Some personal identifiers have been redacted for privacy purposes.

United States of America



















OSHRC Docket No.: 19-1579






Alicia A. Truman, Esq. and Beau Ellis, Esq., Department of Labor, Office of the Solicitor, Denver, CO

For Complainant


Melanie Paul, Esq. and Dion Kohler, Esq., Jackson Lewis, P.C., Atlanta, GA

For Respondent


Before: First Judge Patrick B. Augustine – U. S. Administrative Law Judge



Workplace violence is a unique hazard in the occupational safety and health arena. It is not governed by a specific standard, and many businesses do not have occasion to address it, or only need do so in the most cursory terms. But, in some industries, it is one of the most significant threats to the safety and health of employees covered by the Occupational Safety and Health Act, 29 U.S.C. § 651 et seq. (Act). The hazard is particularly acute in the healthcare industry, where, according to the Bureau of Labor Statistics, U.S. healthcare workers suffered 15,000 to 20,000 workplace-violence-related injuries every year from 2011-2013 and accounted “for nearly as many injuries as all other industries combined.”1 (Ex. C-13 at 3; C-12 at 2-3). The particular threat of violence faced by a healthcare facility depends on the type of services it provides, the patients it sees, and whether it is open to the public. (Ex. C-12)

Inpatient psychiatric hospitals, such as the one Respondent operates in Louisville, Colorado, deal almost exclusively with what is known as Type 2 violence, or “violence committed by patients upon staff members.” (Ex. C-14). This is due, in no small part, to the population Respondent serves. According to Respondent, many of the patients it admits come involuntarily from local emergency departments.2 These patients are typically at a low point in their psychiatric health and need assistance for any number of psychiatric or behavioral conditions, including depression, bipolar disorder, schizophrenia and other schizo-affective disorders. Although data and Respondent’s own experience show many of these patients are not a threat to the safety and health of its employees, some patients, especially those with a history of violent behavior, will perpetrate violence on staff members responsible for providing them with care. In light of this hazard, which Respondent admits is endemic to its workplace and the industry generally, the operative issues in this case are whether (i) Respondent adequately addressed the hazard of workplace violence; and (ii) Complainant established the feasibility of its proposed abatement measures.

Complainant cited Respondent for failing to provide employment and a place of employment free from the recognized hazard of workplace violence under the general duty clause, 29 U.S.C. § 654(a)(1). Complainant identifies multiple aspects of Respondent’s current workplace violence protection plan (WVPP) that are deficient and makes an equivalent number of proposals to abate the workplace violence hazard. While the case is complex in its particulars, at bottom Complainant is alleging Respondent failed to address the hazard of workplace violence in a comprehensive and effective way. Respondent, on the other hand, contends Complainant has not shown Respondent’s methods for addressing the workplace violence hazard were inadequate, nor has he provided any competent evidence the abatement methods he proposes would materially reduce the workplace violence hazard. In particular, Respondent contends the source of workplace violence hazard itself presents multiple difficulties, such as predicting who will perpetrate violence and when, as well as assessing the effectiveness of particular abatement methods. As such, Respondent places significant emphasis on its clinical approach to workplace violence, which it contends is outside the reach of the Occupational Safety and Health Administration (OSHA).

Respondent’s task is a difficult one. It must account for the difficulties inherent to providing a safe environment for psychologically and behaviorally compromised patients to heal while simultaneously ensuring its employees are provided with the most effective means to mitigate potential workplace violence. While this is no doubt a difficult task in an environment where workplace violence is endemic, Respondent is nonetheless obliged to take all necessary and appropriate steps to ensure its employees are provided with “employment and a place of employment which are free from recognized hazards . . . .” 29 U.S.C. § 654(a)(1). Ultimately, the Court finds Complainant has established that Respondent failed to adequately address the hazard of workplace violence in multiple areas and, simultaneously, proved many of its abatement proposals would be effective in reducing the hazard.” Arcadian Corp., 20 BNA OSHC 2001, 2011 (No. 93-0628, 2004). Respondent’s failures can be resolved through the intentional implementation of a comprehensive WVPP, which implements Complainant’s abatement proposals insofar as the Court determines they are feasible.

II.Jurisdiction & Stipulations

A. Stipulations

The parties reached multiple stipulations, which they submitted to the Court prior to the trial in this matter. Those stipulations, which the Court has accepted, provide as follows:

i.Respondent UHS of Centennial Peaks LLC dba Centennial Peaks Hospital (“Centennial Peaks”) is an employer 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. § 652(5).

ii.Centennial Peaks is an in-patient psychiatric hospital.

iii.Centennial Peaks has 5 units with a total of 104 beds.

iv.The Citation and Notification of Penalty underlying this proceeding was issued on May 31, 2019.

v.Respondent timely filed its Notice of Contest on June 4, 2019.

vi.Employees at the worksite are exposed to the hazard of workplace violence, defined in this case as physical threats and assaults by patients toward staff.

vii.The hazard of workplace violence as defined in this case was recognized by Centennial Peaks at the time of the alleged violation.

viii.The hazard of workplace violence is recognized in the industry.

ix.Staff injuries from patient violence and/or assaults by patients against staff could result in serious injuries.

x.The Occupational Safety and Health Review Commission has jurisdiction in this proceeding pursuant to § 10(c) of the Occupational Safety and Health Act (“OSH Act”).

xi.Nurse Manager/House Supervisor Shift Reports were produced by Respondent during discovery for 13 days in 2017, 65 days in 2018, and 57 days in 2019. (J-4)

See Joint Stipulations at 1-2.3


[Remainder of the page intentionally left blank]



Based on the stipulations4, the Court finds Respondent is an employer engaged in interstate commerce and subject to the jurisdiction of the Act. United States v. Lopez, 514 U.S. 549, 558 (1995). See also Slingluff v. OSHRC, 425 F.3d 861 (10th Cir. 2005). The Court also finds the Commission has jurisdiction over the matter pursuant to § 10(c) of the Act. (JS No. 1). See Joel Yandel, 18 BNA OSHC 1623, 1628 n.8 (No. 94-3080, 1999).

III.Procedural History

This case began when one of Respondent’s employees filed an anonymous complaint with OSHA’s Denver Area Office, alleging employees were routinely exposed to physical assaults by patients and Respondent did not have an effective WVPP in place to address those assaults. On receipt of the complaint, Complainant designated Compliance Safety and Health Officer (CSHO) Brian Oberbeck to perform an inspection into the allegations, which began on December 7, 2018. Based on the results of his inspection, CSHO Oberbeck recommended, and Complainant issued, a single-item Citation and Notification of Penalty, which alleged a violation of the general duty clause (§ 5(a)(1) of the Act) and proposed a penalty of $10,229. Complainant filed a timely Notice of Contest, bringing the matter before the Commission.

The parties engaged in extensive motions practice during the pendency of this litigation. Of particular importance to this case are two orders issued by the Court during the course of discovery: (1) Order on Motion to Amend Citation and Complaint, dated Sept. 1, 2020; and (2) Order on Complainant’s Motion for Sanctions and Entry of Sanctions, dated April 19, 2021.

Regarding Complainant’s Motion to Amend the Citation and Complaint, Complainant sought to amend the Citation and Complaint to add UHS-Delaware as a party, based on Complainant’s assertion that UHS-Centennial Peaks and UHS-Delaware were a single entity for the purposes of establishing liability for the violation. Ultimately, the Court determined Complainant failed to properly and timely amend the Citation and Complaint pursuant to Federal Rule of Civil Procedure 15(c)(1)(C) and UHS-Delaware would be prejudiced by an amendment at that stage in the proceedings. See Order on Complainant’s Motion to Amend Citation and Complaint, Docket No. 19-1579 (September 1, 2020).5

As to Complainant’s Motion for Sanctions and Entry of Sanctions, the Court found Respondent failed to comply with its duty to preserve evidence once it was placed on notice of pending litigation. The Court imposed adverse inferences that the unproduced evidence would have shown: (i) a hazard of workplace violence was present in the workplace; (ii) Respondent recognized the presence of the workplace violence hazard; (iii) Respondent had actual knowledge of the workplace violence hazard and was aware its employees were exposed to the recognized hazard of workplace violence; and (iv) such exposure could result in seriously bodily harm or death. See Order on Motion for Sanctions and Entry of Sanctions, Docket No. 19-1579 (April 19, 2021).

The only remaining issue for trial was whether Complainant established a feasible means of abatement. On that topic, however, the parties were limited to a defined time period and were precluded from inquiring into clinical/medical decisions, such as medication levels. (Tr. 194, 1774). At trial, the Court ruled on a motion to establish the timeframe for the admission of relevant evidence. The Court established the period of December 17, 2017, through May 30, 2020, as the relevant period. Id.

After a period of delays owing to the COVID-19 pandemic, a trial in this matter was held over the course of nine days, starting on September 13, 2021, in Denver, Colorado. Both parties submitted post-trial briefs for the Court’s consideration. In addition, on June 13, 2022, the Court ordered supplemental briefing to address the Commission’s March 3, 2022, Decision and Order in UHS of Westwood Pembroke, Inc. and UHS of Delaware, Inc., No. 17-0727, 2022 WL 774272 (O.S.H.R.C., Mar. 3, 2022) appeal docketed, No. 22-1845 (3d Cir. May 2, 2022). The Commission’s decision was issued after the trial in this matter had concluded and post-trial briefs had been filed with the Court but before this Decision and Order was issued. The parties submitted their Supplemental Briefs on this issue certified to the parties6. Based on the evidence presented at trial, the Court issues the following findings of fact and conclusions of law.

IV.Factual Background

A.The Facility and Staff

Respondent runs a stand-alone, behavioral health facility, which provides acute, inpatient psychiatric treatment for adults and adolescents. As noted earlier, psychiatric patients come to Respondent’s facility, often involuntarily,7 for treatment and management of conditions such as depression, anxiety, psychosis, schizophrenia, bipolar disorder, and suicidal thoughts. (Tr. 349). According to Respondent, it admits nearly 4,000 patients each year, which are treated and cared for by a staff of approximately 250 staff members. (Tr. 1786, 2379). Of particular interest to this case are the members of the hospital’s direct care staff, which includes doctors, nurses, mental health technicians (MHTs), social workers, and therapists. (Tr. 94).

The doctors, nurses, therapists, and social workers make up the treatment team, which is responsible for the clinical treatment of patients. (Tr. 1713-1714). The treatment team generates an assessment of each patient and is responsible for their medical treatment and care during their stay at Centennial Peaks. (Tr. 233). MHTs work alongside the nurses and conduct patient observation on a set schedule (observation rounds); observing and managing the milieu;8 transporting patients to meals, recreation, and group sessions; and facilitating group and recreational therapy. (Tr. 434). Although doctors have occasional, direct contact with patients, nurses and MHTs work with and around patients throughout their entire 8-hour work shift.

The hospital facility contains 104 beds that are split into five inpatient care units: Castle, Torrey’s, Sunlight, Pike’s Peak, and Crestone Peak.9 (JS No. 3). Castle treats patients for chemical dependency, Torrey’s treats adult patients with severe anxiety or depression and represent potential suicide risks, and Sunlight treats adolescent patients. (Tr. 1830, 2090, 2301). The remaining two units, Crestone (female) and Pike (male), are known as intensive treatment units, or ITUs, where patients with more severe mental disorders were treated. (Tr. 94, 1119). The ITUs became a principal focus of CSHO Oberbeck’s inspection because most of the workplace violence incidents occurred in those units or involved patients from those units. (Tr. 2093).

All five units each have patient rooms, a nursing station, common areas, a medication room, and a seclusion/quiet room for patients. (Tr. 96). Administrators and executives are located throughout the building, adjacent to and within the units. (Tr. 1766, 1770, 2120-2121; Ex. C-6). According to Respondent, each nurses station is centrally located on the unit to provide a clear line of sight and was designed to permit easy communication between staff and patients. (Tr. 2123; Ex. C-5). Patients come to the nurses station for many reasons other than medical care, including mundane, day-to-day concerns like cups of coffee and telephone calls. (Tr. 269, 377; Ex. C-5). As the photographs and testimony illustrate, the nurse’s station is not enclosed, has relatively short, narrow countertops, and does not have a door to prevent access, which Respondent contends is by design. (Tr. 97; Ex. C-5). Multiple employees identified the nurse’s station as the location for a cluster of violent incidents, including jumping over the counter, grabbing/ripping the computer monitor from the desktop behind the counter, and physical attacks by patients on staff. (Tr. 129, 583-85, 601-604 1634-36; Ex. C-7 at 12, C-19 at 13-16, Ex. C-72).

B.Intake and Admissions

As noted above, many of the patients that come to Respondent’s facility are brought there involuntarily through a referral from a local emergency department (ED) or from the police. Under ideal circumstances, a patient is referred from an ED, and a nurse from Centennial Peaks will communicate with the referring ED nurse prior to the patient’s arrival. (Tr. 1710). This discussion includes information about the patient’s situation, background, treatment, and recommendations, which the nurses referred to as an SBAR. (Tr. 1710-1712).

Upon arriving at Centennial Peaks, the prospective patient is separated from their personal belongings to ensure contraband is not brought onto the premises, including drugs, weapons, or other items the patient could use to harm his/herself or others. (Tr. 2398; Ex. R-10). After the patient is screened for contraband, he/she meets with intake staff, who are trained social workers, in one of the four designated intake rooms. (Tr. 95-96). An intake staff member, often by themselves, typically performs an initial assessment of the patient to determine whether they meet the admission criteria and whether Respondent can provide care that meets the patient’s needs.10 (Tr. 96, 1703-1711; Ex. R-87). The assessment includes a review of a patient’s legal history, previous hospitalizations, diagnoses, medications, and other pertinent medical history, as well as an assessment of the patient’s risk of self-harm or harming others. (Id.). If a determination is made a particular patient is high risk for assaultive behavior, a High-Risk Notification form is supposed to be placed in the patient’s file; however, as will be discussed later, this did not always happen. (Tr. 597, 660, 1654; Ex. C-19, C-23, C-20, R-79). The assessment and other information contained in the file is transmitted to the treatment team and to other treating staff members on subsequent shifts. (Tr. 95-96; Ex. R-87). Though this is the typical process, administrators and staff testified there are occasions when a patient will be directly admitted to the unit to provide immediate medical care. (Tr. 1059, 1719).

