The Osteoarthritis (OA) Criteria Subcommittee of the American Rheumatism Association set out to develop (a) a classification of OA that includes recognised subsets; and (b) subsets of OA identified by a combination of clinical and laboratory features. For the purposes of classification, OA should be specified if of unknown origin (idiopathic, primary) or if related to a known medical condition or event (secondary). Clinical criteria for classification of idiopathic OA of the knee were developed through a multicentre study group involving 130 patients with OA and 107 comparison patients. Comparison diagnoses included rheumatoid arthritis (RA) and other painful conditions of the knee exclusive of referred or para-articular pain. Variables from the history, physical examination, laboratory test results and radiographs were used to develop sets of criteria that serve different investigative purposes: clinical examination (sensitivity 89%; specificity 88%); clinical examination and laboratory tests (sensitivity 88%; specificity 93%); clinical examination, laboratory tests and radiographs (sensitivity 94%; specificity 88%). In contrast to prior classification criteria, the proposed criteria utilise decision trees or algorithms. Clinical criteria for classification of idiopathic OA of the hip are under development. Comparison groups are comprised of patients with other rheumatic diseases (e.g. RA), periarticular pain (e.g. trochanteric bursitis) and referred pain (e.g low back pain). From a method of opinion sampling, OA of the hip may be suggested by a combination of clinical criteria including the following: age greater than 40 years, weight-bearing pain, pain relieved by sitting, antalgic gait, decreased painful range of motion, a normal erythrocyte sedimentation rate (ESR) and a negative rheumatoid factor test.