Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line (2020)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

53 6.1 Introduction The case studies and analyses presented in this chapter introduce primary source employee demographic and wellness program participation data collected from five major metropolitan transit agencies: • The Indianapolis Public Transportation Corporation (IndyGo), in Indianapolis, Indiana; • The Regional Transit Service (RTS), in Rochester, New York; • The Transit Authority of River City (TARC), in Louisville, Kentucky; • The Des Moines Area Regional Transit Authority (DART), in Des Moines, Iowa; and • The Los Angeles Metropolitan Transit Agency (LA Metro), in Los Angeles, California. The analysis involved a review of descriptive literature publicly available from agencies or provided to the project team by the agencies, human resources records, insurance company records, and interviews with administrative, human resources, and health promotion program personnel. The director of human resources administration at IndyGo and the manager of wellness and benefits at RTS also participated in interviews and provided information for these case studies. In some cases, members of agency staff joined a conversation and/or provided data. Details on the project team’s selection method for the sites included in the study are available in Appendices A and B. In the case of IndyGo, RTS, TARC, and DART, the project team conducted an analysis based on individual-level data to determine if statistically measurable benefits were associated with program participation. LA Metro did not provide individual-level data, so regression modeling was not possible for this location. For the analyses, baseline data were collected from before the comprehensive health and health promotion programs began. Also collected before, during, and after the program were individual records of absenteeism (both sick and personal days taken) and workers’ compensation payments. Measures of participation were collected as well. Specifically, the project team examined the relationship between wellness/health promotion programs (screenings, 5K runs, diet) and improved health outcomes (less absenteeism, fewer sick days) in four sites using linear regression analysis. The results for three sites showed no statistically significant measurable benefit, a finding broadly consistent with past studies. In one loca- tion (Des Moines) the participation effect was statistically significant at the 95% level; it was estimated that participation in the program resulted in a 4-hour decrease in absentee hours. This result was reasonably larger, but based on a small sample so it is unclear if it could be repli- cated or should be used to generalize about effective wellness program interventions. The analyses for TARC (Louisville) and DART (Des Moines) were structured somewhat differently from those for IndyGo (Indianapolis) and RTS (Rochester). For IndyGo and RTS, C H A P T E R 6 Case Studies: Health Promotion Programs

54 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line the control/non-participatory group was based on a structural factor, such as whether an employee was insured or not (IndyGo) or worked at a remote location as opposed to working onsite (RTS). By contrast, the participants at TARC and DART volunteered to take part in well- ness activities and wellness screenings. In these two studies, the records on voluntary partici- pation were used to measure the correlations between outcomes and wellness programs. For IndyGo, RTS, TARC, and DART, the data were tracked so that pre- and post-analysis of effects of participation on absenteeism could be properly conducted. Because data were available on gender, race, and day of hire/termination, participation in these programs could be analyzed, as could rates of turnover and other research questions. The LA Metro case study provided information on the prevalence of health conditions from its insurance companies and detailed aggregate data on wellness program participation records. Individual-level data were not made available on absenteeism or workers’ compensation, however, so a multivariate statistical analysis was not possible. This chapter discusses process-based and data-driven benefits, though the two are not always mutually exclusive. Process-based benefits can include the diversity of the wellness committee, the array of programs, and the flexibility of the schedule. Data-driven benefits can include reduced absenteeism or workers’ compensation claims. This chapter begins with the description of the programs and more process-based benefits of the programs and follows with a discussion of estimated data-driven benefits. The case studies include scalable and sustainable strategies that have been implemented by the transit agencies. The programs have multiple features, including workshops on diet and exercise, biometric screenings, targeted education to avoid common injury types (e.g., musculoskeletal), financial planning, fitness challenges, and onsite gyms. 6.2 IndyGo This case study was developed through emails and discussions with the president of Amalgamated Transit Union (ATU) Local 1070, and the director of employee services for IndyGo. 6.2.1 Background IndyGo is a municipal corporation providing public transportation to the city of Indianapolis and surrounding Marion County, Indiana. The agency operates 31 bus routes throughout the county (IndyGo n.d.). As of 2018, it has approximately 680 employees, of whom more than 500 are members of ATU Local 1070 (Russell 2018). 6.2.2 Program Startup and Development The onsite clinic and wellness program were started on January 1, 2010, as part of a binding arbitration award between IndyGo and ATU Local 1070 in response to a pending premium increase of 46% from IndyGo’s health insurer. The steep increase was the provider’s response to the high cost of IndyGo’s medical claims. To control the increase, IndyGo management (together with the agency’s benefits consultant and with agreement from ATU Local 1070) proposed an onsite clinic and wellness program. Given the agreement to offer the onsite clinic and wellness program, the insurance provider dropped the premium increase from the pending 46% to approximately 20%. The overall savings captured by reducing the increases in insur- ance premiums benefited the program in two ways. As an incentive to participation, the agency used some of the savings to reduce the portion of the insurance premiums paid by participating employees, and additional savings helped fund the program itself.

Case Studies: Health Promotion Programs 55 6.2.3 Work Organization/Work Environment Like many other agencies, the majority of operators (approximately 55%) at IndyGo work split shifts. For many operators, this arrangement has a negative impact on their quality of life. Unless operators invest the time and expense to acquire, transport, and store their own food, having access to healthier food choices can be challenging. Onsite vending machines available in the break rooms were not stocked with healthy options. One of the top priorities of ATU Local 1070 has been to provide adequate restroom access for operators. This quality of life issue can have meaningful consequences, both short- and long-term. Before implementing the wellness program, management and union leaders worked together to address this issue. 6.2.4 Health, Wellness, and Safety Concerns From the perspective of IndyGo management, the main health concerns concerning workers’ compensation are musculoskeletal injuries; slips, trips, and falls; and vehicle accidents. According to the aggregate data from claims reports and onsite clinic data, the top health issues on the personal health side are obesity, hypertension, diabetes, prediabetes, and asthma. To address the work-related incidents and injuries, IndyGo has been incorporating ergonomics and prevention of injury into onboarding and in-service training. The union president expressed that diabetes, sleep apnea, and hypertension are the top health and wellness issues of the represented employees. Obesity is also on the rise among frontline employees, according to the local president. 6.2.5 Program Activities/Elements The IndyGo health and wellness program was made available to employees who have insur- ance through IndyGo. In 2016, approximately 88% of all IndyGo employees were covered under group health insurance. Participation was voluntary but incentivized: If employees participated in the program, they paid half of the premium (15% of the total insurance premium) compared with employees who did not participate (30% of the total insurance premium). Because of the incentive, IndyGo reported that 97% of the employees covered under the group health insurance plan elected to participate in the program (Russell 2018). To maintain their discounts on the health premiums, employees must complete the following annually: a physical, a health risk assessment, a biometric screening, a minimum of four coaching sessions, and a health activity. Some of the physical and educational activities include gardening, a Weight Watchers program, onsite exercise classes, walk–run groups, basketball tournaments, a 5K event for runners and walkers, and financial and nutrition classes. Union leadership stated that the approach has been effective because even though parti- cipants have to complete the requirements, the focus is on self-help and learning how to properly care for your health on your own. The union has been particularly pleased with the level of involvement of the onsite clinic provider because they understand the nature of the jobs performed and have developed relationships with the employees. Participants can get advice and care based specifically on the demands of their jobs. Participation in the program primarily occurs while employees are on the clock. According to the agency, getting employees to participate outside of their shifts is difficult. Efforts have been made to hold events and wellness opportunities in the community, but these activities were not well attended. “It’s a great program. I suffer from a lot of ailments and gain weight very easily. The doctor and nurses at the Activate clinic are very personal. They helped me so much and I have seen real progress. They understand how demanding the job is and our eating habits. They define different alternatives. We have good results; people are getting more conscious about fitness. That’s what you’ll hear from most members.” —ATU Local 1070 Financial Secretary

56 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line 6.2.6 Organization The human resources department oversees the onsite clinic and wellness program and all activities related to health program initiatives. IndyGo contracts with a third party that is staffed with two nurse practitioners, a part-time doctor, and medical assistants, and has a wellness committee composed of union and non-union employees that help design new activities and promote the program and initiatives. The program is funded through the IndyGo operating budget, which incorporates funding obtained through agreements with the union and the healthcare insurance provider. The 3-year contract with the current onsite clinic provider costs the transit system approximately $500,000 a year, including staff costs, clinic services (primary and urgent care), and expenses for some prescription drugs (as a one-time fill) (Russell 2018). 6.2.7 Qualitative Program Benefits Although the program did not have strong internal support from frontline workers initially, the president and financial secretary of ATU Local 1070 promoted the program and helped assure workers that information disclosed in the clinic would remain confidential (Russell 2018). Now, agency and union leaders report that there is total support for the program among the employees. Many employees have shared their positive experiences, including being screened for prediabetes or sleep apnea and having access to information about how to improve eating habits and lose weight. 6.2.8 Reported Metrics From 2010–2013, the average cost for health insurance per employee fell from $12,790 to $10,244. Between 2014 and 2017, the insurance provider changed and insurance costs fluctuated. In 2017 (under the new provider) the average insurance cost per employee was $13,004. As shown in Table 29, health claims rose from 2016 to 2017 (fourth column, percentage change) and appeared to be increasing at a similar rate in 2018 (sixth column, percentage change). Additional detailed information on medical claims (e.g., claims broken down by condition or claims dating back before 2015, before the wellness program began) was not received. 6.2.9 Method On June 15, 2018, after preliminary conversations, the project team provided IndyGo with a data use agreement stating that all data—including human resources, payroll, and program participation and other related data—would be used only for the research project, would be handled and protected according to the requirements of the Federal Information Security Management Act (FISMA), and would be destroyed at the end of the research period. Claim Type 2016 a 2017 a PercentageChange January 1– June 30, 2018 Projected Percentage Change b Medical-paid claims $4,257,969 $5,078,484 19.27% $2,538,382 20.0% Prescription-paid claims $1,091,018 $1,494,763 37.01% $791,502 27.1% a Table not adjusted for inflation. The Consumer Price Index (CPI) in 2017 was 1.6% per the U.S. Inflation Calculator; in 2018 it was 1.9%. b Numbers in this column are based on the assumption that the monthly rate in the second half of the year is the same as the monthly rate during the first 6 months of the year. Table 29. Claims and prescriptions reported for IndyGo, 2016–2017.

