To the Editor:
Goetzel et al's1 defense of employment-based wellness programs is timely. Not 1 week after it appeared, Frakt and Carroll2 in their New York Times article added yet another voice to the growing chorus of skeptics about the effectiveness of these programs.3,4
This debate is being fueled by the marketplace with a helping hand from government. Both the number of employers developing wellness programs and the amount of investment they are making have continued to rise sharply, spurred by the Affordable Care Act.5 Under the latter's final rule effective January 1, 2014, the maximum permissible reward or penalty under a plan-based health-contingent program increased from 20% to 30% of the total annual premiums for individual coverage and up to 50% for programs targeting tobacco use.6 All the while, a new round of health care cost increases that are greater than general inflation may soon be in the offing, at least in certain locales.7
In a view that borders on the cynical, Frakt and Carroll2 hold that the best that most wellness programs deliver are results that benefit employers—cost shifting that has employees who do not meet program demands and goals paying more, and camouflaging “gloss” that can be put on undesirable health coverage changes when the latter are implemented around the same time. If any progress at all is made on the nominal goal, cutting costs by improving employee health, the gains will be modest, come only after years of program operation, and be limited to disease not lifestyle management.
To counter such claims, Goetzel et al build on that time-honored academic edifice, Donabedian's structure-process-outcome paradigm, to mount a broad but heavily caveated argument supported by a number of cited studies. They lay out a measured approach that, while not offering explicit financial estimates, is predicated on the idea that programs that adopt best practices and create cultures of health exert positive effects on health and cost outcomes of import to employers.
They note that the field has largely shifted away from cost/benefit analyses, which produce the purely monetary values needed for return-on-investment (ROI) estimates, to analyses of a more cost-effectiveness bent that pit costs incurred relative to the extent of changes achieved on such parameters as lost productivity and disability. It is also important to recognize, though, that it was prior studies by the first author and others on ROI that have done much to establish their “street cred” expertise in this area. The return of some $2 to $10 for every $1 expended (depending on what or who is being targeted)8 reported in these studies has gone a long way toward setting the expectations of what can be achieved by wellness programs that “work.”
Yet, in light of peer-reviewed evidence that has been reported in this journal, one cannot help but wonder if Goetzel et al are not hugely underestimating the potential of these programs to alter the trajectory of the cost burden being borne by employers and employees. Furthermore, just how practical is what they propose when considered in its entirety in light of the real world of tight timelines and constrained resources that so often characterize the business environment? This reality is driven by the financial bottom line. It requires working within the bandwidth that management can devote to such issues. It often curtails capacity to address academic priorities.
Goetzel et al1 lay out qualifier attributes that programs should have if they are to be positioned for success; for example, evidence-based; properly designed/executed/evaluated; adequately resourced; culturally supported; and framed with appropriate expectations. Within this context, the authors:
- Put forth a broad array of questions to ask and measures to assess that focus on the above three components—program structure, delivery process, and expected clinical, utilization/cost, and productivity outcomes.
- List best (and promising) practices characterizing comprehensive programs that range from having management support, grassroots champions, and dedicated staff to meaningful incentives to a communication strategy consistent with corporate culture to regular evaluations using well-defined metrics.
- Call for further studies to differentiate the cost-effectiveness of program components (eg, biometric testing, exercise equipment, Web sites, healthy food subsidies, and incentives) and to determine which program components are most applicable to which populations.
- Cite specific companies with programs that have purportedly been shown to “work” and “not work,” along with lessons learned including the need to track multiple outcomes, employ timelines long enough to detect change (3–4-year window given), maintain transparency in methods, and capture programs with enough clarity to allow readers to assess their adequacy.
