THIS CASE IS NOT A FINAL ORDER OF THE REVIEW COMMISSION AS IT IS PENDING COMMISSION REVIEW

Some personal identifiers have been redacted for privacy purposes.

United States of America

OCCUPATIONAL SAFETY AND HEALTH REVIEW COMMISSION

OSHRC Docket No. 19-0550

Appearances:

Alicia A.W. Truman, Esq. & Beau Ellis, Esq, Department of Labor, Office of the Solicitor, Denver, Colorado

        For Complainant

Melanie L. Paul, Esq. & Dion Kohler, Esq., Jackson Lewis P.C., Atlanta, Georgia

        For Respondent  

Before: Judge Peggy S. Ball – U. S. Administrative Law Judge

DECISION AND ORDER

  In this matter, Complainant alleges Respondent failed to adequately protect its employees from the hazard of workplace violence in an in-patient psychiatric hospital. Although patient-on-employee violence is a hazard Respondent is responsible for preventing, the parties recognize workplace violence is an unavoidable aspect of providing care to individuals with psychiatric disorders. Patients come to Respondent’s facility during difficult times in their lives when they are incapable of self-regulation and need treatment for their illness. In other words, Respondent is responsible for managing and maintaining a therapeutic, yet safe, environment with unpredictable individuals in complicated circumstances.

  Respondent’s obligation to provide a safe environment, however, is not limited to its patients. Under the Occupational Safety and Health Act, Respondent is also obliged to ensure its employees are provided with adequate protections against violence perpetrated by the patients for whom the employees provide care. As will be discussed further below, providing a therapeutic environment for mental health patients while simultaneously ensuring a reasonably safe work environment for employees caring for them is, no doubt, a delicate balance to strike. Notwithstanding the complex nature of the problem, however, the Court finds Complainant proved Respondent failed to achieve the appropriate balance between the treatment of its patients and the safety of its employees. While Respondent argues its clinical approach to patient care is the most effective way to address patient-on-employee violence, the evidence shows this approach resulted in a patchwork of ad hoc policies and programs primarily geared towards patient care with employee safety as an ancillary consideration. While the Court does not doubt Respondent’s desire to protect its employees, the evidence shows Respondent’s workplace violence protection program was ill-suited to that purpose.

  Complainant initiated an inspection of Respondent’s workplace on October 2, 2018, in response to an anonymous complaint about workplace violence and the lack of a workplace violence prevention plan (WVPP). (Tr. 47). The inspection was conducted by Compliance Safety and Health Officer (CSHO) Aimee Stark. CSHO Stark visited the UHS Highlands facility three times over the course of her inspection, during which time she interviewed members of the management team and approximately twenty-two employees. (Tr. 50-54). At the conclusion of her inspection, CSHO Stark recommended, and Complainant issued, a Citation and Notification of Penalty. (Ex. C-1). In the Citation, Complainant alleged Respondent violated the general duty clause, 29 U.S.C. § 654(a)(1), because it failed to adequately protect its employees from the hazard of workplace violence and proposed a penalty of $11,934. (Tr. 68-69; Ex. C-1). Respondent timely filed a Notice of Contest.

  The parties exchanged multiple motions during the course of litigation; however, only one warrants discussion at this point.1 Complainant filed a motion to amend the Complaint to (a) add UHS-Delaware as a party, alleging Respondent and UHS-Delaware were a single entity under the three-part test described in Loretto-Oswego Residential Health Care Facility, 23 BNA OSHC 1356 (No. 02-1164, 2011); and (2) add two additional means of abatement. As to (a), the Court found Complainant failed to establish the two entities functioned as a single-employment entity or had an identity of interest “such that addition of Delaware at this late point in the litigation would not be unduly prejudicial.” As to (b), the Court found the amendments to the abatement section of the Citation did not significantly alter the allegations, nor did they add new facts or issues. Thus, the amendment was granted as to the abatement but not as to the addition of a new party.

  Trial in this matter was originally scheduled to commence on June 1, 2020; however, due to the COVID-19 pandemic, it was rescheduled multiple times. Ultimately, the trial commenced roughly one year later on June 7, 2021, and lasted for 14 days.2 Both parties timely filed post-trial briefs for the Court’s consideration. Based on the evidence presented at trial, and the parties’ respective briefs, the Court issues the following Findings of Fact and Conclusions of Law.  

  Prior to the first trial date in this matter, the parties filed a set of Joint Stipulated Facts. While some basic facts regarding the nature of the hospital are included, the most notable stipulations include those related to jurisdiction and Complainant’s burden of proof. As to jurisdiction, the parties agree Respondent is an employer engaged in a business affecting commerce under Section 3(5) of the Act and that the Commission has jurisdiction over this matter pursuant to Section 10(c). See 29 U.S.C. §§ 652(5), 659(c).

  In addition, the parties also stipulated to required elements of Complainant’s burden of proof in a general duty clause case. Specifically, the parties stipulated: (1) employees were exposed to the hazard of workplace violence; (2) Respondent and its industry recognized the hazard of workplace violence, defined as assaultive or violent behavior of patients toward staff, at the time of the alleged violation; (3) exposure to patient violence and/or assaults by patients could result in serious injuries. The only remaining element at issue before the Court is the question of feasibility of abatement.3 With respect to that particular element, the parties agree “the hazard of workplace violence . . . cannot be eliminated at Highlands through feasible abatement measures.”

Respondent operates a stand-alone behavioral health facility in Littleton, Colorado, which provides in-patient psychiatric care for adults and adolescents ages 11 years and older. (Tr. 58). Most patients are admitted to Highlands involuntarily and are treated for conditions such as depression, anxiety, psychosis, schizophrenia, and suicidal thoughts. (Tr. 3197). The average stay ranges approximately 7-10 days but can be as long as three months. (Tr. 58). According to the former Medical Director, Dr. David Weiss, the individuals admitted to Highlands “tend to be pretty ill” and “are having a break from reality.” (Tr. 3924-3925).

The Highlands hospital building is a three-floor facility with a locked entrance to restrict entry into and exit from the facility. (Tr. 397; Ex. R-2). The first floor of the facility contains the dining room, cafeteria, administrative offices, group therapy rooms, and intake/admissions. (Tr. 397; Ex. R-2). The top two floors are primarily used to house patients. (Id.). There are 86 total beds in the facility spread between three units: Discovery, Horizon, and Compass. (Tr. 397; Ex. R-2). The Discovery unit, which is for adolescents, and the Horizon unit, which is for low-acuity4 adults suffering from illnesses such as depression and bipolar disorder, are both located on the second floor. (Id.). The Compass unit, which is for high-acuity adult patients suffering from schizophrenia and schizo-affective disorders, is located on the third floor. (Id.). The number of patients in the hospital at a given time fluctuates due to new patients being admitted or discharged.

Respondent employs approximately 220 staff members to provide care and services to its patients. (Tr. 58). Those responsible for providing direct care to patients include: doctors, registered nurses, social workers, therapists, and behavioral health associates (BHAs). Though all employees who provide direct care at Highlands are exposed to the hazard of patient-on-employee violence, most of the incidents discussed in this decision include the nurses, BHAs, and individuals who work in intake/admissions.

According to Respondent, its operations are governed by multiple regulatory and accreditation organizations, including the State of Colorado and the Federal Government. (Tr. 3136, 3379). Specifically, Respondent is accredited by a body known as the Joint Commission, licensed by the Center for Medicare and Medicaid Services (CMS), and regulated by the Office of Behavioral Health for Colorado and the Colorado Department of Public Health and Environment (CDPHE). (Tr. 3379, 4022, 4023). As such, any abatement determinations must operate within the boundaries of the accrediting and regulatory frameworks, including EMTALA,5 which requires Respondent to accept all patients with some very narrow exceptions. (Tr. 3467, 4029-4030).

The Court was presented with evidence of multiple incidents of patient-on-employee violence at UHS-Highlands, each of which relate to some aspect of Complainant’s abatement proposals. In lieu of reproducing a recap of each incident of workplace violence, the Court accentuates testimony it finds to be illustrative of the remaining issues central to this Decision. While the following reflect individual instances of violence, every employee who testified indicated he or she had been the subject of or observed multiple incidents of patient-on-employee violence during their tenure which was similar to the incidents exemplified.6

[redacted] was a nurse at Highlands for approximately 8 years, from May 2013 to April 2021. (Tr. 613). While he testified he had been assaulted “a lot” during his tenure, there were two incidents that occurred proximate to the inspection that bear mentioning. First, in December 2017, [redacted] testified he was hit in the face by a patient while he was sitting behind the nurse’s station.7 (Tr. 626-627). Second, in June 2020, [redacted] was attacked by a group of adolescents in the Discovery unit who had barricaded themselves in a room. (Tr. 660). According to [redacted], this event was essentially a repeat of an incident which had occurred the week before involving the same patients.8 (Tr. 660). This time around, [redacted] was repeatedly attacked by the patients as they came out of the room. While the first two were restrained by other staff members, the third and fourth managed to pull [redacted] to the ground and choke him with his lanyard and, subsequently, with a carotid choke hold. (Tr. 660-662). [redacted] suffered serious injuries resulting in two months away from work.

[redacted] worked as a BHA from June 2016 to May 2020. (Tr. 1087-1088). [redacted] suffered two significant injuries during her time at Highlands. In February 2018, her shoulder was dislocated while performing a hold on a large, out-of-control patient, who started the confrontation by reaching over the nurse’s station. (Tr. 1103-1105, 1118). According to [redacted], the patient had exhibited violent outbursts during the period leading up to the injury, as well as during a previous stay at Highlands. (Tr. 1099-1100, 1110; Exs. C-44, C-60). Notwithstanding her injury, [redacted] had to maintain the hold until she could be relieved. In July 2019, [redacted] was kicked in the head by a patient wearing boots, which managed to get onto the unit even though such footwear was not allowed by policy. (Tr. 1126). According to [redacted], the patient, who also had a history of violent behavior, was brought up to the unit by intake even though they were told there was insufficient staff for another patient on the unit. (Tr. 1125). Due to her injuries, [redacted] was out of work from July 2019 until April 2020, when she was cleared by her doctor. (Tr. 1128-1129).

Many other nurses and BHAs testified they had been attacked and/or injured while performing their jobs, including being kicked in the chest, having a dreadlock ripped from their scalp, being spit on, and bitten. (Tr. 801, 939, 948, 1727-1728). While violent incidents could, and did, occur at any time of day under different circumstances, the weight of the evidence suggests many of these incidents occurred in the evening or overnight when staffing levels were at their lowest. (Tr. 147; Ex. R-57). The evidence also indicates that a significant number of the injuries occurred when staff had to go “hands-on”, or place patients in holds to restrain them. (Tr. 86-87; Ex. C-4 at 8).