Case Studies: Health Promotion Programs 57 On June 21, 2018, after a follow-up call with personnel at IndyGo, the project team sent an email requesting the following data: • Excel files (or tab-delimited files) with downloads of the number of personal days and number of sick days with employee names, gender, date of birth, occupational code, and date of hire for 2009–2018 (or whichever historical years were available) for all employees; • Excel files (or tab-delimited files) with downloads of workers’ compensation payments for 2009–2018 with employee names for all employees; • Excel files (or tab-delimited files) with race and employee names for all employees; • Names of participants by year in the health insurance program; and • Names of participants by year in the health wellness program, among those eligible for the health insurance program. IndyGo provided payroll data with individual-level data from 2009 to 2018 on absences, including sick leave, personal leave, family medical leave, and leave without pay, as well as workers’ compensation data from 2012 to 2018. Because IndyGo introduced the health program in 2011, 2010 was established as the baseline year for the analysis, and all requests for data referenced 2010 as the first year. (Based on the initial interview, some early requests were made for 2009 data, but the agency later clarified that the program began in March 2011.) IndyGo further provided insurance information for employees from 2011 to 2018. Using the 97% participation rate in the program among those who carried insurance as a basis, the project team assumed that if employees carried insurance, they participated in the program. No other data were available on participation among those insured. The insurance information was merged with the absentee data based on the employee’s name and birthdate. The data provided 36 categories of job descriptions, with several categories referring to different types of operators (e.g., full-time, part-time), as well as jobs with maintenance, and administrative roles. Employees were categorized as operators, mechanics, and administrative staff based on their job descriptions in the leave data; for example, fixed-route–coach operator and flexible services coach operator were defined as operators. Administrative roles were removed from the analysis because the focus was on the outcomes for frontline employees, which consisted of operators and maintenance staff. For the models, the project team analyzed the full-program effects: comparison of absen- teeism and workers’ compensation measures for 2010 (the baseline year) with measures for 2017 (the last full year of program data) or with the last full year that the employee was at IndyGo before 2017, if the individual’s employment was terminated in 2017 or before. Regression models were run using ordinary least squares to detect any potential correlation between participation in the health program and lower absenteeism. The dependent variable in the models was an overall absentee variable capturing total days of leave, and the inde- pendent variables were participation/insurance (the key explanatory variables) and control variables, including age, race, tenure, gender, and occupation. The regression model was run using alter native dependent variables to measure the robustness of the model and results to different specifications. Two of those alternatives were workers’ compensation dollars and the difference of absenteeism and workers’ compensation before and after the introduction of the health program. 6.2.10 Workforce Characteristics To be included in the analysis, employees had to have been employed with IndyGo for at least 1 full calendar year in 2010 and for 1 full calendar year after the wellness program began in 2011. This qualification applied to 252 records. The workforce under observation was smaller than the total workforce due primarily to missing data and high turnover. In 2010,

58 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line IndyGo had 333 frontline employees. Twenty-one employees were excluded for missing data required in the regression analysis, and 60 employees were excluded because their employ- ment was terminated before the first complete year of the program. This left 252 records avail- able for the analysis. Table 30 presents race, age, and gender breakdowns for the two employee types (operator and maintenance, separate and combined) considered in the analysis. The White population was substantially older than the African-American population: The average age for the 53 White workers was 60.1 years, compared with an average age of 53.6 years for the 220 African-American workers. The men were slightly older than the women, averaging 56.5 years of age for men compared with 52.2 years for women. The maintenance workers were older than the operators, with an average age of 58.5 years compared to 54.7 years, and maintenance workers tended to be male at a higher rate (93.1%) than did operators (56.2%). The analysis examined if outcomes related to absenteeism were related to participation in the program. Thus, the analysis divided the population of frontline employees into two groups: “ever in program” and “never in program.” Table 31 displays the characteristics of these two groups. Based on employees having insurance through IndyGo, the average age of participants in the program was slightly younger (53.7 years) than the average age of non-participants (57.0 years). As Table 31 shows, participants in the program were overwhelmingly operators (only one maintenance worker had been in the program). 6.2.11 Absentee Hours After Program Initiation The data generated from the IndyGo health promotion program provided a wealth of new information in an area where data have been sorely lacking. Figure 5 presents the average annual absentee hours for frontline employees for the 8 years from 2010–2017. The figure illustrates the trend in absentee days, starting with the year before the program began (2010) and extending through the last full calendar year in which data were provided. The graph presents absentee hours over time for all employees (orange line), women (purple), and men Demographic Characteristic Operator Maintenance All Count Percent Age a Count Percent Age a Count Percent Age a African American 191 85.7% 53.9 7 24.1% 50.0 198 78.6% 53.6 White 31 13.9% 59.7 22 75.9% 61.2 53 21.0% 60.1 Other race 1 0.4% 60.0 0 0% N/A 1 0.3% 60.0 Female 87 39.0% 52.5 2 6.9% 48.9 89 35.3% 52.2 Male 136 60.9% 56.2 27 93.1% 59.2 163 64.7% 56.5 Total 223 100.0% 54.7 29 100.0% 58.5 252 100.0% 55.0 a All ages are averages. Table 30. Demographics of IndyGo frontline population, 2010. Demographic Operator Maintenance AllCount Percent Age * Count Percent Age * Count Percent Age * Ever in program 153 68.6% 53.9 1 3.4% 38.6 154 61.1% 53.7 Never in program 70 31.4% 56.4 28 96.6% 59.2 98 38.9% 57.0 Total 223 100.0% 54.7 29 100.0% 58.5 252 100.0% 55.0 * All ages are averages. Table 31. Program participation and age by job category, IndyGo.

Case Studies: Health Promotion Programs 59 (teal), and for African Americans (blue) and Whites (red). Absentee hours were defined as total hours of sick leave, personal leave, and sick unpaid leave. Although year-to-year fluctuations occurred for all six groups, the general trend does not demonstrate much variation. Beginning with an average of 70 hours in 2010, there was a slight increase over the 8-year period to approximately 100 hours at the end (2017), which might reflect an aging workforce. Women on average have slightly higher amount of sick leave than men, which was a trend evident among all the case study populations. Figure 6 shows the trends as plotted for the median annual absentee hours. Figure 7 presents the average annual absentee hours for frontline employees for the eight years from 2010–2017. The graph shows maintenance employees (blue), operator employees (red), and total frontline employees (orange). 6.2.12 Workers’ Compensation Table 32 shows the number of indemnity claims for the years the agency provided— specifically, annual data for frontline employees from 2013 through 2017. These claims could not be matched with individual employees (participants or non-participants). The table 0 20 40 60 80 100 120 140 160 2010 2011 2012 2013 2014 2015 2016 2017 Black or African American White Other Female Male Grand Total Figure 5. Average annual absentee hours, IndyGo frontline employees by race and sex, 2010–2017. 0 10 20 30 40 50 60 70 2010 2011 2012 2013 2014 2015 2016 2017 Black or African American White Female Male Grand Total Figure 6. Median annual absentee hours, IndyGo frontline employees by race and sex, 2010–2017.

60 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line illustrates the variation from year to year in both the number of claims and the average dollar- amount per claim. Figure 8 shows the percentage of (frontline) employees with indemnity claims from 2013 through 2018, including the total of employees with claims (orange), and the percentages for various demographic groups. As discussed above, data on claims before 2013 were not available. The percentage of claims increased over the observed period; thus, there was no evidence of a reduction in claims attributable to the program during this period. The program may have caused a reduction, but other (unobserved) factors would have had to offset that reduction, causing the overall rate to rise. Note that women filed the highest percentage of claims consis- tently throughout the period. 6.2.13 Results Using regression analysis, the project team investigated using several model specifications. The analyses varied the dependent variable, changed the mix of independent variables, and tested several interaction terms. The interaction terms tested how program participation varied by some of the demographic variables. In no case was the coefficient on the effect of program participation statistically significantly different from zero—that is, in no case did participa- tion have a statistically significant effect on health, measured as the change in number of days absent. Variables also were included for operators and maintenance, which would have shown if one occupational group was more likely to have reduced absenteeism days associated with the program than the other group. However, variables in those regressions did not have any statistically significant effects either. 0 20 40 60 80 100 120 2010 2011 2012 2013 2014 2015 2016 2017 Maintenance Operator Grand Total Figure 7. Average annual absentee hours, IndyGo frontline employees by job classification, 2010–2017. Year Sum of Claims Unique Claims Average per Claim 2013 $122,890 125 $983 2014 $228,239 336 $679 2015 $426,234 125 $3,409 2016 $956,551 336 $2,847 2017 $49,534 64 $774 Table 32. Workers’ compensation indemnity claims by year, frontline IndyGo employees, 2013–2017.

Case Studies: Health Promotion Programs 61 Appendix C outlines some of the potential reasons for the lack of significance for the participation variables. In particular, Tables C-1 and C-2 present two regressions that are representative of the variations that were tested, and the corresponding text includes a discus- sion of the analysis. 6.3 RTS This case study was developed with input from the director of well-being and inclusion and the director of people, performance, and development at RTS, and the president of ATU Local 282. 6.3.1 Background RTS is the public transportation agency that provides service to the counties of Monroe, Genesee, Livingston, Ontario, Orleans, Seneca, Wayne, and Wyoming in New York State. The agency serves more than 17 million customers annually and employs more than 900 individuals, of whom approximately 75% are operators and maintenance employees. As the largest subsidiary of the Rochester–Genesee Regional Transportation Authority, RTS has a fleet of 216 buses (of the authority’s total fleet of 404) and has built a reputation for on-time perfor- mance and innovative performance management (Rochester–Genesee Regional Transporta- tion Authority n.d.-c). Approximately two-thirds of RTS employees are based at the agency’s Monroe campus, which is the location of an onsite gym and the hub of the agency’s health promotion activities. The other employees are based at nine offsite locations remote from the Monroe campus and do not have immediate access to the gym. The employees based at the offsite locations have limited access to the agency’s health promotion activities. The data provided by RTS and the Rochester–Genesee Regional Transportation Authority used codes to represent employees at the main locations, including the nine offsite locations: Lift Line, BBS, STS, WATS, WYTS, OTS, CATS, GTC, and LTS. 6.3.2 Program Startup and Development RTS’s health and wellness program, dubbed Healthy U, started in 2011 as a modest and loosely defined program with a focus primarily on physical fitness. It became a more 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 2013 2013.5 2014 2014.5 2015 2015.5 2016 2016.5 2017 2017.5 2018 Black or African American White Female Male Grand Total Figure 8. Percentage of employees with workers’ compensation indemnity claims, by demographic group, IndyGo frontline employees, 2013–2017.

