Multiplicity and sparse data in systematic reviews of anaesthesiological interventions cause increased risk of random errors and lack of reliability of conclusions