However, a program—more an approach—not mentioned in the article was that mounted by Navistar during 1999–2009, the subject of no less than 12 original/fast track papers9–20 in this journal since 2003. As reported correctly in the text of Bunn et al11 (the caption required an erratum21), Fig. 1 compares two trend lines of per capita annualized total direct health care costs during this period—one for Navistar employees/retirees and one for the national average of employees/retirees. The information presented in this figure, it should be emphasized, was the result of the highly scrutinized process required of corporations. The measures—unadjusted for inflation—were independently audited and confirmed by actuaries and presented to the company's Board of Directors annually.
As shown, during the first 4 years (1999–2002)—a window in which the initial and evolving Navistar approach focused largely on benefits only—the company's trend stayed an average of some four percentage points above national trend. In 2003, reflecting an expanded focus that by that point had begun to iteratively incorporate both aggressive disease management and health promotion (including lifestyle management and other) components, the company's increment relative to national trend dropped to one percentage point. In 2004, the company's trend dropped below national trend, and the resulting annual decrement continued to grow such that by 2009 it amounted to 23 percentage points below national trend.
The cost savings that can be inferred from these changes for a company with a covered employee/retiree population size of 65,000 are eye-opening. Although the exact amounts are not available, an estimated $250 million in total annualized direct health care costs were incurred during 1998. Using this estimate as the base and common starting point, the net sum of the annual decrements for Navistar relative to a comparison benchmark reflecting the national average from 1999 to 2009 totals to just more than $342 million when the percentage changes for the two groups are applied and the differences calculated by year (Table 1). A reasonable estimate for the company's expenditures on wellness components across this period above and beyond mechanisms that were already in place before the start of this period (eg, on-site exercise facilities) is $10 million. Under this assumption, the company's investment in wellness yielded an ROI of approximately 34.2 to 1 in total direct health care costs alone in 1999 to 2009. Note that, as documented in Allen et al,9 this ROI was achieved with minimal employee cost-shifting. Also, while the mix of employees to retirees fluctuated, the total number of covered lives in both categories stayed roughly comparable across the time frame.
The ramifications of these trends, however, extend well beyond what is conveyed in Fig. 1. As shown by the results for active employees in Fig. 2 (reported in Allen et al9), the reductions achieved on indirect cost measures from 2001 to 2009 actually exceeded those on direct cost measures (Note: The entries in Figures 2 and 3 are inflation-adjusted and expressed in 2009 dollars, with statistical controls for workforce changes over time in age, gender, percent salaried, and annualized earnings). These reductions not only encompassed drops in the utilization and costs associated with controllable (ie, sickness) absenteeism, presenteeism, workers' compensation, and short- and long-term disability that mirrored and further validated trends on the corresponding measures on the direct cost side; they also meant sharp gains in employee health and productivity as well as organizational profitability—that is, sustained value. Indeed, when the implications of these costs trends are taken into account for corporate contributions to meet requirements for reserve management and Financial Accounting Standard Board reporting (both factors whose specific mechanics are beyond the reach of this letter), it is reasonable to conclude that the Navistar approach toward wellness achieved a return easily in excess of 40 to 1 across the 1999 to 2009 period.
The approach that the company developed to achieve these results had features that can serve as guiding principles for other employers and their wellness approaches. These features spanned demand and supply side management, prevention, purchasing, and measurement—all developed, implemented, and periodically assessed on a worksite by worksite basis—in a framework that adapted the Deming “plan-do-study-act” cycle for quality improvement22 to the corporate setting.
Of note, while this process may in effect have tackled along the way many of the questions that Goetzel et al propose for the best (promising) practices, no comprehensive, systematic effort was undertaken to address the latter. Instead, given the prerequisites of the “real world” business setting at Navistar, the approach that evolved was guided by principles that have already been covered in papers cited above9–11,18–20 and can be summarized as follows:
- Evidence-driven: All decisions, from selecting and prioritizing targets to developing and implementing interventions to determining which components should be retained or refined versus dropped, were driven by the best available evidence accessed from sources both internal and external to the company. When evidence indicated problems, actions were taken.
- Total population: The full spectrum of health status and health risks in the covered population was kept in view. This meant bucketing individuals into groupings ranging from healthy to healthy at risk to those who had acute, chronic, and catastrophic conditions and monitoring the contributions of each to direct and indirect drivers and costs over time.