C.Day-to-Day Operations

The workday at Centennial Peaks is split into three shifts of eight hours per shift. Daytime shift runs from 7:00 a.m. until 3:30 p.m.; night shift runs from 3:00 p.m. until 11:30 p.m.; and overnight shift runs from 11:00 p.m. until 7:30 a.m. (Tr. 1731). The shifts are designed with a 30-minute overlap, which permits the preceding shift to conduct a handoff report with the oncoming shift. (Tr. 167-68, 717, 1731). During the handoff, the oncoming and outgoing MHTs perform rounds together and discuss the status of each patient, incidents of aggression, and unit acuity.11 (Tr. 1731-32). According to Respondent, the information contained in the handoff report is consolidated into the House Supervisor/Nurse Manager report, which takes a broader view of the events occurring at the hospital as a whole, including patient census, staffing, discharge/admission, and whether a Code had to be called. (Ex. R-54; JS No. 4).

During a shift, direct care staff have multiple duties, including performing rounds, handing out medicine, taking patients to group meetings, facilitating recreational therapy, and taking part in meetings. Typically, MHTs perform rounds every 15 minutes to observe patients; however, doctors can issue special orders which increase the frequency of the rounds down to constant supervision, also known as one-to-ones.12 (Tr. 1728-1730; Ex. C-20, R-82). According to Respondent, special orders are typically issued for patients at suicide risk. (Ex. R-82). As for recreational therapy and lunch, a single MHT is typically assigned to transfer the patients in their unit to lunch, the gym, or outside courtyard. (Tr. 435-436).13

In addition to work performed directly with patients, some direct care staff participate in daily and weekly meetings to discuss patient care. Specifically, nurses, doctors, and social workers take part in daily treatment team meetings to discuss the status of each patient, their progress, as well as problems or issues, including those related to aggression. (Tr. 681-682). MHTs do not participate in treatment team meetings, nor, for that matter, were they required or expected to review patient case files. (Tr. 542, 1615, 1825).

Administrators also held meetings to discuss various matters impacting patients and staff, these included the Performance Improvement Committee (PIC), Patient Safety Council (PSC), flash meetings, and leadership meetings. According to Respondent, flash meetings occur every weekday, during which the Chief Executive Officer (CEO), Chief Nursing Officer (CNO), nurse managers, milieu managers,14 and Risk Manager sit down to discuss acts of patient aggression, as well as patients who, themselves, may pose a risk of assault. (Tr. 1748). According to Respondent’s WVPP, the PIC oversees the development and review of “several different [active] performance improvement initiatives, all with a focus on patient care.” (Ex. C-14 at 2). One example of the PIC development and review process was the Seclusion and Restraint Reduction Team, which developed plans to reduce patient restraints as part of the “Be Free” initiative instituted in 2018. (Tr. 2110-2112). The PSC, which meets monthly, consists of a much larger group of administrators, including the CEO, Chief Management Officer (CMO), CNO, Director of Clinical Services, Director of Risk Management, milieu specialists, patient advocates, and nurse managers. (Tr. 2150; Ex. C-47). The PSC is used as a method for tracking risk trends through the review of various reports and action plans developed in response. Respondent states the PSC is used for review of both patient and staff injuries. (Tr. 1699-1700; Ex. C-47)

Although they were invited to PIC and PSC meetings, virtually no non-managerial direct care staff attended. Respondent contends they were encouraged to attend meetings and provide input; however, according to the staff, the primary obstacle to attendance was the lack of adequate staffing in the units to permit them to attend in the first place. (Tr. 544, 1145-46). Some staff also indicated they were unaware meetings were taking place at all. (Tr. 396, 528, 1145, 1172). Respondent also attempted to hold town hall meetings, but they were also sparsely attended due, in part, to staff either being at work in the units or on their day off. (Tr. 529, 1146, 1807-1808). CNO Forster recommended replacing the town hall meetings with unit-based meetings; however, according to [redacted], these meetings did not take place. (Tr. 567-68; Ex. C-25).


  According to Respondent, staffing decisions are made based on the needs of each of the hospital’s five units. As noted above, each unit serves a different population of patients, which have different care needs depending on the number of patients and the severity of their symptoms. Respondent uses a staffing matrix, which sets staffing levels according to three factors: unit, shift, and number of patients. Based on those factors, the matrix supplies a base level number of staff to be allocated.15 Respondent states it adjusts the base number up or down depending on additional factors, including admits/discharges, adult/adolescent, and acuity. Acuity, as mentioned above, is a metric for individuals, for units, and for the hospital as a whole. At bottom, acuity is a measure of the severity of an individual patient’s affliction or the collective severity of the unit or hospital as a whole.

All staffing decisions start with the matrix. Kevin Smith, Respondent’s staffing coordinator and CPI trainer, testified he sets the initial schedule according to the matrix a few weeks in advance. (Tr. 2355). From there, Smith testified he adjusts the schedule based on anticipated admits/discharges, skill/experience of staff, acuity of both individuals and units, and the medical needs of any individual patient. (Tr. 2355-2357). Staffing meetings are held daily, and Smith and the House Supervisors assess staffing levels periodically to assess whether staff should be called in or moved between units. (Tr. 2354). Smith testified if a staff member asked for additional staff, he would discuss it with the staff member, make his own observations to determine whether additional staff was needed, and seek approval from CNO Forster. (Tr. 2262-2263). Smith and Forster testified requests for additional staff were granted if sufficient grounds were identified for adding staff and staff was available to fill in. (Tr. 1744). However, as discussed below, staff was not always available.

Other than the matrix itself, there is no policy or set of standards to determine how adjustments are made, how acuity is taken into account, or how skill/experience is allocated across the units. With respect to acuity, the Court notes the matrices’ allocations of staff are the same regardless of the unit, number of patients, or shift. (Ex. C-74, C-75, R-55). The matrices do not account for baseline acuity differences between the units, notwithstanding Respondent squarely recognizes differences in acuity between the units. Respondent contends the matrix accounts for differencesother than acuitybetween the units, such as the high volume of admits/discharges from Torreys, the additional hands needed to deal with a more active patient population in the adolescent unit, and the importance of MHT support in the ITUs. (Tr. 1850-1858).

Notwithstanding Respondent’s claims regarding how it allocated staff, Respondent’s employees repeatedly identified insufficient staffing as a recurring and significant problem. (Tr. 504, 573, 1031). This included the employees who testified at trial, as well as numerous and repeated complaints about staffing in staff emails, surveys, and reports. (Tr. 569-572, 1607-1613, 1830-1832; Ex. C-24, C-25, C-41, C-41, C-65, R-72). [redacted] and Ashley Mancha testified staff shortages—fewer staff than provided for in the matrix—occurred daily. According to [redacted], her requests for additional staff went unheeded approximately 75% of the time, whereas Mancha testified her success rate was about 50%. (Tr. 504, 509, 560, 573). According to Mancha, staff shortages impacted the available staff’s ability to properly and timely medicate their patients and initiate individual support plans, both of which are important elements of maintaining a healthy milieu. (Tr. 573-574). In addition to having less staff than provided for in the matrix, a review of the matrices themselves illustrates Respondent reduced the number of staff available at certain census levels after the inspection occurred. (Ex. C-74, C-75, R-55).

The problems experienced during the daytime are compounded during the overnight and weekend shifts. The presumption is that most of the patients are sleeping during the overnight shift, so the matrix allocates one fewer staff member during the 11:00 p.m. to 7:30 a.m. shift. (Tr. 1031; Ex. C-74). This presumption does not always hold, as patients in the ITUs can and do stay awake, and, in some cases, can escalate and agitate other patients. (Tr. 1611). Further, patients can still be admitted to the hospital during the overnight shift, which can further affect the staff-patient ratio. The overnight and weekend shifts are also burdened by two additional considerations. First, the weekday and daytime shifts have an abundance of managerial and administrative staff who are trained in both direct care of patients and in methods to de-escalate and restrain patients should the need arise. Those staff members are not included in the matrix and yet are available to assist. The weekend and overnight shifts do not have the same level of backup available to them, because most of the administration and managerial staff have gone home. (Tr. 448, 507, 1094). On weekends and overnights, the only person usually available to fill in is the house supervisor, who is not included on the matrix. (Tr. 572, 1030). According to Erin Ekholm, who served as an overnight house supervisor, staff shortages and a lack of back-up personnel resulted in the house supervisor serving as a de facto float staffer between the units. (Tr. 1025-1027, 1031-1033). This caused staff to miss breaks, and, in some cases, they were left alone on the unit. (Tr. 849-50, 1037).

According to Respondent, it addressed staffing issues in several different ways. In addition to the procedure for staffing described above, Respondent identified staff members who were not included on the matrix but were nevertheless available to provide care and to assist in emergencies. This set of outside-the-matrix employees included milieu managers, nurse managers, and house supervisors.16 When Forster started as CNO, Centennial Peaks had two milieu managers, whose job was to monitoring unit acuity, patient precautions, and Individual Support Plans. At some point in 2018, Forster eliminated the milieu managers and replaced them with one nurse manager, which he believed would better address patient care from a clinical perspective. (Tr. 1602, 1746-1747). Though Respondent asserts these resources were available as outside-the-matrix staff resources, the milieu manager was not a relevant addition to the units until 2020, long after the inspection occurred. (Tr. 1601, 1602, 1748). Neither Mancha nor Eckholm recalled having milieu managers or resource nurses available to them as additional resources beyond the matrix. (Tr. 573, 1030).

Respondent also had other methods to address staffing shortages. According to Forster, employees would regularly call out sick, which placed additional pressure on the scheduling staff to fill in the gaps in coverage. Kevin Smith, the staffing coordinator, testified he sent out text and email blasts to find additional coverage and, if the need was particularly acute, was able to offer bonuses and other incentives to entice employees to come to work. (Tr. 2460-2461). In addition, Smith testified employees could be “mandated” to stay past their originally scheduled hours in order to account for missing staff or late-admitted patients. These additional measures, in the Court’s experience, underscore the point that staffing was inadequate from the start to account for these situations. The Court also observes having staff work back-to-back multiple shifts or being called in results in employee’s not being well-rested, which, in turn, impacts not only patient care but the alertness and stamina needed for staff to properly observed situations which could escalate as well as respond to those circumstances.

The staff claimed shortages impacted their ability to adequately address the acuity of a unit, including one-to-ones, five-minute checks, and patients on assault precautions. (Tr. 1600). Further, staffing shortages also increased the likelihood an employee would be left alone to perform their regular duties. Employees testified safety was impacted by short staffing, including insufficient staff to participate in patient restraints or de-escalation and insufficient staff to even respond to a code during overnight shifts. In at least one instance, patients had to step in to help a staff member whose head was being repeatedly smashed against the ground even though a Code Green (emergency call) had been called over the PA system. (Tr. 449). These concerns were reported to Respondent’s Risk Manager, Kara McArtor, and CNO Forster.

E.Workplace Violence Data and Employee Narratives

The inspection in this matter was commenced because of an anonymous employee complaint that employees at Centennial Peaks were subjected to physical assaults from patients due, in part, to low staff-to-patient ratios in the ITUs. (Ex. C-2). Through his review of Respondent’s OSHA 300 logs, Employee Accident Reports (EARs), and employee interviews, CSHO Oberbeck confirmed Respondent’s employees were exposed to workplace violence. According to the data CSHO Oberbeck reviewed, Respondent experienced a substantial spike in workplace violence injuries in 2018. (Tr. 103-104; Ex. C-6). The OSHA 300 log, alone, indicated Respondent’s employees suffered two injuries, forty-four (44) days of job transfer/restricted duty, and 0 days away from work in 2017 as a result of patient-on-employee violence. (Ex. C-6). Comparatively, Respondent’s employees suffered twelve (12) “reportable” injuries, two hundred eighty-seven (287) days of job transfer/restricted duty, and twelve (12) days away from work in 2018 for the same reason. (Ex. C-6).17 The OSHA 300 logs, however, only tell a portion of the story.

According to CSHO Oberbeck’s research, he found many of the injuries reported in the EARs did not find their way into the OSHA 300 log. (Tr. 108, 123-125; Ex. C-11). In fact, based on his review, CSHO Oberbeck identified 46 reports of employee injuries that occurred in 2018.18 (Ex. C-11). The predominant number of these injuries involved a nurse or MHT and, typically occurred in the ITUs, Pike and Crestone. (Ex. C-11). Of the assaults that led to employee injuries, twenty-five involved police intervention, and most of those involved arrest, though it should be noted the injuries/assaults resulting in arrest may implicate the same patient multiple times. (Ex. C-11). In addition to the documented assaults and injuries in the EAR and OSHA 300 log, employees testified they were regularly exposed to violent, assaultive behaviors that resulted in minor (or no) injuriessuch as being slapped, kicked, or spit onbut did not report such incidents because they accepted it as part of the job. (Tr. 496-497). As noted above, employee acceptance of minor violent behavior as “part of the job” is well documented in the health care profession and in the inpatient mental health arena, in particular. (Ex. C-12, C-13).

As noted by Respondent, some of the injuries occurred in clusters. (Ex. C-11). For example, MC, who had a lengthy history of violent behavior was responsible for multiple assaults over a few days. (Tr. 353-356, 498, 1654-1681; Ex. C-7, C-20, C-72). The same behaviors were also exhibited by RB and RS, who both assaulted multiple staff members during the course of their stay at Centennial Peaks hospital. (Tr. 471- 481, 478-482, 599-608, 634-37; Ex. C-18, C-19).

