Reducing Metabolic Syndrome Risk Using a Personalized Wellness Program : Journal of Occupational and Environmental Medicine

Secondary Logo

Journal Logo


Reducing Metabolic Syndrome Risk Using a Personalized Wellness Program

Steinberg, Gregory MB, BCh; Scott, Adam MBA; Honcz, Joseph MBA; Spettell, Claire PhD; Pradhan, Susil MS

Author Information
Journal of Occupational and Environmental Medicine: December 2015 - Volume 57 - Issue 12 - p 1269-1274
doi: 10.1097/JOM.0000000000000582
  • Open
  • CME Test


The prevalence of metabolic syndrome (Met S) is a costly health problem: an adult with Met S has annual health costs 1.6 times higher than average,1 and workplace participation and productivity are frequently impacted.2 Delaying, preventing, or reversing Met S through healthy lifestyle changes would therefore be expected to result in lower medical costs and reduced prevalence of Met S associated conditions such as hypertension and diabetes.

Met S refers to a constellation of five risk factors. These include out-of-range waist circumference [for women >35 inches (89 cm); men >40 inches (102 cm)], triglyceride levels above 150 mg/dL (1.7 mmol/L), low high-density lipoprotein (HDL) cholesterol levels [women ≤50 mg/dL (1.3 mmol/L); men ≤40 mg/dL (1.04 mmol/L)], high blood pressure [≥130/85 mm Hg (17.3/11.3 kPa)], and an elevated fasting blood sugar level [more than 100 mg/dL (5.5 mmol/L)]. An individual with at least three of these risk factors qualifies as having Met S. If left unchanged, Met S has been shown to increase the risk of diabetes, coronary heart disease, and overall death.3

The most recent data suggest that between 22.9%4 and 25%5 of US adults between the ages of 18 and 65 (44 to 48 million individuals) meet the criteria for Met S, with prevalence higher among females, and increasing significantly with age and body weight.6 An additional 104 million people have one or two out-of-range Met S risk factors. Previous research confirms that out-of-range waist circumference is the most important single factor in determining whether an individual will subsequently develop full Met S.7–9

Employers have invested heavily in wellness intervention programs, which vary significantly in design, engagement, and clinical and financial outcomes.10 Accordingly, it can be difficult to know what drives better clinical outcomes and cost savings, and there is an ongoing debate about whether lifestyle or disease management components of wellness programs generate more value.11

Aetna, in collaboration with a personalized wellness program vendor (Newtopia™), developed a new intervention called the Aetna Personalized Metabolic Syndrome Risk Reduction Program (hereinafter, the Program). The Program was targeted to Aetna employees with at least two out-of-range Met S risk factors, one of which had to be waist circumference. The Program used a high-touch approach to help employees achieve a healthier weight through an integrative and personalized focus on nutrition, exercise, and behavioral well-being. The Program includes voluntary limited genetic screening focused on three specific markers: FTO, MC4R, DRD2. Literature suggests12 that FTO, MC4R,13 and DRD214 influence how diet, exercise, and compulsive behavior impact body weight, body fat, and metabolism. The FTO gene has been linked to obesity and is expressed in adipose tissue and regions of the brain involved in the regulation of energy balance.15 The MC4R gene has been shown to regulate appetite and food intake by initiating satiety signals. Variations in the MC4R gene are associated with increased appetite and food intake.16,17 The DRD2 gene regulates dopamine, the primary chemical messenger of reward in the brain. It has been observed that deregulation of the DRD2 is proportional to higher body mass index.18 In addition to the personalization that the genetic screening provides, participants are grouped into one of eight groups, which represent the permutations of the three genes tested (with or without variation on each gene). Each grouping is assigned a specific starting target of daily dietary percentages of carbohydrate, protein, and fat as well as a target of weekly aerobic and anaerobic exercise. These targets serve as a basis for program coaches and client managers to further personalize behavioral reinforcement strategies to both tie back to targets and leverage genetic predispositions as mechanisms for engagement enhancement.

