Table 3. Results of the intention-to-treat analysis for general and behavioural development.
Control rate | Insulin effect | Total rate | Control rate | Insulin effect | ||||||
---|---|---|---|---|---|---|---|---|---|---|
All ages | All ages | All ages | Age >36 months | Age >36 months | ||||||
C | P-value | C | P-value | C | 95% CI | C | P-value | C | P-value | |
B: cognition | 0.40 | 0.30 | 0.57 | 0.11 | 0.97 | (0.36; 1.58) | 0.19 | 0.64 | 0.75 | 0.046 |
B: receptive language | 1.02 | 0.02 | 0.63 | 0.13 | 1.64 | (1.04; 2.25) | 0.87 | 0.05 | 0.60 | 0.14 |
B: expressive language | 1.35 | 0.002 | 0.05 | 0.92 | 1.39 | (0.85; 1.94) | 1.09 | 0.02 | 0.29 | 0.54 |
B: fine motor function | 0.47 | 0.18 | 0.41 | 0.26 | 0.87 | (0.38; 1.37) | 0.30 | 0.44 | 0.56 | 0.15 |
B: gross motor functiona | 1.82 | 0.01 | −0.07 | 0.92 | 1.75 | (1.12; 2.38) | 1.82 | 0.004 | −0.01 | 0.99 |
V: communication skillsb | 0.11 | 0.80 | 0.57 | 0.19 | 0.68 | (0.004; 1.35) | 0.14 | 0.76 | 0.56 | 0.32 |
V: social skills | −0.12 | 0.88 | 1.41 | 0.06 | 1.29 | (0.11; 2.47) | −0.45 | 0.57 | 1.54 | 0.042 |
V: daily skills | 0.14 | 0.78 | 0.84 | 0.14 | 0.97 | (0.33; 1.61) | 0.04 | 0.94 | 1.10 | 0.07 |
V: motor skills | 0.30 | 0.64 | 1.21 | 0.09 | 1.51 | (0.62; 2.40) | 0.07 | 0.91 | 1.28 | 0.10 |
E: social developmenta | 2.39 | 0.13 | 0.56 | 0.76 | 2.96 | (1.08; 4.83) | 2.31 | 0.18 | 0.93 | 0.64 |
E: emotional developmenta | 0.50 | 0.61 | 0.07 | 0.95 | 0.57 | (−0.59; 1.74) | 0.47 | 0.66 | 0.05 | 0.97 |
Abbreviations: B, Bayley-III-NL; V, Vineland Screener 0-6 and E=ESSEON.
The random coefficients model was reduced to a random intercept model since the variability in the slope was estimated to be zero for ‘all ages' and ‘age >36 months' and
for ‘age >36 months' only. All ages n=25, Age >36 months n=21. The coefficient (C) is an estimate of the change in DAE in months per 6 months for the pre-treatment period (control rate) and treatment period (insulin effect). Observed insulin effects that show a clinical increase in DAE are underlined. The P-value (P) is calculated for the difference between the DAE at the beginning and the end of these respective periods. P<0.05 is considered statistically significant. Significant P-values of insulin effects are bold. The confidence interval (CI) represents the distribution of the individual coefficients.