This poster is published under an
open license. Please read the
disclaimer for further details.
Keywords:
Breast, Mammography, Screening, Audit and standards, Education and training, Quality assurance
Authors:
C. E. Mercer1, K. Szczepura1, J. Kelly2, S. Millington2, E. Denton3, R. Borgen4, B. Hilton4, P. Hogg1; 1Manchester/UK, 2Chester/UK, 3Norwich/UK, 4Lancs/UK
DOI:
10.1594/ecr2015/B-1026
Methods and materials
Building on previous studies12,13 this retrospective evaluation of practitioner compression force variation was conducted over a six year screening cycle in three screening units in the North of England.
To exclude mammography machine breast readout inaccuracies14 data was gathered from one mammogram machine at each location all operating within NHSBSP and manufacturer specifications15,16.
Analogue images were utilised - NHSBSP screening sites had not been converted to digital technology for a six year period at the time of the study.
Clients were selected using consecutive stratified sampling.
Inclusion criteria:
- Three consecutive screening events
- First recorded mammogram experience as their first event
- Four standard projections (left / right CC (cranial-caudal) and left / right MLO (medio-lateral oblique)
Recorded image data available: size of film,
breast compression force value in deca-Newtons (daN) or Newtons (N),
compressed breast thickness (mm) and the practitioner who performed the mammogram.
Exclusion criteria were applied.
The breast density was read by 5 observers in the three screening units using the 4 point BI-RADS® scale (Breast Imaging Reporting and Data System)17 .
Inter and intra observer characteristics ascertained prior to this18.
Near complete intra-observer agreement (Kappa >0.81) and strong or above inter-observer variability was demonstrated (First score Fleiss kappa 0.77 second score 0.65)18.