Effect of A Comprehensive, Employer-sponsored Health and Wellness Program on Medical Costs and Utilization

Objective: To examine the effect of a large company's health and wellness program - including disease management, case management, health risk assessment, lifestyle management, and nurse advice line - on medical costs and utilization. Data Sources: We analyzed administrative claims data of employees and family members spanning the years 2002 to 2007. Our analysis was based upon members >18 years, with two years of baseline data (2002 and 2003) and at least one year of data from the intervention period (2004 to 2007). The final analytical sample consisted of 67,895 members with 181,344 member years. Methods: An intent-to-treat approach was used to define program participation. We measured program effects using difference-in-difference approach based upon a linear mixed model with an individual level random effect, which included controls for health risks, baseline costs and utilization, calendar year, age, gender, and health conditions. Results: Participation in any component of the health and wellness program in the first intervention year led to a per member per month cost increase of $86 (p<0.01) during the same year, but it reduced the second and third year's cost by $92 (p<0.01) and $81 (p<0.01), respectively. Each program component showed similar patterns. Over all three years, the disease management component reduced costs by $206 (p<0.01). The health risk assessment and nurse advice line components were associated with an effect of -$37 (p<0.05) and $153 (p<0.05), respectively. Participation in any program component was associated with a decrease of 63 (p<0.01) ER visits and 34 (p<0.01) IP admissions per 1,000 person years over the study period, mainly due to the disease management component, while the nurse line component dramatically increased ER visits by 376 (p<0.01) per 1,000 person years. Conclusions and Implications: Participation in a comprehensive health and wellness program was associated with a cost and utilization increase in the first year then a decrease in the following years. Our results highlight the need for longitudinal data in order to fully assess the effects of disease management programs on population health.