Science-Based Outcomes

Savings Potential from Prevention and Risk Reduction for the Commercially Insured


/uploadedImages/Success/SavingsPotential.jpgHealth care costs present a substantial threat to the united States economy. U.S. health care spending has risen dramatically as a percentage of gross domestic product (GDP) over the past 30 years and is projected to rise 38% faster than GDP over the next decade, reaching 19.3% of GDP, or $13,387 per U.S. resident, by 2019. Three trends suggest that without significant changes to the demand for or supply of care, these rising costs will become unsustainable.

The Patient Protection and Affordable Care Act, signed into law in March 2010, has a goal of decreasing costs by improving the efficiency and quality of the delivery system. These supply-side approaches are needed, but they are not likely to reduce the rate of growth in spending; rather, these approaches can adjust the level of spending at a point in time. Currently, the uninsured population spends 62% less on health care than the insured population. Much of that difference is attributable to the lack of economic access, raising serious concerns about system capacity and total expenditures when these 41 million individuals secure coverage. For these reasons, the new law also introduces new demand-side strategies, notably including a requirement for the provision of health, wellness and prevention services for the entire insured population.

Ultimately, bending the cost curve will require a substantial decrease in the demand for care. Although the expansion of coverage may help, the underlying cost driver in the system is chronic disease. In 2005, 133 million Americans suffered from one or more chronic conditions, and the prevalence is increasing at an unprecedented rate. Health care for people with chronic diseases already accounts for 75% of the nation’s total health care costs, a number that will only increase without widespread and effective efforts to prevent and manage these conditions as a means to reduce demand and control costs. Strategies may include the expansion of programs that improve the health of the population through the promotion of self-efficacy, behavior change, and the ongoing management of health and disease. Such programs have already proven effective on a small scale. employers who have implemented wellness programs for employees have experienced short-term medical cost savings more than three times what they invested in the programs. However, less is known about the opportunity for long-term reductions in medical costs from health, wellness and prevention initiatives across the entire commercially insured population.

In the present study, which complements our recently reported savings impact of risk reduction in the Medicare population, we determined the average cost of commercially insured health care from birth until entry into Medicare based on levels of health risk. Using standard actuarial modeling, we evaluated the progression of risk and projected health care costs at each age for the cohort of U.S. residents who were under the age of one in 2008. Specific risk reduction scenarios were applied against the net present value of health insurance coverage from birth through age 64 to estimate the long-term savings potential of programs that reduce risk, or the rate of risk progression.


We developed a model to determine average allowed medical costs based on age and health risk using 2003 to 2007 Ingenix benchmark commercial health plan data for approximately 20 million individuals and Vital Statistics age/gender-specific mortality rates. Prescription cost data were not included. Full methods are detailed in Appendix A. In brief, the model was used to project the costs of a cohort of 4,257,000 individuals, the number of U.S. residents under age one in 2007.13 The cohort was stratified into three health status groups—Low Risk, Medium Risk and High Risk— using Resource Utilization Bands (RUBs),  an output from the Johns Hopkins Adjusted Clinical Groups (ACG) system,14, 15 as an approximation of health risk. RUBs represent simplified morbidity categories based on aggregations of ACGs.16 The model used numerical representations of assigned risk level, as well as survival, for all individuals in the data sample. Continuous coverage until age 64, or until death prior to that age, was assumed for all members. Tables based on individual age (“life tables”) were developed for each of the three health status groups. The tables calculate the overall weighted average per member per month (PMPM) and average lifetime costs from age 0 through 64 years. Both risk levels and costs change dynamically as the cohort ages, and these transitions are projected in the model based on the trends experienced in the benchmark data sample.

Risk-reduction scenarios were modeled to evaluate the impact on medical costs from age 0 to 64. The tested scenarios of decreased or delayed risk progression (see box) were chosen as estimates of risk reduction that might be achieved by effective health promotion initiatives, although the model can estimate cost changes based on any userchosen scenario. For each scenario, average PMPM, annual and age 0 to 64 total costs were calculated by the model. As opposed to PMPM costs, evaluation of annual and 0 to 64 total costs incorporated the death rate; however, the low death rate in this age group had minimal impact on costs. Total annual savings were calculated by multiplying annual per member savings by the size of the commercially insured population in 2008 (N = 201 million).17


Population Risk Distribution by Initial Risk Profile

The risk levels and costs for a cohort of over four million individuals were modeled over time, from age 0 to 64, beginning in 2008. The distribution of the population by gender into Low, Medium and High Risk groups shows that risk levels fluctuate until age 25, after which the cohort moves progressively into higher risk categories (Figure 1). Although a greater proportion of males are Low Risk in early adulthood, males show a more dramatic shift into higher risk groups as they age compared with females. Both genders have similar risk distributions at age 64, at which age just over 10% of the population remains at Low Risk status, with over 60% categorized as Medium Risk.

