Evaluating progress in reducing workplace violence - IOS Press

4 downloads 0 Views 484KB Size Report
workers' compensation claims rates,. 1997–2007. Michael Foley∗ and Edmund Rauser. Safety and Health Assessment and Research for Prevention Program, ...
67

Work 42 (2012) 67–81 DOI 10.3233/WOR-2012-1326 IOS Press

Evaluating progress in reducing workplace violence: Trends in Washington State workers’ compensation claims rates, 1997–2007 Michael Foley∗ and Edmund Rauser Safety and Health Assessment and Research for Prevention Program, Washington State Department of Labor and Industries, Olympia, WA, USA

Received 24 November 2009 Accepted 12 July 2010

Abstract. Objectives: This study reports trends in the pattern of injuries related to workplace violence over the period 1997–2007. It tracks occupations and industries at elevated risk of workplace violence with a special focus on the persistently high claims rates among healthcare and social assistance workers. Methods: Industry and occupational incidence rates were calculated using workers’ compensation and employment security data from Washington State. Results: Violence-related claims rates among certain Healthcare and Social Assistance industries remained particularly high. Incidents where workers were injured by clients or patients predominated. By contrast, claims rates in retail trade have fallen substantially. Conclusions: Progress to reduce violence has been made in most of the highest hazard industries within the Healthcare and Social Assistance sector with the notable exception of psychiatric hospitals and facilities caring for the developmentally disabled. State legislation requiring healthcare workplaces to address hazards for workplace violence has had mixed results. Insufficient staffing, inadequate violence prevention training and sporadic management attention are seen as the key barriers to violence prevention in healthcare/social assistance workplaces. Keywords: Assault injury, health care, occupational safety, surveillance

1. Introduction Workplace violence is unevenly distributed across industries and occupations. Some occupations and industries are at higher risk because they involve the exchange of money, face-to-face transactions with the ∗ Address for correspondence: Michael Foley, Safety and Health Assessment and Research for Prevention (SHARP) Program, Washington State Department of Labor and Industries, PO Box 44330, Olympia, WA 98504-4330, USA. Tel.: +1 360 902 5429; Fax: +1 360 902 5672; E-mail: [email protected].

public, working alone or working at night. Such industries as convenience stores, gas stations, hotels, and taxicab services are representative of such risks where the motivation of the assailant is frequently robbery [10]. Other industries are at higher risk for violence because they may combine the above risks with work involving a distressed or constrained population. Health care workers, especially those in psychiatric facilities, are at higher risk of physical assault. Social service workers are also at higher risk of assault because of their more frequent contact with a distressed or impaired population [20].

1051-9815/12/$27.50  2012 – IOS Press and the authors. All rights reserved

68

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

Workers’ compensation data provide one way to track trends over time in injuries from workplace violence (WPV); to identify industries and occupations with elevated risk of violence; to prioritize industries and individual workplaces for prevention and enforcement efforts; and to evaluate the impact of prevention strategies or new regulations designed to prevent workplace violence. The aim of the present study is to report on the industrial distribution of WPV claims over the period 1997–2007, to highlight the industries where progress on reducing claims has been made, and to focus on the persistently high claims rates in certain segments of the health care and social assistance sector. Finally, a discussion of the strengths and limitations of using workers’ compensation data for WPV surveillance and evaluation research is presented. 2. Method Data for workers’ compensation claims related to assaults and violence for the study period were obtained from the Washington State Department of Labor and Industries (L&I) industrial insurance database. 2.1. Workers’ compensation system in Washington State In Washington, employers are required to obtain workers’ compensation insurance through the Department’s Industrial Insurance system unless they are qualified to self-insure. Self-insurance is permitted if a firm is able to set aside sufficient reserves and meet certain guidelines; roughly 480 (chiefly large employers) currently do so. Approximately two-thirds of the workers in Washington State are covered through the State Fund. Excluded from the system are workers covered by another insurance system (e.g. federal employees), the self-employed and a few select groups (e.g. corporate officers, domestic employees) for whom coverage is optional. It should be noted that self-employed workers are not included in these data unless they obtain coverage for themselves through L&I, which is rarely done. But changes in employment status do occur for some occupations. Prior to the passage of Initiative 775 in 2001, most home health workers were considered contract workers and were not covered. Many such workers now have workers’ compensation coverage and therefore appear in much greater numbers in L&I data from 2003 onwards. The Department maintains files for both state fund and self-insured employers,although

the information collected from self-insured companies is more limited. L&I defines claims either as “non-compensable” (for which injured employees are reimbursed for medical treatment costs only) or “compensable” (for which both medical costs and wage-replacement benefits for lost work-days are paid). To qualify for definition as a “compensable” claim, the injury must have resulted in four or more lost work-days. 2.2. Workplace violence case ascertainment Prior to July 1, 2005, the Department of Labor and Industries coded all claims using U.S. Department of Labor Z16.2 codes, developed by the American National Standards Institute. Claims were coded by industry and occupation of the claimant; injury source, nature, body part, and type of event or exposure causing the injury. This coding system does not specifically designate injuries as violence-related. For the purposes of this analysis, violence-related claims were defined as those with type of event/exposure codes 023, 025, 026, 027 or 502 (kicked, struck, struck by another person in the act of a crime, or shot) and any source code except 0200, 0201, 0230, 0240, 0250 or 0270 (assaults by animals) or 5910 (injury caused by victim). In addition to these, claims which were coded as type 028 (stabbed) were defined as assaults if and only if their source was coded as 5900 (person other than injured). This was done in order not to erroneously include many selfinflicted, accidental stabbings. Finally, claims recorded as type 029 (struck by, not elsewhere classified) were also included only if their source was coded as 5900 and only following a review of the claim description text. The code-selection criteria used for claims filed prior to July 1, 2005 are summarized in Appendix A. Following this date, the Department shifted to using the Occupational Injury and Illness Classification System (OIICS), the same coding system used by the Bureau of Labor Statistics for their Survey of Occupational Injuries and Illnesses. Claims were classified as violencerelated if they carried an event code beginning with either 61 or 60 (assaults and violent acts by persons and unspecified) or an event code of 631 (non-venomous bites) together with a source code beginning with 57 (person other than injured worker). A further filtering of these claims is done by excluding those with either a source code beginning with 51 (animals or animal products) or nature code 1331 (heart attacks). The State Fund data base includes claims involving medical treatment and/or time lost from work; the Self-

