Early diagnosis of rheumatoid arthritis (RA) is an important challenge for clinical rheumatologists. This is because there is substantial evidence that early treatment with disease-modifying antirheumatic drugs leads to a better disease outcome. The 1987 American College of Rheumatology classification criteria for RA do not perform well as a diagnostic tool in early arthritis. Therefore, diagnostic studies are needed to develop diagnostic criteria or prediction models that enable clinicians to distinguish RA from other arthritides in an early phase of the disease. Diagnostic studies are hampered by the lack of an independent gold standard for RA. Since the most important clinical features of RA are the persistence of the arthritis and the development of erosions, arthritis outcome is a clinically relevant gold standard. Diagnosis is a phased, multivariable process in which the probability of the presence of disease is updated continuously when new diagnostic information is added to the patient profile. Therefore, besides univariate studies, multivariable diagnostic studies are needed to evaluate the value of current diagnostic practice or to evaluate the added value of a new diagnostic procedure. This chapter describes univariate and multivariate studies that have been performed in early arthritis populations to assess the diagnostic value of medical history, physical examination, laboratory tests and imaging in RA diagnosis.