A comparison of interpretation of screening mammograms by a radiographer, a doctor and a radiologist: results and implications.

In 1986, a project began to assess the feasibility of mobile breast screening, using oblique view mammography, in rural areas near Edinburgh. One of the aims of the project was to determine the feasibility, in terms of accuracy and cost, of non-radiologist prescreeners. The sensitivity and specificity of each reader was calculated (using 12-month interval cancer rates). Sensitivities of 80%, 80% and 83% were found for a radiographer, a non-radiologist doctor and a radiologist respectively, with corresponding specificities of 78%, 84% and 86%. When each reader was costed, prescreening was no cheaper than the radiologist reading all the films himself (79.10 pounds for the radiologist reading 1,000 pairs of films alone, against 84.97 pounds with the radiographer prescreening and 163.93 pounds with the non-radiologist doctor prescreening). We were unable to demonstrate any advantages of prescreening, but found that double reading of the films, though costly, would have increased the sensitivity of the programme.