recognized by safety

experts as feasible, even though it is not of general usage in the industry.” Gen. Dynamics, 599

F.2d at 464. As discussed below, the Secretary identified currently available engineering and

administrative controls that can be feasibly taken and will materially reduce the hazard. See Pelron

Corp., 12 BNA OSHC 1833, 1836 (No. 82-388, 1986) (finding that the Secretary may establish

that an employer's existing safety procedures were inadequate by demonstrating that there were

“specific additional measures” required to abate the hazard).

5. Feasible Means to Abate the Hazard Exist

Despite acknowledging it recognized the hazard and arguing it has taken multiple effective

63 Dr. Welch cited a study that pointedly notes the lack of research supporting the effectiveness of training alone in

reducing workplace violence. (Ex. C-96 at 15.)




steps to abate it, Pembroke claims the hazard is too unpredictable to be abated. (Resp’t Br. at 28-

29.) However, its own expert flatly refuted this contention. She explained “there are certainly

methods” to reduce the risk of workplace violence. (Tr. 1508.) Dr. Welch concurred with her that


there are feasible means to reduce incidents of workplace violence. (Tr. 934-35.) Unpredictable

events do not remove an employer’s obligation to adhere to the general duty clause and address

recognized hazards with feasible means of abatement. SeaWorld, 748 F.3d at 1207, 1215;

Armstrong, 8 BNA OSHC at 1074 (finding the employer’s contention that it should not be found

in violation of the general duty clause when it could not foresee the sequence of events that led to

injury to be without merit).

Respondents also criticize the nature of the Secretary’s multi-prong abatement. Yet, when

a workplace contains a recognized hazard that is likely to cause death or serious harm, the

employer must take “all feasible steps to eliminate or materially reduce the hazard.” Armstrong,

8 BNA OSHC at 1074, citing Gen. Dynamics, 599 F.2d at 464. Feasible abatement may take the

form of a process rather than a single one-time action. In Pepperidge Farm, Inc., 17 BNA OSHC

1993 (No. 89-265, 1997), the hazard could not be abated with a single action, so it was permissible

for OSHA to require the employer to engage in a process approach to abatement to determine what

action or combination of actions would materially reduce the hazard. 17 BNA OSHC at 2033.

See also Pegasus Tower, 21 BNA OSHC 1190, 1191 (No. 01-0547, 2005) (finding that following

a detailed compliance directive was a feasible means of abatement). Similarly, in Integra Health,

64 Dr. Welch acknowledged the role of patient behavior but argued that individual patient factors should not be over-

emphasized at the expense of other factors, particularly environmental or administrative factors well within the

facility’s control. (Tr. 973, 1010-11.) Peer reviewed literature supports his opinion. (Exs. C-95 at 2-3, C-96 at 16.)





several actions were found to be feasible methods for reducing workplace violence hazard. 27

BNA OSHC at 1849-51. The Secretary does not have to prescribe an action or series of actions

that address all situations where the hazard arises. Abatement may still be feasible even if it cannot

be used in all situations. See Wheeling-Pittsburgh Steel Corp., 10 BNA OSHC 1242, 1245 (No.

76-4807, 1981) (consolidated).


While not a process per se, the Secretary has set forth multiple abatement methods here.

He argues that each method is feasible and will materially reduce the hazard. (Sec’y Br. at n.7.)

The Secretary argues that all these actions should be taken. Id. But, if the undersigned concludes

that one or more of the methods are not feasible, the Secretary argues that each remaining method

is feasible and would materially reduce the hazard. Id.

Several of the proposed abatement actions relate to maintaining appropriate staffing at the


1. Cease efforts to interfere with the issuance of medical orders that specify

required staffing arrangements.

2. Ensure that staffing is sufficient to allow the issuance and implementation of

medical orders that specify staffing arrangements.

3. Maintain staffing that is adequate to safely address changes in patient acuity

and new patient admissions. For example, do not decrease staffing levels mid-

shift regardless of the timing of patient discharges.

4. Maintain adequate staffing to support therapeutic activity groups and recreation

periods, thereby engaging patients in activities that reduce patient agitation and

incidents of workplace violence.

5. Provide security staff and/or crisis intervention specialists on all three shifts to

65 The abatement methods proposed for the workplace violence hazard in Integra included: creating a written

workplace violence prevention program, creating a system for reporting and tracking safety concerns, providing

employees with a reliable way to summon assistance when needed, and increasing the number of staff for certain types

of assignments. 27 BNA OSHC 1849-50.

66 The Secretary divides his proposed abatement into seven paragraphs, labeled (a) through (g). (Sec’y Br. at 4.) Some

of the paragraphs require more than one action. Id.



assist in preventing and responding to violent events.

(Sec’y Br. at 4.) Three actions focus on how and where new patients are admitted.

1. Establish a team of nurses and mental health associates (MHAs) that focuses

primarily on the performance of the tasks associated with the admission and

assessment of new patients. This team should not be staffed by employees who

are assigned to care for already-admitted patients.

2. Dedicate a physical area, apart from all other units of the hospital, to the

admission and assessment of new patients.

3. Do not allow new patients into other units until the tasks associated with

admission, including clinical assessments and the provision of medical orders,

are complete.

Id. The final abatement measures relate to the provision of equipment to summon assistance and

implement patient crisis prevention plans:

1. Provide personal panic alarms for all employees who may work in close

proximity to patients, including but not limited to nurses, MHAs, housekeeping

staff, and case workers. Provide training on this equipment and ensure that the

equipment is maintained in working order at all times.

2. Maintain equipment that is sufficient for the implementation of each patient’s

individual crisis prevention plan.

(Sec’y Br. at 4.)

Pembroke does not allege that any of these steps are technologically or economically

infeasible. Notably, Respondents argue that Pembroke already complies with some of the

abatement methods. Nor does Pembroke allege it lacked the authority or control to implement the

proposed abatement measures. Instead, Respondents argue that the Secretary failed to provide

enough detail about what constituted adequate abatement and therefore it was deprived of the fair

notice of the abatement required. (Resp’t Br. at 21.) In their view, besides showing fair notice of

the hazard, the Secretary should also have to prove the employer knew how to materially abate the

hazard before the citation was issued. (Resp’t Br. at 21-23.)

Discussions of “fair notice” in Commission cases involving alleged violations of the



general duty clause typically center on the employer’s notice of an obligation to address a hazard

in the workplace, not on whether the employer had notice about how to abate or materially reduce

a hazard. See e.g., Otis, 21 BNA OSHC at 2206 (discussing fair notice in the context of how the

Secretary defined the hazard); Beverly, 19 BNA OSHC at 1163 (noting that although the complaint

may not have been precise, “at the time of the hearing” the Secretary sufficiently specified the

alleged hazard). Having found above that the Secretary provided sufficient notice of the hazard to

satisfy the Constitution's due process clause, it is time to assess Respondents’ demand that the Act

requires the Secretary to show that employers had fair notice of the required abatement.

The requirement to show feasible means of abatement relates to the Commission’s test for

ensuring that the Act does not impose strict liability. The Commission has not articulated the

abatement test in the manner Respondents seek. Under the Commission’s test, the abatement

prong requires the Secretary to establish the existence of feasible means to eliminate or materially

reduce the hazard. See e.g., Integra, 27 BNA OSHC at 1849 (addressing a violation of the general

duty clause for a workplace violence hazard); Waldon, 16 BNA OSHC at 1062 (affirming a

violation of the general duty clause for a viral exposure risk at a healthcare facility). Showing

feasibility requires the Secretary to put forth abatement methods that are capable of materially

reducing the hazard while also being technologically and economically possible. 27 BNA OSHC

at 1849-51. The Secretary does not have to show that the abatement methods would eliminate the

hazard. Arcadian Corp., 20 BNA OSHC 2001, 2011-13 (No. 93-0628, 2004) (finding two of the

Secretary’s proposed methods of abatement feasible). Nor is he tasked with showing that the

absence of the abatement method was the sole cause of harm. Id. His obligation is to show that

the proposed abatement is capable of being put into place and effective. Acme Energy Servs., 23

BNA OSHC 2121, 2127-29 (No. 10-0108, 2012) (finding the abatement prong met because



prescribed action would materially reduce the harm, even though it would not eliminate the

hazard), aff’d, 542 F.App’x 356 (5th Cir. 2013) (unpublished).