62 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line developed wellness program in 2013 and 2014, addressing a comprehensive set of goals that were defined in 2014. With healthcare costs skyrocketing and premiums rising both for the agency and the covered employees, RTS understood that it had an opportunity and an obli- gation to help employees. The agency hired a full-time health and wellness coordinator to oversee the newly expanded program. As the program was developed, medical claims data were examined to understand the most prevalent and costly medical conditions. Combining this knowledge with information about the demographics of the frontline employees, RTS staff members crafted the initial focus of the program. The agency conducted a survey in 2014 to understand the needs of transit employees, the types of programs they would be most likely to participate in and benefit from, and the most convenient times to hold events. The agency received 153 responses to the survey. In 2018, RTS conducted a similar survey to gather feedback on the wellness program. The latter survey asked respondents whether they had participated in wellness activities in the past and asked respondents to identify what would motivate them to participate in the future and whether there were any barriers that prevented their participation. This survey found that 52 respondents had participated in previous health promotion activities (RTS 2018). 6.3.3 Work Organization/Work Environment At RTS, almost all operators work split shifts (either two or three shifts). Generally, employees are on the clock between 9 hours and 12 hours, including breaks between runs. The maximum amount of time behind the wheel is 12 hours, however, and the maximum shift time is 15 hours. Bus maintenance requires coverage 24 hours a day, 7 days a week. Approximately 10% of bus technicians work split shifts—two sets of scheduled times within a 7-day span. A tech- nician may work, for example, from 11:00 a.m. to 7:00 p.m. for three days, then from 3:00 p.m. to 11:00 p.m. for the remaining days of the work week. Vacation time is based on accrued personal time, and vacation leave is approved and sched- uled for the entire upcoming year. Every employee also is allowed nine sick time and/or unapproved absences per year. If the number of unapproved absences exceeds nine, employees enter disciplinary action. Operators can apply for approved time off by putting in their name and the requested date(s). As long as the employee has sufficient accrued time to cover the requested leave, approved time off does not count as an unapproved absence (RTS 2018). Shift work and varied schedules have an impact on workers’ access to healthy food and sleep patterns. Employees working overnight shifts have access to vending machines onsite, but due to the hours, the availability of alternative healthy food options is limited in the community. Sleep schedules also can be impacted by working overnights. Many bus cleaners who work overnight shifts also work a second job during the daytime, which can result in added stress, limited access to healthy food and healthcare services or support, and irregular meal times. RTS’s health insurance provider issues annual data showing the prevalence of health condi- tions. The top three conditions for 2017–2018 were hypertension (affecting 25.1% of the insured population), cholesterol disorders (16.8%), and back and neck problems (10.2%). The union president considers sleep apnea, diabetes, hypertension, and muscular issues (primarily back and shoulder) as the primary reasons leading to potential medical disqualification among operators (Chapman, personal communication, 2018). The union also cited anxiety and stress and poor nutrition as the top health and wellness concerns.

Case Studies: Health Promotion Programs 63 6.3.4 Program Activities/Elements Healthy U has promoted healthier behavior and habits among RTS employees by providing a comprehensive set of new offerings and services and changing existing services to align with the goals of the program. Many of the adjustments have focused on food because this is an accessible way to build relationships with employees. The new programs and offerings were designed to be convenient and fun to encourage participation (e.g., short workshops in the break room, bowls of fresh fruit, team activities). The program also prompted changes to regular events and services that employees engage with (e.g., by providing healthier choices in vending machines and at employee events). RTS has made efforts to provide services that fit into the daily schedule of its employees. Agency employees have 24/7 access to a wellness center that includes a gym. Employees also can make individual appointments with a health and wellness coordinator. The health and wellness coordinator works full time, which provides some flexibility for operators and other employees with off-hour shifts. The program also offers vouchers that employees can use to obtain produce from a local farmer’s market at their convenience. Employees’ schedules, which are characterized by working shifts around the clock and on weekends, inhibit their participation in various parts of the wellness program. The wellness team and coordinator have tried to create programs that can be used at any time, with the hope of making it as easy as possible to engage all employees, regardless of what shifts or days they work. There is no feasible way to make the program accessible to everyone all the time, however. RTS promotes the Healthy U program through newsletters, posters and flyers, email blasts, paycheck attachments, and home mailings. Employees also can find information on the agency’s intranet (Rochester Business Journal 2016). One of the most effective ways of promotion is through the support and engagement of the agency’s Wellness Committee, whose members keep their coworkers and teams up to date on activities and events—and encourage their coworkers to participate. From each regional property, the RTS regional manager selects one employee (who may have a personal interest in wellness or be interested in a develop- ment opportunity) to participate as a wellness champion. Wellness champions participate in a monthly conference call to share ideas and collaborate on wellness-related topics, outreach, and events. Wellness champions do not receive extra compensation for their participation. The president of ATU Local 282 helps communicate information about the program to the union’s members. 6.3.5 Organization The People Department (Human Resources Department) manages RTS’s health and well- ness program, which employs the full-time wellness coordinator. The Wellness Committee is staffed by representatives from every division and meets once a month to oversee the program. This committee is made up of 16 employees, including one ATU member, and two representa- tives from the agency’s health insurance provider. Participation in the Wellness Committee is voluntary and members are not compensated extra. The President of ATU Local 282 also is personally involved in many health and wellness events organized by the agency. RTS has recently enacted a “Commitment to Diversity and Inclusion,” which the agency posits will impact the overall health and well-being of the organization and all employees by creating a more inclusive atmosphere that favors respect and relationships. A council of 16 employees, of whom 7 are frontline workers and ATU members, is responsible for carrying out the new effort, working in tandem with the wellness committee. Example of Sustainable, Successfully Implemented Strategy • Connecting around food: • Fresh fruit in breakrooms and common areas; • Snack of the month; • Short workshops on nutrition and cooking; • Healthier vending machine choices; • Catered employee events featuring “good-for-you” options; • Voucher program for local farmer’s market and other onsite experiences.

64 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line 6.3.6 Resources The Healthy U program relies on third-party providers and community partners for many of the services offered. RTS funds the program through its operating budget. In fiscal year 2017–2018, the program budget was $24,320, not including the salary for the wellness coor- dinator (Rochester–Genesee Regional Transportation Authority n.d.-a). The budget covers these key categories: blood pressure kiosks onsite; equipment and supplies for the wellness center (onsite fitness facility); health screenings; food for events; promotional items; well- ness initiatives linked to claims management; and other wellness initiatives and employee engagement. For fiscal year 2018–2019, the budget was increased by $17,000 (a substantial 70%) to $41,320. 6.3.7 Qualitative Program Benefits From the perspective of the agency, the wellness program has been a successful endeavor. It has brought the organization together and fostered greater employee engagement. Despite the lack of financial incentives for participation, program engagement and utilization have increased. According to the director of well-being and inclusion, one of the greatest difficulties regarding the participation of operators is scheduling. The majority of RTS bus operators work in shifts with prolonged breaks in between, but the breaks seldom align with planned wellness events. Some success has resulted from efforts to encourage supervisors to communicate with operators and promote the program by word of mouth. According to the union president, operators that have schedules consisting of three shifts participate in wellness events to a lesser extent because of the length of their workdays. Employees with irregular work schedules find it easier to participate in events that are sched- uled on weekends or programs that are available to employees at their discretion (including the produce voucher program). Increased physical activity due to the availability of the gym is the most apparent benefit of the program, though only employees that work at the Rochester campus use it regularly due to the proximity. 6.3.8 Reported Metrics The project team examined statistics from RTS’s health insurance provider (Table 33). As seen in the table, the prevalence rates of most of the major disorders that occur in the transit worker population showed slight increases among RTS’s insured population. Because the aggregate figures provided included administrative workers and covered dependents as well as frontline workers, it was not possible to isolate the effects of participation in the wellness program. Participation in the wellness program may have mitigated increases in prevalence “RTS wants our employees to thrive and live the healthiest lives they can. The RTS Healthy U wellness program fosters a culture of health and well-being within our organization and community by empowering our employees to make healthy lifestyle choices. The strategic initiatives we are implementing for the wellness program will support employees by providing education, resources, support, and access to programs and services that are fun, engaging, and sustainable. Healthy U brings employees together on their wellness journey and celebrates their successes.” —Renee Ellwood, Director of Well-Being and Inclusion Disorder 4 Years Prior Current Change General Population (Excellus) Cholesterol disorder 30.2% 29.0% –1.2% 18.9% Hypertension 38.8% 41.9% 3.1% 23.2% Diabetes 15.4% 16.7% 1.3% 8.1% Back and neck problems 8.5% 12.9% 4.4% 14.9% Depression and anxiety 5.3% 5.9% 0.6% 9.6% Source: Table as provided by RTS via personal communication (Excellus 2018). Table 33. Comparison of rates of major health disorders, RTS insured population to general population, 2012–2017.

Case Studies: Health Promotion Programs 65 among the frontline workers that were part of a more general health trend; however, lacking the necessary granularity in the data, that hypothesis could not be assessed. RTS continues to conduct ongoing review and analysis of the health claims data and monitor wellness initia- tives against claims data (Rochester–Genesee Regional Transportation Authority n.d.-b). In addition to health claims data, new conditions are identified through free, onsite health screenings. RTS seeks to educate and bring awareness to employees about potential health risks and to prevent or manage them. The focus on prevention has resulted in the identifica- tion of more employees with health risks, but this identification has also made it possible for employees to help manage those risks, using the Healthy U wellness program to make healthy lifestyle choices. The program also has focused on the importance of managing health condi- tions and prescriptions, as well as actively using the comprehensive health and wellness benefits provided to employees (e.g., insurance coverage for medical, dental, and vision services, and other employee benefits related to financial wellness and retirement planning). 6.3.9 Method The project team provided RTS with a data use agreement, and data received from RTS associated individuals with their employee ID numbers, thereby protecting their identities. Following conversations with relevant staff members, the project team emailed a list of the absenteeism, workers’ compensation, and demographic data requested. In August 2018, RTS began providing the project team with individual-level payroll data on absences and workers’ compensation. RTS provided absenteeism and workers’ compensation data from 2011 to 2018 for both onsite and offsite employees. Files of employees’ demographic information were provided, as well as hire and termination dates. This information was merged with the absentee data based on the employee ID. Because RTS had indicated that it introduced the health program as a comprehensive program in 2014, 2013 was used as the baseline year for the analysis. Payroll information was made available for more than 1,000 employees who had worked for RTS over the 2010–2017 period. Approximately 650 employees were onsite and had the easiest access to the health program. The entity code “RTS” was used to identify employees who were onsite and had access to the health program, whereas the rest of the employees were combined into a control group of “offsite” employees who were assumed to have limited-to- no-participation in the program. The data provided 282 categories of job descriptions, with several categories referring to different types of operators, maintenance, and administrative roles. Employees were catego- rized using the “Assignment Title” provided with their demographic information. For example, employees with the title bus operator were defined as operators, whereas an employee with the title workforce development manager was defined as administrative. Employees often had multiple assignment titles without a date-of-job change. To determine the job description, the project team selected the last available job title that was not retiree. Trainee was selected as the job title only if it was the only title available. Administrative roles were removed from the analysis, which focused on the outcomes for frontline employees (consisting of operators and maintenance staff). Multiple variables of interest were compared, including use of sick days, unpaid leave, and personal days. At RTS, employees acquire sick leave and personal leave at varying rates based on seniority; up to 120 days of sick leave can be accumulated (Hall, personal communication, 2018). For each variable, the difference in use before and after the introduction of the health pro- gram was examined. Multiple regression and other statistical analyses were run to find a relation- ship between participation in the health program and lower absenteeism.