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

Insurance data base only codes claims involving four or more days of time loss. From each claim record we extracted medical and wage replacement costs, number of lost workdays, age and gender of the claimant, and employer account number. From the employer record we extracted industry, classified by the North American Industry Classification System (NAICS) number. In Washington State the premium charged to employers and employees for workers’ compensation insurance is based partly upon the number of hours worked by covered employees. Therefore, all employers must report the total number of hours worked each quarter. Employer reported hours were extracted and summed by NAICS code at both the sector (2-digit) and industry group (4-digit) levels. This, combined with the number of violence claims submitted by industry, allowed for the calculation of violence claims incidence rates by industry, defined in this report as number of claims per 10,000 full-time equivalent (FTE) workers per year, based upon a working year of 2,000 hours. For selfinsured employers it is not possible to calculate rates for all assault-related claims since only cases involving wage replacement for lost work-time are fully coded in the Industrial Insurance database. So for self-insured companies we can report only rates for “compensable” time-loss claims. 2.3. Occupational incidence rates In order to examine the distribution of violence across occupations, we extracted occupational information (SOC) from individual claims and combined this with occupational employment estimates generated by the Washington State Employment Security Department. These are based on occupation- by-industry percentages from the Occupational Employment Statistics (OES) survey conducted in cooperation with the Bureau of Labor Statistics (BLS). These percentages are applied to the industry employment generated by the Current Employment Statistics (CES) survey, the Quarterly Census of Employment and Wages (QCEW) and self-employment data from the State Population Survey program administered by the state Office of Financial Management to generate the estimates of occupational employment. 2.4. Prevention index Finally, in many tables presented here we use a measure called a “prevention index” to rank industries for purposes of targeting prevention resources. It is ob-

69

tained by ranking each industry by their count of claims and then by their claims rate. The count rank and rate rank of each industry is then averaged to yield their prevention index. It is calculated in order to give equal weight to industries which employ large numbers of workers and industries where the hazard level is high. Industries with both large numbers of claims and high rates of claims will be near the top of the prevention index. Other industries may have low claims rates, but due to their large size they can still be sectors in which many workers are injured. For purposes of prevention, these industries should not be overlooked, as might happen if only a rate-based ranking were used to prioritize prevention efforts. 2.5. Violence typology The circumstances surrounding work-related violence have been categorized by researchers into four main violence types: Criminal Intent (Type I), where the perpetrator has no legitimate relationship to the business or its employee and is usually committing a crime in conjunction with the violence; Customer/Client or Patient (Type II), where the perpetrator has a legitimate relationship with the business and becomes violent while being provided services; Worker on Worker (Type III), where the perpetrator is an employee or past employee of the business; and Personal Relationship (Type IV), where the perpetrator usually does not have a relationship with the business but has a personal relationship with the intended victim [12,25]. 3. Findings 3.1. Claim counts, costs and incidence rates There was an annual average of at least 2,247 claims related to workplace violence (WPV) in Washington State over the period from 1997 to 2007 (Table 1). This total does not include medical-only cases occurring at self-insured employers. Of the total number of WPV claims, 760 were “compensable”, involving four or more lost work days. For the period 1997–2007, total WPV claim costs for State Fund employers averaged more than $17.5 million per year (in 2007 dollars). The median cost per claim was $438, with median compensable claims costs at $5,475 and median medical-only claims costing $304. The highly skewed distribution of claims costs is shown by the fact that the average cost per claim is $8,848 while the average cost per time-loss

70

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence Table 1 Workers’ compensation claims related to assaults and violence, 1997–2007 WPV claims State fund Total WPV claims, 1997–2007 Percent of all claims Total direct cost of WPV claims, 1997–2007

21,849 1.5% $193,322,838

Percent claimants female Average age

56.7% 36.8

Average annual WPV claims rate per 10,000 FTE Average direct cost per WPV claim Compensable Medical-only

13.5 $8,848 $32,963 $751

Median cost per WPV claim Compensable Medical-only

$438 $5,475 $304

Total compensable WPV claims, 1997–2007 Percent of accepted WPV claims

5,492 24.4%

Average yearly compensable WPV claims rate per 10,000 FTEs

3.4

Average time loss days per claim Median time loss days per claim

198.7 26.0

Self-insured Total compensable WPV claims, 1997–2007 Average yearly compensable WPV claims rate per 10,000 FTEs

2,876 4.3

For workplace violence claim definition see Appendix A and Methods. Cost and time-loss days data for self-insured claims is limited and do not permit comparison to State Fund claims. Claims costs are adjusted to 2007 dollar-equivalents using the Consumer Price Index-Urban Wage Earners Seattle-Tacoma-Bremerton series.

