The relationship between real life breast screening and an annual self assessment scheme

Incidence of cancer in the UK NHS Breast Screening Programme (NHSBSP) is relatively low (approximately 7% per 1,000 cases screened). As such, feedback from cancers missed or interval cancers can be a relatively lengthy process (whereby a woman will not present for corroborating imaging for a further three years). Therefore in order to monitor their radiological skill, all breast screening radiologists and technologists read a self-assessed, standard set of challenging mammographic images bi-yearly. This scheme, ‘PERFORMS’ (Personal Performance in Mammographic Screening) has been running since near the inception of the NHSBSP in 1991. Although PERFORMS has functioned as an educational tool for film-readers on the UKBSP for decades, its relation to real life screening in past years has proven to be somewhat equivocal (Cowley & Gale, 1999). The present study investigated the relationship between performance measures in real life and their equivalent on the PERFORMS self assessment scheme namely: Miss Rate (FN), Cases Arbitrated and Returned to Routine screening and Incorrect recall (FP), Specificity (TN) and Cancer Detection (TP). Over 40 individuals from one NHS region in the UK submitted their real life data for comparison with PERFORMS results from the same time frame. Data from this initial study were taken from the year 2005-2006 and compared with the relevant PERFORMS set of cases. Results indicated a significant positive correlation between PERFORMS performance measures and performance measures for real life. These results are discussed in the light of the legitimacy of self-assessment comparative to film-reading skill (during real life clinical practice).