Respondents point to no precedent imposing an obligation on the Secretary to establish an

employer’s knowledge of the proposed abatement methods prior to a citation being issued. In

Chevron Oil, the Commission concluded that the Secretary had to show that feasible means to

eliminate the hazard exist. 11 BNA OSHC at 1330. It was not the Secretary’s burden to show that

the employer or its industry recognized the necessity for proposed safety equipment. Id. at 1330-

31. Instead, the Act requires the Secretary to establish recognition of the hazard and that it can be

materially reduced feasibly. Id. See also Kansas City, 10 BNA OSHC at 1422 (“the recognition

element of an employer’s duty under the general duty clause refers to knowledge of the hazard,

not recognition of the means of abatement); Integra, 27 BNA OSHC at 1849, 1851.

In the present matter, the employer had notice of effective methods of abatement before

the Secretary issued the citation. The methodologies proposed here, while tailored for the issues

identified at Pembroke Hospital during the inspection, align with the OSHA Road Map and

Workplace Violence Guidelines. (Exs. C-97, C-102.) The Road Map specifically discusses

administrative controls to address the hazard, including, assessing patients for violence during

admission, ensuring adequate staffing on all units and shifts, and having policies and procedures

that minimize stress for patients. (Ex. C-102 at 18-19.) It also discusses engineering controls such

as panic buttons. Id.

Besides these publications, a letter sent to Pembroke prior to commencement of the

investigation that lead to the instant Citation outlines similar approaches to abatement as those the

Secretary now seeks. (Ex. R-37.) The letter explains that MHAs are exposed to workplace

violence hazards at Pembroke Hospital. Id. But, as noted above, the letter states that OSHA





elected not to cite the hazard. Still, the letter encourages Pembroke to act and advises that

feasible and acceptable means to reduce the hazards of workplace violence at Pembroke Hospital

include: (1) modifying admissions process, (2) improving responses to crises, (3) making

adjustments to staffing, and (4) providing staff with personal panic alarms:

Carefully review admissions to ensure your staff are equipped to handle acute

patients who have a history of violent behavior. If an admitted patient has a known

history of violent behavior from previous institutions, increase staffing levels for

the unit and inform all workers of the potential of violent behavior.

Conduct a workplace analysis … to find existing or potential hazards for assaults

and workplace violence. This process should involve record analysis, tracking of

injuries, and monitoring trends based on location, shift changes, and staffing levels.

Workplace analysis should include review of personnel ability to respond to unit

crisis on all shifts. Review the staffing on all shifts to ensure that it reflects the

daily acuity of the patient as well [as] patient census and to ensure workers are able

to work without putting themselves at risk for a violent assault.

Change the present grid to increase staffing levels on the acute units and other units

where attacks are occurring. Ensure that extra staff is in place when a patient’s

condition requires a one to one watch and when there has been a history of attacks

in a unit. An increase in staff levels has the potential to decrease injuries in the

workplace. … .

… A silent button that is carried by staff would provide an additional and immediate

way to alert staff, in the event of an escalating situation with a patient and may

reduce response times. Ensure that if purchased, all employees are trained and that

the panic buttons are maintained and tested according to manufacturer instructions.

Id. The letter goes on to point Pembroke to the OSHA Workplace Violence Guidelines and

indicates the facility can request a free on-site consultation to identify other feasible measures of

abatement. Id. The letter does not provide an obligation independent of the Act to take the specific

steps identified therein. Id. Pembroke was free to adopt different feasible abatement measures to

67 In the year prior to this letter, Pembroke had thirteen injuries related to aggressive behavior. (Ex. R-37.) After its

issuance, the number of injuries at Pembroke increased. (Tr. 292, 926; Exs. C-60, C-68, C-69.)



address recognized hazards in its workplace. See Pepperidge Farm, 17 BNA OSHC at 2032

(employers are free to develop solutions different than what the Secretary proposes to render their

workplace “free” of recognized hazards); Brown & Root, Inc., Power Plant Div., 8 BNA OSHC

2140, 2144 (No. 76-1296, 1980); Chevron, 11 BNA OSHC at 1334, n.16 (emphasizing that the

employer could institute “other equally effective methods as long as its alternative methods

achieve at least as great a reduction of the hazard”).

Although the letter did not obligate Pembroke to approach the hazard the way the Secretary

suggests, it also did not provide immunity from complying with the Act. When employers know

existing safety standards do not adequately address hazards, they must take all feasible steps

necessary to protect employees. Int’l Union, United Auto., Aerospace & Agric. Implement

Workers of Am., et al. v. Brock, 815 F.2d 1570, 1577 (D.C. Cir. 1987). “The question is whether

a precaution is recognized by safety experts as feasible, and not whether the precaution's use has

become customary.” Beverly, 19 BNA OSHC at 1191.

Pembroke did not adopt the suggested measures outlined in the letter. The number of

employee injuries from the hazard increased from 2015 to 2017. (Exs. C-68, C-69.) Dr. Welch

identified thirty-two employee accident reports attributable to patient on staff violence in the

Pembroke records he reviewed. (Tr. 926.) The Secretary showed that Pembroke’s alternative

approach to abating the hazard was inadequate. So, rather than a lack of fair notice, this case

represents a situation of notice of both the presence of a hazard (something Respondents do not

dispute) as well as notice of feasible abatement measures.

The OSHA publications and letter to Pembroke alone do not satisfy the Secretary’s burden

on the abatement prong. The Secretary must show that his proposed abatement will materially

reduce the hazard. To meet that obligation, the Secretary largely relies on the expert testimony of




Dr. Welch. Before offering his opinion, Dr. Welch conducted an extensive review of Pembroke

documents and relevant peer reviewed literature. (Tr. 919-23, 1143-44.) He also surveyed how

other similar facilities abated the hazard. Id. Based on his experience and research, he concluded

that each abatement measure outlined in the amended citation was a feasible means by which an

employer could materially reduce the hazard of workplace violence. (Tr. 923-24.) See Arcadian,

20 BNA OSHC at 2011 (“Feasible means of abatement are established if conscientious experts,

familiar with the industry would prescribe those means and methods to eliminate or materially

reduce the recognized hazard”).

Respondents also offered expert testimony on the issue of abatement from Ms. Cooke. Her

testimony was largely based on her experience rather than peer reviewed studies or time spent

reviewing Pembroke Hospital. (Tr. 1458.) Throughout much of her testimony, Ms. Cooke failed

to identify the source material for her expressed beliefs. (Tr. 1458-59.) When asked, Ms. Cooke

did not identify any specific literature she reviewed before evaluating this matter. She said that

she “might’ve (sic)” re-looked at the Workplace Violence Guidelines or the Joint Commission’s

Sentinel Event Alert on Preventing Violence in the Health Care Setting (“Sentinel Alert”). (Tr.

1458; Exs. C-97, C-98.) However, these publications were not identified as documents she

reviewed in her opinion letter. (Tr. 1459.) Similarly, Ms. Cooke indicated that she “attempted”

to look up “some of the research” relied on by Dr. Welch, but does not indicate which studies, if

any, she reviewed either before issuing her opinion letter or before testifying at the hearing. (Tr.