Age 0 to 64 Costs and Impact of Scenario Testing

The modeling of allowed medical costs with no changes to underlying risk levels or transition rates estimated that the average total cost of an insured person prior to Medicare eligibility at age 65 is $243,804 (2008 dollars). Costs for the two genders diverged after adolescence, with females being more expensive until a reversal occurs between age 58 and 59 (Figure 2A). An analysis of cost by risk level for both genders combined demonstrated that, as a general trend, the High and Medium Risk groups become more expensive with age, whereas the Low Risk group maintains extremely low costs at all ages (Figure 2B). It is expected that the Low Risk group maintains low costs because members who develop costly medical conditions transition into higher risk groups. It is noteworthy that the cost of the High Risk group increases steeply, surpassing an average cost of $2,000 per month after age 50. Increasing total expense is driven by two factors: actual costs of increasing utilization and the increase in the number of individuals in the High Risk group as the population ages.

To determine the savings impact of reductions in risk progression over time, we modeled the average PMPM savings that would accrue at each age if risk transitions were reduced by 10% and 25% from age 0  (Figure 3). The savings increase as individuals age, demonstrating the accumulating benefits of health promotion efforts that help to establish healthy behaviors early in life.

Table 1 summarizes the cost estimates for all tested riskreduction scenarios and the change in costs compared to baseline. Risk reduction scenarios applied in this model did not appreciably decrease the death rate (relatively low from age 0 to 64); therefore, percentage changes in PMPM and total costs for each scenario were similar. Each scenario, described in Methods, leads to a reduction in both PMPM and age 0 to 64 total costs. As would be expected, the more pronounced the reduction in risk status, the greater the cost savings. However, even relatively modest and therefore likely achievable reductions in the percentage of the population experiencing upward transitions would result in meaningful savings, as indicated by the 10% decrease scenarios.

Most scenarios were modeled from age 25, since risk trends from this age are more likely a representation of unhealthy behaviors and accumulating chronic disease; however, reduced risk levels from age 25 would be maximized by programs implemented earlier in life that establish healthy behavioral habits as a foundation for the remainder of life.18-20 Estimates of total annual savings are conservative since these values were calculated only for the number of commercially insured U.S. residents in 2008. Much greater savings could ultimately accrue with the expansion of coverage and broadly applied health promotion and management programs.


The success that employers and others have achieved in controlling costs through prevention and disease management programs provides a solid evidence base for expanding such programs to the larger  population.11, 21 Examples of such studies are listed in Appendix B. The work presented here demonstrates the opportunity for reducing the costs associated with commercially insured medical care by making modest changes to the increasing prevalence of health risk and chronic disease as a cohort of Americans age.

Inherent to the use of RUBs as an approximation of health risk, risk levels are affected not only by chronic disease but also by other causes of medical utilization. Since risk levels defined in this manner would be expected to have a weaker relationship to chronic disease at younger ages, we applied most risk-reduction scenarios beginning at age 25, when the steady risk progression that ensues is more likely a representation of declining health stemming from unhealthy behaviors and accumulating chronic disease. We emphasize, however, that achieving the levels of risk reduction modeled from age 25 (or even more aggressive estimates) would likely require that programs begin much earlier to establish the foundation for a lifetime of healthy behavior.18-20 Further, the importance of such programs is becoming increasingly important as children and youth are increasingly burdened with chronic conditions. The magnitude of this problem was revealed in a recent study that followed children age 2 through 8 for six years. The researchers compared a cohort beginning in 1988 to a cohort beginning in 2000 and found that the incidence of chronic disease increased by 9.7% for the first group and 20.4% for the second group.22

Continue reading and print entire article by clicking "Download/Print PDF" on right, top.

Download/Print PDFDownload/Print PDF


The Science of Well-Being

Potential Medicare Savings through Prevention and Risk Reduction

The Impact of Postdischarge Telephonic Follow-Up on Hospital Readmissions

The Impact of a Proactive Chronic Care Management Program on Hospital Admission Rates in a German Health Insurance Society

Impact of Predictive Model-Directed End-of-Life Counseling for Medicare Beneficiaries

Health care expenditures for people with diabetes mellitus, 1992

Follow on