66 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line In the model, the dependent variable was a total of sick days, unpaid leave, and personal days. The key independent variable was participation in the health program. The other inde- pendent variables controlled for age, race, tenure, gender, and occupation. The regression model also was run using the difference of the dependent variable before and after the intro- duction of the health program as the dependent variable. No statistically significant results were found. Details of the analysis are provided in Appendix A. 6.3.10 Workforce Characteristics In 2011, a limited version of the program was introduced, but the comprehensive version of the program was not rolled out until 2014, so the project team chose 2013 to be the base- line year for this analysis. Of the 574 frontline workers (operators and maintenance) who were employed in 2013, 389 workers (approximately 68%) were based at the main location where the gym and wellness programs were held, whereas 185 workers (32%) were based at remote locations. The project team designated the 389 workers at the main location as the participants because they had greater exposure to the program’s core elements. The 185 offsite employees were considered the control group of non-participants. Table 34 presents the averages for the total population of frontline employees, broken down separately for operators and maintenance employees. The first demographic detail that stands out is the same as for IndyGo: The White popu- lation is substantially older than the African-American population. The average age for the 300 White workers that were operators or in maintenance was 60.2 years, compared with the average age of 53.0 years for African-American workers. Men were slightly older than women, with an average age of 57.1 years compared with 54.3 years. As in Indianapolis, the population of maintenance workers was almost all male (106 out of 108 workers). The analysis method was to examine how outcomes related to absenteeism were related to program participation. Thus, the analysis divided the population of frontline employees into two groups: onsite and offsite. These groups represented the workers who participated in the program and those who did not. The characteristics of the two groups are displayed in Table 35. The average age of offsite operators (61.0 years) exceeded that of onsite operators (54.3 years). The average age of offsite maintenance workers (55.0 years) was only slightly lower than that of onsite maintenance workers (55.8 years); however, the vast majority of maintenance workers were onsite, with 101 of the 108 workers on location at the main campus. Calculating the total Demographic Characteristic Operator All Count Percent Age * Count Percent Age * Count Percent Age * White 237 50.9% 61.0 63 58.3% 57.3 300 52.3% 60.2 African American 183 39.3% 53.1 36 33.3% 52.4 219 38.2% 53.0 Hispanic and Latino 2 0.4% 60.5 0 0.0% 0.0 2 0.3% 60.5 Two or more races 2 0.4% 66.0 0 0.0% 0.0 2 0.3% 66.0 Native American 1 0.2% 41.0 2 1.9% 52.0 3 0.5% 48.3 Asian 41 8.8% 48.6 7 6.5% 58.3 48 8.4% 50.0 Female 117 25.1% 54.3 2 1.9% 53.0 119 20.7% 54.3 Male 349 74.9% 57.6 106 98.1% 55.7 455 79.3% 57.1 All 466 100.0% 56.8 108 100.0% 55.6 574 100.0% 56.5 * All ages are averages. Maintenance Table 34. Demographics of RTS frontline population, 2013.

Case Studies: Health Promotion Programs 67 populations of offsite workers (non-participants) and onsite workers (participants), the offsite workers were older (60.8 years) than the onsite workers (54.7 years). (The calculated numbers do not appear in the table.) 6.3.11 The Program Over Time As with IndyGo, data generated from RTS’s health promotion program has provided new information to assess the patterns of absenteeism of transit workers. Absenteeism days are defined as total hours of sick leave, unpaid sick leave, and paid and unpaid personal leave. Between 2011 and 2017, the total hours taken increased from approximately 40 to 60 hours per year. Figure 9 presents the average annual absentee hours for frontline employees for a 7-year period (2011–2017). The baseline in this analysis is 2013 and the comprehensive program began in 2014, but this case study includes some available data from 2011 when the health promotion program was introduced in a limited form. Figure 9 includes absen- teeism data from the early years of the program, before it was fully established (2011–2013), and from the subsequent years (2014–2017) that reflect absenteeism after the program was fully developed. The data in Figure 9 show that Whites had higher rates of absenteeism than did African Americans and that the rate of absenteeism among women was similar to that of men (not greater, as was the case at IndyGo). The data in Figure 9 have not been controlled for age. Figure 10 shows trends related to absenteeism by job category for operations, maintenance and all workers. As seen in the figure, during the period examined (2011–2017), maintenance workers had a higher average number of hours absent than did operators. Factor Operator Maintenance Offsite Onsite Offsite Onsite Number 178 288 7 101 Percentage 38.2% 61.8% 6.5% 93.5% Average age 61.0 years 54.3 years 55.0 years 55.8 years Table 35. Program status and age of RTS frontline population, 2013. 0 20 40 60 80 100 120 2011 2012 2013 2014 2015 2016 2017 Black or African American White Hispanic or Latino Female Male Grand Total Figure 9. Average annual total absentee hours, RTS frontline employees by demographic characteristics, 2011–2017.

68 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line Figure 11 presents the average annual total of absentee hours for RTS frontline employees for 2011–2017. On average during this 7-year period, onsite employees used fewer sick days then did offsite employees. 6.3.12 Workers’ Compensation Whereas Figures 9, 10, and 11 show trends related to absenteeism, Table 36 uses data provided by RTS to illustrate trends related to workers’ compensation over the same period (2011–2017), although data for 2013 and 2014 were not available. Table 36 presents workers’ compensation indemnity claims for the period and the average cost per unique claim. Table 37 includes the estimated number of days of workers’ compensation paid for all claims and the average number of days per claim. The claim percentage rate in 2011 (before program implementation) was 8%, and the percentage rate also was 8% in 2016–2017, well into the program. The percentage rose to 11% in 2012 but was reduced to 5% in 2015 (the next available year). The linear downward trend may indicate some effect due to the wellness programs, particularly if other (undocumented) factors were working at the same time to increase the percentage. It was not possible to statistically test these possibilities. 0 20 40 60 80 100 120 2011 2012 2013 2014 2015 2016 2017 Maintenance Operator Grand Total Figure 10. Average annual total absentee hours, RTS frontline employees by job classification, 2011–2017. 0 20 40 60 80 100 120 2011 2012 2013 2014 2015 2016 2017 Offsite Onsite Grand Total Figure 11. Average annual total absentee hours, RTS frontline employees by work location (onsite/offsite), 2011–2017.

Case Studies: Health Promotion Programs 69 Total workers’ compensation days were calculated based on the 2017 average wage for 731 RTS operators and maintenance workers of $24.32 and the workers’ compensation payment of 66.67% of that wage to fully disabled workers in the state of New York. 6.3.13 Results Several variations of the linear regression were performed, the results of which are presented in Appendix A. The project team varied the dependent variable (e.g., sick day, total leave days), changed the mix of the independent variables, and tried several interaction terms of program participation (e.g., onsite as interacted with various demographic variables). In no case was the coefficient on the effect of program participation statistically significantly different from zero. Similarly, in no case did participation have a statistically significant effect on health, measured as the change in days absent. Appendix C discusses potential reasons for the lack of significance for the participation variables and Table C-3 presents representative regression results of the variations run, accompanied by analysis. 6.4 TARC This case study was developed with the input of the president of ATU Local 1447 and the benefits manager at TARC. 6.4.1 Background TARC provides public transportation to greater Louisville, Kentucky, and the surrounding counties of Clark and Floyd in Indiana. The agency was founded in 1971 after legislation allowed the use of local funding from city and county governments to operate mass-transit systems (TARC n.d.). Table 36. Workers’ compensation indemnity claims at RTS, 2011–2017. Year Sum of Claims Unique Claims Average Cost per Claim 2011 $77,532 36 $2,154 2012 $225,487 47 $4,798 2013 Unavailable Unavailable Unavailable 2014 Unavailable Unavailable Unavailable 2015 $116,875 29 $4,030 2016 $512,173 46 $11,134 2017 $638,591 50 $12,772 Table 37. Workers’ compensation indemnity claims and calculated absentee days at RTS, 2011–2017. Year Unique Claims Frontline Workers Percentage With Claims Total Indemnity Claims Total Workers’ Compensation Days Average Days per Claim 2011 36 447 8% $77,532 604 16.8 2012 47 447 11% $225,487 1756 37.4 2013 N/A 468 N/A N/A N/A N/A 2014 N/A 531 N/A N/A N/A N/A 2015 29 549 5% $116,875 910 31.4 2016 46 582 8% $512,173 3,988 86.7 2017 50 592 8% $638,591 4,972 99.5

70 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line 6.4.2 Program Startup and Development The employee wellness program began in 2015 with the goal of changing workplace culture as prompted by an observed need for smoking cessation programs. The agency was further motivated to start a program that would reduce health insurance claims, which were relatively high. In 2015, TARC began offering smoking cessation classes, bringing together a motivated group of individuals that evolved into a more organized health and wellness committee. 6.4.3 Work Organization/Work Environment Many operators at the agency work split shifts, which can take up a majority of the employee’s time, although breaks can be scheduled that allow for meals or time at the gym. Operators feel that they have time for little else besides resting for the next day. Operators can request specific shifts at three points during the year, when shifts are scheduled. Eligible operators also have the option of working four 10-hour runs and taking weekends off plus one additional day off during the week. Parameters for split runs are governed by the bargaining agreement with the union, and during the period examined by the project team TARC was well under the threshold designated for split runs. From the perspective of the union president, the agency has prioritized restroom access for operators; this issue has improved over time. According to management, the agency has worked to establish ample restroom stops on every route. 6.4.4 Health, Wellness, and Safety Concerns According to the prevalence rates reported by TARC’s health insurance company, the five most prevalent health concerns by number of members (employees and family members) for 2015–2018 were hypertension, hyperlipidemia, back pain, osteoarthritis, and diabetes. In interviews with TARC management, obesity-related diseases were a common concern. TARC reported approximately 15–20 short-term medical disqualifications per year. The disqualifi- cations increased over the period examined, mostly due to non-compliance with sleep apnea requirements. According to the union president, the top health and safety concerns are passenger assaults on operators, operator injury resulting from equipment in the bus or accidents involving the bus, and breathing in harmful fumes. According to the union president, these health concerns are not addressed in the wellness program because they are categorized primarily as “safety” concerns and are dealt with separately under the joint safety committee. (Hamilton, personal communication, 2019). 6.4.5 Program Activities/Elements TARC’s wellness program activities have been based on survey responses from employees indicating the activities they would be interested in. Though the initial program was developed around smoking cessation, this is no longer a primary focus of the program, and was not an item that received interest in the most recent employee survey. Currently, the program consists of events and programs organized around a theme of interest, an annual corporate games weekend, and a boot camp program. Tracking data on participation has been an area of difficulty for the agency, but TARC has seen some success in encouraging participation by offering incentives and prizes to participants. According to the agency, these items are low cost ways to promote participation and camaraderie. TARC has also invested in creating onsite fitness centers at each of the agency’s main facilities, which the agency’s health insurance company has rewarded by issuing a premium refund to the agency and employees. Highest-scoring items from employee interview survey: • Walking to increase physical activity, • Having healthy snacks available for purchase at work, • Increasing my physical activity level, • Participating in “tasting” events, and • Learning about healthier food choices and portions to help manage my weight.

