Research concerning early arthritis and early rheumatoid arthritis (RA) may be considered to have begun with population-based studies in the United Kingdom, the United States and Scandinavia, from the late 1950s to the late 1960s. These studies indicated that the majority of people with clinical findings of RA had no evidence of disease 3-5 years later, and that only about 25% to 30% of people in a population who met the criteria for RA had rheumatoid factor. These findings may have contributed to an underestimation of RA until the severity of long-term outcomes of clinical RA were recognized in the 1980s on the basis of clinical cohorts. The first major early RA clinical cohort was established in 1957-1963 in Bath, England. Although results at 3 and even 11 years were not overly unfavorable, by 15 and 20 years most patients had severe outcomes of functional declines and premature mortality. The Middle-sex (UK) early RA cohort established in 1966-1971 indicated that radiographic abnormalities were observed in about 70% of patients by 2 years of disease, and were seen in most patients initially in the feet. The Memphis (Tennessee, USA) early RA cohort established in 1967-1971 suggested that a progressive course of RA is predicted by a higher number of involved joints at baseline. The Lund (Sweden) early RA cohort established in 1985-1989 indicated rather severe long-term outcomes in patients treated according to traditional conservative approaches to use of disease modifying anti-rheumatic drugs (DMARDs). The early RA study (ERAS) involving nine National Health Service trusts in the UK was established in 1987-93, and showed associations of education level and socioeconomic status with clinical status. The movement towards early arthritis clinics was given great impetus following the work by Emery in the early 1990s. These studies and others described elsewhere in this supplement have contributed to the foundations for the clinical approach to early arthritis in the 21st century.