1458.) Ms. Cooke was not sure how much time she spent on her review but “would venture to


guess” it was 25-30 hours. (Tr. 1461.)

68 “About six hours” of her time was spent at Pembroke Hospital itself. (Tr. 1474.)



Ms. Cooke was offered for only limited purposes. (Tr. 1463.) She offered no opinion on

whether Pembroke could have materially reduced the hazard of workplace violence. (Tr. 1461.)

She was accepted as an expert who evaluated “the suggested feasible abatement methods offered

by the Secretary in this case.” Id. In contrast, Dr. Welch was asked to assess, among other things:

(1) the risk of workplace violence at Pembroke, (2) the recognition of the risk of workplace

violence; (3) what measures would create a material reduction in that risk; and (4) were those

methods identified feasible for Pembroke to implement. (Tr. 902.) While satisfying the minimum

requirements for admissibility, Ms. Cooke’s testimony is given less weight than that of Dr. Welch.

His opinions more fully addressed the contentions at hand and were backed by specifically

identified peer reviewed literature.

a) Staffing

As discussed in the evaluation of the existing methods of abatement in place, staffing at

Pembroke was deficient. Dr. Welch, supported by peer reviewed studies, testified that increasing

staffing could materially reduce the risk from the hazard and that doing so was feasible. (Tr. 903,

937-38, 985; Exs. C-97, C-98, C-101.) Ms. Cooke agreed that examining staffing was relevant to

the cited hazard. (Tr. 1498-99.) She acknowledged that one of the ways to reduce the risk

associated with workplace violence is to both review and have “good staffing.” (Tr. 1510.) She

has advised hospitals that adequate staffing is important to prevent workplace violence. (Tr. 1529.)

The Director of Nursing, Ms. Kent, also recognized the role extra staff can play in diffusing

aggressive events. She explained that the facility experienced more crises on the evening shift and

that adding an additional person was helpful to both prevent and address crises. (Tr. 1437.) The

Secretary established a link between shifts with inadequate staffing and increases in the hazard.

(Exs. C-3, C-78.)




The Secretary showed that it was feasible to increase staffing at Pembroke Hospital. Levels

of staffing at Pembroke Hospital were objectively lower than at similar facilities in the region and

nationally. (Tr. 104-5, 937-38.) For example, depending on the shift, Pembroke Hospital had


approximately 15-30% fewer nurses than a similar psychiatric facility in the same state. (Tr.

937.) AS, who worked both at Pembroke Hospital and another similar facility with the same sized

units, explained that there was less direct care staff at Pembroke. (Tr. 105.) The other facility had

more nurses per shift and no less than three MHAs per unit. Id. In comparison, Pembroke’s units

typically operated with fewer nurses and MHAs. (Tr. 105-6.) The competitor facilities in the

region surveyed by Dr. Welch maintain higher levels of staffing and Respondents failed to counter


his research that implementing the abatement called for by the Secretary was feasible. See Trinity

Indus., 15 BNA OSHC 1481, 1485 (No. 88-2691, 1992) (finding in connection with a violation of

a specific standard that the implementation of a safety measure at other facilities shows the action

was feasible).

The Secretary proposes five methods to address the staffing deficiency. (Sec’y Br. at 4.)

The proposed abatement implicitly acknowledges that there was more than one way to increase

staffing at the facility. Respondents twist this flexibility to argue that the Secretary failed to

establish the specific steps it must take to sufficiently reduce the hazard. In fact, the Secretary

showed that Respondents failed to maintain a sufficient level of staffing: (1) to allow for increased

observation of particular patients; (2) for the acuity of particular units; (3) for anticipated

admissions; (4) to respond to crises, particularly on the evening and overnight shift; and (4) to

69 Pembroke also had more patients per doctor than other facilities surveyed by Dr. Welch. (Tr. 938.)

70 Respondents do not contend that the facility failed to make a profit. Respondents did not introduce any information

regarding the facility’s finances such as budget documents. There is no evidence the facility could not afford any of

the Secretary’s proposed abatement measures. See Capeway Roofing Sys., Inc., 20 BNA OSHC 1331, 1342-43 (No.

00-1968, 2003) (party would have produced evidence if it had been favorable).




engage patients in therapeutic group activities. The Secretary established that bringing the facility

into alignment with the baseline staffing at other similar facilities was feasible and would reduce

the hazard by addressing these deficiencies.

(1) Staffing for Medical Orders

The Secretary argues that Pembroke interfered with the issuance of medical orders for

staffing and failed to ensure there was enough staff to implement such medical orders:

Cease efforts to interfere with the issuance of medical orders that specify required

staffing arrangements. Ensure that staffing is sufficient to allow the issuance and

implementation of medical orders that specify staffing arrangements.

As noted above, when a patient is at high risk for assaultive behavior, a doctor can issue a medical

order for direct observation by assigning an MHA to stay at all times within arm’s reach of a


patient. (Tr. 378-79, 1190-91.) Pembroke considered such orders to be a “last resort” to handling

patients and were used only temporarily. (Tr. 1191, 1300.) Orders for 1:1 observation are

reviewed by the management team. (Tr. 1191.) The focus of these reviews is to assess whether


the patient continues to need the higher level of observation. Id.

Dr. Welch expressed his view that administrators were pressuring staff to remove patients

from 1:1 observation. (Tr. 1037-38.) He explained that interference, by management, with the

medical determination of whether such 1:1 observation is inappropriate and increases the risk of

workplace violence. (Tr. 1038.) His opinion was based on employee statements as well as

71 Sometimes 1:1 observation would be limited to only those times that the patient was awake. (Tr. 379.) Patient

behavior is unpredictable, and patients wake at unexpected times. In such instances, there would be an immediate

need for an employee to observe the patient directly. (Tr. 379-80.) Ms. Devane indicted that in such situations the

supervisor would have the float take over that patient. (Tr. 380.) However, only one person worked as a float on the

overnight shift.

72 A smaller group also assesses the need for such observation at other times, so that orders for 1:1s are reviewed

approximately every eight hours. (Tr. 1191.)




testimony about management’s involvement in the decisions about orders for 1:1s. (Tr. 919-24.)

He was surprised that the medical decision to implement a 1:1 required notification of the CEO

and opined that this practice alone could pressure employees to refrain from seeking a 1:1 even

though safety concerns warranted such an arrangement. (Tr. 1004, 1006, 1038.)

An MHA explained her view that 1:1 observation was not consistently put in place when

called for on the grounds of employee safety. She testified that after a patient attacked her, the


patient was only briefly placed on 1:1 observation for part of the shift. (Tr. 799-801.) Staff also

described pressure to have just enough staff to meet the minimum requirements. (Tr. 1044.) On

some shifts, there were just three employees for a unit. (Tr. 380.) If one of the MHAs was assigned

to observe a single patient, the other two employees had to handle the remaining patients and tasks.

Id. If a second patient required a 1:1, that could not always be accommodated until the next shift.

Id. The CEO described receiving “heat” about his decisions around the number of staff per patient

but asserted that his determinations about staffing were not over-ridden. (Tr. 732-33.) Still,

Pembroke’s approach of having the least number of employees working as possible made fulfilling

staffing requests difficult, particularly during the overnight shift. (Tr. 801, 1044.) Notably,

Pembroke was not able to consistently implement the Plan of Correction it submitted to the DMH

which called for 1:1 assignments to be handled by adding staff. (Tr. 396, 703-4; Exs. C-38, C-47.)

Compared to other facilities, staff to patient ratios were consistently lower at Pembroke

Hospital than at other stand-alone psychiatric hospitals. (Tr. 1046-47.) On average, Pembroke


Hospital had less nursing staff per occupied bed as compared to national averages. (Tr. 1046-

73 Ms. Devane explained that 1:1 observation could be intrusive for patients. (Tr. 1191.) The Secretary does not

dispute this. His contention is limited to situations when 1:1 observation is clinically warranted for assaultive behavior

that creates a risk of violence directed at employees.

