The recent Health Affairs Blog post by Al Lewis, Vik Khanna, and Shana Montrose titled, “Workplace Wellness Produces No Savings” has triggered much interest and media attention. It highlights the controversy surrounding the value of workplace health promotion programs that 22 authors addressed in an article published in the September 2014 issue of the Journal of Occupational and Environmental Medicine titled, “Do Workplace Health Promotion (Wellness) Programs Work?” That article also inspired several follow-up discussions and media reports, including one published by New York Times columnists Frakt and Carroll who answered the above question with: “usually not.”
There are certainly many points of contention and areas for continued discussion on this topic. It turns out that Lewis et al. and I agree on many things, and there are other areas where we see things differently.
Where we agree…
Biometric screenings. Biometric screenings are important for collection of baseline health risk data and are often viewed as an added value by employees participating in workplace health promotion programs. Lewis et al. and I agree that employers should screen their workers for health risks in accordance with guidelines recommended by the U.S. Preventive Services Task Force (USPSTF). These guidelines are clear about the necessity and periodicity of biometric screenings for high blood pressure, obesity, cholesterol, glucose, triglycerides, cervical cancer, colon cancer, breast cancer, and other conditions.
We agree that over-testing people is not a good idea and may lead to false positives, as well as unnecessary medical interventions that are costly and add little value. For readers seeking guidance on biometric screenings in a workplace setting, I refer them to a peer-reviewed article published in the October 2013 issue of the Journal of Occupational and Environmental Medicine.
Incentives. Workplace health promotion programs are not the same as incentive programs. “Smart” incentives are part of a well-designed program, but such programs need to be embedded in healthy company cultures wherein employers encourage and reward healthy behaviors. Comprehensive wellness programs often use financial incentives to attract participation and, in some instances, encourage behaviors that lead to risk reduction.
Most experts in workplace health promotion agree that creating intrinsic motivation for health improvement is an essential component of an effective program. As Daniel Pink points out in his book Drive, people, including workers, are motivated to behave a certain way when they feel a sense of autonomy, when they are able to master certain skills needed to change a behavior, and when they can connect changing that behavior to a larger purpose in life. This applies to individuals wishing to achieve certain health goals such as quitting smoking, being more physically active, and eating a healthier diet.
Paying people to improve their health in an unhealthy work environment is a futile strategy. Workers will expect higher payments each year, will view “non-compliance” as a penalty, and will mistrust their employer for trying to do things to them instead of with them.
To summarize, incentives need to be practical, ethical, and legal. The Affordable Care Act (ACA) legislation should not be used as a vehicle or excuse for “blaming” workers for poor health habits, or to penalize them financially for not achieving certain health outcomes. Employers share responsibility for the health and well-being of workers and can do much to create a healthy company culture.
For readers interested in a more in-depth discussion of health promotion incentive programs, I refer them to a series of Health Affairs Blog posts, and a guidance document prepared by the Health Enhancement Research Organization, American College of Occupational and Environmental Medicine, American Cancer Society and American Cancer Society Cancer Action Network, American Diabetes Association, and American Heart Association.
Culture of health. We also agree that effective workplace health promotion programs need to be embedded within a culture of health that respects workers’ rights to make informed choices about personal health matters. Without question, workplaces need to be safe and employees need to be treated with respect and dignity. Workers also have a right to be in a healthy work environment where positive health behaviors are encouraged and supported. That means making healthy food available in vending machines and cafeterias, encouraging physical activity, prohibiting on-site smoking, offering vaccination programs, and providing health insurance.
The list of programs, policies, and environmental supports for a healthy workplace is long and there are hundreds of environmental and policy interventions available to employers who wish to send a clear message that the company encourages and supports good health. For a more complete discussion of how companies can achieve a healthy culture, see the May 2013 issue of The Art of Health Promotion.
The importance of studying “wellness-sensitive” events…in addition to overall utilization and costs. Lewis et al. highlight the need to focus on “wellness-sensitive” medical events when conducting cost analyses. I agree but ask the authors to address the following questions: What are these events? Where have they been published? Who has reviewed them? Why do they only apply to inpatient claims? Are there not any “wellness-sensitive” events that would appear in outpatient settings?