Case Studies: Health Promotion Programs 71 Specific program activities include weekly yoga classes, 5K runs and participant preparation assistance, periodic weight loss/weight maintenance challenges, walking events, and bioscreen- ings. A point system for participation allows employees to earn small prizes, such as exercise accessories, gear, or gift cards. 6.4.6 Organization The employee wellness program is led by a six-person health and wellness committee made up of representatives from TARC’s Human Resources Department and members of ATU Local 1447, including its president, an operator, and a mechanic. The committee meets every other month to determine upcoming program elements and themes. According to the union president, the relationship between labor and management regarding the program is cooperative. The union encourages participation in wellness program events and activities. 6.4.7 Resources The employee wellness program is funded through TARC’s Human Resources Depart- ment. In fiscal year 2018, $10,000 was budgeted for the agency’s fitness centers and wellness program. The wellness program also has relied on the portion of the health insurance premium refund retained by the agency after premium refunds were distributed to the participating employees. 6.4.8 Qualitative Program Benefits The union president said that the program has been effective in promoting physical activity, although it is not clear whether the employees who have participated are those who would already be active independent of the program. Events are primarily attended by the same group of people, and the program has not broadly affected the employee population. Management at TARC noted that the activities promote team building and encourage a more cooperative work environment. Aside from health outcomes, the program sends a message to the employees that health and wellness are priorities for the agency. 6.4.9 Participation Metrics Participation in several of the activities increased from 2017, the program’s first year, to 2018. For example, participants in the corporate games event rose from 25 in 2017 to 43 in 2018, a significant increase. Data from TARC’s health insurance provider also showed a growing level of involvement since the beginning of the transit agency’s wellness program. Participants are given points for reaching certain levels under the “Humana Go” program (blue, bronze, silver, gold, and platinum). Total participation increased from 84% of all health insurance subscribers (not including dependents and spouses) in 2016 to 94% in 2018. 6.4.10 Workforce Characteristics A total of 338 frontline operators and maintenance workers were employed in 2015 (at the time the program was introduced). Following the program’s rollout in 2016, of these 338 workers 13 employees had attended boot camps, 49 employees had a “high” Humana Go level (i.e., bronze, silver, gold, or platinum level), and 54 employees had attended a bio screening.

72 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line The project team selected these participation variables for analysis because they had the highest participation numbers of the numerous activities included in TARC’s health program. Table 38 presents demographic characteristics for both the total population of TARC’s frontline employees and for the agency’s operators and maintenance employees in 2015. As seen in Table 38, the White population was slightly older than the African-American population. The average age for the 108 White workers was 54.1 years, compared with 51.1 years for the 230 African-American workers. Men were slightly older than women, with an average age of 53.7 years for men compared with 49.8 years for women. The table replicates the pattern observed at IndyGo and RTS, where the majority of the maintenance workers were male. The project team used the data from TARC to further examine how outcomes related to absenteeism were related to participation in the program. To be counted, the workforce under observation in the two analyses performed had to have been employed with TARC since 2015, and their employment had to include at least 1 full calendar year during the period 2016–2017. In this case, multiple measures of participation (key independent variables) were used, and analysis was conducted to see if each one individually was associated with a change in absenteeism. Each category of participation was represented by groups with characteristics, as displayed in Table 39. As shown in Table 39, 13 employees participated in boot camps; these participants had a younger average age (45.3 years) compared to the non-participants (52.3 years). Forty-six employees had an elevated (silver-level or gold-level) Humana Go participation status, and Table 38. Demographics of TARC frontline population, 2015. Demographic Characteristic Operator Maintenance All Employees Count Percent Age * Count Percent Age * Count Percent Age * African American 228 80.6% 51.2 2 3.6% 44.0 230 68.0% 51.1 White 55 19.4% 55.6 53 96.4% 52.6 108 32.0% 54.1 Female 135 47.7% 49.8 1 1.8% 54.0 136 40.2% 49.8 Male 148 52.3% 54.1 54 98.2% 52.2 202 59.8% 53.7 Total 283 100.0% 56.8 55 100.0% 55.6 338 100.0% 56.5 * All ages are averages. Table 39. Program participation and age of TARC frontline population, 2016–2017. Program Participation Operator Maintenance All Count Percent Age * Count Percent Age * Count Percent Age * Key independent variable: participation in boot camps Participated in boot camps 4 1.4% 45.3 9 16.4% 52.4 13 3.8% 50.2 Key independent variable: elevated (gold, silver, or platinum) Humana Go status Base Humana Go status 245 86.6% 52.8 47 85.5% 52.8 292 86.4% 52.8 Elevated Humana Go status 38 13.4% 48.4 8 14.5% 51.8 46 13.6% 48.4 Key independent variable: attended bioscreen Did not attend bioscreen 241 85.2% 52.4 45 81.8% 52.7 286 84.6% 52.5 Attended bioscreen 42 14.8% 50.9 10 18.2% 52.3 52 15.4% 50.9 Total 283 100.00% 52.2 55 100.00% 52.6 338 100.00% 52.5 * All ages are averages.

Case Studies: Health Promotion Programs 73 these participants were younger on average (48.4 years) compared to employees who had a base (blue) level of participation (52.8 years). Fifty-two employees participated in bioscreens, and again had a younger average age (50.9 years) than employees who did not participate (52.5 years). In general, the employees who participated in the wellness program activities tended to be younger than those who did not participate. The Humana Go program was sponsored by the insurer. Employees received points for their participation in wellness program activities, including bioscreens. The points were added up to reach defined levels under the Humana Go program, progressing from blue (the base level) through bronze, silver, and gold, to platinum (the highest level). As an incentive to participation, employees also could earn prizes based on the points they accumulated (partici- pation level). 6.4.11 The Program Over Time Figure 12 presents the average annual number of absentee hours for frontline employees over the 4 years from 2015 through 2018. This analysis used 2015 as the baseline year because the comprehensive wellness program began in 2016. Only one pre-program data point was available, so the data shown in Figure 12 should be interpreted cautiously. Absenteeism days were defined as total hours of sick leave, personal leave, and unpaid leave. As the figure shows, the total average annual absentee hours increased from about 45 hours in 2015 to about 60 hours in 2018. It further shows that women had higher rates of absenteeism, which also was seen in other case studies. Figure 13 presents the average number of annual absentee hours for TARC operators and maintenance workers over the same period, compared to the averages for all frontline employees (Grand Total). As seen in Figure 13, operators consistently had a higher average of total annual absentee hours than did maintenance workers. Again, for all frontline workers, the average annual total absentee hours ranged from about 45 hours in 2015 to about 60 hours in 2018. 6.4.12 Workers’ Compensation Table 40 shows the number of indemnity claims by year from 2015–2017, with partial-year information from 2018 (the 4 years provided by the agency). It was not possible to associate claims with the individual/participant in the health claims, so a regression analysis examining the relationship between changes in claims and participation was not conducted. Figure 12. Average annual total absentee hours, TARC frontline employees, 2015–2018. 0 10 20 30 40 50 60 70 80 90 2015 2016 2017 2018 Black or African American White Female Male Grand Total

74 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line Table 40 demonstrates a trend commonly seen across the case studies, which was that indemnity claims were variable and subject to fluctuation due to factors such as a few high claims. In the case of TARC indemnity claims, the table also shows that the number of unique claims is not much higher than the number of employees with claims, indicating that the instance of repeat claimants was not large. 6.4.13 Results Many variations of the linear regression were performed. The project team varied the dependent variable, changed the mix of the independent variables, and tried several interaction terms of program participation (e.g., boot camp participation or Humana Go level interacted with various demographic variables). In no case was the coefficient on the effect of program participation significantly different from zero, and in no case did participation have a statis- tically significant effect on health (here measured as the change in days absent). Appendix C presents the results of the linear regression analysis, and within the appendix Tables C-4 (Humana Go), C-5 (boot camp), and C-6 (bioscreens) present representative regression results of the various models. 6.5 DART This case study was developed with the input of the human resources manager and chief human resources officer for DART, and the president and business agent of ATU Local 441. 0 10 20 30 40 50 60 70 80 90 2015 2016 2017 2018 Operator Maintenance Grand Total Figure 13. Comparing average annual total absentee hours for TARC operators, maintenance workers, and all TARC frontline workers, 2015–2018. Table 40. Workers’ compensation indemnity claims by year, TARC, 2015–2017 and part of 2018. Year Sum of Claims Unique Claims Average Cost per Claim Employees With Claims 2015 $498,767 63 $7,917 58 2016 $458,357 63 $7,276 60 2017 $1,033,219 76 $13,595 69 2018, January–May $260,673 36 $7,241 35 Total $2,251,016 238 $9,458 222

Case Studies: Health Promotion Programs 75 6.5.1 Background DART is the first regional transit authority in Iowa created under state legislation and was approved in 2005. The agency operates the largest transit system in Iowa, providing more than 15,000 trips per day with a fleet of approximately 145 buses. DART is expanding throughout its service area, introducing more express, shuttle, and weekend service hours. DART also has one of the largest vanpool programs in the Midwest, with more than 100 vans (Iowa DOT n.d.). More than 280 individuals are employed at DART, including its fixed-route and paratransit operators, maintenance and facilities staff, and administration (DART n.d.-b). Taking advantage of a change in leadership within both the labor union and the transit agency management, DART has worked to encourage employees to enroll in the existing health savings account plan and make lifestyle changes. In 2017, DART implemented a comprehensive wellness program for all employees. 6.5.2 Program Startup and Development DART’s annual wellness program began in October 2017. Before developing this program, the agency’s only targeted wellness-related activities were biometric screenings and health risk assessments (HRAs) (McMahon, personal communication, 2018). These programs started 2 years before the current wellness program. DART has promoted a rigorous safety program since 2007 and was recognized by APTA in 2011 for its achievements in building a strong safety cul- ture (DART 2011). The development of the wellness program indicates a shift toward a more holistic approach to the health and safety of its employees. The wellness program was begun for several reasons, including a high number of workers’ compensation claims, low morale, and low employee engagement, and to boost awareness of and participation in the existing wellness screening program and HRAs (McMahon, personal communication, 2018). To structure the program to best suit the needs of the employees, a wellness interest survey was given to employees before the program inception. Ninety percent of employees parti- cipated in the survey; the program was designed and budgeted based on their responses (McMahon, personal communication). A total of 201 survey responses (182 complete, 19 partial) were received in which employees identified desired topics, the length of activities, and most convenient times of the day for activities to take place (McMahon, personal communication). 6.5.3 Work Organization/Work Environment The union has worked with management to improve shifts for operators and therefore reduce the impacts of difficult working hours. The majority (53%) of operators work split shifts, arriving at 5:00 a.m. and working until 8:00 a.m. or 9:00 a.m., after which they break until 2:00, then work again until 6:00 p.m. Efforts have been made to reduce the length of the break between shifts. The union has been bargaining for better scheduling and has worked with management on this issue because it helps with worker retention. Maintenance workers have more standard shifts, working 8-hour or 10-hour shifts with a break scheduled midway during the shift. Restroom access for operators has been a longstanding issue. Management adjusted operator routes due to complaints of urinary tract infections caused by not being able to use the rest- room when needed. Recovery time is now spent at the station, so operators have access to the restroom there. Another issue for operators is proper positioning and type of seat. In 2016, DART bought new seats for their buses and allowed operators to choose the model. The agency also redesigned DART Program Elements • Monthly topics incorporated into the wellness program: back care, cold/flu prevention, diabetes, financial wellness, healthy cooking/ eating, heart health, physical activity, sleep management, stress management, and weight management; • All topics chosen based on survey responses indicating employee interests; • Two to three workshops per month (at DART); • One to three wellness challenges per month that focus on making lifestyle changes (outside of DART); • UnityPoint available at main campus or Central Station location once per month for coaching in operator lounge; and • Gifts/prizes based on participation. DART Survey Response: How Long Should Wellness Activities Last? Most employees believe activities should last between 30–60 minutes, depending on the activity. Averaged across all activities, 41.1% of respondents indicated that activities should last 30 minutes; 20.0% of respondents indicated they should last 45 minutes, and 22.9% of respondents indicated they should last 60 minutes.

