Workplace health and wellness programs are becoming a common employee benefit in the United States. Most recently, the RAND Workplace Wellness Programs Study found about half of employers with at least 50 employees, and more than 90 percent of those with more than 50,000 employees, offered a wellness program in 2012 (Mattke et al., 2013). In addition, a 2011 Aon Hewitt employer survey found that nearly 47 percent of employers without a wellness program planned to add one in the next three to five years (Aon Hewitt, 2012).
Wellness programs commonly screen employees—and, at times, dependents—for health risks through health risk assessment (HRA) surveys and biometric screening. These programs also provide interventions to address health risks and manifest disease, as well as promote healthy lifestyles. Wellness program popularity is mainly driven by employers' expectations that these programs will improve employee health and well-being, lower medical costs, and increase productivity.
As part of the requirements of the Patient Protection and Affordable Care Act (ACA) Section 1201, and sponsored by the Departments of Labor and of Health and Human Services, RAND recently completed a report to Congress on employer-sponsored wellness programs. That report assesses the current status of wellness programs offered by employers in the United States; evaluates the impact of programs on utilization, employee health care costs, health behaviors, and outcomes; and identifies best practices in program implementation. As the largest study on workplace wellness programs, it comprises a review of extant scientific and trade literature, a national survey of employers from the public and private sectors, statistical analyses using health care claims and wellness program data from a sample of employers, and case studies of the wellness programs offered by five heterogeneous employers. While the report fulfilled the requirements of the ACA, the data collected for this and other RAND projects provide a unique opportunity to address additional research questions.
The Office of Policy and Research (OPR) of the Employee Benefits Security Administration (EBSA) within the U.S. Department of Labor (DoL) contracted with RAND to conduct an analysis of existing data on wellness programs. This article describes the results of that analysis. The goals of the project were to leverage existing data to explore patterns of wellness program availability, use of incentives among employers, and program participation and utilization among employees.
We used two sets of data for this project. First, we used the 2012 RAND Employer Survey data set (Mattke et al., 2013), which used a nationally representative sample of U.S. employers that had detailed information on wellness program offerings, program uptake, incentive use, and employer characteristics. We used these data to answer questions on program availability, configuration, uptake, and incentive use. Second, we utilized the health care claims and wellness program data for a large employer. We analyzed these data to predict program participation and changes in utilization and health (for more details, see Mattke et al., 2013).
Analysis of Employer Survey Data
The 2012 RAND Employer Survey solicited detailed information on workplace wellness programs, including the use of incentives, barriers to adopting a wellness program, reasons for discontinuing a wellness program, program evaluation, and program costs. The sampling frame was a stratified random sample of 3,000 public- and private-sector businesses employing at least 50 people, drawn from the 2011 Dun and Bradstreet Employer Data (Hoovers, undated). The response rate was 19 percent, resulting in a final sample of 589 employers.
We focused on two main research questions:
- Which employer characteristics predict program availability, program configuration, and incentive use?
- Are there typical program configurations—such as the combination of certain screening and health management interventions—and how are these configurations related to employee participation?
We also conducted analyses to determine whether framing incentives as rewards or penalties affected program participation rates. The analyses are based on descriptive statistics and logistic regression models that adjust for employer characteristics; all statistical analyses were weighted to make the sample nationally representative.
Which Employer Characteristics Predict Program Availability and Use of Incentives?
In our sample, 69 percent of employers offered a wellness program; of those that offered a program, 75 percent offered incentives to encourage program uptake.
Employer size was the most important predictor of whether the employer offered a program and how the program was configured. We estimated that about a third of the smallest employers (50 to 100 employees) had a wellness program, compared with about four-fifths of the larger employers (more than 1,000 employees), after adjusting for employer characteristics. Similarly, about 60 percent of the smallest employers and 90 percent of other employers used incentives, mostly monetary ones, to promote program uptake.
