A frequent problem in estimating logistic regression models is a failure of the likelihood maximization algorithm to converge. In most cases, this failure is a consequence of data patterns known as complete or quasi-complete separation. For these patterns, the maximum likelihood estimates simply do not exist. In this paper, I examine how and why complete or quasi-complete separation occur, and the effects they produce in output from SAS ® procedures.I then describe and evaluate several possible solutions.

No abstract

Shaken baby syndrome is evaluated in the context of its historical evolution and its veracity in referring to causal injury mechanisms. A rational assessment of the injury causation and consequent pathological states associated with the syndrome is presented. It is now evident that shaken baby syndrome evolved as a result of a faulty application of scientific reasoning and a lack of appreciation of mechanisms of injury. A brief explanation of the commonly understood usage and interface of scientific methodology and reasoning as applied to clinical medicine is given.

Subdural bleeding in the so-called``shaken baby syndrome'' is recognized as a hallmark of this syndrome, and is often noted as chronic in nature, indicating an earlier time of origin than clinical presentation. In infants and neonates, the timeframe for generating such chronic intracranial bleeding is therefore limited. Neurosurgical, obstetric, and pediatric literature all recognize the significance of birth trauma in the generation of intracranial bleeding. This possibility is explored further here, with emphasis on features peculiar to Homo sapiens predisposing to intracranial bleeding during this timeframe. Encephalization and bipedalism combine to render the infant and mother susceptible to injury at birth.

No abstract

No abstract

No abstract