76 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line the wheelchair securement stations so that they require less bending and stooping and allow more room to maneuver (reducing lower back pain). When operators have specific complaints about the seat, they are addressed. Often this is done by readjusting the seat or teaching the operator how to do so. The agency also implemented job offer testing to make sure that opera- tors are physically able to do all functional aspects of the position. Approximately 65%–70% of all frontline employees participate in the agency’s health insurance plan. DART requires an annual bioscreening for every employee enrolled in the health insurance plan. 6.5.4 Health, Wellness, and Safety Concerns The union stated that the most prevalent health concerns among its members were chronic pain from the demands of the job (e.g., back pain, injuries resulting from repetitive motion); high blood pressure; and metabolic disease (e.g., diabetes). The agency had a slightly different perception of the top health and safety issues, stating that the top three were weight management, cardiovascular health, and effects of the job (e.g., ergonomics/fatigue/stress management). The agency stated that their rates of medical disqualification among operators were low, but among those that had been disqualified, the primary reasons were diabetes and soft-tissue injuries, usually occurring in the shoulder due to repetitive movement. 6.5.5 Program Activities/Elements According to DART, the program does not focus on any one aspect of health and wellness but rather on caring for the whole person. To that end, the program is multifaceted, incorporating many different topics and methods of approach. DART has engaged insurance providers, financial planners, and registered nurses to deliver workshops and provide coaching and advice to participants (McMahon, personal communication, 2018). A monthly theme is chosen that corresponds with the interests recorded by employees in the initial survey. To complement the theme, two to three monthly workshops are given at DART, as well as one to three wellness challenges that encourage participants to make lifestyle changes. Participation is incentivized with gifts and prizes ranging from sports equipment to gift certificates. Rewards are given for attending workshops and completing the monthly wellness challenges. Participants receive a reward based on the tier they have reached at the end of the program: Tier 1 is reached by attending three workshops and completing three challenges; Tier 2 is reached by attending six workshops and completing six challenges; and Tier 3 is reached by those who attend all workshops and complete all challenges (McMahon, personal communication, 2018). Outside of the events organized as part of the program, the agency has implemented several policies to improve the work environment. DART has created a new vending program so that fruit, vegetables, eggs, and protein bars are available instead of the more common snacks found in vending machines. The agency also hired specialists to analyze the buses and create cards illustrating the stretches appropriate for operators and their environment. Frontline employees are not paid for the time they spend at wellness events. This has caused some reluctance among operators and maintainers to attend events. Administrative employees attend the events during their workday, and are therefore being paid for their time. 6.5.6 Organization The wellness program relies on the planning and support of a seven-member wellness committee. Positions are open to all departments within DART. The committee meets every month to prepare for the following month and make changes and adjustments to the program as needed. Currently, the committee is staffed by the human resources manager, an operations DART Survey Response: Topics of Interest • Back care, • Cold/flu prevention, • Diabetes, • Financial wellness, • Healthy cooking/eating, • Heart health, • Physical activity, • Sleep management, • Stress management, • Weight management, • Men’s and women’s health, and • Understanding medical insurance and other benefits offered at DART. DART Wellness Committee • Seven members, • Committee members from all departments, and • Monthly meeting to prepare for next month and make changes/ adjustments.

Case Studies: Health Promotion Programs 77 instructor, two fixed-route operators, a maintenance employee, an operations supervisor, and a transit planner (McMahon, personal communication, 2018). The wellness committee and the program have the support and participation of the local union thanks in part to the member- ship of its president, a fixed-route operator, on the wellness committee (McMahon, personal communication, 2018). 6.5.7 Resources DART’s wellness program has a relatively low budget (approximately $5,000 annually) and has relied on existing staff to manage the program rather than hiring dedicated staff. For 2017, approximately half of the budget was used for workshops and the other half for the purchasing of incentives. No additional major capital expenditures have been made. Instead, DART has used existing resources to provide programming. Several workshops have been provided at no cost to DART through leveraging connections with wellness organizations and professionals. 6.5.8 Qualitative Program Benefits Behavioral and cultural shifts have occurred both within the management of DART and within the employee community. A new leadership approach, brought about by a transition in management positions, has been instrumental in changing the environment and focus of the agency. DART appointed a new chief operating officer in October 2015 and a new chief human resources officer in December 2016 (DART n.d.-a). Within the employee community, the inclusion of influential individuals on the wellness committee has been an important component to foster a sense of ownership of the program. There is a focus on the personal participation and commitment to life changes of the committee members. The administration has taken the feedback received at the monthly wellness committee meetings and used it to structure the program and increased the budget for next year of the program based on the input of the wellness committee (McMahon, personal communication, 2018). Participation in the HRA was approximately 20% before the start of the wellness program because of workforce resistance to the biometrics screening, which was a requirement for being enrolled in the agency health insurance plan. Following the start of the program in October 2017, 100% of the agency’s employees participated in the subsequent HRA, which occurred the next month. DART recognized that a lack of clarity about whether the HRA was a required part of the bioscreening contributed to the initially low participation rate. Among the employees, some fear also had existed about what the results of the HRA would be used for. The start of the wellness program prompted more discussion with union leadership and with the employees in general, which led to a shift in perception and an increase in engagement, which was the most important and effective change (McMahon, personal communication, 2018). Despite these successes, management acknowledges that other elements of the wellness program have not reached all employees. Although events were held at different times of day and days of the week in an attempt to boost participation, scheduling remained an issue. The employees who did attend came to many of the events. The small percentage of employees who were very active in the program got the most benefit. From the perspective of the union, the program was beneficial in raising awareness of health issues and there was a general sentiment that it was a good idea. Most employees lacked a will- ingness to participate long-term in the program, however. Frontline employees were asked to attend program events in their free time while administrative employees were often on the clock during events; this created some resentment among frontline employees and exacerbated

78 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line problems of participation. Participation increased when incentives were offered, but the effect of the incentives dwindled over time. Issues related to participation were difficult to address given the varying shifts of frontline workers, particularly operators. The types of shifts that employees worked had an impact on how they engaged with the wellness program. Operators were more likely to participate in the scheduled events (presumably because these events fit into their breaks between shifts), whereas maintainers/mechanics used the onsite gym at a higher rate. After running for 9 months, the wellness program discontinued. Due to several agency staff members leaving who had been instrumental in the vision for the program, the committee was dissolved and no more regular events were scheduled. 6.5.9 Reported Metrics Participation data were collected and recorded for each event (McMahon, personal commu- nication, 2018). The goal for participation in the first year of the wellness program was 30% of employees. Actual participation, measured as having attended at least one activity/workshop, was around 42% of employees. Participation in the HRA and biometrics screenings increased from 20% before the start of the program to 100% in the month following the start of the program (McMahon, personal communication, 2018). Although the program was too brief to be able to measure changes in other metrics, such as workers’ compensation claims and absenteeism, program staff has continued to collect data to help assess the effectiveness of the program. 6.5.10 Workforce Characteristics A total of 245 frontline workers (operators and maintenance) were employed with DART at the beginning of 2016, a year before the program started in 2017. Table 41 presents demo- graphic information for the total population of frontline employees and separate break downs for the operators and maintenance employees. Unlike the other case studies, information on age at the individual level was not provided by DART. Compared to some populations in the other case studies, a larger share of this workforce (84.9%) was male. As in the other agencies, the majority of maintenance workers were male. The analysis method was to examine if and how outcomes related to absenteeism were related to participation in the program. Thus, the analysis divided the population of frontline employees into two groups: those who were recorded as having participated in at least one activity Demographic Characteristic * Operator Maintenance All Count Percent Count Percent Count Percent Asian 9 4.4% 3 7.1% 12 4.9% African American 64 31.5% 13 31.0% 77 31.4% Hispanic or Latino 14 6.9% 8 19.1% 22 9.0% Two or more races 1 0.5% 1 2.4% 2 0.8% White 115 56.7% 17 40.5% 132 53.9% Female 36 17.7% 1 2.4% 37 15.1% Male 167 82.3% 41 97.6% 208 84.9% All 203 100.0% 42 100.0% 245 100.0% * Age-related information at the individual level was unavailable for this case study. Table 41. Demographics of DART frontline population, 2016.

Case Studies: Health Promotion Programs 79 and those who were recorded as having participated in no activities. The characteristics of these two groups are displayed in Table 42. The activities included a series of workshops and wellness challenges. As Table 42 demonstrates, 16 employees (out of a total of 245) were recorded by human resources as having participated in at least one activity. Many of the individuals who participated in at least one activity participated in multiple activities. This analysis did not account for marginal gains associated with participation in multiple activities. 6.5.11 The Program Over Time Figure 14 presents the average number of annual absentee hours for DART frontline employees for a 3-year period (2016–2018). Absenteeism hours were defined based on total hours of sick leave, personal leave, and unpaid leave. Figures 14, 15, 16, and 17 show absentee days over time by demographic characteristics and by job classification. A full year of information was not available for 2018, so for 2018 the full year was estimated by comparing the available months with the previous year and assuming that the difference in hours remained the same between the first and last six months of 2017 and 2018. Figure 15 presents the median annual absentee hours for race and sex. Figure 16 presents the average annual absentee hours for DART frontline employees divided by job type over the same 3-year period (2016–2018). Figure 17 presents the absentee hours over time for all employees and those that participated or did not participate in any activities. On average, employees who participated in any activities Program Participation Operator Maintenance AllCount Percent Count Percent Count Percent Did not participate 191 94.1% 38 90.5% 229 93.5% Participated in at least one activity 12 5.9% 4 9.5% 16 6.5% All 203 100.0% 42 100.0% 245 100.0% Table 42. Program participation of DART frontline population, 2016. 0 50 100 150 200 2016 2017 2018 White Female Hispanic or Latino Male Black or African American Other Grand Total Figure 14. Average annual total absentee hours of DART frontline employees by demographic characteristics, 2016–2018.