  Respondent’s employees testified to several violent incidents to which they have been exposed during their tenure as employees at the Centennial Peaks Hospital. For the purposes of discussing the types of hazards to which Respondent’s employees were exposed and identifying feasible means of abatement to address those hazards, the Court finds the following examples illustrative:

·[redacted] was an MHC at Centennial Peaks from March 2016 to June 2021. During his testimony, [redacted] identified three separate incidents where he was injured. On one occasion, [redacted] was unexpectedly punched in the face by a patient, which resulted in a trip to the emergency room and injuries that kept him out of work for three months. (Tr. 1117-1127).

·[redacted] was also injured at least three times she could recall during her testimony. In all three instances, a physical restraint was either being initiated or in progress. Respondent contends the injuries were largely avoidable because [redacted] attempted to restrain when there were either not enough staff to initiate a hold or because neither the patient nor staff were in immediate danger when [redacted] initiated the intervention. In at least one incident where Respondent questioned [redacted]’s decision-making, [redacted] testified the patient in question had been involved in three prior restraints on the same day and no additional precautions were taken to address the patient’s violent behavior. (Tr. 471-482; Ex. C-18).

·One patient, RB, went behind the nurse’s station eight times over the course of just under a month, during which he assaulted multiple employees and smashed multiple computer monitors. (Tr. 599-608, 634-637; Ex. C-19).

·Another patient, MC, was restrained multiple times over the course of three days, during which time he was responsible for no fewer than five EARs and multiple verbal threats toward staff members. (Tr. 353, 356, 498; Ex. C-7, C-20).

The foregoing incidents are archetypal examples of the types of injuries to which Respondent’s employees were exposed, the situations during which employees were injured, the locations where injuries were likely to occur, and the source of the hazard, i.e., whether such events were isolated and unpredictable or whether other information, such as a patient’s immediate and long-term medical history, indicated a propensity for violence.

F.Respondent’s Methods to Address Workplace Violence

Ultimately, this case is about abatement. Complainant has alleged Respondent failed to address workplace violence in a comprehensive and effective manner and proposed several methods by which Respondent can abate the hazard more effectively. Respondent contends its policies and programs were as effective as possible in addressing workplace violence in the context of an inpatient psychiatric hospital. The following is a recap of the relevant policies and procedures.

i.Workplace Violence Prevention Plan

While Complainant has proposed the development and implementation of a WVPP as one of the proposed abatement methods, in reality a comprehensive WVPP will include or touch upon all of the policies and procedures the parties have discussed during the course of this trial.19 For example, staffing is not, itself, a WVPP program or policy per se; however, the policy must consider factors such as acuity, propensity for violence, special orders, who is designated to respond to codes, and experience of available staff, all of which impact Respondent’s ability to address, deter, prevent and respond to workplace violence. That “adequate staffing” and other complementary issues are discussed in the workplace violence guidance documents issued by OSHA, The Joint Commission (TJC), and Respondent’s own WVPP indicates the scope of a properly implemented WVPP extends beyond administrative and engineering controls designed to prevent a particular type of violent interaction. (Ex. C-12, C-13, C-62).

The initial element to note in Respondent’s WVPP is the definition of workplace violence. According to Complainant’s expert, Dr. Jane Lipscomb, the definition of “workplace violence” is the touchstone upon which a successful WVPP is built, because it guides what needs to be reported, tracked, and ultimately how to address the hazard. (Ex. C-58 at 15-16). Respondent’s WVPP document defines workplace violence quite broadly as Type II, or “violence committed by patients on staff members.” (Ex. C-14 at 1). The specific behaviors that would qualify as workplace violence are not described in the WVPP itself; however, Respondent contends its employees are provided with a clearer definition of workplace violence through training modules, including a PowerPoint presentation, which provides a more precise definition of workplace violence: (i) “WPV is any physical assault, threatening behavior or verbal abuse occurring in the work setting” and (ii) “Physical assault is any attack ranging from slapping, hitting, biting, use of a weapon which results in any degree of injury from no injury to fatal injury.” (Ex. C-58 at 15-16). For the purposes of this case, the parties stipulated the definition of workplace violence is “physical threats and assaults by patients toward staff.” (JS No. 6).

Beyond the initial definition of workplace violence, the WVPP is broken down into five main elements, as well as two additional sections allocating responsibilities and mandating annual evaluation of the plan. The WVPP itself, at least in format and title, mimics the basic elements of a WVPP as described in the OSHA and TJC guidance, which include: (i) Management Commitment; (ii) Employee Participation; (iii) Aggression Analysis & Surveillance; (iv) Hazard Identification and Control; and (v) Staff Education, Training & Competency.

According to the section entitled, “Management Commitment”, the WVPP states generally that senior management will commit sufficient capital and resources to ensure the WVPP is appropriately implemented and continuously updated. (Ex. C-14). Respondent then identifies a series of initiatives, which it contends are supportive of that commitment, including: (i) a 15% reduction in staff injuries resulting from patient aggression in 2018; (ii) monitoring aggression through the use of reports and committees; (iii) encouraging staff to report injuries; (iv) providing post-incident access to medical and psychological care; and (v) additional shows of support, including management commitment to participating in emergency codes alongside staff. (Ex. C-14).

The Employee Participation section identifies some of the strategic committee’s staff are encouraged to participate in but are not necessarily members of. Otherwise, instead of a list of opportunities for employee engagement in the process, this section reads like a series of patient-based obligations. For example, the section states front line staff will be provided with “opportunities to participate in initiatives designed to minimize events of patient-to-staff aggression”, including conducting thorough observation rounds, completing skin checks of patients, and searching patient belongings. (Ex. C-14 at 2-3). While these so-called initiatives are certainly important aspects of a violence prevention program, they have little to do with employee participation and input into the development of a WVPP and instead have the appearance of a job description.

The remaining elements of Respondent’s WVPP address how Respondent will track and trend acts of aggression, identify and respond to specific hazards, and the training it will provide to staff members to execute the elements of the plan. (Ex. C-14). The Court notes, as did Dr. Lipscomb, Respondent’s WVPP focuses on patient-centered interventions, only some of which have an ancillary impact on staff safety. This tendency to focus on the patient side of the equation makes some degree of sense given Respondent’s contemporaneous, yet occasionally conflicting, responsibility to ensure the safety of both its patients and its employees; however, this tension between patient and staff safety highlights some of the systemic problems with Respondent’s WVPP. Principal among those problems is the lack of a targeted definition of workplace violence, which, in turn, impacts Respondent’s ability to effectively develop and implement the WVPP itself.

Finally, the program requires the designated responsible persons to perform an evaluation of the WVPP on an annual basis to ensure it remains effective. (Ex. C-14). The only copy of the WVPP introduced into evidence was neither dated nor signed. CSHO Oberbeck testified there was no evidence Respondent performed an annual evaluation of the WVPP, and neither CNO Forster nor Kara McArtor, who were both designated “responsible persons” under the WVPP, could recall whether they had participated in an evaluation of the WVPP.

ii.Responding to Acts of Aggression

When patients become agitated and start acting out, Respondent has a series of codes its employees can call over the intercom system to ask for help. The two codes discussed extensively at trial were Code Lavender and Code Green. Depending on how agitated the patient has become, staff will request a Code Green or Code Lavender either by calling out loud or, now that walkie-talkies are carried by staff at all times, by calling over the radio. (Tr. 1348). Once the code is called for, the call is repeated three times to ensure it is heard and designated staff respond. (Tr. 2345).

a. Code Lavender

According to Respondent, the Lavender team was created by non-managerial MHT “managers”,20 Kevin Smith and Jordan Larson. (Tr. 2294). The basic idea behind the Lavender team was to create a team of employees, who were particularly skilled at verbal de-escalation, to intervene prior to the point where physical intervention was necessary. The goal was to reduce the number of restraints, which was part of a mandate from PIC called “Be Free”. (Tr. 2295-2298).

The Lavender team is made up of MHTs and nurses and is designated in advance of scheduling. When a Code Lavender is called, the designated members of the team are supposed to respond; however, there are instances where immediate response may be delayed because members of the team may have patient care duties which they cannot leave. (Tr. 326). In order to address this problem, which is equally applicable to Code Green calls, Respondent has designated a Patient Management Leader, or PML. (Tr. 1761). The PML, though not a manager/supervisor per se, is charged with leading the response to a code, which includes interacting with the agitated patient and ensuring continuity of care for the remaining patients on the unit, including observation rounds and special orders. (Tr. 1761, 2315). At any given time, there may be multiple trained PMLs on the schedule; however, only one is designated per shift. (Tr. 2318-2320). According to some employees, the overabundance of PMLs sometimes led to confusion and disagreement as to who was in charge during the calling of a code. (Tr. 518, 1083).

Respondent tracked the Lavender team’s efforts over the course of a year (approximately April 2017 to mid-2018) and determined the initiative showed “significant effectiveness”. (Tr. 1209-1211; Ex. C-36, C-37). In particular, Respondent noted the Lavender team had a 100% success rate in avoiding restraints over a six-month period in the Intake Department and reduced the use of restraints overall over a period of months. (Tr. 2297-2299). Notwithstanding the drop in the use of restraints in early 2018, the number of employee injuries resulting from patient assaults increased dramatically during that period of time. (Tr. 1219-1220; Ex. C-51).



b. Code Green

When an agitated patient becomes physical, or when it becomes clear the enhanced verbal de-escalation techniques are no longer effective, staff members can call a Code Green. A Code Green implies physical management of the patient may be necessary; however, not every Code Green results in a physical restraint. In some cases, the patient can still be verbally de-escalated. If they cannot, however, the Code Green is designed to ensure an adequate number of staff members come to assist in the physical management of the patient using restraint techniques learned in CPI (Crisis Prevention Institute) training.

During the relevant period for this case, a staff member would typically call for a Code Green by verbally calling out to the closest person available. Ideally, this message would be relayed to someone who had access to the telecom/PA system at the nurse’s station, who would, in turn, broadcast the call for a Code Green three times.21 Depending on the situation, however, the individual needing the call may have to run to the nurse’s station to make the call on their own. Once the Code is called, designated responders from each unit are responsible for going to the unit where the Code originated. As with Code Lavender, Code Green responders may have duties which do not permit them to immediately respond, such as one-to-ones, providing medications, or taking patients off-unit. In that instance, PMLs are responsible for ensuring the designated responder’s duties are covered during the pendency of the Code. Sometimes, this includes calling on administrative staff, if available, to fill in; however, this is less likely on the overnight shift because administrators are not typically available at that time.

According to Ashley Mancha, CPI training required four people to initiate a hold in order to maintain safety. (Tr. 562). This assessment was fairly consistent across the board. In some instances, however, staff testified they found themselves in situations where there were not enough people to respond to a Code, or multiple Codes were called simultaneously. (Tr. 517, 562, 1036-37, 1134). Staff testified in those moments it felt as if the Code response was insufficient in terms of the number of people responding, the amount of time it took a sufficient number of people to respond, or both. (Tr. 1036, 1090, 1134). Other than the staff’s subjective perceptions of how long it took other staff members to respond on some occasions, however, there is very little evidence to indicate how long it took for staff to respond, nor is there any indication in the available records that time of response was a problem or reported to upper management. (Tr. 2197). While staff also expressed concerns about the number of staff responding to a Code event, the documentation of those events indicates a robust and timely response. (Tr. 1762, 1765, 2342-2343).

Once the response team arrives at the site of the Code, the on-location staff inform the PML of what has happened to that point. (Tr. 1772, 2153, 2315). Under ideal circumstances, the PML will subsequently debrief each responder as they arrive. (Tr. 2153, 2315). From there, a nurse will supply any needed medication, and the PML will direct the staff on how to properly restrain the patient. (Tr. 1772, 2315-2317).

Once the patient has been restrained or de-escalated, Respondent says it performs debriefs of the incident with all involved parties, including staff and the patient.22 (Tr. 2310). According to Respondent’s policy, debriefs and camera review of the incident are required after a restraint and/or seclusion; however, it does not appear this policy was consistently followed. (Tr. 1204-1205; Ex. C-39). Multiple employees testified they did not participate in a debrief or camera review after they were injured during a restraint or other act of aggression. (Tr. 370, 1082-1090, 1128). Further, even when a camera review took place, employees testified the debriefings took the form of a critique of employee performance instead of an opportunity to both discuss what could be done better and to address employee concerns about what precipitated the incident and how it could have been handled differently. (Tr. 578-580). Debriefing and camera review could also result in retraining or skills update in CPI restraint techniques; however, as with the debrief itself, employees testified such follow-up training would not always occur. (Tr. 485).

In addition to the foregoing, the PML will perform a debrief and maintain a log of the code and its particulars so the data could be used for future trending, tracking, and response. (Tr. 1780; Ex. R-68). This information is also used in Respondent’s Good Catch program, which recognizes employees who are able to manage a potentially dangerous situation, and which provides an additional educational opportunity for other staff members to learn from. (Tr. 2136).

c. Restraint Reduction, “Be Free”, and ITU Expansion

The Seclusion and Restraint Reduction Team (later renamed the Patient Aggression Response Team, aka “PART”) was a subgroup of PIC that was tasked with reducing restraints and seclusions as compared to the UHS corporate benchmark. (Tr. 1185; Ex. C-36). This subgroup was charged with developing risk reduction plans, which, according to Respondent, “focused on reducing the number of staff injuries because of patient restraints.” (Resp’t’s. Br. 36). Ostensibly, this program was designed to reduce staff injuries by reducing the type of incidents Respondent believed to be the most substantial source of staff injury. Notwithstanding Respondent’s successful efforts to reduce the number of restraints, the number of employee injuries went up. This phenomenon coincided with two changes at Centennial Peaks: (1) an increase in the number of patient beds in the ITUs, and (2) instituting the “Be Free” initiative.

In 2017, Respondent added 32 beds to its facility, which Respondent states were dedicated to the ITUs. (Tr. 99). According to Jason Offen, of those 32 beds, eight were added to the Pike ITU. (Tr. 345). Interestingly, both Respondent and Complainant rely on the exact same passages to support these facts. Given that eight beds were added to Pike, for a 30% increase in capacity, there was no explanation as to what happened to the remaining 24 beds. For the sake of clarity and expediency, however, the Court will accept Respondent’s representation that 32 beds were added to the facility, representing a thirty percent (30%) overall increase in capacity, and that those beds were exclusively added to the ITUs. (Tr. 345).

