1. Introduction

From June 2017 onwards, the Office for National Statistics (ONS) and Public Health England (PHE) will be in partnership to produce England cancer survival estimates for:

  • Adult cancer survival (National Statistics)

  • Adult cancer survival by stage at diagnosis (Experimental Statistics)

  • Cancer survival for children (Experimental Statistics)

This report compares the methods currently used (from June 2017) in producing these national survival statistics to methods used previously. It does not consider the methodology applied to sub-national cancer survival estimates.

Back to table of contents

2. Collaboration

This publication is produced in partnership with Public Health England Cancer Survival Team, part of the National Cancer Registration and Analysis Service (NCRAS).

Back to table of contents

3. Relevance to policy

The Official Statistics on cancer survival form an evidence base to inform cancer policy and programmes that aim to improve cancer outcomes. The statistics are commissioned by the Department of Health and are used to:

  • help inform government policy on cancer

  • provide non-government bodies with accurate and timely data on the disease

  • provide citizens with accessible data on the disease to help inform debate

To ensure the data is relevant, ONS and PHE strive to produce estimates that are timely, accurate and accessible.

Back to table of contents

4. Adult cancer survival

Overview of methodological aspects

The following elements of cancer survival methodology are compared; the cancer registrations that are included in the analysis (exclusion / inclusion criteria), the type of survival estimator (net versus relative survival), the weights used to age-standardise the estimates and the version of International Coding Definition (ICD) used to define cancer sites (summary of methods).

Table 1: Summary of methods

Method
Rationale
Outcome

Exclusion/inclusion
criteria

We will continue to use the same criteria that have been historically applied. The method is outlined in the Control of data quality for population – based cancer survival analysis
paper (Li R 2014). These are supplemented by further data quality checks proposed
by the European Network of Cancer Registrations (ENCR)
(Martos C 2014)

No change in method

Survival
estimation method

Since 2012, net survival (Pohar Perme M 2012),
using an unbiased estimator has been used instead of relative survival. The partnership will continue to produce net survival estimates.
This is the most appropriate method for official statistics as it:
• removes the effect of general mortality, so only reflects “excess” mortality due to
diagnosis of cancer
• is internationally comparable, with a growing will within the UK & Ireland
Association of Cancer Registries
to move to the same methods and eventually UK – wide statistics

No change in method


We will continue to produce net survival estimates for 1-, 5- and 10-years of follow - up.


Weights
used to age-standardise the estimates

The weights historically used derived from the Cancer Survival Trends book (Coleman M
1999) based on incidence from 1986 to 1990 (Table A1, Appendix 1).

We will be moving to the International cancer patient population for age standardising survival ratios (ICSS) (Corazziari I 2004) for the publications from June 2017onwards.

ICD
coding version

Currently, the tumour sites in all ONS’s cancer survival, incidence and mortality bulletins are defined using ICD10 - O2 (Table A10, Appendix 4).
Clinical practice and evidence now suggests that a move to ICD - O3 would produce more homogenous cohorts.

No change in method at present

Being considered for future publications

Predicting
10-year survival

We will continue to use the same approach to calculating predictive estimates that have been applied in previous publications, the hybrid method, which is outlined in the Hybrid analysis for up-to-date long-term survival rates in cancer registries with delayed recording of incident cases paper
(Brenner H 2004).

No change in method

In summary, the main change to methods is the adoption of the International Classification of Survival Standard (ICSS) international cancer patient population for age standardising survival ratios.

The reasoning for this change is fourfold:

A summary of the International Cancer Survival Standard for age-standardisation.

When studying diseases present in a population, many of these diseases and their effects are closely related to age. Although comparing survival rates for specific age-groups can be informative, it is often useful to be able to have a summary of survival for every patient in the population diagnosed with a type of cancer. Comparisons using a summary estimate made over time or between geographies can be misleading if the age profiles of the underlying populations are different.

To overcome these potential drawbacks, a weighted average of the age-specific rates is calculated; this process is called age-standardisation. The weights used are independent of geography and time, so they allow for direct comparison of survival estimates over time and between different geographically – defined populations.

Many types of cancer, like lung cancer, are more commonly diagnosed as age increases; others, like Hodgkin’s lymphoma, are more commonly found in younger people. There are also cancers, like brain cancer, that have two peaks in the age distribution where they are commonly diagnosed in the younger and older age-groups but are less commonly diagnosed in middle age. A fourth profile, similar to the first group, is observed for prostate cancer but is far less likely to occur in younger men than other cancers fitting the first profile.