Smaller employers without a wellness program tended to cite lack of financial resources and lack of cost-effectiveness as reasons, whereas larger employers were more likely to cite lack of employee interest. Thus, cost concerns appear to explain the different decisions of smaller employers. This finding has important policy implications related to resources for smaller firms to provide wellness programs, since about 36 percent of Americans work for employers with fewer than 100 employees.
How Are Wellness Programs Configured?
We used cluster analysis to determine five common combinations, or “program configurations,” within wellness programs, with most employers opting for a program with a limited range of services (Table 1). These five configuration categories reflect the services offered in a wellness program within the three main service categories: (1) screening to detect health risks, (2) lifestyle management to reduce health risks and encourage healthy lifestyles, and (3) disease management to support individuals with manifest chronic conditions.
Table 1. Program Configuration Labels and Definitions
| Program Configuration | Definition | Frequency (%) |
|---|---|---|
| Limited | Limited services across all three components | 34 |
| Comprehensive | Extensive services across all three components | 13 |
| Screening-focused | Broad range of screening services, limited other components | 20 |
| Intervention-focused | Broad range of lifestyle and disease management services but limited screening | 21 |
| Prevention-focused | Broad range of screening and lifestyle management services but limited disease management | 12 |
SOURCE: RAND Employer Survey 2012, Mattke et al., 2013.
Limited programs were particularly popular among smaller employers (50 to 100 employees): 70 percent of them offered this configuration, compared with only 41 percent of employers with 101 to 1,000 employees and only 36 percent of employers with more 1,000 employees.
Are Incentives Increasing Program Uptake?
Employers that did not use incentives reported lower participation rates and framing incentives as penalties was associated with higher participation rates. In the absence of incentives, employers reported a median participation rate of only 20 percent (see Figure 1). Uptake appears to increase with the use of monetary or nonmonetary incentives, with a median participation rate of 40 percent. If penalties or surcharges for not participating were used, the median participation rate was 73 percent.
Program configuration also influenced participation. Employers with comprehensive programs reported the highest participation rate (59 percent). Participation in these programs was less sensitive to choice of incentive schemes, as Figure 1 illustrates.
Figure 1. Relationship of Incentives and Program Configuration to Participation Rates
Analysis of Wellness Program Data
We used data from a Fortune 100 employer to examine which employees participate in wellness programs and to estimate the effects of incentives on participation rates, as well as the effectiveness of participation on cost and utilization. The data included health care claims and wellness program data. They covered seven years of the employer's program and two baseline years for nearly 200,000 unique employees and dependents, or 730,000 full person-years.
The employer's wellness program consisted of a health risk assessment with questions on health and health-related behaviors, a lifestyle management component to address health risks, and a disease management component to support employees and dependents with manifest chronic conditions.
We focused on two overarching questions:
- Which employee characteristics predict program participation and are incentives changing these relationships?
- What changes in utilization and health are related to program participation?
To answer those questions, we used descriptive analyses and multivariate regression models. Although we used rigorous econometric techniques common in the program evaluation literature (Imbens and Wooldridge, 2009), we note that the standard caveats for results from observational data analysis apply. In addition, we developed a simulation model to project the health impacts of wellness program participation over a 20-year time frame based on a nationally representative sample of working-age adults.
Which Employee Characteristics Predict Program Uptake?
Overall, only one-fifth to two-fifths of employees annually participated in wellness program components for which they were eligible, and predictors of program uptake varied by program component (e.g., lifestyle management, disease-management, predisease management). Healthier employees and those residing in higher-income ZIP codes were more likely to complete the HRA. For disease management, employees with multiple chronic conditions and older employees were more likely to participate. By contrast, health risk factors—with the exception of increasing body mass index (an indicator of weight status), did not predict lifestyle management and disease management uptake.
How Do Incentives Alter an Employee's Decision to (or Not to) Participate?
As of the fifth program year, the employer had introduced $600 surcharges for smokers who did not participate in a smoking cessation intervention, and for employees who were eligible for disease management but declined participation. The introduction of these penalties allowed us to estimate their effect on participation rates and participant characteristics.
