Minerva: February 2001
Researchers who conduct meta analyses may have to look again at the funnel plot as a way of checking for selection bias. A study of 198 meta-analyses suggests that the shape of funnel plots depends as much on how they are calculated as on the presence or absence of biases (Journal of Clinical Epidemiology 2000;53:477–84). Researchers man. aged to convert 37 asymmetrical plots (indicating bias) into symmetrical ones just by tweaking the calculations.
Two randomised trials in Emergency Medicine Journal, the new incarnation of Accident and Emergency Medicine, compare midazolam and ketamine for sedating children needing emergency procedures. The first concludes that oral ketamine works better than oral midazolam for children with wounds that need suturing under local anaesthetic (2001;18:30–3). The second finds intranasal midazolam inferior to a combination of intravenous ketamine and midazolam (2001;18:39–45). Both teams of authors say that careful monitoring of sedated children is essential, including pulse