The primary goal of the study was to determine whether individuals invited to participate in the year-long Program would demonstrate reduced Met S risk factors and health care costs when compared with a control group not invited to the Program. A secondary objective was to determine whether providing individuals with personalized predictions of their risk of developing Met S in the next year would increase their likelihood of participating in the program.

Learning Objectives

  • Discuss the rationale for and components of the Aetna Personalized Metabolic Syndrome (MetS) Risk Reduction Program evaluated in this study.
  • Summarize the findings of the year-long pilot program, including employee engagement, clinical outcomes, and health care costs.
  • Discuss the implications for taking this approach to MetS risk reduction to scale.


Aetna employees who met the following criteria were recruited by employer e-mail to participate in the study. As reflected in Figure 1, eligible employees included Aetna employees who had previously participated in employer-sponsored Met S biometric screening and had two or more out-of-range risk factors, one of which had to be waist circumference. Individuals were excluded if currently enrolled in another Aetna wellness program or if the employee reported that they were enrolled in an external weight loss/wellness program such as Weight Watchers™. Employees also had to be over 18 years of age and could not be pregnant. Recruitment for the 12-month pilot occurred between June and September of 2013. The intervention occurred between July 2013 and June 2014.

Program sample development.

Study Design

As shown in Figure 1, the 2835 eligible employees were randomly assigned to one of three groups, two Program groups and one control, each with 945 eligible employees, stratified by gender, age group (18–29, 30–39, 40–49, 50–59, and 60–64) and number of Met S risk factors (two risks, three risks, or four to five risks). All 1890 individuals assigned to either of the two Program groups were invited, but due to resource constraints, the Program was artificially limited to the first 600 of these who agreed to register. Of these 600 Program registrants, 445 went on to complete the enrollment process.

  1. Program Group 1: Employees received baseline information about their last Met S results and were invited to participate in the Program.
  2. Program Group 2: Consistent with Program Group 1, Program Group 2 employees received baseline information about their last Met S results, and were invited to the Program. In addition, Program Group 2 employees received a specific 12-month prediction of their future Met S risk based upon the reverse engineering and forward simulation (or REFS) velocity-based predictive model. As discussed in a 2014 article, this model predicts the 12-month future probabilities of an individual developing Met S and each of the specific risk factors related to Met S.7
  3. Control group: Control group employees received the same baseline information about their last Met S results as Program Groups 1 and 2, but were not invited to participate in the Program.

Program Description

The Program was designed to be both highly personalized and high-touch. Employees were provided personal coaches and client care managers to achieve high levels of engagement and sustain behavioral changes. Employees interacted through various channels (eg phone, e-mail, Skype) with their coach and care manager. In addition, employees had access to an online portal and mobile application tailored to the employee's goals and which served as a platform to collect nutritional and activity data. The Program also incorporated a limited genetic profile that, when combined with a more typical psychosocial assessment, allowed for the development of a more personalized treatment plan.

All Program employees received a starter-kit in the mail, which contained a genetic screening kit and a wireless activity tracker. Individuals submitted a saliva sample that was tested for three genes—FTO, MC4R, and DRD2—associated with obesity, appetite, and compulsive behavior, respectively.19 On the basis of these results and an online assessment, individuals received a personalized nutrition and activity plan, and were assigned to a coach trained to work with their specific profile characteristics.

Individuals in Program Group 2 also received information that predicted their subsequent development of new Met S risk factors using the REFS velocity based predictive model previously described.7

Of note, program pricing was negotiated on a per-participant basis and not on an overall population basis (eg per member per month).

Outcome Measures

Outcome measures included program enrollment, engagement, clinical outcomes, and health care costs. Outcomes were measured for the 12-month period from July 2013 and June 2014.


Initial enrollment rates were compared between Program Group 1 and Program Group 2 to determine the impact of providing personalized risk predictions to Program Group 2. Enrollment was defined as employees agreeing to participate in the study and providing pertinent contact information to the Program vendor via the online registration process.


