why is it inappropriate to rely entirely on study abstracts when making clinical decisions?

Clinical decisions about what?