In low-cost health management (HM) interventions, mainly those that use computer-generated health risk feedback, online or web-based “self-service” coaching for all participants, it is generally the case that the more participants the better.  In such cases, costs to HM providers are fixed, and fees to their clients are usually set on a flat fee/per population basis.  So adding participants does not add to costs for clients, while adding to at least the potential for added benefits.

In such cases, the only basis for selecting HM participants is the match between a particular intervention and the targeted persons identified as potentially benefiting themselves and their employer or insurer by participating.  Many may benefit from participating in more than one HM intervention, since the average person typically has more than one health risk, chronic condition, or productivity/performance factor affecting each.

But when costs to HM providers and charges to clients vary significantly by the number of participants in a given intervention, careful selection of targets and of the most cost-effective/efficient ways to secure their enrollment, participation, cooperation and completion of the intervention can be essential.  If the intervention is the same across all participants, selection should focus on ensuring that as many of those targeted for participation as possible have predicted risk/reward levels greater then the costs they add to the intervention.

This is good practice for both HM providers and their clients.  If costs increase per participant, the best way to ensure that the overall ROI ratio and net economic gain or ROI amount of each intervention are acceptable or admirable is to ensure that as many participants as possible have individual ROI ratios and amounts that are at least positive.  This is true whether the type of intervention is the same for all participants or varies by their predicted risk/reward potential.

Individual customers of any business tend to include “most valuables” that generate great profit, “profitables” that at least contribute some profit, “marginals” that cover the fixed costs of serving them but no profit, and “losers” that detract from profits.  Individual employees can be arrayed across a similar scale of value to their employers.  And HM eligibles should be identifiable in terms of the same varying levels of risk/reward potential using predictive analytics.

When those eligible for particular HM interventions are identified through claims analysis or medical records, or through screening tests alone, the risk/reward potential is likely to be based solely on their past costs to insurers or employers.  But if health risk assessments (HRAs) are used, they can add far more information about individuals’ probability of responding well to HM interventions, plus their potential for improving their value to employers, when they are employees.

The key is to identify existing and validated HRAs that determine individuals’ attitudes and perceptions, as well as risks and past and current costs or productivity/performance impairment levels.  These may assessments should identify their levels of concern about and motivation to change their health behaviors and risks, or manage their chronic conditions and the behaviors that can accomplish that.  They may also identify where people are in stages of readiness to change, in internal vs. external locus of control perceptions, perceived barriers to change, etc. – whatever individual characteristics have been shown to affect their probability of responding well to HM interventions.

The combination of individuals’ past and present costs and impairment levels with their probability of changing these in the immediate future will determine their risk/reward potential.  This potential can then be used to:

1.    select targets for a one-size-fits-all intervention
2.    determine what sponsors or providers can afford to spend on  incentives and efforts to promote their participation
3.    graduate interventions to match the risk/reward potential of different cohorts or individuals among those eligible

Since the efforts expended by HM providers or sponsors, as well as any incentives used to entice their participation, cooperation and completion of interventions add to the costs of such interventions, these costs have to be factored into both selection and graduation processes, to ensure that the risk/reward potential of participants exceeds their predicted costs.  And when a combination of general health/wellness promotion, risk behavior, risk condition, and disease management interventions are planned, the potential per individual will vary widely across those eligible.

General health/wellness promotion tends to have the lowest potential in terms of healthcare/insurance costs, though it can reduce worker injuries, for example.  And it can include focused efforts to reduce productivity/performance impairment factors, such as lack of sleep, hydration at work, poor nutrition, etc. that may also add to healthcare costs.  And it can improve employee morale, energy levels and other health and motivation dimensions that may move participants into the range of “positive presenteeism”, where they go beyond expected and “normal” productivity/performance levels and value to employers.

One of the challenges in this type of intervention, aimed at preventing people from getting any risks, or increasing the numbers they already have, is that there is no change to use as the basis for predicting risk/return benefits.  Only the non-emergence of risks represents the source of benefit and ROI.  Fortunately, there is plenty of experience with populations and their tendency to add risks if left “unattended”.  The value of keeping people from moving to a higher risk category should always be added to the value of reducing them to a lower risk category.

For example, in one study, the proportion of people in a low-risk category (having 0-2 risks) who moved to the medium-risk (3-4 risks) level in the following year was 13.7%, while 2.2% moved to high risk (5+ risks).  The added cost of moving from low to medium risk was an increase of 26% in healthcare costs alone, while the added cost of moving to high risk was a 65% increase. Moving from medium to high risk meant an increase of 31%, adjusted for inflation in overall costs. [D. Eddington & S. Musich “The Case for Low-Risk Maintenance” Absolute Advantage 2:5 2003 22-25 (www.welcoa.com)

The benefit of reducing risks from a high to a low level was a 37% reduction in healthcare costs, while remaining at high risks meant a 21% increase. So the total benefit of reducing a person’s risk from high to low amounted to avoiding an 88% increase in costs altogether, counting inflation of costs.  The University of Michigan Health Management Research Center is an excellent source of information on the financial benefits of reducing health risks in employed populations. (www.hmrc.umich.edu)

This Center has also examined the productivity impact of risk reduction and increase.  In one of its studies, it found that the addition or reduction of a single risk was an average of 1.9% loss or gain in productivity due to presenteeism, or impairment among those at work. [W. Burton, et al. “The Association Between Health Risk Change and Presenteeism Change” JOEM 48:3 2006 252-263]  By multiplying such an effect by the average annual compensation of workers, roughly $50,000 nationally among employers who engage in health management, HM providers and clients can gauge the value, e.g. 1.9% of $50,000 = $950.

The value of disease management varies dramatically by the disease in question, in addition to across the individuals who have such a disease.  Congestive heart failure patients, for example, can often have their healthcare costs alone reduced by 30-50%, while diabetes patients may save only 5-10%.  People with depression rarely save on healthcare costs, but can save dramatically in reduced productivity/performance impairment, from 10-25% in many cases.  This kind of potential value can often cover DM costs of hundreds of dollars a month, since 25% of $50,000 = $12,500 a year.

Careful selection of individuals to target for HM interventions, together with prediction of the risk/reward potential of each can make a dramatic difference in the ROI ratios and amounts achieved by such interventions.  The prediction of such potential should be an inherent feature of HM strategies and programs, though it is most important when the costs of such programs vary significantly by the number of participants.  Such prediction also enables informed decisions on incentives and graduated interventions, to make ROI even more likely to be gained.