Estimates of medical need in persons with rheumatoid arthritis (RA) are rarely based on data from population surveys. In a population survey of RA in Stockholm City, 293 RA sufferers were identified in four random samples of four age-groups in the adult population, age-range 31-74 years. The influence of age, sex, and severity of disease on the number of pervious hospital admissions was slight, indicating that diagnosis, rather than rehabilitation, is the impetus to hospital admission. Hospital patient inventories might give quite misleading information as to need. Need for hospital admission, out-atient care, reconstructive surgery, ADL-devices, physical therapy, and rehousing were all considered. 17% of the RA group accounted for 48% of the need. Steinbrocker functional classes I and II contributed to 61% of individuals in need of hospital care greatly exceeded those previously having enjoyed such care and possible availabel resources. Estimates of need for outpatient care for joint disease exceeded provided treatment by only 1.7 times. It is concluded that the psychological distance between doctor's decision to choose inpatient or outpatient treatment in the case of RA is generally fairly short. Figures are given for estimates of need based on a population of 100 000 with known age and sex distribution and prevalence of disease.