The idea of analyzing claims for conditions likely to be most readily influenced by health promotion programs is sensible. In many of our studies, we have analyzed utilization and cost patterns for what we call “lifestyle diagnosis groups” or LDGs. For example, in a 1998 peer-reviewed study, we evaluated Procter and Gamble’s health promotion program and found a 36 percent difference in lifestyle-related costs in the third study year when comparing 3,993 program participants to 4,341 non-participants.
Although it’s important to analyze a subset of diagnoses when evaluating wellness programs, it is equally important to analyze utilization and costs for all conditions. After all, one’s actual well-being and perception of well-being influences health holistically; not just any one particular organ or body system.
Where we disagree…
Whether only randomized trials can determine whether workplace programs are effective. Health services research and the field of epidemiology have had a long track record of studying naturally occurring phenomena and drawing conclusions from observations of those phenomena. That’s how we have learned what causes hospital-acquired infections. We have also learned from long-lasting epidemiological investigations like the Framingham studies that a sedentary lifestyle, smoking, and obesity are causes of heart disease, diabetes, and cancers.
These “natural experiments” inform the scientific community about what happens to individuals or groups “exposed” to a condition whereas others are not. Natural experiments are employed when a randomized controlled trial (RCT) is impractical or unethical.
How does this apply to evaluation of workplace health promotion programs? Imagine trying to convince the head of human resources of a company to approve a double-blinded randomized trial that would test the effectiveness of a wellness program, over three to five years, by randomly assigning some workers to a comprehensive health promotion program that includes health coaching on smoking cessation, weight management, physical activity, and stress reduction while other workers are denied access to the program. Not only that, the HR executive would be asked to allow the researcher to administer a series of blood tests to participants and non-participants, access their medical claims, and ask workers to complete periodic health surveys. The employer would also be prohibited from instituting organizational policies promoting health while this experiment is underway.
It’s hard to imagine a situation in which a company executive would allow this, never mind an institutional review board at a university.
Alternatively, when health services researchers conduct natural experiments, care is taken to control for as many confounding variables and address alternative hypotheses. In our research, we use statistical techniques such as propensity score matching and multivariate regression to compare the health and cost experience of “treatment” workers (those offered health promotion programs) and “comparison” workers (those not offered the programs). Most often, when comparing participants to non-participants, we match entire populations exposed to a program (whether or not individuals participate in that program) and those not exposed. In that regard, we are investigating program impacts on population health and not only comparing outcomes for motivated participants in programs compared to less motivated non-participants.
We publish our analyses in peer-reviewed journals so that the scientific community can review and critique our methods. We are also transparent about the limitations to our research in these peer-reviewed articles.
By the way, there are experimental studies focused on large populations (not necessarily at the workplace) demonstrating the value of health promotion programs. One such trial was recently concluded by the Centers for Medicare and Medicaid Services (CMS) as part of the Senior Risk Reduction Demonstration (SRRD). Two vendors were involved in the demonstration, which lasted two to three years Beneficiaries participating in Vendor A’s risk-reduction programs achieved statistically significant improvements in stress, general well-being, and overall risk, and beneficiaries participating in Vendor B’s program achieved statistically significant improvements in back care, nutrition, physical activity, stress, general well-being, and overall risk.
Interestingly, the interventions were determined to be “cost-neutral,” meaning that Medicare spending for participants in the intervention group was not statistically different from spending for participants in the control group. This was a large scale study where about 50,000 beneficiaries were recruited and approximately 20,000 participated in the health promotion program in any given year. The bottom line: Significant health improvements were achieved at no cost to the government.
Interpreting the data. Lewis et al. highlight mistakes and errors in others’ presentation of results. I have no argument with that. That is, after all, what a peer review process is all about: conduct the study, subject it to peer review, and publish the findings. The problem is that Lewis et al. have not (yet) published any studies in which their interventions are evaluated, nor have their methods been subject to peer review. That is unfortunate because I believe (truly) that all of us can learn from vetted research studies and apply that knowledge to future evaluations.
