Making Cancer Survival Statistics More Relevant for Clinicians, Patients, and the General Public

Making Cancer Survival Statistics More Relevant

for Clinicians, Patients, and the General Public

Paul W Dickman

Department of Medical Epidemiology and Biostatistics,Karolinska Institutet, Stockholm, Sweden

NAACCR Annual ConferencePortland, Oregon

5 June 2012

Overview

Does patient survival have a role in cancer control?

Should we publish survival statistics for patients and theirfamilies?

What measures of cancer patient survival are available and howare they interpreted?

What should a registry include in a report of patient survival?

Why are we publishing survival statistics and who are theintended audience?

Paul Dickman Cancer patient survival NAACCR, 5 June 2012 2

Overview

Does patient survival have a role in cancer control?

Should we publish survival statistics for patients and theirfamilies?

What measures of cancer patient survival are available and howare they interpreted?

What should a registry include in a report of patient survival?

Why are we publishing survival statistics and who are theintended audience?


Measures used in cancer control

The most common measures are incidence, mortality, andsurvival.

By ‘mortality’ we typically mean mortality in the population,whereas ‘survival’ is nothing more than mortality among thepatients (those diagnosed with cancer).

Population-based studies of patient survival provide a measure ofthe effectiveness of the health care system in diagnosing andtreating the cancers that arise in a population.

We should not study any one of these three measures inisolation; in particular we should consider incidence trends wheninterpreting trends in patient survival.


Enstrom and Austin; Science (1977) [1]

the direct data on total survival ratesgiven in Table 1.The same phenomenon holds for indi-

vidual cancer sites even though the sur-vival changes given in Table 2 appear tohave little relationship to concurrentchanges in mortality rates. This is be-cause for most sites the relative changesin survival result from rather small abso-lute changes. For instance, the 29 per-cent relative increase in lung cancer sur-vival is only an absolute increase from 7percent to 9 percent, which is in-significant in view of the inherent errors.So long as incidence and mortality re-main unchanged or change proportion-ately, no genuine change in survival canoccur. This is what has happened forsites such as lung, stomach, rectum, andesophagus. The large increase in theprostatic cancer survival rate has oc-curred because the incidence rate hasincreased markedly while the mortalityrate has declined negligibly (28).An approach advocated by Linden is

helpful in presenting four of the factorsthat should be considered when evaluat-ing progress in cancer control. Figure 1,taken from his work (29), shows fourtrends in cancer of the rectum in whitemales in the United States: the incidencerate, the mortality rate, the 5-year surviv-al rate, and the percentage of cancercases that were localized at the time ofdiagnosis. The same four measurementsare also shown for cancer of the lung.Both sets of data illustrate the burden ofmortality and its close correspondenceto the rate of occurrence. Both also showthe close relationship between the surviv-al rate and the percentage localized atdiagnosis. And for both these sites therisk of getting or dying from cancer haschanged independently of the 5-year sur-vival rates. Whether these changes couldbe taken as measures of "progress" de-pends upon the correct attribution oftheir cause.The fact that mortality rates and in-

cidence rates are highly correlated whileneither is related to survival rates mightlead one to suspect that if progress incancer control is related to how manypeople get cancer and how many peopledie from it, then survival rates are not a

sensitive measure. The suspicion isstrengthened by a different phenomenon,demonstrated in Table 3, in which theEREP 3-year relative survival rates forleukemia for 1955-64 and 1965-69 (21)are summarized by morphologic classifi-cation. Although significant increases inleukemia survival are often cited as evi-dence of great progress in cancer con-trol, and though there have in fact beensurvival improvements for essentially allforms of leukemia, the survival rate forleukemia with all forms combined hasremained constant. This is due to theincreasing proportions of the highly fatalforms of leukemia. If one uses the differ-entiated survival rates as the criterion ofprogress, then in this instance progresshas been negated by a changing in-cidence.

