Surrogate outcomes require adequate validation to prove their usefulness. However, health authorities accept variables with progressively lower standards. Accelerated approvals require confirmatory studies that are often not completed. HIV viral load or blood pressure are examples of adequately validated variables. In contrast, inconsistencies are identified in variables used in the approach to cancer, dementia, osteoporosis and infectious or cardiovascular diseases. Premature access to pharmacological therapies may result in a higher incidence of adverse reactions with little scope for implementing harm mitigation strategies. The abandonment of clinical trials in favour of predictive or observational designs compromises the quality of the evidence obtained.