A dummy variable was introduced to take into account the change in the version of the ICD classification [11], so that we could estimate in the model the conversion factor resulting from this modification.
Convergence problems were probably a small sample size problem, due to the large numbers of parameters to estimate in the models.
The risk estimate in the model that did not include BMI as a covariate was essentially unchanged (RR 1.94, 95% CI 1.68 2.25).
The number of parameters to estimate in the model was limited by using linear reaction norms rather than more sophisticated functions such as polynomials.
16– 19 As a conservative estimate, in the model it has been assumed that all children with unresolved symptoms at 48 hours reconsult their primary care facility.
The MSE of the phase estimation is computed as an average over all MSE of the phase estimates in the model.
Reported accessibility is lower than that estimated in the model.
All estimates in the model were significant at p < 0.001, unless otherwise stated.
We initially investigated a flexible asymptote, with α estimated in the model fitting process.
Pattern coefficients constrained and not estimated in the model = .00 and are presented as blank.
