Breast cancer screening is an integral part of the cancer control strategies of many developedeconomies. In Britain individuals screened in a given year are re-called every three years unless resultsindicate a need for more immediate investigation. This pattern may create a legacy arising from pastdecisions, a legacy that should be considered when examining current decisions. In this paper we use abalanced panel drawn from the British Household Panel Survey of 1,997 women over an 18 yearperiod to examine variations in uptake. A dynamic random effects probit model is used to control forunobserved heterogeneity and the legacy of previous decisions. As might be expected women to whomuniversal screening is offered are more likely to screen than others. Changes during the study period inthe eligible age range saw an increase in uptake among the age group to whom the programme wasextended but not among other groups. Past screening behaviour was found to be a significant predictorof current behaviour. Failure to account for past choices may result in model mis-specification and afailure to develop policies aimed at promoting initial engagement that may compromise the screeningprogramme. Income was not found to be a significant determinant of uptake.