Not long after the expansion, in mid-2018, Respondent instituted a new initiative entitled, “Be Free”, as part of its program to reduce seclusions and restraints. (Tr. 1797). The philosophy statement of the Be Free initiative was “that Centennial Peaks is a place of healing that is committed to delivering restraint-free services to every individual that [they] support.” (Ex. C-36 at 8). Notwithstanding the patient-centered philosophy behind “Be Free”, Respondent asserts the focus was still staff safety, which is why it developed a presentation and training sessions to roll out the initiative. Respondent claims the initiative came with several enhancements to the protocol for responding to patient-on-staff aggression, including: (i) post-incident assessments, (ii) risk investigation, including debrief and camera review, interviews, and determination of whether additional clinical intervention is needed; and (iii) a second medical opinion for any patient who had multiple restraints.23 (Tr. 1800; Ex. C-39). CNO Forster also testified he added staff to the ITUs in order to account for the expansion and Be Free initiative. (Tr. 1800-1801).

CNO Forster testified he added staff after the expansion of the ITUs; however, the Matrices for 2017 and 2019 tell a different story. (Tr. 1800-1801). According to the Court’s analysis of the 2019 Matrix, as compared to 2017, Respondent reduced the number of required staff by one at the following census levels in the Pike Unit: 20 patients, 24-27 patients, and 28-30 patients. (Ex. C-74, C-75). For example, in 2017, a census of 20 patients required five staff during the day and evening shifts and four staff during the overnight shift. (Ex. C-74). In 2019, the same census required four staff during the day and evening shifts and three during the overnight. (Id.). Similar changes were imposed in the Crestone, Torrey, and Sunlight units. (Ex. C-74, C-75).

As noted above in the discussion regarding Code Lavenders and PART, Respondent anticipated a reduction in staff injuries through reducing the number of restraints. Unfortunately, as the data illustrated, that did not happen. Employees testified there were a couple of explanations for the failure of the Be Free/Restraint Reduction initiatives. First, according to some employees, it was not clear the program was designed with staff safety in mind; rather, consistent with the philosophy statement, the program had the appearance of focusing on patient safety with staff injuries being an ancillary consideration. (Tr. 575, 1099). This understanding appeared to be confirmed by meeting minutes discussing the initiative. Specifically, the data presentations from a Seclusion and Restraint Committee meeting indicate Respondent was tracking seclusions, restraints, and special orders, but no data or analysis was presented on the topic of employee injuries resulting from such incidents. (Tr. 1213-1215; Ex. C-37). Second, both Mancha and Wells testified the Be Free program, as presented, created confusion, and some staff were reluctant to participate in restraints without prior permission from Forster, McArtor, or Smith. This hesitation, they believed, resulted in more injuries. (Tr. 576-582).

Respondent identified a couple of reasons for the significant increase in employee injuries in 2018. First, Respondent claims the unexpected rise was due, in part, to the nature of the patients who had to be restrained after enhanced de-escalation efforts were undertaken. Specifically, Respondent contends these patients were more likely to be physically aggressive and, presumably, cause more injuries, which is consistent with one of its primary arguments that staff injuries are the result of confrontations with a small population of outliers. (Tr. 1799-1800). Second, Respondent contends the increase in injuries was attributable to an increased emphasis on reporting employee injuries during the relevant time period. Based on those conclusions, Respondent ended the Be Free initiative around the same time it was determined the Lavender team was successful in its mission to reduce restraints—and, thus, stopped data tracking—and instead chose to focus its efforts on clinical management of patient aggression. (Tr. 1228; Ex. C-54).

d. CPI and WVPP Training

As part of the new employee orientation, direct care staff were required to take CPI training, which was spread out over the course of two days and included training in verbal de-escalation and physical restraint techniques.24 (Tr. 412, 2373-2376; Ex. R-48). Kevin Smith, the CPI coordinator and master trainer, was responsible for providing instruction, certifying competencies, and tracking and providing follow-up training to ensure staff maintained their CPI certification. (Tr. 2306-2307).

The basic gist of CPI training is to utilize verbal de-escalation and, only if necessary, restraint maneuvers, which are based on the biomechanics of the human body. (Tr. 2318, 2376-77; Ex. R-45). According to Smith, premising the physical restraints on biomechanics means CPI-based restraint maneuvers are not dependent on the size or strength of the person initiating the hold and should be able to be performed with as few as two staff members. (Tr. 2318, 2390). If performed correctly, the principles of CPI should serve to reduce the likelihood of injury for both staff and patients, whether through verbal de-escalation or using physical holds. (Tr. 207).

In 2017, a new curriculum was developed by the Crisis Prevention Institute, entitled Advanced CPI. (Tr. 2381). In June 2018, a UHS Clinical Training Coordinator recommended implementation of Advanced CPI because Respondent was using a particular hold that was not from an approved curriculum. (Ex. C-46 at 2). Respondent began implementing the advanced CPI program in September 2018. (Tr. 2380). By the time of the inspection, roughly one-half to three-quarters of Respondent’s employees had completed this training. (Tr. 2378-79). Although Complainant questions why all of Respondent’s staff had not been trained in Advanced CPI at the time of the inspection, there is no evidence to suggest the content of either regular CPI or Advanced CPI is insufficient in terms of preventing injury vis-à-vis any other training program designed for the same purpose, nor was there any evidence to suggest the way Respondent presented it was insufficient.

To support CPI training, as well as Code Green responses, Respondent instituted Code Green drills and utilized actual Code Green drills as an avenue for training/retraining the concepts learned in orientation and refresher training. Kevin Smith sought to improve upon the Code Green drills, and created a CPI drill, which was purportedly more structured, involved actual scenarios, and was done more frequently. (Tr. 2115, 2308-2309; Ex. R-56). Even though Smith testified the drills happened more frequently, there is only documentation of these drills happening in October and November 2018. (Ex. R-56).

Staff members were not as complimentary in their assessment of the effectiveness of CPI and the training regime that accompanied it. [redacted] testified she did not believe the CPI techniques were sufficient when attempting to restrain a patient of significant strength or size. (Tr. 457, 521). Another complaint levied against the program is the fact that some staff members could not properly utilize CPI techniques or simply refused to perform those duties. (Tr. 393, 1135). Finally, employees testified refresher training was often mandated but not usually provided. (Tr. 484-85).

Respondent has a documented WVPP, which it says was made available via its intranet site. (Tr. 2262). None of the employees who testified at trial remembered seeing the WVPP document, nor did they recall ever receiving training specific to the WVPP, itself. (Tr. 1096, 1145. According to Respondent, employees were trained on the component pieces discussed in the WVPP; however, neither McArtor, Smith, or Forster could recall whether Respondent ever held a training session discussing the WVPP and how it connects the disparate programs targeted towards reducing workplace violence. (Tr. 2300)

e. Incident and Injury Tracking

If an employee was injured during an encounter with an aggressive patient, Respondent had a policy requiring reporting of all employee injuries, regardless of severity. (Ex. R-43). The policy lays out the respective responsibilities of the injured employee, supervisors, human resources, and loss control and prevention. (Ex. R-43). In the event of an accident, the WVPP requires the injured employee to call the Sedgwick hotline to report the injury.25 (Id.). Sedgwick is a claims management company, which provides clinical consultation for the injured employee, including whether the employee should pursue medical care for the injuries. (Tr. 2239, 3842). Sedgwick is also responsible for compiling the injuries into the OSHA 300 log. (Tr. 2239-2240). Respondent also utilized other methods for compiling information about employee assaults, including EARs, Healthcare Peer Reviews (HPRs) (also referred to as MIDAS reports), and shift handoff reports.

In addition to reporting the injury to Sedgwick, employees (or their supervisors) are expected to fill out an EAR. (Tr. 106; Ex. R-43). In many instances, however, these reports were not filled out because employees did not perceive their injuries rose to the level of reporting. In other instances, EARs did not get filled out because employees did not receive training on how, when, or whether to fill out an EAR. (Tr. 126-27, 584-86). This was more or less confirmed by Respondent’s Education Coordinator, Carolyn Walker, in response to a query from the HR Director, Erika Donnellan. (Tr. 2257; Ex. C-83).

Respondent also had another series of reports called HPRs, which are also referred to as MIDAS reports. The HPRs are primarily designed for reporting patient-related incidents; however, McArtor testified they can also be used to collect data on employee injuries. (Tr. 1162, 2094). The MIDAS system did not have a data entry point for employee injuries. (Tr. 1162, 2209). Instead, HPRs had a narrative section where injuries could be described; however, as noted by McArtor and Mancha, that depended on the individual filling out the report and was not always done. (Tr. 2209; Ex. C-72 at 32). Those individual reports were then compiled into a summary log. (Ex. C-72). While Respondent claims the HPRs were filled out not only for restraints and seclusion, but also for acts and threats of violence, the data and evidence show otherwise. In particular, the HPR summaries show general categories, such as “patient out of control” or “patient attacked staff—for which there was no discernible distinction—but do not necessarily indicate whether a staff member was injured, threatened, or otherwise unless included in the narrative. As noted by Complainant, in one instance on July 6, 2018, four employees were injured during a restraint of patient SL, but the HPR summary entry does not include any information regarding the injuries. (Ex. 2209-2210; Ex. C-72 at 32).

According to Respondent, the HR Director was responsible for tracking and trending all employee injuries, which she shared with the Risk Manager. This information, in turn, was discussed at various meetings, including daily flash meetings and PIC meetings. See Resp’t’s Br. 35 (citing Tr. 2242-2243). According to the HR Director, Erika Donnellan, she was responsible for maintaining employee files and reporting employee injuries but was not responsible for any portion of the WVPP or policies related to patient-on-employee violence. (Tr. 2260, 2262). Instead, she testified she was responsible for employee-on-employee violence. (Tr. 2263). Kara McArtor, who is a designated responsible person under the WVPP, did not receive EARs (because they were sent to Donnellan), was not formally notified of employee injuries, and did not know whether it was mandatory for employees to report all injuries. (Tr. 1179, 1240; Ex. C-14 at 5). Nevertheless, McArtor was responsible for data reporting in the PSC meetings, which she culled from the HPRs, and Donnellan was responsible for EAR reporting at PIC meetings. (Tr. 1179-80, 2262).

CSHO Oberbeck found the system of injury/incident reporting lacked consistency and Respondent’s system of review lacked cohesion, which he attributed to a confusing system of managerial responsibility for data collection and reporting, as well as the failure to train staff on accident reporting protocol. (Tr. 126, 158).


The facts of this case illustrate the hazard of workplace violence in an inpatient psychiatric/behavioral facility is difficult to assess and wide-ranging in terms of the considerations an employer like Respondent must make to address it. For this reason, the guidance from OSHA, TJC, and the relevant scientific community indicates the response to the workplace violence hazard must be comprehensive in its scope, yet adaptable in its particulars. (Ex. C-12 at 11). This requires a systematic, process-based approach to violence that goes beyond individual, ad hoc abatement strategies and, instead, establishes a program through which all policies and procedures are developed and implemented. The answer, according to industry guidance, is a WVPP. (Ex. C-12, C-13).

The Court is mindful that, due to the nature of the problem, complete elimination of the hazard is an admirable, yet perhaps not entirely achievable, goal.26 Indeed, even Complainant admits it “is not proposing that employee injuries from patient aggression must be entirely eliminated in order for Respondent to adequately protect employees from the hazard of workplace violence.” Compl’t’s Br. 54, n.6. However, just because complete elimination of the hazard is not possible does not lessen Respondent’s obligation under the general duty clause. See 29 U.S.C.

§ 654(a)(1).

Crafting a response to a multi-faceted hazard requires a multi-faceted approach. Indeed, for as much as Respondent pounds its fist about the scientific reliability of any individual method of abatement, the science supports a comprehensive and systematic approach to workplace violence that is inclusive of many of the recommended abatement measures proposed by Complainant. Notwithstanding this fact, Respondent, as it reiterated numerous times at trial and in its brief, focuses primarily on the clinical approach to patient-on-staff violence. Based upon these statements, it is not surprising Respondent’s WWVP places its focus on patient clinical care. Unfortunately, as illustrated by this case, the clinical approach is not enough on its own to completely address the hazard, regardless of its importance to preventing patient aggression.27

The Court finds Respondent’s policies as well as its approach to workplace violence are insufficient. While Respondent undoubtedly has a policy document entitled “Workplace Violence Protection Plan”, the following analysis shows neither the program itself, nor the initiatives purportedly carried out under its banner, are effective at abating the workplace violence hazard. Accordingly, the Court shall affirm the Citation in the manner discussed below.

A.The General Duty Clause

The general duty clause requires employers to “furnish to each of his employees’ employment and a place of employment which are free from recognized hazards that are causing or likely to cause death or serious physical harm . . . .” 29 U.S.C. § 654(a)(1) (also referred to as § 5(a)(1)). See Austin Bldg. Co. v. Occupational Safety & Health Review Comm’n, 647 F. 2d 1063, 1069 (10th Cir. 1981). In order to prove a violation of § 5(a)(1) of the Act, Complainant must show: (i) there was an activity or condition in the employer’s workplace that constituted a hazard to employees; (ii) either the cited employer or its industry recognized that the condition or activity was hazardous; (iii) the hazard was causing or was likely to cause death or serious physical harm; and (iv) there were feasible means to eliminate the hazard or materially reduce it. Baroid Div. of NL Indus., Inc. v. Occupational Safety & Health Review Comm’n., 660 F.2d 439, 444 (10th Cir. 1981). See also 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. See Otis Elevator Co., 21 BNA OSHC 2204, 2207 (No. 03- 1344, 2007).