Engagement was measured each month and defined as a participant tracking their nutrition or physical activity (manually, electronically, or via activity tracker) for at least 12 days per month, and/or participating in at least one coaching or care manager session (telephonic, e-mail, or video).

Clinical Outcomes

Clinical outcomes related to Met S factors (waist circumference, triglycerides, HDL, blood pressure, and fasting blood sugar) were obtained from clinical laboratory results captured in Aetna's administrative systems from the contracted vendor who provided the annual biometric screenings. Individuals with two sets of measurements (pre and post the Program timeframe) were included in these analyses. Across the five Met S factors, the percentage of individuals who had two sets of measurements ranged between 70% and 78%. In addition, the Program vendor provided information for participants regarding reported weight loss during the study period.

Health Care Costs

Total medical costs were calculated on a per-employee per-month basis during the 12-month study period. Total medical costs were capped at the 98.5th percentile ($45,000 per employee per year) to minimize the impact of extreme outliers. Inpatient, outpatient, emergency room, and pharmacy costs were calculated on a per-employee per-month basis during the 12-month study period. For each of these specific cost categories, costs were also capped at the 98.5th percentile.

Statistical Analyses

Analyses were conducted at a number of levels depending on the outcome measure of interest. Enrollment and engagement rates were compared between Program Groups 1 and 2 for employees invited to participate in the Program. These comparisons were done both from an intent-to-treat perspective (invited employees vs controls) and from an as-treated perspective (participants vs controls).

Clinical outcome measures were compared for individuals in the Program and Control Groups who had biometric screening results from before and after the study. Cost measures were compared for all employees recruited to the study for the 12 months (July 2013 to June 2014). Program Groups 1 and 2 were combined for the clinical outcome and cost comparisons relative to the Control Group.

Z-tests of proportions were used to compare enrollment and engagement rates between Program groups. Chi-square tests were used for comparing discrete variables among groups. Two-tailed t-tests were conducted to assess differences in continuous variables between groups. The level of statistical significance for all comparisons was set at P value less than 0.05. We performed all analyses with SAS 9.4 software (SAS institute Inc., Cary, NC).


Baseline Characteristics

Table 1 summarizes that the three groups were similar to each other in demographic and geographic characteristics. In addition, they were similar to each other in overall comorbidity risk score, number of Met S risk factors, and prevalence of chronic conditions related to Met S identified through prior medical claims.

Comparison of Baseline Characteristics of Groups

Enrollment and Engagement Levels

Enrollment and engagement rates between Program Group 1 and Program Group 2 were compared to determine the effect of providing the individuals in Program Group 2 with a 12-month prediction of their future Met S risk. The hypothesis was that this additional information would increase both Enrollment and Engagement levels in Program Group 2.

As summarized in Table 2, enrollment was higher for Program Group 1 than for Program Group 2 who received personalized predictions of Met S risk (26% vs 21%, P = 0.03). Of those enrolled, the percentage who remained engaged throughout the study period was similar for the two Program Groups (50% for Program Group 1 vs 49% for Program Group 2, P = 0.73).

Enrollment and Engagement Rates of Program Groups 1 and 2

Of the total of 445 individuals who enrolled in the program (Program Group 1 and 2 combined), 221 or 50% demonstrated sustained engagement over the course of the Program.

Clinical Outcomes

Weight Loss

Weight loss was calculated from participant self-report. Of the 445 Program enrollees, 421 or 95% reported their pre and post-Program weight. Of these, 318 or 76% lost weight. The average per person pre-Program weight was 220 pounds (99.8 kg), and the average per person post-Program weight was 210 pounds (95.2 kg), an average loss of 10 pounds (4.5 kg), or 4.3% of the pre-Program value (P < 0.001).

