Wolfson Institute of Preventive Medicine - Barts and The London

Analysis and modelling of cancer statistics

Centre of Cancer Prevention

Project Investigator:  Peter Sasieni


Often cancer registries in low to middle income countries do not have access to a statistician, while some epidemiologists are unable to use packages such as Nordpred or SEER*Stat that require a high level of technical ability.

This project aims to produce a user-friendly web-based programme for the analysis of cancer incidence data in conjunction with Dr Bray at the International Agency for Research on Cancer (IARC).

There is currently a lack of capacity for supporting registries to analyse their data in a standardised way. This project plans to fill that gap and would be highly beneficial in increasing a registry's ability to analyse and interpret their data for reports and for peer-reviewed research articles.

The software will have one module in basic descriptive statistics (including age-standardised rates, lifetime-risk, population pyramids, and graphs of age-specific rates). A second module for time trends in cancer incidence would include simple graphs (trends with calendar period and birth cohort) as well as options to fit age-period-cohort models and to make both short-term (1-5 years) and medium term (5-20 years) projections.

The projection module will build on the previous work of the two research groups.

Further Information

Relevant methodological problems include:

  • Developing confidence intervals for projections that take into account model uncertainty as well as probabilistic uncertainty
  • Developing random-effects models to perform age-period-cohort modelling in registries that we would anticipate would have similar age-effects and potentially similar trends (period and cohort effects).
  • Integrating the projections of cancer incidence, survival and mortality to ensure that they are consistent

Progress in these areas should enable us to produce better quality-control tools to identify potential artefacts related to registration or other factors.


Peter Sasieni
Email: p.sasieni@qmul.ac.uk
Phone: +44 (0) 20 7882 3544