Specifically, Complainant alleges Respondent violated § 5(a)(1) of the Act as follows:

UHS of Centennial Peaks LLC., dba Centennial Peaks Hospital, at 2255 S 88th Street, Louisville, CO: On and proceeding 12/7/18, the employer did not furnish employment and a place of employment which were free from recognized hazards that were causing or likely cause death or serious physical harm to employees in that employees were exposed to physical threats and assaults by patients. Employees, including nurses and mental health counselors, were exposed to incidents of violent behavior by patients that resulted in serious injuries including bites, sprains, lacerations, bruising, scratches, concussion, and injuries to the head, torso and legs from punches, kicks, forceful grabs, pushes, and tripping. Employees were exposed to the hazard of physical threats and assaults during routine interactions with patients who had known histories of violent behavior. The employer had not established or implemented effective measures to protect employees from assaults or other physical violence in the workplace.


See Citation and Notification of Penalty, Citation 1, Item 1.

As stated above, due to Respondent destroying videos which the Court found Respondent had an obligation to preserve as relevant evidence, the Court issued sanctions against Respondent and imposed adverse inferences that the destroyed videos would have established all elements of the prima facie case including knowledge except for the feasibility prong of the general duty clause analysis. Based on the adverse inferences, Complainant only has to prove Respondent’s policies did not adequately address the hazard of workplace violence and whether there were feasible means to eliminate the hazard or materially reduce it.

Complainant must establish this element by preponderance of the evidence. See Hartford Roofing Co., 17 BNA OSHC 1361 (No. 92-3855, 1995). “Preponderance of the evidence” has been defined as:

The greater weight of the evidence, not necessarily established by the greater number of witnesses testifying to a fact but by evidence that has the most convincing force; superior evidentiary weight that, though not sufficient to free the mind wholly from all reasonable doubt, is still sufficient to incline a fair and impartial mind to one side of the issue rather than the other.


Preponderance of the Evidence, BLACK’S LAW DICTIONARY (10th ed. 2014).

i.Respondent Was Aware of the Hazard

Though the parties did not stipulate to the issue of knowledge, the Court’s adverse inference applies to each element of Complainant’s prima facie case, with the exception of the element of feasibility. Specifically, the Court’s adverse inference on knowledge found the destroyed videos would have provided Respondent with actual knowledge. Additionally, the Court finds this element is not subject to reasonable dispute. Everyone from the MHTs to the CNO, COO, and CEO (which were supervisors/managers) of Centennial Peaks was aware of the workplace violence hazard, which is memorialized in the testimony presented in this case, the reports submitted to Centennial Peaks administration, the employee surveys, EARs, HPRs, and many other documents recounting the same. The question in this case is not whether anyone was aware of the hazard but whether Respondent’s response to the hazard was adequate. Thus, it is unnecessary for the Court to engage in a lengthy analysis about whether a particular individual was aware or could have been aware of the hazard, whether that person was a supervisor, or whether it was foreseeable the person in question would engage in unsafe conduct. See, e.g., Mountain States Tele. & Tele. v. OSHRC, 623 F.2d 155 (10th Cir. 1980).28 Accordingly, the Court finds Respondent had actual knowledge of the hazard and such knowledge can be imputed to Respondent. Id.


ii.Feasibility Prong

Prior to assessing a particular abatement proposal, however, Complainant must show “as a threshold matter, that the methods undertaken by the employer to address the alleged hazard were inadequate.” Integra Health Mgmt., Inc., No. 13-1124, 2019 WL 1142920, at *12 n.14 (O.S.H.R.C., Mar. 4, 2019); see also Sea World of Fla., LLC v. Perez, 748 F.3d 1202 (D.C. Cir. 2014) (holding adequacy of employer’s precautions is measured against precautions taken by “a reasonably prudent employer familiar with the circumstances of the industry”). In making the threshold determination that Respondent’s existing abatement regime is insufficient, the Court is guided by the D.C. Circuit’s analysis in BHC Nw. Psychiatric Hosp. LLC, 951 F.3d 558, 564 (D.C. Cir. 2020). In that case, the employer challenged the ALJ’s conclusion that “the recommended measures would materially reduce the hazard of patient violence beyond what Brooke Glen’s existing safety program already achieved. BHC, 951 F.3d at 564. In particular, the employer asserted the ALJ should have placed more emphasis on its expert’s testimony, which, for the most part, focused on whether the hospital’s existing abatement measures were consistent with the industry’s response to workplace violence. Id. In response, the D.C. Circuit reiterated its prior holdings, which state, “‘[A] safety precaution’ that ‘is recognized by safety experts’ need not find general usage in an industry or have become customary for its absence [to] give[ ] rise to a violation of the General Duty Clause. Id. (citing Nat’l Realty, 489 F.2d 1257, 1266 n.37 (D.C. Cir. 1973).

If Respondent’s existing means of abatement are determined to be inadequate, or plain absent, as compared to precautions taken by a reasonably prudent employer, Complainant must then show: (1) its proposed measures are capable of being put into effect, and (2) those methods would be effective in materially reducing the incidence of the hazard.” See Integra., 2019 WL 1142920, at *12 (citing Beverly Enters., Inc., 19 BNA OSHC 1161, 1190 (No. 91-3344, 2000) (consolidated)). Indeed, in many of the recent Commission and circuit court cases discussing the general duty clause, one of the primary considerations was whether OSHA established the effectiveness of the proposed method of abatement. For example, in BHC, the employer argued the Secretary failed to produce evidence of how the proposed abatement measures would reduce the employer’s “already low rate of patient aggression.” BHC, 951 F.3d at 564. The D.C. Circuit noted two problems with the employer’s argument—one factual and the other legal. First, the circuit court cited with approval the ALJ’s conclusion that Respondent’s reporting process was flawed and thus “diminish[ed] the significance of its recorded rate of patient aggression.” Id. Second, the court noted that, even were it to accept the incidence rate documented by the employer, “[A] low rate of workplace accidents cannot alone establish compliance with the General Duty Clause.” Id. Rather, as it articulated previously in National Realty, [H]azardous conduct need not actually have occurred, or have occurred at any particular rate, for an employer to be liable. Id. (quoting Nat’l Realty, 489 F.2d at 1267). Instead, the key question will always be “whether the Secretary’s measures would appropriately safeguard employees by materially reduc[ing] the hazard of patient-on-staff violence.” Id. at 565 (quoting SeaWorld, 748 F.3d 1202, 1215 (D.C. Cir. 2014)). Ultimately, the court determined the employer’s existing safety protocols were incomplete and/or inconsistently implemented and that a “comprehensively considered and applied program would materially reduce the hazard.” Id.

  The standard for assessing the effectiveness of a particular abatement proposal is not necessarily mathematical. In BHC, the D.C. Circuit held:

Contrary to Brooke Glen’s assertions . . . , the Secretary need not quantify the extent to which that program and its component parts “would have materially reduced the likelihood” of patient-on-staff violence, Nat’l Realty, 489 F.2d at 1267. Instead, the Secretary satisfied the General Duty Clause’s test by establishing that a comprehensive workplace safety program would more effectively and consistently apply measures designed to reduce patient-on-staff violence than Brooke Glen’s present system did.

BHC, 951 F.3d at 565. In that respect, the Commission has found expert testimony, insofar as it is reliable, is sufficient to establish a particular abatement method would materially reduce the hazard. See Integra, 2019 WL 1142920, at *13-14 (finding expert testimony regarding material reduction of hazard of workplace violence sufficient without requiring expert to quantify extent of reduction). The effectiveness of a particular abatement proposal can also be established through the identification of the successful use of a similar approach elsewhere and compliance with industry standards.29 See Pepperidge Farm Inc., 17 BNA OSHC 1993 (No. 89-265, 1997).

iii.Abatement as Process vs. Alternative Abatement Proposals

One final issue the Court must dispose of before addressing the details of Complainant’s abatement proposals is the manner in which Complainant has proposed them. Relying on the Commission’s recent decision in A.H. Sturgill Roofing, Respondent contends Complainant has couched these abatement proposals as alternative forms of abatement. See A.H. Sturgill Roofing, Inc., No. 13-0224, 2019 WL 1099857 (O.S.H.R.C., Feb. 28, 2019). Specifically, Respondent suggests it need not implement every proposed abatement measure, because just one of Complainant’s proposals would be sufficient to materially reduce the hazard. Complainant, on the other hand, argues the abatement measures proposed in the Citation are elements of a process-based approach to abatement. Compl’t’s Br. 61. The Court agrees with Complainant.

According to the Commission, “[W]here actual injury is present and substantial causation has been shown, the Secretary may require [an employer] to engage in an abatement process, the goal of which is to determine what action or combination of actions will eliminate or materially reduce the hazard.” Pepperidge Farm Inc., 17 BNA OSHC 1993, 2034(No. 89-0265, 1997). This approach has been recognized in recent cases addressing the same subject matter as the present case. See, e.g., UHS of Westwood Pembroke, Inc., No. 17-0737, 2022 WL 774272, at *8 (The Secretary’s approach in this regard aligns with the nature of workplace violence, which as alleged here arises in different contexts and conditions at Pembroke, necessitating different abatement measures.”); but see Sturgill, 2019 WL 1099857, at *9 (finding abatement measures to be alternatives where “any one of them would constitute abatement of the alleged violation”). On the flip side of this coin, if abatement measures are proposed as alternatives, then proof Respondent implemented any of the individual abatement proposals constitutes abatement of the violation. See id.

Complainant made clear he proposed the suggested abatement measures as a process by which Respondent could achieve a material reduction in the hazard of workplace violence. In fact, much of Complainant’s case was premised on the concept of a thoroughgoing and comprehensive workplace violence prevention plan that incorporated most, if not all, of the proposed measures. The primary thrust of Dr. Lipscomb’s testimony and report, upon which Complainant substantially relied, was the importance of a comprehensive, process-based approach to workplace violence prevention. (Tr. 1310-1311; Ex. C-58). Although Respondent argues Complainant’s proposals were alleged as alternatives, there is no evidence to support this. This matter was tried by Complainant as a process by which Respondent could achieve a material reduction in the hazard of workplace violence, which Respondent addressed through direct- and cross-examination testimony and exhibits. Accordingly, the Court finds Complainant has properly proposed the abatement proposals as a process-based approach to workplace violence.

The characterization of the abatement proposals either as alternatives or as components of a process of abatement dictates Complainant’s burden of proof with respect to this element. As noted above, if a set of proposals is couched in the alternative, then Complainant must prove Respondent has implemented none of the measures. See Sturgill, 2019 WL 1099857, at *9. However, where Complainant has proposed a series of abatement measures as a process or as components of a single means of abatement, Complainant need only show Respondent failed to implement one of them. Id.; see also UHS, No. 17-0737, 2022 WL 774272, at *8 (“[T]he Secretary need only prove that at least one of the measures he proposed was not implemented and that the same measure is both effective and feasible in addressing the alleged hazard.”); BHC Nw., 951 F.3d at 564 (Secretary proposed “menu” of abatement options to materially reduce workplace violence hazard alleged under general duty clause). Thus, in Pembroke, the Commission limited its analysis to two abatement proposals—equipment for summoning assistance and equipment for de-escalationwhen it determined UHS Pembroke violated the general duty clause. See UHS, No. 17-0737, 2022 WL 774272, *9, n. 11. Notwithstanding the Commission’s holding, the Court will address each of the proposed forms of abatement below in Section V.d.

B.Credibility/Persuasiveness of Experts

A substantial amount of time, roughly three days, was dedicated to expert testimony in this matter. As noted above, the Commission has held reliable expert testimony is sufficient itself to establish a particular abatement method would materially reduce the hazard even if the expert cannot specifically quantify the extent of the reduction. See Integra, 2019 WL 1142920, at *13-14. The Court analyzed the credibility of each expert based on their demeanor, motivation, clarity, and reliability. Although each was properly qualified as an expert in this matter, the Court finds their opinions are not entitled to equal weight. See i4i Ltd. Partnership v. Microsoft Corp., 598 F.3d 831, 852 (Fed. Cir. 2010) (“When the methodology is sound, and the evidence relied upon sufficiently related to the case at hand, disputes about the degree of relevance or accuracy (above this minimum threshold) may go to the testimony’s weight, but not its admissibility.”), aff’d 564 U.S. 91 (2011).

i.Dr. Jane Lipscomb, PhD

Dr. Lipscomb was qualified as an expert in workplace violence in healthcare and behavioral health settings. She has been studying workplace violence for over thirty (30) years, including independent research, publications, and consultation with both government and private entities. (Ex. C-57). Her career started when she worked as a staff nurse at University Hospital in Boston from 1976-78, which she followed up with a stint as an Epidemic Intelligence Service Officer and Nurse Epidemiologist with the National Institute of Safety and Health (NIOSH) from 1981-1984. (Tr. 1288). Dr. Lipscomb earned her Ph.D. in epidemiology from the University of California, Berkeley, during which time she also taught epidemiology at the University of California, San Francisco (UCSF). (Tr. 1290). After her stint at UCSF, she returned to NIOSH, where she advised the Director on a host of issues, including workplace violence, which she had written about during her time in California. (Tr. 1291). After NIOSH, Dr. Lipscomb went back to teaching at the University of Maryland, where she stayed for twenty (20) years, during which time she spent the better part of ten years studying workplace violence full time and even developed a course entitled “Violence as a Public Health Problem”. (Tr. 1292).

According to her Curriculum Vitae, Dr. Lipscomb has published over twenty-five (25) articles addressing the issue of workplace violence in the healthcare setting, including a handful dealing specifically with violence in a behavioral healthcare context. (Ex. C-57). In addition, Dr. Lipscomb has served as principal investigator, co-investigator, or consultant for seven different research and evaluation studies of workplace violence; presented papers/speeches on workplace violence at ten conferences/seminars; and has been honored multiple times for her work in that area. (Ex. C-57). Dr. Lipscomb’s research was also instrumental in developing the OSHA Guidelines for Workplace Violence and subsequent publications related thereto. (Ex. C-12, C-13).