The following discussion addresses the ways Respondent attempted to address patient aggression through policy and procedures. These policies and procedures address how incoming patients are handled during the admissions and intake process, how employees are trained to act in response to an act of aggression, and how the information about particular acts of aggression is used for subsequent data collection, communications, treatment, and training.

Everyone testified their first response to a patient who is exhibiting signs of aggression is to engage in what is called Verbal De-escalation. Direct care staff are trained to first attempt to calm and redirect the patient from the source of their aggravation. (Ex. R-30). In some instances, however, verbal de-escalation is not possible or successful, and an employee will need assistance restraining an aggressive patient. In those instances, the employee can call for a Code Green, which is, for lack of a better term, a distress call to all available members of the hospital staff. Code Greens could be called in through the use of panic alarms (intake), verbally calling out to other members of the unit, walkie-talkies when they are available, or the intercom system. (Tr. 3132-3133).

While all staff members are expected to respond, if possible, there are often limitations on who can respond. For example, if an employee is engaged in one-on-one care with a patient under doctor’s orders, that employee cannot respond to a Code Green. (Tr. 95, 708, 3294). The same is true of staff members doing patient rounds, which have to occur every 15 minutes, and staff transporting patients around the facility for lunch and other activities. (Tr. 95, 708, 831, 2530, 3294). Even if staff members are designated as Code Green responders for their shift, there is no policy preventing them from being assigned to one-on-one care of a patient or other activities that would prevent them from responding to a Code Green. (Tr.377, 2350). Thus, the employees testified there were instances where: (1) calls for a Code Green were not responded to or required repeated calls; and (2) multiple Code Greens were called simultaneously, and there were not enough people to respond. (Tr.413, 961, 1149, 2296). This is so even though the staff roster for each shift has designated responders.

Respondent contends the typical response time to a Code Green is “within seconds”; however, now-former Director of Risk Management, Amy Petschauer testified she did not track response times. (Tr. 3495). Further, while many members of management downplayed the number of Code Greens called for and the average response to those Code Greens, staff members repeatedly complained of how long it took staff to respond to Code Greens. (Tr. 149, 464). This gap in perception is due, in no small part, to the fact that most management employees work during the daytime when both staff and management are available to respond. Comparatively, the evening and overnight shift employees do not have additional employees or management at their disposal.

Respondent is adamant that clinical management is the most effective way to materially reduce the incidence of patient-on-employee violence because the source of the violence is the patient’s mental illness. While Dr. Lipscomb agreed clinical management is a necessary aspect of reducing workplace violence, she noted it was not, of itself, sufficient. This is illustrated by numerous aggression propensity assessment forms and policies introduced by Respondent. Some of the forms were only partially implemented, implemented for short periods of time, or never used at all.

For example, there were programs such as the Managing Aggressive Patient Protocol, or MAPP, which was designed to identify aggressive tendencies of a patient prior to being admitted to Highlands. (Ex. C-13). This form was only in use for a short period of time, because employees did not find it helpful. (Tr. 714). Further, it appears the MAPP references other plans/programs that were never implemented. (TR. 591). Other forms, such as the Conduct Expectation for Aggressive Behaviors, Promoting Patient Safety Review, and Source of Aggression, were similarly short-lived or introduced but never actually used. (Tr. 715, 1156, 3221; Ex. C-15, R-23).

In response to an act of aggression requiring staff to place a patient in a hold, staff were required to fill out a Restraint/Seclusion packet. (Tr. 3265; Ex. R-9, C-42, C-46). The packet contains a treatment plan; however, the treatment plans were basically the same for all patients and, ultimately, did not change the course of treatment for an individual patient or patient management as it related to employee safety. (Tr. 419-420; Ex. C-42, C-46). This is illustrated in the cases of NS and HM, who had multiple documented restraints. (Ex. C-42, C-46 The treatment plans in their files are largely pro-forma with the exception of the specific event that precipitated the hold. (Ex. C-42; C-46). There is no indication changes were made to these patients’ treatment plans to address their “aggressive behaviors.” (Ex. C-42).

As will be discussed later, Respondent is not being cited for implementing programs that ultimately failed. However, it is noteworthy these attempts to address patient aggression on a purely individual, clinical level resulted in a patchwork of ad hoc solutions instead of being addressed by a comprehensive plan to deal with patient-on-employee violence, which is clearly endemic to the business of running a mental health facility.

Respondent contends its clinical approach to addressing patient-on-staff violence starts with the admissions process. Patients are typically admitted to Highlands on a referral from an emergency department and usually arrive by ambulance, but some come to the facility as walk-ins. (Tr. 4025). Once the prospective patient is allowed into the building, they are immediately separated from their belongings, which are checked with a metal detector, to ensure the prospective patient is not carrying prohibited items, including anything that could be used as a weapon. (Tr. 3331; Ex. R-34). Staff is also required to perform a “non-invasive body search” for the same reason. (Ex. R-34).

After the initial screening for contraband is complete, prospective patients meet with intake staff. (Tr. 3212). This consultation takes place in a consult room just down the hallway from the main entrance and is recorded on CCTV; though there is no indication these consultations are monitored in real time. (Tr. 67, 4026). Respondent says staff are required to wear a panic alarm during intake as a standard precaution for the possibility a prospective patient becomes violent; however, CSHO Stark was told the intake staff used their personal judgment regarding when to wear them. (Tr. 65-66; 4027). Intake staff conduct an assessment, which includes legal history, past hospitalizations, diagnoses, medications, and drug screening (if one has not been done at the referring emergency department). (Ex. R-7). Part of the reason for this assessment is to determine whether the patient is at risk of self-harm or harm to others. (Tr. 2894). If a patient was determined to be at high risk for aggression, a High Risk/High Alert Notification form, which is bright pink in color, was supposed to be filled out and included in the patient’s file. (See R-5 at Highlands 00624). Respondent also stated intake staff will contact the unit if the patient has been admitted to Highlands before. (Tr. 2898). Ultimately, the decision to admit is up to the on-call psychiatrist and is informed by current capacity and the medical needs of the patient, which may not be appropriate for Highlands.9 (Tr. 4024-4025).

While the policy governing Admissions/Intake appears thorough and designed to prevent violent behaviors, the reality of the process was much different. Direct care staff were not always consulted about re-admitting a previous patient, even when the patient had a history of violent behavior at Highlands. For example, Patient JH was re-admitted to Highlands even though he had assaulted [redacted], who had pressed charges against JH. (Tr. 637-638). In addition, intake would sometimes admit a patient to a unit that was either understaffed or, at that particular moment, was not ready to admit the patient to the unit. In one instance, [redacted] was seriously injured by an extremely aggressive patient admitted before the unit was prepared to admit him/her. (Tr. 1124).

Even if a patient’s Comprehensive Assessment indicated a history of aggressive/assaultive behavior, this was not indicated on the High Risk/High Alert Notification form. For example, Patient JH assaulted an employee in December 2017, but upon being readmitted six months later, his HR/HA Notification form did not indicate an assault risk. (Tr. 646, C-43 at 11). Patient MG’s assessment indicated a history of violence and aggression but was assessed as low- to no-risk on the HR/HA form. (Tr. 1241-1243; Ex. R-5 at Highlands 00624). Both JH and MG ended up assaulting staff after they had been readmitted. (Ex. C-43 at 12; Ex. R-5 at Highlands 00626).

According to Respondent’s policy, the only time a staff debrief is required is after a patient has been restrained. (Tr. 163-64). According to Respondent, this is required by CMS regulations and the State of Colorado. See Resp’t Br. at 32. The staff debriefing can be found in the Restraint and Seclusion packet, which was included in the numerous redacted patient files submitted into evidence. (See, e.g., C-40 to C-49). Respondent contends debriefing also occurs after any incident of patient-on-staff violence; however, the evidence shows staff debriefs, if they occurred, were cursory at best. Respondent’s policy also indicates camera review of such incidents should occur for the purposes of training. (Ex. C-11).

Two of the staff members, Mr. Harriott and [redacted], stated they had not participated in a debrief after particularly violent encounters even though the documentation indicated a debrief had occurred. (Tr. 808, 813, 1110). This was the case for many of the staff members who testified. (Tr. 637, 663, 954, 1279, 1747, 2221-2222). In [redacted]’s case, the debrief indicated a camera review had occurred, but she testified she was not included in any such review. (Tr. 1120-1123; Ex. C-26). Further, on two separate occasions, [redacted] asked to see video of the encounters but was told it was not available to her. (Tr. 1115, 1135). To the extent debriefings did occur, they were often cursory checks to ensure no one got hurt during the hold. (Tr. 1738). Indeed, Respondent’s own review of its files indicated staff debriefings did not include all parties involved in the incident. (Tr. 2410-2411; Ex. C-35 at 2).

It appears management debriefings and camera reviews are, at best, lip service to regulatory requirements, as opposed to an integral part of a robust workplace violence prevention program. Nearly every employee who testified indicated the debrief was little more than a check to ensure there were no injuries rather than a method to evaluate how a situation could have been handled more safely or identify needed changes in procedure. According to Mark Helferich, this was likely because there were “not enough staff to entertain those things”. (Tr. 943-944). As for camera reviews, Mr. Helferich testified he had participated in at least 50 Code Greens and had participated in only one camera review. (Tr. 954). Lori Ayala, a house supervisor, testified, at least as of February 2020, she had not participated in a camera review as part of the debriefing process. (Tr. 3315).

Respondent’s reporting system for employee injuries during the relevant period was an amalgam of multiple, different sources without any consistency as to where information should be documented, what should be reported, or how the data should be compiled. This lack of consistency is reflected in both the data and the testimony provided at trial.

