incidence, mortality and survival - cancer institute nsw · survival by aboriginal status and major...

Incidence, mortality and survival

CANCER IN NSW ABORIGINAL PEOPLES

September 2012

C A N C E R I N N S W A B O R I G I N A L P E O P L E S

Contents

Tables 1Figures 2

Message from the Chief Cancer Officer 4

Executive summary 5

Introduction 6

Results 7Cancer incidence by Aboriginal status 7

Incidence of major cancers from multiple imputation 8

Mortality from major cancers 21

Age-specific incidence and mortality from major cancers 30

All cancers 30

Major cancers 31

Degree of spread by Aboriginal status for major cancers 34

Survival by Aboriginal status and major cancer type 49

All cancers 50

Survival by major cancers 51

Methods 65Cancer Registration in NSW 65

Sources of Aboriginal and/or Torres Strait Islander Status recording in the CCR 66

Missing data 67

Complete case analysis 67

Multiple imputation (MI) 68

Comparison of analyses with and without multiple imputation 71

Age standardisation 72

Survival analysis 72

Conclusions 73

Appendix 74Appendix 1: Cancer incidence and mortality to the world population 74

Appendix 2: Glossary 83

Appendix 3: Abbreviations 84

Appendix 4: References 85

National Library of Australia Cataloguing–in–Publication data:Cancer in NSW Aboriginal peoples: incidence, mortality and survival

SHPN (CI) 120185ISBN 978-1-74187-724-3 Key words: Cancer, Aboriginal, New South Wales, Australia.

Suggested citation:Cancer in NSW Aboriginal peoples: incidence, mortality and survival Sydney: Cancer Institute NSW, 2012.

Cancer Institute NSWPO Box 41Alexandria NSW 1435

Telephone (02) 8374 5600Facsimile (02) 8374 5700

Email [email protected] www.cancerinstitute.org.au

Copyright © Cancer Institute NSW 2012.

This work is copyright. It may be reproduced in whole orpart for study or training purposes subject to the inclusion ofacknowledgement of the source. It may not be reproducedfor commercial usage or sale. Reproduction for purposesother than those indicated above requires written permissionfrom the Cancer Institute NSW.

I N C I D E N C E , M O RTA L I T Y A N D S U RV I VA L

1

Tables

Table 1: Cancer incidence in Aboriginal males NSW, 1999–2007 11

Table 2: Cancer incidence in non-Aboriginal males, NSW, 1999–2007 12

Table 3: Cancer incidence in Aboriginal females, NSW, 1999–2007 13

Table 4: Cancer incidence in non-Aboriginal females, NSW, 1999–2007 14

Table 5: Cancer incidence in Aboriginal persons, NSW, 1999–2007 15

Table 6: Cancer incidence in non-Aboriginal persons, NSW, 1999–2007 16

Table 7: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal people compared to NSW, 1999–2007 19

Table 8: Cancer mortality in Aboriginal and Non-Aboriginal males, NSW, 1999-2007 24

Table 9: Cancer mortality in Aboriginal and Non-Aboriginal females, NSW, 1999–2007 25

Table 10: Cancer mortality in Aboriginal and Non-Aboriginal persons, NSW, 1999–2007 26

Table 11: Standardised mortality ratios, Aboriginal peoples compared to NSW, 1999–2007 29

Table 12: Age-standardised incidence rates by degree of spread at diagnosis, males, NSW, 1999–2007 43

Table 13: Age-standardised incidence rates by degree of spread at diagnosis, females, NSW, 1999–2007 44

Table 14: Age-standardised mortality rates by degree of spread at diagnosis, males, NSW, 1999–2007 45

Table 15: Age-standardised mortality rates by degree of spread at diagnosis, females, NSW, 1999–2007 46

Table 16: Standardised incidence ratios by degree of spread, Aboriginal people compared to NSW, 1999–2007 47

Table 17: Standardised mortality ratios by degree of spread, Aboriginal people compared to NSW, 1999–2007 48

Table 18: Percentage surviving all cancers, by Aboriginal status and years since diagnosis 50

Table 19: Percentage surviving head and neck cancer, by Aboriginal status and years since diagnosis 51

Table 20: Percentage surviving oesophageal cancer, by Aboriginal status and years since diagnosis 52

Table 21: Percentage surviving stomach cancer, by Aboriginal status and years since diagnosis 53

Table 22: Percentage surviving large bowel cancer, by Aboriginal status and years since diagnosis 54

Table 23: Percentage surviving liver cancer, by Aboriginal status and years since diagnosis 55

Table 24: Percentage surviving pancreatic cancer, by Aboriginal status and years since diagnosis 56

Table 25: Percentage surviving lung cancer, by Aboriginal status and years since diagnosis 57

Table 26: Percentage surviving melanoma, by Aboriginal status and years since diagnosis 58

Table 27: Percentage surviving breast cancer, by Aboriginal status and years since diagnosis 59

Table 28: Percentage surviving cervical cancer, by Aboriginal status and years since diagnosis 60

Table 29: Percentage surviving uterine cancer, by Aboriginal status and years since diagnosis 61

Table 30: Percentage surviving ovarian cancer, by Aboriginal status and years since diagnosis 62

Table 31: Percentage of surviving prostate cancer, by Aboriginal status and years since diagnosis 63

Table 32: Percentage of surviving kidney cancer, by Aboriginal status and years since diagnosis 64

Table 33: Predictors used in multiple imputation model 69

Table 34: Odds-ratio estimates from logistic regression modelling of covariates used to inform imputation of Aboriginal status 70

Table 35: Comparison of estimated incidence from complete case and multiple imputation approaches, NSW female breast cancer, 1999–2007 71 Table A1: Cancer incidence in Aboriginal males NSW, 1999–2007 74

Table A2: Cancer incidence in Non-Aboriginal males NSW, 1999–2007 75

Table A3: Cancer incidence in Aboriginal females NSW, 1999–2007 76

Table A4: Cancer incidence in Non-Aboriginal females NSW, 1999–2007 77

Table A5: Cancer incidence in Aboriginal persons NSW, 1999–2007 78

Table A6: Cancer incidence in Non-Aboriginal persons NSW, 1999–2007 79

Table A7: Cancer mortality in Aboriginal and Non-Aboriginal males NSW, 1999–2007 80

Table A8: Cancer mortality in Aboriginal and Non-Aboriginal females NSW, 1999–2007 81

Table A9: Cancer mortality in Aboriginal and Non-Aboriginal persons NSW, 1999–2007 82


2

Figures

Figure 1: Age-standardised incidence rates for all cancers, by multiple imputation (MI) and complete case analysis, Aboriginal and non-Aboriginal people, NSW, 1999–2007 7

Figure 2: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal males, NSW, 1999–2007 9

Figure 3: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal females, NSW, 1999–2007 9

Figure 4: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal persons, NSW, 1999–2007 10

Figure 5: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal male compared to NSW, 1999–2007 17

Figure 6: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal female compared to NSW, 1999–2007 17

Figure 7: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal people compared to NSW, 1999-2007 18

Figure 8: Directly age-standardised mortality rates, all cancers, Aboriginal and non-Aboriginal people, males and females, NSW, 1999–2007 20

Figure 9: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal males, NSW, 1999–2007 22

Figure 10: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal females, NSW, 1999–2007 22

Figure 11: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal persons, NSW, 1999–2007 23

Figure 12: Standardised mortality ratios, Aboriginal male compared to NSW, 1999–2007 27

Figure 13: Standardised mortality ratios, Aboriginal female compared to NSW, 1999–2007 27

Figure 14: Standardised mortality ratios, Aboriginal people compared to NSW, 1999–2007 28

Figure 15: Age-specific incidence and mortality rates for all cancers, NSW, 1999–2007 30

Figure 16: Age-specific incidence and mortality rates for large bowel cancer, NSW, 1999–2007 31

Figure 17: Age-specific incidence and mortality rates for lung cancer, NSW, 1999–2007 31

Figure 18: Age-specific incidence and mortality rates for melanoma, NSW, 1999–2007 32

Figure 19: Age-specific incidence and mortality rates for breast cancer, NSW, 1999–2007 32

Figure 20: Age-specific incidence and mortality rates for cervical cancer, NSW, 1999–2007 33

Figure 21: Age-specific incidence and mortality rates for prostate cancer, NSW, 1999–2007 33

Figure 22: Age-standardised all-cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 35

Figure 23: Age-standardised all-cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 35

Figure 24: Age-standardised stomach cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 36

Figure 25: Age-standardised stomach cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 36

Figure 26: Age-standardised large bowel cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 37

Figure 27: Age-standardised large bowel cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 37

Figure 28: Age-standardised lung cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 38

Figure 29: Age-standardised lung cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 38

Figure 30: Age-standardised melanoma incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 39

Figure 31: Age-standardised melanoma mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 39

Figure 32: Age-standardised breast cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 40


3

Figure 33: Age-standardised breast cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 40 Figure 34: Age-standardised cervical cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 41

Figure 35: Age-standardised cervical cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 41

Figure 36: Age-standardised prostate cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 42

Figure 37: Age-standardised prostate cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007 42

Figure 38: Kaplan-Meier survival curve for all cancers by Aboriginal status, NSW, 1999-2007 50

Figure 39: Kaplan-Meier survival curve for head and neck cancer by Aboriginal status, NSW, 1999–2007 51

Figure 40: Kaplan-Meier survival curve for oesophageal cancer by Aboriginal status, NSW, 1999–2007 52

Figure 41: Kaplan-Meier survival curve for stomach cancer by Aboriginal status, NSW, 1999–2007 53

Figure 42: Kaplan-Meier survival curve for large bowel cancer by Aboriginal status, NSW, 1999–2007 54

Figure 43: Kaplan-Meier survival curve for liver cancer by Aboriginal status, NSW, 1999–2007 55

Figure 44: Kaplan-Meier survival curve for pancreatic cancer by Aboriginal status, NSW, 1999–2007 56

Figure 45: Kaplan-Meier survival curve for lung cancer by Aboriginal status, NSW, 1999–2007 57

Figure 46: MI Kaplan-Meier survival curve for melanoma by Aboriginal status, NSW, 1999–2007 58

Figure 47: Kaplan-Meier survival curve for female breast cancer by Aboriginal status, NSW, 1999–2007 59

Figure 48: MI Kaplan-Meier survival curve for cervical cancer by Aboriginal status, NSW, 1999–2007 60

Figure 49: MI Kaplan-Meier survival curve for uterine cancer by Aboriginal status, NSW, 1999–2007 61

Figure 50: MI Kaplan-Meier survival curve for ovarian cancer by Aboriginal status, Females, NSW, 1999–2007 62

Figure 51: Kaplan-Meier survival curve for prostate cancer by Aboriginal status, Males, NSW, 1999–2007 63

Figure 52: Kaplan-Meier survival curve for kidney cancer by Aboriginal status, NSW, 1999–2007 64

Figure 53: Comparison of estimated incidence from complete case and multiple imputation approaches, NSW female breast cancer, 1999–2007. 71


4

Message from the Chief Cancer Officer

The health issues experienced by Aboriginal peoples are well documented and include: 12 per cent of the total burden of disease and injury from smoking in Australia; lower rates of access to acute care investigations and procedures; and lower likelihood of being treated for and surviving cancer.*

The high rates of disease and poorer survival from cancer shows us that our priority is to improve the responsiveness of the health system and close the gaps for Aboriginal people with cancer, their carers and their communities.

That’s why improving cancer care and survival in Aboriginal communities across NSW is a key feature of the NSW Cancer Plan 2011–15 and we are already looking at new programs and initiatives that will help to reduce the gap in cancer outcomes in the years ahead.

To be able to improve these outcomes, we need to know more about the impact of cancer on Aboriginal communities. Historically, we haven’t had an accurate reading about cancer in these communities. As a result, the Cancer Institute NSW is working on improving our data recording in NSW through the Central Cancer Registry.

We are now also starting to see more people identify as Aboriginal which, combined with improvements to our Registry, is providing more robust data and a clearer picture of their health outcomes.

From our statistics, we’ve found that much of the excess cancer incidence and mortality in Aboriginal peoples is the result of higher incidence of lung, and head and neck cancers in men and women; stomach and kidney cancers in men; and uterine and cervical cancer in women. Clearly, as most of these cancers are preventable, targeted programs potentially can mean better outcomes for Aboriginal peoples at risk of, and with, cancer.

Our collaboration with Aboriginal peak bodies, stakeholders and patients in setting the goals for the Cancer Plan has been an eye-opening experience and has presented an insight into the history, the conflict and triumphs of Aboriginal peoples in Australia. We have also identified some of the language; resources and tools we can use to more effectively reach Aboriginal peoples and ultimately improve their health outcomes into the future.

David Currow

Chief Cancer Officer and CEO

Cancer Institute NSW

*http://www.coag.gov.au/intergov_agreements/federal_financial_relations/docs/national_partnership/NP_closing_the_Gap_indigenous_health_outcomes.pdf


5

Executive summary

It is essential to have accurate and complete collection and recording of health information on Aboriginal and Torres Strait Islander people if we are to make informed decisions about future service provision. Historically, health information about Aboriginal peoples, including cancer statistics, has been limited due to under-recording of Aboriginal status in administrative health data collections.

Over the past decade or so, a range of initiatives have been undertaken to improve the recording of Aboriginal status and much of this work has now flowed on to cancer statistics.

Mortality data from 1994 onwards is considered of sufficient quality to use as recorded by the NSW Central Cancer Registry (CCR). However, in order to analyse cancer incidence and survival, and minimise recording bias in NSW Aboriginal peoples, we used multiple imputation methods for cancer diagnosed from 1999 onwards (see the Methods section in Appendix 1 for more information). This has resulted in an extra 146 and 140 cancer cases in Aboriginal men and women respectively over 1999–2007.

After imputation for unstated Aboriginal status, Aboriginal peoples were found to have 7 per cent higher incidence of cancer than for NSW overall (9 per cent higher in men, 6 per cent higher in women), which was statistically significant. Specific cancers with significantly higher incidence rates in both Aboriginal men and women include head and neck, lung and bronchus, oesophagus and liver. Men also had significantly higher rates of cancer of the stomach, pancreas and unknown primary. Women had significantly higher rates of cervical cancer.

All-cancer mortality was also statistically significantly higher in Aboriginal peoples than for NSW overall, with excess cancer mortality estimated to be 69 per cent overall (68% in males, 73% in females). The differential between excess incidence and mortality indicates lower survival in Aboriginal peoples diagnosed with cancer. Significantly higher male and female Aboriginal cancer mortality was found for lung, head and neck, oesophagus, stomach, liver, pancreas, and kidney. Aboriginal men also had significantly higher mortality from bowel, prostate, and unknown primary. In Aboriginal women, excess mortality occurred from breast, cervical and uterine cancers. Outcomes for Aboriginal men with prostate cancer are particularly poor: despite significantly lower incidence of prostate cancer than NSW overall, Aboriginal men have 86 per cent higher mortality from prostate cancer.

Stage of disease at diagnosis is a major prognostic indicator for cancer outcomes, and differences in degree of spread (the modified CCR staging) at diagnosis between Aboriginal and non-Aboriginal people explain some of the differences in cancer mortality and survival. Overall, Aboriginal peoples have lower incidence rates of localised cancer; and in men this difference is significant. Correspondingly, they have higher rates of regional and distant cancers. Aboriginal peoples also have higher mortality across each degree of spread category – despite having lower incidence of localised cancer, they have higher mortality from localised cancer.

These mortality differentials are reflected in five-year cancer survival for Aboriginal peoples of 53 per cent, compared to 65 per cent for non-Aboriginal people. Five-year survival for Aboriginal men and women is 49 per cent and 57 per cent respectively, compared to 64 per cent and 67 per cent in non-Aboriginal men and women.


6

Introduction

The collection and recording of accurate and complete registration details about all patients and clients receiving publicly provided health services in NSW is vital for making informed decisions about services provided for the population, including Aboriginal and Torres Strait Islander peoples. Historically, statistics on Aboriginal Australians have been limited across a range of health issues due to poor identification of Aboriginal people in the various health and health-related administrative data collections. The reasons are complex and include lack of awareness and training of staff; staff reluctance to ask the question; refusal to answer the question; staff perceptions about Aboriginal people not wanting to disclose their status; perceived lack of privacy when answering the question; minimal checking and validating of the data and follow up of missing information; inadequate data management systems; and a mistrust of health services.1

This issue also applies to cancer statistics, which in the past have been limited because of under-reporting of Aboriginal status in cancer registrations; Aboriginal people diagnosed with cancer are registered, but many of them are not identified as Aboriginal in cancer recording. In the past the cancer registers of the Northern Territory, Western Australia and South Australia have been viewed as having Aboriginal status of sufficient quality to allow publication of Aboriginal cancer incidence and mortality breakdowns, but this has not been the case for cancer registries in other jurisdictions.2, 3 Thus, despite considerable detailed knowledge on cancer epidemiology in the NSW population and the broader Australian population, information on cancer in Aboriginal and Torres Strait Islander people is limited.

In the interests of ensuring that information on the health status of Aboriginal Australians is accurately reported and enables more informed decisions about the planning and delivery of health services, a range of initiatives were undertaken in NSW to assess data quality and improve Aboriginal status recording in administrative health data collections. In the mid-1990s, NSW Health undertook an assessment of collection of Aboriginal status when patients presented to a NSW public hospital or other health care facility. As a result, staff were trained extensively and guidelines were developed to improve the recording of Aboriginal status at NSW Healthcare facilities.4

Consequently, identification of Aboriginal people attending NSW hospitals improved significantly and hospital statistics are now considered of sufficient quality to report by Aboriginal status.

The asking of, and recording of, Aboriginal status is mandatory for the NSW Health data collection systems. Adherence to the NSW Health 2005 policy (a re-issue of a 2002 Policy) on identifying Aboriginal status is mandatory in all facilities and Health Networks. This policy is being reviewed and new projects aimed at improving Aboriginal identification are being planned for implementation in 2011, 2012 and 2013 as part of the National Partnership Agreement Closing the Gap in Indigenous Health Outcomes.5

A 2007 survey undertaken by NSW Health of nearly 3,000 admitted patients in 20 hospitals drawn from metropolitan, inner regional, outer regional and remote locations of NSW found that admitted patient data were 89 per cent correct in relation to Aboriginality, ranging from 82 per cent in metropolitan hospitals to 100 per cent in remote hospitals. This survey was repeated in 2010, and preliminary figures show a result similar to the 2007 survey. (NSW Health, personal communication Dr Kerry Chant).

As a large proportion of notifications for cancer come from the NSW hospital system, the effects of these activities would be expected to flow through to cancer statistics in Aboriginal people. The Cancer Institute NSW has therefore undertaken an assessment of data quality in the NSW Central Cancer Registry to determine which data on Aboriginal people with cancer are now of sufficient quality to report on.

This project has been undertaken in two parts: the first report describes the quality and completeness of the Aboriginal status variable in the NSW Central Cancer Registry, and describes which subsets of the data are suitable for analysis by this variable. Potential biases are discussed as well as methods to minimise them. This second report presents estimates of incidence and mortality of cancer in the NSW Aboriginal population, based on an imputation approach to assigning Aboriginal status to records in which Aboriginal status is unknown.

As a final note, the Aboriginal population, both Aboriginal and Torres Strait Islander People, are referred to in this report as ‘Aboriginal’ in recognition of the fact that Aboriginal people are the original inhabitants of NSW.


7

Results

Cancer incidence by Aboriginal status

For 1999–2007, 146 extra cases of cancer were assigned as Aboriginal by multiple imputation in men and 140 extra cases in women. The total number of cancer cases (excluding non-melanoma skin cancer) estimated by multiple imputation for 1999–2007 in Aboriginal men was 1,347, and in Aboriginal females was estimated to be 1,257. Mean annual incidence of all cancers was estimated by multiple imputation to be 660 per 100,000 compared to 569 per 100,000 in non-Aboriginal men (Figure 1, Table 1) The mean annual incidence rate for Aboriginal women was estimated as 462 per 100,000 person years, compared to 396 per 100,000 in non-Aboriginal women. For both men and women, these incidence rate differences are statistically significant. An Aboriginal man was estimated as having a one in 2.6 risk of a cancer diagnosis from birth to age 75 years, compared to 2.8 in non-Aboriginal men. In Aboriginal women, the risk of cancer was estimated to be 3.5 compared to 3.8 in non-Aboriginal women.

Mean annual cancer incidence for all Aboriginal peoples was estimated to be 545 per 100,000 compared to 472 per 100,000 in non-Aboriginal people (Tables 5 & 6). As expected, the non-imputed estimates for each subgroup were lower, but relative differences were similar, as indicated by Aboriginal : non-Aboriginal incidence ratios of 1.15 and 1.20 from imputed and non-imputed estimates respectively.

Figure 1: Age-standardised incidence rates for all cancers, by multiple imputation (MI) and complete case analysis, Aboriginal and non-Aboriginal people, NSW, 1999-2007*

* Directly standardised to the Australian 2001 population

0

100

200

300

400

500

600

700

800

MI_Males MI_Females MI_Persons Males Females Persons

Per 1

00,0

00

Aboriginal people Non-Aboriginal peopleAboriginal people

Per

100,

000

non-Aboriginal people

MI_males MI_females MI_persons Males Females Persons


8

Incidence of major cancers from multiple imputation

Aboriginal men

During 1999–2007 Aboriginal men had significantly lower incidence of melanoma (34 per 100,000 person years) and non-significantly lower incidence of prostate cancer (145 per 100,000), non-Hodgkin’s lymphoma (17 per 100,000) and brain cancer (5.5 per 100,000) than non-Aboriginal men (Figure 2; Table 1). These represented standardised incidence ratios (SIR) of 0.4, 0.8, 1.0 and 0.7, respectively (compared to NSW overall, Figure 5, Table 7). Cancers with significantly higher incidence in Aboriginal than non-Aboriginal men were: head and neck (38 per 100,000; SIR=2.0), lung and bronchus (111 per 100,000; SIR=1.9) and stomach (26 per 100,000; SIR=1.8). Aboriginal men had non-significantly higher incidence of most other cancers including oesophageal, liver, gall bladder, pancreas, kidney, bladder and cancer of unknown primary (30 per 100,000), mesothelioma (7 per 100,000; SIR=1.5), kidney (24 per 100,000; SIR=1.4), and leukaemia (24 per 100,000; SIR=1.4) (Figure 5). Most other cancers in Aboriginal peoples either had similar incidence rates to non-Aboriginal people or incident numbers were too small to judge with certainty the extent or direction of any differences.

Aboriginal women

Aboriginal women had significantly lower incidence rates of melanoma of the skin (20 per 100,000; SIR=0.5) and thyroid cancer (7 per 100,000; SIR=0.5) than non-Aboriginal women (Figures 3 & 4; Tables 3 & 7). Non-Hodgkin’s lymphoma (11 per 100,000; SIR=0.7) was borderline significantly lower, but the estimate for breast cancer, expected to be somewhat lower, was very similar to that for non-Aboriginal women (116 per 100,000 compared to 113 per 100,000). Significantly higher cancer incidence occurred for cancer of the head and neck (13 per 100,000; SIR=2.0), cervix (19 per 100,000; SIR=2.4) and lung and bronchus (69 per 100,000; SIR=2.4). Liver cancer was borderline significantly higher (6 per 100,000; SIR=2.2), with incidence of most remaining cancers non-significantly higher than in non-Aboriginal women. Lung cancer incidence in female Aboriginal people exceeded that in non-Aboriginal men (cf. Table 2).