In this Court’s opinion, Dr. Lipscomb’s testimony credibly and directly addressed the sufficiency of Respondent’s abatement methods and the feasibility of Complainant’s proposed measures. Further, with perhaps one exception, Dr. Lipscomb’s testimony went largely unrebutted by fact witnesses or Respondent’s expert, Dr. Marc Cohen, whose testimony was largely based on personal experience and research he found for the purposes of this case, but which is not related to his primary field of study, forensics. Dr. Lipscomb’s opinions, on the other hand, are not only based on studies she personally performed but also upon studies that are within her field of expertise. This, above all, sets her apart from Dr. Cohen and Dr. Argumedo, whose testimony was largely dependent upon their individual, personal experiences and studies they found but do not relate specifically to their field of study. Dr. Lipscomb’s testimony was based on an intimate understanding of the problem of workplace violence due to a career spent studying it. As such the Court has relied extensively on her conclusions in this matter. Accordingly, the Court assigned the testimony of Dr. Lipscomb controlling weight except for the one instance noted below.

ii.Dr. Monica Argumedo

Dr. Argumedo was qualified as an expert in evaluating and clinical management of psychiatric patients and in the efficacy of abatement measures to address workplace violence. (Tr. 777-779). She has been practicing psychiatry for over seventeen (17) years, including receiving board certification in general psychiatry in 2010 and forensic psychiatry in 2011. (Tr. 758-59; Ex. C-27). Dr. Argumedo received her medical degree from the Illinois College of Medicine in 2004 and completed a four-year residency program at Georgetown University Medical Center, during which time she became chief resident in charge of the hospital’s inpatient psychiatry unit. (Tr. 755-56). Since that time, Dr. Argumedo has primarily worked in the field of forensic psychiatry, performing assessments of those charged with crimes to determine whether they are fit to stand trial, as well as serving as an expert witness. (Tr. 761-62; Ex. C-27). However, in addition to her forensic practice, Dr. Argumedo has intermittently served at inpatient psychiatric facilities, in which she has served as director of intake and provided general psychiatric services. (Tr. 763-69; Ex. C-27).

For the most part, Dr. Argumedo’s testimony was based on her personal experiences as a psychiatrist in inpatient facilities; she has not engaged in independent research on the topic of workplace violence other than studies she found for the purposes of this case. That said, to the extent Dr. Argumedo’s opinionswhether based on research or her own experiencealign with Dr. Lipscomb’s, the Court finds such agreement should be due substantial, but not controlling, weight.

iii.Dr. Marc Cohen

Dr. Cohen was qualified as an expert in the field of psychiatry. (Tr. 1895). He received his medical degree from the University of Southern California in 2003 and did his residency at the UCLA/San Fernando Valley Psychiatry Residency Program, during which time he became chief resident from 2006 to 2007. (Ex. R-1). He also took part in the UCLA/VA Forensic Psychiatry Fellowship program. Like Dr. Argumedo, Dr. Cohen is board certified in general psychiatry and forensic psychiatry. (Ex. R-1). During his career, Dr. Cohen has served as an Associate Clinical Professor at the UCLA School of Medicine in the Department of Psychiatry and Biobehavioral Sciences; staff psychiatrist for the Twin Towers Correctional Facility, the VA, and Olive View-UCLA Medical Center; psychiatry expert in mental health court; and as a forensic psychiatrist. (Ex. R-1). He has also given two presentations on workplace violence to the Employment Law Institute and Practicing Law Institute; however, neither addressed the specific issue of workplace violence in an inpatient psychiatric unit. (Ex. R-1).

Like Dr. Argumedo, Dr. Cohen’s testimony was premised, in large part, on his personal experiences and on studies he found specifically for the purposes of this litigation. On those bases, Dr. Cohen opined that the efficacy of the proposed abatements, taken individually, were not supported by peer-reviewed studies. As will be shown in more detail below, many of the studies Dr. Cohen relies on (i) are either limited in their scope, such as the nurse’s station study, which focused on a single nurse’s station in a single inpatient unit; (ii) are anecdotal, such as the Due article, which consisted of 34 total hours of the researchers’ recorded observations; or (iii) are dependent on limited self-reporting, like the Shannon article, which looked at only 19 reports over the course of 4 years and based its conclusion about the counter-therapeutic effect of security presence on a single anecdote. (Tr. 2028, 2047-2060; Ex. R-2). This stands in stark contrast to the peer-reviewed studies Dr. Lipscomb relied upon, which showed that a comprehensive WVPP, consisting of many of Complainant’s proposed measures, would be materially effective at reducing the hazard by up to sixty percent (60%). (Ex. C-58). Further undermining Dr. Cohen’s opinion is the fact that he admitted he neither reviewed, nor was he charged with offering opinions on, the elements of Respondent’s WVPP. (Tr. 2074-75). To the extent he offered an opinion on such matters, it was limited to a conclusory statement that the existing regime was consistent with his understanding of industry practice and was based on his review of deposition testimony. (Tr. 1895). In some cases, however, Dr. Cohen readily admitted certain proposals would have a positive effect on workplace violence, such as personal alarms or walkie-talkies and limiting access to nurse’s stations (without fully enclosing them). (Tr. 2019).

While there are areas where the Court relied on Dr. Cohen’s assessment, particularly with respect to the question of hazard transference from staff to security, the Court did not place significant weight on Dr. Cohen’s testimony with respect to the effectiveness of Respondent’s policies and procedures or with respect to his assessment of the effectiveness of Complainant’s proposals for the reasons stated below.

C.The OSHA Guidelines & Roadmap for Healthcare Facilities

Most, if not all, of Complainant’s proposals can be found in, or are at least based on, a set of three publications: OSHA’s Guidelines for Preventing Workplace Violence (Guidelines); Preventing Workplace Violence: A Road Map for Healthcare Facilities (Road Map); and The Joint Commission’s Sentinel Event Alert: Physical and Verbal Violence Against Health Care Workers (Sentinel Alert). (Ex. C-12, C-13, C-62). As noted by the parties, neither of the foregoing documents present mandatory standards, nor do they create new legal obligations. (Ex. C-12 at 4). Nevertheless, regardless of whether the Guidelines are mandatory standards or not, they recognize employers are bound by the general duty clause and by its reasonable and prudent employer standard. (Ex. C-12 at 4).

The Guidelines were originally produced in 1996, were updated again in 2004, and updated most recently in 2018. (Ex. C-12). They were devised specifically for health care and social service settings, including psychiatric facilities. (Ex. C-12 at 7). According to the Overview, the Guidelines were developed based on “industry best practices and feedback from stakeholders and provide recommendations for developing policies and procedures to eliminate or reduce workplace violence in a range of healthcare . . . settings.” (Ex. C-12 at 7). Those recommendations ultimately form the basis of a model workplace violence protection program, which is broken down into its constituent parts and shows how it would apply in various settings using charts and checklists. (Ex. C-12 at 20, 24). According to the Guidelines, the recommendations reflect the variations that exist in different settings and incorporate the latest and most effective ways to reduce the risk of violence in the workplace.” (Ex. C-12 at 7). The Guidelines’ recommendations are informed by an extensive bibliography of workplace violence studies, including studies by the Center for Disease Control (CDC), international standards associations, state departments of health, multiple publications from Dr. Lipscomb, and the Sentinel Event Alert from TJC. (Ex. C-12 at 40-45).

The Road Map essentially provides the same framework as the Guidelines, upon which it is based. (Ex. C-13). The difference, however, is the Road Map uses concrete examples of how various healthcare facilities, including inpatient psychiatric facilities, have successfully utilized workplace violence policies and procedures, like those discussed in the Guidelines and Road Map. (Ex. C-13). The exemplar facilities were selected, in part, from publications; however, in most cases, OSHA performed site visits, held meetings, and conducted interviews. (Ex. C-13). Of those facilities, they all acknowledged a WVPP were works in progress subject to continuous improvement. (Ex. C-13 at 4).

The third document is the Sentinel Event Alert on workplace violence against healthcare workers, which is published by The Joint Commission (TJC). TJC is a non-profit accrediting and standards-setting body, of which Respondent is a member. (Tr. 1315, 1499; Ex. C-62). TJC produces Sentinel Event Alerts, which identifies specific types of sentinel and adverse events and high-risk conditions, describes their common underlying causes, and recommends steps to reduce risk and prevent future occurrences.” (Ex. C-62). In response, “Accredited organizations should consider information in a Sentinel Event Alert when designing or redesigning processes and consider implementing relevant suggestions contained in the alert or reasonable alternatives.” (Ex. C-62). The Sentinel Event Alert does not explicitly adopt the five-part breakdown of a WVPP discussed in the Guidelines and Road Map; however, a closer inspection of the recommendations reflects a near-identical set of expectations, including: (i) management commitment; (ii) employee participation; (iii) tracking and trending of workplace violence data; (iv) identifying source(s) of the hazard and developing a response; and (v) providing training on de-escalation, self-defense, and responding to emergency codes. (Tr. C-62). In addition, the Sentinel Event Alert identifies factors associated with violence, including, as is relevant to this case: (i) understaffing; (ii) inadequate security; (iii) staff working in isolation; and (iv) lack of emergency communication devices. (Ex. C-62 at 3).

There is a substantial amount of overlap between these publications, including cross-references to each within the respective bibliographies or reference pages, which shows the abatement measures proposed by Complainant (which are based, in significant part, on these publications) are not merely ad hoc responses to workplace violence but are industry-recognized methods for addressing a complex hazard. The overlap not only includes internal cross-references to the documents themselves, but the Sentinel Alert, for example, includes references to the same exemplar hospital systems referenced in the Road Map that use the WVPP Guidelines generated by OSHA. (Ex. C-62 at 4).

Respondent claims neither the Guidelines nor the Road Map are applicable to its facility. This argument is rejected. Not only do the Guidelines specifically reference inpatient psychiatric facilities, but the Road Map uses inpatient psychiatric/behavioral facilities as examples to illustrate the effectiveness of the guidelines. (Ex. C-12, C-13). Further, Respondent has not elucidated any cogent distinction between itself and other in-patient facilities to suggest the engineering and administrative controls suggested by the Guidelines, TJC Sentinel Alert, or Roadmap would not be effective. Nor, for that matter, does it explain why it chose to model its WVPP after the Guidelines’ recommendations.

In addition to the above, the Court notes Dr. Lipscomb testified the Sentinel Event Alert introduced into evidence “really very closely follows the OSHA Guidelines”. (Tr. 1316). Indeed, looking at the Sentinel Event Alert itself, the Court notes it repeatedly references the OSHA Guidelines, as well as other OSHA publications related to the topic of workplace violence. (Ex. C-62 at 2-8).

D.Evaluation of Individual Abatement Proposals

In this section, the Court will address the abatement proposals Complainant included in the Citation. Although the Court will discuss them individually, as noted above, the entire list constitutes an abatement process, of which each numbered proposal is a part. In lieu of addressing these proposals in the order in which Respondent lists them in the Citation, the Court starts with Complainant’s WVPP proposal since, according to expert testimony, it is the touchstone from which other abatement proposals naturally flow.

1.Respondent’s WVPP.

The Court starts with its evaluation of Complainant’s recommended abatement by examining Respondent’s WVPP because it is linked to nearly every other proposal at issue in this case and is the umbrella under which any serious attempt to address workplace violence must fall. Indeed, as noted by Complainant, one of Respondent’s key problems is the failure to connect its various efforts to address workplace violence under the comprehensive rubric of a WVPP. While Respondent’s WVPP uses the terminology discussed in the Guidelines, Road Map, and Sentinel Alert, and even provides a list of connected programming, neither it, nor many of Respondent’s other efforts, were materially effective at reducing the hazard of workplace violence.

The “effectiveness” of Respondent’s approach to workplace violence, which is primarily focused on clinical, patient-based solutions, is indicated by a significant rise in workplace violence events and injuries in 2018. At or around that time, Respondent: (i) increased the capacity of its facility by roughly 30 percent; (ii) implemented an initiative to reduce restraints; and (iii) at certain census levels, reduced the number of staff required under the matrix. While none of these actions are problematic when viewed individually, the manner and timing of their implementation highlights Respondent’s failure to approach the problem of workplace violence in a comprehensive manner, which Dr. Lipscomb, Dr. Argumedo, the Guidelines, Road Map, and the studies they rely upon show is effective at reducing the workplace violence hazard. Respondent contends the rise in reported injuries and incidents is likely due to an emphasis on increased reporting30 and the patients that had to be restrained tended to be more aggressive. (Tr. 1799-1800, 1863). While that may be true at a base level for those injuries actually reported on the OSHA 300, the reporting explanation does not account for the incident and injury numbers reported by CSHO Oberbeck, who culled multiple data sources to find a more accurate picture of accident/injury data that more closely aligned with the interviews he took. His findings highlighted both the unreliability of the data collected and relied upon by Respondent and the lack of a centralized system of data collection for employee injuries.


i.Develop and Implement a Comprehensive WVPP

According to Dr. Lipscomb, an effective WVPP starts with a clear definition of workplace violence. (Tr. 1331; Ex. C-58). According to the Guidelines, NIOSH defines workplace violence as “violent acts (including physical assaults and threats of assaults) directed toward persons at work or on duty.” (Ex. C-12 at 8). The parties have stipulated to this definition for the purposes of this case. In addition, a WVPP contains five key building blocks: (i) Management Commitment/Employee Participation, (ii) Worksite Analysis, (iii) Hazard prevention and control, (iv) safety and health training, and (v) Recordkeeping and program evaluation. The Guidelines state:

The components are interdependent and require regular reassessment and adjustment to respond to changes occurring within an organization, such as expanding a facility or changes in managers, clients, or procedures. And, as with any occupational safety and health program, it should be evaluated and reassessed on a regular basis.

(Ex. C-12 at 11). This interdependency underscores the importance of a comprehensive approach to workplace violence: a failure in one aspect of the program has a trickle-down effect on other elements. For example, the definition of “violence” impacts what events are reported and when they should be reported. (Tr. 1332). The events that are reported, in turn, impacts data collection, which impacts employer’s ability to fashion a targeted and effective response.

