The development and evaluation of new drugs often rely on surrogacy. An intermediate outcome becomes a surrogate outcome if it fulfils certain criteria, it should be easier to measure compared with the clinical outcome, a statistical relationship should exist between the clinical outcome and the surrogate outcome, a relation should exist allowing prediction of the degree of clinical effect based on the measured effect on the surrogate outcome. Development and authorization of drugs today often rely on so-called surrogate outcomes. Is this use sound? The validity of such outcomes has been reviewed in different therapeutic areas: hypertension, venous thromboembolism, AIDS, osteoporosis, hepatitis C. Based on this review, a pragmatic strategy is proposed which allows for the validation and proper use of surrogate outcomes.