Retrospective cohort design.
Tertiary-care US Veterans' Hospital, July 2007 to July 2012, using pre-existing database.
Prediabetic, adult veterans (n = 372), mostly men (94.4%, n = 351).
Visits with existing nutrition education classes were collected.
Primary outcome: diabetes status; predictors: visits/encounters, age, body mass index, weight change, and hemoglobin A1c.
Cox proportional hazards method, χ2 test, and logistic regression.
In this sample, prediabetic veterans who received nutrition education were less likely to develop diabetes when compared with prediabetic veterans who did not receive nutrition education (hazard ratio, 0.71; 95% confidence interval, 0.55–0.92; P < .01). This difference remained significant after adjusting for body mass index and weight change.
Nutrition education was significantly associated with preventing the progression from prediabetes to diabetes in US Veterans participating in a nutrition education intervention at the Michael E. DeBakey Veterans Affairs Medical Center.