QUALITY AND RETRIEVAL OF OBSTETRICAL-ANESTHESIA RANDOMIZED CONTROLLEDTRIALS

Purpose: Randomized controlled trials (RCTs) are suitable for meta-ana lysis and systematic reviews provided they are of high quality and are easy to retrieve. We determined these attributes of RCTs in obstetric al anaesthesia in a sample of available journals that are indexed in M EDLINE. Source: Randomized controlled trials published between January 1985 and December 1994 in seven anaesthetic and three obstetric journ als were identified by a MEDLINE search, and by handsearch of the same journals. Each RCT was assigned a quality score by a blinded rater us ing a reliable and validated scale. The quality of each RCT was descri bed and compared over time and by journal. The comparative yield of ME DLINE and handsearch was evaluated. Principal findings: Three hundred and forty RCTs were retrieved by MEDLINE and handsearch. Two hundred a nd twenty seven (65%) were identified by the MEDLINE search and 333 by the handsearch (98%). The median quality score was 3/5. There was no difference in score over time. Anesthesiology had the highest median s core, Anaesthesia had the lowest (P < 0.05). Conclusions: Care must be taken when reviewing obstetrical anaesthesia research. Strategies in addition to a MEDLINE search must be used to identify RCTs since more than one third were missed using MEDLINE alone. Poor quality RCTs are more likely to be biased in favour of a new treatment. Therefore, to i ncrease the validity of reviews sensitivity analyses based on quality should be done.