Based on CSHO Stark’s review, Respondent’s workplace violence injuries were supposed to be documented in-house in Employee Accident Reports (EARs) until mid-2018, when it began using a third-party, called Sedgwick, to record employee injuries. (Tr. 86; Ex. R-35). Using Sedgwick requires either the employee or their supervisor to call a hotline to report the injury. The problem, as noted by some of the employees at trial, is aggressive or violent behavior “was so normalized” it did not occur to them to report injuries, including those not requiring immediate medical attention. (Tr. 1901, 1977, 2080). Respondent also said it uses the MIDAS system to record injuries; however, closer review of that system reveals it is a patient-focused system with no section for employee injuries, except for the “Comments” section, and is entirely dependent on the individual entering the data. (Tr. 2400). A clear illustration of the inadequacy of the MIDAS system for recording employee injuries was illustrated by two examples: (1) when [redacted] was kicked in the head by RM, the report simply coded the patient as “out of control”, but did not mention the injury; and (2) when one patient managed to punch a BHA, slap a charge nurse, bite a staff member, and spit in the faces of two others, the MIDAS report indicated “no injury involved” because the outcome refers to the patient only. (Tr. 2397-2401; Ex. C-51). Thus, there is no data point to be gleaned from those reports elucidating consequences to staff members in the same manner as records of patient injury in the same report. (See, e.g., Ex. C-43 at 2, C-47 at 1, C-51).

Exacerbating the problem of inconsistent reporting in multiple sources and formats, CSHO Stark found employee injuries that were either not recorded or were improperly recorded in Respondent’s OSHA 300 logs and EARs that were made available to her. (Tr. 87-88). This included [redacted]’s injuries, neither of which were coded as lost workdays or restricted duty, notwithstanding both requiring long-term absences and restrictions upon her return. (Tr. 1113-1114, 1132; Ex. C-7 at 3, C-8).

Given the inconsistencies discussed herein, the Court is skeptical about Respondent’s claims about the downward trend in injuries found by CSHO Stark. (Tr. 77, 375; Ex. C-4). Those numbers were based purely on the OSHA 300 logs, which, as shown above, do not reflect all the injuries that occurred, nor do they accurately reflect the injuries that are reported. Ms. Petschauer testified injury tracking was the responsibility of the Human Resources Department. (Tr. 3415). While Respondent’s current Human Resources director, Susan Coulter, testified about tracking and auditing injury reports, she noted she did not begin at Highlands until May 24, 2019, which was after the Citation was issued. (Tr. 3866). She has not reviewed or audited the OSHA logs prior to the beginning of her employment. (Tr. 3887). Thus, prior to Ms. Coulter’s tenure, the data is questionable at best. This, like other matters discussed herein, highlights another problem with the lack of a comprehensive WVPP: lack of accurate data, tracking, and, subsequently, trend analysis.

In addition to Code Greens, there have been numerous instances where the violent behavior of a patient or group of patients exceeded the staff’s ability to handle the situation. In such cases, the police have been called. Exhibit C-10 shows a summary of the of the police reports filed for a period of three years leading up to the inspection. (Tr. 88-89; Ex. C-10). Some employees testified management discouraged calling law enforcement; however, it does not appear anyone was prevented from doing so. That said, at the very least, there appears to be inconsistent communications and confusion regarding the proper protocol for contacting law enforcement, which is another indication Respondent’s WVPP is not being properly or clearly communicated.

A consistent refrain from nearly every employee who testified was a lack of adequate staff to properly carry out their jobs, provide care, and feel safe. (Tr. 464, 440, 696, 820, 830, 958, 960, 1290, 2097). This sentiment was echoed in numerous communications to management, between management, and even from a UHS-Delaware Divisional Nursing Director, who audited UHS-Highlands’ operations and found, “[a]t [the levels provided for in the matrix] it may be difficult for staff to get breaks, handle multiple admissions and/or manage an acute unit.” (Ex. C-25 at 5). Complaints about a lack of adequate staff were also included in multiple staff surveys and daily house supervisor reports. (Tr. 703-705; Ex. C-17, C-20, C-23, C-37). While employee suggestions and complaints are not, of themselves, sufficient indicators of the need for additional staff, the following discussion, as well as the legal argument provided later in this opinion, illustrates the level of staffing provided for by Respondent was insufficient for the same reasons identified by the UHS-Delaware Divisional Nursing Director.

Respondent sets staffing levels for each unit based on a staffing matrix, which sets the level of BHAs and nurses according to the number of patients, or census. (Tr. 146; Ex. C-16). Respondent employs a staffing coordinator, who is responsible for ensuring adequate staffing levels and creating daily assignment sheets. (Tr. 3529). This position was held by Jesus Gaspar during the period leading up to the inspection until he was replaced by Krystal Garnhart in 2020.10 (Tr. 3525, 4021). According to Ms. Garnhart, she maintained the same procedures Mr. Gaspar employed during his tenure as staff coordinator. (Tr. 3526). Ms. Garnhart testified she used the matrix as the baseline for staffing and would adjust staffing levels upward based on additional factors including acuity, planned admissions, and patients requiring special precautions, such as one-to-one observation. (Tr. 3529-3530). The house supervisor may further adjust the assignments based on skill and experience. (Tr. 3229).

The matrix establishes a general standard but does not account for acuity adjustments or special orders, such as one-to-one observations,11 nor is there a policy or procedure for such adjustments. (Tr. 697, 1291, 2314). Further, the matrix itself is the same for each of the three units, notwithstanding the baseline differences in acuity between the different populations. (Tr. 2328; Ex. C-16). Interestingly, the matrix reduces staff down to one nurse and one BHA for each unit during the overnight shift based on an expectation most patients are asleep. (Ex. C-16). Based on the testimony of the nurses and BHAs, as well as the statistics regarding confrontations, however, it appears a substantial number of confrontations occur during the evening and overnight shifts. (Ex. R-57). Complicating matters staff-wise, Respondent would count staff on restricted duty or performing one-to-ones as part of the staff count, even though they were not able to perform the full panoply of duties required to run the unit. (Tr. 146, 664, 720, 1112, 1920).

Further complicating matters, employees would frequently call in sick or would not come in for various reasons. (Tr. 1142-1143). Respondent said it would send out text blasts and, in some cases, offer incentives for people to come in during their days off; however, this was not always successful, and employees would have to work short-staffed or, on some occasions, alone. (Tr. 432-433, 1142-1143).

The impact of being short-staffed was felt acutely by the staff. According to [redacted], Harriot Helferich, [redacted], Broadbent, Panza, and Graumann, the units were, on occasion, understaffed even according to the matrix. (Tr. 432-433, 696, 820, 958, 1140, 1290, 2087, 2223). This was documented in multiple survey comments, and in daily house reports, which noted staff felt discouraged, unsafe, and unable to complete the tasks expected of them. (Tr. 703, 2318; Ex. C-17, C-22, C-23, C-37). Also, as previously noted, this was noted by the Divisional Director of Nursing in March 2018, who observed the matrix levels would negatively impact the nurses’ and BHAs’ ability to handle the basics of their job, let alone be able to take breaks without leaving their coworkers alone. (Ex. C-25).

Difficulties described by the Divisional Director of Nursing’s assessment regarding staff played out at Highlands in different ways. In some instances, there were not enough employees to respond to Code Greens. (Tr. 830, 1103, 1294). Multiple employees testified they either had to work on a unit alone during an overnight shift or would be left alone on the unit to perform observation rounds or other duties while their unit colleague would take groups of patients to meals or group therapy or had to respond to a Code Green on another unit. (Tr. 698, 1783, 1913, 2008, 2094). The short-staffing problem not only impacted staff’s ability to perform their jobs, but also was noted to negatively impact patient care, which had an ancillary impact on employee safety. For example, [redacted] testified group therapy sessions and activities would be cancelled because staffing was insufficient. (Tr. 1117). Dr. Argumedo testified these activities are instrumental to clinical management and the failure to engage patients in a meaningful way, especially among the adolescent population, can result in restlessness and acting out in unpredictable ways. (Tr. 1444, 1507). This was noted by staff members at trial. (Tr. 1137).

A particularly illustrative example of the impact on short-staffing was summed up by a staff nurse, Roseanne Tulley, in an email that was sent to Jill Orr, now-former Assistant/Interim Director of Nursing, who, in turn, forwarded the report to both the CEO, Amy Alexander, and Aaron Van Dam, the director of nursing. (Ex. C-33). According to Ms. Tulley, “[S]taff has to continually open and close doors for bathroom trips. Nurse’s station desk is bombarded with patient requests. Staff cannot spend even five minutes with a child who needs some processing and it’s impossible to keep an eye on 20 plus patients safely.” (Ex. C-33). In her estimation, Respondent needed one staff at the desk, one at the nurse’s station, one in the day room, one in group sessions, and one-to-two staff members doing five- and ten-minute rounds. (Tr. 2348; Ex. C-33). Even then, Ms. Tulley noted staffing at her suggested levels was not sufficient for any one-to-one orders which may be in place. The Court finds this compelling, not so much because Ms. Tulley’s assessment of how many staff members were needed should be adopted in its entirety, but as an illustration of the real-life, day-to-day duties staff were called upon to carry out in addition to the primary charge of providing therapeutic care. Her account was echoed by other BHAs and nurses, who talked about having to man the desk for mundane tasks, such as providing telephone access, bathroom breaks, and smoke breaks, all of which needed to be carried out in addition to doing observation rounds (whether in 5-, 10-, or 15-minute increments), one-on-ones, medicine dispensing, charting, and other medically required duties. (Tr. 804, 928, 1513-1514).

Respondent attempted to address the issue of short-staffing using multiple stratagems, including text blasts and bonuses/incentives. Respondent’s most direct attempt to address staffing issues in a concrete way was to implement a position known as a Qualified Safety Responder, or QSR. (Tr. 1143, 2342). The guiding idea behind the QSR position was to have a BHA who could float between units and provide assistance as needed. Unfortunately, however, given the problems with call-outs, the QSR was often used as a fill-in. (Tr. Id.). Thus, the only remaining float personnel on overnight shift was the house supervisor. (Tr. 2335, 3284). The difference between daytime and evening/overnight is important because, as noted above, the daytime shifts had multiple administrators and management personnel available. Those same people were not available at night. Further, according to Ms. Orr the QSR position was not consistently implemented, even at night, and was ultimately scrapped due to purported budgetary issues, not because it was ineffective. (Tr. 2344-2345).

Lori Ayala provided unintentional, yet compelling, testimony with respect to the QSR position and staffing matrix. In response to a question about the matrix only requiring two overnight staff, Ms. Ayala stated this would only happen when the census on a unit fell below eleven, which only happened three times in the last two years. (Tr. 3286). Apparently, the matrix was changed at some point after the inspection in 2018. (Tr. 3285). Prior to that, a unit could have up to 17 patients covered by only two staff during the evening/overnight shifts. (Tr. 3285-3286; Ex. C-16).

