We propose that diagnostic tests should only be used when their results could affect treatment and prognosis of the diagnosed condition, taking costs into account. The cost-effectiveness of MRI has been evaluated in a randomised clinical trial in which 356 patients were divided over one group in which MRI was done within two weeks after experiencing trauma and one group that received standard care. No difference in QALYs was found between the groups. The number of referrals did not decrease in the MRI group, whereas costs for treatment increased. The point-of-care CRP-test for dealing with patients with acute cough presenting in daily general practice is one example of the opposite. CRP led to a decrease of antibiotic prescriptions and was found to be cost-effective. We conclude that more experts on diagnostics should be consulted in general practice and that diagnostic tests and procedures should receive attention in the medical curriculum and vocational training for general practice.