BREAST SURGERY ANZJSurg.com
Breast cancer survival in New Zealand women Ian D. Campbell,* Nina Scott,† Sanjeewa Seneviratne,* James Kollias,‡§¶ David Walters,‡ Corey Taylor,** Fleur Webster,†† Helen Zorbas†† and David M. Roder††‡‡ *Waikato Clinical School, Faculty of Medical and Health Sciences, University of Auckland, Hamilton, New Zealand †Te Puna Oranga, Maori Health, Waikato District Health Board, Hamilton, New Zealand ‡National Breast Cancer Audit Steering Committee, Breast Surgeons of Australia and New Zealand, Sydney, New South Wales, Australia §Breast Endocrine and Surgical Oncology Unit, Royal Adelaide Hospital, Adelaide, South Australia, Australia ¶Department of Surgery, University of Adelaide, Adelaide, South Australia, Australia **Breast Surgeons of Australia and New Zealand Quality Audit, Royal Australasian College of Surgeons, Adelaide, South Australia, Australia ††Population Health, Cancer Australia, Sydney, New South Wales, Australia and ‡‡School of Population Health, University of South Australia, Adelaide, South Australia, Australia
Key words breast cancer, ethnicity, risk factor, survival. Correspondence Professor David M. Roder, Cancer Australia, Locked Bag 3, Strawberry Hills, NSW 2012, Australia. Email: [email protected] I. D. Campbell MB, FRACS; N. Scott FNZCPHM, MBChB; S. Seneviratne MBBS, MD; J. Kollias FRACS, MD; D. Walters MBBS, DDU, FRACS; C. Taylor BSc, GDipPsych; F. Webster MPH, MScVSc; H. Zorbas MBBS, FASBP; D. M. Roder MPH, DDSc. Accepted for publication 16 August 2014.
Abstract Background: The Quality Audit (BQA) of Breast Surgeons of Australia and New Zealand includes a broad range of data and is the largest New Zealand (NZ) breast cancer (BC) database outside the NZ Cancer Registry. We used BQA data to compare BC survival by ethnicity, deprivation, remoteness, clinical characteristic and case load. Methods: BQA and death data were linked using the National Health Index. Diseasespecific survival for invasive cases was benchmarked against Australian BQA data and NZ population-based survivals. Validity was explored by comparison with expected survival by risk factor. Results: Compared with 93% for Australian audit cases, 5-year survival was 90% for NZ audit cases overall, 87% for Maori, 84% for Pacific and 91% for other. Conclusions: BC survival in NZ appears lower than in Australia, with inequities by ethnicity. Differences may be due to access, timeliness and quality of health services, patient risk profiles, BQA coverage and death-record methodology.
doi: 10.1111/ans.12851
Introduction The Quality Audit (BQA) undertaken by Breast Surgeons of Australia and New Zealand (BreastSurgANZ) includes data on breast cancer (BC) and its treatment since 1998.1 BQA data are collected to provide feedback to participating surgeons on their patients, the cancers they encounter and clinical practices and outcomes.1,2 The data support quality audit by surgeons of their practices, allow outcomes to be assessed and provide opportunities for research. Operational details have been described in earlier publications.1,2 Surgeons obtain from the BQA database feedback on their own clinical practices, and aggregated data for all surgeons participating in the BQA, allowing self-audit against key quality indicators.2,3 Originally, members of the Breast Section of the Royal Australasian College of Surgeons were encouraged to submit their data to the ANZ J Surg 85 (2015) 546–552
BQA, but since late 2010, it has been a requirement for full membership of BreastSurgANZ that all patients are audited through the BQA. The BQA database includes linked death data, enabling BC survival to be analysed by patient factors, access, timeliness and quality of treatment, case load and other descriptors.1 Clinical outcomes can be compared across the patient population and internationally. These practices accord with objective risk stratification practices and outcome measures recommended by safety and quality organizations for clinical quality assessment.4 In this report, data from the New Zealand (NZ) component of the BQA are used to describe patient and BC characteristics and survival. Death data were obtained by linking BQA data to NZ death records. This is a process routinely undertaken between cancer registry and death data around the world to produce survival data, including with NZ Cancer Registry data.5–9 © 2014 Royal Australasian College of Surgeons
Breast cancer survival
BQA data were not population based during the roll-out of the BQA. The database covered around 50% of early BC in NZ across the 1998–2010 study years, increasing to 80% by 2008. The BQA database is the largest nationwide repository of NZ BC data outside the NZ Cancer Registry. BQA data provide a unique opportunity to analyse clinical features and how survival differs by patient factors, such as ethnicity, socio-economic status of residential area and remoteness, and by conventional cancer prognostic factors such as size, nodal status and grade.10,11 Differences in BC survival are known to exist in NZ,9 but BQA data on BC cancer survival by ethnicity have not been previously published. Results are discussed in the context of populationbased survival data from the NZ Cancer Registry,9 Australian BQA data1 and expected survival patterns by established clinical risk factors.10,11