In summary, survival rates should notbe used as a sole or primary measure ofprogress in cancer control, because fac-tors unrelated to the efficacy of treat-ment play an important role in the deter-mination of those rates and their trends.If cancer control is related to how manypeople get and die from cancer, thenprogress can better be measured by theuse of incidence and mortality rates.

References and Notes

1. D. S. Greenberg, Columbia Journalism Rev. 13,40 (1975); F. J. Rauscher, Congr. Rec. 121,S3463 (7 March 1975); G. R. Newell, Ca 25, 338(1975).

2. D. S. Greenberg, N. Engl. J. Med. 292, 707(1975).

3. S. J. Cutler, M. H. Myers, S. B. Green, ibid.293, 122 (1975).

4. E. Frei, III, ibid., p. 146.5. National Cancer Program: The Strategic Plan,DHEW Publ. No. (NIH) 74-569 (GovernmentPrinting Office, Washington, D.C., 1973).

6. National Cancer Program: Digest of ScientificRecommendations for the National Cancer Pro-gram Plan, DHEW Publ. No. (NIH) 74-570(Government Printing Office, Washington,D.C., 1973).

7. National Cancer Program: The OperationalPlan, DHEW Publ. No. (NIH) 75-777 (Govern-ment Printing Office, Washington, D.C., 1974).

8. Invasion is the process of infiltration of a malig-nant neoplasm into adjacent normal tissues. Me-tastasis is the seeding of a malignant neoplasmto some site distant from its organ of origin. Forfurther background, see J. Cairns, Sci. Am. 233(No. 5), 64 (1975).

9. The proportion of cases diagnosed in each stagevaries from hospital to hospit.l and changesover time depending on the popuiation served,prevailing medical practice, availability of diag-nostic facilities, national news items, and localpolitical issues.

10. M. Merrell and L. E. Shulman,J. Chron. Dis. 1,12 (1955).

11. F. Ederer, L. M. Axtell, S. J. Culter, Natl.CancerInst. Monogr. 6, 101 (1961).

12. G. Linden,Am. J. Public Health 59, 267 (1969).13. , K. U. Bragg, L. Breslow, Cancer Regis-

traton and Survival in California (CaliforniaState Department of Public Health, Berkeley,1963).

14. M. B. Shimkin, in The Physiopathology ofCan-cer, F. Homburger, Ed. (Hoeber-Harper, NewYork, ed. 2, 1959), pp. 855-871.

15. , Public Health Rep. 79, 1 (1964).16. , Proceedings of the Third National Can-

cer Conference, Detroit, 1956 (Lippincott, Phila-delphia, 1957), pp. 835-838.

17. S. J. Cutler and H. B. Latourette, J. Natl.Cancer Inst. 22, 633 (1959).

18. S. J. Cutler and F. Ederer, Eds., End Results inCancer, Report No. 1 [Natl. Cancer Inst.Monogr. 6, 1(1961)1.

19. S. J. Cutler et al., Eds., End Results in Cancer,Report No. 2, PHS Publ. No. 1149 (Govern-ment Printing Office, Washington, D.C., 1964);L. M. Axtell, M. H. Myers, E. M. Shambaugh,Eds., Treatment and Survival Patterns forBlack and White Cancer Patients Diagnosed1955 through 1964, DHEW Publ. No. (NIH)75-712 (Government Printing Office, Washing-ton, D.C., 1975).

20. S. J. Cutler et al., Eds., End Results in Cancer,Report No. 3, NIH Publ. No. 30 (Govern-ment Printing Office, Washington, D.C., 1968).

21. L. M. Axtell, S. J. Cutler, M. H. Myers, Eds.,End Results in Cancer, Report No. 4, DHEWPubl. No. (NIH) 73-272 (Government PrintingOffice, Washington, D.C., 1972).

22. L. M. Axtell and M. H. Myers, Eds., RecentTrends in Survival of Cancer Patients, 1960-1971, DHEW PubI. No. (NIH) 75-767 (Govern-ment Printing Office, Washington, D.C., 1974);End Results Evaluation Program, NationalCancer Institute, Washington, D.C., 1975, un-published data.