Our estimates suggest that the surcharge was associated with a statistically significant increase in smoking cessation program uptake of 8.5 percentage points, but the participation rate remained well below 30 percent. Counterintuitively, the introduction of the surcharge in the disease management program was associated with a significant decrease in program uptake, about 18 percentage points. With the exception of higher disease burden, we found that few employee characteristics moderated the effect of the nonparticipation surcharges on program uptake.
Do Health Care Utilization Patterns Change Following Program Participation?
In the 2013 report, we found that participation in lifestyle management programs was not associated with significant changes in overall cost or utilization. Examining this relationship at a more granular level in this report, we still find no significant cost savings or reduction in utilization. Participants in the telephonic coaching component of the lifestyle management did not exhibit lower rates of all hospital admissions, hospital admissions for wellness sensitive conditions, or emergency room visits.
Is There a Differential Effect of Selected Program Components on Medical Costs?
We hypothesized that participation in lifestyle management components, which targeted individuals with higher health risks, might be more likely to achieve reductions in health care cost. To test this hypothesis, we analyzed data for participants in the smoking cessation intervention and in the so-called predisease program for employees who were on the verge of developing a manifest chronic disease. We found no evidence of cost savings among participants of either the smoking cessation or the predisease management program.
Is There a Dose-Response Effect on Medical Costs in the Intensity of Program Interventions?
Greater exposure to the wellness program, through participating in more telephonic counseling sessions, also was not associated with greater health care cost savings. In fact, per-member-per-month health care costs were approximately $20 higher for high-intensity participants (those who attended five or more sessions per year) than for their lower-intensity counterparts.
What Are the Long-Term Effects of Lifestyle Management on Health and Cost?
Lower cardiovascular event rates due to wellness program participation reduced costs, but these savings did not come close to offsetting cumulative costs of participation. We used estimates from our previous report on lifestyle management program uptake and effect on health risks—such as smoking, weight, and cholesterol levels—to simulate the impact on a working-age population over a 20-year horizon. The results, under realistic assumptions, simulated a modest reduction of 257 cardiovascular deaths and 1,796 nonfatal cardiovascular events in a population of 100,000 people, at an estimated cost of about $40,000 per avoided event. While improving employee health certainly has additional benefits, these numbers suggest that employers will find it difficult to achieve financial gains from saved health care costs alone from lifestyle management.
Conclusions
The findings in this article underscore the increasing prevalence of worksite wellness programs. About four-fifths of all U.S. employers with more than 1,000 employees are estimated to offer such programs. For those larger employers, program offerings cover a range of screening activities, interventions to encourage healthy lifestyles, and support for employees with manifest chronic conditions.
Smaller employers, especially those with fewer than 100 employees, appear more reserved in their implementation of wellness programs. They are less likely to offer any program, have typically limited program offerings, and voice concerns about the business rationale.
In spite of widespread access, the actual use of wellness programs by eligible employees and/or dependents remains limited. Our analysis of data from the large employer shows that only 20 to 40 percent of eligible individuals participate in a program in any given year. Participation rates as reported in our survey suggest a median rate of 40 percent.
Thus, it is not surprising that employers are attempting to boost program uptake with rewards and penalties. And employers state that incentives have the intended effect: In our survey data, median program participation rates were reported to be 20 percent for employers who did not use incentives, compared with 40 percent for employers that used rewards and 73 percent for employers that used penalties and/or rewards.
However, the analysis of actual participation data from a large employer paints a less optimistic picture regarding the effectiveness of penalties, as a $600 penalty increased participation in a smoking cessation program by only 8 percent and failed to raise uptake of a disease management program. Combined with the observation from the survey that participation rates in comprehensive wellness programs appear to vary less with incentive use, this study suggests that employees consider factors in addition to incentives, such as program design and accessibility, when contemplating whether to join.