Met S Risk Factors

Table 3 summarizes the changes in the five Met S factors from the start of the program. Employees invited to the Program demonstrated a trend for a greater reduction in the waist circumference compared to the Control group (−0.77 vs −0.48 inches, P = 0.06). From the as-treated perspective, the Program participants from the two groups combined showed significantly greater reduction in waist circumference relative to the Control Group (−1.06 inches vs −0.48 inches, P = 0.02). Improvements were also seen in triglyceride levels for those employees invited to the Program compared with the Control Group (−8.12 mg/dL vs −2.56 mg/dL, P = 0.05), and for Program participants compared with the Control Group (−18.47 mg/dL vs −2.64 mg/dL, P = 0.01). HDL levels also improved for the Invited and Participant groups compared with the Control group, although the difference was only statistically significant for the Participant group (2.81 mg/dL vs 1.44 mg/dL, P = 0.02).

Clinical Outcomes

Health Care Costs

As summarized in Table 4, from the intent-to-treat perspective, the Program groups had a trend for lower mean total medical costs compared to the Control group ($389 vs $434 PMPM, P < 0.07). Although component medical cost categories were also lower for the Program group than for the Control group, these differences did not reach statistical significance. From the as-treated perspective, the Program participants had significantly lower mean total medical costs versus the Control group ($312 PMPM vs $434 PMPM, P < 0.02). Significantly lower costs were also observed for each of the medical cost subcategories for Participants than for Controls.

Health Care Cost Outcomes


As the health of American workers declines, employers have invested in a variety of wellness programs to improve the health and productivity of employees, and reduce the associated health care costs. Wellness programs vary in design, ranging from generic, “one size fits all” programs purchased from outside vendors, to targeted, high-intensity interventional programs. Programs also vary by duration, and the type of incentives and rewards employees may gain. Consequently, there are limited high-quality clinical and financial data that allow comparison of wellness programs and their relative impact over similar timelines. Commentators20 have noted that the variation of outcomes used, lack of transparency as to research methods, and differing durations of reports on wellness programs make comparative evaluation difficult.

Nonetheless, recent meta-analyses concluded that wellness programs can generate a return on investment exceeding 3.25:1 after 3 years,21 and researchers are understandably interested in finding the factors that account for such savings. However, teams have reached different conclusions about what drives those savings: some have achieved positive results from lifestyle-focused programs, while others believe disease management programs drive savings.

The Aetna Personalized Metabolic Syndrome Risk Reduction Program pilot study was undertaken to add more rigor and transparency to the design and reporting of wellness program outcomes. This time-limited, lifestyle-focused wellness program was targeted to individuals with or at an increased risk for Met S. Met S and its component risk factors can often be precursors to developing significant medical conditions such as diabetes, coronary heart disease, and stroke.

The Program used a genetic screen focused on three specific markers—FTO, MC4R, and DRD2—to help individualize and personalize the design of the coaching program. We believe that this contributed to the high and sustained engagement rates of program enrollees.

The lifestyle changes adopted by employees (improved nutrition, and increased activity primarily) delivered clinical and economic impact in just 12 months. In our study, 95% of the 445 Program enrollees reported pre and post-Program weights, and of these 76% (318 of 421) lost weight, with an average weight loss of 10 pounds (4.5 kg) or 4.3% of their initial average weight (P < 0.001). Several Met S component risk factors also improved—waist circumference, triglycerides, and HDL. The improved clinical results were associated with reductions in total health care costs of $122 per participant per month, for a total savings of over $600,000 for those engaged in the program. Given that program fees were on a per-participant basis, this resulted in a positive net return on investment for the Program in its first year.

As noted earlier, previous wellness studies have taken several years to demonstrate benefit and it is possible that we would see even greater benefit in subsequent years of the Program.10,11,21 Preliminary data from year 2 of the Program show that participants from the study who were engaged in a lower intensity “maintenance” program sustain their weight loss during the second year.

Of interest was the lack of any significant positive effect on either enrollment or engagement when individuals were provided with specific evidence-based information about their future risk of Met S. We believe that this is consistent with other data that demonstrate that individuals often appear to be irrational decision-makers when presented with evidence-based information on the risks and safety of various consumer products such as cigarettes and alcohol.22 This issue is worthy of further study, as it has implications for future wellness program designs.

