I would mistrust individual studies for all the reasons you mention, but these are also the same reasons that meta-analyses exist, and they will look for those things and more. Generally, they will try to find and correct for systematic biases, and (although I haven't read this one) they will often completely discard other studies if their results are useless.