Aboriginal peoples

For Aboriginal peoples overall, the lower incidence rates of non-Hodgkin’s lymphoma in males and females becomes statistically significant, and the incidence of stomach, oesophageal, liver, pancreatic and kidney cancers is significantly higher than for non-Aboriginal people. (Table 5 cf. Table 6).

Prostate cancer is the most commonly diagnosed cancer in Aboriginal men (as in non-Aboriginal men), followed by lung cancer, unlike non-Aboriginal men where the second most common is large bowel cancer (Figure 2). In women, the pattern is similar: breast cancer is the most common in both Aboriginal and non-Aboriginal people, but lung cancer is the next most common in Aboriginal women, while large bowel cancer is the next most common in non-Aboriginal women (Figure 3). Lung cancer is the most commonly diagnosed cancer in Aboriginal peoples overall, followed by large bowel, breast and prostate (Figure 4).

In summary, all-cancer incidence in Aboriginal peoples is 7 per cent higher than for NSW overall (statistically significant), but is double or more for Aboriginal women compared to all NSW women for cancers of the cervix, lung, liver, head and neck and oesophagus, despite female all-cancer incidence being 6 per cent higher than NSW (Table 7; Figures 5 & 6). In Aboriginal men, incidence is 70 per cent or higher than all NSW males for cancers of the head and neck, lung, stomach and oesophagus, with 9 per cent higher all-cancer incidence.


9

Figure 2: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal males, NSW, 1999-2007*

0

20

40

60

80

100

120

140

160

180

Pros

tate

Lung

Larg

e bo

wel

Hea

d &

neck

Mel

anom

a of

ski

n

Unk

nown

prim

ary

Stom

ach

Kidn

ey

All l

euka

emia

s

Non

-Hod

gkin

's ly

mph

oma

Per 1

00,0

00

Aboriginal males Non-Aboriginal males


Figure 3: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal females, NSW, 1999-2007*

0

20

40

60

80

100

120

140

Brea

st

Lung

Larg

e bo

wel

Ute

rus

Mel

anom

a of

ski

n

Cer

vix

Unk

nown

prim

ary

Panc

reas

Hea

d &

neck

Kidn

ey

Ova

ry

Per 1

00,0

00

Aboriginal females Non-Aboriginal females


Aboriginal males

Aboriginal females

non-Aboriginal males

non-Aboriginal females

Per

100,

000

Per

100,

000


10

Figure 4: Multiple imputation estimates of incidence of common cancers, Aboriginal and non-Aboriginal persons, NSW, 1999-2007*

0

20

40

60

80

100

120

Lung

Larg

e bo

wel

Brea

st

Pros

tate

Mel

anom

a of

ski

n

Hea

d &

neck

Unk

nown

prim

ary

Kidn

ey

All l

euka

emia

s

Stom

ach

Non

-Hod

gkin

's ly

mph

oma

Per 1

00,0

00

Aboriginal persons Non-Aboriginal persons


Aboriginal persons non-Aboriginal persons

Per

100,

000


11

Table 1: Cancer incidence in Aboriginal males NSW, 1999-2007(a)

Cancer site

Complete case method Missing (%)

Multiple imputation method

Cases ASR(b) 95%CI Risk to age 75 1 in

Cases ASR(b) 95%CI Risk to age 75

1 in

Head & neck 96 35.2 (27.0-44.8) 36 6.4 103 37.9 (28.5-47.2) 34

Oesophagus 27 13.1 (7.8-20.3) 88 2.2 28 13.6 (7.5-19.7) 86

Stomach 46 25.2 (17.1-35.3) 55 3.1 47 25.5 (16.7-34.3) 54

Large bowel 147 71.0 (57.6-86.1) 19 6.4 158 75.4 (60.6-90.3) 17

Liver 29 11.5 (7.1-17.3) 83 2.1 30 11.7 (6.9-16.6) 81

Gallbladder 10 5.0 (1.7-10.3) 417 2.8 10 5.1 (1.1-9.1) 393

Pancreas 32 17.5 (10.7-26.2) 88 1.1 32 17.7 (10.2-25.1) 87

Lung 200 108.3 (91.0-127.5) 13 2.1 206 111.0 (92.6-129.4) 12

Melanoma of skin 37 23.1 (14.9-33.6) 78 41.2 67 33.6 (21.3-45.9) 48

Mesothelioma 11 7.3 (3.1-14.0) 216 1.7 11 7.4 (2.4-12.5) 211

Prostate 183 112.7 (94.4-133.2) 13 20.0 239 144.5 (120.0-169.0) 10

Kidney 50 23.1 (15.6-32.3) 64 9.1 53 24.1 (15.9-32.4) 60

Bladder 30 22.6 (14.0-33.9) 111 5.3 31 23.3 (13.7-33.0) 105

Brain 20 5.3 (2.4-9.3) 273 2.8 21 5.5 (2.2-8.8) 262

Non-Hodgkin's

lymphoma

47 15.1 (10.4-20.8) 69 10.2 53 17.1 (11.4-22.7) 62

Multiple myeloma 17 8.4 (4.2-14.5) 142 6.3 18 9.1 (3.9-14.2) 130

All leukaemias 51 21.9 (14.2-31.3) 78 8.6 56 23.8 (15.0-32.5) 70

Unknown primary 59 29.3 (20.7-39.7) 49 4.7 61 30.1 (20.7-39.5) 47

Myelodysplasia 27 16.8 (9.8-26.2) 129 6.1 29 18.1 (9.7-26.4) 118

All cancers (c) 1,201 595.2 (554.1-638.1) 2.9 13.6 1,347 660.2 (613.7-706.6) 2.6

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


12

Table 2: Cancer incidence in non-Aboriginal males, NSW, 1999-2007(a)

Cancer site

Complete case Method Missing (%)

Multiple Imputation Method


1 in


1 in

Head & neck 5,413 18.6 (18.1-19.1) 63 6.4 5,778 19.8 (19.3-20.3) 59

Oesophagus 2,158 7.6 (7.2-7.9) 179 2.2 2,206 7.7 (7.4-8.0) 174

Stomach 3,586 12.6 (12.2-13.1) 111 3.1 3,699 13.0 (12.6-13.4) 107

Large bowel 20,080 70.2 (69.3-71.2) 19 6.4 21,451 75.0 (74.0-76.0) 17

Liver 2,135 7.4 (7.1-7.8) 175 2.1 2,180 7.6 (7.3-7.9) 172

Gallbladder 716 2.5 (2.3-2.7) 549 2.8 737 2.6 (2.4-2.8) 532

Pancreas 3,194 11.3 (10.9-11.7) 125 1.1 3,231 11.4 (11.0-11.8) 124

Lung 16,422 57.9 (57.0-58.8) 23 2.1 16,776 59.1 (58.2-60.0) 23

Melanoma of skin 10,113 35.4 (34.7-36.1) 40 41.2 17,201 59.7 (58.8-60.6) 23

Mesothelioma 1,408 5.0 (4.7-5.2) 282 1.7 1,433 5.1 (4.8-5.3) 276

Prostate 35,733 124.1 (122.9-125.4) 10 20.0 44,644 155.1 (153.7-156.6) 8

Kidney 4,503 15.6 (15.2-16.1) 81 9.1 4,955 17.2 (16.7-17.7) 73

Bladder 4,808 17.2 (16.7-17.7) 93 5.3 5,078 18.1 (17.6-18.6) 87

Brain 2,311 8.0 (7.6-8.3) 154 2.8 2,376 8.2 (7.9-8.5) 150

Non-Hodgkin’s lymphoma 5,673 19.8 (19.3-20.3) 68 10.2 6,314 22.0 (21.4-22.5) 60

Multiple myeloma 2,041 7.2 (6.9-7.5) 195 6.3 2,179 7.7 (7.3-8.0) 180

All leukaemias 4,525 15.9 (15.5-16.4) 92 8.6 4,947 17.4 (16.9-17.8) 83

Unknown primary 5,534 19.7 (19.1-20.2) 81 4.7 5,809 20.6 (20.1-21.2) 77

Myelodysplasia 2,889 10.4 (10.0-10.7) 170 6.1 3,076 11.0 (10.6-11.4) 157

All cancers (c) 140,586 491.7 (489.1-494.3) 3.2 13.6 162,813 568.5 (565.7-571.2) 2.8

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


13

Table 3: Cancer incidence in Aboriginal females, NSW, 1999-2007(a)

Cancer siteComplete case Method Missing

(%)Multiple Imputation Method


1 in


1 in

Head & neck 36 12.1 (8.1-17.2) 87 8.0 39 13.1 (8.5-17.6) 81

Oesophagus 14 5.7 (2.8-9.9) 207 1.6 14 5.8 (2.4-9.1) 207

Stomach 18 7.9 (4.3-13.0) 181 3.6 19 8.2 (4.0-12.5) 173

Large bowel 110 52.1 (41.7-64.0) 31 7.7 120 56 (44.4-67.6) 29

Liver 15 6.4 (3.2-11.1) 214 1.7 15 6.4 (2.8-10.1) 213

Pancreas 28 14.1 (8.9-21.0) 121 1.5 29 14.4 (8.5-20.2) 119

Lung 162 66.8 (55.8-79.2) 20 3.1 167 68.7 (57.0-80.3) 19

Melanoma of the skin 30 11.4 (7.0-17.2) 148 46.7 62 20.2 (12.9-27.4) 74

Breast 304 102.9 (90.1-116.8) 14 12.2 346 115.7 (101.2-130.3) 12

Cervix 65 17.6 (13.0-23.1) 73 12.4 71 19 (13.7-24.4) 69

Uterus 43 18.8 (13.0-26.0) 75 8.7 50 21 (14.2-27.8) 65

Ovary 36 11.4 (7.5-16.4) 115 4.5 38 12 (7.5-16.4) 109

Kidney 31 12 (7.6-17.8) 114 10.0 34 12.9 (7.8-17.9) 105

Brain 22 6.7 (3.7-10.9) 177 2.5 23 6.8 (3.4-10.2) 174

Thyroid 24 5.5 (3.4-8.5) 199 12.2 28 6.5 (3.7-9.3) 173

Non-Hodgkin’s lymphoma 26 8.9 (5.4-13.6) 119 11.4 33 11.2 (6.4-16.1) 96

All leukaemias 30 9.3 (5.2-14.6) 254 8.3 33 10.3 (5.5-15.0) 203

Unknown primary 44 17.3 (12.0-24.0) 77 4.1 47 18.1 (12.1-24.2) 73

Myelodysplasia 14 6.3 (3.1-11.2) 307 8.2 16 6.9 (2.9-11.0) 272

All cancers (c) 1,117 419.7 (391.8-448.8) 3.8 12.1 1,257 462.2 (431.9-492.6) 3.5

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 females.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


14

Table 4: Cancer incidence in non-Aboriginal females, NSW, 1999-2007(a)

Cancer site Complete case Method Missing (%)



1 in


1 in

Head & neck 1,999 6.1 (5.8-6.3) 211 8.0 2,173 6.6 (6.3-6.9) 193

Oesophagus 1,129 3.2 (3.0-3.4) 530 1.6 1,148 3.2 (3.1-3.4) 521

Stomach 1,957 5.7 (5.4-6.0) 266 3.6 2,030 5.9 (5.7-6.2) 253

Large bowel 16,567 48.8 (48.0-49.5) 28 7.7 17,953 53.0 (52.2-53.7) 26

Liver 905 2.7 (2.5-2.8) 540 1.7 920 2.7 (2.5-2.9) 533

Pancreas 3,153 9 (8.7-9.4) 169 1.5 3,199 9.2 (8.9-9.5) 166

Lung 9,159 27.2 (26.7-27.8) 46 3.1 9,454 28.1 (27.6-28.7) 44

Melanoma of the skin 6,392 19.8 (19.3-20.3) 67 46.7 11,991 37.7 (37.1-38.4) 34

Breast 31,354 98.8 (97.7-99.9) 13 12.2 35,698 112.5 (111.3-113.7) 11

Cervix 1,911 6.1 (5.9-6.4) 213 12.4 2,184 7.1 (6.8-7.4) 186

Uterus 4,433 13.6 (13.2-14.0) 85 8.7 4,855 15.0 (14.5-15.4) 77

Ovary 3,537 10.8 (10.5-11.2) 117 4.5 3,704 11.4 (11.0-11.7) 112

Kidney 2,772 8.3 (8.0-8.6) 159 10.0 3,078 9.2 (8.9-9.6) 140

Brain 1,639 5.2 (4.9-5.4) 241 2.5 1,680 5.3 (5.0-5.5) 235

Thyroid 3,201 10.6 (10.2-10.9) 122 12.2 3,645 12.0 (11.6-12.4) 107


All leukaemias 3,114 9.4 (9.1-9.7) 159 8.3 3,397 10.3 (9.9-10.6) 142

Unknown primary 5,475 15.6 (15.2-16.1) 107 4.1 5,707 16.4 (15.9-16.8) 101

Myelodysplasia 2,212 6.3 (6.0-6.5) 282 8.2 2,407 6.9 (6.6-7.2) 249

All cancers(c) 113,879 345.8 (343.8-347.8) 4.3 12.1 129,625 395.6 (393.4-397.8) 3.8

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 females.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


15

Table 5: Cancer incidence in Aboriginal persons, NSW, 1999-2007(a)

Cancer site Complete case Method Missing (%) Multiple Imputation Method


1 in


1 in

Head & neck 132 22.7 (18.5-27.6) 52 6.8 142 24.4 (19.7-29.2) 48

Oesophagus 41 9 (6.1-12.6) 128 2.0 42 9.2 (6.0-12.5) 125

Stomach 64 15.4 (11.3-20.3) 88 3.3 66 15.7 (11.3-20.2) 86

Large bowel 257 61.1 (52.7-70.3) 24 7.0 278 65.3 (56.1-74.4) 22

Liver 44 8.8 (6.1-12.3) 124 2.0 45 9.0 (5.9-12.0) 122

Pancreas 60 15.6 (11.4-20.8) 104 1.3 61 15.9 (11.3-20.5) 102

Lung 362 84.3 (74.7-94.8) 16 2.5 373 86.5 (76.3-96.8) 15

Melanoma of the skin 67 16.2 (11.9-21.4) 105 43.5 130 25.7 (18.6-32.8) 59

Breast 306 57.0 (49.6-65.0) 25 12.2 348 63.9 (55.5-72.2) 22

Cervix 65 9.4 (6.9-12.5) 138 12.4 71 10.2 (7.2-13.1) 130

Uterus 43 10.5 (7.2-14.7) 140 8.7 50 11.7 (7.8-15.6) 123

Prostate 183 48.6 (41.0-57.0) 27 20.0 239 62.5 (52.3-72.6) 21

Kidney 81 16.9 (12.7-21.7) 84 9.4 87 17.8 (13.3-22.3) 78

Bladder 39 11.4 (7.6-16.1) 167 5.2 41 11.9 (7.7-16.1) 158

Non-Hodgkin’s lymphoma 73 12 (9.0-15.5) 88 10.7 86 14.2 (10.5-17.9) 76


All leukaemias 81 14.8 (10.7-19.5) 124 8.5 89 16.1 (11.6-20.7) 108

Unknown primary 103 22.5 (17.6-28.0) 61 4.4 107 23.3 (18.0-28.5) 58

Myelodysplasia 41 10.7 (7.2-15.2) 188 7.0 45 11.6 (7.4-15.7) 170

All cancers (c) 2,318 493.4 (469.8-517.6) 3.3 13.0 2,604 545.2 (518.8-571.7) 3.0

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 persons.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


16

Table 6: Cancer incidence in non-Aboriginal persons, NSW, 1999-2007(a)




1 in


1 in

Head & neck 7,412 12.0 (11.7-12.3) 98 6.8 7,951 12.9 (12.6-13.2) 92

Oesophagus 3,287 5.2 (5.1-5.4) 270 2.0 3,354 5.3 (5.2-5.5) 264

Stomach 5,543 8.9 (8.6-9.1) 159 3.3 5,729 9.2 (8.9-9.4) 152

Large bowel 36,647 58.6 (58.0-59.2) 23 7.0 39,404 63.1 (62.4-63.7) 21

Liver 3,040 4.9 (4.7-5.1) 268 2.0 3,100 5.0 (4.8-5.2) 262

Pancreas 6,347 10.1 (9.8-10.3) 144 1.3 6,430 10.2 (10.0-10.5) 142

Lung 25,581 40.9 (40.4-41.4) 31 2.5 26,230 42.0 (41.5-42.5) 30

Melanoma of the skin 16,505 26.7 (26.3-27.1) 50 43.5 29,191 47.6 (47.1-48.2) 27

Breast 31,601 51.6 (51.0-52.1) 24 12.2 35,984 58.7 (58.1-59.3) 21

Cervix 1,911 3.2 (3.0-3.3) 420 12.4 2,184 3.6 (3.5-3.8) 368

Uterus 4,433 7.2 (6.9-7.4) 166 8.7 4,855 7.8 (7.6-8.1) 151

Prostate 35,733 57.0 (56.5-57.6) 20 20.0 44,644 71.3 (70.7-72.0) 16

Kidney 7,275 11.7 (11.5-12.0) 108 9.4 8,033 13.0 (12.7-13.2) 97

Bladder 6,608 10.5 (10.2-10.7) 148 5.2 6,974 11.1 (10.8-11.3) 138


Multiple myeloma 3,630 5.8 (5.6-6.0) 240 5.9 3,859 6.2 (6.0-6.4) 221

All leukaemias 7,639 12.4 (12.1-12.7) 117 8.5 8,344 13.5 (13.2-13.8) 105

Unknown primary 11,009 17.5 (17.1-17.8) 92 4.4 11,517 18.3 (17.9-18.6) 88

Myelodysplasia 5,101 8.1 (7.9-8.3) 214 7.0 5,483 8.7 (8.5-8.9) 194

All cancers (c) 254,465 409.6 (408.0-411.2) 3.7 13.0 292,438 471.5 (469.8-473.2) 3.2

(a) Directly standardised to the Australian 2001 population.(b) Australian standardised rate (ASR) per 100,000 persons.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


17

Figure 5: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal male compared to NSW, 1999–2007

0.0 0.5 1.0 1.5 2.0 2.5 3.0

All cancersUnknown primary

All leukaemiasNon-Hodgkin's lymphoma

ThyroidKidney

ProstateMelanoma of skin

LungPancreas

LiverLarge bowel

StomachOesophagusHead & neck

Standardised incidence ratio

Figure 6: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal female compared to NSW, 1999–2007


Head and neck

Oesophagus

Stomach

Large bowel

Liver

Pancreas

Lung

Melanoma of skin

Prostate

Kidney

Thyroid

Non-Hodgkin’s lymphoma

All leukaemas

Unknown primary

All cancers

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5



ThyroidKidneyOvaryUterusCervixBreast

Melanoma of skinLung

PancreasLiver

Large bowelStomach

OesophagusHead & neck


Head and neckOesophagusStomachLarge bowelLiverPancreasLungMelanoma of skinBreastCervixUterusOvaryKidneyThyroidNon-Hodgkin’s lymphomaAll leukaemasUnknown primaryAll cancers



18

Figure 7: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal people compared to NSW, 1999-2007

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5



ThyroidKidneyTestis

ProstateOvaryUterusCervixBreast

Melanoma of skinLung

PancreasLiver

Large bowelStomach

OesophagusHead & neck

standardised incidence ratio

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Head and neck

Oesophagus

Stomach

Large bowel

Liver

Pancreas

Lung

Melanoma of skin

Breast

Cervix

Uterus

Ovary

Prostate

Testis

Kidney

Thyroid


All leukaemas

Unknown primary

All cancers



19

Table 7: Standardised incidence ratios of major cancers from multiple imputation, Aboriginal people compared to NSW, 1999-2007

Cancer site SIR(a) 95%CI SIR(a) 95%CI SIR(a) 95%CI

Head & neck 2.00 (1.66-2.35) 2.05 (1.64-2.46) 2.01 (1.37-2.65)

Oesophagus 1.91 (1.32-2.49) 1.85 (1.15-2.54) 2.16 (1.03-3.29)

Stomach 1.65 (1.25-2.06) 1.85 (1.32-2.38) 1.35 (0.73-1.98)

Large bowel 1.00 (0.88-1.13) 1.06 (0.88-1.23) 0.95 (0.77-1.13)

Liver 1.79 (1.26-2.32) 1.67 (1.06-2.27) 2.21 (1.09-3.33)

Pancreas 1.50 (1.12-1.88) 1.53 (1.00-2.06) 1.49 (0.94-2.04)

Lung 2.10 (1.88-2.32) 1.93 (1.66-2.19) 2.43 (2.06-2.81)

Melanoma of skin 0.45 (0.34-0.56) 0.44 (0.29-0.58) 0.47 (0.31-0.62)

Mesothelioma 1.13 (0.49-1.77) 1.32 (0.54-2.10) 0.51 (0.00-1.49)

Breast 0.97 (0.86-1.08) 0.94 (0.84-1.05)

Cervix 2.51 (1.89-3.13) 2.43 (1.83-3.03)

Uterus 1.20 (0.85-1.55) 1.16 (0.82-1.51)

Ovary 1.14 (0.76-1.51) 1.11 (0.74-1.47)

Prostate 0.79 (0.68-0.91) 0.82 (0.71-0.94)

Kidney 1.30 (1.02-1.59) 1.31 (0.95-1.66) 1.33 (0.87-1.80)

Bladder 1.01 (0.69-1.32) 1.07 (0.69-1.45) 0.91 (0.33-1.49)

Brain 0.88 (0.62-1.15) 0.74 (0.42-1.07) 1.08 (0.63-1.53)

Thyroid 0.58 (0.38-0.78) 0.69 (0.24-1.14) 0.54 (0.33-0.75)

Non-Hodgkin’s

lymphoma

0.86 (0.67-1.06) 0.95 (0.68-1.22) 0.76 (0.48-1.04)

Multiple myeloma 1.07 (0.65-1.49) 1.24 (0.65-1.83) 0.87 (0.30-1.44)

All leukaemias 0.98 (0.77-1.20) 1.06 (0.77-1.35) 0.89 (0.58-1.20)

Unknown primary 1.46 (1.17-1.75) 1.63 (1.21-2.05) 1.29 (0.90-1.68)

Myelodysplasia 1.27 (0.88-1.66) 1.50 (0.94-2.07) 1.00 (0.47-1.53)

All cancers (b) 1.07 (1.03-1.12) 1.09 (1.02-1.15) 1.06 (1.00-1.13)

(a) Standardised incidence ratio (SIR) in Aboriginal people compared to total NSW population. (b) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


20

Cancer mortality by Aboriginal status

During 1999–2007, 1,229 Aboriginal peoples were recorded as dying from cancer, 657 men and 572 women. Mean annual all-cancer mortality (excluding non-melanoma skin cancers) in NSW Aboriginal people was estimated to be 66 per cent higher than in non-Aboriginal people (295 versus 178 per 100,000 persons) (Table 10; Figure 8). The standardised mortality ratio (SMR, with respect to all NSW) was 1.7 (Table 11; Figure 14).

The 70 per cent excess cancer mortality in Aboriginal peoples compared to NSW is considerably higher than the excess of cancer incidence (7%), signifying lower survival in Aboriginal peoples when they are diagnosed with cancer. Aboriginal male cancer mortality was 373 per 100,000 (Table 8; Figure 8, SMR=1.7) and in women was 240 per 100,000 (Table 9; Figure 8, SMR=1.7). The Aboriginal female cancer mortality rate exceeded that of non-Aboriginal males.