Our results are in marked contradistinction to several recent studies, including one published in January 2014 in Health Affairs,11 which questions the benefit, if any, of wellness programs, particularly in the absence of concomitant disease management. The obvious question is how one can reconcile these markedly different assessments of the value of wellness programs. To address this, we feel that it may be useful to compare our study in more detail with that of Caloyeras et al.11 They evaluated the cost impact of both lifestyle and disease management programs at PepsiCo, and concluded that, after 7 years, only the disease management program component was associated with lower costs.

There are several significant methodological differences between our study and that of Caloyeras et al.11 First, in their study, control patients were obtained via propensity matching from the pool of individuals who had elected not to participate in the wellness programs offered by PepsiCo, whereas they were randomized in our study. This might have introduced some degree of unknown bias. Second, unlike the present study, it does not appear that the Lifestyle component of the PepsiCo study was specifically targeted to higher risk individuals. In fact, the PepsiCo study authors noted a marked relative benefit for the combination of Disease Management along with Lifestyle versus Disease Management alone, and remarked that this “suggested that proper targeting can improve program performance.” Third, program participation definitions appear to be different and were more stringent in our study. Lastly, there were significant differences in underlying wellness program design and implementation between our study and the PepsiCo study.

In sum, we believe that these methodologic and programmatic differences are sufficient to explain the marked discrepancies noted between the results of the two studies. The implications are that for lifestyle wellness programs to be successful, they need to be targeted to appropriate higher risk individuals, and be well designed and implemented.

Study Limitations

The study has a number of limitations, but we believe none of them are material enough to detract from the overall positive results of the Program relative to engagement, clinical outcomes, and costs. Firstly, the study was limited to a single large employer. Secondly, as noted previously, due to resource constraints, study registration was artificially limited to the first 600 qualified individuals. Thirdly, the results relative to weight loss were based on self-reported data and are therefore subject to criticism. However, the fact that 95% of all 445 program enrollees reported their pre and post-Program weights mitigates the likelihood of significant reporting bias. Finally, it is not known whether any individuals in either the study or control groups were engaged in additional external programs or efforts that could impact their results and health profile. Measures of employee productivity were not examined in this study. This is an important area for future investigation.

The improved clinical outcomes and health cost reductions generated by the Program demonstrate that significant clinical and cost benefits can be derived from addressing Met S and its risk factors through appropriately designed wellness programs that focus on weight management. Such programs, if implemented at scale and maintained, would be expected to produce additional marked beneficial effects on downstream risk factors and events such as hypertension, diabetes, myocardial infarction, stroke, and heart failure, along with their associated costs.


Lifestyle-focused wellness programs can be effective vehicles for change to both improve the health of individuals and reduce health care costs. The Met S Engagement Program described here shows that a clinically targeted, personalized wellness program can result in significant improvement in engagement, clinical outcomes related to Met S risk, and costs within just 1 year.


The authors thank Jeff Ruby and the management team at Newtopia for its insights related to the program and for reporting program data. We are thankful for the insight and guidance from Michael Palmer, Chief Innovation and Digital Officer at Aetna, as well as the support of Jessica Jacobs and Kyle Niejadlik, both of the Aetna Innovation Labs. We are appreciative of the contributions of our colleagues Richard Kobylinski and Alan Borthwick, for study design, implementation, and outcome analyses. We are further thankful to our colleagues Nancy Lusignan, Jenifer Yakey-Ault, and Karen Ryan, each of whom helped coordinate the program with Aetna employees. We are grateful for support received from Accenture's Health & Public Services practice, specifically Brian Kalis and Kerry Vincent, who provided insights into organizations providing novel interventions for Met S risk reductions. Finally, we appreciate the editorial assistance of KC Spears of Compel Communications and Lana Berson of Aetna.


