Problems with the measurement of radiological change

Area of Science:

• Occupational Health
• Radiology
• Environmental Medicine

Background:

• Chest X-rays were collected from asbestos workers in British factories between 1971-1978.
• The study involved 36,592 X-rays from 26,471 workers, following the 1969 Asbestos Regulations.

Purpose of the Study:

• To evaluate the consistency and validity of pneumoconiosis classification from chest X-rays in an occupational cohort.
• To identify factors contributing to inconsistencies in reading radiographic films.

Main Methods:

• Utilized the 1971 International Labour Organization (ILO) Classification of Radiographs of Pneumoconioses.
• Compared results from a pilot study using a 'side-by-side' reading method with a main survey where films were read independently and shortly after collection.

Main Results:

• Serious inconsistencies were observed in the main survey compared to the pilot study.
• A bias towards classifying films into lower ILO/UC categories was evident as the survey progressed.
• Independent reading of different X-rays from the same worker, potentially by different readers, contributed to variability.

Conclusions:

• Accumulating a set of films and reading them together over a short period is essential for obtaining valid results in pneumoconiosis assessment.
• Future re-reading of a sample of collected films will incorporate these findings to improve accuracy.