Complainant characterized Respondent’s WVPP as a “paper exercise that contained some elements of a proper WVPP but was otherwise incomplete in both its content and implementation. Respondent contends it has developed and implemented an effective WVPP and Complainant’s concerns are directed to the form, not the substance, of its WVPP. Specifically, Respondent points out its program is based on the Guidelines and incorporated numerous clinical and non-clinical policies and practices, which address the hazard of workplace violence at least as effectively as Complainant’s proposals. The Court finds Complainant established Respondent’s WVPP was not effective in some of the subitems identified in Abatement Item #5 of the Citation and, therefore, needs to be changed consistent with the proposed abatement. To the extent Complainant failed to meet its burden with respect the other subitems, the Court has noted those as well.

  Subitem 1: Evaluation and assessment of WPV incidents or trends, including a worksite-specific hazard analysis that analyzes the worksite for risks, including but not limited to, potential weapons, potential for employees to be cut off from communication, delays in activating alarm systems, potential for physical entrapment of employees.” (Ex. C-1 at 7). There are a couple of layers to this subitem which bear discussion. First, Respondent’s WVPP has policies and procedures related to the collection of incident/injury data, Respondent utilizes a committee system through which it collects and analyzes the data, and it performs hazard analyses of the facility and activities occurring therein. Thus, the question with respect to this subitem is whether Respondent’s system was effective. The Court finds it was a mixed bag.

  At the most basic level, the Court finds Respondent’s system of data collection, upon which its system of evaluation and assessment is based, is flawed. As noted, a couple of times above, Respondent’s injury data was contained in multiple sources (OSHA 300 logs, EARs, HPRs), each of which were under the purview of a different manager. The HR Manager, who is not responsible for patient-on-employee violence, is nonetheless responsible for gathering employee accident reports and supplying that data to the Risk Manager, who is responsible for patient-on-employee violence but only reviews HPRs, which are patient-focused documents. In the end, as illustrated by the testimony of CSHO Oberbeck, this led to inaccurate data and inconsistent reports. The Court finds there was no central point of collection for workplace violence data, nor was there a responsible person designated to analyze it, summarize it, and disseminate the conclusions reached on the basis of that data. This underscores the lack of a “coordinated approach” required under the Guidelines, the Roadmap, Dr. Lipscomb’s expert testimony and cited publications.

  That said, the Court finds Respondent performed environmental and data-driven assessments of the facility for hazards, including how furniture or other items could become potential weapons. (Tr. 1492, 2475, 2477). In response to incidents involving furniture as weapons, Respondent purchased new, weighted furniture to reduce the possibility a patient would be able to pick it up, and in response to an incident where a patient threw computer monitors, Respondent bolted them down. (Tr. 2475-2477). When that barrier was overcome, Respondent purchased smaller monitors that could fit under the overhang of the station desk. (Tr. 2496). This illustrates Respondent, at least with respect to this issue, was already implementing the proposed abatement measures.

  As to the potential for employees to be alone, cut off from communication, delays in code activation, and possible entrapment, the Court finds Respondent had a blind spot. Whether through surveys, shift reports, HPRs, EARs, or even policy, Respondent was at least constructively aware its employees would be left to cover a unit alone due to understaffing and, based on their responsibilities, could be cut off from communication because portable communications devices were not required unless off unit. (Tr. 2127, 2213-14). At the least, Respondent was aware MHTs were alone when they transferred patients off-unit for recreation or meals and that intake personnel were often alone, without surveillance, with potentially violent patients during the intake process. (Tr. 2127-28). Each of these situations, including when staff had to enter patients’ rooms for rounds or special orders, presented the possibility of staff entrapment and subsequent delay in initiating a Code.

Subitem 2: “Provide clear written procedures for how employees should respond to patients making threats, showing aggression, and assaults.” With respect to this item, the Court recognizes Respondent had policies covering Code Lavender and Code Green and the use of CPI techniques, all of which are effective and, to an extent, necessary aspects of a successful WVPP. However, there were three areas in which Respondent’s procedures were not clear: (i) when to call police; (ii) when to report an incident of workplace violence; and (iii) establishing a chain of command during a Code involving multiple PMLs. As to the procedures for calling police, staff members on the whole felt as if they were discouraged from calling the police. (Tr. 482, 712, 1047, 1101). Management did not do much to dispel this notion when they second-guessed an assaulted employee’s decision to press charges; changed policy to create additional barriers to contacting police; and questioned the wisdom of calling police, despite the extreme danger, because “encouraging police involvement is not where we need the analysis of the incident to be directed.” (Ex. C-44). Thus, there was a fair amount of confusion amongst staff members regarding the policy.

Regarding the question of when to report an incident of workplace violence, the WVPP is silent. Indeed, the WVPP talks about encouraging employees to submit EARs and that EARs, HPRs, and other data points will be used to inform policy; however, there is no discussion about when or under what conditions an employee should report an incident of workplace violence, including those not resulting in injury.31 This, as mentioned above, has a lot to do with the lack of clear definition of workplace violence. While the parties agreed to a definition of workplace violence which includes threats of violence, the policy’s definition is an overly broad relational description: “violence committed by patients upon staff members”. (Ex. C-14 at 1). Respondent contends it gives a more precise definition during training, which is found in Respondent’s WVPP PowerPoint presentation. (Ex. C-58). According to Dr. Lipscomb, while the definitions provided in the PowerPoint give a clearer understanding of workplace violence, her review of Respondent’s Risk Management Worksheets described numerous incidents fitting the definition of workplace violence but no corresponding reports in any of the known data sources, which she believed undermined the assertion these definitions were understood by or known to staff members.32 (Ex. C-58 at 15-16). Given the reporting discrepancies noted above, as well as the testimony of employees, who testified to accepting some level of violence as “normal, the Court finds employees were given little, or at least inconsistent, guidance on when to report an incident of workplace violence. This is especially so for incidents of violence not rising to the level of an injury, so-called, for which there appears to be no system of reporting or information gathering.

Finally, staff members testified there were occasions during Code Greens and Lavenders where conflicts arose amongst responders as to who should lead the code. (Tr. 518, 1083, 1091). Emmett Wells, who worked as a social worker at Centennial Peaks, testified he believed the breakdown in communication in these circumstances led to three of his injuries during responses to codes and, ultimately, caused him to opt out of responding to codes.33 (Tr. 1083, 1091). As noted above, PMLs are either nurses or MHTs designated to lead crisis/code response. According to Smith, there are designated PMLs for code greens, but, due to their training, more than one of the designated responders are likely to be trained as PMLs. (Tr. 2318-2319). Considering the PMLs also had duties that might prevent prompt response, such as one-to-ones or other special patient care responsibilities, the ground was ripe for conflict based on who responded first. Respondent did not present any evidence, nor could the Court find any, to indicate a procedure to prevent clashes over who should lead the code.

Subitem 3: “The WVPP must also provide for the participation of direct care staff such as MHCs and RNs, e.g., through the committees that discuss WPV incidents.” Complainant has two complaints about this element: (i) the section of the WVPP entitled “Employee Participation” primarily consisted of tasks for staff to complete that are part of patient care, not initiatives to participate in; and (ii) direct care staff did not have a standing seat on any of the committees discussed above—PIC, PSC—or even knew they existed. (Tr. 528, 1145, 1175). Respondent argues; however, its direct care employees have a standing invitation to participate in both PSC and PIC meetings. In addition, Respondent highlighted initiatives, policies, and procedures, which were initiated at the suggestion of direct care staff, including Patient Management Leads (PMLs), the Lavender Response Team, and even the WVPP was developed with the assistance of Kevin Smith and Jordan Larsen, who were MHTs/non-managerial milieu managers at the time. (Tr. 2099, 2299; Ex. C-14); see Resp’t’s Br. 72. Even Mancha and [redacted] testified to occasions where their suggestions were adopted. (Id.).

While the Court finds Respondent provided its employees with opportunities to participate in developing initiatives to address workplace violence, the Employee Participation program of Respondent’s WVPP is, for the most part, a set of responsibilities and duties employees owe to patients. This is not just a criticism of form, but the WVPP’s current construction suggests a misunderstanding of the role direct care staff have in identifying, assessing, and implementing policies and procedures to prevent workplace violence. (Ex. C-12 at 13). This conclusion is made stronger by the fact that most of the employees who testified stated they had never seen the document before. (Tr. 394, 1095, 1145). The WVPP should be amended to reflect the purpose of employee participation in its continuing development.

  Subitem 4: Implement administrative and engineering controls to reduce or eliminate WPV hazards, including an assessment of appropriate staffing levels for each unit.” The Court finds the substance of this subitem is addressed in multiple other sections, from the modification of the nurse’s station and the use of walkie-talkies or panic alarms to Respondent’s assessments of proper staffing levels. While Complainant’s criticism as to this item is directed to the contents of the physical plan itself, the Court finds these sub-proposals are ultimately duplicative of other proposals specifically targeted at engineering controls and administrative procedures, such as staffing levels. Thus, the Court will discuss those specific abatement proposals elsewhere.

  Subitem 5: Annually review the WVPP and update as necessary. Solicit employee feedback during the process.” According to CSHO Oberbeck, there was no evidence the WVPP had been reviewed or updated since its creation. (Tr. 166). Neither Forster nor McArtor recalled participating in an annual review of the WVPP, even though both are designated as responsible persons under the plan. (Tr. 1219, 1693; Ex. C-14). Respondent contends the policies and procedures set forth in the plan “were constantly being revised and new mitigation efforts were developed and implemented on a continuous basis.” Resp’t’s Br. 72 (citing Tr. 1340). While policies and procedures may have been addressed on an individual basis, there is no sense—indeed, no evidence—Respondent reviewed any of those policies collectively as part of its WVPP. The fact that two of the primary responsible parties under the WVPP had no recollection of participating in a review of the program, even after the need for such a review was identified by UHS Corporate Loss Control in a site visit at the beginning of 2019, suggests the WVPP program is ultimately a list of duties and patient-based programs or, as Complainant put it, a “paper exercise”.

  Through the foregoing discussion, the Court has addressed many of the Guidelines’ so-called building blocks of a WVPP, including Employee Participation/Management Commitment, Hazard Analysis, Injury Tracking, and Evaluation.34 The Court identified areas where Respondent’s WVPP complied with the letter and spirit of the Guidelines and also where Respondent’s existing regime was ineffective at addressing the hazard. At bottom, however, the Court finds Respondent’s WVPP, taken as a collective whole, is ineffective at its primary role. There is no sense in which the various programs and/or initiatives that purportedly make up the WVPP are connected to one another. Further, there appears to be a lack of buy-in into the program from either side of the employment coin: designated responsible parties in management are not aware of when or even whether a review of the WVPP took place, and most of the employees who testified are not aware of the existence of an overarching WVPP, let alone their ability to participate in initiatives and programs created under its guise. In that way, the Court agrees Respondent’s WVPP, in its current form, is little more than a paper exercise and is, therefore, deficient, and ineffective at preventing workplace violence.

ii.Feasibility of Complainant’s Proposal Regarding WVPP and Analysis of Whether it Would Result in Material Reduction of the Hazard

There is little debate over the feasibility or effectiveness of this particular proposal; indeed, Respondent developed a WVPP with at least a nominal nod to the components discussed in the Guidelines. Nor, for that matter, does Respondent contend a WVPP would not be effective. Instead, Respondent spent significant time arguing Complainant could not prove any of the individual abatements, or specific combination thereof, would be effective at reducing the hazard. This approach is contrary to the current state of workplace violence research, which is, in the words of Dr. Lipscomb, a “multi-factorial” problem. (Tr. 1322-23). What researchers discovered is that a multi-factorial problem requires a multi-factorial, yet adaptable, solution. (Ex. C-58, C-59; C-61 at 2-3).

  In the study relied upon by Dr. Lipscomb, researchers conducted a randomized, controlled intervention study (the so-called “gold standard”) of implementing a “data-driven, worksite-based intervention based on the OSHA Guidelines. (Ex. C-58 at 12). The study found hospital units, including psychiatric units, that implemented a comprehensive program based on the Guidelines and adapted it according to the data, was effective at reducing patient-to-worker violence-related injuries by 60% within 2 years of implementation. (Tr. 1324; Ex. C-58 at 21). This conclusion was borne out in other contexts, including examples highlighted in both the Road Map and Sentinel Event Alert. (Ex. C-13, C-62). Dr. Lipscomb highlighted Aria Jefferson hospital, which is referenced in the TJC publication, and a trade journal called Healthcare Executive. (Ex. C-61, C-62 at 4). According to these sources, Aria Jefferson was able to use a data-driven, comprehensive approach to workplace violence to reduce violence related injuries by 70 percent over the course of four years.35 Indeed, throughout the Road Map, there were multiple examples of healthcare facilities that experienced success in reducing workplace violence through the implementation of a comprehensive program to address workplace violence. (Ex. C-13).

  Respondent’s attempts to undermine these conclusions are not persuasive. In particular, Respondent points out the size of the hospital system, perhaps to suggest its conclusions are not applicable to a small, stand-alone facility like Respondent’s; however, Respondent has not identified anything about the 42 individual units studied within that system to highlight a meaningful difference between them and Respondent. Respondent also points out the researchers tracked a more broadly defined hazard, which included verbal abuse, bullying, and harassment; however, the problem for Respondent is the study may have used a more inclusive definition of workplace violence, but the ultimate effect was a reduction in patient-to-staff injuries. In other words, a more inclusive definition arguably made the program more successful not in the identification and reduction of workplace violence generally, but in the reduction of injuries, which data likely showed was the result of a broad range of violent encounters. Finally, Respondent points out the study was not specific “as to which interventions were used on which units and it cannot be determined from the study which interventions improved reducing the rate of workplace violence incidents. Resp’t’s Br. 64.