23. G. Linden and K. Bragg, Cancer Mortality,Morbidity and Survival in California (CaliforniaState Department of Public Health, Berkeley,1968).

24. California Tumor Registry, 1974, unpublisheddata.

25. J. W. Berg, R. Ross, H. B. Latourette, Cancer(Philadelphia) 39, 265 (1977).

26. "Trends in cancer incidence," Stat. Bull. Met-rop. Life Insur. Co. 56, 2 (October 1975); E.Silverberg and A. I. Holleb, Ca 25, 2 (1975); H.Seidman et al., ibid. 26, 2 (1976).

27. S. J. Cutler and S. S. Davesa, in Host Environ-ment Interactions on the Etiology of Cancer inMan, R. Doll and I. Vodopija, Eds. (Inter-national Agency for Research on Cancer,Lyons, 1973), pp. 15-34.

28. This phenomenon is most likely due to the factthat most prostatic cancer is 'occult" and isrelatively benign in its course. The increasingrate of autopsies and operative procedures forother causes makes for an increase in the num-ber of these occult tumors that are found, lead-ing to an increased recorded incidence. Thosediscovered incidentally in patients being treatedfor other conditions are also followed, and theserelatively slow-growing malignancies contributeto an increased survival rate. The effect of theincreasing rate of discovery of occult prostaticcancer has caused the stage distribution (in situ,localized, regional, distant, and unknown) ofprostatic cancers reported to the CTR to changefrom 0, 41, 16, 28, and 15 percent, respectively,for the years 1942 to 1954, to 0, 63, 13, 19, and 5percent, respectively, in 1955 to 1969.

29. G. Linden, History of Cancer Control Project,NCI contract No. NOI-CN-55172, personal com-munication, 1976.

30. S. J. Cutler, Natl. Cancer Inst. Monogr. 15, 1(1964).

31. We thank Morton A. Klein of the Department ofHealth, Education, and Welfare for initiatingour interest in this subject and for giving manyuseful comments and suggestions. We alsothank George Linden for providing us with un-published data.

4 MARCH 1977 851

on

May

16,

201

2w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from


100120

Lung cancer incidence, mortality and survival (age-standardised) England, 1982-2008, by sex

90

100

110

r yea

r

Mortality

Incidence

70

80

80

90

al (%

)

100,

000

per Mortality

50

60

60

70

ve s

urvi

va

y ra

te p

er 1

Men

30

40

40

50

ear r

elat

iv

or m

orta

lity Men

Women

20

30

20

30

Five

-ye

ncid

ence

o Women

0

10

0

10

1982 1986 1990 1994 1998 2002 2006

In

Survival WomenMen

1982 1986 1990 1994 1998 2002 2006

Year of death / Year of diagnosis


All-cause mortality for males with colon cancer and

for the Finnish male population

0

10000

20000

30000

40000

50000

60000

70000

80000

Mor

talit

y R

ate

(per

100

,000

per

son

year

s)

20 40 60 80 100Age

General populationColon cancer patients


Relative survival

As an alternative to cause-specific mortality we can estimateexcess mortality: the difference between observed (all-cause) andexpected mortality.

excess = observed − expectedmortality mortality mortality

Relative survival is the survival analog of excess mortality — therelative survival ratio is defined as the observed survival in thepatient group divided by the expected survival of a comparablegroup from the general population.

relative survival ratio =observed survival proportion

expected survival proportion


Relative merits of relative survival

A major advantage of relative survival (excess mortality) is thatinformation on cause of death is not required, therebycircumventing problems with the inaccuracy [2] or nonavailabilityof death certificates.

Provides a measure of the excess mortality experienced bypatients with cancer, irrespective of whether the excess mortalityis directly or indirectly (e.g., treatment complications)attributable to the cancer.