In this context, it is important that our results contribute to the literature that documents an inability of lifestyle management portions of workplace wellness programs to reduce health care cost. We have previously shown that lifestyle management participation is associated with reduction of health risks (such as smoking and being overweight), but not with lower cost. In this study, we analyzed whether cost savings might be realized in higher-risk employees and in those who are more engaged in the program, but we found no evidence to support this hypothesis.
In addition, we extrapolated the impact of the estimated health effects of lifestyle management programs on the risk of cardiovascular events and found that an employer with 100,000 employees would see only 1,796 fewer cardiovascular events and 257 fewer cardiovascular deaths over a period of 20 years under current estimates for program participation and effect.
While we have to acknowledge the limitations of our study, such as the nonexperimental design and the fact that we analyzed data from only one employer, our findings contribute to evidence on the currently prevailing type of lifestyle management programs that are offered by employers; i.e., telephonic coaching for employees with health risks. If similar findings were reproduced in further research, they would imply that screening large numbers of individuals for health risks combined with education and one-on-one coaching for those with risks appears not to be effective or cost-effective enough to have a meaningful impact on the health of America's workers and the cost of health coverage.
Our assessment contrasts with a widely quoted meta-analysis by Baicker et al. that estimated a reduction in health care cost of about $3 for every dollar invested in workplace wellness programs (Baicker, Cutler, and Song, 2010). This study, however, has been criticized for including studies that were several decades old and had substantial methodological weaknesses. Our findings also appear surprising when compared to the analysis of the Johnson & Johnson program, one of the best-known and longest-running workplace programs in the United States (Henke et al., 2011). Its most recent evaluation suggested a return on investment in a range of $1.88–$3.92 saved for every dollar spent. It should be kept in mind, though, that this evaluation compared health care cost for Johnson & Johnson to those of similar employers, correcting for workforce differences. Consequently, the effect estimates are reflective of the overall differences in health care coverage and health management between Johnson & Johnson and the reference companies and effect estimates may partly be attributable to unobservable firm-level differences, such as health care coverage, health management, or culture between Johnson & Johnson and the reference companies.
Because we compared program participants to statistically matched nonparticipants, our design is more reflective of the actual incremental effect of participating in the telephonic coaching interventions under a workplace wellness program. Of note, nonindustry-sponsored studies that use a similar design tend to arrive at similar conclusions. For example, a 2012 evaluation of the University of Minnesota's wellness program found no savings to be associated with lifestyle management program participation.
Apart from replicating our findings in a larger sample of employers, future research should investigate the potential of “personalized” wellness programs, which match intervention modality, intensity, and objectives more closely to an individual's beliefs, attitudes, and preferences, and of “public health” type programs, which aim to create a culture of health in the workplace rather than targeting individuals.
References
Aon Hewitt, 2012 Health Care Survey, Chicago, Ill., 2012.
Baicker, K., D. Cutler, and Z. Song, “Workplace Wellness Programs Can Generate Savings,” Health Affairs, Vol. 29, No. 2, 2010, pp. 304–311.
Henke, Rachel M., Ron Z. Goetzel, Janice McHugh, and Fik Isaac, “Recent Experience in Health Promotion at Johnson & Johnson: Lower Health Spending, Strong Return on Investment,” Health Affairs, Vol. 30, No. 3, 2011, pp. 490–499.
Hoovers, a Dun & Bradstreet company, Index, web page, undated (subscription required).
Imbens, Guido W., and Jeffrey M. Wooldridge, “Recent Developments in the Econometrics of Program Evaluation,” Journal of Economic Literature, Vol. 47, No. 1, 2009, pp. 5–86.
Mattke, S., H. Liu, J. P. Caloyeras, et al., Workplace Wellness Programs Study, Santa Monica, Calif.: RAND Corporation, RR-254-DOL, 2013. As of May 1, 2015:
http://www.rand.org/pubs/research_reports/RR254.html
This research was conducted in RAND Health Advisory Services, the consulting practice of RAND Health.
