A multi-choice Patient Knowledge Questionnaire (PKQ) was developed for use with patients with rheumatoid arthritis (RA). Test/re-test was used to test its stability (r = 0.81), and Kuder Richardson formula 20 (r = 0.72) for internal consistency. Seventy randomly selected RA patients then completed the PKQ in a rheumatology out-patient clinic of a large teaching hospital. There was a wide variation in total scores ranging from 3 to 28 out of 30. Total scores correlated with years of general education (P less than 0.05) but not with disease duration or age. Sixty-two per cent of patients knew that the cause of RA is, as yet, unknown but 27% thought it could be caused by injury and 11% by cold damp weather. Fifty-two per cent had no idea why they had blood tests. All but four patients were taking some form of medication but there was widespread confusion about disease-modifying drugs and non-steroidal anti-inflammatory drugs (NSAIDs). Exercise was reasonably well understood but many patients were unable to differentiate between methods of energy conservation and joint protection. This study highlights the need for careful individual knowledge assessment by use of tools such as the PKQ and effective patient education programmes.