The cumulated risk of dying from cancer over 0-75 years in Aboriginal males is 1 in 4.7 compared to 7.5 for non-Aboriginal males; in females the corresponding risks are 1 in 6.5 versus 1 in 11.1.

Figure 8: Directly age-standardised mortality rates, all cancers, Aboriginal and non-Aboriginal people, males and females, NSW, 1999-2007*

0

50

100

150

200

250

300

350

400

450

Males Females Persons

Per 1

00,0

00

Aboriginal people Non-Aboriginal people


Per

100,

000

Aboriginal people

Males Females Persons

non-Aboriginal people


21

Mortality from major cancers

Lung cancer is the leading cause of cancer-related mortality in Aboriginal men and women. It is also the leading cause of cancer death in non-Aboriginal men, but not in non-Aboriginal women, where breast cancer is the leading cause of cancer death (Figures 9 & 10). Lung cancer mortality in Aboriginal men (91 per 100,000), is 1.9 times that of non-Aboriginal men (Table 8), and the SMR is 2.0 (Table 11; Figure 12). In Aboriginal women, lung cancer mortality (58 per 100,000) is 2.7 times that of non-Aboriginal women. As with lung cancer incidence, the lung cancer mortality rate in Aboriginal women exceeded that in non-Aboriginal men.

Aboriginal men also had significantly higher mortality from head and neck, prostate, stomach, oesophageal and unknown primary cancer than non-Aboriginal men (Figure 9). Despite lower incidence of prostate cancer (Figure 2; Tables 1 & 2), prostate cancer mortality (57 per 100,000) is 85 per cent higher in Aboriginal than in non-Aboriginal men (Table 8). Other major cancers with significantly higher mortality, and showing similar patterns of higher mortality than incidence ratios (compared to NSW overall), are: head and neck (SMR=3.2 vs. SIR=2.0), oesophagus (SMR=2.3 vs. SIR=1.9), large bowel (SMR=1.4 vs. SIR=1.1) leukaemia (SMR=1.4 vs. SIR=1.1), kidney (SMR=1.7 vs. SIR=1.3), and cancer of unknown primary (SMR=1.9 vs. SIR=1.6). These differences are borne out in lower survival from these cancers in Aboriginal men (see the survival section of this report).

Aboriginal women have significantly higher mortality from cancer of the lung and bronchus, breast, pancreas, cervix, head and neck, and kidney than non-Aboriginal women (Figure 10, Table 9). Lung cancer is the leading cause of cancer mortality in Aboriginal women, followed by breast and large bowel cancer (Figure 10). In non-Aboriginal women, the leading cancer cause of death is from breast cancer, followed by lung and large bowel.

The picture of mortality relative to incidence is similar to men for major cancers. Despite similar breast cancer incidence to non-Aboriginal women, mortality from breast cancer in Aboriginal women is 50 per cent higher than for NSW (SMR=1.5 vs. SIR=1.0); 374 per cent higher from cervical cancer compared to 140 per cent higher incidence (SMR=4.7 vs. SIR=2.4); 221 per cent higher from cancer of the head and neck compared to 100 per cent higher incidence (SMR=3.2 vs. SIR=2.0); 105 per cent higher from uterine cancer compared to 20 per cent higher incidence (SMR=2.1 vs. SIR=1.2); 136 per cent higher from kidney cancer compared to 30 per cent higher incidence (SMR=2.4 vs. SIR=1.3); 98 per cent higher mortality from stomach cancer despite 35 per cent higher incidence (SMR=2.0 vs. SIR=1.4); and 82 higher from pancreatic cancer despite 50 per cent higher incidence (SMR=1.8 vs. SIR=1.5).

In summary, Aboriginal peoples not only have significantly higher incidence of most major cancers, they also have higher mortality, such that it exceeds that expected from the higher background level of cancer incidence. That is, survival is lower in Aboriginal peoples when diagnosed with most major cancers than in non-Aboriginal people.


22

Figure 9: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal males, NSW, 1999-2007*

0

20

40

60

80

100

120Lu

ng

Pros

tate

Larg

e bo

wel

Unk

nown

prim

ary

Stom

ach

Hea

d &

neck

Panc

reas

All l

euka

emia

s

Oes

opha

gus

Kidn

ey

Per 1

00,0

00

Aboriginal males Non-Aboriginal males


Figure 10: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal females, NSW, 1999-2007*

0

10

20

30

40

50

60

70

80

Lung

Brea

st

Larg

e bo

wel

Panc

reas

Unk

nown

prim

ary

Cer

vix

Ova

ry

Stom

ach

Kidn

ey

Hea

d &

neck

Pper

100

,000

Aboriginal females Non-Aboriginal females


Per

100,

000

Aboriginal males

Aboriginal females

Non-Aboriginal males

Non-Aboriginal females

Per

100,

000


23

Figure 11: Directly age-standardised mortality rates for common cancers in Aboriginal and non-Aboriginal persons, NSW, 1999-2007*

0

10

20

30

40

50

60

70

80

90

Lung

Larg

e bo

wel

Pros

tate

Brea

st

Unk

nown

prim

ary

Panc

reas

Stom

ach

Hea

d &

neck

All l

euka

emia

s

Kidn

ey

Per 1

00,0

00



Per

100,

000



24

Table 8: Cancer mortality in Aboriginal and Non-Aboriginal males, NSW, 1999-2007(a)

Cancer sites

Aboriginal Non-Aboriginal

Deaths ASR(b) 95%CI Risk to age 75

1 in

M:I Ratio(c)


1 in

M:I Ratio(c)

Head & neck 54 20.6 (14.5-28.0) 56 0.6 2,325 8.1 (7.8-8.4) 155 0.4

Oesophagus 25 12.9 (7.6-20.1) 86 1.0 1,684 5.9 (5.7-6.2) 241 0.8

Stomach 34 21.2 (13.5-31.1) 71 0.8 2,535 9.0 (8.6-9.3) 165 0.7

Large bowel 68 38.6 (28.0-51.2) 45 0.5 8,012 28.4 (27.8-29.0) 52 0.4

Liver 20 7.8 (4.2-12.8) 135 0.7 1,571 5.5 (5.2-5.8) 250 0.7

Pancreas 30 19.0 (11.6-28.7) 94 1.1 2,806 9.9 (9.6-10.3) 147 0.9

Lung 164 91.1 (75.3-108.8) 14 0.8 13,706 48.6 (47.7-49.4) 29 0.8

Melanoma of skin 12 8.1 (3.5-15.0) 174 0.4 2,560 9.0 (8.7-9.4) 173 0.3

Prostate 65 57.1 (42.6-74.5) 51 0.5 8,392 30.8 (30.1-31.4) 86 0.2

Kidney 19 10.5 (5.4-17.8) 159 0.5 1,664 5.9 (5.6-6.2) 254 0.4

Bladder 10 7.3 (2.8-14.6) 366 0.3 1,886 6.8 (6.5-7.2) 316 0.4

Brain 14 4.4 (1.7-8.4) 346 0.8 1,813 6.2 (6.0-6.5) 191 0.8

Non-Hodgkin’s

lymphoma

18 7.9 (4.0-13.6) 169 0.5 2,478 8.8 (8.5-9.2) 176 0.4

All leukaemias 27 16.0 (9.3-25.1) 115 0.7 2,424 8.6 (8.3-9.0) 185 0.5

Unknown primary 48 25.1 (17.1-35.0) 57 0.9 4,339 15.5 (15.0-15.9) 106 0.8

All cancers(d) 657 372.9 (338.9-409.0) 4.7 0.6 63,177 224.8 (223.0-226.6) 7.5 0.5

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 males.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


25

Table 9: Cancer mortality in Aboriginal and Non-Aboriginal females, NSW, 1999-2007(a)

Cancer sites

Aboriginal Non- Aboriginal

Deaths ASR(b) 95%CI Risk to age 75 1 in

M:I Ratio(c)


M:I Ratio(c)

Head & neck 18 6.8 (3.8-11.1) 135 0.6 811 2.4 (2.2-2.5) 608 0.4

Oesophagus 11 4.5 (2.0-8.5) 316 0.8 874 2.4 (2.3-2.6) 787 0.8

Stomach 17 7.9 (4.2-13.2) 235 1.0 1,407 4.0 (3.8-4.2) 433 0.7

Large bowel 49 23.6 (16.7-32.2) 81 0.5 6,604 18.8 (18.4-19.3) 87 0.4

Liver 10 4.6 (2.0-8.9) 260 0.7 696 2.0 (1.9-2.2) 768 0.7

Pancreas 29 15.0 (9.6-22.2) 118 1.1 2,800 8.0 (7.7-8.3) 201 0.9

Lung 134 57.8 (47.4-69.6) 25 0.9 7,281 21.4 (20.9-21.9) 62 0.8

Melanoma of skin 9 4.1 (1.6-8.2) 325 0.4 1,182 3.5 (3.3-3.7) 442 0.2

Breast 99 35.6 (28.0-44.4) 37 0.3 7,887 23.7 (23.2-24.2) 59 0.2

Cervix 31 10.9 (6.8-16.2) 123 0.6 680 2.1 (1.9-2.2) 663 0.3

Uterus 13 6.4 (3.1-11.5) 259 0.3 982 2.9 (2.7-3.0) 490 0.2

Ovary 22 8.3 (4.8-13.1) 157 0.7 2,281 6.7 (6.4-7.0) 203 0.6

Kidney 16 6.9 (3.6-11.7) 188 0.6 1,160 3.3 (3.1-3.5) 469 0.4

Brain 15 5.5 (2.7-9.7) 208 0.8 1,244 3.8 (3.6-4.0) 315 0.7

Non-Hodgkin’s

lymphoma

13 5.3 (2.5-9.6) 213 0.6 2,164 6.1 (5.9-6.4) 277 0.4

All leukaemias 15 5.4 (2.5-9.6) 389 0.6 1,705 4.9 (4.7-5.2) 345 0.5

Unknown primary 35 14.4 (9.4-20.8) 103 0.8 4,491 12.7 (12.3-13.1) 138 0.8

All cancers (d) 572 240.3 (218.5-263.5) 6.5 0.6 48,932 142.0 (140.7-143.3) 11.1 0.4

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 females.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


26

Table 10: Cancer mortality in Aboriginal and Non-Aboriginal persons, NSW, 1999-2007(a)

Cancer sites



1 in

M:I Ratio(c)


1 in

M:I Ratio(c)

Head & neck 72 13.2 (9.9-17.0) 80 0.6 3,136 5.0 (4.9-5.2) 250 0.4

Oesophagus 36 8.2 (5.4-11.8) 143 0.9 2,558 4.1 (3.9-4.2) 374 0.8

Stomach 51 13.6 (9.6-18.5) 114 0.9 3,942 6.3 (6.1-6.5) 242 0.7

Large bowel 117 30.1 (24.0-37.1) 59 0.5 14,616 23.2 (22.8-23.6) 66 0.4

Liver 30 6.2 (3.9-9.2) 181 0.7 2,267 3.6 (3.5-3.8) 382 0.7

Pancreas 59 16.7 (12.2-22.2) 106 1.1 5,606 8.9 (8.7-9.1) 171 0.9

Lung 298 72.0 (62.9-81.8) 19 0.9 20,987 33.5 (33.0-33.9) 40 0.8

Melanoma of skin 21 5.7 (3.2-9.1) 239 0.4 3,742 6.0 (5.8-6.2) 252 0.2

Breast 100 20.1 (15.7-25.3) 69 0.4 7,944 12.7 (12.5-13.0) 114 0.2

Cervix 31 6.0 (3.7-9.1) 229 0.6 680 1.1 (1.0-1.2) 1,305 0.3

Ovary 22 4.6 (2.6-7.3) 297 0.8 2,281 3.6 (3.5-3.8) 396 0.6

Prostate 65 23.2 (17.5-30.2) 113 0.5 8,392 13.1 (12.8-13.4) 177 0.2

Kidney 35 8.4 (5.5-12.2) 173 0.5 2,824 4.5 (4.3-4.7) 331 0.4

Bladder 16 4.4 (2.2-7.6) 431 0.4 2,764 4.3 (4.2-4.5) 492 0.4

Brain 29 5.1 (3.0-7.8) 250 0.8 3,057 5.0 (4.8-5.2) 239 0.8

Non-Hodgkin’s

lymphoma

31 6.6 (4.1-9.8) 185 0.6 4,642 7.4 (7.2-7.6) 217 0.4

All leukaemias 42 9.8 (6.5-14.0) 186 0.7 4,129 6.6 (6.4-6.8) 243 0.5

Unknown primary 83 19.0 (14.5-24.3) 75 0.8 8,830 14.0 (13.7-14.2) 120 0.8

All cancers(d) 1,229 295.3 (276.4-315.1) 5.5 0.6 112,109 178.2 (177.2-179.3) 9.0 0.4

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 persons.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


27

Figure 12: Standardised mortality ratios, Aboriginal male compared to NSW, 1999–2007

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

All cancers

Unknown primary

All leukaemias

Non-Hodgkin's lymphoma

Kidney

Prostate

Melanoma of skin

Lung

Pancreas

Liver

Large bowel

Stomach

Oesophagus

Head & neck

Standardised mortality ratio

Figure 13: Standardised mortality ratios, Aboriginal female compared to NSW, 1999–2007

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

All cancers

Unknown primary

All leukaemias


Kidney

Ovary

Uterus

Cervix

Breast

Melanoma of skin

Lung

Pancreas

Liver

Large bowel

Stomach

Oesophagus

Head & neck

Standardised mortality ratioStandardised mortality ratio

Standardised mortality ratio

Head and neck

Oesophagus

Stomach

Large bowel

Liver

Pancreas

Lung

Melanoma of skin

Prostate

Kidney

Thyroid


All leukaemas

Unknown primary

All cancers

Head and neckOesophagus

StomachLarge bowel

LiverPancreas

LungMelanoma of skin

BreastCervixUterusOvaryKidney

ThyroidNon-Hodgkin’s lymphoma

All leukaemasUnknown primary

All cancers


28

Figure 14: Standardised mortality ratios, Aboriginal people compared to NSW, 1999–2007

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

All cancers

Unknown primary

All leukaemias


Kidney

Prostate

Ovary

Uterus

Cervix

Breast

Melanoma of skin

Lung

Pancreas

Liver

Large bowel

Stomach

Oesophagus

Head & neck

Standardised mortality ratioStandardised mortality ratio

Head and neck

Oesophagus

Stomach

Large bowel

Liver

Pancreas

Lung

Melanoma of skin

Breast

Cervix

Uterus

Ovary

Prostate

Testis

Kidney

Thyroid


All leukaemas

Unknown primary

All cancers


29

Table 11: Standardised mortality ratios, Aboriginal peoples compared to NSW, 1999–2007

Cancer siteSMR(a) 95%CI SMR(a) 95%CI SMR(a) 95%CI

Head & neck 3.13 (2.45-3.94) 3.18 (2.39-4.15) 3.21 (1.90-5.07)

Oesophagus 2.30 (1.61-3.18) 2.31 (1.50-3.42) 2.40 (1.20-4.29)

Stomach 2.02 (1.50-2.65) 2.09 (1.45-2.92) 1.98 (1.15-3.16)

Large bowel 1.32 (1.09-1.58) 1.40 (1.08-1.77) 1.24 (0.92-1.64)

Liver 1.83 (1.24-2.62) 1.74 (1.06-2.68) 2.19 (1.05-4.03)

Pancreas 1.75 (1.33-2.25) 1.70 (1.15-2.42) 1.82 (1.22-2.61)

Lung 2.23 (1.99-2.50) 1.99 (1.69-2.31) 2.73 (2.29-3.23)

Melanoma of skin 0.77 (0.47-1.17) 0.67 (0.35-1.17) 1.00 (0.46-1.89)

Mesothelioma 1.20 (0.60-2.15) 1.23 (0.56-2.34) 1.25 (0.15-4.52)

Breast 1.59 (1.30-1.94) 1.54 (1.25-1.88)

Cervix 4.89 (3.32-6.94) 4.74 (3.22-6.73)

Uterus 2.12 (1.13-3.62) 2.05 (1.09-3.51)

Ovary 1.39 (0.87-2.10) 1.35 (0.84-2.04)

Prostate 1.78 (1.37-2.26) 1.86 (1.43-2.37)

Testis 1.82 (0.22-6.59) 1.88 (0.23-6.79)

Kidney 1.93 (1.34-2.68) 1.70 (1.02-2.66) 2.36 (1.35-3.82)

Bladder 1.22 (0.70-1.98) 1.12 (0.54-2.06) 1.51 (0.55-3.29)

Brain 0.94 (0.63-1.34) 0.76 (0.42-1.28) 1.20 (0.67-1.98)

Thyroid 2.39 (0.65-6.12) 1.36 (0.03-7.57) 3.21 (0.66-9.38)

Non-Hodgkin’s lymphoma 1.04 (0.71-1.48) 1.09 (0.64-1.72) 1.00 (0.53-1.72)

Multiple myeloma 1.12 (0.61-1.88) 1.14 (0.49-2.26) 1.11 (0.41-2.42)

All leukaemias 1.32 (0.95-1.79) 1.44 (0.95-2.10) 1.17 (0.66-1.94)

Unknown primary 1.60 (1.27-1.98) 1.86 (1.37-2.46) 1.35 (0.94-1.88)

Myelodysplasia 0.64 (0.13-1.88) 1.13 (0.23-3.29) 0.00 (0.00-1.53)

All cancers (b) 1.69 (1.60-1.79) 1.68 (1.55-1.81) 1.73 (1.59-1.88)

(a) Standard mortality ratio in Aboriginal people compared to total NSW population. (b) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


30

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30- 3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at diagnosis (yr)

Age specific incidence rate NSW (1999-2007)

Aboriginal malesNon-Aboriginal malesAboriginal femalesNon-Aboriginal females

0

500

1,000

1,500

2,000

2,500

3,000

3,500

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)

Age-specific mortality ratesNSW(1999-2007)


Age-specific incidence and mortality from major cancers

Due to small numbers of cases of cancer in the Aboriginal population of NSW, age-specific rates for most cancers cannot reliably be calculated. Even for cancer with high incidence such as lung cancer, age-specific rate curves exhibit substantial variability. However, differences in age patterns of cancer incidence and mortality can indicate which demographic groups are contributing most to overall differences in Aboriginal and non-Aboriginal cancer.

All cancers

For cancers overall, Aboriginal incidence rates were higher across most age groups. Aboriginal versus non-Aboriginal differences were closer in men than women (Figure 15); in the 40–44 yr age group, Aboriginal women actually had higher incidence than Aboriginal men and non-Aboriginal men and women.

Figure 15: Age-specific incidence and mortality rates for all cancers, NSW, 1999–2007

The age differences for all cancers widened with mortality rates. Aboriginal female cancer mortality rates were higher than males and non-Aboriginal males and females in the 40–44 year age group (reflecting incidence), but were also higher than both male and female non-Aboriginal people from 40 to 74 years.

Per

100,

000

Per

100,

000

Aboriginal males Non- Aboriginal males Aboriginal females Non-Aboriginal females

Age-specific mortality rates NSW (1999–2007)

Age-specific incidence rate NSW (1999–2007)

Age at diagnosis (yr) Age at death (yr)


31

Major cancers

Bowel cancer

For large bowel cancer, Aboriginal versus non-Aboriginal age-specific rates were very similar for incidence and for female mortality (Figure 16). However, mortality was higher in Aboriginal men aged 40–74 years compared to non-Aboriginal people.

Figure 16: Age-specific incidence and mortality rates for large bowel cancer, NSW, 1999–2007

0

50

100

150

200

250

300

350

400

450

500

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00




0

50

100

150

200

250

300

350

400

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)



Lung cancer

Aboriginal age-specific incidence and mortality from lung cancer was substantially higher across all 40+ year age groups in both men and women(Figure 17). Moreover, Aboriginal female age-specific incidence and mortality exceeded that of non-Aboriginal men in all age groups excepting 65+ years.

Figure 17: Age-specific incidence and mortality rates for lung cancer, NSW, 1999–2007

0

100

200

300

400

500

600

700

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)



0

100

200

300

400

500

600

700

800

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00








Per

100,

000

Per

100,

000

Per

100,

000

Per

100,

000



Age at death (yr)

Age at death (yr)

Aboriginal men Non- Aboriginal men Aboriginal women Non-Aboriginal women



32

Melanoma

While melanoma rates are lower in Aboriginal people across most age groups, in males they exceeded non-Aboriginal females by age 65+ years. Also, mortality from melanoma in Aboriginal men exceeded that in non-Aboriginal men from age 70 years (Figure 18).

Figure 18: Age-specific incidence and mortality rates for melanoma, NSW, 1999–2007

0

50

100

150

200

250

300

350

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00




0

20

40

60

80

100

120

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)



Breast cancer

Breast cancer incidence in Aboriginal peoples was similar to or lower in most age groups than in non-Aboriginal people, but higher in the 75+ year women (Figure 19). Breast cancer mortality rates were higher in all 35+ year age groups.

Figure 19: Age-specific incidence and mortality rates for breast cancer, NSW, 1999–2007

0

100

200

300

400

500

600

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00



Aboriginal females


0

20

40

60

80

100

120

140

160

180

200

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)


Aboriginal females





Age specific mortality rates NSW (1999-2007)


Age specific mortality rates NSW (1999–2007)

Per

100,

000

Per

100,

000

Per

100,

000

Per

100,

000



Age at death (yr)

Age at death (yr)


Aboriginal women Non-Aboriginal women


33

Cervical cancer

For cervical cancer, age-specific incidence and mortality are substantially higher in Aboriginal than non-Aboriginal women, from age 25 years on (Figure 20).

Figure 20: Age-specific incidence and mortality rates for cervical cancer, NSW, 1999–2007

0

10

20

30

40

50

60

70

80

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00



Aboriginal females


0

10

20

30

40

50

60

70

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)


Aboriginal females


Prostate cancer

Despite lower prostate cancer incidence across most 45+ year age groups in Aboriginal men, mortality was somewhat higher from age 50 years, with the excess accelerating from age 70 years (Figure 21).

Figure 21: Age-specific incidence and mortality rates for prostate cancer, NSW, 1999–2007

0

200

400

600

800

1,000

1,200

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00



Aboriginal males


0

100

200

300

400

500

600

700

800

0-4

5-9

10-1

4 15

-19

20-2

4 25

-29

30-3

4 35

-39

40-4

4 45

-49

50-5

4 55

-59

60-6

4 65

-69

70-7

4 75

+

Per 1

00,0

00

Age at death (yr)


Aboriginal males






Per

100,

000

Per

100,

000

Per

100,

000

Per

100,

000



Age at death(yr)

Age at death (yr)

Aboriginal women Non-Aboriginal women

Aboriginal men Non-Aboriginal men


34

Degree of spread by Aboriginal status for major cancers

Survival from cancer is determined by prognosis of the cancer at diagnosis and adequacy of treatment, along with other factors including cancer type (histology), co-morbidities and age. For a similar background incidence of cancer and overall mortality, differing survival patterns can result from a key prognostic indicator of cancer; namely stage at diagnosis. For reporting purposes the NSW Central Cancer Registry reports validated summary measures of cancer stage, which characterise a cancer’s degree of spread. These are categorised as localised, regional, distant (i.e. metastatic) and unknown degree of spread. Differences in degree of spread at diagnosis between Aboriginal and non-Aboriginal people can shed some light on differences in cancer mortality and survival between Aboriginal and non-Aboriginal people.

