In a recent search of PubMed I was surprised to see a chastening letter by Nakao and colleagues published in the Archives of Internal Medicine that criticised a 2010 meta-analysis by Sciarretta et al on antihypertensive treatment and development of heart failure in hypertension published in the same journal. Althought the paper in question searched PubMed and Embase as well as checking the references of a 2009 meta-analysis, it did not publish the search strategy. Even worse, it appears to have missed significant and recent studies (e.g. the CASE-J trial, the Kyoto Heart study and the HIJ-CREATE study) that it ought to have included.
While I haven’t investigated this for myself, it is interesting to see a complaint about search strategies not being published. This is a brick wall where there should be transparency. Perhaps researchers and publishers need to overcome their reluctance to print what may look like gibberish (diab$ adj3 oedema?.ti,ab anyone?) to some. Or perhaps many researchers still don’t even see a lack of explicit search strategy as a problem. A study selection algorithm is great, but it is no replacement for a proper foundation to this calibre of research.