Methods The study included 10 824 early invasive BC diagnosed in NZ women between 1 January 1998 and 31 December 2010 who were treated by breast surgeons participating in the BQA.1 Cases excluded 267 where data collected were incomplete and insufficient for analysis. The distribution of cases by three-yearly diagnostic epoch was 11.1% prior to 2001; 19.5% in 2001–2003; 27.0% in 2004–2006; and 35.9% in 2007–2009; plus a further 6.5% already entered for 2010. Death data were obtained from the NZ National Mortality Collection by linking with the BQA database using the National Health Index (NHI).12 The NHI comprises a unique alphanumeric identifying number for each New Zealander.12 Death data were traced to 31 December 2010. Results were compared with data from the Australian BQA and survival recorded by the NZ Cancer Registry. The BQA is directed at early BC defined similarly to local and regional spread of disease, as recorded by the US Surveillance Epidemiology and End Results (SEER) Program and the NSW and NZ Cancer Registries.1,7,9,13 However, regional cancers with fixed nodes, which are regarded as regional by SEER and the NSW and NZ Registries, are not included in the BQA definition of early BC.13 BQA data in this study included (i) person descriptors – age at diagnosis (years), ethnicity (Maori, Pacific, other), geographic remoteness of residence from a regional cancer centre classified as 0–9 km, 10–100 km and >100 km, and deprivation index of residential area (ordinal scale from 1 (least deprived) to 4 (most deprived));14 (ii) tumour descriptors – histology type (ductal, lobular, other), tumour size (mm), grade (low, intermediate, high), nodal status, oestrogen and progesterone receptor status, HER-2 receptor status, phenotype (basal (triple negative), luminal A, luminal B, ER-/PR-/HER-2 + , other), vascular invasion, tumour location (ICDO-3 classification), whether unilateral or synchronous bilateral tumour, and number of cancer foci; and (iii) other descriptors – diagnostic year, referral type (symptomatic, BreastScreen, other) and annual surgeon case load (≤10, 11–30, 31–100, 101+). Traditional Kaplan–Meier disease-specific survival from BC was calculated, regarding as a BC death those deaths where BC was recorded as the direct or antecedent cause on the death certifi© 2014 Royal Australasian College of Surgeons
547
cate.4,15,16 Disease-specific survival using this definition is a valid epidemiological option for population-based studies when there is an adequate standard of death certificate recording. This option has been shown to give similar results to relative survival in populationbased studies of BC.17–19 Disease-specific survival may be preferred where patients in clinical studies may not be entirely representative of the population and may have risks of causes of death that do not equate with population norms. Traditional disease-specific survival also was used in an earlier study of the Australian component of the BQA database and its use for the NZ study was designed to produce comparative data.1 Survival times were calculated from date of diagnosis to 31 December 2010 or time of death, whichever occurred first. Hazard ratios for BC death (95% confidence limits) were calculated by patient and cancer descriptor for the follow-up period, using Cox proportional hazards regression, and by applying the same censoring criteria as for the Kaplan–Meier analyses.15,16 Survival analyses were repeated using only cases with complete data for all variables (i.e. ‘complete case analyses’) and by using competing risk regression.20 Results were very similar to those presented from the traditional Kaplan–Meier and proportional hazards regression methods and are not presented in this report. In particular, hazard ratios for variables with higher numbers of missing values (due to not being core items throughout the study period) were similar in the complete case analysis and full analysis. Imputation of missing values was not undertaken due to similarities of the results from the full and complete case analyses.
Results Survival outcomes (unadjusted predictors) Overall BC survival was 90% at 5 years from diagnosis and 84% at 10 years. Differences by patient and cancer characteristic, referral type and surgeon case load are shown in Table 1. Hazard ratios (95% confidence limits) indicated the following differences: • By demographic characteristic: ○ Age at diagnosis: lower hazard ratios for 40–69-year-olds than the reference age (
Breast cancer survival in New Zealand women Ian D. Campbell,* Nina Scott,† Sanjeewa Seneviratne,* James Kollias,‡§¶ David Walters,‡ Corey Taylor,** Fleur Webster,†† Helen Zorbas†† and David M. Roder††‡‡ *Waikato Clinical School, Faculty of Medical and Health Sciences, University of Auckland, Hamilton, New Zealand †Te Puna Oranga, Maori Health, Waikato District Health Board, Hamilton, New Zealand ‡National Breast Cancer Audit Steering Committee, Breast Surgeons of Australia and New Zealand, Sydney, New South Wales, Australia §Breast Endocrine and Surgical Oncology Unit, Royal Adelaide Hospital, Adelaide, South Australia, Australia ¶Department of Surgery, University of Adelaide, Adelaide, South Australia, Australia **Breast Surgeons of Australia and New Zealand Quality Audit, Royal Australasian College of Surgeons, Adelaide, South Australia, Australia ††Population Health, Cancer Australia, Sydney, New South Wales, Australia and ‡‡School of Population Health, University of South Australia, Adelaide, South Australia, Australia
Key words breast cancer, ethnicity, risk factor, survival. Correspondence Professor David M. Roder, Cancer Australia, Locked Bag 3, Strawberry Hills, NSW 2012, Australia. Email: [email protected] I. D. Campbell MB, FRACS; N. Scott FNZCPHM, MBChB; S. Seneviratne MBBS, MD; J. Kollias FRACS, MD; D. Walters MBBS, DDU, FRACS; C. Taylor BSc, GDipPsych; F. Webster MPH, MScVSc; H. Zorbas MBBS, FASBP; D. M. Roder MPH, DDSc. Accepted for publication 16 August 2014.
