'B-readers' and asbestos medical surveillance

Our plain-English summary of 'B-readers' and asbestos medical surveillance, by A.M. Ducatman, W.N. Yang, S.A. Forman (Journal of Occupational Medicine., 1988).

The study

• Authors: A.M. Ducatman, W.N. Yang, S.A. Forman
• Journal: Journal of Occupational Medicine.
• Published: 1988;30(8):644-7
• Cited by: 45 articles, as counted by NIH iCite
• Read it: PubMed · Publisher
• In J-M's library: Johns-Manville's company medical library held a copy (J-M no. 000902).

Abstract

"B-readers" certified in International Labor Office methodology interpret large numbers of randomly distributed asbestos medical surveillance roentgenograms of US Navy employees. Analysis of 23 participating observers, interpreting more than 105,000 radiographs, demonstrated a 300-fold prevalence range of perceived "definite" pulmonary parenchymal abnormalities. There was an evident geographic component to interpretation habits, with East and West Coast observers more likely to interpret films as abnormal than observers from the midcontinent. The most expert observers, a group who instruct the course leading to National Institute for Occupational Safety and Health certification in International Labor Office methodology, also perceived fewer abnormalities than other readers or coastal observers. Instructors still exhibited a sevenfold prevalence range of positive interpretation. Under usual surveillance conditions, the habits of B-readers appear to have a major impact upon the diagnosis of asbestosis from roentgenograms. Certification in B-reading should not be the only quality assurance for radiographic surveillance programs, medical decision-making, epidemiologic comparisons, nor related legal activities.

In plain English

This study examined how much certified 'B-readers' differed when reading chest X-rays from the US Navy's Asbestos Medical Surveillance Program. Each film was read by only one of 23 civilian B-readers, assigned at random and blinded to worker details. The authors analyzed each worker's most recent X-ray, 105,029 in all, counting only findings in the lung tissue itself (parenchymal findings). The share of films a reader called definitely abnormal (ILO category 1/1 or higher) ranged from 0.02% to 7.55%, a 377-fold range. One reader, H, read 4.4% of films but made 33% of those calls. Readers from Pacific and Atlantic coast states called films abnormal more often than midcontinent readers. Instructors of the NIOSH certification course read 26.4% of films but found only 14.9% of the definite abnormalities. Six instructors still ranged from 0.17% to 1.29%. No two readers read the same film, so the authors called these inferential data. The authors concluded that B-reader habits appear to have a major impact on diagnosing asbestosis from X-rays, and that certification should not be the only quality assurance.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Analysis of routine surveillance X-ray interpretations, inferring reader habits. Each film was read once by a single, randomly assigned, blinded B-reader; the same films were not re-read by others.
• Who was studied: The most recent chest radiograph of 105,029 individuals (US Navy employees) in the US Navy's Asbestos Medical Surveillance Program. The films were interpreted by 23 independent civilian B-readers working under three-year contracts awarded since 1978.
• Exposure: No fiber types or exposure levels were studied. The variable compared was the B-reader who interpreted the film, grouped by the reader's region (East, Pacific, midcontinent) and by whether the reader was an instructor of the NIOSH certification course. Only coded parenchymal (lung tissue) findings were analyzed.
• Main results:
  • Across the 23 B-readers, the share of films classified as definite parenchymal abnormality (1/1 or higher) ranged from 0.02% to 7.55%, a 377-fold range.
  • Reader H read 4.4% of all films but contributed 33% of films classified 1/1 or higher, and more than half of those classified 2/1 or higher.
  • Excluding the two most extreme readers (C and H) reduced the range to 28-fold.
  • Pacific-state readers classified films as 1/1 or higher about four times as often as midcontinent readers. This ratio dropped from 4.5:1 to 1.8:1 when the one outlying reader's films were removed. Eastern seaboard readers did so about twice as often as midcontinent readers (2.1:1).
  • In one Pacific Coast city (25,118 films), 1.97% were read as 1/1 or higher. In one midcontinent city (15,235 films), 0.20% were, a ratio of 9.9:1.
  • Instructors of the NIOSH certifying course read 26.4% of the films but found only 14.9% of the definite (1/1 or higher) abnormalities. The six instructors individually ranged from 0.17% to 1.29%, a sevenfold range that rose to 15-fold at the 1/0 threshold.
  • Lowering the threshold from 1/1 to 1/0 reduced the overall range among readers from 377-fold to 219-fold, but increased the apparent differences between regions.
  • For the more serious categories 2 and 3, differences among regions were far less marked.
• The authors' conclusion: The authors concluded that, under usual surveillance conditions, B-reader habits appear to have a major impact on the diagnosis of asbestosis from X-rays. They stated that B-reading certification should not be the only quality assurance for surveillance programs, medical decision-making, epidemiologic comparisons, or related legal activities. They said individual diagnoses, legal decisions and population assessments ought to rely on multiple readings.
• Limitations the authors state:
  • The data are inferential: no two readers interpreted the same film, so the study does not directly measure interobserver variability.
  • The available data did not allow the randomization of initial films to be analyzed. Randomization was shown to have worked for follow-up films.
  • The 23 readers may not be representative of the more than 450 B-readers on the NIOSH list.
  • There is no gold standard for B-reading in surveillance settings, and the cutoff used for comparisons was arbitrary.

Topics

The National Library of Medicine indexes this study under subjects our site covers: Asbestosis, The Navy.

Medical terms in this study, explained in our Lexicon: B-Reader, Habit.

Related studies

• Variability in interpretation among B-readers in the U.S. Navy Asbestos Medical Surveillance Program — J.G. Muller, J.M. Lieske, et al. (2008)
• Navy Asbestos Medical Surveillance Program 1990-1999: demographic features and trends in abnormal radiographic findings — D.M. Dell, B.K. Bohnker, et al. (2006)
• Trends in pleural radiographic findings in the Navy Asbestos Medical Surveillance Program (1990-1999) — J.G. Muller, B.K. Bohnker, et al. (2005)
• Navy Asbestos Medical Surveillance Program (1984-1990): linear regression analysis for effect of asbestos exposure on pulmonary function testing — B.K. Bohnker, L.S. Betts, et al. (2004)
• Navy Asbestos Medical Surveillance Program (1991-1999): linear regression analysis for the effect of asbestos exposure on pulmonary function testing — B.K. Bohnker, L.S. Betts, et al. (2004)
• Comparison of "B" readers' interpretations of chest radiographs for asbestos related changes — J.N. Gitlin, L.L. Cook, et al. (2004)

Browse more research in Journal of Occupational Medicine., or return to medical research.