At least one systematic reviewer must be a disease expertBMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e8011 (Published 03 December 2012) Cite this as: BMJ 2012;345:e8011
- Andrew Whitelaw, professor of neonatal medicine1
- 1Department of Neonatal Neuroscience, University of Bristol, St Michael’s Hospital, Bristol BS2 8EG, UK
Gøtzsche and Ioannidis recommended that “teams performing systematic reviews should not include content area experts as authors” but must include experts in systematic review.1 Given the influence of systematic reviews on clinical guidelines and funding, patients need reviewers experienced with the condition and treatment. A general medical education may be adequate in common diseases. Expertise in systematic review methods is essential in diseases with large numbers of trials, but just having a medical degree, even with such expertise, is not adequate for rare serious conditions, for surgical treatments, and for complex interventions.
For example, reviewing therapeutic hypothermia for neonatal hypoxic-ischaemic brain injury requires knowledge of diagnosis of brain injury, cooling methods, and neurodevelopmental assessment of infants. If treatment aims to prevent “serious disability,” it is vital to understand the definition of serious disability in each trial before deciding whether outcomes can be meta-analysed together.
Reviewers of treatment of intraventricular haemorrhage need knowledge of neuroimaging, neurosurgical procedures, and neurodevelopmental outcomes.
Developmental care involves multiple interventions to provide stimulation, reduce stress, and help parental bonding. These interventions may have similar names but be different or have different names but be similar.
Experts in systematic reviewing cannot on their own apply review software to neonatal subjects and produce a review that helps clinicians.
The Cochrane Collaboration has greatly improved treatment decisions for millions of patients, and the founders were right in insisting that at least one reviewer must be an expert in the condition and treatment being reviewed.
Cite this as: BMJ 2012;345:e8011
Competing interests: AW has undergone Cochrane systemic review training and has published five Cochrane reviews.