Volume 47 | Number 6 | December 2012

Abstract List

Todd P. Gilmer Ph.D., Patrick J. O'Connor M.D., M.P.H., JoAnn M. Sperl‐Hillen M.D., William A. Rush Ph.D., Paul E. Johnson Ph.D., Gerald H. Amundson B.S., Stephen E. Asche, Heidi L. Ekstrom M.A.

Background and Objective

Medical groups have invested billions of dollars in electronic medical records (s), but few studies have examined the cost‐effectiveness of ‐based clinical decision support (). This study examined the cost‐effectiveness of ‐based for adults with diabetes from the perspective of the health care system.

Data Sources/Setting

Clinical outcome and cost data from a randomized clinical trial of ‐based were used as inputs into a diabetes simulation model. The simulation cohort included 1,092 patients with diabetes with A1c above goal at baseline.

Study Design

The nited ingdom Prospective Diabetes Study Outcomes Model, a validated simulation model of diabetes, was used to evaluate remaining life years, quality‐adjusted life years (s), and health care costs over patient lifetimes (40‐year time horizon) from the health system perspective.

Principal Findings

Patients in the intervention group had significantly lowered A1c (0.26 percent,  = .014) relative to patients in the control arm. Intervention costs were $120 ( = 45) per patient in the first year and $76 ( = 45) per patient in the following years. In the base case analysis, ‐based increased lifetime s by 0.04 ( = 0.01) and increased lifetime costs by $112 ( = 660), resulting in an incremental cost‐effectiveness ratio of $3,017 per . The cost‐effectiveness of ‐based persisted in one‐way, two‐way, and probabilistic sensitivity analyses.


Widespread adoption of sophisticated ‐based has the potential to modestly improve the quality of care for patients with chronic conditions without substantially increasing costs to the health care system.