Article Text

Download PDFPDF

157 Occupation and cancer: an alternative tool for evaluating the evidences
  1. O E Oddone1,
  2. C P Crosignani2,
  3. S A Scaburri2,
  4. B E Bai2,
  5. M C Modonesi2,
  6. I M Imbriani1
  1. 1University of Pavia, Pavia, Italy
  2. 2Italian National Cancer Institute, Milan, Italy


Objectives Meta-analysis and pooled analysis are considered as a gold standard to summarise and to sift the biomedical evidences. In the framework of the Occupational Cancer Monitoring (OCCAM), it was developed a tool to assess the existence of an association between industrial sectors and cancer risk of a specific site so called “Literature Matrix” (LM), storing only positive literature results. The aim of this work is to challenge the adjusted results of this tool with those of meta-analysis studies.

Methods To evaluate the effectiveness of LM to provide useful summary risk estimates, we compared formal meta analyses with the set of “positive” results provided by the matrix. Among the several associations provided in LM, some were selected for present study. For this exploratory study we limited comparisons to 7 areas: agriculture and hematopoietic cancers as a whole, agriculture and non-Hodgkin lymphoma (NHL), agriculture and leukemia, agriculture and multiple myeloma, transports sector and female breast cancer (compared with two different meta-analysis), transports sector and lung cancer and painters and bladder cancer.

Results LM data, after adjustment with trim and fill method, show a substantial agreement with the correspondent meta-analysis, although results driven from LM data tend to an expected overestimation, but in most cases very slight. Results from LM data are always included between the limits of 95% CI of the correspondent meta-analysis, with the exception of Agriculture/NHL and Agriculture/All hematopoietic cancers. Results from LM data not adjusted are constantly farer from meta-analytic results, with the only exception of Painter/Bladder cancer analysis.

Conclusions The collection of only positive results derived from the scientific literature, and the use of an appropriate statistical correction, allows for a useful estimation of cancer risk by site and economic branch and it can be used for interpreting results of surveillance systems and for public health purposes.

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.