In Search of Financial Savings from Disease Management: Applying the Number-Needed-to-Decrease Analysis to a Diabetic Population

Background: In a previous article using population-level data, an a priori number-needed-to-decrease (NND) analysis was conducted to determine if there is potential opportunity in a given population for a disease management program to achieve financial effectiveness. Critics of that study have suggested that analysis at the entire population level does not account for differential enrollment trends. They also contend that reviewing disease-only hospitalization data disregards changes in acute utilization for comorbidities of the primary condition. This article responds to these two criticisms by critically examining the hypothesis that evaluating a specific diseased cohort elicits more reasonable projections of the financial effectiveness of a disease management program than when the analysis is conducted at the population level. To do this, this article reports the results of an a priori NND analysis of hospitalizations conducted on a diabetes mellitus cohort. Methods: An NND analysis was conducted on a diabetes cohort that was identified in a health plan population using Health Plan Employer Data and Information Set (HEDIS(R)) criteria. Hospitalizations were categorized in three groups: diabetes-only; diabetes plus comorbid conditions; and diabetes plus comorbid conditions and diagnoses possibly associated with diabetes. Results: To cover fees alone, it is estimated that a disease management program would have to reduce diabetes-only hospitalizations by 74%; hospitalizations for diabetes and comorbid conditions by 39%; or hospitalizations for diabetes, comorbid conditions, and diagnoses possibly associated with diabetes by 26%. Conclusions: The findings of the present study indicate that when performing the NND analysis at the cohort level as opposed to at the population level, even more stringent levels of performance are required to break even. Given that program fees is the only variable that can truly be manipulated a priori by the disease management program under the current model to improve the likelihood of achieving economic effectiveness, alternative approaches to this dilemma are discussed.