claim is $32,963. The average annual WPV claims rate in the State Fund was 13.5 per 10,000 FTEs. The State Fund time-loss WPV claims rate was 3.4 per 10,000 FTEs, while that for self-insured employers was 4.3 per 10,000 FTEs. 3.2. Demographics of workers injured Workers who file claims for workplace violence are on average older and more likely to be female, than is typical of claimants in general. Over the period of study the percentage of WPV claimants who are female averaged 57%. This is a substantially higher proportion than for all claims regardless of type, where the percentage female is about 32%. The average age of the WPV claimants was 37 for the State Fund, and 40 for the self-insured category. This is older than the median age of claimants for all claims combined of 32 years. The most frequent OIICS code type of injury was “struck or beaten by fellow worker, patient, etc.”, reflecting the large number of violence-related injuries to health and social service workers. 3.3. Distribution by assault type A sample of nonfatal violence-related claims for injury-year 2003 was drawn and the claim description

text was reviewed so as to classify the claim by violence “type”. Of the 404 claims analyzed, 19% could not be classified based upon the brief claim description text provided. Type I claims accounted for 20% of the total; Type II were 45% of the total; Type III were 17% and Type IV were less than 1% of the total. When we broke these figures out by gender, we found that workplace violence where the victim is male is much more likely involve a co-worker (Type III), whereas if the victim is female it is more likely to involve a customer, patient or client (Type II). 3.4. WPV incidence rate trends Across all industries in the State Fund there was a decrease over the period 1997–2007 from 14.8 claims per 10,000 FTEs to 10.4. The regression-estimated time trend for the State Fund WPV claims rate was −5.0% per year (95% CI: −3.3%, −6.6%). This was not significantly different from the time trend over this period for State Fund claims for all injury events combined, which was −4.9% per year (95% CI: −4.2%, −5.6%). However, the time trend for State Fund WPV time-loss claims was flatter, at −2.7% per year (95% CI: −4.5%, −0.9%), than that for all State Fund time-loss claims

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

71

Table 2 Industry sectors with the highest WPV claims rates, claim count and prevention index ranks, Washington State, 1997–2007 NAICS code 62 92 61 72 71 56 81 44 53 48 45 23

Industry sector description Health care and Social assistance Public administration Educational services Accommodation and Food services Arts, entertainment and recreation Waste management remediation services Other services (except public administration) Retail trade Real estate and rental and leasing Transportation and Warehousing Retail trade Construction All industries

Average rate 75.5

Average count 1,053.0

Rate trend 1997–2007 (% per Rate Count year) rank rank −7.2% 1 1

Prevention index 1

29.9 15.0 14.4

224.5 102.4 180.5

−3.2% −4.8% −8.4%

2 4 5

2 4 3

2 4 4

15.1

29.9

−4.5%

3

8

5.5

9.7

81.4

−3.3%

6

5

5.5

8.1

60.1

−7.0%

7

7

7

6.1 6.0

78.0 29.3

−7.2% −11.0%

8 9

6 9

7 9

4.9

16.8

−6.8%

10

12

11

4.5 2.2 13.4

23.6 28.3 1,986

−10.7% −9.2% −5.0%

11 13

11 10

11 11.5

“Rate trend” is the regression-estimated time trend of WPV claims per 10,000 FTEs for the study period.

Fig. 1. Workplace violence claims rates by major industry sectors, 1997–2007.

72

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence Table 3 Industry groups with the highest WPV claims rates, claim count and prevention index ranks, Washington State, 1997–2007 NAICS code 6222 6232 6231 6233 6239 6243 7213 9221 6241 9211 5616 6221 6216 6111 6244

Industry group description Psychiatric and substance Abuse hospitals Residential mental retardation, mental health and substance abuse facilities Nursing care facilities Community care facilities for the elderly Other residential care facilities Vocational rehabilitation services Rooming and boarding houses Justice, public order, and safety activities Individual and family services Executive, legislative, and other general government Investigation and security services General surgical and medical hospital Home Health Care Services Elementary and secondary schools Child day care services All industries

Rate trend 1997–2007 (% per Rate Count year) rank rank 1.8% 1 1

Average rate 875

Average count 248

Prevention index 1

749

201

−2.3%

2

2

2

127 92

143 117

−11.5% −9.7%

4 7

3 4

3.5 5.5

167

54

−3.3%

3

10

6.5

111

63

−14.7%

5

8

6.5

99

68

−5.6%

6

7

6.5

64

105

−1.7%

11

6

8.5

75

63

−1.9%

9

9

9

42

107

−5.1%

14

5

9.5

65

45

−4.3%

10

12

11

77

15

−6.1%

8

17

12.5

47 39

41 54

−17.0% −1.9%

12 15

14 11

13 13

37 13.4

42 1,986

−10.7% −5.0%

16

13

14.5

Industry groups with at least 10 claims related to workplace violence per year, 100 or more full-time employees per year, and at least twice the average overall rate of 13.4 WPV claims per 10,000 FTEs.

combined, where the time trend was −3.7% per year (95% CI: −4.1%, −3.4%). This difference, however, is not statistically significant. 3.5. WPV claims distribution by industry and occupation Major industry sectors in the State Fund with the highest rankings on the prevention index for assault claims are shown in Table 2. Trends over the period 1997–2007 are depicted in Fig. 1. Health Care and Social Assistance leads all sectors with a WPV claims rate of 75.5 claims per 10,000 full-time-equivalent workers, with Public Administration at a rate of 29.9, and Educational Services with a rate of 15.0 incidents per 10,000 FTEs. It is also true that WPV claims rates fell more quickly over the period 1997–2007 for Health Care and Social Assistance (−7.2% per year) ,for Accommodation and Food Services (−8.4% per year), and for Retail Trade (NAICS 44: −7.2%; NAICS 45:

−10.7%) than is true for all sectors combined, where the WPV claims rate fell by 5.0% per year. For the entire period, over half of all WPV claims filed occurred in Health Care and Social Assistance, and this sector had an average rate of WPV claims over five times greater than that for all industrial sectors combined. So as to focus more specifically on the industry groups with the greatest risk of workplace violencerelated injury, Table 3 shows the ranking by prevention index of the top 15 industry groups as defined by their four-digit NAICS code. Figure 2 depicts the trend over time in WPV claims rates for the top five industry groups. Ten industry groups out of the top fifteen by prevention index are in the Health Care and Social Assistance sector. Psychiatric and Substance Abuse Hospitals are at the top of the rankings in both count of WPV claims (248 claims per year) and claims rate (875 per 10,000 FTEs). This is more than 65 times the overall WPV claims rate of 13.4 for all industries combined. Furthermore, the rate of WPV claims at

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

73

Fig. 2. Workplace violence claims rates by industry groups, 1997–2007.

Psychiatric Hospitals and Residential Mental Retardation/Mental Health and Substance Abuse Facilities combined was more than twice as high as that of the next highest industry: Nursing Care Facilities, which had a rate of 127 WPV claims per 10,000 FTE. Of the top three industries, only Nursing Care Facilities saw any substantial drop in their claims rate, which fell by over 11% per year over this period. Residential Mental Retardation/Mental Health and Substance Abuse Facilities had been making sustained progress in reducing its rate of WPV claims up until the last two years of the study period, when rates rose substantially back to their 1997 level. This reflected an increase in claims, rather than a decrease in FTEs. Other Residential Care Facilities (WPV claims rate of 167 per 10,000 FTE), an industry comprising child group foster homes and group homes for the disabled, showed a decrease until 2003 followed by an increase through 2007. It should be noted that most of the decrease in claims rate for the top industries occurred before 2003, and rates have generally been flat from 2003 through 2007. Exceptions to this rule are Vocational Rehabilitation Services (Rate: 111, Trend: −14.7%) and Home Health Care Services (Rate: 47, Trend: −17.0%), which both saw substantial declines in rates over the entire period. The

only industry showing an increase in claims rate was Psychiatric and Substance Abuse Hospitals, rising by an average of 1.8% per year. Because the State of Washington does not receive sufficient data from self-insured employers to allow for the coding of medical-only claims by injury event, Table 4 shows only time loss claims rates and counts for self-insured employers. These employers typically are large corporations, non-profits or the larger units of local government such as the city of Seattle. Among selfinsured companies there was a decline in the overall compensable claims rate for WPV, which fell at an annual rate of −3.5% per year (95% CI: −9.1%, +2.1%). However, most of this decline came in the last three years of the period, and cannot be seen yet as representing a real trend, given the wide variation in annual rates. Overall rates for time-loss WPV were higher among self-insured employers (4.3 time-loss claims per 10,000 FTE) than for employers in the State Fund (3.4 time-loss claims per 10,000 FTE). As expected, the industries at the top of the list are largely healthcare and social assistance-related industries, with time-loss rates ranging from 10 (General, Medical and Surgical Hospitals) to 26 claims per 10,000 FTE (Community Care Facilities for the Elderly). The exception of note

74

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence Table 4 Self-insured industry groups with the highest time-loss claims rates, claim counts and prevention index ranks, Washington State 1997–2007 NAICS code 9211 6231 6233 6221 6219 6111 4851 8131 5616 4461

Industry group description Executive, legislative, and other general government Nursing care facilities Community care facilities for the elderly General, medical and surgical hospitals Other ambulatory health care services Elementary and secondary schools Urban transit systems Religious organizations Investigation and security services Health and personal care stores All industries

Rate trend 1997–2007 (% per Rate Count year) rank rank −3.7% 3 1

Average rate 19

Average count 97

Prevention index 2

24

7

−12.1%

2

4

3

26

2

−5.1%

1

7

4

10

58

−2.2%

6

2

4

13

2

−4.7%

4

6

5

7

54

−2.8%

9

3

6

10 7 9

1 3 1

7.0% −10.6% 6.9%

5 8 7

8 5 10

6.5 6.5 8.5

5

1

6.9%

10

9

9.5

4.3

262

−3.5%

Industry groups with at least 3 time-loss WPV claims per year, 100 or more full-time employees per year, and had an average overall rate above 4.3 time-loss claims per 10,000 FTEs. Only time-loss claims are included.

is Executive, Legislative and Other General Government, with a rate of 19 time-loss claims per 10,000 FTE. This industry comprises all of Washington State’s larger cities and counties and, as such, represents a highly diverse distribution of occupations, ranging from utility workers and firefighters to office personnel dealing directly with customers and clients. This ranked distribution of industries is similar to that for State Fund industry groups, with the exception of the prominence of General, Medical and Surgical Hospitals, which ranked twelfth among State Fund industries, but ranked fourth among the self-insured. This perhaps reflects the fact that the largest hospitals in the large cities and more urbanized counties are included in the self-insured category. Occupations with the highest numbers and rates of time-loss WPV claims are shown in Table 5. For both State Fund and self-insured employers, occupations in the health care, social assistance and security-related services ranked highest by prevention index. In order of their ranking on the prevention index, these include Police and Sheriff’s Patrol Officers (76 time-loss claims per 10,000 FTE; annual claim count: 59), Nursing Aides, Orderlies, and Attendants (claims rate: 43; claim count: 99), Counselors (claims rate: 964; claim count: 22); Psychiatric Technicians (claims rate: 264;

claim count: 24), Correctional Officers/Jailers (claims rate: 45; claim count: 29), and Licensed Practical Nurses (claims rate: 25; claim count: 26) and Registered Nurses (claims rate: 11; claim count: 54). It should be noted that the compensable claims rates for two health care and social assistance occupations – Psychiatric Technicians and Counselors – were several times higher than that for the next highest risk occupation: Police/Sheriff’s Patrol Officers. The rate for Nursing Aides and Orderlies was comparable to that for Correctional Officers and Jailers. Finally, it should be noted that exposure to the risk of workplace violence appears to be growing along with employment share in the health-care and social assistance sector. Employment in all State Fund industries combined grew by about 20% over the period 1997– 2007, whereas employment in the higher risk industry groups represented in Table 3 grew by nearly 35%.