The ICSS system of weights (Table A2) results from analysing which cancer types best fit one of the population profiles described above. An optimal set of weights is then determined that provides the largest number of estimates with good agreement between the standardised and raw survival estimates. There are four sets of weights, reflecting the four commonly observed age profiles of diagnosing different cancers and the weights are set out in Table A2.

Please note that the results throughout this paper have been presented to allow a direct comparison to the previously published estimates. Therefore, allowing the sole impact of the change in age-standardisation methodology to be assessed. These adult cancer survival estimates do not observe all the quality controls for robustness that will be applied to the National Statistics, where there is the potential for some cancer sites having age-groups combined prior to age-standardisation or the age-standardised rate being suppressed in cases where the quality criteria is not fully met. Further information is detailed in the Cancer survival statistical bulletins Quality and Methodology Information paper.

What is the impact of changes in aspects of the methodology on survival estimates?

To assess the impact of the change in the age-standardised weighting, we have run the new method on the same cohort of cancer registrations that was used to produce the survival estimates published in 2016 (Difference A column, Table 2). We also ran the new method on an updated cohort of cancer registrations from PHE, to assess the impact of change in data quality (Difference B column, Table 2), which focuses on 1-year survival as an example. Comparisons between 1- , 5- and 10-year survival for 24 common cancer sites are available in Appendix 2. The age-weights associated with both analyses can be found in Tables A1 and A2 of Appendix 1.

Table 2 shows that the impact from changing age-weights when using the same underlying data. The absolute differences, in Difference A, range from 0.1% to 10.9% in males (melanoma of skin and brain cancer) and from 0% to 9.7% in females (melanoma of skin and brain cancer).

However, when the updated data and the same ICSS weights are used (Difference B), the absolute differences are much smaller than those found when comparing the different weights (Difference A).

In males, the differences resulting from updating the data range from 0% (Hodgkin lymphoma, kidney, lung, mesothelioma myeloma, non-Hodgkin lymphoma, oesophagus, pancreas, stomach cancer) to 0.4% (brain, leukaemia and liver cancer) (Difference B).

In females, the differences resulting from updating the data range from 0% (Hodgkin lymphoma, myeloma, non-Hodgkin lymphoma, oesophagus, stomach and thyroid cancer in females) to 0.5% in females (brain cancer) (Difference B).

This indicates that small changes to the underlying data as a result late registration of cancer diagnoses do not substantially impact the survival estimates, but the choice of age-weights does substantially impact the survival estimates.

The full extent of the differences of the previously published estimates and the estimates obtained from using current cancer registration data and the ICSS weights are shown in Table A3 (1-year survival), Table A4 (5-year survival) and Table A5 (10-year survival) in the appendix.

Why does the change in age-weighting systems result in these differences in the survival estimates?

As a result of adopting the ICSS weighting, when we consider the combinations of cancer site and gender that form the top 24 cancers by incidence, 31 cancer sites had an absolute difference of less than 2%, 9 cancer sites had an absolute difference between 2% and 5%, and 3 cancer sites had an absolute difference greater than 5% (Table 2, Difference A).

The cancer sites that had absolute differences greater than 5% are ovary (females) and brain (males, females). These are cancer sites with age profiles that are substantially different to the common profile of increasing incidence by age. The differences are similar for 5-year survival estimates, with the same sites continuing to have absolute differences greater than 5% with the addition of cervix (females) (Table A4).

The observed differences may partially be explained by the categorisation of age-groups. The new analysis uses five age-groups and weights set out by the ICSS, whereas the previous analyses used six age-groups and weights provided in Table A1 (Coleman M 1999). Consequently, the age-group boundaries differ, as does the relative amount of weighting given to younger and older patients.

The difference in the number of age-groups means that the age-weightings are not the same and therefore results are not directly comparable. However, when we examine the number of patients from each age-group diagnosed in recent years and compare it with the weight applied to the corresponding age-group (see Table A3), we observe that:

(a) for ovarian cancer, the ICSS weighting gives approximately the same weighting as the incidence proportion except for the final age-group where a higher weighting is given by the ICSS weighting system;

(b) for testicular cancer, the ICSS weighting gives a higher weighting in those aged 55 and over even though the incidence proportion is substantially lower for these age-groups in relative terms, whereas the previous weighting system gives approximately the same weighting as the incidence proportion; and

(c) for brain cancer, the ICSS weighting gives more weight to the youngest age-group and less weight to the eldest age-group in contrast to the incidence proportion which is the other way around, whereas the previous weighting system gives more weight to the middle age-groups and very low weight to the eldest age-group.