A consistent refrain from experts, administrators, and staff alike was the importance of communication among direct care staff relating to patient aggression. The Court has already addressed how information regarding patient aggression is communicated from intake to the unit, which includes the comprehensive assessment, High Risk/High Alert Notice, a pre-admission nurse-to-nurse form, and a safety plan form. (Ex. R-21, R-29). The nurse-to-nurse form provides nurses at Highlands with contact information of the patient’s previous provider to gain a clearer picture of the patient’s history, and the safety plan form contains therapeutic information provided by the patient, such as what helps them calm down and what activities they like to do. Equally essential are communications between shifts about the state of the milieu and aggressive patients.12

Respondent creates a snapshot of the unit using a census board and “Picture of a Safe Milieu” form. (Tr. 3219, 3966; Ex. R-24). Both contain information about each patient on the unit, including their doctor, social worker, special precautions, and appropriate interventions. Prior to and shortly after the inspection in October 2018,13 the primary communications regarding patient aggression between staff and/or shifts were documented in the shift hand-off report, which provided the oncoming shift with information about patient aggression issues on the previous shift. (Ex. R-13).

BHAs and nurses noted several problems with the communications at shift changeover. Everyone agreed information was exchanged; however, staff noted they only had about 15 minutes to communicate information about the entire unit in that period of time. (Tr. 142). Specifically, staff had to convey information about medical issues, medications, interpersonal issues between patients, suicide precautions, fall risks, as well as incidents of patient aggression towards staff. (Tr. 386, 2231). Further compounding the problem, BHAs typically did not review patient records, the comprehensive assessment, the previous days’ observation rounds, nor, according to Ms. Orr, were they required to do so. (Tr. 143, 1099, 1154, 2230, 2360).

Another key source of information about the state of the milieu and patient aggression were the various meetings held by the treatment teams and administrators. While BHAs participated in meetings at shift changeover, they did not attend treatment meetings or any other meetings directly or tangentially related to staff safety.14 (Tr. 2230, 3291). Various meetings which were held to discuss safety included the Daily Nurse’s meeting, treatment plan team meetings, daily flash meetings, Performance Improvement Committee meetings, and Patient Safety Committee meetings.

The daily nurse’s meeting was essentially a nurse-specific version of the shift changeover exchange of information. (Tr. 3915-3916). The treatment plan team meetings included discussions of each patient and the course of treatment, but it would also include a discussion of any aggressive behaviors and how to address those behaviors in a clinical way. (Tr. 3228, 3918; Ex. R-28). If appropriate, a behavioral plan would be developed to address those behaviors and included in the patient’s chart. (Tr. 2941, 3921; Ex. R-28). Daily Flash meetings, which were typically an administration affair, took a birds-eye view of the facility as a whole, but also addressed patient aggression incidents and staffing concerns. (Tr. 3392, 4020).

While the nurse’s, treatment plan, and flash meetings all addressed specific incidents of aggression from the standpoint of treatment and staffing, the Performance Improvement Committee (PIC) and the Patient Safety Committee (PSC) used those incidents to address policy and procedure. The PIC reviewed injuries to determine root causes, whether an injury could have been prevented, and whether any patterns emerged. (Tr. 3388, 3868). The PSC is comprised of administration and management (CEO, CMO, CNO, director of nurse management) and was intended to review big picture issues, including month-to-month and year-to-year trends, major incidents, errors, restraints, and discharge, amongst other things. (Tr. 3381-3382). Respondent contends staff members are encouraged to attend these meetings but seldom do so.

According to staff members and management, staff did not attend these meetings, nor were they expected to. (Tr. 481, 846, 1308, 1926, 2017, 2304). Further, while Respondent held town hall meetings, which were intended to include staff, they were held during the day, which prevented the overnight staff from being able to attend. (Tr. 483). Staff also indicated they raised safety issues and other matters with management but never received follow-up or felt as if their suggestions were taken seriously. (Tr. 722, 2110). As was the case with staffing, direct care workers repeatedly identified lack of communication as a problem in the employee surveys from 2016 to 2018. (Ex. C-37). Specifically, employees noted “communication is lacking: across levels of the organization [and] in team meetings”, staff were “not involved in decision making and not told of changes”, “my opinion doesn’t matter”, and “safety is not seen as an issue.” (Tr. 2436-2437; Ex. C-37).

Complainant argues, although Respondent had a WVPP in place at the time of the inspection, such plan had multiple shortcomings, including (1) a lack of meaningful review of the policy since it was created; (2) a lack of clarity regarding which department is responsible for implementing and providing training on it; (3) a lack of a clear definition of workplace violence, which leads to difficulties in tracking trends and data; (4) reliance on outdated, inappropriate, or non-existent forms or documentation of patient aggression; (5) failure to follow through on program requirements, including camera reviews of violent incidents; and (6) characterizing patient-based initiatives as elements of employee participation in the plan. Respondent, on the other hand, argues Complainant has put “form over substance” and lists all the clinical-based interventions it uses to control patient aggression. The problem, as will be shown later in this decision, is not that Respondent is missing a piece of paper called a Workplace Violence Prevention Program, or that it utilizes a clinical approach to patient-on-employee violence. Rather, the problem is Respondent’s WVPP is insufficient in design and, in practice, is a patchwork amalgam of ad hoc interventions with no clear line of responsibility for its implementation. Thus, this is a problem of both form and substance. The following are just a few examples of the program’s elements and their insufficiency to accomplish the goal for which they are designed.

The WVPP states its focus is on Type 2 violence, which is “violence committed by patients upon staff members.” (Ex. C-11). This is the extent of the definition of workplace violence.15 According to Dr. Jane Libscomb, this definition is inadequate because it does not indicate what constitutes an act of violence, which, in turn, impacts Respondent’ ability to prevent and track violent acts. (Tr. 3655; Ex. C-62 at 25). This failure of clarity can be seen in the section on Management Commitment, which sets goals for reductions in violent incidents and identifies how “incidents of patient aggression, restraint and seclusion will be reported”, including HPRs (MIDAS reports) and Employee Accident Reporting. (Ex. C-11). As recounted above, however, MIDAS reports are patient-based and the Employee Accident Reporting procedure was stopped at some point in 2018 when Respondent switched over to the Sedgwick reporting line. (Tr. 85). The plan also calls for camera review of all incidents resulting in restraint or seclusion and a debrief of all staff involved; however, the evidence shows camera reviews rarely occurred and debriefs were not consistently conducted with all staff involved.

With respect to Employee Participation, the WVPP lists a series of purported initiatives, which, in reality, read like job requirements and have little or nothing to do with how employees should participate in the process of developing violence prevention initiatives. (Ex. C-11). Regarding Aggression Analysis and Surveillance, Respondent again refers to HPRs (MIDAS) and EARs, which are either inadequate for tracking employee injuries or are no longer in use. As an illustration of this, Ms. Petschauer created a Risk Management Report for the first quarter of 2019 based on an analysis of HPRs, but no information was included regarding employee injuries or patient aggression toward staff. (Tr. 2414-2415; Ex. C-36). Ultimately, it appears Respondent has a written WVPP, but neither the document itself nor its implementation appear to be coherent, comprehensive, or effective to address workplace violence toward workers.

According to Respondent, all its employees receive training during orientation and periodically throughout the year at regularly scheduled “competency fairs”, where employees sign up for refresher training and develop practical skills (HWC was not available to take during the fair). (Tr. 3119-3220; Ex. R-49). All direct care staff are trained in Verbal De-Escalation and Handle With Care, which are the two primary programs for dealing with an in-progress act of aggression. During orientation, employees are provided classroom training in both disciplines. Employees must take an assessment and be certified in both verbal and physical management of patients before they can work on one of the units. (Tr. 3109-3111; Ex. R-32). The training itself is provided by Respondent’s resident master trainer, Mike Carstens. According to Mr. Carstens, the HWC program is designed to account for differences in size between the patient and the staff member initiating the hold. (Tr. 3114). Carstens also testified he attempts to respond to all Code Greens in order to either participate or coach the staff members engaging in a hold. (Tr. 3066-3067). The training includes self-defense techniques and counsels avoidance of a conflict if at all possible, including retreating to a locked room if engaging in a hold is not feasible. (Tr. 3082-3083).

Many of the employees took issue with how the training was carried out, the techniques taught, the lack of buy-in from certain members of the staff, and the lack of follow-up training. According to multiple employees, the practice methods used in training were not at all like the real-life violent situations they confronted while on the unit. (Tr. 1160, 2291). Specifically, they claimed placing holds on Mr. Carstens was not realistic when compared to dealing with the unpredictable nature of a psychiatric patient having a violent episode. Further, employees complained they were instructed at least two employees were necessary for every hold but experienced a number of situations where they had to go “hands-on” without the assistance of another staff member. (Tr. 1785, 3116). Employees also testified they were aware of staff who were reluctant or even refused to participate in holds. (Tr. 720, 1920). This included [redacted], who refused to participate in holds because of her age and the condition of her knees. (Tr. 1920). She testified Mr. Carstens was aware of this.

Respondent claims it provides refresher training during its training fairs, after debriefs identifying improper use of HWC technique, and at the request of employees. (Tr. 3061-3062). Respondent’s own internal communications belie this claim. In an email dated October 2018, around the time of the inspection at issue in this case, Mr. Carstens noted many members of their “veteran team” had not re-certified in verbal de-escalation or HWC “maybe since their [new employee orientation] however many months/years ago that was”. (Ex. C-29; Ex. 2369-2371, 3150-3151). Thus, notwithstanding Respondent’s claims regarding its tracking software for refresher training, it appears a fair number of employees slipped through the cracks. (Ex. C-25 at 4). Further, in response to a purported assessment of one of her holds, [redacted]’s file indicated she had received refresher training in HWC; however, at trial, [redacted] testified she had not received any such follow-up training nor had she been told how or why the method she employed was insufficient.16

As regards the WVPP, Respondent’s training program is patchwork, at best. CSHO Stark found, and Respondent’s employees confirmed, Respondent does not provide training specific to the WVPP. (Tr. 137, 845, 988, 1150). Indeed, two of the people responsible for the program’s implementation, Ms. Orr and Ms. Petschauer, testified they did not know whether the WVPP was covered during new employee orientation. Thomas Braswell, the director of facilities, apparently provides training on doing environmental rounds, which are used to identify unsafe objects in the milieu, and workplace violence coming from outside the facility. (Tr. 3341). As an illustration of its training, Respondent introduced a presentation entitled, “Preventing Workplace Violence”; however, upon closer examination of the presentation, there is nothing specific to Respondent’s worksite as a stand-alone psychiatric facility nor does it specifically address the principal cause of injury to employees, which is patient-on-staff violence. (Ex. C-12). Further, when Complainant presented it to staff members at trial, many did not recall having received that particular training. (Tr. 1151, 1297).