For cancer overall, Aboriginal men and women have lower incidence rates of localised cancer than non-Aboriginal men and women, but not significantly so (Figure 22). Correspondingly, Aboriginal men and women have significantly higher incidence rates of regionalised and distant cancers. In terms of standardised incidence ratios (comparing Aboriginal peoples with NSW overall), the standardised incidence rate (SIR) for localised cancer in Aboriginal men and women was 0.8 (Table 16); for regional cancer the male and female SIR was 1.3 and 1.2 respectively; for distant cancer the SIR was 1.7 for both men and women.

The mortality pattern is somewhat different: mortality rates are significantly higher for each degree-of-spread category, including unknown degree of spread, in both Aboriginal men and women than non-Aboriginal counterparts (Figure 23). Notably, mortality is significantly higher for Aboriginal peoples with localised cancer and substantially higher in Aboriginal men for unknown degree-of-spread cancer, despite lower incidence of localised cancer and similar incidence of cancer with unknown degree of spread. This is reflected in significantly higher standardised mortality ratios of 1.6 (males) and 1.7 (females) for localised cancer, 1.7 (males) and 1.8 (females) for regionalised cancer, and 1.8 (both) for distant cancer (Table 17).

In proportional terms, 33 per cent of cancers in Aboriginal men were localised at diagnosis, lower compared to 40 per cent in non-Aboriginal men (Table 12). Correspondingly higher proportions of regional and distant cancers were diagnosed in Aboriginal men (17% and 19%, respectively), compared to non-Aboriginal men (15% and 14% respectively). There is no significant Aboriginal/non-Aboriginal difference in unknown degree of spread. The incidence pattern is similar in women. The proportion of localised cancers is 36 per cent in Aboriginal women and 42 per cent in non-Aboriginal women, with correspondingly higher proportions of regional and distant cancer (26% and 19%, respectively) compared to non-Aboriginal women (23% and 15%, respectively) (Table 13).

By contrast, degree-of-spread proportions by cancer mortality were not substantially or significantly different between Aboriginal and non-Aboriginal people (Tables 14 & 15).


35

Figure 22: Age-standardised all-cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

50

100

150

200

250

300

Local Regional Distant Unknown

Per 1

00,0

00

Degree of spread at diagnosis

Aboriginal males


Aboriginal females


Figure 23: Age-standardised all-cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

20

40

60

80

100

120

140


Per 1

00,0

00


Aboriginal males


Aboriginal females


Aboriginal men

Non- Aboriginal men

Aboriginal women

Non-Aboriginal women





Aboriginal men

Non- Aboriginal men

Aboriginal women


Per

100,

000

Per

100,

000


36

Stomach cancer

For stomach cancer, there was substantially higher incidence of localised and distant cancer in Aboriginal men but these differences were not statistically significant. Differences in women were less pronounced (Figure 24). These differences were reflected in stomach cancer mortality, except that the substantially higher male Aboriginal mortality from distant stomach cancer was statistically significant (Figure 25). Small event numbers in Aboriginal stomach cancer incidence and mortality preclude inferences about Aboriginal/non-Aboriginal differences in proportions of stomach by degree-of- spread (Tables 12–15).

Figure 24: Age-standardised stomach cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

2

4

6

8

10

12

14

16


Per 1

00,0

00


Aboriginal males


Aboriginal females


Figure 25: Age-standardised stomach cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

2

4

6

8

10

12

14

16

18


Per 1

00,0

00


Aboriginal males


Aboriginal females






Aboriginal men

Non- Aboriginal men

Aboriginal women



Per

100,

000

Per

100,

000


37

Large bowel cancer

For bowel cancer, there are no substantial or significant differences by Aboriginal status in incidence by degree of spread (Figure 26) and this is reflected in bowel cancer mortality (Figure 27). However, 27 per cent of bowel cancers in Aboriginal men were localised at diagnosis, compared to 34 per cent in non-Aboriginal men (Table 12). Most of the corresponding excess occurred in distant spread of disease: 21 per cent versus 16 per cent for Aboriginal and non-Aboriginal people, respectively. In women, the proportional differences were closer, with 30 per cent of bowel cancers localised at diagnosis compared to 33 per cent in non-Aboriginal women. As with men, most of the corresponding increased proportion occurred in distant cancer, 20 per cent in Aboriginal women versus 16 per cent in non-Aboriginal women.

Figure 26: Age-standardised large bowel cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

5

10

15

20

25

30

35

40

45

50


Per 1

00,0

00


Aboriginal males


Aboriginal females


Figure 27: Age-standardised large bowel cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

5

10

15

20

25

30


Per 1

00,0

00


Aboriginal males


Aboriginal females


Aboriginal men

Non- Aboriginal men

Aboriginal women


Aboriginal men

Non- Aboriginal men

Aboriginal women






Per

100,

000

Per

100,

000


38

Lung cancer

Lung cancer incidence is considerably and significantly higher across all degree-of-spread categories in male and female Aboriginal people compared to non-Aboriginal people; lung cancer incidence is higher in Aboriginal women across all degree-of-spread categories than for non-Aboriginal men, although these differences are not statistically significant (Figure 28).

Figure 28: Age-standardised lung cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

5

10

15

20

25

30

35

40

45


Per 1

00,0

00


Aboriginal males


Aboriginal females


Due to low survival from lung cancer, the Aboriginal/non-Aboriginal differences in mortality closely reflect those in incidence (Figure 29).

Figure 29: Age-standardised lung cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

5

10

15

20

25

30

35

40

45


Per 1

00,0

00


Aboriginal males


Aboriginal females


Aboriginal men

Non- Aboriginal men

Aboriginal women


Aboriginal men

Non- Aboriginal men

Aboriginal women






Per

100,

000

Per

100,

000


39

Melanoma

Aboriginal peoples had significantly lower incidence of localised melanoma than non-Aboriginal people but incidence rates of regional and distant degree-of-spread melanoma were not significantly different from non-Aboriginal people (Figure 30).

Figure 30: Age-standardised melanoma incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

10

20

30

40

50

60


Per 1

00,0

00


Aboriginal males


Aboriginal females


Small numbers prevent definitive conclusions for mortality from melanoma, but female mortality from localised melanoma is higher (non-significant) than in non-Aboriginal women, despite significantly lower incidence (Figure 31, cf. Figure 30).

Figure 31: Age-standardised melanoma mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

2

4

6

8

10

12


Per 1

00,0

00


Aboriginal males


Aboriginal females


Aboriginal men

Non- Aboriginal men

Aboriginal women


Aboriginal men

Non-Aboriginal men

Aboriginal women






Per

100,

000

Per

100,

000


40

Breast cancer

Aboriginal women had lower incidence of localised breast cancer and significantly higher incidence of distant degree-of-spread breast cancer (12 per 100,000), compared to non-Aboriginal women (6 per 100,000) (Figure 32).

In proportional terms, localised breast cancer comprised 45 per cent of diagnoses in Aboriginal women compared to 52 per cent in non-Aboriginal women (Table 13). Correspondingly, 40 per cent and 9 per cent of breast cancers in Aboriginal women were regional and distant respectively, compared to 34 per cent and 5 per cent in non-Aboriginal women.

Figure 32: Age-standardised breast cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

10

20

30

40

50

60

70


Per 1

00,0

00


Aboriginal females


In contrast, mortality rates in Aboriginal women were higher across all the degree-of-spread categories (Figure 33). Mortality from distant degree-of-spread breast cancer was significantly and substantially higher, at 9 per 100,000 compared to 3.3 per 100,000 (Table 15). By proportion, the mortality pattern by degree of spread was similar for regional spread, lower in Aboriginal women for localised spread (22%) compared to non-Aboriginal women (27%); and correspondingly higher by distant degree of spread (22% versus 14%).

Figure 33: Age-standardised breast cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

2

4

6

8

10

12

14

16

18

20


Per 1

00,0

00


Aboriginal females


Aboriginal women


Aboriginal women


Local Regional Distant UnknownDegree of spread at diagnosis

Local Regional Distant UnknownDegree of spread at diagnosis

Per

100,

000

Per

100,

000


41

Cervical cancer

Incidence of localised and regional cervical cancer in Aboriginal women is significantly and substantially higher than in non-Aboriginal women (Figure 34). It is almost double (6.3 versus 3.4 per 100,000), more than triple (6.4 versus 1.9 per 100,000) and over five times (2.5 versus 0.5 per 100,000) that in non-Aboriginal women for localised, regional and distant degree-of-spread cancer, respectively (Figure 34, Table 13).

Figure 34: Age-standardised cervical cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

2

4

6

8

10

12


Per 1

00,0

00


Aboriginal females


The mortality pattern is worse: Aboriginal women suffer more than five times the mortality rate from localised cervical cancer (3.1 versus 0.5 per 100,000), over four times that from regional cervical cancer (3.7 versus 0.8 per 100,000), and almost five times that from distant cancer (1.8 versus 0.4 per 100,000) (Figure 35, Table 15).

Figure 35: Age-standardised cervical cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

1

2

3

4

5

6

7

8


Per 1

00,0

00


Aboriginal females


Aboriginal women


Aboriginal women






Per

100,

000

Per

100,

000


42

Prostate cancer

Incidence of localised prostate cancer is significantly lower in Aboriginal men than non-Aboriginal men, and is significantly higher for distant degree of spread prostate cancer (Figure 36).

Figure 36: Age-standardised prostate cancer incidence rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

10

20

30

40

50

60

70

80

90


Per 1

00,0

00


Aboriginal males


The mortality pattern for prostate cancer is somewhat different: Aboriginal mortality is higher across all degree-of-spread categories, and for unknown degree of spread, it is significantly so (Figure 37). Of note is the higher mortality from localised prostate cancer in spite of significantly lower incidence.

Figure 37: Age-standardised prostate cancer mortality rates by Aboriginal status and degree of spread at diagnosis, NSW, 1999–2007

0

5

10

15

20

25

30

35

40

45

50


Per 1

00,0

00


Aboriginal males


Aboriginal men

Non-Aboriginal men

Aboriginal men

Non-Aboriginal men





Per

100,

000

Per

100,

000


43

Table 12: Age-standardised incidence rates by degree of spread at diagnosis, males, NSW, 1999-2007(a)

Cancer sites


Degree of spread

Number Per cent (%)

ASR(b) 95% CI Number Per cent (%)

ASR(b) 95% CI

All cancers(c) Localised 444 33.0 207.5 (182.3-232.7) 65,620 40.3 227.2 (225.4-228.9)

Regionalised 235 17.5 113.1 (94.6-131.5) 24,005 14.7 83.3 (82.3-84.4)

Distant 260 19.3 129.4 (109.9-148.8) 21,952 13.5 77.1 (76.0-78.1)

Unknown 408 30.3 210.2 (181.4-239.0) 51,236 31.5 180.9 (179.3-182.5)

Stomach Localised 11 24.2 7.3 (2.3-12.4) 938 25.3 3.3 (3.1-3.5)

Regionalised 15 32.3 6.1 (2.4-9.8) 1,259 34.0 4.4 (4.2-4.6)

Distant 13 27.9 9.2 (3.5-15.0) 963 26.0 3.4 (3.2-3.6)

Unknown 7 15.5 2.8 (0.5-5.2) 540 14.6 1.9 (1.8-2.1)

Large bowel Localised 43 27.2 20.3 (12.6-28.0) 7,313 34.1 25.6 (25.0-26.2)

Regionalised 67 42.3 34.0 (23.7-44.3) 8,456 39.4 29.5 (28.9-30.1)

Distant 33 21.1 14.1 (8.2-20.1) 3,504 16.3 12.2 (11.8-12.6)

Unknown 15 9.5 7.0 (2.6-11.4) 2,178 10.2 7.7 (7.3-8.0)

Lung Localised 47 23.1 26.2 (17.0-35.4) 3,754 22.4 13.3 (12.8-13.7)

Regionalised 35 17.0 19.9 (12.1-27.7) 2,977 17.7 10.4 (10.0-10.8)

Distant 70 34.3 33.4 (24.1-42.7) 5,818 34.7 20.3 (19.8-20.9)

Unknown 53 25.7 31.5 (21.1-41.8) 4,227 25.2 15.1 (14.7-15.6)

Melanoma of

skin

Localised 54 79.5 26.4 (15.4-37.4) 14,093 81.9 48.8 (48.0-49.6)

Regionalised 5 7.2 3.7 (0.0-7.7) 1,268 7.4 4.5 (4.2-4.7)

Distant 5 7.6 1.6 (0.0-3.2) 929 5.4 3.2 (3.0-3.5)

Unknown 4 5.7 1.9 (0.0-4.3) 910 5.3 3.2 (3.0-3.4)

Prostate Localised 108 45.3 57.2 (43.6-70.8) 21,303 47.7 72.8 (71.8-73.8)

Regionalised 11 4.7 5.8 (1.7-9.9) 2,514 5.6 8.5 (8.1-8.8)

Distant 20 8.4 17.8 (9.2-26.3) 1,865 4.2 6.7 (6.4-7.1)

Unknown 99 41.5 63.8 (46.0-81.5) 18,963 42.5 67.1 (66.1-68.1)

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


44

Table 13: Age-standardised incidence rates by degree of spread at diagnosis, females, NSW, 1999-2007(a)

Cancer sites


Degree of spread

Number Per cent (%)


ASR(b) 95% CI

All cancers (c) Localised 452 36.0 159.9 (142.1-177.8) 54,094 41.7 168.1 (166.6-169.5)

Regionalised 327 26.0 115.1 (100.6-129.6) 29,193 22.5 90.2 (89.1-91.2)

Distant 241 19.2 93.8 (80.2-107.4) 18,733 14.5 55.8 (55.0-56.6)

Unknown 238 18.9 93.4 (78.6-108.2) 27,606 21.3 81.5 (80.6-82.5)


Regionalised 4 22.2 2.5 (0.0-5.1) 616 30.3 1.8 (1.7-2.0)

Distant 7 38.0 2.1 (0.5-3.7) 498 24.5 1.5 (1.4-1.6)

Unknown 2 12.7 1.0 (0.0-2.4) 389 19.1 1.1 (1.0-1.2)


Regionalised 49 40.4 22.3 (15.0-29.6) 7,305 40.7 21.6 (21.1-22.1)

Distant 24 20.1 9.5 (5.1-13.8) 2,835 15.8 8.4 (8.1-8.8)

Unknown 12 9.6 6.6 (2.3-10.9) 1,909 10.6 5.5 (5.2-5.7)

Lung Localised 36 21.3 15.3 (9.7-20.9) 2,186 23.1 6.5 (6.2-6.8)

Regionalised 34 20.6 11.5 (7.4-15.7) 1,613 17.1 4.9 (4.6-5.1)

Distant 56 33.8 25.3 (17.8-32.8) 3,305 35.0 10.0 (9.6-10.3)

Unknown 41 24.3 16.5 (10.9-22.1) 2,351 24.9 6.8 (6.5-7.1)

Melanoma of skin Localised 53 84.7 17.2 (10.2-24.2) 10,151 84.7 32.1 (31.5-32.7)

Regionalised 5 8.3 2.1 (0.0-4.3) 703 5.9 2.1 (2.0-2.3)

Distant 2 3.3 0.4 (0.0-1.1) 478 4.0 1.5 (1.3-1.6)

Unknown 2 3.7 0.6 (0.0-1.7) 659 5.5 2.1 (1.9-2.2)

Breast Localised 155 44.8 52.9 (42.9-62.9) 18,531 51.9 58.5 (57.6-59.3)

Regionalised 138 39.8 41.5 (33.5-49.5) 12,291 34.4 39.3 (38.6-40.0)

Distant 31 8.9 11.7 (7.0-16.5) 1,901 5.3 5.9 (5.6-6.1)

Unknown 23 6.5 9.6 (4.5-14.7) 2,974 8.3 8.8 (8.5-9.1)

Cervix Localised 27 38.0 6.3 (3.6-9.0) 1,041 47.7 3.4 (3.2-3.6)

Regionalised 21 30.0 6.4 (3.2-9.7) 598 27.4 1.9 (1.8-2.1)

Distant 10 14.0 2.5 (0.0-.) 165 7.6 0.5 (0.0-.)

Unknown 13 18.1 3.8 (1.2-6.3) 380 17.4 1.2 (1.1-1.4)

(a) Directly standardised to the Australian 2001 population. (b) Per 100,000 females. (c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


45

Table 14: Age-standardised mortality rates by degree of spread at diagnosis, males, NSW, 1999-2007(a)

Cancer sites


Degree of spread

Number Per cent (%)


ASR(b) 95% CI

All cancers c) Localised 135 20.5 65.7 (52.7-80.4) 13,727 21.7 44.6 (43.9-45.4)

Regionalised 126 19.2 66.8 (53.3-82.3) 11,254 17.8 36.1 (35.4-36.7)

Distant 217 33.0 101 (85.2-118.5) 18,166 28.8 58.2 (57.3-59.0)

Unknown 179 27.2 108.1 (89.9-128.4) 20,030 31.7 65.5 (64.6-66.4)


Regionalised 14 41.2 6.6 (3.2-11.8) 902 35.6 3.2 (3.0-3.4)

Distant 12 35.3 8.7 (3.9-16.2) 830 32.7 2.9 (2.7-3.1)

Unknown 2 5.9 0.8 (0.1-2.9) 376 14.8 1.4 (1.2-1.5)


Regionalised 25 36.8 16.8 (9.6-26.5) 3,289 41.1 11.7 (11.3-12.1)

Distant 29 42.6 13.3 (8.0-20.3) 2,742 34.2 9.6 (9.3-10.0)

Unknown 7 10.3 3.4 (0.9-7.9) 851 10.6 3.1 (2.9-3.3)

Lung Localised 39 23.8 20.8 (13.7-29.9) 2,483 18.1 8.9 (8.5-9.2)

Regionalised 26 15.9 18.8 (11.4-28.8) 2,394 17.5 8.4 (8.1-8.8)

Distant 62 37.8 30.7 (22.2-41.0) 5,324 38.8 18.6 (18.1-19.2)

Unknown 37 22.6 20.7 (13.5-29.9) 3,505 25.6 12.6 (12.2-13.0)

Melanoma of

skin

Localised 6 50.0 3.3 (0.8-8.0) 1,342 52.4 4.8 (4.5-5.0)

Regionalised 3 25.0 3.2 (0.6-9.6) 407 15.9 1.4 (1.3-1.6)

Distant 2 16.7 0.9 (0.1-3.5) 650 25.4 2.3 (2.1-2.5)

Unknown 1 8.3 0.6 (0.0-3.6) 161 6.3 0.6 (0.5-0.7)

Prostate Localised 16 24.6 13.1 (6.8-22.3) 2,281 27.2 8.4 (8.0-8.7)

Regionalised 4 6.2 2.5 (0.3-7.2) 507 6.0 1.8 (1.7-2.0)

Distant 14 21.5 10.8 (5.3-19.1) 1,478 17.6 5.4 (5.1-5.7)

Unknown 31 47.7 30.6 (20.0-44.5) 4,126 49.2 15.2 (14.7-15.6)

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


46

Table 15: Age-standardised mortality rates by degree of spread at diagnosis, females, NSW, 1999-2007*

Cancer sites


Degree of spread

Number Per cent (%)


ASR(b) 95% CI

All cancers (c) Localised 113 19.8 47.3 (38.0-58.0) 9,684 19.8 25.5 (25.0-26.0)

Regionalised 146 25.5 50.9 (41.9-61.0) 10,989 22.5 29.5 (29.0-30.1)

Distant 190 33.2 70.4 (59.5-82.5) 14,875 30.4 39.7 (39.0-40.3)

Unknown 123 21.5 51.0 (41.4-62.1) 13,384 27.4 34.2 (33.6-34.8)


Regionalised 3 17.6 2.0 (0.4-6.1) 432 30.7 1.3 (1.1-1.4)

Distant 7 41.2 2.1 (0.8-4.3) 437 31.1 1.3 (1.2-1.4)

Unknown 2 11.8 1.2 (0.1-4.6) 284 20.2 0.8 (0.7-0.9)

Large bowel Localised 6 12.2 3.9 (1.3-8.8) 841 12.7 2.3 (2.2-2.5)

Regionalised 20 40.8 9.5 (5.3-15.3) 2,656 40.2 7.6 (7.3-7.9)

Distant 20 40.8 8.2 (4.6-13.3) 2,248 34.0 6.6 (6.4-6.9)

Unknown 3 6.1 2.0 (0.4-6.1) 859 13.0 2.3 (2.2-2.5)

Lung Localised 21 15.7 9.7 (5.6-15.3) 1,285 17.6 3.7 (3.5-3.9)

Regionalised 26 19.4 8.8 (5.5-13.3) 1,198 16.5 3.6 (3.4-3.8)

Distant 52 38.8 23.3 (16.7-31.5) 2,946 40.5 8.8 (8.5-9.1)

Unknown 35 26.1 15.9 (10.6-22.7) 1,852 25.4 5.3 (5.0-5.5)

Melanoma of skin Localised 8 88.9 3.9 (1.5-8.1) 629 53.2 1.9 (1.7-2.0)

Regionalised 159 13.5 0.5 (0.4-0.5)

Distant 1 11.1 0.2 (0.0-0.9) 294 24.9 0.9 (0.8-1.0)

Unknown 100 8.5 0.3 (0.2-0.3)

Breast Localised 22 22.2 9.5 (5.5-15.1) 2,164 27.4 6.4 (6.2-6.7)

Regionalised 43 43.4 12.6 (8.7-17.5) 3,383 42.9 10.4 (10.1-10.8)

Distant 22 22.2 9.0 (5.2-14.3) 1,110 14.1 3.3 (3.1-3.5)

Unknown 12 12.1 4.4 (2.1-8.1) 1,230 15.6 3.5 (3.3-3.7)

Cervix Localised 9 29.0 3.1 (1.2-6.4) 180 26.5 0.5 (0.5-0.6)

Regionalised 9 29.0 3.7 (1.4-7.6) 261 38.4 0.8 (0.7-0.9)

Distant 7 22.6 1.8 (0.7-3.9) 123 18.1 0.4 (0.3-0.5)

Unknown 6 19.4 2.2 (0.6-5.2) 116 17.1 0.3 (0.3-0.4)

(a) Directly standardised to the Australian 2001 population.(b) Per 100,000 females.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


47

Table 16: Standardised incidence ratios by degree of spread, Aboriginal people compared to NSW, 1999-2007

Cancer siteDegree of Spread

Ratio(a) 95%CI Ratio(a) 95%CI Ratio(a) 95%CI

All cancers (b) Localised 0.83 (0.77-0.89) 0.82 (0.74-0.91) 0.84 (0.76-0.93)

Regionalised 1.22 (1.12-1.33) 1.27 (1.10-1.44) 1.18 (1.05-1.32)

Distant 1.66 (1.52-1.81) 1.67 (1.47-1.88) 1.67 (1.46-1.88)

Unknown 1.09 (0.99-1.19) 1.13 (1.00-1.25) 1.07 (0.92-1.22)

Stomach Localised 1.75 (0.89-2.61) 1.91 (0.78-3.04) 1.54 (0.19-2.88)

Regionalised 1.40 (0.77-2.02) 1.65 (0.81-2.48) 0.94 (0.02-1.87)

Distant 1.81 (1.01-2.60) 1.85 (0.84-2.85) 1.80 (0.47-3.13)

Unknown 1.81 (0.61-3.01) 2.35 (0.60-4.10) 1.09 (0.00-2.61)

Large bowel Localised 0.86 (0.66-1.06) 0.86 (0.59-1.13) 0.87 (0.55-1.20)

Regionalised 1.03 (0.83-1.23) 1.12 (0.84-1.40) 0.94 (0.66-1.22)

