The effects of treatment in early rheumatoid arthritis (RA) and the consequences of delayed therapy represent important areas for research. The concept of a 'window of opportunity' is now well established and considerable attention has been paid to when it might close. However, in order to study how long the window of opportunity lasts, the timing of its opening must be precisely defined. An analysis of definitions of 'onset' in clinical studies reveals imprecision and heterogeneity, making accurate assessment of this important concept of the 'window of opportunity' very difficult. In this paper we propose that, in clinical trials in early RA, data on durations since onset of symptoms and onset of joint swelling as well as disease duration based on fulfilment of classification criteria should be routinely presented.