Number needed to succeed in disease management.

THE IDEA IS BOTH SIMPLE and logical, but there has been a good deal of controversy over the use of “Number Needed to Decrease” (NND) analysis (which here I label as “Number Needed to Succeed” [NNS] analysis) as a tool in considering proposed disease management (DM) investments. This analysis method first considers the costs of DM interventions compared to the cost savings potential for each one, then determines how many “cases” of cost-generating episodes of care (eg, inpatient admissions, emergency room [ER] visits) would have to be eliminated from future utilization in order to cover the costs of DM (“breakeven”) or yield a desired return on investment (ROI) ratio and net savings. In his first published article on this model, its creator described the NND approach and illustrated its use relative to 4 different chronic diseases: asthma, coronary heart disease, diabetes, and congestive heart failure.1 He based his analysis of savings on hospital admissions and ER visits and concluded that because of the low rates of these events among patients with these conditions, significant reductions would be needed to generate significant savings. He based his cost of the DM intervention on fees per plan member, not per DM prospect or participant, and because of the low incidence of these conditions in most populations, this created a high cost per person with the conditions. [Note: While charging a fee for every member of a population makes costs more predictable for payors and vendors, it is likely to increase the costs compared to the benefits of DM interventions. If vendors charge $2.00 per member per month for a given DM program, while only 2% of the members have the disease it addresses, then the cost per individual who has any likelihood of having reduced health care costs is $2.00 divided by 2% or $100 per person affected per month. This makes it necessary that each person affected generate savings of at least $1200 per year just to break even. If only half of those affected actually participate in the DM program, each participant will have to generate savings of $2400 per year to break even. By contrast, if vendors were to charge based on the number of people in the population with a given disease, and charge enough to cover the costs to the vendor of that number of participants, even if that meant raising fees to $200 per year per participant, each participant would have to generate only $200 to break even, or $400 to achieve an ROI of $2.00:1. Even if vendors charged per person affected, and only half those affected participated, this would still require savings of only $400 per participant to break even and $800 each to achieve a $2.00:1 ROI.] The combination of the limited cost focus and high imputed costs per person affected yielded projected requirements for between 10% and 20% decreases in utilization of hospi-