4. Discussion The pattern of non-fatal injuries across industries and occupations that emerges from Washington State workers’ compensation data shows that Health Care and Social Assistance ranks as the highest risk major

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

75

Table 5 Occupations with the highest rate of time loss claims, claim counts, and prevention index, Washington State, 2000–2005 SOC code 333051 311012 211019 292053 333012 292061 339032 291111 211015 211014 533021 253099 259041 311011 319099

Occupation Police and Sheriff’s patrol officers Nursing aides, orderlies, and attendants Counselors, all other Psychiatric technicians Correctional officers and jailers Licensed practical or vocational nurses Security guards Registered nurses Rehabilitation counselors Mental health counselors Bus drivers, transit and intercity Teachers and instructors, all other Teacher assistants Home health aides Healthcare support workers

Average claim count 2000–2005 59

Rate per 10,000 workers* 76.4

Count rank 2

Rate rank 3

Prevention index 2.5

99

42.7

1

5

3

22 24 29

964.1 263.9 44.6

7 6 4

1 2 4

4 4 4

26

25.2

5

7

6

22 54 15 8 12

14.8 11.2 14.8 26.1 14.4

7 3 11 16 13

8 13 8 6 10

7.5 8 9.5 11 11.5

21

8.9

9

15

12

18 10 7

5.5 9.4 13.2

10 14 18

16 14 11

13 14 14.5

*Full-time equivalent employment is not reported by occupation in the Washington State industrial insurance database. Source: Washington Occupational Employment Projections, June 2007. Washington State Employment Security Department, Labor Market and Economic Analysis Branch. Link to occupational employment estimates: http://www.workforceexplorer.com/admin/uploadedPublications/1647 longoccupt.xls.

sector, followed by Public Administration and Educational Services. Despite the fact that Health Care and Social Assistance comprises less than 15% of the working population in Washington State, it accounts for over 50% of the violence-related claims filed. The riskiest occupations were related to psychiatric and long-term health-care, social-services, security services and retail [13,26]. Among staff at the large state psychiatric facilities, WPV claims rates averaged nearly 1 case for every 10 workers each year. Unlike work-related homicides, non-fatal incidents are much more likely to be Type II cases, involving clients or patients. A study of fatalities in Washington State classifying homicides by assault type found that in workplace homicides 59% of the incidents involved an assailant who was either a stranger, a co-worker or was a personal relation of the victim. This contrasts with non-fatal incidents, where such types accounted for only 38% of the total. [27]. 4.1. Risk factors for increased WPV in health care and social assistance Nonfatal assaults in Health Care and Social Assistance are primarily encounters between caregivers

and patients or between social service providers and clients [13,18,23]. Staffing levels in psychiatric hospitals, where full-time-equivalent staff numbers have been flat over the study period, are of particular importance. Two studies have shown that assault rates in psychiatric care settings have been higher in areas where staff-patient ratios and the costs expended per patient are lower [14,15]. Policies of de-institutionalization of patients, which place social service workers at greater risk, are also faulted for increasing workplace risks for health care workers by increasing the proportion of the institutionalized population with more serious diagnoses. The effect of staff shortages and the resulting strategies that facilities use to cover for temporary absences of staff may be contributing factors to higher WPV risks. A survey of staff in 142 Veterans Administration hospitals found over 13% had experienced at least one assault in the previous year. The results showed that the risk of assault increased when staff worked in geriatrics, mental health, or in rehabilitative care. It also found that working as floating staff or on shift/switch or mandatory overtime schedules increased the risk of

76

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

experiencing assault [11]. In another study a survey of general hospital staff in the UK tested the relationship between levels of anxiety, coping styles and burnout among patient-caregivers and the frequency with which they experienced aggression from patients in the prior year. Only burnout levels were found to be directly correlated with risk of assault [32]. The authors propose a cyclical model wherein violence increases the level of emotional exhaustion in staff which in turn leads to depersonalization as a coping mechanism. This manifests itself in negative behaviors toward patients, and finally greater patient aggression. They suggest among the organizational factors heightening the risk of burnout is inadequate staffing levels. Inadequate staffing levels may heighten the risk of assault indirectly by increasing the frustration level of patients, leading to physical confrontation. Staffing shortages have been recognized recently as playing a role in the persistently high levels of assault at the largest psychiatric facility operated by Washington State. An interview with the director of the Mental Health Division at the Department of Social and Health Services revealed that inadequate staffing levels at Western State Hospital has meant that this facility relies more on physical restraints and patient seclusion than is typical in such facilities [16]. A review of studies evaluating WPV prevention programs in healthcare settings found only mixed evidence of effectiveness [31]. Partly this was due to problems typical of cross-sectional study design such as uncontrolled confounders and reporting bias. Partly this was due to the variability of the outcomes tracked and the multiplicity of the interventions introduced. This study finds that although patient-management training would seem an appropriate component of prevention, more research is needed to identify practical methods which have the greatest impact on reducing violence. Prevention guidelines and materials have been developed by CDC-NIOSH and by OSHA for industries at risk for workplace violence, such as healthcare and retail trade and [20,21,29,30]. For the healthcare sector, these resources provide specific strategies incorporating environmental design, such as enclosed nurses’ stations, administrative controls, such as adequate staffing to minimize patient frustration, and behavior modification, such as training in non-violent communication and deescalation techniques. At the same time NIOSH has recognized the need for better evidence-based strategies and for research to establish how to implement best practices for a range of workplaces [22]. Recently it was suggested that as the general population in the US ages over the next few decades, a greater