In addition to policy-based interventions, Complainant identified several physical implements/interventions it believes Respondent could institute to materially reduce the hazard of workplace violence. Each of Complainant’s suggestions is essentially a modification to existing engineering controls.

During CSHO Stark’s inspection, she discovered a number of violent incidents occurred around the nurse’s station, which is located in the center of each unit. (Ex. R-2). According to Respondent, the station is so designed in order to give nurses and BHAs a clear view of the entire unit from one location. The station itself is not enclosed, and the counter is approximately four to five feet above the ground. (Tr. 62-64; Ex. C-4 at 12). Nurses and BHAs enter the station through a swinging door that does not lock, which means patients can just as easily enter the nurse’s station. (Tr. 62, 476, 627; Ex. C-4). On many occasions, patients have done just that. (Tr. 981, 3364; Ex. C-7 at 16). In addition to entering through the doorway, patients have also climbed and jumped over the counter to attack staff members or grab items such as pens, pencils, and other job-necessary items to use as weapons. (Tr. 409, 478, 673, 838, 1304, 3364). The Divisional Nursing Director’s audit noted similar problems with patient access to the nurse’s station, especially when staff is conducting rounds or engaged in other duties requiring them to be away from the station. (Ex. C-25 at 6).

Respondent appears to downplay these concerns in two ways. First, Respondent contends a change to the nurse’s station would not necessarily prevent aggressive actions toward staff once they leave the nurse’s station, thus calling into question whether the change would be materially effective. Second, Respondent contends fully enclosing the nurse’s station could have a deleterious effect on the therapeutic milieu, because it would place a physical barrier between patients and staff, thereby setting the two parties in opposition to one another. As will be discussed later in this opinion, the evidence paints a different picture.

Complainant identified incidents of aggression where patients picked up furniture, such as chairs and desks, and used it as a weapon. (Tr. 410, 667, 1788). Respondent was aware of these incidents and agreed it would improve staff safety if all furniture was secured; however, it also noted securing or weighing down all furniture may be impractical for sanitary reasons, such as cleaning the cafeteria chairs after meals. That said, at some point in 2019, Respondent purchased weighted furniture and found ways to increase the weight of existing furniture to reduce the likelihood it would be used to harm staff and other patients. (Tr. 2887).

At the time of the inspection, Respondent had approximately 100 cameras located throughout its facility. (Tr. 3346-47; Ex. R-2). The camera feeds can be viewed at nurse’s stations, intake, and in administrators’ offices. (Tr. 3346, 4026). At the time of the inspection, the cameras did not possess zoom capability or sound recording; however, since the inspection, Respondent upgraded the cameras to include zoom capability and added 10 additional cameras. (Tr. 3346-42). While the cameras have been used incidentally for surveillance, for the most part it appears they were used for review of violent incidents according to Respondent’s WVPP.

Although there was one instance where an employee saw an encounter developing over the CCTV system and was capable of responding to a violent incident prior to a Code Green being called, this appears to have been a singular event. Otherwise, Complainant’s evidence was, more or less, limited to the staff’s belief that additional surveillance would make them safer.

At the time of the inspection, each unit had two walkie-talkies, but employees were only required to carry them when they went off unit for meals, groups, or activities. (Tr. 65, 390). Intake employees were provided with panic alarms, but, as noted above, Respondent did not specifically mandate their use and they were not made generally available to other staff members. The only other devices available for hospital-wide staff communication were the telephones behind the nurse’s stations. Approximately a year or so after the inspection, Respondent made the decision to increase the number of radios so employees could always carry them. (Tr. 3160, 3358).

The analysis of any workplace violence case is a complex and nuanced exercise requiring the input of experts in the industry in question. This is particularly the case here, where the Court is asked to: (1) consider the effectiveness of actions taken by Respondent to guard against the hazard of workplace violence in an environment where acts of aggression are endemic, and (2) based on that determination, consider whether Complainant’s proposed methods of abatement would be materially effective at reducing the hazard without impacting Respondent’s core mission.

The following is just a brief snapshot of the experts’ respective backgrounds and the testimony they provided in support of, or contrary to, (1) and (2).

Dr. Argumedo is a board-certified psychiatrist in general psychiatry and forensic psychiatry. (Ex. C-63). After medical school, Dr. Argumedo completed a four-year residency at Georgetown University Medical center, during which time she became chief resident and was responsible for the hospital’s inpatient psychiatry unit, and a forensic psychiatry fellowship at the Medical College of Wisconsin. (Tr. 1421-1425). After her residency and fellowship, Dr. Argumedo has served in multiple capacities, including performing competency assessments for criminal courts and serving as a testifying expert. (Tr. 1427-1428). Dr. Argumedo also worked for a free-standing psychiatric hospital, acting as service director for its Consultation Liaison Service, as well as its intake department, where she worked on the hospital’s intake procedures. (Tr. 1428-1431).

Dr. Argumedo’s focus at trial was Respondent’s clinical methodology for managing patient aggression. Ultimately, Dr. Argumedo concluded Respondent’s clinical focus was insufficient for the purposes of addressing patient-on-employee violence and found Complainant’s proposed abatement methods would materially reduce the hazard. Respondent contends Dr. Argumedo’s discussion and analysis of Respondent’s clinical methodology is an inappropriate area of inquiry for OSHA. The problem with this argument, however, is Respondent also asserted clinical management is the principal, if not only, method for mitigating patient-on-employee violence. As such, it has placed the clinical management of its patients at issue. Respondent cannot be allowed to use its patient management as both sword and shield.

Of the three experts providing testimony, Dr. Lipscomb was the most well-versed in the arena of workplace violence. Dr. Lipscomb has conducted studies, researched, written, and consulted on this topic for over 30 years. (Ex. C-61). Dr. Lipscomb started her career as a nurse after receiving a B.S. from Boston College, which she followed shortly thereafter with a Master’s in occupational health nursing from Boston University/Harvard School of Public Health and a PhD in epidemiology and occupational health from UC-Berkley. (Ex. C-61). Dr. Lipscomb has served as a professor, researcher, CDC/NIOSH Senior Scientist, and consultant to various government agencies. Of note to this case, Dr. Lipscomb has made patient-on-worker violence in behavioral health facilities a particular focus of her research, noting it is “such a high risk setting within all of healthcare.” (Tr. 3591). This includes researching the implementation of OSHA’s Guidelines on Violence Prevention in Mental Health in New York State mental health facilities and consulting the Maryland Department of Health and Mental Hygiene to help them improve their own workplace violence training in inpatient psychiatric facilities, amongst others. (Tr. 3594, 3599).

Dr. Lipscomb, much like Dr. Argumedo, analyzed Respondent’s methods for preventing workplace violence; however, rather than focusing on the clinical emphasis of Respondent’s prevention plan, Dr. Lipscomb was concerned with how Respondent’s methodology compared to industry studies and guidelines related to violence against healthcare workers. Further, Dr. Lipscomb discussed the importance of a comprehensive and internally coherent WVPP and illustrated how Respondent’s plan fell short in multiple areas.

Respondent argues Dr. Lipscomb’s testimony is not persuasive because she did not conduct or cite a study that unequivocally proves Complainant’s proposed abatement measures, taken individually, were insufficient to materially reduce the hazard of workplace violence. In some instances, Respondent argues its own expert, Dr. Marc Cohen, identified studies where some of the proposals supported by Dr. Lipscomb’s were contraindicated in the research, especially in the arenas of staffing and how certain abatements may impact the therapeutic milieu. These reservations are addressed herein.

Dr. Marc Cohen received his medical degree from the University of Southern California School of Medicine. Somewhat similar to Dr. Argumedo, Dr. Cohen completed a four-year psychiatric residency, followed by an additional year of training in forensic psychiatry. (Tr. 2441). Dr. Cohen testified he has been recognized as a permanent expert witness in Los Angeles County and is currently associate clinical professor in the Department of Psychiatry at UCLA School of Medicine. (Tr. 2446). In addition to serving as a professor, Dr. Cohen trained residents and treated veterans on an outpatient or partial hospitalization basis at the Veterans Affairs Hospital in Los Angeles, which is one of the UCLA teaching sites. (Tr. 2447). Following that position, Dr. Cohen worked at Olive View Medical Center, which was another teaching facility through UCLA. During his tenure at Olive View, Dr. Cohen briefly oversaw a secondary inpatient unit and worked in the psychiatric emergency room. (Tr. 2451-2452). Ultimately, Dr. Cohen testified his background in forensic psychiatry uniquely situated him to assess a patient’s violence risk, which is in line with Respondent’s claims regarding a clinical approach to patient-on-staff violence prevention. (Tr. 2456).

Dr. Cohen’s testimony, for the most part, focused on the engineering and control-based abatements proposed by Complainant and how they could negatively impact the therapeutic milieu. He also highlighted studies regarding the potential negative impact of increased staffing on patient-on-staff violence and the difference between staff perception of safety resulting from a chosen abatement and the actual impact a chosen abatement has. Dr. Cohen also opined Respondent’s WVPP exceeded industry standards for such a program.

Although Dr. Cohen clearly has extensive experience with clinical management of psychiatric conditions and forensic psychiatry, he does not have substantial experience working in an inpatient psychiatric unit and testified he had virtually no experience performing psychiatric intake or developing a comprehensive WVPP in the context of an inpatient psychiatric hospital. Thus, while he was well-suited to providing an analysis of Respondent’s clinical management program and methodologies for assessing individual violence risk, workplace violence in an inpatient setting is not an area Dr. Cohen has done research, lectured, or studied outside the confines of this case. (Tr. 2627, 2641, 2775-2776).

  Complainant alleged a serious violation of the Act in Citation 1, Item 1 as follows:

OSH ACT of 1970 Section 5(a)(1): The employer did not furnish employment and a place of employment which were free from recognized hazards that were causing or likely to cause death or serious physical harm to employees in that employees were exposed to physical threats and assaults by patients:

Among other methods, feasible and acceptable means to abate the WPV hazards at HBHS include implementation of a comprehensive and effective WPV prevention program which includes engineering and administrative controls, as well as training to reduce WPV hazards. These controls include:

Engineering Controls:

Administrative Controls

Training

See Citation and Notification of Penalty at 6-8.