80 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line Figure 15. Median annual total absentee hours of DART frontline employees by race and sex, 2016–2018. 0 20 40 60 80 100 120 140 160 2016 2017 2018 Black or African American White Female Male Grand Total Figure 16. Average annual total absentee hours of DART frontline employees by job classification, 2016–2018. 0 20 40 60 80 100 120 140 160 180 2016 2017 2018 Maintenance Operator Grand Total Figure 17. Total absentee hours, DART frontline employees, 2016–2018. 0 20 40 60 80 100 120 140 2016 2017 2018 No Participation Participated in Activities Grand Total

Case Studies: Health Promotion Programs 81 used fewer sick days than did non-participating employees; however, because only 16 employees are recorded as participating in any activities, this is probably a case of self-selection bias. 6.5.12 Results A set of 115 observations were available to test for whether program participation had an effect on health. The principal regression model was used to examine the relationship between program participation and absentee hours (see Table 43). The model included controls for race, gender, and type of employee (mechanic or operator). The coefficient estimate of –3.9 was statistically significant at the 95% confidence level. Thus, the project team estimated parti cipation in the program resulted in a 4-hour decrease in absentee hours. Further, race was found to be statistically significant at the 95% confidence level. Specifically, if an employee was White, then absentee hours decreased by 16 hours. No other variable was statistically significant. Because the sample of workers was small—only 12 operators and 4 mechanics participated in at least one activity—the results may be meaningful, but should be interpreted with caution. The results from this model demonstrate that it is possible to find an impact of a wellness program on one of the measures (absentee hours) that often is available at transit agencies. Other agencies may be able to use this approach to evaluate the effectiveness of their wellness programs. 6.6 LA Metro This case study was developed through emails and discussions with the International Union of Sheet, Metal, Air, Rail and Transportation Workers (SMART)–Metropolitan Transportation Authority (MTA) wellness manager and the vice-general chairman of SMART Local 1565. 6.6.1 Background LA Metro serves the 9.6 million residents of Los Angeles County, California, with 165 bus routes and a fleet of 2,308 buses alongside four light rail and two subway lines. Bus and rail operators at LA Metro totaled 4,397 employees in 2018. These occupations were represented by the United Transportation Union (UTU) until 2008, when the UTU merged with SMART. LA Metro also employs 2,370 mechanics, who are represented by the ATU. There are 9,817 total full-time employees at LA Metro. Together, SMART and the MTA manage a trust fund that administers benefits for bus and light rail operators. Ordinary Least Squares: Change in Hours Estimate t-Stat Intercept 38.0 3.60 Participation -3.9 -2.13 Male -5.9 -0.55 White -16.2 -2.32 Mechanic -3.5 -0.40 Observations 115.0 -- R2 0.078 -- Adjusted R2 0.046 -- Table 43. Effect of program participation on absentee hours, DART frontline employees, 2016–2017.

82 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line 6.6.2 Program Startup and Development In 2006, LA Metro piloted a health and wellness pilot program at two locations. The agency started the pilot to produce positive effects on absenteeism and workers’ compensation claims. At the end of the pilot program in 2009, LA Metro determined that it did not have the resources to continue to fund the program long-term; however, the value of the program had been recognized and the SMART-MTA trust fund stepped in to manage a permanent program. The program was expanded to 24 locations, including all of the main facilities. 6.6.3 Work Organization/Work Environment Currently, almost 50% of operators work split shifts, with the remainder split evenly between operators who work three shifts and those who work more traditional hours. Because of the demands of their schedules, fatigue can be an issue preventing operators from attending wellness program events. This type of schedule also can be seen as a benefit, however, because operators can use their breaks between shifts as an opportunity to rest, exercise, and/or participate in wellness activities. Since 2017, SMART has worked with the University of California, Irvine, to assess the workplace and job tasks that positively and negatively impact employees’ health and behaviors. The health program plans to use the findings from the university assessment to create program- ming to improve the overall health of employees and their families. The close attention to employees has helped uncover issues that can be resolved by influencing positive changes in corporate culture, policies, and procedures at LA Metro (e.g., schedule changes). 6.6.4 Health, Wellness, and Safety Concerns According to the SMART-MTA Wellness Program Strategic Plan 2018–2020, diabetes, hyper- tension, and cancer are targeted for disease management programs. The vice-general chairman at SMART cited diabetes, high blood pressure/hypertension, and stress as the top three health and safety issues with which the union and its members are most concerned. He shared that stress contributes to many of the health issues that operators suffer from, such as heart condi- tions and sleep apnea, which are causes for medical disqualification of operators, though the instances are low (Wormley, personal communication, 2019). 6.6.5 Program Activities/Elements The primary focus of the health and wellness program is to assist operators, but all employees, regardless of union affiliation, can participate. One program feature cited by the agency and union as leading to its success is the use of ambassadors. Ambassadors are selected from among the frontline workers to promote the wellness program. Each location has an ambassador, and large locations may have multiple ambassadors. Most locations have two wellness ambassadors from SMART (usually a main and an alternate). On specified days (called Wellness Wednesdays), the wellness ambassadors are given 8 hours of release time to engage employees in program activities. Ambassadors also are given hours of release time for offsite events, which are primarily weekend events. Compensation for ambassadors’ time spent on wellness program duties is covered by the MTA. LA Metro’s wellness program runs year-round and features disease management and edu- cation, seminars and table topics, fitness challenges, health fairs and screenings, free family sporting events, and a monthly wellness newsletter. Each year, eight health fairs are held at different facilities on a rotating schedule with the result that over 3 years, all locations hold a

Case Studies: Health Promotion Programs 83 health fair. Wellness activities occur mainly on Wednesdays and are scheduled to coincide with operator breaks between split shifts. The wellness program also has an incentivized weight loss program, called the Metrofit Club. The program is optional and requires a commitment of 10–12 weeks. Participants weigh in every other week with their wellness ambassador and receive assistance in their efforts through education on calculating caloric intake, recipe preparation, and basic nutrition. The program is incentivized with monetary rewards of up to $100 for losing a certain percentage of body weight. A concerted effort has been made to promote LA Metro’s wellness program. This has been done in several ways, including the presence of wellness ambassadors; union, employer, and health plan communication channels; incentives, rewards, kickoff events, challenges, and contests; a consistent theme and key messages; and mail, posters, email, newsletters, and social media marketing and testimonials. 6.6.6 Program Organization The health and wellness program is managed by a full-time coordinator. A health and wellness committee also provides input on programming and goals. The committee meets quarterly, is chaired by the wellness coordinator, and is represented equally by staff and labor members, though two unions working with LA Metro are not represented on the committee (the ATU and the Teamsters Union, which represents security guards). SMART is working toward a goal of including the Teamsters Union and the ATU on the committee, representing mechanics. Ambassadors are chosen jointly by union leadership and management. Every January, the ambassador roles and responsibilities are reviewed, and ambassadors are asked if they want to renew their contract. Training for new ambassadors occurs every quarter. 6.6.7 Program Resources Program costs, not including ambassador pay and the salary of the wellness program manager, amount to approximately $55,000 annually. Health insurance providers contribute to the budget as part of the services offered to employees in exchange for premiums; however, the insur- ance provider does not control the program fund itself. A union trust fund covers the ambas- sador pay (about $275,000 annually) and also covers the salary of the wellness program manager. 6.6.8 Qualitative Program Benefits The project team’s analysis indicated that the pilot program produced benefits. Injury- related claims decreased at some locations, and employees reported better sleep, weight loss, and reduced stress. Participation was tracked by employee badge number and showed that 382 employees participated in some element of the program during the pilot. Since 2009, the expanded program has seen increased levels of participation. Between 2009 and 2012, long-term goals of the program were to reach 10% partici pation and limit health insurance premium rate increases to no more than 5%. Increased partici- pation in the expanded program meant that more employees accessed services covered by the health insurance provider, which led to higher premium rates. Although the increased participation was a positive step, it negatively impacted the premium; for this reason, the well- ness committee changed the goal respecting premiums to maintaining a cost “less than the Southern California healthcare trends.” Program Promotion • Wellness ambassadors; • Use of all communication channels (union, employer, and health plan); • Incentives, rewards, kick-off events, challenges, and contests; • Consistent theme, key messages; and • Mail, posters, email, newsletters, social media, and testimonials.

84 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line SMART uses program data from its health insurance provider to tailor the program and counter cost trends. For example, high numbers of emergency room (ER) visits led to program education on how to avoid using the ER by scheduling appointments and going first to primary care. Results from biometrics screenings performed at the health fairs and data from LA Metro on employee metrics also are used to inform programmatic elements. Results are communicated in a newsletter for members with highlights of the changes in different measures. The program has resulted in policy changes that signify management and union willingness to work together toward the health of employees and ensure that the program receives the proper support and attention. When the program first started and the concept of well- ness ambassadors was introduced, management at LA Metro agreed to provide time off for the ambassadors’ participation in Wellness Wednesday activities. As the program progressed, wellness ambassadors were given paid time for Wellness Wednesdays, and eventually were given paid time off for events outside of work hours, including weekend events. This shift has raised the status of wellness ambassadors and the program generally. According to the program coordinator, the employees’ attitudes toward the program have changed from indifference to more overtly positive sentiments. Employees actively seek out elements of the program and are more willing to provide their success stories, which are published in the wellness newsletter. Union leadership expressed the view that members are genuinely excited about the program and appreciate the involvement of the union and management because it shows that both the union and management have taken a concern in the operators’ health. 6.6.9 Reported Metrics For this analysis, participation was defined as a person attending at least one event within a year. From 2012 through 2017, participation data from LA Metro showed generally positive trends, rising to 38% by 2014 and remaining close to that percentage in later years (Figure 18). Through the Metrofit Club, SMART has tracked the weight loss of participating employees. Figure 19 shows a peak in pounds lost in 2014, after which the amount of weight lost declined in 2015 and 2016. The drop-off may not be a negative trend, however, as the amount of weight lost in 2014 might mean that many participants had already reached or were approaching a healthy weight. Figure 20 shows the number of health club participants per year. The peak was in 2014 at 584 participants. There was a drop-off in 2016, but a slight increase in 2017 brought the number 38%37%38%38% 33% 28% 0% 5% 10% 15% 20% 25% 30% 35% 40% 201720162015201420132012 Figure 18. Participation in health program as a percentage of total LA Metro employee population, 2012–2017.