Problematic if patients are not representative of the generalpopulation (see Howlader et al. [3]).


Should we use relative survival or cause-specific

survival to estimate net survival?

Both methods involve assumptions specific to the approach:

Cause-specific Accurate classification of cause-of-deathRelative Appropriate estimation of expected survival

We choose the approach for which we have the strongest beliefin the underlying assumptions.

For population-based studies this is typically relative survival butevery study must be evaluated on its specific merits.


Commonly reported survival statistics

In a hypothetical world where breast cancer is the only possiblecause of death, 70% of women diagnosed with breast cancerwould survive 5 years or more.

If the age distribution in Oregon was the same as theInternational Cancer Survival Standard (ICSS) then the 5-yearrelative survival of patients diagnosed in Oregon would be 75%.

Among patients diagnosed 10 years ago, 10-year relative survivalwas 65%.












Should we publish survival statistics aimed at

patients?

Many patients devote considerable time learning about theirdisease, including prognosis.

Ideally, the information they consume should be discussed withtheir physician.

In practice, this is not always the case.

Half of the visitors to the NCI web pages are members of thegeneral public.


Crude and Net Probabilities

Net Probabilityof Death

Due to Cancer=

Probability of death due to cancerin a hypothetical world, where thecancer under study is the only

possible cause of death

Crude Probabilityof Death

Due to Cancer=

Probability of death due to cancerin the real world, where you may die

of other causes before thecancer kills you

In competing risks, the crude probability as a function of time isknown as the cumulative incidence function.


Net (left) and crude (right) probabilities of death in men with localized

prostate cancer aged 70+ at diagnosis (Cronin and Feuer [4])


Natural frequencies presented using infographics


Cancer Survival Query System (Rocky Feuer)


Later at this meeting

09 Estimating Probability of Death for Cancer Patients: In Presenceof Competing RisksNadia Howlader, NCI

27 How Can We Make Cancer Survival Statistics More Useful forPatients and Clinicians — An Application Using LocalizedProstate Cancer in Sweden.Sandra Eloranta, Karolinska Institutet


‘Avoidable’ premature deaths

It is difficult to grasp the implications of one population having,for example, a relative survival that is 5 percent units lower thananother.

‘Avoidable’ deaths attempts to answer the question,

How many deaths would be avoided if we could reduce theexcess mortality rate in one group to that observed inanother group?

The approach has been used to understand the importance ofsurvival differences between countries [5, 6, 7], betweengeographic regions within a county [8], and betweensocioeconomic groups [8, 9].


The concept of avoidable premature deaths [5]

0

2000

4000

6000

8000

10,000

12,000

14,000

16,000

Annual no. of deaths within 5 years

of diagnosis

Excess

Expected

Avoidable

Total

Expected

27%

73%

Excess deaths reduced to zero – theoretical level if allcancer patients were cured

Excess deaths reduced to thelevel seen in a population withhigher survival

Excess deaths among cancerpatients in Britain

Figure 1 Partition of the annual number of deaths in cancer patientswithin 5 years of diagnosis into the number expected from backgroundmortality and the excess deaths (attributable to cancer), showing theproportion of the excess deaths that would be avoidable (27%) if relativesurvival had reached the higher level seen in a comparator population.Note: numbers are of deaths occurring in cancer patients, not deathscertified as due to the cancer in question (see text).


Quantifying differences in breast cancer survival between England and Norway

Paul C. Lambert a,b,*, Lars Holmberg c, Fredrik Sandin d, Freddie Bray e,f, Karen M. Linklater g,Arnie Purushotham h, David Robinson g, Henrik Møller g

a Centre for Biostatistics and Genetic Epidemiology, Department of Health Sciences, 2nd Floor Adrian Building, University of Leicester, University Road,