number will reside in long-term care facilities. As this happens, greater pressure will be exerted on the staff of these facilities unless adequate levels of funding are committed to raising staff levels and training to cope with the influx. The same will likely occur in hospital emergency departments, which have become the entry point into the health care system for the increasing numbers of uninsured persons [8]. 4.2. Policy responses to WPV in health care in Washington State The legislature in Washington State has recognized the issue of workplace violence in the healthcare setting and enacted a requirement that such workplaces develop a violence prevention program. Legislation based upon NIOSH recommendations for reducing WPV in healthcare workplaces was first introduced in 1997 and the new regulations have been phasing in since 1999 in both general and psychiatric hospitals. The new rule required all healthcare employers to develop a hazard assessment; implement staff training on WPV risk factors, patient violence predictors, de-escalation techniques and post-incident procedures; and to develop and maintain a WPV incident tracking system. Evaluation of how this new set of rules is being implemented will be crucial since, after an initial drop, persistently high claims rates in the psychiatric facilities show that, as of 2007, there does not yet appear to have been a sustained reduction in WPV claims among state psychiatric facilities (Fig. 2). This is in contrast to the substantial progress which appears to have been made in other segments of the healthcare and social services sector such as in Nursing Care Facilities, Home Health Care Services and Vocational Rehabilitation Services. The substantial, but short-lived, drop in WPV claims rate in the psychiatric hospitals in Washington State in 2000 may be related to the initial commitment of funding and management attention to implementing the staff training that was required under the new WPV regulations. Unfortunately, funding cuts in 2001 eliminated support for this effort before it could be adequately evaluated. One industry which had appeared to be making substantial progress in bringing down WPV claims rates was Residential Mental Retardation facilities (NAICS 6232). As Fig. 2 shows, from 1999 through 2005, rates declined by about half. Then in 2006 the rate rose substantially back to its pre-1999 level. Discussions with management at one large facility in Washington State centered on two factors underlying this trend: a continuing increase in the acuity level of the resident

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

population, and a large decrease in the average level of experience among the direct care staff. In past decades, developmentally disabled children comprised a large share of the resident population in this industry. This population is now largely deinstitutionalized and cared for by families or in group homes. However, upon reaching their early twenties, many such persons are admitted to these facilities when their families can no longer care for them. This, in the opinion of managers at the facility, has led to a higher risk of assault for caregivers. In addition, a surge of retirements and new hires after 2003, with a concomitant drop in the average experience level of the direct care staff, was cited as an additional factor increasing the risk of assaults. The risk of work-related violence in healthcare extends beyond the walls of the traditional healthcare work setting. In 2006 the legislature enacted a rule to reduce the risks faced by mental health professionals who go to clients’ residences to conduct evaluations for involuntary treatment. The legislation, called the Marty Smith Act, was motivated in part by the murder of a mental health worker by a client during a solo home visit. The act, which went into effect in July 2007, will require that mental health professionals who visit clients’ homes must be accompanied by a second person if they request; not face retaliation for refusing to go alone; requires providers to have written training plans; back-up staffing, and communication plans for outreach staff. The act also requires staff to have a cell phone and access to information about the history of potentially dangerous clients. In addition, the Department of Social and Health Services is charged with monitoring compliance with the act as part of the provider licensing process. One question that will need to be addressed is whether the declining rates of WPV claims seen in all healthcare workplaces except psychiatric hospitals is due in some part to the existence of the new regulations. A recently published study comparing hospital employee assault rates in California, where a WPV regulation covering hospitals was enacted in 1995, to those in New Jersey found evidence of significant policy impact [4]. A similar comparison of Washington State with other states that do not have such rules would be useful. One challenge to evaluating the impact of the new WPV prevention rules in healthcare is that the industry focus of occupational safety enforcement in Washington State has not emphasized health care settings. While over 11% of total employment in Washington State is in the Healthcare and Social Assistance industry only about 1% of all DOSH inspections occurred

77

at workplaces in this sector. About 9% of the inspections L&I performed from 2002 to 2007 had a workplace violence special emphasis code or cited a workplace violence-related violation. This fraction was stable over this time period. This means that only the general deterrent effect of the new rules is likely to be observable, and even this is probably weakened by non-compliance with the new regulations. 4.3. Policy responses to WPV in retail trade in Washington State In retail trades, such as grocery stores, gasoline service stations and restaurants, increased risk has been associated with: working late at night; working alone and exchanging money with the public [6]. Wassell’s review of WPV intervention effectiveness research found that environmental design interventions, such as lighting entrances and exits, using cash-drop boxes and maintaining good external visibility were associated with a significant decrease in robberies and worker injuries [31]. In February of 1990 a regulation was instituted in Washington State which required employers operating late night retail establishments to offer crime prevention training to their employees, to provide adequate lighting levels, assure a clear view of the cash registers from the street and to limit access to safes and cash. Among other factors, it is possible that the drop in violence-related claims rate in the Grocery Stores industry from 21 per 10,000 FTEs in 1997 to 10 in 2007, and in the Gasoline Stations industry from 21 to 14 over this period, might be the result of increasing awareness of and compliance with this regulation. One continuing issue highlighted by the Bureau of Labor Statistics is the low prevalence of formal programs or policies to address violence in workplaces [28]. This appears even in spite of specific rules requiring such programs. In the case of the 1990 Washington State rule covering late night retail establishments, Nelson conducted a survey in 1995 that found that 35% of establishments covered by the Late Night Retail rule did not report having violence prevention training for their employees, and that only 6.5% of establishments covered by the rule were even aware of its existence. Awareness of the rule was, however, found to be associated with offering the required prevention training [19]. 4.4. Limitations In two respects this study does not fully capture the true scale of the problem of work-related assault and

