2045. Baseline Characteristics And Randomization: The VA Dental Diabetes Study
Judith A Jones, DDS,MPH,DScD, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, RI Garcia, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, CJ Wehler, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, SE Rich, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, DR Miller, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, J Rothendler, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health, C Christiansen, CHQOER, Bedford VAMC and VA Boston Healthcare System and Boston University Schools of Dental Medicine and Public Health

Objectives: In randomized trials, it is critical to examine differences in characteristics: a) between participants and the sample frame and b) between treatment arms. We assessed differences in our ongoing trial of the efficacy of periodontal therapy on glycemic control.

Methods: We used administrative, (baseline HbA1c, comorbid conditions, age), survey (SES, dental care use, self-reported health, stress, activity, diabetes duration, BMI, smoking and alcohol use) and clinical exam data (# of teeth, CPITN, exudate, Gingival Index, periodontal probing, recession). We compared means (continuous variables) and frequencies (categorical variables) using t-tests and chi square tests (a=0.05).

Results: Results: Compared to those not randomized (n=2534), the 193 randomized participants were younger (58 vs. 64 years), had slightly higher HbA1c (10.2 vs. 9.8%) and were more likely to be depressed, drug dependent, and obese. Group A participants (early treatment) were more likely than group B (deferred treatment) to have a history of kidney disease (7% vs 1%) and transient ischemic attacks (5.3% vs. 0%), higher self-reported stress (mean=5.8 {out of 10} vs. 4.9), and were less likely to be male (95% vs. 100%), current smokers (20% vs. 32%) or former smokers (50% vs. 55%). The 63 other parameters showed no significant differences between groups and no differences are significant at a=0.01.

Conclusions: Randomization was successful in this study. Adjustment of some covariates (stress, smoking, gender) may be important in the analyses of efficacy.

Impact: The results of the VA DDS will inform VA about the efficacy of periodontal therapy on control of HbA1c.