The use of hospital care was assessed in a 13-year follow-up of 293 persons with rheumatoid arthritis (RA) and a healthy contrast group of 456 persons. The groups were identified in an epidemiological survey of RA made in Stockholm in 1965-67. Data on deaths, migration and use of in-patient care during the follow-up were obtained by a linkage with the Stockholm County Medical Information System. Follow-up shows that the persons in the RA group were admitted to hospital more than twice as often as the controls and had three times as many hospital days. However, only 15% of the individuals in the RA group were ever admitted with a diagnosis of rheumatoid arthritis and only 1% of the RA group's total number of hospital days were spent at specialist rheumatology departments. The amount of psychiatric care was also low. The use of hospital care in the RA group was closely correlated to the severity of the disease prior to follow-up, as measured by Steinbrocker functional grading, number of criteria for RA fulfilled and perceived severity of joint symptoms.