Rapid adoption of disease management has outpaced systematic evaluation of its net value in improving health outcomes and mitigating healthcare cost. This article identifies areas in which outcomes research in disease management is needed to demonstrate its value or to enhance its performance. Patient identification for disease management relies on administrative database queries but the trade-offs in sensitivity, specificity, and predictive value of alternative queries are not well known. Large-scale deployment, rapid patient engagement, and repeated interactions between patients and nurses could be important attributes for attaining measurable improvements in quality and cost reduction over short periods of time, but these hypothesis need to be tested. There is a trend toward integration of multiple chronic disease management programs onto a single platform. To support this trend, there is a need for a corresponding set of integrated clinical guidelines or "meta-guidelines" that combine the contents of individual practice guidelines. The relative contribution of various disease management interventions in improving clinical results, lowering costs, and their respective ease of implementation is not known. Research leading to a better understanding of tradeoffs could lead to more rational resource allocation and better overall outcomes. Coordination between disease management programs and physician practices is lacking. Research aimed at defining operational and technical interfaces and cultural and behavioral professional adjustments necessary to achieve integration and coordination is needed. The lack of a consistent analytical framework for evaluating clinical and financial outcomes has made comparisons of reported results impossible and has rendered many reports unreliable. Theoretical work on a standard methodology that integrates clinical and financial outcomes and empiric validation is needed.