How frequently should workplace spirometry screening be performed? Optimization via analytic models

Philip Harber, Jessica Levine, Siddharth Bansal

Research output: Contribution to journalArticlepeer-review

7 Scopus citations

Abstract

Background: Our objective was to determine how to select the optimal frequency of workplace spirometry screening using diacetyl-exposed workers as an example. Methods: A Markov model was constructed to assess the likelihood of progressing from healthy status to early or advanced disease, starting from four different exposure levels, and performing longitudinal or cross-sectional interpretation of spirometry results over time. Projected outcomes at 10 years were evaluated to inform the optimal frequency of workplace spirometry testing. Results: The optimal screening interval depends on the population risk and is highly sensitive to the real-life impact (utility) associated with false-positive results (eg, related to the availability of alternative work). Screening interval is particularly important for high-risk individuals with rapid transition from early to advanced disease, where the 10-year prevalence of advanced disease would be reduced from 5.3 to 2.5% using a 6-month interval rather than a 12-month interval. Longitudinal test interpretation, based on observing trends within each person over time, is marginally preferable to traditional cross-sectional spirometry interpretation. Conclusions: There is no single best screening interval. For high-risk populations, annual testing may be too infrequent.

Original languageEnglish (US)
Pages (from-to)1086-1094
Number of pages9
JournalCHEST
Volume136
Issue number4
DOIs
StatePublished - Oct 1 2009
Externally publishedYes

ASJC Scopus subject areas

  • Pulmonary and Respiratory Medicine
  • Critical Care and Intensive Care Medicine
  • Cardiology and Cardiovascular Medicine

Fingerprint

Dive into the research topics of 'How frequently should workplace spirometry screening be performed? Optimization via analytic models'. Together they form a unique fingerprint.

Cite this