78

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

violence. First, workers’ compensation data underestimate the number of injuries related to assaults and violence. The workers’ compensation data do not include incidents from among the self-insured employers that do not result in at least four lost work-days. Neither does the workers’ compensation data source include self-employed workers nor those covered under federal workers’ compensation programs. A workplace assault only results in a workers’ compensation claim if it results in an injury at least requiring medical treatment. In addition, workers’ compensation is subject to the many disincentives for employees and employers to report workplace injuries [2,7]. The scale of this underreporting can be appreciated by comparing the estimates of assault frequencies from different sources. For example, in a survey of workers at two Washington psychiatric facilities, researchers found that the ratio of assaults reported in the survey to that found in the “incident reports” filed by the hospitals was 5:1. They also found that the number of workers’ compensation claims filed by staff was less than 5% of all assaults reported in the survey [3]. A large population-based survey, the Minnesota Nurses Study, looked at rates of assault among nurses and found that only 15% of incidents of physical assault were ever reported. Another study in a long-term psychiatric care facility reported that only one assault in ten was recorded on the facility’s OSHA log [18]. Nonphysical incidents, such as threats, were even less likely to be reported in spite of their potential to eventuate in a physical assault or their impact on the nurses’ psychological well-being. Nurses listed among the reasons for non-reporting: lack of need for first-aid or lost work-time, fear of supervisor disapproval, that it was time-consuming, and that they didn’t expect anything to change as a result. Over 40% didn’t report because they believed the risk of physical assault was just “part of the job” [9]. In order for any prevention effort to be properly evaluated, the reporting and tracking of assault incidents must be improved. Accurate recordkeeping and an emphasis on reporting both physical and verbal assaults in employee training should be part of any WPV prevention program or new regulation. In the most recently published report on the National Crime Victimization Survey (NCVS) in 2005 it was reported that the highest rates of violent crime by occupation were experienced by police officers, correctional officers and taxi cab drivers [5]. These were followed by bartenders, mental health workers, special education teachers and gas station attendants. The highest-risk occupations are somewhat different from that reported

in the workers compensation data. This may be due to the fact that a large proportion of police and corrections officers work for self-insured city and county governments. In these cases we do not receive fully-coded medical claims, and thus miss a large fraction of assault cases for these occupations. This study also omits the work-related assaults suffered by self-employed workers. In the taxi industry it is a common labor practice for the taxicab owners to maintain an arms-length relationship with the drivers, who are considered independent contractors rather than employees. A final factor differentiating the NCVS, a household survey, from the claims database is indicated by the striking contrast in the gender ratios found in each study. In the NCVS the rate of male victimization was 56% higher than that for women. The Washington State claims data reports that women account for 55– 58% of violence-related claims. It may be the case that there is a gender- or even an occupation-based difference in injury-reporting behavioral norms. A prospective study in a New England long-term care facility reported that, controlling for job- and exposure-related predictors, although women were no more likely to experience an assault, they were more likely than men to report the incident [18]. Further research into this question would be very useful. In another respect, the burden of even the reported injuries is understated by the data presented in this report. The cost of just the State Fund portion of violencerelated claims in Washington State over the study period exceeded $193 million. However, this represents only the reimbursement of medical costs, wage replacement and, for some, disability pensions. It does not cover the material losses incurred by society in the form of lost production, for retraining the injured worker, for training replacement workers and for administrative costs. Estimates of the ratio of these indirect costs to the insurance costs reported in Table 1 are highly variable across industries and occupational classes, but are commonly suggested to range from 1.5:1 to upwards of 20:1 [1]. Beyond this are the costs borne by the worker, the worker’s household and by society. These include long-term productivity loss, increased stress and more frequent job change. Recent survey research has shown greatly increased levels of emotional and financial disruption in the lives of those who have suffered an assault while at work [17,24]. While violence-related injury in Washington workplaces seems to be declining it continues to be a significant contributor to workplace morbidity and mortality, particularly in healthcare workplaces. This analy-

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

sis suggests that investments in evidence-based interventions such as increased critical-needs staffing, deescalation training, risk communication and return-towork support teams can succeed in reducing this burden. Tracking workers’ compensation claims can be one part of that evidentiary base. A decade of workers’ compensation claims shows how stable is the pattern of industries where workers are most at risk. As the trend in late night retail businesses has shown, appropriately designed prevention steps may be effective in reducing violence. It is also clear, however, that simply having a WPV regulation is not enough: maintaining management commitment and investment in prevention is crucial and requires continuous oversight. Tracking the patterns of WPV claims in the workers’ compensation database can help to prioritize the state’s enforcement and consultation resources to where prevention efforts are lagging. References [1] [2] [3]

[4]

[5] [6] [7]

[8] [9]

[10] [11]