The general duty clause requires every employer to provide its employees with a workplace “free from recognized hazards that are causing or are likely to cause death or serious physical harm.” 29 U.S.C. § 654(a)(1). Typically, in order to prove a violation of Section 5(a)(1) of the Act, Complainant must show: (1) there was an activity or condition in the employer’s workplace that constituted a hazard to employees; (2) either the cited employer or its industry recognized that the condition or activity was hazardous; (3) the hazard was causing or was likely to cause death or serious physical harm; and (4) there were feasible means to eliminate the hazard or materially reduce it. Waldon Health Care Ctr., 16 BNA OSHC 1052, 1058 (No. 89-3097, 1993). The evidence must also show the employer knew, or with the exercise of reasonable diligence, could have known of the hazardous condition. See Otis Elevator Co., 21 BNA OSHC 2204, 2207 (No. 03- 1344, 2007).

As noted above, the parties stipulated to all but the fourth element of a general duty clause violation: whether there were feasible means to eliminate or materially reduce the hazard. Prior to addressing the feasibility and efficacy of Complainant’s proposed methods of abatement, however, Complainant must show “as a threshold matter, that the methods undertaken by the employer to address the alleged hazard were inadequate.” Integra Health Management, Inc., 2019 WL 1142920 at *12 n.14 (No. 13-1124, 2019); see also Sea World of Florida, LLC v. Perez, 748 F.3d 1202 (D.C. Cir. 2014) (holding adequacy of employer’s precautions against a hazard is measured against precautions taken by “a reasonably prudent employer familiar with the circumstances of the industry”). If Respondent’s existing means of abatement are determined to be inadequate as compared to precautions taken by a reasonably prudent employer, Complainant must then show: (1) the 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 Health Management, Inc., 2019 WL 1142920 at *12 (citing Beverly Enters., Inc., 19 BNA OSHC 1161, 1190 (No. 91-3344, 2000) (consolidated)).

  Complainant asserts Respondent failed to take reasonably prudent measures to materially reduce the hazard of patient-on-staff violence. Complainant argues the inadequacy of Respondent’s attempts to mitigate the hazard are illustrated by the “frequent extent to which employees continue to be injured as a result of patient aggression.” Compl’t Br. at 55 (citing SeaWorld, 748 F.3d 1202, 1215 (D.C. Cir. 2014)). As noted above, and as agreed to by the parties, complete elimination of the hazard of workplace violence is not the appropriate metric under the circumstances; and the standard does not require complete elimination of the hazard. (Jt. Stip. No. 11). Respondent is only required to rid its workplace of preventable hazards. See Nat’l Realty & Constr. Co. v. OSHRC, 489 F.2d 1257, 1266–67 (D.C. Cir. 1973). Notwithstanding the unpredictable nature of psychiatric patients under duress, the Court finds Respondent’s method of abating the hazard of patient-on-staff violence is inadequate.

  Respondent contends Complainant failed to prove the inadequacy of its program in two principal ways: (1) the existence of injuries alone is not sufficient to establish Respondent’s program is inadequate, and (2) its WVPP contains the requisite building blocks OSHA recommends for developing an effective WVPP. With respect to (2), Respondent discusses the WVPP’s component parts, especially as it relates to its clinical management of patients, which it claims is “the best way to mitigate the hazard of Type 2 workplace violence.” Resp’t Br. at 50. The Court will address each position in turn.

  One of the primary problems with Respondent’s argument regarding the incidence of injuries is Respondent’s failure to consistently collect good data. As discussed at length above, the way Respondent collected staff injury data was haphazard and inconsistent. At any given point during the inspection period and afterward, it appears there were multiple potential sources of employee injury data: Employee Accident Reports, which were discontinued and inconsistently used; HPR (MIDAS) reports, which are patient-based forms and only incidentally track employee injuries insofar as that information is entered into the “Comments” section; the Sedgwick hotline, which, more often than not, requires the employee to make the report; and the OSHA 300 logs, which should serve as a snapshot of all injuries but failed to include all injury data. (Ex. C-4 at 7 to 9). The result is an incomplete picture of the number and severity of injuries that have occurred at UHS-Highlands. This jigsaw puzzle of data leads to a number of problems, including: injuries that were not recorded; injuries that were documented but missing information; and, perhaps most importantly, insufficient data to establish trends and, thus, measure effectiveness of a particular abatement/intervention. (See, e.g., Ex. C-4 at 6 to 7). As an example, CSHO Stark noted Respondent provided her with 4 EARs for 2018, only one of which corresponded with an injury listed on the OSHA 300 log for that year. (Ex. C-4 at 9). Given this inconsistency and considering CSHO Stark identified 31 EARs filed in 2017, she reasonably concluded the data for 2018 was, at best, incomplete. (Ex. C-4).

  Further, even if the Court accepts as accurate the data as Respondent collected and reported it (it does not) and as CSHO Stark summarized, there is no adequate foundation on which the Court could reasonably conclude Respondent’s focus on clinical management of the hazard has been effective in reducing patient-staff violence. What Respondent characterized as a “slight reduction” in workplace injuries is a change from 8 recorded injuries in the OSHA 300 log to 7.17 While a reduction in injuries is certainly a worthy goal and evidence of effectiveness, the number alone is not particularly convincing when the number of record injuries was reduced by one, but the number of workdays lost to such injuries multiplied by three over the same time period. (Ex. C-4). Additionally, though Complainant did not make comparisons to national averages, or its own internal benchmarks, the Court does not find any such comparison would have been useful given the unreliability of Respondent’s data and any conclusions derived from it. As noted by the D.C. Circuit, “‘[H]azardous conduct need not actually have occurred,’ or have occurred at any particular rate, for an employer to be liable.” BHC NW Psychiatric Hosp. LLC v. Sec’y of Labor, 951 F.3d 558, 565 (D.C. Cir. 2020) (quoting Nat’l Realty, 489 F.2d at 1267). More specifically, the court held the general duty clause is not concerned with how an employer’s accident rate compares internally or to the industry as a whole; rather, the key question, as always, is “whether a reasonably prudent employer familiar with the circumstances of the industry would have protected against the hazard in the manner specified by the Secretary’s citation.” Id. (citing SeaWorld, 743 F.3d at 1207). The existence of injuries was not the sole basis supporting the Complainant’s assertion the Respondent’s workplace prevention program was inadequate as is more fully explicated herein.

   Respondent repeatedly asserted clinical management of patients is the most effective method to prevent patient-on-staff violence in an inpatient psychiatric hospital. Included under the rather broad umbrella of Respondent’s methodology of patient management to curb aggression are: proper evaluation and diagnosis, clinical care, monitoring patients and managing the milieu, de-escalation, medication management, and treatment by physicians and other direct care staff. Resp’t. Br. at 51. Respondent also claims its WVPP exceeds the recommendations in the OSHA Guidelines for Preventing Workplace Violence for Healthcare and Social Service Workers. (Ex. C-65). The Court has addressed many of these issues in Section III.C.4, supra.

  Dr. Argumedo and Dr. Lipscomb, as well as Respondent’s direct care staff, agreed effective patient management is vital to addressing workplace violence; however, they also testified patient management is only one piece of a comprehensive WVPP and identified a number of areas where patient management alone was insufficient, or, at the very least improperly carried out. For example, as noted above, treatment plans were generated for patients as part of the mandatory restraint and seclusion paperwork; however, a review of those treatment plans showed they were mostly pro-forma, identical for nearly each patient for whom they were completed, and not reviewed by staff. (Tr. 419-420, 686; Ex. C-42). In some instances, multiple treatment plans were generated for the same patient with no change in the form or documentation within the same file. (Ex. C-42 at 17, 26). Respondent pointed to its use of individualized safety plans as a clinical-based method for controlling violent behavior; however, Dr. Argumedo only found one such plan out of 80 files she reviewed in this case. (Tr. 1489, 1600). She also noted the plans themselves were not timely implemented nor timely modified when it was apparent the plan did not work. (Tr. 1489, 1499-1502, 1600). In one instance, a behavior plan was not implemented until a patient had already committed 10 acts of aggression towards staff and, even after its implementation, the only actual changes were related to medication. (Tr. 1494-1502; Ex. C-49). Consistent with Complainant’s criticism of Respondent’s program being patient-focused but not designed to address care staff safety, the behavior plans directed how a patient should manage his/her own behavior, not how staff should work with the patient. (Tr. 1495).

  Dr. Argumedo also identified failures in the admissions process, including numerous patients whose history indicated a violence risk but whose file indicated no- or low-risk of violence, a lack of consistency in how forms were filled out and inadequacies in the transmission of information regarding an incoming patient to direct care staff. (Tr. 1469-1474; Exs. C-13, R-8). Dr. Argumedo was also critical of how the milieu was managed to the extent that acuity did not appear to be a consideration in admissions and identifying examples of highly acute patients being admitted to a unit already housing multiple patients with high acuity and low staffing. (Tr. 1449-1459).

  As argued by Complainant, neither OSHA nor this Court are positioned to second-guess how patients are managed for the purposes of providing therapeutic care. That said, while Respondent’s methodologies for patient management might be sufficient for the purposes of providing medical care to its patients, those methods alone are not adequate for addressing the discrete hazard of patient-on-staff violence. This is best illustrated by the fact that multiple patients were involved in repeated acts of aggression towards staff and yet their files do not indicate changes to their therapeutic care or management designed to prevent the repeated acts of violence, some of which were directed at the same staff member who was the subject of the initial act of aggression. (Tr. 438, 662-663, 1725-1745, 3302; Ex. C-42, C-46).

  In addition to clinical management, Respondent touts its training program, including its workplace violence power point presentation, which Respondent argues is fundamental to its clinical approach towards identifying signs of patient aggression which may lead to assaultive behavior and corresponding methods of early intervention and management to prevent such behavior. (Ex. C-12). Per the guidelines, Respondent also provides de-escalation and self-defense/restraint training (HWC) to its employees to deal with an aggressive patient.

  Respondent appears to suggest any problem with its training program was not its failure to comport with the OSHA Guidelines or that its methodologies were inadequate for the task, but only its employees’ failure to understand certain aspects of their training were related to the WVPP. Not only does this highlight a key problem with Respondent’s WVPP—namely, the patchwork nature of its presentation and implementation—but it does not begin to identify the inadequacies of its program.

  First, as explained by Dr. Lipscomb, Respondent’s training cannot be limited to De-escalation, HWC, and clinical management of its patients; instead, Respondent’s training program must be comprehensive. Not only should training include the WVPP itself, training which neither Ms. Orr nor Ms. Petschauer could confirm took place, but it should set a clear standard for how violence is defined, how it should be reported, when it should be reported, how staff can participate in the process to improve the WVPP consistently with milieu experience, and provide examples of patient-on-staff violence and how to properly respond. (Tr. 3695, Ex. C-62 at 51). According to Dr. Lipscomb, simply training staff on de-escalation and HWC is insufficient without generating an understanding of their context within the entire scheme of workplace violence prevention, and the Court agrees. (Ex. C-62 at 51). To suggest, as Respondent has, the failure to understand the context of workplace violence training inures to its employees is to fundamentally misunderstand an employer’s role to ensure its employees are properly trained.

