Chest Radiography

There are four necessary components for attainment of reliable classification of chest radiographs for the pneumoconioses: 1) appropriate methods for image collection and viewing, 2) reader competency, 3) commitment to ethical classification, and 4) proper radiographic reading methods.

Recommendations are available for addressing the first component: obtaining chest images of sufficient quality for pneumoconiosis classification and for optimizing viewing conditions (ILO, 42 CFR 37.41 external link). The NIOSH B Reader Program addresses the second component: identification of competency in the classification of radiographs for abnormalities due to dust. The B Reader Code of Ethics addresses the third component in more detail.

The fourth of these components, proper radiographic reading methods, is addressed by the recommended best practices summarized below. Individually designated for each of three principal settings involving chest radiography for pneumoconiosis, these recommended best practices represent practical, real-world approaches tailored to the unique needs of each setting. Details are provided elsewhere on two other settings: Medical Diagnosis and Government Programs. In general, ILO classification of the chest radiograph is not necessary for the former setting, while the latter has specialized guidance that is summarized for each program.

Worker Monitoring and Surveillance  

1. ILO system: For consistency with accepted definitions of radiographic appearances of the pneumoconioses and associated disorders, and to ensure uniform standards within a monitoring program and comparability with other data, it is necessary that chest radiographs be classified using the ILO system.
2. Remuneration: To avoid biased classifications, it is necessary that a reader’s remuneration not be related in any way to classification outcomes.
3. Reader selection: Selection of B Readers for monitoring programs is generally based on practical considerations, such as availability, access to radiographs, and timeliness.
4. Number of readers and summary classifications: A single B Reader classification of each chest radiograph is generally sufficient; additional independent classifications may be needed to ensure reliability within the program. Where multiple readings exist, it is desirable that the summarization approach for a specific abnormality be essentially unbiased, with any bias favoring disease detection. (For more details, see Classification of Chest Radiographs: Practices for Worker Monitoring and Surveillance.)
5. Blinding: In order to facilitate disease detection in environments where individuals are potentially at risk, blinded classification is undesirable.
6. Quality assurance: Formal quality assurance procedures are useful for maintaining accuracy and precision within a program over time. (For more details, see Classification of Chest Radiographs: Practices for Worker Monitoring and Surveillance.)
7. Notification: It is ethically necessary to inform individuals of their personal findings, and, if surveillance identifies potential risks of occupational disease within the workplace, inform workers and employers of those risks.

For more details, see Classification of Chest Radiographs: Practices for Worker Monitoring and Surveillance.

Epidemiologic Research

1. ILO system: For consistency with accepted definitions of radiographic appearances of the pneumoconioses and associated disorders, and to ensure comparability with other scientific information, it is necessary that chest radiographs be classified using the ILO system.
2. Remuneration: To avoid biased classifications, it is necessary that a reader’s remuneration must not be related in any way to classification outcomes.
3. Reader selection: To achieve scientifically-accepted standards of precision independent classifications by B Readers are necessary. To maximize the likelihood of unbiased classifications, selection of readers should follow scientifically-based criteria established beforehand. (For more details, see Classification of Chest Radiographs: Practices for Epidemiologic Research .)
4. Number of readers and summary classification: The ILO recommends at least two but preferably three independent classifications. For the highest flexibility and quality of data, three or more classifications of each radiograph are desirable. To avoid bias, a scientifically acceptable summarization approach is necessary. (For more details, see Classification of Chest Radiographs: Practices for Epidemiologic Research.)
5. Blinding: To avoid bias, blinded classification is necessary.
6. Quality assurance: To achieve scientifically-accepted standards of accuracy, formal quality assurance procedures are necessary. (For more details, see Classification of Chest Radiographs: Practices for Epidemiologic Research.)
7. Notification: Whenever possible and especially when medical findings are pertinent to maintaining and protecting health, it is ethically necessary to inform the individual of their radiographic findings. (For more details, see Classification of Chest Radiographs: Practices for Epidemiologic Research.) There is ethical justification in notifying member of occupational cohorts and their employers of results of scientific investigations in which they participate.

For more details, see Classification of Chest Radiographs: Practices for Epidemiologic Research.

Contested Proceedings

1. ILO System: Use of the ILO system provides an accepted means of standardizing disease assessment, and thus is necessary to ensure fairness and equity in contested proceedings.
2. Remuneration: To avoid biased classifications it is necessary that a reader’s remuneration not be related in any way to classification outcomes.
3. Reader selection: To ensure accuracy in classification, it is necessary to select readers at random from the largest pool of B Readers available, following a procedure that is defined and documented at the outset. (For more details, see Classification of Chest Radiographs: Practices in Contested Proceedings.)
4. Number of readers and summary classifications : To avoid any implication of bias, it is necessary to specify from the outset the number of readers that will be used. It is not acceptable for any party to obtain additional classifications beyond those specified at the beginning of the classification process. NIOSH recommends a minimum of two independent classifications by appropriately selected readers, with a third classification if the first two disagree. The latter is used for resolving disagreement when this is critical to the outcome of the contested proceedings. To permit a full review of the impact of radiograph quality on the reliability of the classifications, it is necessary to report the radiograph quality information from each classification separately. To avoid bias when deriving summary classifications for parenchymal and pleural abnormalities, it is necessary to employ appropriate procedures defined and documented at the outset. (For more details, see Classification of Chest Radiographs: Practices in Contested Proceedings.)
5. Blinding: To avoid bias, blinded classification is necessary.
6. Quality assurance: The Contested Proceeding setting poses a particular challenge for achieving reliable readings, indicating the necessity of incorporating adequate quality assurance procedures in the classification process. (For more details, see Classification of Chest Radiographs: Practices in Contested Proceedings.)
7. Notification: Whenever possible and especially when medical findings are pertinent to maintaining and protecting health, it is ethically necessary to inform the individual of their radiographic findings. Medical follow-up should be recommended where appropriate. (For more details, see Classification of Chest Radiographs: Practices in Contested Proceedings.)

For more details, see Classification of Chest Radiographs: Practices in Contested Proceedings.

References

International Labour Office (ILO). Guidelines for the Use of the ILO International Classification of Radiographs of Pneumoconioses, Revised Edition 2000 (Occupational Safety and Health Series, No. 22). International Labour Office: Geneva, 2002.

Specifications for medical examinations of underground coal miners. 42 CFR 37 External link: http://www.access.gpo.gov/nara/cfr/waisidx_02/42cfr37_02.html

Issues in Classification of Chest Radiographs

Resources