Distant 1.23 (0.91-1.55) 1.33 (0.87-1.78) 1.13 (0.68-1.59)

Unknown 0.98 (0.55-1.41) 1.02 (0.44-1.61) 0.95 (0.35-1.55)

Lung Localised 2.09 (1.63-2.54) 2.02 (1.43-2.60) 2.25 (1.49-3.01)

Regionalised 2.08 (1.59-2.58) 1.74 (1.16-2.32) 2.69 (1.78-3.59)

Distant 1.93 (1.59-2.27) 1.78 (1.36-2.20) 2.23 (1.64-2.81)

Unknown 2.41 (1.91-2.91) 2.25 (1.63-2.86) 2.76 (1.90-3.63)

Melanoma of skin Localised 0.44 (0.31-0.56) 0.42 (0.26-0.58) 0.47 (0.29-0.64)

Regionalised 0.59 (0.20-0.98) 0.48 (0.01-0.94) 0.78 (0.10-1.47)

Distant 0.58 (0.15-1.02) 0.66 (0.08-1.23) 0.47 (0.00-1.12)

Unknown 0.35 (0.00-0.79) 0.47 (0.00-1.11) 0.22 (0.00-0.80)

Breast Localised 0.85 (0.70-0.99) 0.82 (0.68-0.97)

Regionalised 1.03 (0.86-1.21) 1.00 (0.83-1.18)

Distant 1.80 (1.17-2.43) 1.70 (1.09-2.31)

Unknown 0.97 (0.52-1.42) 0.95 (0.51-1.40)

Cervix Localised 1.88 (1.15-2.61) 1.82 (1.11-2.52)

Regionalised 3.05 (1.75-4.36) 2.96 (1.69-4.22)

Distant 5.55 (2.66-10.21) 5.40 (2.59-9.93)

Unknown 2.49 (0.97-4.01) 2.42 (0.94-3.90)

Prostate Localised 0.69 (0.54-0.83) 0.71 (0.56-0.86)

Regionalised 0.58 (0.21-0.94) 0.60 (0.22-0.97)

Distant 2.13 (1.20-3.06) 2.22 (1.25-3.19)

Unknown 0.87 (0.66-1.08) 0.91 (0.69-1.12)

(a) Standard incidence ratio in aboriginal people compared to total NSW population. (b) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


48

Table 17: Standardised mortality ratios by degree of spread, Aboriginal people compared to NSW, 1999-2007

Cancer siteDegree of Spread Ratio(a) 95%CI Ratio(a) 95%CI Ratio(a) 95%CI

All cancers (b) Localised 1.65 (1.45-1.87) 1.61 (1.35-1.91) 1.72 (1.42-2.07)

Regionalised 1.75 (1.55-1.97) 1.71 (1.43-2.04) 1.80 (1.52-2.11)

Distant 1.80 (1.63-1.98) 1.80 (1.57-2.05) 1.82 (1.57-2.10)

Unknown 1.56 (1.39-1.74) 1.58 (1.36-1.83) 1.56 (1.30-1.87)

Stomach Localised 3.07 (1.53-5.50) 2.61 (0.96-5.68) 4.09 (1.33-9.54)

Regionalised 1.84 (1.07-2.95) 2.23 (1.22-3.75) 1.06 (0.22-3.11)

Distant 2.04 (1.23-3.19) 2.06 (1.06-3.59) 2.10 (0.84-4.32)

Unknown 1.27 (0.35-3.26) 1.06 (0.13-3.84) 1.63 (0.20-5.89)

Large bowel Localised 1.28 (0.68-2.18) 1.20 (0.48-2.48) 1.41 (0.52-3.07)

Regionalised 1.26 (0.92-1.69) 1.26 (0.81-1.86) 1.29 (0.79-1.99)

Distant 1.42 (1.05-1.88) 1.57 (1.05-2.26) 1.26 (0.77-1.95)

Unknown 1.18 (0.57-2.17) 1.54 (0.62-3.18) 0.78 (0.16-2.27)

Lung Localised 2.74 (2.09-3.52) 2.84 (2.02-3.89) 2.67 (1.65-4.08)

Regionalised 2.19 (1.64-2.88) 1.75 (1.15-2.57) 3.04 (1.99-4.46)

Distant 1.97 (1.62-2.36) 1.75 (1.34-2.24) 2.39 (1.78-3.13)

Unknown 2.41 (1.88-3.03) 1.99 (1.40-2.75) 3.21 (2.23-4.46)

Melanoma of skin Localised 0.99 (0.54-1.66) 0.66 (0.24-1.45) 1.62 (0.70-3.19)

Regionalised 0.79 (0.16-2.29) 1.09 (0.23-3.20) 0.00 (0.00-2.97)

Distant 0.39 (0.08-1.15) 0.39 (0.05-1.42) 0.41 (0.01-2.30)

Unknown 0.56 (0.01-3.09) 0.89 (0.02-4.95) 0.00 (0.00-4.56)

Breast Localised 1.38 (0.87-2.09) 1.35 (0.85-2.04)

Regionalised 1.42 (1.03-1.91) 1.39 (1.00-1.87)

Distant 2.65 (1.68-3.98) 2.48 (1.56-3.76)

Unknown 1.52 (0.79-2.66) 1.49 (0.77-2.60)

Cervix Localised 5.50 (2.52-10.44) 5.33 (2.44-10.13)

Regionalised 3.74 (1.71-7.10) 3.62 (1.66-6.88)

Distant 5.54 (2.23-11.42) 5.39 (2.17-11.11)

Unknown 5.80 (2.13-12.63) 5.63 (2.07-12.25)

Prostate Localised 1.66 (0.95-2.70) 1.74 (0.99-2.82)

Regionalised 1.51 (0.41-3.87) 1.57 (0.43-4.03)

Distant 2.04 (1.11-3.41) 2.12 (1.16-3.56)

Unknown 1.78 (1.21-2.53) 1.86 (1.26-2.64)

(a) Standard mortality ratio in aboriginal people compared to total NSW population. (b) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer


49

Survival by Aboriginal status and major cancer type

Survival differences in cancer in different populations occurs due to differences in prognostic factors at cancer diagnosis (degree of spread, cancer histology), treatment efficacy, age and sex, along with other factors including co-morbid conditions. However, as the previous sections have shown, not all the excess Aboriginal cancer mortality is due to higher incidence, which in any case is not relevant to survival, nor is all of it explained by higher proportions of late-stage cancers. This is highlighted by lower Aboriginal incidence of localised cancers overall, yet significantly higher mortality from localised cancer (cf. Figures 22 and 23). These differences indicate the extent of differences in cancer survival, best quantified by Kaplan-Meier survival analyses.

For cancers overall, survival in Aboriginal peoples is significantly lower than in non-Aboriginal people, and the survival deficit is greater in men than women (Figure 38). The proportion of Aboriginal persons surviving cancer at five years following diagnosis was 53 per cent compared to 65 per cent of non-Aboriginal people (Table 18). In men, these proportions are 49 per cent and 64 per cent, and in women 57 per cent and 67 per cent. Survival differences at one, two, three and four years were similar.

Part of the difference in survival is explained by differences in degree of spread at diagnosis. In Aboriginal peoples, localised cancers were a smaller proportion of cancers than in non-Aboriginal people, and correspondingly proportions of regional and distant cancer were higher (cf. Tables 12 & 13) and this was also the case with degree of spread by mortality breakdowns (cf. Tables 14 & 15). However, when incidence versus mortality rates by degree of spread at diagnosis are compared, mortality is considerably higher in Aboriginal people for each degree of spread category, including by localised cancer where incidence in Aboriginal people is lower than non-Aboriginal people (cf. Figures 22 & 23).


50

All cancers

Figure 38: Kaplan-Meier survival curve for all cancers by Aboriginal status, NSW, 1999-2007

0.2

5.5

.75

1

Surv

ival

pro

babi

lity

12 24 36 48 60Months since diagnosis

P-value: <0.0001

Persons

0.2

5.5

.75

112 24 36 48 60

Months since diagnosis

P-value: <0.0001

Females

0.2

5.5

.75

1


95% CI 95% CIAboriginal Non_Aboriginal

P-value: <0.0001

Males

Number of cases(a)

Aboriginal non-Aboriginal Aboriginal non-Aboriginal Aboriginal non-Aboriginal

2, 582 288, 366 1,333 160,768 1,249 127,598

Year since

diagnosis

1 72.2 ( 70.4 - 74.0) 79.6 ( 79.4 - 79.7) 69.4 ( 66.8 - 72.0) 78.8 ( 78.6 - 79.0) 75.2 ( 72.8 - 77.7) 80.5 ( 80.3 - 80.8)

2 62.0 ( 59.9 - 64.0) 73.1 ( 72.9 - 73.2) 58.4 ( 55.5 - 61.3) 72.0 ( 71.8 - 72.2) 65.7 ( 62.8 - 68.5) 74.4 ( 74.1 - 74.6)

3 58.2 ( 56.0 - 60.3) 69.5 ( 69.3 - 69.7) 54.0 ( 51.0 - 57.1) 68.3 ( 68.1 - 68.6) 62.4 ( 59.5 - 65.3) 70.9 ( 70.6 - 71.2)

4 54.8 ( 52.5 - 57.0) 67.1 ( 66.9 - 67.3) 50.7 ( 47.5 - 53.9) 65.9 ( 65.7 - 66.2) 59.0 ( 55.9 - 62.0) 68.6 ( 68.4 - 68.9)

5 52.6 ( 50.3 - 55.0) 65.4 ( 65.2 - 65.5) 48.8 ( 45.5 - 52.1) 64.1 ( 63.8 - 64.4) 56.5 ( 53.4 - 59.7) 66.9 ( 66.6 - 67.2)

Log-rank test <0.0001 <0.0001 <0.0001

Percentage of survival is survival estimates from Kaplan-Meier analysis with 95% confidence interval (CI). (a) Cancer cased found by post-mortem and death certificate only are excluded.

Table 18: Percentage surviving all cancers, by Aboriginal status and years since diagnosis


51

Survival by major cancers

Survival in Aboriginal peoples from cancer of the head and neck (Figure 39), stomach (Figure 41), large bowel (Figure 42), lung (Figure 45), breast (Figure 47), cervix (Figure 48) and prostate (Figure 51) is significantly worse, and not significantly different for oesophagus, liver, pancreatic cancer, melanoma, uterus, ovary and kidney, compared to non-Aboriginal people.

Cancer of the head and neck

The proportion of Aboriginal peoples surviving cancer of the head and neck five years after diagnosis is 49 per cent, significantly lower than 66 per cent survival in non-Aboriginal people (Figure 39, Table 19).

Most of the Aboriginal survival deficit has occurred in males, where 41 per cent survive to five years compared to 65 per cent in non-Aboriginal men (Figure 39, Table 19).

In Aboriginal women, short-term survival (to three years) from head and neck cancer is considerably less, but by five years the proportion surviving (68%) is similar to non-Aboriginal women (Figure 39, Table 19). However, small numbers prevent definitive conclusions about long-term survival differences between Aboriginal and non-Aboriginal women.

Figure 39: Kaplan-Meier survival curve for head and neck cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1

Surv

ival

pro

babi

lity


P-value: <0.0001

Persons

0.2

5.5

.75

1



P-value: <0.0001

Males

0.2

5.5

.75

1


P-value: 0.8429

Females

Table 19: Percentage surviving head and neck cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


142 7,901 103 5,750 39 2,151

Year since diagnosis

1 76.6 ( 69.4 - 83.8) 83.5 ( 82.7 - 84.3) 74.6 ( 65.9 - 83.3) 83.5 ( 82.5 - 84.5) 81.7 ( 69.3 - 94.0) 83.5 ( 81.9 - 85.1)

2 58.4 ( 49.6 - 67.2) 74.6 ( 73.6 - 75.6) 54.7 ( 44.1 - 65.2) 74.3 ( 73.2 - 75.5) 67.7 ( 52.4 - 83.0) 75.4 ( 73.6 - 77.3)

3 56.3 ( 47.3 - 65.3) 70.7 ( 69.6 - 71.7) 51.7 ( 40.8 - 62.5) 70.3 ( 69.1 - 71.6) 67.7 ( 52.4 - 83.0) 71.6 ( 69.6 - 73.6)

4 52.2 ( 42.7 - 61.8) 67.4 ( 66.3 - 68.6) 45.8 ( 34.2 - 57.4) 66.7 ( 65.3 - 68.0) 67.7 ( 52.4 - 83.0) 69.4 ( 67.3 - 71.5)

5 48.7 ( 38.4 - 58.9) 65.6 ( 64.4 - 66.8) 40.7 ( 28.2 - 53.2) 64.7 ( 63.3 - 66.2) 67.7 ( 52.4 - 83.0) 67.7 ( 65.5 - 69.9)

Log-rank test <0.0001 <0.0001 <0.8429



52

Oesophagus

Survival from oesophageal cancer was not substantially or significantly different between Aboriginal and non-Aboriginal people (Figure 40).

Figure 40: Kaplan-Meier survival curve for oesophageal cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y

12 24 36 48 60 72Months since diagnosis

P-value: 0.1320

Persons

0.2

5.5

.75

1



P-value: 0.1986

Males

0.2

5.5

.75

1


P-value: 0.4360

Females

Table 20: Percentage surviving oesophageal cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


42 3,301 28 2,182 14 1,119


1 30.2 (15.6 - 44.9) 41.7 (40.0 - 43.4) 32.9 (14.5 - 51.3) 41.2 (39.0 - 43.3) 24.3 (0.7 - 47.9) 42.7 (39.7 - 45.7)

2 22.6 (6.3 - 39.0) 27.6 (26.0 - 29.3) 21.6 (0.4 - 42.8) 26.9 (24.9 - 28.8) 24.3 (0.7 - 47.9) 29.2 (26.4 - 32.0)

3 20.4 (6.8 - 34.0) 22.5 (21.0 - 24.1) 17.2 (0.7 - 33.7) 21.6 (19.7 - 23.5) 24.3 (0.7 - 47.9) 24.3 (21.6 - 27.0)

4 15.6 (2.1 - 29.0) 20.0 (18.5 - 21.5) 9.3 (0.0 - 24.3) 18.9 (17.0 - 20.7) 24.3 (0.7 - 47.9) 22.1 (19.4 - 24.8)

5 10.8 (0.0 - 23.2) 18.3 (16.8 - 19.9) 9.3 (0.0 - 24.3) 17.4 (15.5 - 19.2) 12.4 (0.0 - 33.2) 20.2 (17.5 - 22.8)

Log-rank test 0.132 0.1986 0.436



53

Oesophagus

Survival from oesophageal cancer was not substantially or significantly different between Aboriginal and non-Aboriginal people (Figure 40).

Figure 40: Kaplan-Meier survival curve for oesophageal cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.1320

Persons

0.2

5.5

.75

1



P-value: 0.1986

Males

0.2

5.5

.75

1


P-value: 0.4360

Females

Table 20: Percentage surviving oesophageal cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


42 3,301 28 2,182 14 1,119


1 30.2 (15.6 - 44.9) 41.7 (40.0 - 43.4) 32.9 (14.5 - 51.3) 41.2 (39.0 - 43.3) 24.3 (0.7 - 47.9) 42.7 (39.7 - 45.7)

2 22.6 (6.3 - 39.0) 27.6 (26.0 - 29.3) 21.6 (0.4 - 42.8) 26.9 (24.9 - 28.8) 24.3 (0.7 - 47.9) 29.2 (26.4 - 32.0)

3 20.4 (6.8 - 34.0) 22.5 (21.0 - 24.1) 17.2 (0.7 - 33.7) 21.6 (19.7 - 23.5) 24.3 (0.7 - 47.9) 24.3 (21.6 - 27.0)

4 15.6 (2.1 - 29.0) 20.0 (18.5 - 21.5) 9.3 (0.0 - 24.3) 18.9 (17.0 - 20.7) 24.3 (0.7 - 47.9) 22.1 (19.4 - 24.8)

5 10.8 (0.0 - 23.2) 18.3 (16.8 - 19.9) 9.3 (0.0 - 24.3) 17.4 (15.5 - 19.2) 12.4 (0.0 - 33.2) 20.2 (17.5 - 22.8)

Log-rank test 0.132 0.1986 0.436


Stomach

Survival from stomach cancer is significantly lower in Aboriginal than in non-Aboriginal people (Figure 41, Table 2.24): at five years since diagnosis, 17 per cent of Aboriginal peoples have survived stomach cancer, compared to 30 per cent of non-Aboriginal people.

Despite lower survival from stomach cancer in Aboriginal men, this difference is not significant from non-Aboriginal men. At five years following diagnosis, 20 per cent of Aboriginal men have survived, compared to 30 per cent of non-Aboriginal men. While this difference is considerable, small numbers prevent definitive conclusions; however the size of the difference does suggest that larger numbers would not alter the difference substantially but simply confirm it.

This is evident in survival from stomach cancer in Aboriginal women, which is significantly lower than for non-Aboriginal women (Figure 41). At five years, the proportion of Aboriginal women surviving stomach cancer is estimated to be 11 per cent, compared to 30 per cent in non-Aboriginal women (Table 2.24).

Figure 41: Kaplan-Meier survival curve for stomach cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.0130

Persons

0.2

5.5

.75

1

12 24 36 48 60 72 84 96Months since diagnosis


P-value: 0.2622

Males

0.2

5.5

.75

1


P-value: <0.0020

Females

Table 21: Percentage surviving stomach cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


65 5,661 46 3,665 19 1,996


1 48.1 (35.4 - 60.8) 53.5 (52.2 - 54.8) 52.1 (37.1 - 67.1) 54.0 (52.4 - 55.7) 38.1 (14.9 - 61.2) 52.5 (50.3 - 54.8)

2 29.8 (17.8 - 41.8) 39.7 (38.4 - 41.1) 33.9 (19.2 - 48.7) 39.9 (38.2 - 41.6) 19.8 (0.5 - 39.1) 39.4 (37.2 - 41.7)

3 23.0 (11.3 - 34.7) 33.8 (32.5 - 35.2) 24.3 (10.1 - 38.5) 33.8 (32.1 - 35.5) 19.8 (0.5 - 39.1) 33.9 (31.7 - 36.1)

4 17.1 (5.6 - 28.6) 30.9 (29.6 - 32.3) 19.9 (5.7 - 34.1) 30.9 (29.3 - 32.6) 11.4 (0.0 - 28.9) 31.0 (28.7 - 33.2)

5 17.1 (5.6 - 28.6) 29.6 (28.3 - 30.9) 19.9 (5.7 - 34.1) 29.6 (28.0 - 31.3) 11.4 (0.0 - 28.9) 29.5 (27.3 - 31.8)

Log-rank test 0.013 0.2622 0.002



54

Large bowel

Survival from bowel cancer in Aboriginal peoples is significantly and substantially lower than for non-Aboriginal people (Figure 42, Table 22). Survival in the short term (to 18 months) is similar but the largest survival drop is apparent from 21 to 30 months (Figure 42). Five-year survival from bowel cancer in Aboriginal peoples is estimated as 51 per cent, compared to 65 per cent in non-Aboriginal people. Most of the survival deficit in bowel cancer is explained by higher proportions diagnosed as distant cancer.

Survival from bowel cancer in Aboriginal men at five years is estimated to be 48 per cent compared to 65 per cent in non-Aboriginal men (Figure 42, Table 22). Most of the survival deficit occurs in the medium term (18–30 months).

Survival from bowel cancer in Aboriginal males at five years is estimated to be 48 per cent compared to 65 per cent in non-Aboriginal males (Figure 42, Table 22). Most of the survival deficit occurs in the medium term (18–30 months).

Survival from bowel cancer in Aboriginal women is somewhat longer than in males (55% at five years compared to 48%), but is considerably lower than in non-Aboriginal women (65%) (Table 22). The survival pattern is similar to men: most of the survival deficit occurs over 21–30 months (Figure 42).

Figure 42: Kaplan-Meier survival curve for large bowel cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: <0.0001

Persons

0.2

5.5

.75

1



P-value: 0.0019

Males

0.2

5.5

.75

1


P-value: <0.0080

Females

Table 22: Percentage surviving large bowel cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


277 39,066 158 21,319 119 17,747


1 81.8 (77.1 - 86.5) 84.0 (83.7 - 84.4) 83.0 (76.9 - 89.0) 84.8 (84.3 - 85.3) 80.3 (73.0 - 87.6) 83.1 (82.5 - 83.7)

2 67.4 (61.5 - 73.4) 75.9 (75.5 - 76.3) 68.5 (60.6 - 76.3) 76.4 (75.8 - 77.0) 66.0 (56.9 - 75.1) 75.3 (74.6 - 76.0)

3 57.9 (51.4 - 64.5) 70.7 (70.2 - 71.2) 56.8 (47.9 - 65.6) 70.9 (70.3 - 71.6) 59.4 (49.6 - 69.2) 70.4 (69.6 - 71.1)

4 52.9 (45.9 - 59.9) 67.4 (66.8 - 67.9) 50.4 (40.9 - 59.9) 67.3 (66.6 - 68.0) 56.3 (46.0 - 66.7) 67.4 (66.6 - 68.1)

5 50.7 (43.5 - 57.9) 64.9 (64.4 - 65.4) 47.9 (38.1 - 57.7) 64.6 (63.8 - 65.3) 54.6 (43.8 - 65.3) 65.3 (64.5 - 66.1)

Log-rank test <0.0001 0.0019 0.008



55

Liver

Survival from primary liver cancer appears to be similar in Aboriginal and non-Aboriginal people, although small event numbers prevent definitive conclusions (Figure 43; Table 23).

Figure 43: Kaplan-Meier survival curve for liver cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y

12 24 36 48 60 72 84 96 108 120Months since diagnosis

P-value: 0.5167

Persons

0.2

5.5

.75

1



P-value: 0.3037

Males

0.2

5.5

.75

1


P-value: 0.7833

Females

Table 23: Percentage surviving liver cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


45 2,981 30 2,107 15 874


1 39.2 (24.4 - 54.1) 41.9 (40.0 - 43.7) 39.5 (21.1 - 58.0) 43.3 (41.1 - 45.5) 39.9 (15.1 - 64.6) 38.5 (35.2 - 41.8)

2 30.8 (16.4 - 45.3) 30.6 (28.8 - 32.4) 30.2 (11.9 - 48.4) 32.1 (30.0 - 34.2) 31.9 (7.7 - 56.1) 27.1 (24.0 - 30.2)

3 19.0 (4.8 - 33.2) 25.2 (23.5 - 27.0) 21.3 (1.2 - 41.3) 26.3 (24.2 - 28.4) 16.0 (0.0 - 35.7) 22.7 (19.7 - 25.8)

4 13.4 (0.0 - 27.4) 21.8 (20.0 - 23.5) 21.3 (1.2 - 41.3) 22.4 (20.3 - 24.5) 16.0 (0.0 - 35.7) 20.3 (17.2 - 23.3)

5 13.4 (0.0 - 27.4) 19.8 (18.1 - 21.5) 21.3 (1.2 - 41.3) 20.1 (18.0 - 22.2) 16.0 (0.0 - 35.7) 19.1 (16.0 - 22.2)

Log-rank test 0.5167 0.3037 0.7833



56

Pancreas

Survival from pancreatic cancer in Aboriginal peoples is similar and not significantly different from non-Aboriginal people (Figure 44, Table 24), and is similar in men and women. While small numbers prevent definitive conclusions, the similarity of the survival point estimates and the generally low survival from pancreatic cancer suggest that the similarity is valid.

