In order to suggest therapy modifications with the aim of diminishing the risk of therapy-induced long-term distressful symptoms in cancer survivors, data are needed relating details of therapy to the long-term symptom situation. In this article, the concepts and means used to assess the latter while developing the Radiumhemmet scale for symptom assessment are described. The focus is on the subjective long-term situation, and symptoms as a perceived abnormality are defined. For conceptual clarity, one symptom at a time is considered, excluding scales in which items are summarized. Moreover, measures of disease occurrence in the population are translated (epidemiologically) into measures of symptom occurrence in an individual. Nature distinguishes one long-term symptom from another. Occurrence of a symptom in an individual is measured by an incidence (e.g. number of defecations per week) or prevalence rate (e.g. urinations with involuntary cessation divided by the total number of urinations). Any scale expressing symptom intensity is arbitrary, be it 'verbal' (no/little/moderate/much pain) or visual (analogue or with integers). A time period describes symptom duration. The relevance of a symptom to emotions and social activities, sometimes cited as the associated symptom-induced distress, is a separate issue from symptom occurrence, intensity, and duration.