Background: To investigate cost differences between intervention arms of a telephone-delivered self-management training for African Americans with diabetes.

Methods: This RCT used a 2x2 factorial design, randomizing African American adults with poorly controlled diabetes to: 1) knowledge only, 2) skills training only, 3) combined knowledge and skills training, or 4) control group receiving general health education. The primary outcome for cost effectiveness was glycemic control (HbA1c) at 12-months post-randomization. Costs for healthcare utilization were measured using Medical Expenditure Panel Survey (MEPS) estimates and costs for lost wages were estimated using interval regression. All costs were adjusted to 2012 dollars. Cost utility was quality-adjusted life years (QALYs) estimated via the area under the curve technique using utilities obtained from respondents’ answers to SF-12 questions and mortality estimates based on CDC lifetables. Empirical joint distribution of incremental costs, QALYs, and cost effectiveness acceptability curves were generated using non-parametric bootstrapping.

Results: There was no significant difference in cost (p=0.44) or change of QALYs between arms (p=0.36). With no statistically significant difference in cost, the incremental cost effectiveness ratios (ICER) were not significant compared to control, however, the knowledge only arm was $660 less expensive, the behavioral skills arm was $813 more expensive, and the combined arm was $3,630 less expensive than the control arm to achieve a 0.6% decrease in HbA1c.

Conclusions: Combined with the clinical findings, these results suggest telephone-delivered education and behavioral skills training is a cost-effective way to improve glycemic control. In particular, the combined knowledge and skills training intervention arm resulted in costs $3,600 lower than the control, and dropped HbA1c by 0.6%, achieving a clinically significant change for lower cost.


L.E. Egede: Research Support; Self; National Institute of Diabetes and Digestive and Kidney Diseases. Advisory Panel; Self; Novo Nordisk Inc.. C. Dismuke: None. C. Eiler: None. R.J. Walker: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at