Skip to main content
. 2020 Nov 5;22(11):e20322. doi: 10.2196/20322

Table 3.

Effectiveness of the intervention on coprimary outcomes.

Coprimary outcomes and analysis models and Arms Between arm differences at 6 months (95% CI) P value Between arm differences at 12 months (95% CI) P value Estimated mean changes between baseline and 6 months (95% CI)a P value Estimated mean changes between baseline and 12 months (95% CI)a P value
Glycated hemoglobin (%)

Primary (Intention-to-treat) modelb


Intervention 0.06 (−0.35 to 0.47) .78 −0.04 (−0.45 to 0.36) .84 −0.20 (−0.49 to 0.09) .17 −0.33 (−0.62 to −0.04) .03


Control Reference N/Ac Reference N/A −0.26 (−0.55 to 0.03) .08 −0.28 (−0.57 to 0.00) .05

Adjusted modeld


Intervention 0.06 (−0.35 to 0.46) .79 −0.04 (−0.44 to 0.37) .87 −0.20 (−0.49 to 0.09) .18 −0.32 (−0.61 to −0.03) .03


Control Reference N/A Reference N/A −0.25 (−0.54 to 0.03) .09 −0.28 (−0.57 to 0.00) .05

Per-protocol modele


Intervention −0.05 (−0.47 to 0.37) .81 −0.14 (−0.56 to 0.28) .52 −0.26 (−0.59 to 0.07) .12 −0.40 (−0.73 to −0.06) .02


Control Reference Reference −0.21 (−0.47 to 0.05) .11 −0.26 (−0.51 to −0.00) .05
Health Related Quality of Life: Assessment of Quality of Life-8D utility score

Primary (Intention-to-treat) modelb


Intervention 0.05 (0.01 to 0.08) .006 0.04 (0.00 to 0.07) .039 0.04 (0.01 to 0.07) .002 0.04 (0.01 to 0.06) .007


Control Reference N/A Reference N/A −0.01 (−0.03 to 0.02) .48 0.00 (−0.03 to 0.02) .92

Adjusted modeld


Intervention 0.05 (0.01 to 0.08) .005 0.03 (0.00 to 0.07) .047 0.04 (0.01 to 0.06) .002 0.03 (0.01 to 0.06) .009


Control Reference N/A Reference N/A −0.01 (−0.03 to 0.01) .46 −0.00 (−0.02 to 0.02) .93

Per-protocol modele


Intervention 0.06 (0.02 to 0.09) .002 0.06 (0.02 to 0.09) .003 0.05 (0.02 to 0.08) .001 0.05 (0.02 to 0.08) .001


Control Reference N/A Reference N/A −0.01 (−0.03 to 0.01) .49 −0.01 (−0.03 to 0.02) .63

aMean changes in outcomes were estimated based on the linear mixed-effect regression model.

bFor HbA1c, the intraclass correlation coefficient (ICC) for the primary model was 0.551 (95% CI 0.465-0.634). For HRQoL, the ICC for the unadjusted model was 0.847 (95% CI 0.806-0.880). Number of participants with valid data at each time point: n for HbA1c (intervention vs control): 93 vs 94 at baseline, 78 vs 78 at 6 months, and 77 vs 79 at 12 months. Number of participants at each time point for HRQoL (intervention vs control): 93 vs 94 at baseline, 67 vs 77 at 6 months, and 60 vs 78 at 12 months.

cN/A: not applicable.

dThe adjusted model adjusted baseline values of variables that were either imbalanced by intervention allocation by chance (baseline age and depression score) or associated with loss to follow-up (baseline AQoL-8D utility value and HADS Anxiety score).

eThe per-protocol analysis considered participants who had completed more than 6 chats with Laura as following the study protocol.