- Iterative, integrated development: As priorities were identified, an integrated mix of health promotion and lifestyle/disease management programs and benefits policies was iteratively enacted. The focus was threefold—primary prevention (keeping disease from developing; eg, weight control), secondary prevention (modifying risk factors to avert disease; eg, high blood pressure), and tertiary prevention (rehabilitation from a disease state; eg, cardiovascular event management).
- Relentless monitoring: As much attention was paid to analysis and reporting as optimizing the particulars of these programs. Monthly “in-house” dashboard reports on direct and indirect costs and drivers were provided to—and used to help assess the performance of—management and operations personnel. Periodically, supplemental studies by external researchers to provide more sophisticated analyses and strategic thinking were conducted, including those focusing on disease conditions (eg, autoimmune; chronic obstructive pulmonary disease; allergies; low back pain), the impact of overtime, and the validity of self-report and other methodological issues, which have appeared in this journal as cited above.
Implementation of these principles eventually led to demand-focused components that ranged from a well-staffed company health program on-site at all 26 company locations charged with managing an array of primary, secondary, and tertiary prevention initiatives to 16 disease management programs focusing on major disease categories, to on-site health clinics and individual case management support capabilities buttressed by nurse-based telephonic and e-mail support, to value-based benefit design with tiered-pricing for prescription drugs. Supply-focused components included periodic review of provider contracts and analyses of provider charges.
The monitoring of these components, it can be noted, did not prioritize tying observed changes in costs and utilization to specific programs and policies—a decision that ran counter to Goetzel et al's call for the need to “better understand which ... components are most cost-effective in providing value to workers, driving participation, and achieving specific outcomes.” Given the integrated nature of the approach, when reductions in costs were linked to a given disease, the likelihood was not only that the disease management program for that condition contributed, but also that multiple other initiatives (eg, medication formularies and self-care educational and health incentives) had an effect. In light of business realities, the proper evaluation elements needed to quantify the specific contributions of these components were simply not feasible.
But nor were they deemed necessary because it became increasingly clear that the integrated approach was achieving its objectives, not only among the sick but also across the entire health continuum. Figure 3 summarizes trends for selected groups in total direct and indirect costs from 2001 to 2009 already reported.9 Cost reductions occurred not only for those with hypertension, allergies, and back problems diagnoses, but also for healthy employees.
It remains a bit of mystery why the Navistar experience has not been brought up more for discussion by advocates for the effectiveness of corporate wellness programs. Whatever the case, notwithstanding the uncertainty of the methods used for the above ROI calculations, the body of work on Navistar's experience offers a well-documented referent point for the employment-based wellness debate. Not every wellness program implemented will by itself generate such savings—surely a prudent caveat for any such undertaking. But prudence in this instance does not necessitate precluding what has been shown to be possible. This debate is more fully joined when it incorporates a replicable roadmap for achieving the top end of what has been reported to date under peer review to work in an actual business setting. A concerted approach to wellness, amplified by a focus on workforce productivity and bolstered by a “stay the course” commitment on the part of senior management, can have a huge impact in reducing direct and indirect costs while also generating major improvements in employee health and sustained value. Indeed, these results make it clear that an integrated program focused on wellness, health and productivity can have an ROI that exceeds almost any other corporate investment or program goal. For employers that are challenged by intractable health care cost dynamics and have decided that they need to be proactive, the well-documented Navistar experience offers a viable benchmark for comparison.