1. Boudreau DM, Malone DC, Raebel MA, et al.. Health care utilization and costs by metabolic syndrome risk factors. Metab Syndr Relat Disord 2009; 7:305–314.
2. Edington DW, Schulz AB. Metabolic syndrome in a workplace: prevalence, co-morbidities, and economic impact. Metab Syndr Relat Disord 2009; 7:459–468.
3. National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). Final report. Circulation 2002; 106:3143–3421.
4. Beltrán-Sánchez H, Harhay M, McElligott S. Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999–2010. J Am Coll Cardiol 2013; 62:697–703.
5. Fitch K, Pyenson B, Iwasaki K. Metabolic Syndrome and Employer Sponsored Medical Benefits: An Actuarial Analysis. Milliman. Available at: Published March 2006. Accessed September 14, 2015.
6. Ervin RB. Prevalence of Metabolic Syndrome among Adults 20 Years of Age and Over, by Sex, Age, Race and Ethnicity, and Body Mass Index: United States, 2003–2006. National Health Statistics Report, vol 13. CDC website. Available at: Accessed May 5, 2009.
7. Steinberg G, Church B, McCall C, et al. Novel predictive models for metabolic syndrome risk: a big data analytic approach. Am J Manag Care 2014; 20:e221–e228.
8. Janssen I, Katzmarzyk P, Ross R. Waist circumference and not body mass index explains obesity-related health risk. Am J Clin Nutr March 2004; 79:379–384.
9. Klein S, Allison D, Heymsfield S, et al.. Waist Circumference and Cardiometabolic Risk: A Consensus Statement from Shaping America's Health: Association for Weight Management and Obesity Prevention; NAASO, The Obesity Society; the American Society for Nutrition; and the American Diabetes Association. Available at: Accessed September 14, 2015.
10. Cyboran S, Paralkar S. Wellness Program ROI Depends on Design and Implementation. Alexandria, VA: Society for Human Resource Management; July 2013.
11. Caloyeras J, Liu H, Exum E, et al. Managing manifest diseases, but not health risks, saved PepsiCo money over seven years. Health Aff 2014; 33:124–131.
12. Thorleifsson G, Walters GB, Gudbjartsson DF, et al. Genome-wide association yields new sequence variants at seven loci that associate with measures of obesity. Nat Genet 2009; 41:18–24.
13. Qi L, Kraft P, Hunter D, et al. The common obesity variant near MC4R gene is associated with higher intakes of total energy and dietary fat, weight change and diabetes risk in women. Hum Mol Genet 2008; 17:3502–3508.
14. Johnson P, Kenny P. Dopamine D2 receptors in addiction-like reward dysfunction and compulsive eating in obese rats. Nature Neurosci 2010; 13:635–641.
15. Frayling T, Timpson NJ, Weedon MN, et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 2007; 316:889–894.
16. Haupt A, Thamer C, Heni M, et al. Impact of variation near MC4R on whole-body fat distribution, liver fat, and weight loss. Obesity (Silver Spring) 2009; 17:1942–1945.
17. Chambers JC, Elliott P, Zabaneh D, et al. Common genetic variation near MC4R is associated with waist circumference and insulin resistance. Nat Genet 2008; 40:716–718.
18. Wang GJ, Volkow ND, Logan J, et al. Brain dopamine and obesity. Lancet 2001; 357:354–357.
19. Blum K, et al. Increased prevalence of the Taq I A1 allele of the dopamine receptor gene (DRD2) in obesity with comorbid substance use disorder: a preliminary report. Pharmacogenetics 1996; 6:297–305.
20. Goetzel R, Mosher Henke R, Tabrizi M, et al. Do workplace health promotion (wellness) programs work? J Occup Environ Med 2014; 56:927–934.
21. Baicker K, Cutler D, Song Z. Workplace wellness programs can generate savings. Health Aff (Millwood) 2010; 29:304–311.
22. Thaler RH, Sustein RS. Nudge: Improving Decisions about Health, Wealth and Happiness. New Haven, CT: Yale University Press; 2008.

Aetna; biometric screening; blood pressure; DRD2; engagement; fasting blood sugar; FTO; genetic; high-density lipoprotein; inpatient; lifestyle; MC4R; medical costs; metabolic; metabolic syndrome; Newtopia; outpatient; personal; personalize; risk; triglyceride; waist; waist circumference; weight; weight loss; wellness

Copyright © 2015 by the American College of Occupational and Environmental Medicine