Effect of screening on cervical cancer mortality in England and WalesBMJ 1999; 319 doi: https://doi.org/10.1136/bmj.319.7209.581 (Published 28 August 1999) Cite this as: BMJ 1999;319:581
Interactive effects and causation must be considered
- Wai-Ching Leung, senior registrar in public health medicine
- Northern Region Public Health Medicine Training Scheme, County Durham Health Authority, Durham DH1 5XZ
- Department of Mathematics, Statistics, and Epidemiology, Imperial Cancer Research Fund, London WC2A 3PX
EDITOR—From their modelling of mortality data Sasieni and Adams estimated the number of deaths from cervical cancer prevented as a result of screening.1 Two issues must be considered.
Firstly, they did not consider possible interactive effects between the age effect and the cohort effect. Factors such as increases in life expectancy, possible improvements in the treatment of cervical cancer, and different secular changes in sexual behaviour among different age groups could all have produced such interactive effects. The authors have not explained why they ignored this possibility.
Secondly, even if there were a reduction of deaths from cervical cancer, it could have been due to reasons other …