The “best (and promising) practices” that Goetzel et al propose make an important contribution. These practices, for example, clearly have much potential when applied post hoc to figure out why programs are not producing intended results and what to do about it. But, what emerges as key from this case study are the real benefits from wellness approaches that may well require a longer timeframe to become quantifiable than Goetzel et al suggest—after all, it took into at least the fifth year before Navistar's investment first recouped a below–national trend estimate. To realize the savings that the company recouped beyond this point, a proactive, integrated, and sustained quality improvement–oriented approach may also be required—one that gives first priority not to questions trained on Donabedian's paradigm but rather to evolving iteratively to maximize the fit between population health characteristics that are observed and available resources that can be brought to bear to manage these characteristics. Being driven by evidence, oriented toward the total population, focused on primary, secondary, and tertiary prevention, and continuously vigilant with respect to total costs will increase the odds for success.
Harris Allen, PhD
From the Harris Allen Group, LLC Brookline, Mass.
The author gratefully acknowledges the input of William B. Bunn, MD, JD, MPH and Robin Bouvier when drafting this letter.
1. Goetzel R, Henke R, Tabrizi M, et al. Do workplace health promotion (wellness) programs work? J Occup Environ Med. 2014;56:927–934.
4. Caloyeras J, Liu H, Exum E, Broderick M, Mattke S. Managing manifest diseases, but not health risks, saved Pepisco money over seven years. Health Aff. 2014;33:124–131.
8. Goetzel R, Ozminkowsk R, Villagra A, Duffy J. Return on investment in disease management: a review. Healthcare Financ R. 2005;26:1–19.
9. Allen H, Rogers W, Bunn W, Pikelny D, Naim A. Reducing total health burden from 2001 to 2009: a major self-insured employer's counter-trend success story and implications for health care reform. J Occup Environ Med. 2012;54:904–916.
10. Allen H, Wright M, Craig T, et al. Tracking low back problems in a major self-insured workforce: toward improvement in the patient's journey. J Occup Environ Med. 2014;56:604–620.
11. Bunn W, Allen H, Stave G, Naim A. How to align evidence-based benefit design with the employer bottom-line: a case study. J Occup Environ Med. 2010;52:956–963.
12. Bunn W, Pikelny D, Paralkar S, Slavin T, Borden S IV, Allen HM Jr. The burden of allergies—and the capacity of medications to reduce this burden—in a heavy manufacturing environment. J Occup Environ Med. 2003;45:941–955.
13. Allen H, Bunn W. Using self-report and adverse event measures to track health's impact on productivity in known groups. J Occup Environ Med. 2003;45:973–983.
14. Allen H, Borden S, Pikelny D, Paralkar S, Slavin T, Bunn WB III. An intervention to promote appropriate management of allergies in a heavy manufacturing setting: an evaluation of health and productivity outcomes. J Occup Environ Med. 2003;45:956–983.
15. Allen H, Bunn W. Validating self-reported measures of productivity at work: a case for their credibility in a heavy manufacturing setting. J Occup Environ Med. 2003;45:926–940.
16. Allen H, Slavin T, Bunn W. Do long workhours impact health, safety and productivity at a heavy manufacturer? J Occup Environ Med. 2007;49:148–171.
17. Allen H, Woock C, Barrington L, Bunn W. Age, overtime and employee health, safety and productivity outcomes: a case study. J Occup Environ Med. 2008;50:873–894.
18. Bunn W, Stave G, Allen H, Naim A. Evidence-based benefit design: toward a sustainable healthcare future for employers. J Occup Environ Med. 2010;52:951–955.
19. Allen H, Bunn W, Naim A. The self-reported health and productivity burden of auto immune inflammatory disorders at a major self-insured employer. J Occup Environ Med. 2012;54:1049–1063.
20. Allen H, Rogers W, Bunn W. Managing the burden of COPD on workforce health and productivity: upping a leading employer's game. J Occup Environ Med. 2012;54:1064–1077.
21. Bunn W, Allen H, Stave G, Naim A. How to align evidence-based benefit design with the employer bottom-line: a case study—erratum. J Occup Environ Med. 2014;56:e142.
22. Moen R, Norman C. Circling back: clearing up myths about the Deming cycle and seeing how it keeps evolving. Available at: http://apiweb.org/circling-back.pdf
. Published 2010. Accessed September 24, 2014.