Case Studies: Health Promotion Programs 85 of participants back up to 446. The pattern seems to be fairly stable and all other years were higher than the initial year of 2012. Data from LA Metro was only provided in the aggregate, so the project team was unable to conduct regression modeling as was done with the other case studies. The information learned from this case study was based on LA Metro’s reports of employee participation in the activities and the overall weight lost by employees, but could not be correlated with outcomes such as absenteeism or controlled for race, gender, or type of position. 6.7 Summary of Case Studies The work organization and environment at each case study location was unique, and each agency faced different health, wellness, and safety challenges. Many commonalities were found across the locations, however: For example, at all five locations, the majority of bus operators worked split shifts, and some operators worked irregular schedules. The varied scheduling patterns impacted operators’ access to healthy food and their sleep patterns, and limited their ability to participate in certain health and wellness program activities. 6.7.1 Program Development and Work Environment The health and wellness programs examined were developed for various reasons and to meet various needs. For example, IndyGo added an onsite clinic as a way to avoid steep insurance premium increases. RTS began with a focus on physical fitness, but added more goals after several years, eventually hiring a full-time health and wellness coordinator. TARC’s 1,500 1,345 1,580 2,039 1,799 1,162 0 500 1000 1500 2000 2500 201720162015201420132012 Po un ds Figure 19. Weight lost per year (in pounds), Metrofit Club participants, 2012–2017. 446420 537 584564 305 0 100 200 300 400 500 600 700 201720162015201420132012 Figure 20. Number of Metrofit Club participants per year, 2012–2017.

86 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line initial focus was on smoking cessation, but the program was expanded to include more general wellness goals. DART encouraged employees to take advantage of all existing employee benefit programs, including a health savings account, and implemented a comprehensive wellness program for all employees. LA Metro’s program began as a pilot in two locations and expanded to 24 locations. At all the sites, at least half of the operators worked split shifts. This presented some chal- lenges for staff, including accessing healthy food choices and finding time for regular exercise. Irregular shifts also contributed to sleep deprivation. An issue emphasized at most sites by staff and union representatives was restroom access. 6.7.2 Health, Wellness, and Safety Concerns Comparing the top three health, wellness, and safety concerns expressed by management and labor representatives and examining insurance claims data, the project team found hyper- tension, musculoskeletal injuries (back and neck pain), and diabetes to be the most commonly found concerns for frontline transit workers across the locations (Table 44). Other areas of concern included sleep apnea, cardiovascular diseases (heart conditions), injuries from bus accidents, obesity, stress/fatigue, and cholesterol disorders (hyperlipidemia). Table 44 lists the top three health issues for each of the health and wellness programs discussed in the case studies and breaks down each issue by three sources: management, labor, and claims data. Not all locations provided data from all three sources. In several cases, management used analysis from insurance claims data to respond to the question about their top health, wellness, and safety concerns. At all five case study locations, labor listed diabetes as a major concern—indeed, in two of the five locations, it was the top concern. Hypertension also was named by labor in four of the five agency locations. Claims data added obesity (including hyperlipidemia) and back pain to the list of top health issues. Management, on the other hand, was more concerned with musculoskeletal injuries, weight management/cardiovascular health, and vehicular accidents. Figure 21 graphs the information presented in Table 44. Again, the most commonly mentioned issue was hypertension, followed by diabetes and musculoskeletal injuries. Areas with only one mention were included in the “Other” category. The distribution of concerns in Figure 21 Program Constituent Priority of Health/Safety Concern Primary Secondary Tertiary IndyGo Management Musculoskeletal injuries Slips, trips, and falls Vehicle accidents Labor Diabetes Sleep apnea Hypertension Claims data Obesity Hypertension Diabetes RTS Labor Sleep apnea Diabetes HypertensionClaims data Hypertension Cholesterol disorders Back and neck problems TARC Labor Operator assault Operator injury from accidents Breathing in harmful fumes Claims data Hypertension Hyperlipidemia Back pain DART Management Weight management Cardiovascular health Ergonomics/fatigue/stress management Labor Chronic pain from the job Hypertension Metabolic disease (e.g., diabetes) LA Metro Joint trust fund Diabetes Hypertension CancerLabor Diabetes Hypertension Stress Table 44. Comparison of top three health, wellness, and safety concerns at five case study locations.

Case Studies: Health Promotion Programs 87 broadly follows the data presented in Chapter 4 regarding the most prevalent health and safety issues for transit workers, with other key conditions also represented. Given the variations in data-supported or perceived health and wellness concerns, program design elements such as activities offered, facility needs, incentives for participation, staffing, organization of committees, and selection of champions were distinct from location to location. Chapter 7 presents process-driven strategies based on these case studies that transit systems can use to maximize program effectiveness. 6.7.3 Program Activities and Elements The programs offered various voluntary activities to employees, though it was common to provide incentives for participation. At IndyGo, participants were required to undergo a physical, health assessment, biometric screening, a minimum of four coaching sessions, and participate in a health activity to qualify for an insurance discount. Health activities might include gardening, Weight Watchers, exercise classes, walk–run groups, 5Ks, basketball tournaments, and/or financial or nutrition classes. RTS offered short workshops onsite, fresh fruit, team activities, different choices in their vending machines, blood pressure kiosks, health screenings, and a wellness center that includes a gym. TARC’s program began with a focus on smoking cessation but later expanded to provide events and programs organized around themes of interest, an annual corporate games week- end, and a fitness-oriented boot camp. The agency has offered some incentives and prizes to participants, but nothing systematic. TARC has provided its employees access to two onsite fitness centers, where they can participate in weekly yoga classes, 5K runs and participant preparation assistance, periodic weight loss/weight maintenance challenges, walking events, and bioscreens. Figure 21. Most common health, wellness, and safety concerns at five case study locations.

88 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line DART’s program did not focus on any one aspect of health and wellness but had a different theme each month that corresponds with the interests recorded by employees in the initial survey. DART has offered two to three workshops and one to three wellness challenges every month. Rewards were given for attending workshops. LA Metro has used wellness ambassadors and provided incentives, rewards, kickoff events, challenges, and contests. The wellness program featured disease management and education, seminars and table topics, fitness challenges, health fairs and screenings, free family sporting events, and a monthly wellness newsletter. Following a rotating schedule, eight annual fairs were held at different facilities so that, over 3 years, all locations had held a health fair. Wellness activities were scheduled to coincide with operator breaks between split shifts. The wellness program also had an incentivized weight loss program. 6.7.4 Organization Most programs were overseen by human resources departments and used third-party vendors to provide services. Several programs had full-time coordinators and volunteers (or paid employees) who served as wellness “coordinators” or “ambassadors.” The funding came from a mix of operating budgets, and agencies were able to detail staff who were already employed with the agency to serve the programs. All the sites examined had a wellness committee that was staffed with a mix of management, union representatives, and frontline staff. Committees met regularly and helped determine the activities and goals of the programs. These programs worked best when there was a cooperative relationship between management and the union. The programs demonstrated a wide range of budgets and operating processes. The best-funded of the case studies was IndyGo, which staffed a clinic with two nurse practitioners, a part-time doctor, and medical assistants. During the assessed period, IndyGo operated with a budget of $500,000 per year. RTS employed one full-time wellness coordinator and funded the program through the agency’s operating budget, using third-party vendors, spending approximately $41,000 per year. TARC’s program was funded by the agency’s human resources office with a budget of approximately $10,000 per year, though the program received additional funds via a premium refund from their health insurance carrier. DART had a relatively low budget of approximately $5,000 annually. Dart relied on existing staff members to manage the program rather than hiring dedicated personnel. Finally, LA Metro spent approximately $55,000 annually, not including the salary for the program coordinator. The health insurance provider contributed to the program budget through a negotiated premium arrangement, though the fund itself was not controlled by the health insurance provider. A union trust fund covered the ambassadors’ pay and the salary of the wellness program manager. 6.7.5 Workforce Characteristics Overall, a racial and gender divide was evident based on job roles. The majority of operators were male, but some gender diversity could be found, with one site having a male population of “only” 52.3% (see Table 45). Maintenance workers were overwhelmingly male, with no site lower than 93%. At all sites, at least three-quarters of maintenance employees were White. The demographics of the populations that are eligible or participate in the wellness programs can help agencies decide on how to focus their activities and how to market them effectively. Figure 22 shows the annual total average absentee hours for each of the case study sites. A great deal of variability can be seen across the agencies, which leads to the conclusion that each must be considered in a local context. Absenteeism seems to be a much greater issue in

Case Studies: Health Promotion Programs 89 some places than others: DART, in particular, experienced such high rates that it is possible to suspect some data discrepancy may explain it, though our discussions with the agency did not suggest this. IndyGo experienced a fairly steady rise in absenteeism beginning in 2014, which might be attributable to a structural change. RTS and TARC have more level numbers, but also seem to have experienced slight rises in absentee hours. This issue is one that agencies will want to continue to monitor. Although absenteeism seems like a good outcome variable for evaluation, it is open to many potential causes that a wellness program will not be able to address. 6.7.6 Conclusions The newly available primary source employee data from this study has provided informa- tive descriptive statistics and statistical results. Details have been included about how pro- grams were developed, the organizing process, and the services and activities offered by each site. Where possible, the project team gathered individual-level data on workforce character- istics, participation rates, and program metrics. The metrics gathered included claims data, data on specific disorders, prescription claims, absentee hours, and workers’ compensation claims. These data have provided a big picture understanding of workforce patterns—and how variable they are. The data examined in this chapter adds to findings from the literature review on the effective- ness of health promotion programs. Although these studies may not have produced measurable Agency Operator MaintenanceBlack Male Age a Black Male Age a IndyGo 85.7% 60.9% 54.7 24.1% 93.1% 58.5 RTS 39.3% 74.9% 56.8 33.3% 98.1% 55.6 TARC 80.6% 52.3% 56.8 3.6% 98.2% 55.6 DART 31.5% 82.3% Unavailable 31.0% 97.6% Unavailable LA Metro b Unavailable Unavailable Unavailable Unavailable Unavailable Unavailable a All ages are averages. b LA Metro did not share individual-level data. Table 45. Summary of wellness program participant characteristics. Figure 22. Comparison of average annual absentee hours across sites. 0 20 40 60 80 100 120 140 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 INDY GO RTS TARC DART

90 Improving the Health and Safety of Transit Workers with Corresponding Impacts on the Bottom Line effects that translate to quantifiable cost savings for agencies, the absence of concrete statistical evidence does not mean the cases lack substantial value. The data that was collected and analyzed adds a host of new information on employee patterns of absence/sick leave and how it relates to participation in health and wellness programs among different segments of the employee population. The project team could not identify a direct relationship between the programs offered and the outcomes examined, but the process followed offers a good way to understand how agencies may undertake such evaluations regarding their own programs. Having clear data available on participants, what programs they have participated in, and for how long, could make future research easier to undertake and interpret.