In the assessment of outcome in rheumatic diseases a number of factors must be taken into account. It is important to make an accurate diagnosis, so that the response to treatment is not confused by heterogeneity of the population. The meaning of outcome needs to be defined. The quality of life over a prolonged period is just as important as the ultimate outcome. Subjective symptoms are important to the patient. Pain is the most important, followed by disability and then stiffness. Despite attempts to produce numerical values for pain, particularly visual analogue scales, patients' accuracy in recalling pain experienced more than an hour previously is dubious. In an attempt to quantify this aspect we have measured disturbance of sleep by changes in the EEG and in the motility of the patient. Objective clinical measures are desirable for accuracy. Arthrographs of the knees and metacarpophalangeal joints have produced useful data for physical stiffness. It is doubtful, however, whether they truly reflect the subjective stiffness of which the patient complains. That is more likely to be due to limitation of motion. Grip strength is commonly measured by a pneumatic dynomometer, but a pinch/hand grip analyser promises to give more extensive information. Active movement has been measured goniometrically. The value of electrogoniometers should be enhanced by telemeterization of the apparatus. Passive movement has been measured with a hyperextensometer in patients with hypermobility. Ligamentous laxity of the knee can be measured on the Leeds Knee Analyser and differentiates collateral ligament damage and anterior cruciate ligament damage. Laboratory variables are important in a patient model system in which potential antirheumatoid drugs can be screened and their mechanism of action investigated. Correlation matrices separate second-line agents from NSAIDs. Although mini-matrices have been produced, it would not appear that any single biochemical test will suffice to differentiate these two classes of drugs. A therapeutic index, in which the efficacy is expressed as a ratio of the toxicity of the drug, is important in determining its value. The nearest we can get to serial assessment of the pathological changes in the joint is by X-ray assessment. Changes radiologically correlate to some extent with the height of the ESR, and their progress with changes in the ESR. Functional impairment is important to the patient in the long term, and the Disability Index devised by the Stanford group commends itself for extensive long-term studies.