Table 3

Longitudinal analysis: estimated means, regression coefficients, and 95% CIs for study outcomes by study arm and time*

Outcome variables by study armEstimate mean (95% CI)Between-arm difference (arm × time interaction)
BaselineImmediately after6 Months12 MonthsBaselineP valueImmediately afterP value6 MonthsP value12 MonthsP value
Primary outcome
 HbA1c (%)
  Standard DSME10.46 (10.07, 10.86)9.91 (9.50, 10.32)10.26 (9.84, 10.68)10.38 (9.96, 10.81)0.02 (−0.55, 0.58)0.955−0.61 (−1.19, −0.03)0.038−0.45 (−1.04, 0.15)0.139−0.77 (−1.38, −0.17)0.013
  Adapted DSME10.48 (10.08, 10.89)9.30 (8.90, 9.70)9.81 (9.40, 10.23)9.61 (9.17, 10.04)
 HbA1c (mmol/mol)
  Standard DSME90.8 (86.6, 95.2)84.8 (80.3, 89.3)88.6 (84.1, 93.2)90.0 (85.4, 94.7)0.2 (−6.0, 6.3)0.955−6.7 (−13.0, −0.3)0.038−4.9 (−11.4, 1.6)0.139−8.4 (−15.1, −1.9)0.013
  Adapted DSME91.1 (86.7, 95.5)78.2 (73.8, 82.5)83.7 (79.2, 88.3)81.5 (76.7, 86.2)
Secondary outcomes
 BMI (kg/m2)
  Standard DSME31.46 (30.50, 32.42)31.49 (30.53, 32.46)31.39 (30.42, 32.36)31.32 (30.35, 32.29)−0.88 (−2.26, 0.49)0.205−0.84 (−2.22, 0.54)0.234−0.42 (−1.81, 0.97)0.552−0.54 (−1.93, 0.85)0.447
  Adapted DSME30.58 (29.60, 31.55)30.66 (29.68, 31.63)30.97 (29.99, 31.95)30.78 (29.80, 31.77)
 Total cholesterol (mg/dL)
  Standard DSME189.08 (181.97, 196.19)182.64 (175.35, 189.94)175.84 (168.30, 183.37)180.79 (173.24, 188.34)−7.69 (−17.81, 2.43)0.136−12.50 (−22.82, −2.18)0.0192.85 (−7.77, 13.48)0.598−9.91 (−20.76, 0.94)0.073
  Adapted DSME181.39 (174.25, 188.52)170.15 (162.92, 177.37)178.69 (171.27, 186.12)170.88 (163.15, 178.61)
 HDL (mg/dL)
  Standard DSME36.99 (35.04, 38.93)35.63 (33.63, 37.63)35.49 (33.46, 37.52)35.58 (33.53, 37.62)−0.56 (−3.33, 2.21)0.6911.90 (−0.92, 4.72)0.1863.85 (0.97, 6.72)0.0091.92 (−1.02, 4.87)0.201
  Adapted DSME36.43 (34.47, 38.39)37.53 (35.56, 39.50)39.34 (37.31, 41.36)37.50 (35.39, 39.60)
  • Statistically significant P values are in boldface type.

  • *Estimates from linear mixed-effects regression models are adjusted for baseline data: sex, age, education, marital status, employment status, use of diabetes medication, and households containing multiple participants.

  • †The estimated arm contrast of the changes from baseline corresponds to the estimates of the arm-by-time interaction.