74 Pembroke Hospital operated shifts with 15 to 20 percent fewer nurses than other similar psychiatric facilities in the

same state. (Tr. 1047.)



47.) Dr. Welch opined that it would be feasible for Pembroke to maintain a higher level of baseline

staff to bring it in line with other similar facilities. Id. Such staffing would permit 1:1s to be put

in place more quickly rather than waiting for the next shift to begin. (Tr. 1044-47.) While Ms.

Cooke indicated staffing was adequate, she did not dispute or counter the evidence relied on by

Dr. Welch in reaching his conclusion that increasing the level of staff would reduce the hazard.

As for the other part of this proposed abatement, the interference with medical orders

concerning staffing arrangements, the record reveals that Pembroke was already abiding by this.

Rather than interference with the issuance of orders for 1:1s, the issue was Pembroke’s ability to

staff the assignments. The Secretary established that having a greater staff to patient ratio would

permit the prompt, appropriate use of 1:1 observation to materially reduce the hazard.

(2) Staffing for Acuity and New Admissions

The Secretary argues that to materially reduce the hazard, Respondents need to maintain

adequate staffing to address changes in patient acuity and new patient admissions:

Maintain staffing that is adequate to safely address changes in patient acuity and

new patient admissions. For example, do not decrease staffing levels mid-shift

regardless of the timing of patient discharges.

(Sec’y Br. at 4.) Dr. Welch, supported by peer reviewed literature, explained the need for

appropriate staffing, particularly to handle admissions, which are a “high risk time for acuity.”

(Tr. 1058.) Increased admissions are associated with workplace violence incidents. (Ex. C-101.)

Respondents agree with the need to staff for acuity and admissions. Still, as discussed,

Pembroke largely only met the minimum number of staffing hours per patient. There’s no debate

that this minimum number was not the appropriate average ratio of staff per patient for all

circumstances. In other words, the DMH’s view is that it is never appropriate to have less than the

ratio of staff per patient called for by the grid. However, no one contends that the level of staffing




called for by the grid is always adequate to address the hazard of patient on staff violence. There

is no dispute that to keep workers safe, each unit will routinely require higher staff to patient ratios

to address acuity.

While the staffing records showed 1:1 assignments, there was little evidence of increased

staffing to address overall acuity. Increasing the number of patients on 1:1 observation caused the

remaining employees to be responsible for a higher number of patients. (Tr. 1045.) For example,

on some shifts, there were just three employees for a unit. (Tr. 380.) If one of the MHAs was

assigned to observe a single patient, the other two employees had to handle the remaining patients

and tasks. Id. Ordering 1:1 observation helped to minimize risks associated with a single patient

in acute distress, but it left fewer employees to monitor the rest of the patients for agitation or other

signs of potential violence. (Tr. 871, 1010-11.) Each employee was responsible for more patients,


making it more difficult to work on de-escalation. Id.

Maintaining adequate staff for admissions was also a feasible abatement method which the

Secretary showed would reduce the hazard. Respondents’ expert, Ms. Cooke, acknowledged that

new patients “really need a lot of attention and a lot of care.” (Tr. 1467-68.) She described

admissions as “very difficult” and “critical junctures.” Id. In her view, they present “an

opportunity for danger.” (Tr. 1468.) Patients may be harmful to others during the transition into

the facility. Id. Adequate staffing was necessary to address the potential of violence from the

75 Dr. Welch explained that verbal de-escalation requires several people. (Tr. 966, 1010.) It requires enough staff

who can both quickly identify who is escalating and then change tasks so there can be a quick intervention. Id. Even

then, patients may not have sufficient connection with reality for verbal intervention to work. (Tr. 1010-11.) So, a

program needs additional forms of abatement to be effective. (Tr. 1011.) Similarly, Mr. Quinn, the crisis intervention

specialist, indicated that he can spend considerable amounts of time trying to talk to a patient to find reasonable

alternatives to their violent behavior. (Tr. 1250.) However, there are six patient units and typically only one crisis

intervention specialist on one shift for all the units. (Tr. 1250-53.) KS, an MHA, explained that often there were

multiple agitated patients and frequently there were not enough people to de-escalate two or three agitated patients.

(Tr. 871-72.)



arriving patient. Admissions also triggered the need for adequate staff to work with the already

admitted patients to minimize risks and respond quickly if assaultive behavior occurred. (Tr. 947,

963, 966, 1010, 1057-58.)

Dr. Welch opined that there were simply not enough employees to prevent workplace

violence incidents and minimize the consequences of such occurrences. (Tr. 963, 1031, 1047-48.)

Pembroke argues it staffed for acuity and admissions but did not offer documentary evidence

showing a pattern of higher staffing levels to accommodate acuity or anticipated admissions. In

contrast to the management witnesses, direct care employees testified that Pembroke did not

consistently staff for admissions or acuity. (Tr. 132, 839, 863, 871-72, 879.) The number of

admissions varied greatly. On at least one shift during the inspection period 14 patients were

admitted. (Tr. 1008.) KS recalled having seven patients admitted to one unit during a shift. (Tr.

863.) Another MHA explained that admissions took a patient to staff ratio that was “already

stretched too thin” to an even lower level. (Tr. 84.)

Records reveal consistently low staffing levels, and even some occasions where the level

of staff did not meet the DMH minimum requirements. (Exs. C-29, C-47, C-78.) The testimony

from direct care providers about the failure to staff appropriately for admissions and acuity is

credited over the management witnesses on this issue. Management employees could have offered

documentary evidence to support their claims, if it was available, but chose not to do so. See

Capeway, 20 BNA OSHC at 1342-43. Dr. Welch explained that Pembroke could feasibly maintain

sufficient staffing to accommodate changes in patient acuity and new patient admissions. He noted

that another similar facility maintains a much higher patient to staff ratio for newly admitted

patients. (Tr. 1117-18.) The higher ratio allows more staff to safely respond to assaultive behavior

as opposed to Pembroke’s approach of having a single employee in an isolated room handle




searches and other tasks with the new patients. (Tr. 947, 1031, 1118.) The Secretary showed it is

feasible for Pembroke to maintain staffing that is adequate to safely address changes in patient

acuity and new patient admissions.

As for the second part of this abatement method, ending the practice of sending employees

home mid-shift, employees explained that when the ratio of patients to staff exceeded the minimum

requirement set out in the grid, an employee would be sent home before the end of his or her shift.

(Tr. 109, 329, 838-39.) AS and KS both described instances where staff was sent home, and then

a new admission would leave the unit short staffed for the remainder of the shift. (Tr. 109, 838-

39.) According to AS, the other psychiatric facility he was familiar with did not adjust staffing


downward partway through shifts. (Tr. 109.) Pembroke acknowledges that this occurred and

recognizes that the facility “may or may not” be able to add staff after the start of a shift if

admissions increase the census to the point that additional staff is needed either by terms of the

grid or because of acuity. (Tr. 382.) Dr. Welch viewed the practice of mid-shift reductions as

unsafe, particularly given the low baseline level of staffing at Pembroke. (Tr. 1060-61.) The

practice made it more likely that Pembroke would be left unable to respond appropriately to

prevent or mitigate violent situations if admissions or the overall level of acuity changed. Id.

Pembroke had no separate staff focused on admissions, so being understaffed when a new patient

arrived, made it particularly difficult for staff to handle potentially violent patients who may be in


acute distress. (Tr. 1059-61.) The Secretary established that ending the practice of sending

76 The other facility was a state-run hospital. (Tr. 103.) It has fewer units, but the unit size is the same as at Pembroke

Hospital. (Tr. 103-4.)

