Sickness certification is one of the most common tasks performed in general practice. This review describes and discusses concepts and terms used in earlier studies. 'Sickness certification' is defined, and related to the issues of 'absence from work' and 'sickness absence'. The use of measurements and results reported are emphasized according to patient- and doctor-related variables. Great variations are found, and some of the reasons may be differences in morbidity patterns, diagnostic procedures or sickness benefit acts. However, in studies from general practice, the number of sickness certificates is related to different denominators without describing the real population at risk, that is those of the patients who were employed or entitled to sickness benefits. Further studies are needed on the epidemiology of sickness certification, and the duration of the episodes. Analysis of the basis for the doctors' decisions, the patients' viewpoint, inter-doctor variations and doctors' attitudes should also be emphasized in the future. There is a need to discuss the reliability and validity of the measurements used, and theoretical considerations of the doctor's sickness certification practice are called for.