  Second, although Respondent has a program to track training, it appears most of Respondent’s veteran staff was overdue for HWC and verbal de-escalation training by months if not years without evidence of adequate effort on Respondent’s part to bring them up to planned training level. This is particularly problematic because over 60% of staff injuries occurred during the process of implementing a hold. (Ex. C-4). Along the same lines, Respondent did not present any evidence showing the power point presentation on the WVPP was presented to its employees, and many employees who testified did not recall seeing the document at any point during their tenure at UHS-Highlands.

  The WVPP has a substantial problem at its foundation which impacts Respondent’s ability to identify, track, and respond to workplace violence: the lack of a clear definition of workplace violence. Respondent’s WVPP targets what is characterizes as “Type 2” violence, which is defined as “violence committed by patients on staff members”.18 (Ex. C-11 at 1). However, as noted by Dr. Lipscomb, without an adequate definition of workplace violence, Respondent’s ability to track, predict, and respond to violent incidents is limited. Interestingly, Respondent’s Power Point presentation, which was purportedly introduced during orientation—though employees dispute ever having seen it—includes a more comprehensive definition of workplace violence, which includes physical attacks “resulting in any degree of injury”, verbal or written threats “that express the intent to cause physical harm”, and stalking, harassment, or intimidation. Short of an attack resulting in injury that was self-reported by an employee or that warranted management review, none of these more thoroughly included acts of violence were tracked, nor does it appear actions were taken to prevent them from occurring or recurring.

  An overarching issue in this case is whether Respondent provided adequate staffing. Respondent contends Complainant failed to show “specific data” to establish understaffing was a systemic issue or that having more staff would have materially reduced the hazard of patient-on-staff violence. Respondent also argues Complainant failed to show its staffing levels were below what is provided at comparable facilities or what would have been adequate given the circumstances at UHS-Highlands. Although Respondent contends the testimony of its staff members, alone, is not sufficient to show it was understaffed or that more staff would have materially reduced the hazard, the Court disagrees. First, the complaints regarding insufficient staffing were not limited to a couple of dilettantes unhappy with the amount of work they had to perform as alleged by the Respondent. Staffing inadequacy was a constant refrain in years’ worth of surveys, shift handoff paperwork, resignation letters, and employee complaints to management. (Ex. C-22, C-21, C-22, C-33). Nearly every employee who testified before this Court stated they had worked on shifts, typically evening and overnights, where there were as few as one or two people on a unit. This resulted in Code Greens that were not responded to or that required multiple calls for help because too many staff were designated to perform one-to-ones or other tasks which prevented them from responding to codes; only one or two employees performing holds even though they were trained such holds required at least three to perform; and, as a more extreme example, situations such as those described by Ms. Graumann, who was required to work a shift alone on more than one occasion. (Tr. 600). Though these employees could not identify specific dates,19 it is clear to this Court these were not isolated occurrences but were merely part of the job as experienced by Respondent’s employees. Respondent attributes many of these hazardous situations described by its employees to uncontrollable call-outs, where employees canceled or failed to show up for scheduled shifts, but this merely elucidates the challenges encountered by the Respondent in the development of its WVPP. It does not illustrate how its WVPP is designed to address such hazards. Respondent knew it had a problem with call-outs and could have addressed that specific problem through other means. Additionally, from what the Court can infer, many of the call-outs were a result of the staffing problem in the first instance. Finally, the fact that a UHS-Delaware Divisional Nursing Director’s audit found even perfect staffing in accordance with the matrix would have made it difficult for Respondent’s employees to perform basic duties on a consistent basis, let alone “handle multiple admissions and/or manage an acute unit”, identifies the WVPP’s shortcomings both from a staffing perspective as well as a clinical one.  

  In summary, Respondent’s attempts to abate the hazard of patient-on-staff violence were inadequate for several reasons.20 Respondent’s primary response to the hazard of patient-on-staff violence is to focus on the patient, but this has resulted in a patchwork of policies, some of which are only tangentially related to the hazard. This is best illustrated by the multitude of forms Respondent introduced into evidence to demonstrate its commitment to reducing workplace violence which were either (1) the same for every patient, thereby undermining its claim such forms had value as individualized responses to particular patient behaviors, or (2) were left blank. As noted by Complainant, Respondent introduced several blank forms that purported to exemplify its methods to manage aggressive patient behavior, including Managing Aggressive Patient Protocol, Promoting Patient Safety Review, Expectations for Aggressive Behaviors, Picture of a Safe Milieu, Assaultive/Aggressive Individual Treatment Plan, Safety Plan, and the Situational Awareness Communication Tool. (Ex. R-20, R-23, R-22, R-24, R-25, R-29, R-36). Without any testimony as to how these forms were used in any specific case or related to any specific behaviors encountered in the workplace, or even evidence to establish consistent implementation, it is impossible for the Court to conclude they effectively served the purpose for which they were designed. Accordingly, on this, and the other grounds discussed above, the Court finds Respondent failed to properly address the hazard of patient-on-staff violence.

  As stated previously, Complainant must show: (1) the 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 Health Management, Inc., 2019 WL 1142920 at *12 (citing Beverly Enters., Inc., 19 BNA OSHC 1161, 1190 (No. 91-3344, 2000) (consolidated)). For the most part, Respondent does not seriously challenge whether the proposed abatements are capable of being put into effect; in fact, Respondent has implemented, or begun to implement, many of these proposals since the inspection. Instead, the remaining key issue is whether the proposed abatements would be materially effective at reducing, albeit not eliminating, the hazard.

  The standard for assessing whether a particular abatement proposal would be materially effective at reducing the hazard is not necessarily mathematical. In BHC Northwest Psychiatric Hospital, 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 NW Psych., 951 F.3d at 565. In that respect, the courts have found expert testimony, insofar as it is reliable, is sufficient to establish a particular abatement method would materially reduce the hazard. See Integra Health Mgmt., 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). Effectiveness of a particular abatement proposal can also be established through successful use of a similar approach elsewhere and compliance with industry standards. See Pepperidge Farm Inc., 17 BNA OSHC 1993 (No. 89-265, 1997).

  Complainant does not attempt to specify a fully evolved violence prevention plan for the Respondent to implement, but rather leaves discretion in the hands of the Respondent to adapt, try, and modify suggested abatement measures to address the needs and circumstances of its facility. It is acceptable for Complainant to propose a process-based approach to abatement. See Pepperidge Farm Inc., 17 BNA OSHC 1993 (“[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.”). Complainant asserts the package of abatement measures proposed, either as a whole or in part, will serve to materially reduce the hazard of workplace violence even if the Court determines any specific abatement proposals do not meet the two-prong test for feasibility. Accordingly, the Court does not find affirmation of the Citation will rise or fall based upon Complainant’s ability to prove each and every one of the proposed abatement measures will result in significant reduction of workplace violence standing alone, but rather considers whether each is a feasible element of a more effective WVPP under the general duty clause.

  As discussed at length above, CSHO Stark discovered and staff testified to multiple acts of aggression and injury that occurred behind the nurse’s station: patients have climbed/jumped over the counter to attack staff, patients have reached across the counter to stab or punch staff, and patients have grabbed the phone and used it as a weapon. Further, the door to the nurse’s station just swings open and does not lock, which allows patients to enter without having to jump or climb over the counter. As noted by the Divisional Nursing Director, this allows patients an opportunity to access items behind the station when staff is away from the desk performing other duties. Respondent contends, consistently with OSHA and Joint Commission Guidelines, it determined the nurse’s station configuration is appropriate based on a balance between the need to protect employees from violent acts and the need to maintain a therapeutic atmosphere for its patients. Further, Respondent contends there is no scientific study to indicate changing the nursing station’s configuration would materially reduce the incidence of violence or assault; rather, Respondent suggests such a barrier would only change the location of the assault.

  As regards feasibility, there is no evidence to show changing the configuration of the nurse’s station is not capable of being put into effect. The Court does not perceive, nor does Respondent argue, any barrier to modifications such as raising the counter height, installing a lock on the entrance, or other change to decrease the possibility of a patient accessing the station. In fact, according to Lori Ayala, at the time of the trial, the process was already under way to modify the nurse’s station to raise the counter height. (Tr. 3293).

  The Court also finds modifying the nurse’s station to prevent entry or reaching/jumping over the counter would materially reduce the hazard of patient-on-staff violence. This conclusion is based on the testimony of all the experts, as well as a dose of common sense. Both Dr. Argumedo and Dr. Lipscomb agreed a change to the nurse’s station to limit patient access would have prevented many of the violent interactions documented by CSHO Stark and testified to by Respondent’s staff. Even Dr. Cohen acknowledged an enclosed nurse’s station with locking doors would have prevented some of the assaults. (Tr. 2684-2692). While Respondent spent much of the trial attacking the concept of an enclosed station because it would have a negative impact on the therapeutic milieu, this characterization amounts to a straw man. Although enclosure of the station, including full barrier installation, was considered, it was not part of the recommended abatement measure; rather, the focus was on a targeted modification which prevented the types of violent acts presently occurring at the nurse’s station.

  Respondent also seems to suggest any such modification would not reduce violent assaults because a barrier will only slow or delay a patient who is determined to cause harm to staff. Resp’t Br. at 61. This is not a convincing argument to oppose any of the proposed abatements, because it seems to suggest there is no way to prevent violent acts in the first instance, which similarly calls into question the efficacy of Respondent’s methods to prevent violent acts, including clinical management. The parties have stipulated complete elimination of patient-to-staff violence is not feasible, but the Court does not accept this as a reason not to take reasonable steps to reduce the incidence or severity. The Court finds, at a very basic level, a modification to the nurse’s station would prevent at least some of the type of assaults identified in this case by limiting the opportunity or temptation for such assaults to occur and\or delaying patient activity long enough for staff to implement other preventative measures.