Figure 44: Kaplan-Meier survival curve for pancreatic cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.8398

Persons

0.2

5.5

.75

1



P-value: 0.6152

Males

0.2

5.5

.75

1


P-value: 0.8204

Females

Table 24: Percentage surviving pancreatic cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


60 6,127 31 3,106 29 3,021


1 27.4 (15.7 - 39.0) 23.6 (22.5 - 24.7) 33.8 (16.8 - 50.7) 24.7 (23.1 - 26.2) 20.1 (4.7 - 35.5) 22.5 (21.0 - 24.0)

2 4.4 (0.0 - 10.3) 12.1 (11.2 - 13.0) 8.0 (0.0 - 18.4) 11.7 (10.5 - 12.9) 8.5 (0.0 - 19.5) 12.4 (11.2 - 13.7)

3 4.4 (0.0 - 10.3) 8.7 (7.9 - 9.5) 8.0 (0.0 - 18.4) 8.2 (7.1 - 9.3) 8.5 (0.0 - 19.5) 9.1 (8.0 - 10.3)

4 4.4 (0.0 - 10.3) 7.4 (6.7 - 8.2) 8.0 (0.0 - 18.4) 6.8 (5.8 - 7.9) 8.5 (0.0 - 19.5) 7.9 (6.9 - 9.0)

5 4.4 (0.0 - 10.3) 6.7 (6.0 - 7.5) 8.0 (0.0 - 18.4) 6.2 (5.2 - 7.2) 8.5 (0.0 - 19.5) 7.2 (6.1 - 8.3)

Log-rank test 0.8398 0.6152 0.8204



57

Lung

Lung cancer survival in Aboriginal peoples is significantly lower than for non-Aboriginal people (Figure 45, Table 25). The survival difference is evident in the short term and becomes significant by about 18 months. At five years following diagnosis of lung cancer, 9 per cent of Aboriginal peoples are alive, compared to 16 per cent of non-Aboriginal people.

In Aboriginal men, survival from lung cancer is significantly lower than non-Aboriginal men, and most of the survival deficit occurs 9–12 months following diagnosis (Figure 45). At five years, 9 per cent of Aboriginal men are alive compared to 15 per cent of non-Aboriginal men.

In Aboriginal women, the survival deficit from lung cancer is more highly significant than men, as reflected in the lower p-value for the difference with non-Aboriginal women (0.0080), compared to the male difference (0.0165). Somewhat in contrast to the men, most of the survival deficit occurs from nine months and extends to 18 months following diagnosis (Figure 45). At five years, 9 per cent of Aboriginal women are alive from lung cancer, compared to 19 per cent of non-Aboriginal women. Aboriginal women have lower survival from lung cancer than non-Aboriginal men (Table 25).

Despite similar proportions of lung cancer incidence by degree of spread in Aboriginal peoples versus non-Aboriginal people, the proportional contribution of localised cancer to Aboriginal lung cancer mortality is higher in males (24%) compared to non-Aboriginal males (18%) (Table 14).

Figure 45: Kaplan-Meier survival curve for lung cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.0008

Persons

0.2

5.5

.75

1



P-value: 0.0165

Males

0.2

5.5

.75

1


P-value: 0.0080

Females

Table 25: Percentage surviving lung cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


367 25,570 201 16,376 166 9,194


1 34.9 (29.8 - 40.0) 40.7 (40.1 - 41.3) 31.9 (25.2 - 38.5) 38.9 (38.1 - 39.6) 38.6 (30.9 - 46.3) 43.8 (42.8 - 44.9)

2 17.9 (13.6 - 22.2) 26.4 (25.8 - 26.9) 18.0 (12.2 - 23.8) 24.8 (24.1 - 25.5) 17.8 (11.2 - 24.3) 29.1 (28.2 - 30.1)

3 13.8 (9.7 - 17.9) 20.6 (20.1 - 21.2) 13.7 (8.3 - 19.1) 19.2 (18.6 - 19.9) 14.0 (7.7 - 20.3) 23.1 (22.1 - 24.0)

4 9.8 (6.0 - 13.7) 18.1 (17.5 - 18.6) 10.3 (5.2 - 15.4) 16.7 (16.1 - 17.4) 8.9 (2.8 - 15.1) 20.4 (19.5 - 21.4)

5 9.2 (5.4 - 13.1) 16.3 (15.8 - 16.9) 9.3 (4.3 - 14.3) 15.0 (14.4 - 15.7) 8.9 (2.8 - 15.1) 18.7 (17.7 - 19.6)

Log-rank test 0.0008 0.0165 0.008



58

Melanoma

Survival from melanoma in Aboriginal peoples is quite similar to that in non-Aboriginal people (Figure 46, Table 26), and is similar in men and women. Five-year survival from melanoma is >90 per cent for Aboriginal men and women (Table 26).

Figure 46: MI Kaplan-Meier survival curve for melanoma by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.6039

Persons

0.2

5.5

.75

1



P-value: 0.6708

Males

0.2

5.5

.75

1


P-value: 0.9560

Females

Table 26: Percentage surviving melanoma, by Aboriginal status and years since diagnosis

Number of cases(a)


130 29,151 68 17,181 62 11,970


1 97.7 (95.0 - 100) 96.9 (96.7 - 97.1) 95.4 (90.2 - 100) 96.2 (95.9 - 96.4) 100 98.0 (97.7 - 98.2)

2 94.0 (89.6 - 98.5) 94.6 (94.3 - 94.9) 93.5 (87.0 - 99.9) 93.4 (93.0 - 93.8) 94.7 (88.7 - 100) 96.4 (96.0 - 96.7)

3 91.9 (86.6 - 97.2) 92.7 (92.4 - 93.0) 89.0 (80.1 - 97.9) 91.1 (90.6 - 91.5) 94.7 (88.7 - 100) 95.1 (94.7 - 95.5)

4 90.6 (84.7 - 96.5) 91.4 (91.1 - 91.8) 89.0 (80.1 - 97.9) 89.5 (89.0 - 90.0) 92.1 (84.3 - 99.9) 94.2 (93.7 - 94.6)

5 90.6 (84.7 - 96.5) 90.4 (90.0 - 90.8) 89.0 (80.1 - 97.9) 88.3 (87.7 - 88.8) 92.1 (84.3 - 99.9) 93.3 (92.8 - 93.8)

Log-rank test 0.6039 0.6708 0.956



59

Breast

Breast cancer survival in Aboriginal women is significantly lower than for non-Aboriginal women (Figure 47, Table 27). A substantial proportion of the survival deficit occurs at 40 months (Figure 47). At five years, 79 per cent of Aboriginal breast cancer cases are alive compared to 88 per cent of non-Aboriginal cases (Table 27).

The contribution to lower breast cancer survival in Aboriginal women comes from the lower proportion of localised cancer diagnosed compared to non-Aboriginal women, and correspondingly higher distant cancers diagnosed (cf. Table 13). This is reflected in the mortality by degree-of-spread breakdown (cf. Table 15).

Figure 47: Kaplan-Meier survival curve for female breast cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y



P-value: <0.0001

Females

Table 27: Percentage surviving breast cancer, by Aboriginal status and years since diagnosis

Number of cases(a)

Aboriginal non-Aboriginal

345 35,492


1 94.7 (92.3 - 97.1) 97.3 (97.1 - 97.5)

2 91.1 (88.0 - 94.2) 94.5 (94.2 - 94.7)

3 87.7 (84.0 - 91.5) 92.0 (91.7 - 92.3)

4 82.7 (78.2 - 87.2) 89.8 (89.5 - 90.1)

5 79.1 (74.0 - 84.2) 87.8 (87.4 - 88.2)

Log-rank test <0.0001



60

Cervix

Survival from cervical cancer in Aboriginal women is significantly and substantially lower than in non-Aboriginal women (Figure 48, Table 28). The survival deficit occurs mostly over 0–24 months but continues to widen up to 40 months following diagnosis (Figure 48). At five years, 59 per cent of Aboriginal women survive cervical cancer compared to 74 per cent of non-Aboriginal women (Table 28).

A large part of lower survival from cervical cancer in Aboriginal women is explained by lower proportions of cases diagnosed as local cancer, and correspondingly higher proportions with distant degree of spread (cf. Table 13). However, Aboriginal/non-Aboriginal mortality ratios in each degree-of-spread category are substantially higher than corresponding incidence ratios (cf. Figures 34 & 35, and Tables 16 & 17 for SIR versus SMR comparisons).

Figure 48: MI Kaplan-Meier survival curve for cervical cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y



P-value: 0.0051

Females

Table 28: Percentage surviving cervical cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


70 2,173


1 80.1 (70.6 - 89.6) 87.9 (86.5 - 89.3)

2 70.4 (59.1 - 81.7) 81.3 (79.7 - 83.0)

3 65.2 (53.2 - 77.2) 77.4 (75.6 - 79.3)

4 59.0 (46.0 - 72.0) 75.2 (73.3 - 77.2)

5 59.0 (46.0 - 72.0) 73.6 (71.5 - 75.6)

Log-rank test 0.0051



61

Uterus

Survival from uterine cancer in Aboriginal women is lower than in non-Aboriginal women, but the difference is not statistically significant (Figure 49, Table 29).

Figure 49: MI Kaplan-Meier survival curve for uterine cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y



P-value: 0.2690

Females

Table 29: Percentage surviving uterine cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


49 4,802


1 87.6 (78.3 - 97.0) 92.2 (91.4 - 92.9)

2 83.2 (72.5 - 94.0) 87.3 (86.3 - 88.2)

3 80.7 (69.1 - 92.2) 84.5 (83.4 - 85.6)

4 77.7 (65.0 - 90.3) 82.6 (81.5 - 83.8)

5 74.1 (60.2 - 88.0) 81.2 (80.0 - 82.4)

Log-rank test 0.269



62

Ovary

Survival from ovarian cancer in Aboriginal women is somewhat higher than in non-Aboriginal women, but not significantly so (Figure 50, Table 30).

Figure 50: MI Kaplan-Meier survival curve for ovarian cancer by Aboriginal status, Females, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y



P-value: 0.6771

Females

Table 30: Percentage surviving ovarian cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


38 3,642


1 84.0 (72.2 - 95.8) 75.7 (74.3 - 77.1)

2 67.2 (51.7 - 82.7) 63.2 (61.6 - 64.8)

3 60.8 (44.3 - 77.2) 54.3 (52.6 - 56.0)

4 57.3 (40.4 - 74.3) 47.4 (45.6 - 49.2)

5 44.8 (26.4 - 63.3) 43.3 (41.4 - 45.1)

Log-rank test 0.6771



63

Prostate

Prostate cancer survival in Aboriginal men is significantly and substantially lower than for non-Aboriginal men (Figure 51, Table 31). The survival deficit begins soon after diagnosis and continues to widen at a constant rate thereafter (Figure 51). At five years, the proportion of Aboriginal men surviving prostate cancer is estimated to be 77 per cent compared to 88 per cent of non-Aboriginal men diagnosed with prostate cancer (Table 31).

A contributing factor to lower survival from prostate cancer in Aboriginal men is that distant degree-of-spread cancer comprises 8.4 per cent of diagnoses compared to 4.2 per cent of those in non-Aboriginal men (cf. Table 12). However, mortality from localised prostate cancer is higher in Aboriginal than non-Aboriginal men, despite lower incidence of localised prostate cancer.

Figure 51: Kaplan-Meier survival curve for prostate cancer by Aboriginal status, Males, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y



P-value: <0.0001

Males

Table 31: Percentage of surviving prostate cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


238 44,215


1 91.4 (87.7 - 95.1) 96.6 (96.4 - 96.7)

2 87.0 (82.5 - 91.6) 93.8 (93.6 - 94.0)

3 83.4 (78.2 - 88.7) 91.4 (91.1 - 91.7)

4 79.6 (73.6 - 85.7) 89.4 (89.1 - 89.7)

5 77.6 (70.9 - 84.3) 87.7 (87.3 - 88.1)

Log-rank test <0.0001



64

Kidney

Survival from kidney cancer in Aboriginal peoples is somewhat lower than non-Aboriginal people, but not significantly so (Figure 52, Table 32). Most of the lower survival occurs in men, while survival in Aboriginal and non-Aboriginal women is very similar.

Figure 52: Kaplan-Meier survival curve for kidney cancer by Aboriginal status, NSW, 1999–2007

0.2

5.5

.75

1Su

rviv

al p

roba

bilit

y


P-value: 0.2097

Persons

0.2

5.5

.75

1



P-value: 0.1609

Males

0.2

5.5

.75

1


P-value: 0.7379

Females

Table 32: Percentage of surviving kidney cancer, by Aboriginal status and years since diagnosis

Number of cases(a)


86 7,897 52 4,884 34 3,013


1 75.5 (66.1 - 84.9) 81.7 (80.8 - 82.6) 75.4 (63.2 - 87.6) 83.4 (82.3 - 84.4) 75.8 (61.0 - 90.5) 79.0 (77.6 - 80.5)

2 68.4 (57.9 - 78.8) 75.5 (74.6 - 76.5) 65.3 (51.2 - 79.3) 77.3 (76.1 - 78.6) 72.5 (57.0 - 88.0) 72.6 (71.0 - 74.2)

3 68.4 (57.9 - 78.8) 72.2 (71.1 - 73.2) 65.3 (51.2 - 79.3) 73.9 (72.6 - 75.2) 72.5 (57.0 - 88.0) 69.4 (67.6 - 71.1)

4 64.2 (52.8 - 75.5) 69.9 (68.8 - 71.0) 65.3 (51.2 - 79.3) 71.3 (69.9 - 72.7) 65.0 (47.8 - 82.2) 67.7 (65.9 - 69.5)

5 64.2 (52.8 - 75.5) 68.7 (67.6 - 69.8) 65.3 (51.2 - 79.3) 70.0 (68.6 - 71.4) 65.0 (47.8 - 82.2) 66.5 (64.7 - 68.4)

Log-rank test 0.2097 0.1609 0.7379



65

Methods

Cancer Registration in NSW

The NSW Central Cancer Registry

The NSW Central Cancer Registry (CCR) receives notifications of all malignant cancers diagnosed in NSW. The CCR is managed by the Cancer Institute NSW for the NSW Ministry of Health (NSW Health), and operates under the authority of the Public Health Act of 1991. The Registry maintains a record of all malignant cancer cases diagnosed in NSW residents since 1972. Information collected on patients with cancer includes name and address, date and country of birth, clinical details about the cancer, the notifying institution, treatment information, treating doctor and general practitioner. Personal identifying details such as name and address are needed to ensure that accurate information is recorded for each person and that each new cancer is only counted once in the statistics. All information sent to the Registry is kept confidential in a highly secure environment and is protected by the Public Health Act 1991 and the Health Records Information Privacy Act 2002.

Notification of cancer in NSW

Notification of malignant neoplasms has been a statutory requirement for all notifying institutions in NSW since 1986. These institutions include public and private hospitals, departments of radiation oncology, nursing homes, pathology laboratories, outpatient departments and day procedure centres. When any of these institutions diagnose or treat someone with malignant cancer, they are required by law to notify the NSW Central Cancer Registry. Notifications of cancer in NSW residents are also received from cancer registries in other states and territories.

Purpose of the Registry

The CCR contributes to the prevention, control and treatment of cancer in the population of NSW. The aims of the NSW Central Cancer Registry are to:

• monitor the number of new cases of cancer and deaths from cancer in NSW

• produce regular reports about cancer incidence and mortality

• provide descriptive analyses of cancer incidence and mortality

• utilise cancer data for epidemiological and clinical research

• evaluate screening programs to determine their effectiveness

• assist in planning and monitoring services for the control of cancer and the care of cancer patients in NSW

• make CCR data available for use by health providers, planners, educators and research scientists.

Demographic Variables in the CCR Database

The CCR database records the year and month of birth, death and diagnosis of each cancer case in NSW, plus basic demographic variables including sex, country of birth and residential address. Identifying information is kept secure and is not available to researchers unless ethical approval is obtained.

For recording of Aboriginal status the NSW CCR includes a variable named ABTSI (Aboriginal status indicator), coded as follows:

1 – Aboriginal2 – Torres Strait Islander3 – Aboriginal and Torres Strait Islander4 – Neither Aboriginal nor Torres Strait Islander9 – Not stated

This classification replicates Aboriginal and Torres Strait Islander status recording in the ABS population census, and is in keeping with National best practice guidelines for recording of Aboriginal status.7 However, as with many other health data collections, the coding of Aboriginal status in cancer registration makes no distinction between cases where the patient has not, or refused, to state their Aboriginal status and where the information was not collected (e.g. not asked) or recorded.

Clinical Variables in the CCR Database

The cancer site recorded in the central cancer registry is the bodily site or organ where the cancer originated, (i.e. the topography of the cancer). The method of diagnosis and histology of the cancer are also recorded.

Degree of spread in the NSW CCR is a summary measure based on cancer staging at first presentation. It is derived by the CCR from the maximum extent of disease based on all reports and notifications dated within four months of the date of diagnosis. Degree of spread reported here follows the international coding guidelines for summary stage adopted by several international groups including the World Health Organization and the International Association of Cancer Registries.8

Degree of spread is grouped as:

i. local (assumed to predominantly consist of Tumour Node Metastases (TNM) Stage i but may include some Stage ii)

ii. regionalised (assumed to be predominantly TNM Stage ii and most of Stage iii)

iii. distant (predominantly TNM Stage iv cancers)

iv. Some cancers are classified as unknown degree of spread for which staging information is inadequate or has not been collected.


66

Sources of Aboriginal and/or Torres Strait Islander Status recording in the CCR

The coding and registration of cancer notifications is a complex process in which diagnostic and demographic information from multiple sources, including pathology laboratories, hospitals and the NSW Registrar of Births, Deaths and Marriages (RBDM); is accumulated over an approximate 6-month period for each cancer notification before it is finalised as completed CCR record. Information on Aboriginal status comes from hospital treatment episodes and the RBDM, but not from notifying pathology laboratories. The Aboriginal status field in the CCR and the equivalent in NSW hospital and death notifications denotes ‘Neither Aboriginal nor Torres Strait Islander’, ‘Aboriginal’, ‘Torres Strait Islander’, ‘Aboriginal and Torres Strait Islander’, along with ‘Declined to respond’ and ‘Unstated’. The more cancer notifications received from hospitals, the higher the chance that Aboriginal status will be non-missing; since the mid-1990s there is also a very low chance of Aboriginal status missing on NSW death certifications. Aboriginal information from the RBDM can originate from either part of the death certificate; that completed by the medical certifier or that completed (usually) by the funeral director.

The algorithm used by the CCR for coding Aboriginal status is one of ‘progressive’ positive identification in which a single notice from any notifying source indicating more information on Aboriginal status overwrites less information. In brief, if 1=‘Aboriginal’, 2=‘Torres Strait Islander’, 3=‘Aboriginal and Torres Strait Islander’, 4=‘Neither Aboriginal nor Torres Strait Islander’, 8=‘Declined to respond’, and 9=‘Unstated’, then the coding hierarchy is as follows: 1 replaces 4, 8, or 9; 2 replaces 1, 4, 8 or 9; 3 replaces 1, 2, 4, 8 or 9; 4 replaces 8 or 9; and 8 replaces 9. For the purposes of this report, however, Aboriginal codes indicating Aboriginality are grouped into a binary yes/no category, and ‘unstated’ or ‘declined to respond’ are regarded as ‘Unknown’.

Death registration/certificates

The main source of death information for linkage with the CCR is the NSW RBDM which provides identified death information to the CCR fortnightly. Cancer records are also linked annually after receipt of ABS death data with all causes of death coded according to the International Classification of Diseases, 10th edition (ICD-10). The ABS has provided cause of death data files to the registry annually since 1994, which include Aboriginal status.

Prior to 2002 most cancer deaths in the NSWCCR had no Aboriginal status recorded. In 2002/2003, the NSW CCR used ABS-recorded Aboriginal status from death registration data to retrospectively update records of those who had died from cancer between 1994 and 2002. In 2010, the NSW CCR used ABS-recorded Aboriginal status from death registration data to retrospectively update all its deaths. If a person’s Aboriginal status was not known or missing on

the cancer registry before the update, it was replaced by the value from the ABS-coded death information; otherwise the original value was retained. If the ABS death file had non-cancer deaths with Aboriginal status recorded, they were also used to update Aboriginal status on the CCR. Additional linkage with the National Death Index is carried out every 3 years, which enables matching with NSW residents whose deaths may have been registered interstate. The last linkage between the CCR and the NDI was performed in 2010 for cases diagnosed up to December 31st 2006. This is likely to affect Aboriginal status recording in the CCR when ascertainment is via death registration. A small proportion of cancers (1–5%) are notified to the CCR via a Death Certificate Only (DCO).

Hospitals

Episode of care data are also collected on the CCR. For each cancer case, care episodes by hospital admission, chemotherapy, radiotherapy and other treatment are recorded. Each of these episodes represents an opportunity for Aboriginal status to be recorded. A recent review of the quality of Aboriginal identification in NSW hospitals showed that 93per cent of Aboriginal people were correctly identified in the admission record, and only 1per cent of patients had a ‘not reported’ status.9

Summary

Information on Aboriginal status in cancer registration comes from multiple sources, from episodes of care to death registration. Cancer records of cases subsequently found to be Aboriginal are updated retrospectively.

Despite the numerous sources of notification of Aboriginal status, completeness does not necessarily mean that status is recorded accurately, nor does it mean that Aboriginal people have notified that they are Aboriginal.

Notification of Aboriginal status from death certification will contribute more to status completeness in those cancers with low survival. This means there is potentially a disproportionate representation of deaths from these cancers amongst Aboriginal people. Conversely, cancers such as melanoma, which predominantly have only a pathology notification reported to the CCR, potentially will show a lower incidence in Aboriginal people than is the case.


67

Missing data

Missing data are the result of a ‘missing value mechanism’, which may not be clearly identified. In the case of surveys, and administrative health information, data can be missing for six generic reasons:

1. The question may not have been asked of a respondent or patient.

2. The question was asked but the subject did not respond for various reasons including refusal or ‘no comment’.

3. The question was asked and the respondent did not know the answer but there was no ‘not known’ response category in the data collection instrument or database.

4. The question was asked and the respondent responded but for some reason the response was not recorded.

5. The question was asked, the respondent answered, the response was recorded but the data were not entered.

6. The data are missing legitimately, as in a ‘not applicable’ response, but this category may not be in the data collection instrument or database.

In the case of recording Aboriginal status in the health and vital registration system, a policy of minimising (i) above was implemented in NSW and, as evidenced in the first quality and completeness report, has resulted in substantially reduced proportions of missing Aboriginal status in cancer registrations. However, these reductions have not been uniform: missing Aboriginal status remains high in some cancers, notably skin cancer and to a lesser extent urogenital cancer.

Missing data may cause two problems. First, missing data reduces statistical power and reduces the probability of rejecting a null hypothesis when it is false. Statistical power is reduced due to the loss of data. Second, missing data may produce biased parameter estimates.10 As a consequence, missing data may increase the probability of rejecting the null hypothesis when it is true.

Data can be missing in different ways. Data are said to be missing completely at random (MCAR) if the probability that an observation is missing does not depend on either the observed values or the missing values. Under MCAR, the cases with missing values for a variable are like a random sub sample of the original sample. Therefore the distribution of missing values is the same as the distribution of the observed values. Inferences from such data, using the observations with non-missing values, would be based on lower power but the estimates are unbiased.