Leicester LE1 7RH, United Kingdomb Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, PO Box 281, SE-171 77 Stockholm, Swedenc King’s College London, School of Medicine, Cancer Epidemiology Unit, 3rd Floor Bermondsey Wing, Guy’s Hospital, Great Maze Pond, London SE1 9RT, United Kingdomd Regional Oncological Centre, Uppsala/Orebro Region, University Hospital, 751 85 Uppsala, Swedene Department of Clinical and Registry-based Research, Cancer Registry of Norway, Postboks 5313 Majorstuen, 0304 Oslo, Norwayf International Agency for Research on Cancer, Lyon, Franceg King’s College London, School of Medicine, Division of Cancer Studies, Thames Cancer Registry, 1st Floor Capital House, 42 Weston Street, London SE1 3QD, United Kingdomh King’s College London, Department of Research Oncology, 3rd Floor Bermondsey Wing, Guy’s Hospital, Great Maze Pond, London SE1 9RT, United Kingdom

1. Introduction

A number of studies have indicated that survival from breastcancer is lower in the UK compared to other European countries[1,2]. These studies have been criticized due to potential biases,including incomplete collection of incident cases and incompleteascertainment of death [3]. Cancer survival in the UK is mostappropriately compared with countries with high quality nationalcancer registries such as the Nordic countries [3,4]. Here we report

a comparison of breast cancer survival in England and Norway withemphasis on mortality differences at different points in time sincediagnosis and with age at diagnosis. Much of the survivalvariability between European countries can be explained byfactors operating within the first year or so after diagnosis [5,6].Relative survival tends to be lower as age at diagnosis increases [7].Nearly one third of women diagnosed with breast cancer are overthe age of 70 years; these women have been observed to be lesslikely to receive standard therapy [8,9].

International differences in cancer survival could be due to (i)differences in disease and patient characteristics (e.g. stage ofdisease and co-morbidity), (ii), differences in availability and use oftherapy and care and (iii) bias due to systematic differences in datacollection [4]. We have selected two countries, England andNorway, with complete geographical coverage via population-based cancer registration systems and a robust means ofascertaining the deaths that occur among those registered. Thecountries have similar population age distributions and lifeexpectancies and have publicly financed health care systems with

Cancer Epidemiology 35 (2011) 526–533

A R T I C L E I N F O

Article history:

Received 22 February 2011

Received in revised form 10 April 2011

Accepted 15 April 2011

Available online 24 May 2011

Keywords:

Breast Cancer

Survival

Relative survival

England

Norway

Crude mortality

A B S T R A C T

Background: Survival from breast cancer is lower in the UK than in some other European countries. We

compared survival in England and Norway by age and time from diagnosis. Methods: We included

303,648 English and 24,919 Norwegian cases of breast cancer diagnosed 1996–2004 using flexible

parametric relative survival models, enabling improved quantification of differences in survival. Crude

probabilities were estimated to partition the probability of death due to all causes into that due to cancer

and other causes and to estimate the number of ‘‘avoidable’’ deaths. Results: England had lower relative

survival for all ages with the difference increasing with age. Much of the difference was due to higher

excess mortality in England in the first few months after diagnosis. Older patients had a higher

proportion of deaths due to other causes. At 5 years post diagnosis, a woman aged 85 in England had

probabilities of 0.35 of dying of cancer and 0.32 of dying of other causes, whilst in Norway they were 0.26

and 0.35. By eight years the number of ‘‘avoidable’’ all-cause deaths in England was 1020 with the

number of ‘‘avoidable’’ breast cancer related deaths 1488. Conclusion: Lower breast cancer survival in

England is mainly due to higher mortality in the first year after diagnosis. Crude probabilities aid our

understanding of the impact of disease on individual patients and help assess different treatment

options.

Crown Copyright � 2011 Published by Elsevier Ltd. All rights reserved.

* Corresponding author at: Centre for Biostatistics and Genetic Epidemiology,

Department of Health Sciences, 2nd Floor Adrian Building, University of Leicester,

University Road, Leicester LE1 7RH, United Kingdom. Tel.: +44 116 229 7265;

fax: +44 116 229 7250.

