Rheumatic diseases in a health interview survey and in in-patient care. A record linkage study on two registers.

A record linkage was performed between the in-patient register of Stockholm County and the Stockholm County sample of a national health interview survey. The purpose was to describe and compare characteristics of the two registers, particularly as regards rheumatic diseases. All persons included in the health interview survey of 1977-78 were sought in the in-patient register for the same period. 37% of the population had reported long-term illness in the health interview survey and 22% had been hospitalized, according to the in-patient register. 11% reported a rheumatic disease in the health interview survey, but only 2% were hospitalized with a rheumatic diagnosis. Persons who reported a rheumatic disease in the health interview survey were compared with persons hospitalized with a rheumatic diagnosis by using items in the health interview survey on perceived health, physical mobility and work capacity. The most important difference between the groups was a much higher proportion of disabled persons in the hospitalized group. Perceived health was reported very similarly in the two groups. The reports on hospital episodes in the health interview survey were validated against the in-patient register. An under-reporting of 13% and several inaccuracies in the timing of the hospital episodes were found.