Infant feeding practices and ulcerative colitis in childhood.

Abstract

TABLE A2 -Independent variables signtficantly associated with still-birth or neonatal death excludedfrom model because of covariance Height of mother Number of previous premature births Haemoglobin concentration Duration of first stage of labour Miaternal age Fetal distress (cord/prolapse/ meconium/fetal heart) Social and economic status Fever/tonsillitis/laryngitis Urinary tract infection Type of resuscitation Prescription of drugs to baby Lobelline Streptomycin included in the model that optimally estimates the log odds of stillbirth or neonatal death in the whole British perinatal mortality survey ( 1958) cohort. This list of variables includes the names of categorical variables together with the design variables that are generated (for example, "parity" is a categorical variable with four categories and hence three design variables labelled as "1", "2", "3"), as well as their coefficients. Categorical variables with two categories have been listed in table A2 as having only a single entry in the model. No interactions were included in the model owing to the computational problems that these would have generated for little improvement of the model.