D. Andreoni, The cost of occupational accidents and diseases, Occupational Health and Safety Series, no. 54. Geneva, International Labor Office. 1986. L.S. Azaroff, C. Levenstein and D. Wegman, Occupational injury and illness surveillance: conceptual filters explain underreporting, Am J Public Health 92 (2002), 1421-1429. L. Bensley, N. Nelson, J. Kaufman, B. Silverstein, J. Kalat and J. Shields, Injuries due to assaults on psychiatric hospital employees in Washington State, Am J Ind Med 31 (1997), 92-99. C. Casteel, C. Peek-Asa, M. Nocera, J. Smith, J. Blando, S. Goldmacher, E. O’Hagan, D. Valiante and R. Harrison, Hospital employees assault rates before and after enactment of the California Hospital Safety and Security Act, Ann Epidemiol 19 (2009), 125-33. D. Duhart, Violence in the Workplace, 1993–1999. Special Report 190076, (2001) Washington, DC: U.S. Department of Justice. R. Erickson, Retail employees as a group at risk for violence, Occup Med 11 (1996), 269-276. Z. Fan, D. Bonauto, M. Foley and B. Silverstein, Underreporting of work-related injury or illness to workers’ compensation: individual and industry factors, J Occup Environ Med 48 (2006), 914-22. D. Gates, The epidemic of violence against healthcare workers, Occup Environ Med 61 (2004), 649-650. S. Gerberich, T. Church and P. McGovern, An epidemiological study of the magnitude and consequences of work-related violence: the Minnesota Nurses’ study, Occup Environ Med 61 (2004), 495-503. L. Hashemi and B. Webster, Non-fatal workplace violence workers’ compensation claims, 1993–1996, J Occup and Environ Med 40 (1998), 561-567. M. Hodgson, R. Reed, T. Craig, F. Murphy, L. Lehmann, L. Belton and N. Warren, Violence in healthcare facilities: Lessons from the Veterans Health Administration, J Occup Environ Med 46 (2004), 1158-1165.

[12] [13]

[14] [15] [16] [17] [18] [19]

[20]

[21]

[22]

[23] [24] [25] [26] [27] [28] [29]

[30]

79

J. Howard, State and local regulatory approaches to preventing workplace violence, Occup Med: State of the Art Reviews 11 (1996), 293-302. S. Islam, R. Surender, P. Mujuru, M. Priscah, E. Doyle and A. Ducatman, Risk factors for physical assault: state-managed workers’ compensation experience, Am J Prev Med 25 (2003), 31-37. S. Lee, S. Gerberich, L. Waller, A. Anderson and P. McGovern, Work-related assault injuries among nurses, Epidemiology 10 (1999), 685-691. L. Lehmann, R. McCormick and K. Kizer, A survey of assaultive behavior in Veterans Health Administration facilities, Psychiatr Serv 50 (1999), 384-389. J. Martin, “Attacks up sharply on workers at state psychiatric hospital”, Seattle Times. December 10, 2006. J. Miller-Burke, M. Attridge and P. Fass, Impact of traumatic events and organizational response, J Occup and Environ Med 41 (1999), 73-83. D. Myers, D. Kriebel, R. Karasek, L. Punnett and D. Wegman, Injuries and assaults in a long-term psychiatric care facility, AAOHN J 53 (2005), 489-498. N. Nelson, C. Mendoza, B. Silverstein and J. Kaufman, Washington State’s late night retail worker crime protection regulation. Relationships with employer practices, J Occup Environ Med 39 (1997), 1233-9. NIOSH, Health workers at risk for workplace violence, U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, Public Health Reports 111 (1996), 477-485. NIOSH, Violence: Occupational hazards in hospitals. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health. Publication No. 2002-101. NIOSH, Workplace violence prevention strategies and research needs. Report from the conference: Partnering in workplace violence prevention, Transferring research to practice. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health. Publication No. 2006-144. C. Peek-Asa and J. Howard, Workplace violence investigations by the California Division of Occupational Safety and Health, 1993–1996, J Occup Environ Med 41 (1999), 647-653. C. Peek-Asa, C. Runyan and C. Zwerling, The role of surveillance and evaluation research in the reduction of violence against workers, Am J Prev Med 20 (2001), 141-148. C. Peek-Asa and E. Jenkins, Workplace violence: How do we improve approaches to prevention? Clin Occup Environ Med 3 (2003), 659-672. S. Richardson and J. Windau, Fatal and nonfatal assaults in the workplace, 1996–2000, Clin Occup Environ Med 3 (2003), 673-689. P. Spielholz, Fatal Occupational Injuries in Washington State, 2000–2005. State of Washington Department of Labor and Industries. Olympia, WA, 2001–2006. USDOL/BLS, Survey of workplace violence prevention, 2005, United States Department of Labor, Bureau of Labor Statistics, Publication 06-1860, (2006). USDOL/OSHA, Guidelines for preventing WPV for health care and social services workers, United States Department of Labor, Occupational Safety and Health Administration, Publication 3148-01R. (2004). USDOL/OSHA, Recommendations for workplace violence prevention programs in late night retail establishments, United

80

[31]

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence States Department of Labor, Occupational Safety and Health Administration., Publication 3153. (2009). J. Wassell, Workplace violence intervention effectiveness: A systematic literature review, Safety Science 47 (2009), 10491055.

[32]

S. Winstanley and R. Whittington, Anxiety, burnout and coping styles in general hospital staff exposed to workplace aggression: a cyclical model of burnout and vulnerability to aggression, Work Stress 16 (2002), 302-315.

M. Foley and E. Rauser / Evaluating progress in reducing workplace violence

Appendix A: Case definition for workplace violence claims

81