The effectiveness results were as follows:
After three months, the mean (SD) HbA1c level was 8.80 (0.16%) for the inpatient group and 9.56 (0.20%) for the outpatient group.
Covariance analysis showed equivalent glycemic control at 3 months in both groups (adjusted means (SD) with respect to the inclusion values: 9.00 (1.14%) %) for the inpatient group versus 9.37 (1.14%) for the outpatient group; equivalence hypothesis: p=0.01).
A low and similar incidence of episodes of hypoglycaemia or hyperglycaemia with ketonuria was observed (67.9% for the inpatient group versus 57.4% for the outpatient group (p=0.2) for proportion of patients with one or more episodes of hypoglycaemia and 0% for the inpatient group versus 3.7% for the outpatient group (p=0.24) for hyperglycaemia with ketonuria).
Average weight gain was similar in both groups (2.8 kg).
Adverse events, probably or possibly attributable to the treatment, were essentially local reactions at the injection site, which were evenly distributed between both groups.
Analysis of the number of patients who reached at least a 15% reduction of their HbA1c 3 months after starting insulin showed no difference between the groups (21 in each study group), and the number of hypoglycaemic episodes was comparable in these subgroups.