77 Dr. Welch also explained that when staff was tied up with admissions tasks, other patients have “unmet needs.”

(Tr. 1030-31.) The frustration associated with unmet needs is a well-documented pre-cursor to violent acts by patients.

(Tr. 946-47, 962, 1033; Ex. C-101.) KS explained that this occurred at Pembroke Hospital. (Tr. 871-72.) There were

times when the number of agitated patients exceeded the number of caregivers resulting in staff being unable to timely

address patient agitation. Id.




employees home mid-shift was one way to increase the number of workers available to address


changes in unit acuity and handle new admissions.

(3) Provide security staff and/or crisis intervention specialists

The Secretary proposes security staff “and/or” crisis intervention specialists on all three

shifts. (Sec’y Br. at 4.) Dr. Welch, supported by peer reviewed literature, opined that the presence

of security personnel reduces the rate of assaults. (Tr. 980; Ex. C-96.) Similarly, having

inadequate security is a risk factor for workplace violence. (Tr. 985, 995; Ex. C-97, C-99.)

Pembroke did not have “security” personnel. (Tr. 96.) Instead, all individuals with direct

patient care responsibilities were trained in a de-escalation technique, called Handle with Care.

(Tr. 1189; Exs. R-10, R-32.) Only on the evening shift did Pembroke employ a “crisis intervention

specialist.” (Tr. 1233.) The crisis intervention specialist is not assigned to any particular unit and

does not have direct patient care tasks. (Tr. 964, 1437.) The employee rounds through the units

to identify problems and address them before they escalate. Id. The individual is also responsible

for making sure doors are appropriately locked. (Tr. 964.)

In proposing security or crisis intervention specialists for all shifts, the Secretary does not

insist that Pembroke specifically designate employees as “security”—the proposal permits use of

crisis intervention specialists instead. (Sec’y Br. at 4.) The abatement method calls for there to

be employees without specific direct care tasks focused on safety who could immediately respond

to actual or threatened incidents of violence. The Secretary argues that having such people

available is feasible and would materially reduce the hazard.

Dr. Welch conceded that the crisis intervention specialist on the evening shift was effective.

78 Ending the practice of having employees leave mid-shift was not the only way to maintain sufficient staffing to

manage increases in acuity and/or admissions.



(Tr. 1009, 1051.) However, it was only one employee on one shift. The one employee was

responsible for six units, including making sure multiple doors were locked and secured. (Tr.

1009.) Dr. Welch opined that there should have been enough people available on all three shifts

to ensure safety and to prevent and minimize the hazard. (Tr. 1010, 1051-52.) He explained how

abating the hazard required regular rounds focused on security. (Tr. 995.) And, he emphasized

the importance of having staff dedicated to responding to immediate crises. (Tr. 1052, 1087.)

At Pembroke Hospital, MHAs and nurses had numerous patient care tasks, such as

dispensing medications, running groups, housekeeping, and coordinating care. (Tr. 1087.) Many

tasks could not be instantly stopped to permit the person to respond to a call for assistance. Id. In

addition, MHAs were responsible for routinely taking patients off units for fresh air breaks and

meals in the cafeteria. (Tr. 55-56, 863.) When conducting such tasks, MHAs could not leave

patients alone to respond to requests for assistance. (Tr. 863, 949.)

Dr. Welch also explained that some staff did not feel physically capable of managing very

large, violent patients --- particularly on their own. (Tr. 1087.) In his view, Pembroke needed

employees physically capable of assisting with restraining large patients and who were focused on

safety as opposed to other tasks. (Tr. 1087, 1093-94.) The single person in the role of crisis

intervention specialist on one shift was not adequate for the degree of patient aggression and the

number of workplace violence incidents. (Tr. 1051.) Dr. Welch opined that other behavioral

health facilities were able to have staff dedicated primarily to safety and opined that Pembroke

could feasibly do the same. (Tr. 1053.)

While Ms. Cooke had concerns with maintaining a security staff, she acknowledged the

value of having additional staff available to respond to crises and manage situations. (Tr. 1492,

1495-96.) She noted that managing patient behavior is particularly difficult on the evening shift.




(Tr. 1496.) However, in her view, having only one person on the evening shift as a crisis

intervention specialist was not problematic. As for the lack of someone in such a role on the night

and day shift, Ms. Cooke cited the float position at night and the availability of other staff during

the day. (Tr. 1478.) However, the float could be assigned patient care tasks, including a 1:1

assignment that would prevent him/her from being available to respond to a crisis. (Tr. 374-75,

379, 1437, 1478.)

Importantly, Ms. Cooke did not dispute Dr. Welch’s opinion that having staff focused on

security was feasible. She did not specifically review the studies relied on by Dr. Welch in

reaching his conclusions. (Tr. 1518.) Also, she admitted awareness of other facilities that had

building security. (Tr. 1493-94.) In particular, she acknowledged being familiar with facilities

that have security respond to a code, such as to “be manpower” and “help restrain” patients or de-


escalate situations. (Tr. 1492-93.) Ms. Cooke did not address the two concerns with Pembroke’s

approach cited by Dr. Welch: (1) the lack of anyone assigned only to safety on the day and night

shifts; and (2) the lack of physical capabilities of some of the individuals tasked with controlling

violent patients. Additionally, her testimony failed to rebut the Secretary’s evidence of feasibility

and effectiveness. Altogether, Ms. Cooke’s testimony did not rebut the Secretary’s evidence that

having security staff “and/or” crisis intervention specialists on all three shifts was feasible and

would materially reduce the hazard.

(4) Staffing for Therapeutic Activity Groups and Recreation Periods

The Secretary proposes that Pembroke maintain sufficient staffing for “therapeutic activity

79 Pembroke’s Code Green policy directed employees to contact 911 for “psychiatric emergencies or as deemed

appropriate.” (Ex. R-9 at 3-4.) Ms. Cooke argued that uniformed personnel could increase a patient’s agitation but

did not discuss Pembroke’s policy of contacting outside emergency responders for psychiatric emergencies. (Tr.





groups and recreation periods.” (Sec’y Br. at 4.) Dr. Welch, supported by peer reviewed literature,

explained that canceling therapeutic activities can increase the level of acuity for the unit. (Tr.

969-70; Ex. C-101.) Boredom and insufficient activities are also precipitants for violence. (Tr.

967-68, Ex. C-101.) Structured organized activities, along with the reliability and predictability

of their occurrence promotes positive engagement and reduces acuity. (Tr. 820-21, 949-50, 1062.)

Conversely, unmet needs can trigger violence. (Tr. 946-47, 949-50, 960, 1030; Ex. C-101 at 9.)

If an MHA can engage some patients in an activity, this frees staff to better intervene with sicker

patients. (Tr. 970.) Dr. Welch, informed by the employee statements, noted that the lack of

activities was particularly detrimental for the adolescent units. Id. Teenagers “tend to be very

stimulus prone” and failing to engage them consistently can lead to adverse consequences such as

agitation. (Tr. 965, 970-71.) Engaging patients in therapeutic group activities and recreation

reduces patient agitation and incidents of workplace violence, making it an effective method to


materially reduce the hazard. (Tr. 967-68, Ex. C-101.)

As discussed, therapeutic groups were routinely canceled due to inadequate staffing. (Tr.