The 2016 published data provided 10-year survival estimates for bladder (male), breast, colorectal, kidney, melanoma, prostate, non-Hodgkin lymphoma, leukaemia and colon (male). Not every combination of sex and cancer type produces robust estimates of 10-year survival, which is why this group is a selection of those forming the main body of the publication. Of the comparable survival estimates that were provided (N=13), 7 were within 5% absolute difference and 6 had an absolute difference greater than 5%. The sites that had a difference greater than 5% were breast cancer (negative 7.3%), kidney cancer (negative 10.4% males, negative 11.7% females), leukaemia (negative 6.7% males, negative 6.8% females), non-Hodgkin lymphoma (negative 10.5%) (Table A5).

Back to table of contents

5. Childhood cancer survival

What methodological changes are being implemented?

Current and historical estimates use the same methodology.

The methodology for children’s cancers is less complicated to implement than for adult cancers. This is due to the childhood methodology taking an overall survival approach. In adults, the net survival indicator is used in order to compensate for mortality from other causes, which may be considerable. For children with cancer, by contrast, overall survival is considered a reliable estimator of cancer survival because, unlike in adults, death within 5 years of diagnosis is almost always due to the cancer. Therefore, no life tables are used to adjust for background mortality. Overall survival for children aged 0 to 14 are calculated using the Kaplan-Meier method (Kaplan EL 1958).

When re-estimating previously published estimates we can observe minor changes; these have resulted from using updated cancer registration data, rather than methodological changes in the way data are analysed (Figure 1). These estimates are compared in Tables A8 (5-year, age-standardised, unsmoothed childhood cancer survival) and A9 (10-year, age-standardised, unsmoothed childhood cancer survival) in Appendix 3.

Back to table of contents

6. Which publications will be impacted?

To date ONS has published a suite of cancer survival publications, summarised below:

  • Cancer survival in England: Adults

  • Geographic patterns of cancer survival

  • Index of cancer survival for Clinical Commissioning Groups

  • Childhood cancer survival

We have also published in partnership with PHE:

  • Cancer survival by stage at diagnosis for England

The method changes outlined in this paper will only impact the Cancer survival in England: Adults publication. Our intention is to rationalise the number of publications. Therefore, we intend to bring all the national level estimates together into one publication which will first be published on 29th June 2017. This publication will contain the estimates for:

  • Adult cancer survival for England

  • Cancer survival by stage

  • Childhood cancer survival for England

The combined publication will contain the same analysis as previously published in the historic publications, but by bringing them together the results are more timely and it will be easier for users to find and compare the analyses.

Back to table of contents

7. References

Brenner H, Rachet B. 2004. “Hybrid analysis for up - to - date long - term survival rates in cancer registries with delayed recording of incident cases.” European Journal of Cancer 40: 2494 -2501.

Coleman M, et al. 1999. Cancer Survival Trends in England and Wales 1971 - 1995 Deprivation and NHS Region. Office for National Statistics

Corazziari I, et al. 2004. “Standard cancer patient population for age standardising survival ratios.” European Journal of Cancer 15: 2307-2316.

Kaplan EL, Meier P. 1958. “Nonparametric Estimation from Incomplete Observations.” Journal of the American Statistical Association 53: 457-481.

Li R, et al. 2014. “Control of data quality for population - based cancer survival analysis.” Cancer Epidemio 38: 314-20.

Martos C, et al. 2014. "A proposal on cancer data quality checks: one common procedure for European cancer registries". European Commission, Joint Research Centre.

Pohar Perme M, Stare J, Estève J. 2012. “On estimation in relative survival.” Biometrics 68:113 -20.

Back to table of contents

8. Appendices

Appendix 1: Standardised weight tables

Appendix 2: Adult survival comparisons

Appendix 3: Comparison of childhood estimates

Appendix 4: Definition of Cancers

Back to table of contents

Contact details for this Methodology

You might also be interested in: