This review examines some of the current perspectives in evidence-based laboratory medicine. From evidence-based medicine, which has concentrated mainly on treatments, we have learned that systematic reviews of the literature are the major way of producing high-quality evidence. This is because systematic reviews search all the literature, apply quality standards to ensure that only the best evidence is available, and aggregate a sufficient weight of information to inform on both the direction of a result (that a treatment works) and the magnitude of the result (how good the treatment is). In laboratory medicine, we have few systematic reviews. Even those we do have give little insight into features of studies which provide the closest view of the 'truth'. It may well be that current studies of laboratory tests are just not good enough. Often, they are too limited in scope, too small, and choose inappropriate patient populations. In the future, comprehensive, prospective and large studies should examine patient demographics and clinical history, as well as laboratory findings. This will better predict those features that are most closely associated with the correct diagnosis in a particular clinical situation. Clinical decision rules combining all available data will be the best way forward.