  Finally, as regards this and other proposed methods of abatement, the Court would like to discuss the Blando study referenced by Dr. Cohen, which suggests a disparity between staff members’ perception of safety based on abatement measures and the actual impact of those safety measures. (Ex. R-73). While the study found “elements of the workplace that might make nurses feel safe might not actually result in a lower risk of assault”, the study itself considered disparate perceptions of safety between two populations of nurses: those working in emergency departments and those in psychiatric departments. Ultimately, the key finding was that variances in perceptions of safety were based on personal experience, noting in particular the perceptions of emergency room nurses: “[Emergency Department] nurses and psychiatric nurses perceive their safety differently, which is likely the result of the particular characteristics of the environment where they work, their background and how each group defines ‘violence’.” (Ex. R-73 at 8). The Court places little weight on the conclusions derived from this study as it relates to any particular abatement proposal in that it does not meaningfully assess the basis for differences between perceptions of nurses and verifiable impact on safety in any meaningful way.

  As with the modification of nurse station configuration proposal, the Court finds this abatement not only capable of being put into effect, but also likely to materially reduce the incidence of injury from workplace violence. First, since the inspection, Respondent has purchased additional radios and made them more readily available to direct care staff, which belies any argument the proposed abatement is not capable of being put into effect. Second, the Court finds any argument this would not materially reduce the likelihood of injury from patient attacks is specious, especially under the facts of this case. For example, [redacted] testified she needed to send a patient to get help when she was in the process of performing a hold on a violent patient. (Tr. 1148). Likewise, Jennifer Ranus testified she and two other employees needed help controlling a patient in the exam room but could not get the attention of other staff on the unit even though they were screaming, resulting in one of the employees being injured. (Tr. 1737-1738). Further, multiple employees testified there were times they were left alone on a unit and would not be able to call out to other employees or use the phone at the nurse’s station if they were performing rounds or other duties. If each employee was equipped with a panic alarm or radio, the employees discussed in the foregoing scenarios could have more quickly summoned help, which, in at least one instance, could have prevented, or at least lessened, the possibility of injury.

  As with the previous abatement discussion regarding the nurse’s station, Respondent argues the ability to summon assistance quickly and effectively “would not necessarily prevent a patient from becoming violent.” Resp’t Br. at 64. For much the same reasons, the Court finds this argument unavailing. The abatement proposals under consideration are not solely about preventing an individual from being violent, but also about preventing injury and/or exposure to such hazard. If an employee is capable of summoning help quickly and effectively, this will reduce the likelihood of having to engage a violent patient alone, which the evidence shows impacts an employee’s ability to safely perform a hold. Given the staffing issues identified throughout this case, it is not unusual for one of Respondent’s employees to be alone, rendering apparent the need for each employee to have a communications device available to him/her at all times. Accordingly, the Court finds this method of abatement is feasible under the general duty clause.

  The OSHA Guidelines recommend securing furniture and other items that could be used as weapons. Also, like the discussion regarding the nurse’s station, the Guidelines note it is important to strike a balance between creating an appropriate atmosphere for the services provided and securing furniture to prevent acts of violence. With respect to this recommendation, the Court finds the abatement proposal is both capable of being put into effect and likely to reduce the probability and severity of injury from a patient using furniture as a weapon. Respondent does not necessarily disagree, because it already adds weight to pre-weighted chairs and replaces and/or updates its furnishings “as design and industry standards change.” Respondent’s primary issue with this abatement proposal is that certain pieces of furniture, such as cafeteria chairs, should not be bolted down due to sanitary reasons, such as the need for frequent floor cleaning. This is Respondent’s prerogative according to the Guidelines and something that might require situational planning room by room, but it in no way detracts from a finding of feasibility. The evidence shows patients have used furniture as projectiles, which could injure staff. Thus, the Court finds this abatement proposal would be feasible.

  The evidence as to the effectiveness of this abatement proposal was minimal. The Court does not see any particular barriers to implementing the proposal; however, the answer to that question is mostly related to staffing. As for effectiveness, the Court finds there is little evidence to show a dedicated individual or individuals watching security camera footage from 120 different cameras simultaneously for 24 hours every day would enable Respondent’s staff to respond to acts of violence more quickly than ensuring each team member has an adequate means to call for assistance. With the exception of the one instance where Mr. Harriot, who just happened to be looking at the monitors at the right time, was able to respond quickly to an assault prior to the Code Green being called, there was little to no evidence presented to show the proposed monitor would be able to (a) identify an imminent assault without the aid of sound, which the cameras do not have; (b) respond more quickly than employees on the unit from an as-yet-undefined location where they were watching what would no doubt be an unwieldy bank of monitors; or (c) call for a Code Green more quickly than the employee engaged with the patient, especially if, as discussed above, the direct care employees are consistently equipped with a communication device, such as a walkie-talkie or panic alarm.

  The metric for feasibility is whether a particular form of abatement will be “materially effective” at reducing the hazard. While there may be some minor, potential benefit, the evidence presented at trial failed to convince the Court this element of Complainant’s comprehensive approach would provide a material reduction of the hazard when compared to an increase in staff and on-hand communications devices.

  In addition to the shortcomings identified above in Sections II.C.4 and others, supra, the Court notes a few additional problems with Respondent’s WVPP to highlight the importance of a comprehensive program. Dr. Lipscomb took great pains to identify specific aspects of a WVPP which are essential to both preventing patient-on-staff violence and explain how those specific elements work together to form a comprehensive approach to patient-on-staff violence. Respondent contends its program in place at the time of the inspection was based on the OSHA Guidelines, was comprehensive, and, according to Dr. Cohen, was above the standard of care for acute psychiatric facilities. For the reasons identified by Dr. Lipscomb in her testimony and expert report, including the OSHA Guidelines, the Court finds Respondent’s WVPP fell short in numerous respects and that Complainant’s proposed abatement is feasible.

  Clearly, the proposed abatement measure is capable of being implemented, as Respondent technically has a WVPP, albeit one which is inadequately constructed and poorly implemented. Thus, as with most of the proposed abatement methods, the question is whether Complainant’s abatement proposal “would more effectively and consistently apply measures designed to reduce patient-on-staff violence” than Respondent’s WVPP. BHC NW Psych., 951 F.3d at 565. The Court finds the following factors illustrate that it would.

  According to Dr. Lipscomb, management commitment to the WVPP is a foundational element of the safety program because it supports other program elements, such as risk analysis, hazard control, training, and evaluation/recordkeeping. (Ex. C-62 at 12). Without management commitment, Dr. Lipscomb opined the WVPP is just a paper exercise. (Tr. 3633). The most notable deficiencies in Respondent’s current workplace violence prevention procedures relate to Respondent’s implementation of its own program elements. Respondent’s WVPP sets a goal to reduce restraints and seclusion below the UHS benchmark and to reduce employee injuries resulting from patient aggression by 20%. (Ex. C-11 at 1). The problem, as noted repeatedly above, is Respondent does not adequately track employee injuries, nor does it consistently define workplace violence such that there is a benchmark to identify what, exactly, constitutes an act of violence and how the incidence has been accurately recorded and measured. This failure prevents any meaningful collection of data to show a reduction in staff injuries resulting from patient aggression. Further, given the disparate sources of data, Respondent is unable to adequately comply with the second element of Management Commitment, which is to record, report, and monitor all acts of patient aggression.

  Dr. Lipscomb and the OSHA Guidelines identify Employee Participation as another fundamental element of a comprehensive WVPP. In particular, Dr. Lipscomb noted the risk of workplace violence is not monolithic and reduction of the risk requires multiple perspectives, including the viewpoints of the individuals who confront that risk on a daily basis. (Tr. 3643). Similarly, the Guidelines indicate employees can provide useful input through committees discussing reports of violent incidents, facility design, and policies and procedures which impact safety. (Ex. C-65 at 12). The “Employee Participation” section in Respondent’s WVPP includes none of these elements but is instead predominantly focused on staff obligations with respect to patients. Respondent’s witnesses testified about the number of committees and meetings it holds with respect to workplace violence, including the PIC and PSC committees, but direct care staff were not members of those committees. Instead, they were told they could come to those meetings, even though the staffing levels on the units likely prevented their ability to do so. Further, though Respondent distributed employee surveys, there is no indication management entertained, let alone responded to, its employees’ repeated entreaties regarding staff and communication inadequacies21

  In order to assess the effectiveness of the program as a whole, as well as any particular intervention within the program, Dr. Lipscomb highlighted the importance of methodical hazard analysis. As noted previously, this requires both a clear definition of workplace violence and the establishment of a clear and effective policy for reporting and documenting incidents of violence. (Tr. 3627-3630). The evidence discussed throughout this Decision illustrates Respondent’s WVPP lacked both or, at the very least, its protocol for reporting and documenting incidents of violence was poorly implemented. For example, employees testified about injuries and incidents that either went unreported or unrecorded. The platforms for reporting incidents of workplace violence were not designed for that purpose, i.e., HPRs (MIDAS) reports. See, e.g., BHC NW. Psych Hosp., 951 F.3d 558 (D.C. Cir. 2020). In that respect, not only is the underlying data questionable, but there is a total lack of tracking and trending of workplace violence. Contrary to Ms. Petschauer’s claims, the only documented reports of trends came from CSHO Stark’s analysis of OSHA 300 logs, EARs, and MIDAS. (Ex. C-4). Ultimately, Respondent’s WVPP uses the language of the OSHA Guidelines and Joint Commission’s Sentinel Alert, but its components bear little relationship to the actual purpose of a comprehensive WVPP.

  Dr. Lipscomb’s testimony was not limited to identifying the problems with Respondent’s WVPP, nor did she rely solely on the OSHA guidelines and industry standards to illustrate the necessary components of a comprehensive WVPP. Dr. Lipscomb also supported her conclusions by reference to a comprehensive study,22 which showed implementing the OSHA guidelines was effective in reducing the risk of patient-on-staff violence by 60% within 24 months of their implementation. (Tr. 3613; Ex. C-62 at 32, C-68). In addition to this study, Dr. Lipscomb identified other behavioral health facilities that experienced reductions in staff injuries due to patient assaults after fully implementing a comprehensive WVPP. (Tr. 3663-3664; Ex. C-62 at 32, C-70). Respondent did not introduce any persuasive countervailing evidence. Accordingly, the Court finds Complainant established the implementation of a comprehensive WVPP is feasible.

  To start, it is important to note what, exactly, Complainant is proposing. Respondent categorically interprets this abatement proposal to mandate the presence of a uniformed, armed security force, whose sole responsibility is to respond to acts of patient aggression, without giving due consideration to other options. While that is one method by which a security staff could be utilized, another method would be for Respondent to re-institute its QSR program in a sustained and carefully designed manner. Because the QSR program has been used in the past, it is capable of being implemented. The fact that Respondent ended the program does not speak to either its ability to be implemented or its effectiveness, because, as discussed above, it was never truly implemented as it was designed, and was only ended due to purported budgetary restraints.