I am the first to admit that the methods we use to evaluate wellness programs have evolved over time and are still undergoing revisions as we learn from our mistakes. I invite Lewis et al. to reveal their methods for evaluating workplace programs and to publish those methods in peer-reviewed publications — we can all benefit from that intelligence.
One mistake Lewis et al. point to is a study conducted by Health Fitness Corporation (HFC) for Eastman Chemical Company, which earned the company the C. Everett Koop Award. (In the spirit of full disclosure, I am the President and CEO of The Health Project. which annually confers the Koop prize to organizations able to clearly and unambiguously document health improvements and cost savings for their employees.) Eastman Chemical’s application is online and subject to review.
In their analysis of the Eastman Chemical application, Lewis at al. complain that costs for participants and non-participants diverged in the baseline years of the program and therefore it was not the program that explains cost savings. Here’s the real story. Eastman Chemical’s program has been in place since the early 1990s. The chart found on the website (unfortunately mislabeled) shows participant and non-participant medical costs at baseline (2004), in subsequent years, and the final year of the study (2008).
The study compares medical and drug claims for minimally engaged (non-participant) and engaged (participant) employees matched at baseline (using propensity score matching) on age, gender, employee status, insurance plan, medical costs, and other variables. No significant differences were found between participant and non-participant costs at baseline, but their claims experience differed significantly at follow up. Although not a perfect study, the economic results, coupled with significant and positive health improvements in many of the health behaviors and risk factors examined for a multi-year cohort of employees, convinced The Health Project Board of Directors that Eastman Chemical earned the C. Everett Koop prize in 2011.
Whether return-on-investment (ROI) is the only metric for evaluating workplace health promotion programs. It seems that too much of the debate and controversy surrounding workplace health promotion is focused narrowly on whether these programs save money. If that were the aperture by which we judged medical care in general, we would withhold treatment from almost every patient and for almost every procedure, with the exception of a few preventive services that are either cost-neutral or minimally cost-saving. That makes no sense for a compassionate society.
In a February 2009 Health Affairs article, I argued that prevention should not be held to a higher standard than treatment; both should be evaluated on their relative cost-effectiveness (not cost-benefit) in achieving positive health outcomes and improved quality of life.
Take a simple example of two employees. One has just suffered a heart attack and undergoes a coronary bypass procedure. If the individual is willing, he is then engaged in counseling that encourages him to quit smoking, become more physically active, eat a healthy diet, manage stress, take medications to control blood pressure, and see the doctor for regularly scheduled preventive visits. For that individual, I would be surprised if an employer providing medical coverage would demand a positive ROI.
How about a second employee? That person is overweight, smokes cigarettes, eats an unhealthy diet, is sedentary, experiences stress at work, and has hypertension. He has not (yet) suffered a heart attack, although most would agree he is at high risk for such an event. To justify a health promotion program for that employee and others in the company, many employers insist on a positive ROI. Why is that a requirement? If a well-designed program can achieve population health improvement (as demonstrated using valid measures and an appropriate study design), and the program is cost-neutral or relatively inexpensive, why wouldn’t an employer invest in a wellness program, especially if is viewed as high value to both workers and their organization?
It’s time to change the metric for success. Instead of demanding a high ROI, employers should require data supporting high engagement rates by workers, satisfaction with program components, population health improvement, an ability to attract and retain top talent, fewer safety incidents, higher productivity, and perceived organizational support for one’s health and well-being. That’s where program evaluations should be focused, not simply on achieving a positive ROI.
I appreciate the reality that some employers may still require an ROI result. Fortunately, there is evidence, published in peer-reviewed journals, that well-designed and effectively executed programs, founded on best practices and behavior change theory, can achieve a positive ROI. I won’t re-litigate this point, other than to point newcomers to a large body of literature showing significant health improvements and net cost savings from workplace heath promotion programs. (See, for example, studies for Johnson & Johnson, Highmark, and Citibank and several literature reviews on the topic).
I challenge proponents and opponents of workplace wellness to direct their energy away from proving an ROI metric to measuring one or several important outcomes of interest to employers, as listed above. Achievement of these outcomes is only possible when management and labor work toward a mutually beneficial goal — creating a healthy workplace environment. Health promotion programs require time to take root and be self-sustaining, but the benefits to employees and their organizations are worth the effort.