819-20.) Dr. Welch explained that increased staffing would allow therapeutic group activities to

occur consistently as scheduled. (Tr. 1062-63.) He noted that another similar facility has a

dedicated staff for conducting therapeutic group activities. (Tr. 1063.) Because the employees

running the groups have no direct care responsibilities, the need to handle admissions or

emergencies did not result in therapeutic activities being canceled. Id. Ms. Cooke recognized that

group activities were regularly canceled but alleged that the number of cancellations was not “an

80 The proposed abatement calls for both group activities and recreation periods. (Sec’y Br. at 4.) MHAs took patients

off the units for fresh air breaks. (Tr. 55-56.) The Secretary failed to show that the few times during the inspection

period when these breaks did not occur materially contributed to the hazard. Pembroke appeared to have sufficiently

implemented the need for fresh air breaks.



overabundance.” (Tr. 1503.) She did not refute Dr. Welch’s view that therapeutic groups

materially reduce the hazard. Nor did she address whether Pembroke could maintain a higher staff

to patient ratio to ensure groups more consistently occurred as scheduled. The Secretary

established that it was feasible to maintain a higher level of staffing which would reduce the

frequency with which groups were canceled. The Secretary also established that maintaining

therapeutic group activities would materially reduce the hazard.

b) Change how admissions occur

The next proposed abatement method is to have an admissions area separate from the on-

going care units for admissions and requiring that new patients remain in the separate area until all

clinical assessments are complete. Specifically, the Secretary proposes that Pembroke:

1. Establish a team of nurses and mental health associates (MHAs) that focuses

primarily on the performance of the tasks associated with the admission and

assessment of new patients. This team should not be staffed by employees

who are assigned to care for already-admitted patients.

2. Dedicate a physical area, apart from all other units of the hospital, to the

admission and assessment of new patients.

3. Do not allow new patients into other units until the tasks associated with

admission, including clinical assessments and the provision of medical

orders, are complete.

(Sec’y Br. at 4.) The first method calls for an admissions team focused solely on the tasks

associated with admitting new patients. The Secretary showed there was an overall need for

additional staff at the facility to reduce the cited hazard. (Tr. 1058.) Carefully screening newly

admitted patients was critical to addressing the hazard. The initial assessments looked at the

patients’ history of violent behavior and sought information on how patients could be de-escalated.

(Exs. R-1, R-9.) As part of this assessment, MHAs searched patient belongings to identify

materials that could harm staff. Id. Most of the admissions process took place in a small room




with the MHA alone with the patient even before the belongings search was completed. (Tr. 81.)

The room had no phone to commence a Code Green and even with a radio it was difficult to call

for help from the room. (Tr. 82.)

The record reveals that admissions tasks are time consuming and that there is an increased

risk for an employee to experience violence during this time. (Tr. 1006-7, 1011-13.) The overall

level of staffing was often too low to handle the acuity of existing patients and more than a few

admissions. Admissions took time away from existing patients and sometimes introduced an acute

patient into the unit population before he/she had adequate medication. (Tr. 1030-31.)

In those instances where there was enough staff to adequately handle the admissions tasks

and care for existing patients, the Secretary did not adequately explain how precluding employees

who handle admissions from other care responsibilities would materially reduce the hazard. In

other words, the Secretary showed that Pembroke failed to maintain adequate staffing levels but

he did not show how implementing the additional abatement method of precluding admissions

staff from having other care responsibilities would materially reduce the hazard if staffing was

already adequate to handle admissions. Unlike other facilities, Pembroke Hospital does not accept

“walk-ins,” i.e. patients that have not yet been assessed by someone outside of the facility. (Ex.

R-9.) The information obtained during these pre-admissions calls is often incomplete. Still, the

facility had some information about the patients and sometimes the patients were stabilized before


their arrival at Pembroke. (Tr. 1468-69; Ex. R-9.) Dr. Welch indicated that admissions could

81 When Ms. Cooke was discussing admissions, she was not always clear about whether she was describing the

situation at Pembroke or her experience at other facilities. (Tr. 1469.) For example, she indicated that at “most” in-

patient facilities she was aware of, or worked act, they would not accept “highly psychiatrically unstable” patients.

(Tr. 1468.) She did not clarify whether that was true of Pembroke. (Tr. 1468-69.) Further, when describing the

threats associated with newly admitted patients, she said that “we mitigate the risk.” (Tr. 1468.) When asked by

counsel whether she meant “Pembroke” when she used the term “we,” Ms. Cooke explained that she meant Pembroke

or was using the term “we” in the sense of the “behavioral health world.” (Tr. 1469.)



take as many as twenty-two hours of nursing time per day. (Tr. 1008.) On average, however,

much less time was necessary for admissions at Pembroke. To reach twenty-two hours of nursing

time, there would have to be fifteen admissions, and each would have to take ninety minutes. (Tr.

1008.) While some admissions took the full ninety minutes and sometimes there were many

admissions, the record indicates that in 2016, on average, Pembroke had approximately seven

admissions per day. (Tr. 1436.)

Regarding the admission and assessment of new patients in a dedicated physical area, Dr.

Welch opined that a separate admissions unit would materially reduce the risk of workplace

violence. (Tr. 1028-30.) In his view, the existing spaces on the units used for some of the

admissions tasks were unsafe because they were isolated and small. (Tr. 1031.) Further, patients

went directly from the small rooms into the common area with existing patients. (Tr. 1033.) The

newly admitted patients find themselves in uncertain situations, may be frustrated, and could lack

proper medication. (Tr. 1032-34.) Dr. Welch explained that Pembroke’s approach creates a

“destabilizing situation.” (Tr. 1034.) AS and KS also asserted that a separate unit would be helpful

to abate the hazard. (Tr. 83-84, 861-65.) AS explained that while a separate unit would not change

patient behavior it would eliminate “a big risk factor” in terms of the new patient interacting with

existing patients. (Tr. 164.) Similarly, Ms. Wollner, the Director of Admissions, acknowledged

that a separate admissions unit would reduce the burden on the units. (Tr. 1386-87.)

In terms of technical feasibility, the facility had space apart from the units that could be

used for admissions. (Tr. 93, 1034.) Other similar facilities also had dedicated admissions areas.

(Tr. 104-5.) Dr. Welch discussed other similar facilities at which the assessment, evaluation, and

medication of new patients occur in dedicated areas apart from where existing patients are located.

(Tr. 936, 1027, 1031-32, 1034, 1101.) The Secretary presented evidence that it was



technologically feasible to dedicate a physical area apart from the other units to handle assessments

before patients are moved into the treatment units.

However, the Secretary falls short on persuading how moving admissions from the units

to this separate area would materially reduce the hazard if Pembroke was adequately staffed to

handle such admissions and addressed the difficulties with summoning assistance. Again, the

Secretary showed that Pembroke failed to consistently staff in anticipation of admissions and

addressing this deficiency would materially reduce the hazard. Rather than being a separate

abatement method, the proposal to have a separate admissions area appears to be one approach to

handling the inadequate staffing and insufficient means to summon assistance. In other words, if

Pembroke had enough employees on the units and employees could quickly summon assistance,

then a separate admissions unit might not be necessary to further reduce the hazard. See Cerro

Metal Prods. Div. 12 BNA OSHC 1821, 1822-23 (No. 78-5159, 1986) (noting that the abatement

actions must be “necessary and valuable”).

As for Pembroke’s practice of allowing new patients into the units before all assessments

and the provision of medical orders are complete, the evidence reveals that this increased acuity

and created workplace violence risks. The evidence of record also reveals that increasing staff to

appropriately address acuity and 1:1 assignments would materially reduce the risk. The Secretary

established that waiting to integrate new patients into the units was one way to handle the acuity

new admissions bring. However, if the facility had adequate staff and means to summon assistance

it is not clear that segregating the patients would further reduce the hazard. See Cerro, 12 BNA

OSHC at 1822-23.

The Secretary proved that establishing a team of nurses and MHAs focused primarily on

the performance of the tasks associated with the admission and assessment of new patients, was



feasible and would materially reduce the hazard. He failed to show that if the other identified

abatement was in place, (1) that the employees on this team could not have any other

responsibilities; (2) that admissions had to occur in an area apart from all other units of Pembroke

Hospital, or (3) that patients had to remain segregated until all of the tasks associated with

admission were complete.

c) Equipment

Finally, the Secretary calls for two types of equipment: personal panic alarms and

equipment to implement patient crisis prevention plans. (Sec’y Br. at 4.)

(1) Personal Panic Alarms

The Secretary proposes that Pembroke provide “personal panic alarms for all employees

who may work in close proximity to patients, including but not limited to nurses, MHAs,

housekeeping staff, and case workers.” Id. Pembroke’s approach to summoning help in the event

of a workplace violence incident was ineffective. JS, a nurse, discussed her experience with

personal panic alarms at two other behavioral health hospitals. (Tr. 241.) She found them to be

an effective way to seek immediate assistance. (Tr. 241-42.) Dr. Welch explained that personal

panic alarms are more effective at summoning assistance than Pembroke’s current approach, and

that providing them would reduce the likelihood of staff members becoming victims of workplace

violence. (Tr. 937, 974-75.) Unlike radios and the overhead paging system, panic alarms allow

for discreet requests during emergencies. (Tr. 1049.) Dr. Welch relied on studies which found

that access to panic alarms correlated with significantly lower rates of assault against staff in a

healthcare setting. (Tr. 972-74, 995, Ex. C-95 at 3-10, C-96.) Workers without access to such

equipment were victims of workplace violence more often. (Ex. C-95 at 10.) Regarding

feasibility, other similar facilities provide personal panic alarms and literature discusses their





successful use in healthcare settings. (Tr. 974, 1049; Ex. C-95.)

Ms. Cooke argued that there was no evidence to support the efficacy of panic alarms in

reducing risk. (Tr. 1481.) It is unclear if she reviewed the studies Dr. Welch cited in reaching his

conclusion that panic alarms are effective at materially reducing the hazard. (Tr. 1458.) She only

indicated that she “might’ve (sic)” re-looked at the Sentinel Alert. Id. The Sentinel Alert describes

physical and verbal violence against health care workers. (Ex. C-99 at 1-2.) It cites, among other

things, lack of access to emergency communication such as a “call bell” as a contributing factor to


violence against healthcare workers. (Ex. C-99 at 3, 6.) Ms. Cooke acknowledged there were

times when panic alarms may be easier, but said they were not “standard.” (Tr. 1485-86, 1521.)

The general duty clause requires employers to “take all feasible steps” to protect against

recognized hazards. Gen. Dynamics, 599 F.2d at 464. Employers must implement abatement

measures safety experts agree are feasible even if they are “not of general usage in the industry.”

599 F.2d at 464. See also Beverly, 19 BNA OSHC at 1191 (finding abatement measures may be

required even if the practice is not yet customary in the industry). Ms. Cooke did not testify that

panic alarms were technologically or economically infeasible and agreed they may be an easier

way to obtain assistance with a violent patient. (Tr. 1483.) No witness testified they were too

expensive or would not work at Pembroke for technological reasons. Ms. Cooke’s testimony does

not refute the evidence establishing that personal panic alarms would materially reduce the hazard

and constitute a feasible means of abatement.

82 One peer reviewed study Dr. Welch relied on, Simha F. Landau, Personnel Exposure to Violence in Hospital

Emergency Wards: A Routine Activity Approach, 34 Aggressive Behavior 88, 91 (2008), described access to an

“emergency button” as a “widely used” and “important physical protective device.” (Ex. C-95.)

83 The Sentinel Alert was issued after the inspection occurred and was not considered as evidence of Respondents’

knowledge of the hazard. (Ex. C-99.) However, it was reviewed by Dr. Welch and possibly by Ms. Cooke to assess

the feasibility of the proposed abatement.



The Secretary’s burden is to establish that the abatement in place, a limited number of

unreliable radios, was ineffective and that this issue could be addressed feasibly. Integra, 27 BNA

OSHC at 1849; Armstrong, 8 BNA OSHC at 1073-74 (upholding a violation of the general duty

clause when the employer failed to maintain an abatement measure it implemented). Dr. Welch’s

opinion that personal panic alarms are feasible and effective at reducing the likelihood a worker

would be injured by the hazard was well supported by the testimony of other witnesses and

numerous studies. (Tr. 241, 309-10; Exs. C-95, C-96, C-99.) The Secretary established that

personal panic alarms are a feasible and effective means to abate the hazard by making it easier

for additional people to respond to violent incidents or threats of violence.

(2) Equipment for Crisis Intervention Plans

The final abatement method proposed calls for Pembroke to “maintain equipment that is

sufficient for the implementation of each patient’s crisis prevention plan.” (Sec’y Br. at 4.)

Pembroke asks patients to complete forms about what helps to calm them down and what helps

them to cope with situations. (Tr. 512, 1423, 1432, 1439; Ex. R-18.) Some patients decline to

complete it. (Tr. 512, 1432.) However, staff can add information to the plans even if a patient

does not provide any. (Tr. 512-13, 1249, 1431-32.) The plans assist staff with knowing how to

deescalate a patient who becomes aggressive. (Tr. 512-13, 1432.) Most of the plans call for the

use of music to de-escalate patients. (Tr. 73, 830, 1434.)

Dr. Welch opined that having music playing equipment would significantly reduce the risk

of workplace violence and indicated that the equipment necessary for patients to have access to

music could be affordably obtained. (Tr. 1064.) He explained that patients most often cite music

as a calming technique, and he had seen it work effectively. (Tr. 962.) Dr. Welch asserted that

providing access to calming modalities would decrease the hazard. (Tr. 1064-65.) In his




experience, the headsets for playing music could be affordably maintained to allow them to be

given out by request, rather than having waiting lists as was the case at Pembroke during the

inspection period. (Tr. 963.) Employees agreed that providing music worked effectively at

Pembroke to soothe and calm aggressive patients. (Tr. 73-74, 666, 831.) A nurse considered it a

great help in handling patients who were distressed or agitated. (Tr. 666.) Ms. Cooke did not

dispute the evidence regarding the usefulness of crisis prevention plans or providing music when

patients indicate it has helped calm them in the past. Rather, her testimony revealed that she was

unaware of any studies that proposed a specific number of devices based on the patient count. (Tr.

1505-6.) The Secretary established that being able to implement patient crisis prevention plans by

providing music to calm distressed patients would materially reduce the hazard.

As for technical feasibility of implementing the abatement, by the time of the hearing,

Pembroke had improved the ability of caregivers to routinely provide music consistent with de-

escalation plans. Fed. R. Evid. 407 (evidence of subsequent remedial measures may be admitted

to show “the feasibility of precautionary measures”); SeaWorld, 748 F.3d at 1215. Accordingly,

there is no dispute that providing music as part of implementing patient crisis prevention plans


was technologically and economically feasible.

IV. Relationship of UHS DE and Pembroke

The Secretary contends that despite being separate corporate entities, UHS DE and

Pembroke should be treated as a single employer of the workers exposed to the hazard. Related

employers are treated as a single entity when “they share a common worksite, have interrelated

84 To the extent that the proposed abatement could be interpreted as requiring other types of equipment, the Secretary

failed to sufficiently identify such equipment or establish how it would materially abate the hazard.




and integrated operations, and share a common president, management, supervision, or


ownership.” Loretto-Oswego Residential Health Care Facility, 23 BNA OSHC 1356, 1359 (No.

02-1174, 2011) aff’d, 692 F.3d 65 (2d Cir. 2012); C.T. Taylor Co., 20 BNA OSHC 1083, 1086

(No. 94-3241, 2003) (consolidated); Trinity Indus., Inc., 9 BNA OSHC 1515, 1518 (No. 77-3909,

1981). The Secretary fails to satisfy these criteria.

