Validation of a modelling approach for estimating the likely effectiveness of cancer screening using cancer data on prevalence screening and incidence
2011 (English)In: Cancer Epidemiology, ISSN 1877-7821, Vol. 35, no 2, 139-144 p.Article in journal (Refereed) Published
Purpose: This study aims to validate a biostatistical approach to predict the likely effectiveness of screening in reducing advanced disease in the absence of data on incident screen and interval cancers. Methods: We derived the predicted relative reduction in advanced stage disease following screening from the expected proportion of advanced disease following screening and the observed proportion of advanced disease detected clinically among the controls. We compared the predicted estimates to those observed in a randomised trial. Results Using our method, the predicted estimates of relative reduction in node positive breast cancer following screening were comparable to the observed estimates for the age groups 50-59 and 60-69 in the screening study (predicted 32% vs. observed 40% (p = 0.274) and predicted 34% vs. observed 45% (p = 0.068), respectively). However, for the age groups 40-49 and 70-74 the predicted values were overestimates of the likely effectiveness of screening compared to the observed values (predicted 38% vs. observed 16% (p = 0.014) and predicted 34% vs. observed 0% (p = 0.001), respectively). Conclusion: When the number of cancer cases is more than hundred, the method of prediction using only prevalence screen data may be accurate. Where cancers are less common, for example in small populations or young age groups, further data from interval cancers or incidence screens may be necessary.
Place, publisher, year, edition, pages
2011. Vol. 35, no 2, 139-144 p.
Screening, Prevalence screen, Incident cancer, Relative reduction, Advanced cancer, Overdiagnosis, Prediction
Medical and Health Sciences
IdentifiersURN: urn:nbn:se:uu:diva-153645DOI: 10.1016/j.canep.2010.07.012ISI: 000290067800006PubMedID: 20719587OAI: oai:DiVA.org:uu-153645DiVA: diva2:417403