Monday-Friday
08:30-17:30 classroom teaching (lectures and exercise sessions)

Saturday (no new material will be presented)
09:00-11:30 practical session
11:30-12:00 joint presentation ceremony for all classes at the school
12:00-13:00 lunch

All teaching (lectures and practical exercise sessions) will be in the same room and the exact timetable will vary from day to day. There will be a greater focus on lectures during the early part of the course and a greater focus on other activites (exercises and discussion) during the latter days. No new material will be presented on Saturday.

There will be 30-minute coffee breaks each morning and afternoon and a 1-hour lunch break each day from 13:00-14:00.

Our goal is to provide each participant with individual instruction. A large amount of time will be devoted to exercise sessions where 5 faculty members will be available to work with participants individually or in small groups. We will provide an extensive set of exercises with fully-worked solutions but the exercise sessions will also provide an opportunity for participants to discuss their own research projects with the faculty (and with each other).

We are continually improving the course based on our own experiences and participant feedback. We have further developed and improved the lectures and exercises, added new exercises, and restructured the order in which content is presented. The course content has been updated to reflect recent developments in the field and in our own research. In particular:

• The course has been restructured to be useful to participants without a specific interest in cancer patient survival. The methods we teach (e.g., competing risks, modelling (especially flexible parametric modelling), regression standardisation etc.) can be applied to any area.

• Expanded coverage of various marginal measures, their estimation, and their advantages and disadvantages. We will discuss marginal relative survival, marginal crude probabilities, marginal life expectencies. See, for example, our recent paper on marginal measures and causal effects using the relative survival framework.

• Expanded coverage of the use of regression standardisation (e.g., for age standardisation) and the link with the literature on causal inference and mediation analysis.

• Therese Andersson, Mark Rutherford and Paul Lambert worked on projects within the International Cancer Benchmarking Partnership (ICBP) and will discuss how differences in registration practice and data quality can impact on survival differences.

• Expanded coverage of methods for estimating loss in expectation of life and proportion of expected life lost due to cancer, along with the utility of these measures for quantifying differences between, for example, regions or socioeconomic groups, and the potential savings in person-years that could be achieved by eliminating differences.

• What is 'population-based cancer survival analysis' and what makes it special compared to other applications of survival analysis?
• Net survival; cause-specific survival; relative survival; relative merits of cause-specific survival and relative survival for population-based cancer registry data;
• Estimating patient survival using the actuarial and Kaplan-Meier methods.
• Non-parametric estimation of net survival (including the Pohar Perme estimator);
• Impact of (erroneously) including cancer patients in the population mortality file when estimating expected survival.
• Obtaining, constructing, and extending population mortality rates for the purpose of estimating expected and relative survival;
• Age standardisation of net survival, including model-based standardisation;
• Reference-adjusted and standardized all-cause and crude probabilities;
• Cohort, complete, period and hybrid approaches to estimation;
• Poisson regression, Cox regression, flexible parametric (Royston-Parmar) models and their application to modelling cause-specific mortality and excess mortality;
• Assessing the proportional hazards assumption; non-proportional hazards and how to adjust for them;
• Cure models for relative survival - estimating and modelling the cure proportion; flexible parametric cure models;
• Direct modelling of marginal net survival;
• Regression standardisation (g-computation) and links to causal inference;
• Estimation of life expectation and proportion of expected life lost;
• Estimation in the presence of competing risks (in both cause-specific and relative survival frameworks);
• Methods for analysing data with missing covariates (lecture notes but no lecture);
• Estimating the number of avoidable premature deaths;
• Discussion of what to include in a (cancer registry) report of cancer patient survival (e.g., the relative merits of various approaches for various target audiences);
• Impact of data quality, completeness, stage migration, screening and lead-time bias;
• Potential biases in estimates or patient survival;