Respondent’s last argument misses the point: what matters is that the approach be based on the data, specific to the worksite, and comprehensive in its scope. The fact the authors of the study could not specify some magic formula of interventions culled from the Guidelines makes sense considering the assessment is data-driven and site-specific. What these studies show is the importance of a process driven by the Guidelines. Respondent’s position, though incorrect, makes sense given its reliance on Dr. Cohen’s testimony, which focuses on evidence-based analyses of individual interventions. This focus also highlights one of the reasons why the Court placed more weight on Dr. Lipscomb’s testimony: she is more attuned to the state of the science with respect to workplace violence. Dr. Cohen is clearly a qualified clinical and forensic psychiatrist, but he, like Respondent, is focused on clinical interventions at an individual level, which are more easily controlled than changes to administrative or engineering controls impacting a large population of patients with differing diagnoses and levels of acuity implemented by healthcare professionals in different settings. Dr. Lipscomb, on the other hand, has made the study of workplace violence a significant part of her life’s work; she is not only familiar with the state of the research but was also instrumental in conducting studies supporting the comprehensive approach described above. The Court finds Complainant, through Dr. Lipscomb’s testimony, established the proposed comprehensive WVPP would be materially effective at reducing workplace violence in Respondent’s hospital.

Finally, Respondent also attempts to undermine Complainant’s claim regarding effectiveness by suggesting many of the incidents of workplace violence were “unpredictable” and, thus, no amount of intervention, comprehensive or otherwise, would be adequate to address the hazard. Respondent goes as far as to claim [a]ggression which is caused by mental illness rather than bad intent is by its very nature unpredictable and preventing it is not feasible because it is almost always unforeseeable. Resp’t’s Br. 64 (citing Integra, 2019 WL 1142920, at *24. (MacDougall, H., concurring) (Workplace violence is a “condition that results from the acts of third parties, engaged in unpredictable human behavior, and outside of an employer’s control; no matter how sound an employer’s safety program might be for the particular circumstances in its workplace, it is possible the employer cannot free the workplace of the hazard.”) While the Court agrees there were examples of staff being injured by patients without warning, this made up a small portion of the cases discussed. Further, the Court finds Respondent overstates the case regarding the predictability of incidents of aggression.

Just because a small universe of outbursts may be “unpredictable, such does not justify the refusal to implement strategies that have been successfully implemented by others and are recommended based on joint efforts between OSHA, NIOSH, and the Joint Commission, which itself approves of the Guidelines/Roadmap as a starting point for an effective WVPP. Further, if the Court were to accept Respondent’s characterization of workplace violence, it begs the question as to why Respondent uses tools to assess violence risk and implement initiatives to reduce that risk if they serve no purpose.36 The fact is, while individual acts viewed in isolation may appear unpredictable, the evidence presented at trial shows there were multiple data points available to Respondent, which would have illustrated the existence of injury clusters related to a location, activity, or type of patient.

Respondent lamented the characterization of its facility as a hazardous workplace because of the presence of mentally ill patients when, it claims, the data shows a substantial amount of the workplace violence incidents was connected to a discrete population of high acuity patients with a substantial history of prior violent acts. Looked at another way, however, Respondent had data available to it indicating a specific subset of patients was responsible for an outsized number of incidents, which presented an opportunity for assessment and evaluation of policies and procedures related to this subset of patients, which Respondent did not do. The individual files of MC, RB, and LJ illustrate this failure: there was a lack of documentation (or at least a lack of notification/warning) about a patient’s prior history of violence even though the information was available at intake. (Ex. C-19, C-20, C-23). Further, with respect to Respondent’s claims regarding the efficacy of a clinical approach to violence from this select subset of patients, the Court finds such interventions were insufficient, on their own, to address the hazard, as illustrated by the number of attacks towards staff occurring in close succession despite clinical interventions occurring in the interim. (Id.). Along similar lines, the Court finds proper implementation of a comprehensive program, driven by data, would also have revealed the cluster of incidents occurring at the nurse’s station and presented opportunities for prophylactic measures to prevent these “unpredictable” acts of aggression. Had Respondent properly analyzed its own data, whether with respect to high acuity patients or the nurse’s station, the next logical step would have been for Respondent to take appropriate actions to eliminate or substantially reduce this risk through proper planning and implementation of a comprehensive WWVP focused on such populations and locations. Respondent cannot simply put its head in the sand, plead a certain class of patients are more prone to such conduct and unpredictable, and then do nothing.

The Court can rely on expert testimony to determine the feasibility of proposed abatement as well as whether it would materially reduce the hazard if implemented. The Commission has found expert testimony, insofar as it is reliable, is sufficient to establish a particular abatement method would materially reduce the hazard. See Integra., 2019 WL 1142920, at *13-14 (finding expert testimony regarding material reduction of hazard of workplace violence sufficient without requiring expert to quantify extent of reduction). The Court previously found Dr. Lipscomb’s testimony was entitled to controlling weight. Dr. Lipscomb testified Respondent’s WVPP was flawed and provided reasons for her conclusions as well as studies which supported her testimony that a comprehensive, linked WVPP would result in a material reduction of workplace violence instances in Respondent’s workplace.

Based on the foregoing, the Court finds Complainant established the feasibility of Complainant’s abatement proposal #5, recommending the implementation of a comprehensive WVPP. In addition, the Court also finds the testimony supports Dr. Lipscomb’s conclusion that implementation of a comprehensive and coordinated WVPP would result in a material reduction of workplace violence incidents and, therefore, would materially reduce the overall hazard of workplace violence as recognized by both Respondent and its industry.

The Court could stop its analysis at this point and end the discussion. As discussed above, the Court found Complainant proposed a series of abatement measures as a process. Thus, Complainant need only show Respondent failed to implement one of them. See Sturgill, 2019 WL 1099857, at *9; see also UHS Pembroke, 2022 WL 774272 at *8 (“[T]he Secretary need only prove that at least one of the measures he proposed was not implemented and that the same measure is both effective and feasible in addressing the alleged hazard.”); BHC Nw., 951 F.3d at 564 (Secretary proposed “menu” of abatement options to materially reduce workplace violence hazard alleged under general duty clause). However, for the purposes of clarity and completeness, the Court will continue its analysis on the remaining proposed abatement measures in light of the pending appeal of the Commission's decision in UHS Pembroke.

2.Reliable and readily available communications devices

Prior to the inspection, Respondent’s employees only carried communication devices when they transported patients off-unit. When they were on unit, staff only had two radios available, which were kept at the nurse’s station, along with the telephone, which served as the primary device by which codes were broadcasted to the rest of the hospital. (Tr. 97, 389). Otherwise, while on unit, staff would have to verbally call for a code. At full staffing capacity, this might not be a problem; however, there were numerous examples provided by employees where they were alone or could not immediately be heard. This included intake personnel, who are typically alone while performing their assessment and are not usually monitored remotely, even though cameras were available.

The Court finds Respondent’s prior regime, which relied on calling out to other employees in the hopes of getting a message relayed to the PA system, was ineffective. Under the best of circumstances, response time could be delayed by the logistics of the situation: the individual calling the code needs to be heard by someone physically occupying the nurse’s station (though Respondent apparently frowned on staff remaining behind the desk), who could relay the message out to the hospital as a whole. Any potential delay in response can, and did, result in employees receiving serious injuries. For example, [redacted] testified she and a co-worker were repeatedly assaulted during the pendency of a Code Green call because they could not easily make the call for a code, which resulted in other patients intervening to prevent further injury. (Tr. 448-49). Respondent’s Corporate Loss Control recognized this shortcoming, at least with respect to the intake process, and recommended the implementation of panic alarms during the intake process. (Ex. C-51 at 2).

The Court also finds implementing the use of walkie-talkies and/or personal panic alarms would be feasible and effective. First, Respondent already had two walkie-talkies present on each unit at the time of the inspection, which its staff used in specific circumstances, indicating its employees were familiar with using the walkie-talkies for this purpose. See Integra 2019 WL 1142920, at *12. Second, after the inspection occurred, Respondent increased the number of radios on unit to make them available to each employee, which both Forster and Smith testified improved response times to Code Greens and had the ancillary effect of allowing codes to be called at night without disturbing sleeping patients on the unit and unnecessarily affecting acuity of the milieu. (Tr. 2347-48). See SeaWorld, 748 F.3d at 1215; FMC Corp., 12 BNA OSHC 2008 (Nos. 83-488, 83-489, 1986). Third, each of the experts, including Dr. Cohen, testified the use of walkie-talkies and/or panic alarms would be effective in reducing the likelihood or severity of the injury resulting from an act of patient aggression. (Tr. 1520, 2019). Fourth, the Guidelines refer to using communication devices, such as panic alarms or cellular phones to summon assistance more effectively. (Ex. C-12).

All three experts testified the use of walkie-talkies would result in a material reduction in the identified hazard. See Integra, 2019 WL 1142920, at *13-14 (finding expert testimony regarding material reduction of hazard of workplace violence sufficient without requiring expert to quantify extent of reduction). Accordingly, the Court finds Complainant proved the use of personal communication devices and/or panic alarms would be a feasible and effective method to materially reduce the hazard.37

3.Reconfiguring the Nurses Station

According to the testimony, the nurses’ station was a locus for patient aggression. (Ex. C-72). With no door to prevent entry, and a fairly low and narrow countertop, the nurse’s station provided virtually no barrier between direct care staff and patients. (Ex. C-5). The incidents described by employees include: (i) patients reaching over the top of the counter to rip computer monitors off the desk even when they were bolted down; (ii) patients walking behind the nurses desk and physically assaulting staff; (iii) a patient attempting to barricade a staff member behind the station, numerous threats; and (iv) destruction of property in the area of staff. (Tr. 129, 583-85, 601-604 1634-36; Ex. C-7 at 12, C-19 at 13-16, Ex. C-72). Many of these incidents could have been prevented by modest improvements at creating an actual barrier to entry. If the behavior of psychiatric patients is so unpredictable, as Respondent has argued, employees should be afforded more substantial protection when working behind the desk, during which time their focus may be on paperwork or other duties which may divert attention from an unpredictable source of harm. The recurrence of these types of incidents at the nurses’ station during the relevant period indicates Respondent’s current layout is not effective, and the Court so finds.

Complainant proposes a performance-oriented solution to the nurse’s station. Contrary to Respondent’s characterization of the abatement proposal, the Citation does not mention a particular solution; instead, Respondent is expected to reconfigure the station to prevent the types of incidents described above: (i) jumping over the counter; (ii) walking through the door; and (iii) grabbing items that could be used as weapons. (Ex. C-1 at 7). At trial, of course, this idea was fleshed out to include solutions such as a locking door and higher and deeper countertops. (Tr. 145). The employees who were attacked at the nurses station all testified they believed a more substantial barrier, such as a door or higher counter, would have prevented the types of attacks described above, as did the experts, including Dr. Cohen. (Tr. 1352-56, 2032). This consensus is reflected in the Guidelines and Road Map, which recommend deep counters at the nurses’ stations up to enclosure as a means of barrier protection, but also states employers and workers must determine the appropriate balance between creating a suitable atmosphere for the services provided at the hospital and the type of barrier being used. (Ex. C-12 at 21).

Respondent contends the number of injuries and incidents at the station is overblown, citing only two injuries at the station in 2018. (Tr. 143). Respondent downplays incidents where patients entered the station without injury, including when one patient “pulled a monitor from its base” without mentioning what happened to the monitor afterwards; or when another patient “grabbed a badge and keys”, which just so happened to be attached to a staff member; or when a patient entered the station and “grabbed a bin of pens/paperclips”, which she subsequently used to throw at staff. (Ex. C-63, C-72 at 33). Respondent casts this proposal as a straw man and suggests Complainant is requiring it to fully enclose the station. While the Road Map mentions enclosure as an option, Complainant did not include it among the options it was pursuing. Respondent also argues it previously attempted to modify a nurses’ station but notes it was poorly received by staff, who claimed it interfered with performing their duties. This argument is equally unconvincing because it only suggests that particular iteration of the nurses’ station was inconvenient; however, there was no evidence Respondent attempted to address this, and, for that matter, inconvenience is not a defense to the element of feasibility. See Tri-State Roofing & Sheet Metal, Inc. v. OSHRC, 685 F.2d 878, 881 (4th Cir. 1982) (“The particular views of workmen are not necessarily, and often times are not, the best determination as to what is safe and what is unsafe. Convenience rather than safety considerations often dictates a worker's perspective.”).

Finally, Respondent yet again relies on a fatalist perspective on patient-on-staff violence, when it argues, “[W]hen dealing with patients with emotional disorders, enclosing the nurses’ station does not prevent an assault; it simply changes its location as the patient will commit the assault elsewhere if not clinically addressed. Resp’t’s Br. 68. Simply because the possibility of a future assault exists does not justify failing to respond in a concrete way to an easily identifiable problem. Preventing an assault in the moment can give the staff member a moment to prepare and the patient an opportunity to decompress, depending on the nature of the “dispute”. In this instance, something as simple as a door or making the counters higher and deeper would likely have prevented at least a handful of the encounters described.

All three experts testified that the reconfiguration of the nurse’s station would result in a material reduction of the identified hazard. See Integra, 2019 WL 1142920, at *13-14 (finding expert testimony regarding material reduction of hazard of workplace violence sufficient without requiring expert to quantify extent of reduction). The Court finds the foregoing abatement feasible and effective at materially reducing the identified hazard.

4.Staffing Issues

In this section, the Court will address two separate, but interrelated abatement proposals: (i) designating staff, with no duties to prevent immediate response to a code, to respond to violent incidents; and (ii) ensuring staffing adequate for census and acuity, taking into account patient aggression and history of violence.38 As the following will show, the Court finds Respondent has already implemented Complainant’s proposal to designate a response team for acts of patient aggression. Complainant failed to show Respondent’s current protocol for Code Green or Code Lavender response is insufficient in terms of time or adequacy of response. In nearly every report of a Code Green, the treatment records indicate a robust response in terms of the number of employees responding. (See e.g., Tr. 1762, 1765, 2342-2343; Ex. C-18 at 2, 8, 14; C-19 at 9, 23, 31,