Data are said to be missing at random (MAR) if the probability that the value for one variable for an observation is missing depends on the observed values of another variable but not on the missing values. Under MAR, the cases with missing values for a variable are like a simple random sample of observations within subgroups defined by the observed values. That is, within each subgroup the distribution of missing values is the same as the distribution of observed values. For the special case of uniform non-response within strata, unbiased overall estimates and valid inferences are derived from weighted stratum-specific estimates (eg stratum means) from the observations with non-missing values of the variable of interest in each stratum.

Data are said to be missing not at random (MNAR) if the probability that an observation is missing depends on the missing values after controlling for observed values.11, 12 Unfortunately, it is extremely difficult with MNAR data to establish the missing value mechanism, and valid inferences and unbiased estimates depend on knowledge of the missing data.

Complete case analysis

Complete case analysis (also known as listwise deletion) is commonly used due to its simplicity and accessibility. It is also the default in most standard statistical software, which simply excludes cases with any missing variable values from an analysis. These cases are called incomplete cases. While using only the complete cases has its simplicity, we lose information on the incomplete cases. Also, this approach ignores possible systematic differences between the complete cases and the incomplete cases and thus the resulting inferences may not be applicable to the population.10, 11, 13

Complete case analysis may be preferred where the proportion of missing data is small and the missing data mechanism is approximately MCAR.14 Based on the data presented in the quality and completeness report on the proportion of cases with missing Aboriginal status, cancer mortality from 1999 is of sufficient quality to be analysed using the Aboriginal status variable as is, based on complete case analysis. However, analysis of incidence using complete case analysis would be inappropriate since the proportion missing is too high, and systematic differences in missing status indicate that the data are not MCAR.


68

Multiple imputation (MI)

In order to properly analyse cancer incidence and survival in Aboriginal people in NSW, we used a technique that effectively handles missing data, so that differences in incidence, mortality and survival between people with known and unknown Aboriginal and/or Torres Strait Islander status are accounted for. This means that any potential recording bias is minimised.

Multiple imputation (MI) was developed in the mid-1970s by Donald Rubin as a robust general solution for making valid inferences in the presence of missing data.15 Multiple imputation has become increasingly popular as computational power has increased,16-21 and has become a standard procedure in popular statistical software packages, such as SAS.

In addition, complete-case analysis does not yield valid inferences when the missing status of covariates is related to survival, but if the proportion of missing status is low then multiple imputation can be used.22 The primary advantage of the MI method is that it leads to valid statistical inference in the presence of non-response. In addition, MI produces approximately unbiased parameter estimates under reasonable assumptions. In the case of Aboriginal status in cancer data, mortality information has been complete since the late 1990s, so imputation was applied to cancer incidence data only.

MI is a means for representing uncertainty in the missing data and produces a distribution of plausible values for a missing variable in a record, given the values of that record’s non-missing covariates. The contribution of the non-missing covariates to the distribution of missing values is based on modelling the non-missing co-variates in complete cases to predict the likely value of the variable in question in records where it is missing.

Markov chain Monte Carlo (MCMC) simulation is then used to create a small number (typically 5-10) of independent draws of records with a range of imputed values from the predictive distribution. The draws represent a random sample of the missing values and are then used for multiple-imputation inference such that a number of data sets of complete cases equal to the number of draws (with each data set incorporating a single draw) are then analysed using standard statistical analyses. The results from the ‘complete’ data sets are then combined to produce inferences. Standard errors for the missing variable reflect the uncertainty due to the missing values, thus leading to valid statistical inferences.

The key assumption of most methods for handling missing data is that of Missing Completely at Random (MCAR) – the observations that are missing are a simple random sample of all observations. This is clearly not the case for Aboriginal status, as was demonstrated in detail in the previous section.

However, MI requires only the more relaxed assumption of Missing at Random (MAR), which means that the missing data is a random sample of all the data after adjusting for all other observed (non-missing) variables.19 In other words, systematic differences between missing cases and the complete cases are allowed so long as all (or most) of the differences can be explained by the remaining observed variables.

MI is thus particularly suitable for estimating Aboriginal cancer rates as there are numerous covariates available in the Central Cancer Registry (such as area of residency, sex, age, country of birth and clinical variables) which makes it quite likely that the MAR assumption would be met.


69

Imputation model

The statistical modelling approach used to develop the imputation model for the MI procedure was logistic regression. No covariates other than Aboriginality had missing values, so a monotonic missing data pattern could be assumed, thus simplifying the estimation procedure. The covariates included in the model were age, the percentage of Aboriginal people living in the area of residence, remoteness of the area of residence, sex, an indicator for one-year survival, area health service, and indicators for type of cancer (Table 33).

The logistic regression model was fitted to cancer incidence data with a binary Aboriginal status (1, Aboriginal or Torres Strait Islander; 0, otherwise) as the outcome, and demographic and clinical variables available to the CCR as covariates. The following variables were used as predictors in the modelling (Table 33):

Table 33: Predictors used in multiple imputation model

Predictor Variable type Description

Age group Categorical Age at diagnosis categorised into 5-year age groups: 0-4, 5-9,…, 70-74,75+.

The 75+ age group was used as the reference category.

Sex Binary Male or female.

Female was used as the reference category.

Country of Birth Categorical Country of birth was grouped into 3 categories: Australia, Not Australia, and Unknown.

Australia was used as the reference category.

Degree of Spread Categorical Degree of Spread at diagnosis, coded as: 1=localised, 2=regionalised, 3=distant, and

9=unknown. Localised was used as the reference category.

Clinical Grouping Categorical The cancer site was grouped into one of 14 clinical groupings:

Skin, Head and Neck, Upper Gastrointestinal, Colorectal, Respiratory, Bone and

other connective tissue, Breast, Gynaecological, Eye, Neurological, Thyroid and other

endocrine, Lymphohaematopoeitic, Ill-defined and unknown primary sites, and Urogenital.

Respiratory was used as the reference category.

One-year survival Binary Indicator for survival of at least 12 months after diagnosis

Area Health Service Categorical Area Health Service of residence (2005 definition). The residential address was grouped

into one of the 8 Area Health Services of NSW: Sydney South West, South Eastern

Sydney & Illawarra, Sydney West, Northern Sydney & Central Coast, Hunter & New

England, North Coast, Greater Southern, Greater Western.

South Eastern Sydney & Illawarra was used as the reference category.

Percentage of local

population identifying as

Aboriginal.

Continuous

(0% – 100%)

The percentage of the population of the local government area of residence that identifies

as Aboriginal or Torres Strait Islander for that age group, sex and year, as estimated by the

Australian Bureau of Statistics population projections.*

*Source: Australian Bureau of Statistics population data (HOIST).

The MI procedure in SAS v. 9.1.3 was used to perform the analyses via the logistic regression method for monotone missing data.23

Evaluation of Imputation model

Most of the odds ratio estimates from the logistic regression model applied to incidence records with known Aboriginal status were statistically significant (Table 34). The concordance for the model was 74.4 per cent, and the percentage discordance was 20.0 per cent, suggesting that the model correctly classified Aboriginal status in most cases.


70

Table 34: Odds-ratio estimates from logistic regression modelling of covariates used to inform imputation of Aboriginal status.

Predictor Odds Ratios 95% CI

Age group Age Group 0-4 3.73 (2.48-5.61)

Age Group 5-9 4.43 (2.58-7.58)

Age Group 10-14 2.13 (1.07-4.24)

Age Group 15-19 3.45 (2.13-5.60)

Age Group 20-24 3.54 (2.34-5.36)

Age Group 25-29 3.35 (2.35-4.77)

Age Group 30-34 2.85 (2.08-3.90)

Age Group 35-39 4.05 (3.23-5.07)

Age Group 40-44 3.88 (3.20-4.69)

Age Group 45-49 3.30 (2.78-3.92)

Age Group 50-54 3.16 (2.71-3.69)

Age Group 55-59 2.55 (2.20-2.96)

Age Group 60-64 2.46 (2.14-2.84)

Age Group 65-69 1.88 (1.63-2.18)

Age Group 70-74 1.59 (1.38-1.85)

Age Group 75+ 1 referent

Sex Male 0.93 (0.85-1.02)

Female 1 referent

Country of Birth Not Australia 0.22 (0.19-0.25)

Unknown 0.90 (0.66-1.22)

Australia 1 referent

Degree of Spread Degree of Spread Regional 1.20 (1.07-1.33)

Degree of Spread Distant 1.27 (1.12-1.43)

Degree of Spread Unknown 1.05 (0.94-1.18)

Degree of Spread Localised 1 referent

Clinical Grouping Skin 0.24 (0.19-0.31)

Head and Neck 1.11 (0.91-1.35)

Upper Gastrointestinal 0.87 (0.74-1.01)

Colorectal 0.54 (0.46-0.62)

Bone and other connective tissue 0.53 (0.34-0.82)

Breast 0.58 (0.50-0.69)

Gynaecological 0.90 (0.74-1.09)

Urogenital 0.56 (0.48-0.64)

Eye 0.46 (0.21-1.00)

Neurological 0.57 (0.42-0.76)

Thyroid and other endocrine 0.44 (0.32-0.62)

Lymphohaematopoeitic 0.65 (0.54-0.77)

Ill-defined and unknown primary sites 0.64 (0.52-0.79)

Respiratory 1 referent

One-year survival Death within 1 year from Diagnosis 1.36 (1.25-1.49)

Survival of 1 year from Diagnosis 1 referent

Area Health Service Sydney South West AHS 0.74 (0.63-0.87)

Sydney West AHS 0.86 (0.73-1.00)

Northern Sydney & Central Coast AHS 0.59 (0.50-0.70)

Hunter & New England AHS 1.29 (1.13-1.47)

North Coast AHS 1.28 (1.11-1.49)

Greater Southern AHS 1.21 (1.03-1.41)

Greater Western AHS 1.83 (1.56-2.13)

South Eastern Sydney & Illawarra AHS 1 referent

% Aboriginal in LGA Percentage Aboriginal in Local Government Area 1.13 (1.12-1.15)


71

Comparison of analyses with and without multiple imputation

A recent report from the Australian Institute of Health and Welfare (AIHW) compared breast cancer incidence between the Aboriginal and non-Aboriginal populations in four Australian states and territories (Western Australia, South Australia, Northern Territory and Queensland).24 The method used to deal with missing data was case exclusion which affects the numerator of an incidence rate but not the denominator, and thus results in an underestimation of incidence in each sub-category. The reported incidence of both Aboriginal and non-Aboriginal females was less than that of the total, which includes unknown Aboriginal status in the numerator.

To demonstrate the effect for NSW, the same analysis was performed on NSW breast cancer using complete cases to compare with estimates from the multiple imputation method used in this report. As is evident, inappropriate treatment of missing data can produce misleading and inconsistent estimates (Table 35, Figure 53). The complete-case estimates do not sum to the total due to exclusion of missing data.

Table 35: Comparison of estimated incidence from complete case and multiple imputation approaches, NSW female breast cancer, 1999–2007

Complete case method Multiple Imputation Method

Rate1 95% CI Rate1 95% CI

Aboriginal 102.9 (90.1-116.8) 115.7 (101.2-130.3)

Non-Aboriginal 98.8 (97.7-99.9) 112.5 (111.3-113.7)

Total 112.5 (111.3-113.6) 112.4 (111.3-113.6)

Per 100,000, directly age standardised to 2001 Australian population

Figure 53: Comparison of estimated incidence from complete case and multiple imputation approaches, NSW female breast cancer, 1999–2007.

0

20

40

60

80

100

120

140

Aboriginal Non-Aboriginal Total

Per 1

00,0

00

Complete case Method Multiple Impution Method

Per

100,

000

Aboriginal Non-Aboriginal Total

Multiple imputation methodComplete case method


72

Age standardisation

As cancer risk increases rapidly with age, it is necessary to adjust for age differences between populations being compared. While differences in age distributions are responsible for differences in cancer incidence and mortality, such differences are not of primary interest when examining underlying reasons for different cancer rates in different populations.

The age distribution of the Aboriginal and non-Aboriginal populations of NSW age are markedly different. (25) In order to compare cancer rates between these two populations not attributable to age, direct age standardisation was used to account for the different age distributions. Rates presented in this section have been age- and sex-standardised to the Australian 2001 standard population. Rates standardised to the World Health Organisation standard population are also included (Appendix 1).

In the case of the Aboriginal population with its substantially younger age distribution compared to the 2001 Australian population standard, incidence and mortality rates in older Aboriginal age groups will be given a greater weight in calculations of direct standardised rates. Direct age standardisation essentially estimates event numbers, and a rate, that would occur in a standard population if it experienced the same age-specific rates as the study population. Consequently, age-standardised incidence and mortality rates in Aboriginal people generally will be considerably higher than corresponding crude rates. Indirect age standardised mortality ratios obviate this effect and are reported here also. Indirect age standardisation uses age-specific rates from the standard population to estimate expected event numbers in the study population. The indirectly standardised incidence or mortality ratio is then the observed events in the study population divided by the expected events for that population. In addition, cumulative risk to age 75 years was calculated, which is useful both for its ease of interpretation and for also being unaffected by wide differences in age distributions between populations compared.

Survival analysis

Cancer-specific survival analyses were conducted using Kaplan-Meier estimates, and differences between Aboriginal and non-Aboriginal strata were tested using the log-rank test. Kaplan-Meier (K-M) survival curves are convenient for displaying proportions of cases surviving over time, especially differences between different groups. K-M curves are estimated by calculating the proportion of individuals who survive to, or beyond, each failure time. The resultant curves are a step function of the surviving proportion of individuals against the time that has elapsed. The log-rank test, a non-parametric statistical test with an approximate chi-square distribution, is commonly used to test whether the survival curves are significantly different, and is used in this report.

For the survival analyses presented here, the outcome of interest is death from the cancer at diagnosis. Cancer cases that die of another cause, including from a cancer other than that diagnosed originally, are treated as censored observations and the date of death is the censoring date. Those still alive at the end of the study period (31 December 2008) are treated as censored observations and their date of censoring is 31 December 2008.


73

Conclusions

Until now, cancer incidence and mortality in NSW Aboriginal peoples has not been reliably estimable due to inadequate designation and recording of Aboriginal status, either when cancers are diagnosed and treated or when an Aboriginal person dies. Measures to address this problem were implemented with death registration in 1994 and throughout the NSW health system by 1999. Consequently, the proportion of unknown Aboriginal status in cancers has fallen considerably, so much so that we can now produce reliable estimates of Aboriginal cancer incidence and mortality for NSW. However, despite improved rates of recording of Aboriginal status in NSW cancer data, there is no guarantee that the quality of recording has improved: misclassification of Aboriginal status among those people with known Aboriginal status may not have changed at all, even though the proportion of people with known Aboriginal status may have increased.

The reduction in proportions of Aboriginal status unknown in NSW cancer data, to around 13 per cent of recorded cancer cases overall from 1999, has been the result of combined trends in Aboriginal recording rates. Rates of designation of Aboriginal status at death registration remain high (>95%), while recording rates in the health system have increased and continue to do so. The 19–29 per cent of live cancers with unknown Aboriginal status has been the outcome of these trends. Prior to 1999 the proportion of live cancers with unknown Aboriginal status was near 80 per cent.

Most of the excess cancer incidence and mortality in Aboriginal peoples is the result of higher incidence of lung, head and neck, stomach and kidney cancer in men, and lung, head and neck, uterine, and cervix cancer in women. Moreover, where cancer incidence has been similar or lower in Aboriginal peoples, survival has been worse, including from prostate and breast cancer. Most excess cancer incidence in Aboriginal peoples is uniform across age groups. While most of the mortality excess is due either to higher background incidence or later stage of cancer at diagnosis, excess mortality, and lower survival, also occurs in cancers of the same grade.

Clearly, the higher incidence and mortality from preventable cancers among Aboriginal peoples requires commitment and action by many to ‘close the gap’. However, from the age-specific analyses of cancer incidence and mortality (Part III above), it is also evident that the widest differential between Aboriginal and non-Aboriginal people is occurring in the oldest age groups, especially people aged 70+ years. Given that Aboriginal peoples also suffer substantially higher mortality from cardiovascular, endocrine and other non-communicable chronic disease at younger ages than the general population, then the excess cancer incidence and mortality in the older Aboriginal age groups can be viewed as that occurring in a surviving population denominator already considerably diminished by non-cancer causes of death. Premature mortality from non-cancer causes is thus contributing to higher rates of cancer incidence and mortality in Aboriginal peoples.

As set out in the NSW Cancer Plan 2011–15, the Cancer Institute NSW aims to reduce variations in cancer incidence and outcomes among Aboriginal peoples throughout NSW. Our strategies to achieve this objective include: ensuring programs and activities addressing the needs of Aboriginal peoples are developed by or in partnership with Aboriginal communities and peak bodies; developing the skills of the workforce, which provide services to Aboriginal people; developing culturally appropriate cancer services; generating evidence of effective cancer control strategies; integrating cancer into the primary health care for Aboriginal peoples; and the use of data and information to drive program and service performance improvement. To be effective, these efforts will take a holistic and system-wide approach and be guided by the principles of self-determination, equal partnership and mutual respect.


74

Appendix

Appendix 1: Cancer incidence and mortality to the world population Table A1: Cancer incidence in Aboriginal males NSW, 1999-2007(a)




1 in


1 in

Head & neck 96 27.3 (21.7-34.0) 36 6.4 103 29.3 (22.9-35.8) 34

Oesophagus 27 9.4 (6.0-14.1) 88 2.2 28 9.8 (5.8-13.8) 86

Stomach 46 17.2 (12.2-23.5) 55 3.1 47 17.4 (12.0-22.9) 54

Large bowel 147 51 (42.4-60.8) 19 6.4 158 54.4 (44.6-64.1) 17

Liver 29 8.9 (5.8-13.0) 83 2.1 30 9.1 (5.6-12.6) 81

Pancreas 32 11.7 (7.6-17.0) 88 1.1 32 11.8 (7.3-16.4) 87

Lung 200 75.2 (64.3-87.3) 13 2.1 206 77.1 (65.5-88.7) 12

Melanoma of skin 37 15.1 (10.2-21.4) 78 41.2 67 23.4 (15.4-31.4) 48

Mesothelioma 11 4.9 (2.3-8.9) 216 1.7 11 5 (1.9-8.1) 211

Prostate 183 76 (64.7-88.6) 13 20.0 239 98.1 (82.9-113.3) 10

Testis 21 4.2 (2.3-6.7) 431 12.8 25 4.9 (2.6-7.2) 344

Kidney 50 16.5 (11.7-22.3) 64 9.1 53 17.3 (12.1-22.6) 60

Bladder 30 13.8 (8.9-20.2) 111 5.3 31 14.3 (8.8-19.9) 105

Brain 20 4.4 (2.4-7.2) 273 2.8 21 4.6 (2.3-7.0) 262

Non-Hodgkin's lymphoma 47 12.4 (8.9-16.8) 69 10.2 53 14 (9.7-18.3) 62


All leukaemias 51 15.5 (10.8-21.2) 78 8.6 56 16.9 (11.5-22.3) 70

Unknown primary 59 20.7 (15.2-27.3) 49 4.7 61 21.3 (15.3-27.2) 47

Myelodysplasia 27 10.9 (6.8-16.4) 129 6.1 29 11.7 (6.8-16.7) 118

All cancers (c) 1,201 419 (393.1-446.0) 2.9 13.6 1,347 466.6 (437.0-496.2) 2.6

(a) Directly standardised to the WHO population.(b) Per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


75

Table A2: Cancer incidence in Non-Aboriginal males NSW, 1999-2007(a)





Head & neck 5,413 14.1 (13.7-14.5) 63 6.4 5,778 15.1 (14.7-15.4) 59

Oesophagus 2,158 5.3 (5.1-5.5) 179 2.2 2,206 5.4 (5.2-5.7) 174

Stomach 3,586 8.8 (8.5-9.1) 111 3.1 3,699 9.1 (8.8-9.4) 107

Large bowel 20,080 49.7 (49.0-50.4) 19 6.4 21,451 53.2 (52.5-53.9) 17

Liver 2,135 5.4 (5.2-5.7) 175 2.1 2,180 5.6 (5.3-5.8) 172

Pancreas 3,194 7.8 (7.5-8.0) 125 1.1 3,231 7.9 (7.6-8.1) 124

Lung 16,422 39.9 (39.3-40.5) 23 2.1 16,776 40.8 (40.1-41.4) 23

Melanoma of skin 10,113 25.8 (25.3-26.3) 40 41.2 17,201 45.1 (44.4-45.8) 23

Mesothelioma 1,408 3.4 (3.2-3.6) 282 1.7 1,433 3.4 (3.3-3.6) 276

Prostate 35,733 88.1 (87.2-89.0) 10 20.0 44,644 110.2 (109.1-111.2) 8

Testis 1,676 5.7 (5.5-6.0) 245 12.8 1,920 6.6 (6.3-6.8) 214

Kidney 4,503 11.5 (11.2-11.8) 81 9.1 4,955 12.7 (12.3-13.1) 73

Bladder 4,808 11.3 (11.0-11.6) 93 5.3 5,078 12 (11.7-12.3) 87

Brain 2,311 6.5 (6.2-6.8) 154 2.8 2,376 6.7 (6.4-7.0) 150

Non-Hodgkin's lymphoma 5,673 14.7 (14.3-15.1) 68 10.2 6,314 16.4 (16.0-16.8) 60

Multiple myeloma 2,041 5 (4.8-5.2) 195 6.3 2,179 5.3 (5.1-5.6) 180

All leukaemias 4,525 12.1 (11.8-12.5) 92 8.6 4,947 13.3 (12.9-13.7) 83

Unknown primary 5,534 13.2 (12.9-13.6) 81 4.7 5,809 13.9 (13.6-14.3) 77

Myelodysplasia 2,889 6.8 (6.6-7.1) 170 6.1 3,076 7.3 (7.1-7.6) 157

All cancers (c) 140,586 352.9 (351.0-354.8) 3.2 13.6 162,813 411 (409.0-413.0) 2.8

(a) Directly standardised to the WHO population.(b) Per 100,000 males.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


76

Table A3: Cancer incidence in Aboriginal females NSW, 1999-2007(a)




1 in


1 in

Head & neck 36 9.4 (6.5-13.2) 87 8.0 39 10.2 (6.8-13.6) 81

Oesophagus 14 4.2 (2.3-7.2) 207 1.6 14 4.3 (2.0-6.6) 207

Stomach 18 5.6 (3.2-9.0) 181 3.6 19 5.9 (3.1-8.7) 173

Large bowel 110 35.1 (28.6-42.6) 31 7.7 120 38.1 (30.7-45.5) 29

Liver 15 4.6 (2.5-7.6) 214 1.7 15 4.6 (2.2-7.0) 213

Pancreas 28 9.3 (6.1-13.6) 121 1.5 29 9.5 (5.9-13.1) 119

Lung 162 48.3 (40.9-56.6) 20 3.1 167 49.7 (41.8-57.6) 19

Melanoma of skin 30 8.1 (5.3-11.7) 148 46.7 62 15.2 (10.0-20.4) 74

Breast 304 78.1 (69.1-87.9) 14 12.2 346 88.4 (78.0-98.8) 12

Cervix 65 14.3 (10.9-18.4) 73 12.4 71 15.5 (11.5-19.5) 69

Uterus 43 13.2 (9.4-18.0) 75 8.7 50 14.9 (10.4-19.5) 65

Ovary 36 8.9 (6.1-12.5) 115 4.5 38 9.4 (6.2-12.7) 109

Kidney 31 8.7 (5.8-12.6) 114 10.0 34 9.4 (6.0-12.8) 105

Brain 22 5.2 (3.1-8.2) 177 2.5 23 5.3 (2.9-7.8) 174

Thyroid 24 4.8 (3.0-7.3) 199 12.2 28 5.7 (3.4-8.0) 173

Non-Hodgkin's lymphoma 26 6.8 (4.3-10.1) 119 11.4 33 8.6 (5.2-11.9) 96

All leukaemias 30 6.6 (4.1-9.8) 254 8.3 33 7.4 (4.4-10.4) 203

Unknown primary 44 12.6 (9.0-17.1) 77 4.1 47 13.2 (9.1-17.3) 73

Myelodysplasia 14 4.3 (2.3-7.4) 307 8.2 16 4.8 (2.2-7.4) 272

All cancers (c) 1,117 306.5 (287.8-326.0) 3.8 12.1 1,257 340.5 (319.7-361.3) 3.5

(a) Directly standardised to the WHO population.(b) Per 100,000 females.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


