Diagnostic discordance: we cannot agree when to call knee pain 'osteoarthritis'.

BACKGROUND Joint pain is common in community-dwelling older adults. Osteoarthritis is the most likely cause and the most common diagnosis made in this age group by GPs. However, the level of agreement between patients and their GPs in this diagnosis is questionable and may have important implications for clinical care. OBJECTIVES Our aim was to determine the level of agreement between GP diagnosis of knee osteoarthritis and patients' own attribution of osteoarthritis in older adults consulting their GP with knee pain. METHODS Forty-five patients aged > or =50 years were recruited retrospectively from consecutive knee pain attenders at two general practices in North Staffordshire. All patients were assessed by trained research physiotherapists using standardized assessment procedures. Patients' attribution was assessed by the open-ended question "What do you think is the matter with your knee now?" Blind to the findings of the assessment, a researcher extracted the most recent knee-related Read code recorded by the patient's GP from the medical records to determine GP diagnosis. The physiotherapists also classified patients as osteoarthritis or non-osteoarthritis according to the American College of Rheumatology's clinical classification criteria. RESULTS Agreement between GP diagnosis of 'knee osteoarthritis' or 'osteoarthritis NOS' and patient diagnostic attribution of knee 'arthritis' or 'osteoarthritis' was poor [kappa=-0.03; 95% confidence interval (CI) -0.32 to 0.26]. Neither the GP diagnosis nor the patient diagnostic attribution was strongly related to clinical classification using standard criteria (kappa=0.28; 95% CI -0.01 to 0.56; and kappa=-0.39; 95% CI -0.66 to -0.13, respectively). CONCLUSIONS The level of agreement between a GP diagnostic code of osteoarthritis and patients' use of the term '(osteo)arthritis' is no greater than would be expected by chance alone. Discordance may be high in the use of the label osteoarthritis but is not direct evidence of a lack of shared understanding of this condition.