Article Text

Download PDFPDF
Fast and specific detection of moderate long-term changes in occupational blood exposures
  1. Isabelle Chaillol1,2,3,4,
  2. René Ecochard1,2,3,4,
  3. Marie-Agnès Denis5,6,
  4. Jean Iwaz1,2,3,4,
  5. Phrem Khoueiry7,
  6. Alain Bergeret6,8
  1. 1Hospices Civils de Lyon, Service de Biostatistique, Lyon, F-69003, France
  2. 2Université de Lyon, F-69000, Lyon
  3. 3Université Lyon I, Villeurbanne, F-69100, France
  4. 4CNRS, UMR 5558, Laboratoire de Biométrie et Biologie Evolutive, Equipe Biostatistique Santé, Pierre-Bénite, F-69310, France
  5. 5Hospices Civils de Lyon, Hôpital Edouard Herriot, Médecine du Personnel, Lyon, F-69008, France
  6. 6UMR T9405 – INRETS Unité Mixte de Recherche Epidémiologique et de Surveillance; Université Lyon 1; Institut de Veille Sanitaire, Bron, F-69500, France
  7. 7Hospices Civils de Lyon, Service de la Gestion, du Budget et des Effectifs, Lyon, F-69003, France
  8. 8Hospices Civils de Lyon, Centre Hospitalier Lyon-Sud, Service Central de Médecine et Santé au Travail du Personnel, Lyon, F-69002, France
  1. Correspondence to Dr René Ecochard, Service de Biostatistique, Hospices Civils de Lyon, 162, Avenue Lacassagne, F-69003, Lyon, France; rene.ecochard{at}


Objectives Hospital surveillance systems have been established to monitor occupational blood exposures. We compare short-term monitoring with long-term monitoring of data analysis over 11 years and 21 institutions to identify variations in the number of reported exposures.

Methods Short-term monitoring examines the current number of exposures compared to their average over previous years. Long-term monitoring detects trends over several years by various exposure characteristics (place, staff, procedure, etc) through estimating rates of change and using the best linear unbiased predictors (BLUPs) to prevent artefactual trends due to the many categories for each characteristic. Graphical representations of estimated rates help detect change and differences in rates of change.

Results Annual monitoring allowed detection of significant changes in the number of reported exposures. Long-term monitoring identified moderate trends over time. The BLUP corrected the estimate of each specific annual rate of change and allowed all other rates to reduce the random variability around the mean change for more specificity. League tables showed significant increases or decreases compared to no change. League tables for two-by-two comparisons allowed reliable comparisons between estimates of the rates of change, although with spurious ranking. Funnel plots enabled quick detection of changes in trends within specified confidence intervals. Long-term trends agreed with the dominant type of annual changes over the 11 years but were not as sensitive.

Conclusions The two methods have different uses. Both are helpful for assessing short-term sudden and long-term minor changes in number of exposures, possibly reflecting the success or otherwise of introducing specific safety devices or guidelines.

  • Occupational exposure
  • bloodborne pathogens
  • hospital personnel
  • hygiene/occupational hygiene
  • occupational health practice
  • statistics
  • mathematical models
  • preventive medicine

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • This work should be attributed to the Hospices Civils de Lyon. The analysis was carried out at the Department of Biostatistics on data collected at the Department of Occupational Medicine. There was no external funding.

  • Competing interests None.

  • Provenance and peer review Not commissioned; not externally peer reviewed.