77

Table A4: Cancer incidence in Non-Aboriginal females NSW, 1999-2007(a)



Cases ASR b) 95%CI Risk to age 75 1 in

Cases ASR b) 95%CI Risk to age 75 1 in

Head & neck 1,999 4.6 (4.4-4.8) 211 8.0 2,173 5 (4.8-5.2) 193

Oesophagus 1,129 2.1 (2.0-2.2) 530 1.6 1,148 2.1 (2.0-2.3) 521

Stomach 1,957 3.9 (3.8-4.1) 266 3.6 2,030 4.1 (3.9-4.3) 253

Large bowel 16,567 34.6 (34.0-35.1) 28 7.7 17,953 37.7 (37.1-38.3) 26

Liver 905 1.9 (1.8-2.0) 540 1.7 920 1.9 (1.8-2.0) 533

Pancreas 3,153 6.1 (5.9-6.3) 169 1.5 3,199 6.2 (5.9-6.4) 166

Lung 9,159 19.6 (19.2-20.0) 46 3.1 9,454 20.3 (19.8-20.7) 44

Melanoma of skin 6,392 15.5 (15.1-15.9) 67 46.7 11,991 30.6 (30.0-31.1) 34

Breast 31,354 78.6 (77.7-79.5) 13 12.2 35,698 89.6 (88.6-90.5) 11

Cervix 1,911 5.1 (4.9-5.4) 213 12.4 2,184 6 (5.7-6.2) 186

Uterus 4,433 10.5 (10.1-10.8) 85 8.7 4,855 11.5 (11.2-11.8) 77

Ovary 3,537 8.3 (8.0-8.6) 117 4.5 3,704 8.8 (8.5-9.1) 112

Kidney 2,772 6 (5.8-6.3) 159 10.0 3,078 6.8 (6.5-7.0) 140

Brain 1,639 4.3 (4.1-4.5) 241 2.5 1,680 4.4 (4.2-4.6) 235

Thyroid 3,201 9.3 (8.9-9.6) 122 12.2 3,645 10.5 (10.2-10.9) 107


All leukaemias 3,114 7.4 (7.1-7.7) 159 8.3 3,397 8.1 (7.8-8.4) 142

Unknown primary 5,475 10.5 (10.2-10.8) 107 4.1 5,707 11.1 (10.8-11.4) 101

Myelodysplasia 2,212 4.2 (4.0-4.4) 282 8.2 2,407 4.7 (4.5-4.9) 249

All cancers (c) 113,879 261.2 (259.6-262.8) 4.3 12.1 129,625 301.5 (299.8-303.2) 3.8

(a) Directly standardised to the WHO population.(b) Per 100,000 females.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


78

Table A5: Cancer incidence in Aboriginal persons NSW, 1999-2007 (a)




1 in


1 in

Head & neck 132 17.8 (14.7-21.3) 52 6.8 142 19.1 (15.7-22.6) 48

Oesophagus 41 6.6 (4.6-9.0) 128 2.0 42 6.8 (4.6-8.9) 125

Stomach 64 10.7 (8.1-13.9) 88 3.3 66 11 (8.1-13.8) 86

Large bowel 257 42.7 (37.3-48.6) 24 7.0 278 45.8 (39.9-51.8) 22

Liver 44 6.6 (4.7-9.0) 124 2.0 45 6.7 (4.6-8.8) 122

Pancreas 60 10.4 (7.8-13.6) 104 1.3 61 10.6 (7.7-13.4) 102

Lung 362 59.8 (53.5-66.7) 16 2.5 373 61.5 (54.8-68.2) 15

Melanoma of skin 67 11 (8.3-14.2) 105 43.5 130 18.6 (13.8-23.5) 59

Breast 306 42.4 (37.5-47.9) 25 12.2 348 47.9 (42.2-53.6) 22

Cervix 65 7.6 (5.7-9.8) 138 12.4 71 8.2 (6.1-10.4) 130

Uterus 43 7.3 (5.1-9.9) 140 8.7 50 8.2 (5.7-10.7) 123

Prostate 183 33.4 (28.6-38.8) 27 20.0 239 43.2 (36.7-49.7) 21

Kidney 81 12.2 (9.5-15.4) 84 9.4 87 13 (10.0-15.9) 78

Bladder 39 7.2 (5.0-10.0) 167 5.2 41 7.6 (5.1-10.1) 158

Brain 42 4.9 (3.4-6.8) 209 2.6 44 5.1 (3.4-6.8) 203

Non-Hodgkin's lymphoma 73 9.6 (7.4-12.2) 88 10.7 86 11.3 (8.5-14.0) 76

All leukaemias 81 10.5 (8.0-13.5) 124 8.5 89 11.6 (8.7-14.5) 108

Unknown primary 103 16.1 (13.0-19.8) 61 4.4 107 16.7 (13.3-20.2) 58

Myelodysplasia 41 7.1 (5.0-9.8) 188 7.0 45 7.7 (5.2-10.3) 170

All cancers (c) 2,318 354.6 (339.3-370.4) 3.3 13.0 2,604 394.3 (376.6-412.0) 3.0

(a) Directly standardised to the WHO population.(b) Per 100,000 persons.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


79

Table A6: Cancer incidence in Non-Aboriginal persons NSW, 1999-2007(a)





Head & neck 7,412 9.1 (8.9-9.3) 98 6.8 7,951 9.8 (9.6-10.1) 92

Oesophagus 3,287 3.6 (3.5-3.8) 270 2.0 3,354 3.7 (3.6-3.8) 264

Stomach 5,543 6.2 (6.0-6.4) 159 3.3 5,729 6.4 (6.2-6.6) 152

Large bowel 36,647 41.6 (41.2-42.1) 23 7.0 39,404 44.9 (44.5-45.4) 21

Liver 3,040 3.6 (3.4-3.7) 268 2.0 3,100 3.7 (3.5-3.8) 262

Pancreas 6,347 6.9 (6.7-7.1) 144 1.3 6,430 7 (6.8-7.2) 142

Lung 25,581 28.8 (28.5-29.2) 31 2.5 26,230 29.6 (29.2-29.9) 30

Melanoma of skin 16,505 20.2 (19.9-20.5) 50 43.5 29,191 37.2 (36.8-37.7) 27

Breast 31,601 40.6 (40.2-41.1) 24 12.2 35,984 46.3 (45.8-46.8) 21

Cervix 1,911 2.6 (2.5-2.7) 420 12.4 2,184 3 (2.9-3.2) 368

Uterus 4,433 5.4 (5.3-5.6) 166 8.7 4,855 6 (5.8-6.1) 151

Prostate 35,733 41.2 (40.8-41.6) 20 20.0 44,644 51.5 (51.1-52.0) 16

Kidney 7,275 8.6 (8.4-8.8) 108 9.4 8,033 9.6 (9.4-9.8) 97

Bladder 6,608 6.9 (6.8-7.1) 148 5.2 6,974 7.4 (7.2-7.6) 138

Brain 3,950 5.3 (5.2-5.5) 188 2.6 4,056 5.5 (5.3-5.7) 184


All leukaemias 7,639 9.6 (9.4-9.8) 117 8.5 8,344 10.5 (10.3-10.7) 105

Unknown primary 11,009 11.8 (11.6-12.0) 92 4.4 11,517 12.4 (12.1-12.6) 88

Myelodysplasia 5,101 5.4 (5.2-5.6) 214 7.0 5,483 5.9 (5.7-6.0) 194

All cancers (c) 254,465 301.9 (300.7-303.1) 3.7 13.0 292,438 350.3 (349.0-351.6) 3.2

(a) Directly standardised to the WHO population.(b) Per 100,000 persons.(c) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


80

Table A7: Cancer mortality in Aboriginal and Non-Aboriginal males NSW, 1999-2007(a)

Cancer sites



1 in

M:I Ratio(c)


1 in

M:I Ratio(c)

Head & neck 54 15.9 (11.6-21.2) 56 0.6 2,325 5.8 (5.6-6.0) 155 0.4

Oesophagus 25 9.2 (5.8-13.9) 86 1.0 1,684 4.1 (3.9-4.3) 241 0.8

Stomach 34 14.1 (9.4-20.1) 71 0.8 2,535 6.1 (5.9-6.4) 165 0.7

Large bowel 68 25.6 (19.2-33.2) 45 0.5 8,012 19.2 (18.7-19.6) 52 0.4

Liver 20 6.1 (3.5-9.6) 135 0.7 1,571 3.9 (3.7-4.1) 250 0.7

Pancreas 30 12.3 (7.9-18.0) 94 1.1 2,806 6.8 (6.5-7.0) 147 0.9

Lung 164 63 (53.0-74.2) 14 0.8 13,706 32.8 (32.3-33.4) 29 0.8

Melanoma of skin 12 5.1 (2.4-9.2) 174 0.4 2,560 6.3 (6.0-6.5) 173 0.3

Prostate 65 33.6 (25.5-43.4) 51 0.5 8,392 18.3 (17.9-18.7) 86 0.2

Kidney 19 7.1 (4.0-11.5) 159 0.5 1,664 4 (3.8-4.2) 254 0.4

Bladder 10 4.5 (1.9-8.7) 366 0.3 1,886 4.2 (4.0-4.4) 316 0.4

Brain 14 3.5 (1.7-6.2) 346 0.8 1,813 4.9 (4.7-5.1) 191 0.8

Non-Hodgkin's

lymphoma

18 5.9 (3.3-9.6) 169 0.5 2,478 6 (5.8-6.2) 176 0.4

All leukaemias 27 10.6 (6.6-16.0) 115 0.7 2,424 6 (5.8-6.3) 185 0.5

Unknown primary 48 17.5 (12.4-23.7) 57 0.9 4,339 10.3 (9.9-10.6) 106 0.8

All cancers (d) 657 251.1 (230.4-273.1) 4.7 0.6 63,177 150.7 (149.6-151.9) 7.5 0.5

(a) Directly standardised to the WHO population.(b) Per 100,000 males.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


81

Table A8: Cancer mortality in Aboriginal and Non-Aboriginal females NSW, 1999-2007a)

Cancer sites



M:I Ratio(c) Deaths ASR(b) 95%CI Risk to age 75 1 in

M:I Ratio(c)

Head & neck 18 5.1 (3.0-8.2) 135 0.6 811 1.7 (1.5-1.8) 608 0.4

Oesophagus 11 3.2 (1.5-5.9) 316 0.8 874 1.6 (1.5-1.7) 787 0.8

Stomach 17 5.3 (3.0-8.7) 235 1.0 1,407 2.7 (2.5-2.8) 433 0.7

Large bowel 49 15.6 (11.4-20.9) 81 0.5 6,604 12.6 (12.3-13.0) 87 0.4

Liver 10 3.3 (1.5-6.1) 260 0.7 696 1.4 (1.3-1.5) 768 0.7

Pancreas 29 9.9 (6.5-14.3) 118 1.1 2,800 5.3 (5.1-5.5) 201 0.9

Lung 134 40.9 (34.0-48.7) 25 0.9 7,281 15 (14.6-15.4) 62 0.8

Melanoma of skin 9 2.8 (1.3-5.5) 325 0.4 1,182 2.5 (2.3-2.6) 442 0.2

Breast 99 26.3 (21.1-32.3) 37 0.3 7,887 17.4 (17.0-17.8) 59 0.2

Cervix 31 8 (5.3-11.6) 123 0.6 680 1.6 (1.5-1.7) 663 0.3

Uterus 13 4.3 (2.2-7.4) 259 0.3 982 2 (1.8-2.1) 490 0.2

Ovary 22 6.2 (3.8-9.5) 157 0.7 2,281 4.8 (4.6-5.0) 203 0.6

Kidney 16 4.9 (2.7-8.0) 188 0.6 1,160 2.2 (2.1-2.3) 469 0.4

Brain 15 4.1 (2.1-6.9) 208 0.8 1,244 3 (2.8-3.2) 315 0.7

Non-Hodgkin's

lymphoma

13 3.7 (1.9-6.5) 213 0.6 2,164 4.1 (3.9-4.3) 277 0.4

All leukaemias 15 3.9 (2.0-6.6) 389 0.6 1,705 3.5 (3.3-3.6) 345 0.5

Unknown primary 35 10.2 (7.0-14.4) 103 0.8 4,491 8.4 (8.1-8.6) 138 0.8

All cancers (d) 572 169 (154.8-184.2) 6.5 0.6 48,932 98.4 (97.5-99.3) 11.1 0.4

(a) Directly standardised to the WHO population.(b) Per 100,000 females.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


82

Table A9: Cancer mortality in Aboriginal and Non-Aboriginal persons NSW, 1999-2007 (a)

Cancer sites



1 in

M:I Ratio(c)


1 in

M:I Ratio(c)

Head & neck 72 10.2 (7.9-13.0) 80 0.6 3,136 3.6 (3.5-3.7) 250 0.4

Oesophagus 36 5.9 (4.1-8.3) 143 0.9 2,558 2.7 (2.6-2.8) 374 0.8

Stomach 51 9.1 (6.7-12.2) 114 0.9 3,942 4.3 (4.1-4.4) 242 0.7

Large bowel 117 20 (16.3-24.3) 59 0.5 14,616 15.6 (15.4-15.9) 66 0.4

Liver 30 4.6 (3.0-6.7) 181 0.7 2,267 2.6 (2.5-2.7) 382 0.7

Pancreas 59 10.9 (8.2-14.2) 106 1.1 5,606 6 (5.8-6.1) 171 0.9

Lung 298 50.4 (44.6-56.8) 19 0.9 20,987 23.1 (22.8-23.4) 40 0.8

Melanoma of skin 21 3.7 (2.2-5.8) 239 0.4 3,742 4.2 (4.1-4.4) 252 0.2

Breast 100 14.5 (11.6-17.9) 69 0.4 7,944 9.2 (9.0-9.4) 114 0.2

Cervix 31 4.4 (2.9-6.4) 229 0.6 680 0.8 (0.8-0.9) 1,305 0.3

Ovary 22 3.3 (2.0-5.2) 297 0.8 2,281 2.6 (2.4-2.7) 396 0.6

Prostate 65 13.9 (10.6-17.8) 113 0.5 8,392 7.9 (7.7-8.1) 177 0.2

Kidney 35 5.8 (4.0-8.2) 173 0.5 2,824 3.1 (2.9-3.2) 331 0.4

Bladder 16 2.9 (1.6-4.8) 431 0.4 2,764 2.7 (2.6-2.8) 492 0.4

Brain 29 3.9 (2.4-5.7) 250 0.8 3,057 3.9 (3.8-4.1) 239 0.8

Non-Hodgkin's

lymphoma

31 4.8 (3.1-6.9) 185 0.6 4,642 5 (4.8-5.1) 217 0.4

All leukaemias 42 6.7 (4.7-9.3) 186 0.7 4,129 4.6 (4.5-4.8) 243 0.5

Unknown primary 83 13.4 (10.5-16.8) 75 0.8 8,830 9.2 (9.0-9.4) 120 0.8

All cancers (d) 1,229 203.7 (191.7-216.2) 5.5 0.6 112,109 121.7 (121.0-122.4) 9.0 0.4

(a) Directly standardised to the WHO population. (b) Per 100,000 persons.(c) M:I Ratio - Mortality rate versus Multiple Imputation incidence rate(d) All cancers include C00-C96, D45-D47 and exclude non-melanoma skin cancer.


83

Appendix 2: Glossary

Glossary and Terminology

Term Definition

Age-specific rate A rate of incidence or mortality for a specified age group. The numerator and denominator refer to the

same age group. For cancer, it is usually expressed per 100,000 people in the population per year.

Age-standardised rate Age standardisation is a set of techniques used to remove as far as possible the effects of differences in age

when comparing rates for two or more populations.

Cancer incidence Cancer cases diagnosed in a defined population during a specified period, usually expressed as a rate per

100,000 people.

Cancer mortality Deaths from cancer in a defined population during a specified period, usually expressed as a rate per

100,000 people.

Cases These are individual cancers. A person may have more than one cancer giving rise to multiple cases in the

same person. Second cancers in one person are only counted if they are of different cell type or originate

in a different organ. Metastatic cancers are not counted as cases.

Confidence Interval A range of values for rate (or other statistic) than has a 95% probability of including the true value of the

rate (or other statistic).

Cumulative rate A weighted sum of age-specific incidence rates, up to a specified maximum age. The weights are the widths

of the age groups. The cumulative rate is not a rate as such, but a dimensionless number most conveniently

expressed as a percentage.

Cumulative risk The probability that an event of interest (such as diagnosis or death from a particular cancer) will occur for

a person over a specified age range in the absence of any competing cancer diagnosis or cause of death.

Degree of Spread This is the maximum extent of disease within four months of the first date of diagnosis. This is a summary

stage that combines information from clinical and pathological documentation on the extent of disease.

Lifetime risk The risk that a person will be diagnosed with or die from cancer during their life. In this report, cumulative

risk for the age period 0 to 74 years is used to approximate lifetime risk.

Significant difference Rates are considered significantly different if their confidence intervals do not overlap, although some

differences are significant when confidence interval overlap.

Site The place in the body (e.g., and organ) where the cancer occurs.

Topography The site where the cancer originated.


84

Appendix 3: Abbreviations

Abbreviation Definition

ABS Australian Bureau of Statistics

AIHW Australian Institute of Health and Welfare

CCR Central Cancer Registry

DCO Death Certificate Only notification to the Central Cancer Registry

ICD-10 International Classification of Diseases version 10

K-M Kaplan Meier calculation (see methods)

MAR Missing at Random (see methods in relation to Multiple Imputation)

MCMC Markov Chain Monte Carlo modelling (see methods in relation to Multiple Imputation)

MCAR Missing Completely at Random (see methods in relation to Multiple Imputation)

MNAR Missing Not at Random (see methods in relation to Multiple Imputation)

MI Multiple Imputation (See methods)

NDI National Death Index

NSW New South Wales

RBDM Registrar of Births Deaths and Marriages

SIR Standardised Incidence Ratio (see methods)

SMR Standardised Mortality Ration (see methods)

TNM Tumour Node Metastases


85

Appendix 4: References

1. Australian Institute of Health and Welfare. National best practice guidelines for collecting Indigenous status in health data sets.: Canberra : AIHW.2010.

2. South Australian Cancer Register. Cancer incidence mortality and case survival in the South Australian Aboriginal population. Epidemiology of cancer in South Australia, 1977 to 1996. South Australian Cancer Registry Adelaide: South Australian Health Commission1997.

3. Threlfall TJ, JR. T. Cancer incidence and mortality in Western Australia, 1996. Perth: Health Department of Western Australia1998.

4. NSW Department of Health. Better Practice Guidelines to Improve the Level of Aboriginal and Torres Strait Islander Identification in the NSW Public Health System. Sydney: NSW Health2000.

5. NSW Health: Centre for Aboriginal Health. National Partnership Agreement on Closing the Gap in Indigenous Health Outcomes: Implementation Plan Jurisdiction: New South Wales. Sydney: NSW Health; 2008 [cited 2011 28 March ]; Available from: http:// www.health.nsw.gov.au/resources/publichealth/aboriginal/national_partnership_agreement_closing_the_gap_in_indigenous_health_ outcomes_implementation_plan.pdf.

6. NSW Department of Health. Communicating positively; A guide to appropriate Aboriginal terminology. Sydney: NSW Health. Available from http://health. nsw.gov.au/pubs/2004/pdf/aboriginal_terms.pdf. Accessed 28 March 20112004.

7. Australian Institute of Health and Welfare. National best practice guidelines for collecting Indigenous status in health data sets. Cat. no. IHW 29. Canberra: AIHW; 2010a.

8. International Agency for Research on Cancer WHO, & International Association of Cancer Registries. Coding. In D. Esteban, S. Whelan, A. Laudico, & D. Parkin (Eds.),. Manual for Cancer Registry Personnel, IARC Technical Report No 10. Lyon1995.

9. Australian Institute of Health and Welfare. Indigenous identification in hospital separations data-quality report. Canberra: AIHW2010b.

10. Allison PD. Missing Data:Paper Series on Quantitative Applications in the Social Sciences,. In: Sage, editor. Thousand Oaks, CA: Sage University; 2001. p. 07-136.

11. Schafer JL, Graham JW. Missing data: Our view of the state of the art. Psychological methods. 2002;7(2):147-77.

12. Zhang P. Multiple imputation: theory and method. International Statistical Review. 2003;71(3):581-92.

13. Little JR. Regression with Missing X’s: A Review. Journal of the American Statistical Association 1992;87(December):1227-37.

14. King G. Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation. American Political Science Review. 2001; 95(1):49-69.

15. Rubin DB. Multiple Imputation After 18+ Years. Journal of the American Statistical Association. 1996;91(434):473-89.

16. Barnard J, Meng X. Applications of multiple imputation in medical studies: from AIDS to NHANES. . Statistical Methods in Medical Research. 1999;8(1):17-36.

17. Clogg CC, Rubin DB, Schenker N, Schultz B, Weidman L. Multiple Imputation of Industry and Occupation Codes in Census Public- Use Samples Using Bayesian Logistic Regression. Journal of the American Statistical Association. 1991;86(413):68-78.

18. Horton NJ, Kleinman KP. Much ado about nothing: A comparison of missing data methods and software to fit incomplete data regression models. The merican statistician. 2007;61(1): 79-90.

19. Schafer JL. Multiple imputation: a primer. Statistical Methods in Medical Research. 1999;8(1):3-15.

20. Schenker N, Raghunathan TE, Chiu PL, Makuc DM, Zhang G, Cohen AJ. Multiple imputation of missing income data in the National Health Interview Survey. Journal of the American Statistical Association. 2006;101(475):924-33.

21. Stuart EA, Azur M, Frangakis C, P L. Multiple Imputation With Large Data Sets: A Case Study of the Children’s Mental Health Initiative. American Journal of Epidemiology. 2009;169(9):1133-9.

22. Burton A, Altman DG. Missing covariate data within cancer prognostic studies: a review of current reporting and proposed guidelines. British Journal of Cancer. 2004;91(1):4-8.

23. SAS Institute. SAS/STAT 9.2: Logistic Regression Method for Monotone Missing Data [cited 2010 2/11/2010]; Available from: http:// support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_mi_sect023.htm.

24. Australian Institute of Health and Welfare & National Breast and Ovarian Cancer Centre. Breast Cancer in Australia: An Overview 2009. Canberra.: AIHW2009.

25. Australian Bureau of Statistics. Population Distribution of Aboriginal and Torres Strait Islander Australians. Canberra: Australian Bureau of Statistics2006.

incidence, mortality and survival - cancer institute nsw · survival by aboriginal status and major...

Documents