Does the assessment of general movements without video observation reliably predict neurological outcome?