Pembroke and UHS DE have different worksites: UHS DE operates out of Pennsylvania,

while Pembroke operates in Massachusetts. (Tr. 724.) As the First Circuit explained in A.C.

Castle, the common worksite element requires consideration of the location of where the

employees work and are exposed to the hazard. 882 F.3d at 42. The Secretary does not allege that

any UHS DE employees were exposed to the cited hazard. Nor does he allege that employees

were exposed to the hazard of workplace violence at locations other than Pembroke Hospital. The

record is clear that UHS DE and Pembroke have different business addresses and there is no

allegation of workplace hazards at both locations.

With regard to operations, Gary Gilberti, the Senior Vice President in the Behavioral

Health Division for UHS DE, testified that UHS DE and Pembroke were related, but distinct

businesses. (Tr. 724.) Pembroke provides direct patient care, while UHS DE is a management


and consulting business. Id. Pembroke primarily develops and sets its own budgets. (Tr. 643-

44, 1326.) See Loretto-Oswego, 23 BNA OSHC at 1359 (finding healthcare facility and

management company were not a single employer). There is no evidence that Pembroke lacked

sufficient capital or other resources to address worker health and safety.

85 The parties agree that this is the appropriate test to assess whether two legal entities functioned as a single employer.

(Am. Pre-Hr’g Stmt. at 16; Resp’t Br. at 6.)

86 Respondents assert that Pembroke set its own budget. (Resp’t Br. at 42.) However, Ms. Gilmore, a UHS DE

employee testified that she handled the workers’ compensation budget for Pembroke Hospital. (Tr. 550; Stip. 10.)



As for personnel, Pembroke Hospital’s CEO (Dr. Hickey) was employed by UHS DE. (Tr.

695; Stips. 10-11.) Other UHS DE employees routinely visited the Pembroke Hospital but did not

provide daily oversight of Pembroke Hospital. (Tr. 533-34.) UHS DE had its own management

structure with a separate CEO, chief financial officer, and management group. (Tr. 741.) Dr.

Hickey had no managerial duties for UHS DE. (Tr. 741.) Cf. C.T. Taylor, 20 BNA OSHC at 1085

(finding two entities to be a single employer when both companies were owned and controlled by

the same person and operated from the same office). The Secretary failed to establish that UHS

DE and Pembroke operate as a single entity. Accordingly, the finding of a violation and

assessment of a penalty are against UHS of Westwood Pembroke only. The Secretary’s allegation

against UHS DE is DISMISSED.

V. Repeat Characterization

Respondents argue that if UHS DE and Pembroke are not a single employer, then the

citation cannot be characterized as repeat. (Resp’t Br. at 42.) This misconstrues the record

evidence of Respondents’ own corporate structure. The Secretary does not rely on a citation issued

to UHS DE to support the repeat characterization. Instead, he relies on one issued to Pembroke

itself. (Exs. C-14 thru C-17.)

A violation is repeated if the same employer was previously cited for a substantially similar

violation. Potlach Corp., 7 BNA OSHC 1061, 1063 (No. 16183, 1979); Lake Erie Constr. Co.,

21 BNA OSHC 1285, 1289 (No. 02-0520, 2005) (similarity of hazards is a “principle factor” in

assessing the appropriateness of a repeat characterization). A violation of the general duty clause

may be “found to be repeated on the basis of either a prior section 5(a)(1) or section 5(a)(2)

violation.” 7 BNA OSHC at 1064. When relying on previous general duty citation, the Secretary

must show substantial similarity based on the circumstances surrounding the hazard. GEM Indus.,



Inc., 17 BNA OSHC 1861, 1865-66 (No. 93-1122, 1996) (declining to rely on a previous 5(a)(1)

citation to support characterizing a subsequent violation of a specific standard as repeat) aff’d, 149

F.3d 1183 (6th Cir. 1998).

The Secretary argues that the hazard cited in 2016 at Pembroke’s Lowell Treatment Center

and the one cited in 2017 at Pembroke Hospital are substantially similar. (Sec’y Br. at 74.) But,

the Secretary failed to offer enough information about the circumstances surrounding the cited

hazard at Lowell Treatment Center. The limited information in the record reveals only that the

cited workplaces had significant differences related to the hazard. For example, while the

Secretary proposed some similar abatement measures, others were notably different. (Ex. G-14.)

The undersigned finds that while the hazards appear to share some commonality the record does

not establish that the two violations are sufficiently similar to support a repeat characterization.

GEM, 17 BNA OSHC at 1866 (declining to conclude that two violations were substantially similar

even though both involved fall hazards).

The Secretary has met his burden to establish, by a preponderance of the evidence, that

UHS of Westwood Pembroke violated 29 U.S.C. § 654(a)(1) of the Act and such violation was

“serious” rather than “repeat”.

VI. Penalty

“Section 17(j) of the Act, 29 U.S.C. § 666(j), requires that when assessing penalties, the

Commission must give due consideration to four criteria: the size of the employer's business, the

gravity of the violation, the employer’s good faith, and any prior history of violations.” Hern Iron

Works, Inc., 16 BNA OSHC 1619, 1624 (No. 88-1962, 1994). When determining gravity, the

Commission considers the number of exposed employees, the duration of their exposure, whether

precautions could have been taken against injury, and the likelihood of injury. Capform, Inc., 19



BNA OSHC 1374, 1378 (No. 99-0322, 2001), aff’d, 34 F.App’x 85 (5th Cir. 2000

(unpublished). Gravity is typically the most important factor for determining the penalty. Id.

When initially issued, the Citation included a proposed penalty of $12,675. In the

Amended Complaint, the Secretary increased the proposed penalty to $25,350. This amount could

only be imposed if the violation was characterized as repeat rather than serious. Under the Act, as

amended by the Federal Civil Penalties Inflation Adjustment Act Improvements Act of 2015,

Public Law 114-74, sec. 701, the maximum penalty for a violation cited in 2017 and characterized

as “serious” was $12,675.00.

Having considered the four factors, the undersigned finds that a penalty of $12,675 is

appropriate. The hazard caused serious injury and was capable of causing death. Many employees

were exposed to the hazard, with several suffering serious injuries. (Ex. C-60.) There is no

evidence that Pembroke is a small employer. While it took some steps to mitigate the hazard, it

failed to implement feasible abatement measures even after they were identified. In terms of

history, it was previously cited for a violation related to workplace violence. (Ex. G-14.) Although

the prior citation is not enough to support a repeat characterization, it is relevant to the evaluation

of Pembroke’s history.

As to good faith, as noted above, employees felt pressure to be less than forthcoming with

investigators. (Ex. C-1; Tr. 883-84, 886-89.) Also, it is worth noting that MHAs were provided

with a “cheat sheet” on how to communicate with regulators and then told to destroy it after

reviewing it. (Tr. 125-26.) However, there is no evidence of actual obstruction. The MHA AS

explained that they were to be direct and honest with surveyors, including OSHA. (Tr. 181-82.)

Further, Respondents had a safety program and, as noted, had taken steps to minimize the hazard.





On this record, neither an increase nor a decrease for good faith is warranted.


The foregoing constitutes the findings of fact and conclusions of law in accordance with

Rule 52(a) of the Federal Rules of Civil Procedure.


Based upon the foregoing findings of fact and conclusions of law, it is ORDERED that:

Citation 1, Item 1 for a violation of section 5(a)(1) of the Act is AFFIRMED as

SERIOUS, and a penalty of $12,675 is ASSESSED against UHS of Westwood




Keith E. Bell

Administrative Law Judge, OSHRC

Dated: February 19, 2020

Washington, D.C.

87 Respondents limit their argument on the penalty amount to asserting that if the violation is affirmed it should not be

characterized as repeat. (Resp’t Br. at 43.)