E-mail addresses: [email protected] (P.C. Lambert), [email protected]

(L. Holmberg), [email protected] (F. Sandin), [email protected] (F. Bray),

[email protected] (K.M. Linklater), [email protected]

(A. Purushotham), [email protected] (D. Robinson), [email protected]

(H. Møller).

Contents lists available at ScienceDirect

Cancer EpidemiologyThe International Journal of Cancer Epidemiology, Detection, and Prevention

jou r nal h o mep age: w ww.c an cer ep idem io log y.n et

1877-7821/$ – see front matter . Crown Copyright � 2011 Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.canep.2011.04.003


Estimates are heavily dependent on follow-up time

0

100

200

300

400

500A

void

able

Dea

ths

0 2 4 6 8

Years since Diagnosis

All Cause Deaths

Cancer Related Deaths

Avoidable breast cancer deaths in England if Norway’s relativesurvival ratios are applied. Only those aged 80+ are included.


Application using crude probabilities

0

5,000

10,000

15,000

Num

ber

of D

eath

s

0 2 4 6 8Years from Diagnosis

All Cause

Deaths due to cancer

Deaths due to other causes

England Norway Rates


Loss in Expectation of Life (Therese Andersson)

Loss in Expectation of Life

0.0

0.2

0.4

0.6

0.8

1.0S

urvi

val

Years from Diagnosis

Expected SurvivalObserved Survival


Loss in Expectation of Life (Therese Andersson)

0.0

0.2

0.4

0.6

0.8

1.0S

urvi

val

Years from Diagnosis

Expected SurvivalObserved Survival


What should one include in a report of cancer

patient survival?

Relative survival or cause-specific survival?

All-age summary or age-specific estimates?

Period, cohort, or both?

Methodological approach to age standardisation?

Which standard population(s)?

Net survival or crude probabilities of death?

Other measures (e.g., avoidable deaths, proportion of expectedlife lost, conditional survival)?

Inclusion criteria?


References

[1] Enstrom JE, Austin DF. Interpreting cancer survival rates. Science 1977;195:847–851.

[2] Percy CL, Stanek E, Gloeckler L. Accuracy of cancer death certificates and its effect oncancer mortality statistics. American Journal of Public Health 1981;71:242–250.

[3] Howlader N, Ries LAG, Mariotto AB, Reichman ME, Ruhl J, Cronin KA. Improvedestimates of cancer-specific survival rates from population-based data. J Natl Cancer Inst2010;102:1584–1598.

[4] Cronin KA, Feuer EJ. Cumulative cause-specific mortality for cancer patients in thepresence of other causes: a crude analogue of relative survival. Statistics in Medicine 2000;19:1729–1740.

[5] Abdel-Rahman M, Stockton D, Rachet B, Hakulinen T, Coleman MP. What if cancersurvival in Britain were the same as in Europe: how many deaths are avoidable? BritishJournal of Cancer 2009;101 Suppl 2:S115–S124.

[6] Møller H, Sandin F, Bray F, Klint A, Linklater KM, Purushotham A, et al.. Breast cancersurvival in England, Norway and Sweden: a population-based comparison. InternationalJournal of Cancer 2010;127:2630–2638.

[7] Lambert PC, Holmberg L, Sandin F, Bray F, Linklater KM, Purushotham A, et al..Quantifying differences in breast cancer survival between England and Norway. CancerEpidemiology 2011;35:526–533.


References 2

[8] Dickman PW, Gibberd RW, Hakulinen T. Estimating potential savings in cancer deaths byeliminating regional and social class variation in cancer survival in the Nordic countries. JEpidemiol Community Health 1997;51:289–298.

[9] Pokhrel A, Martikainen P, Pukkala E, Rautalahti M, Seppa K, Hakulinen T. Education,survival and avoidable deaths in cancer patients in Finland. British Journal of Cancer 2010;103:1109–1114.