Abstract Background: The Quality Audit (BQA) of Breast Surgeons of Australia and New Zealand includes a broad range of data and is the largest New Zealand (NZ) breast cancer (BC) database outside the NZ Cancer Registry. We used BQA data to compare BC survival by ethnicity, deprivation, remoteness, clinical characteristic and case load. Methods: BQA and death data were linked using the National Health Index. Diseasespecific survival for invasive cases was benchmarked against Australian BQA data and NZ population-based survivals. Validity was explored by comparison with expected survival by risk factor. Results: Compared with 93% for Australian audit cases, 5-year survival was 90% for NZ audit cases overall, 87% for Maori, 84% for Pacific and 91% for other. Conclusions: BC survival in NZ appears lower than in Australia, with inequities by ethnicity. Differences may be due to access, timeliness and quality of health services, patient risk profiles, BQA coverage and death-record methodology.
doi: 10.1111/ans.12851
Introduction The Quality Audit (BQA) undertaken by Breast Surgeons of Australia and New Zealand (BreastSurgANZ) includes data on breast cancer (BC) and its treatment since 1998.1 BQA data are collected to provide feedback to participating surgeons on their patients, the cancers they encounter and clinical practices and outcomes.1,2 The data support quality audit by surgeons of their practices, allow outcomes to be assessed and provide opportunities for research. Operational details have been described in earlier publications.1,2 Surgeons obtain from the BQA database feedback on their own clinical practices, and aggregated data for all surgeons participating in the BQA, allowing self-audit against key quality indicators.2,3 Originally, members of the Breast Section of the Royal Australasian College of Surgeons were encouraged to submit their data to the ANZ J Surg 85 (2015) 546–552
BQA, but since late 2010, it has been a requirement for full membership of BreastSurgANZ that all patients are audited through the BQA. The BQA database includes linked death data, enabling BC survival to be analysed by patient factors, access, timeliness and quality of treatment, case load and other descriptors.1 Clinical outcomes can be compared across the patient population and internationally. These practices accord with objective risk stratification practices and outcome measures recommended by safety and quality organizations for clinical quality assessment.4 In this report, data from the New Zealand (NZ) component of the BQA are used to describe patient and BC characteristics and survival. Death data were obtained by linking BQA data to NZ death records. This is a process routinely undertaken between cancer registry and death data around the world to produce survival data, including with NZ Cancer Registry data.5–9 © 2014 Royal Australasian College of Surgeons
Breast cancer survival
BQA data were not population based during the roll-out of the BQA. The database covered around 50% of early BC in NZ across the 1998–2010 study years, increasing to 80% by 2008. The BQA database is the largest nationwide repository of NZ BC data outside the NZ Cancer Registry. BQA data provide a unique opportunity to analyse clinical features and how survival differs by patient factors, such as ethnicity, socio-economic status of residential area and remoteness, and by conventional cancer prognostic factors such as size, nodal status and grade.10,11 Differences in BC survival are known to exist in NZ,9 but BQA data on BC cancer survival by ethnicity have not been previously published. Results are discussed in the context of populationbased survival data from the NZ Cancer Registry,9 Australian BQA data1 and expected survival patterns by established clinical risk factors.10,11
Methods The study included 10 824 early invasive BC diagnosed in NZ women between 1 January 1998 and 31 December 2010 who were treated by breast surgeons participating in the BQA.1 Cases excluded 267 where data collected were incomplete and insufficient for analysis. The distribution of cases by three-yearly diagnostic epoch was 11.1% prior to 2001; 19.5% in 2001–2003; 27.0% in 2004–2006; and 35.9% in 2007–2009; plus a further 6.5% already entered for 2010. Death data were obtained from the NZ National Mortality Collection by linking with the BQA database using the National Health Index (NHI).12 The NHI comprises a unique alphanumeric identifying number for each New Zealander.12 Death data were traced to 31 December 2010. Results were compared with data from the Australian BQA and survival recorded by the NZ Cancer Registry. The BQA is directed at early BC defined similarly to local and regional spread of disease, as recorded by the US Surveillance Epidemiology and End Results (SEER) Program and the NSW and NZ Cancer Registries.1,7,9,13 However, regional cancers with fixed nodes, which are regarded as regional by SEER and the NSW and NZ Registries, are not included in the BQA definition of early BC.13 BQA data in this study included (i) person descriptors – age at diagnosis (years), ethnicity (Maori, Pacific, other), geographic remoteness of residence from a regional cancer centre classified as 0–9 km, 10–100 km and >100 km, and deprivation index of residential area (ordinal scale from 1 (least deprived) to 4 (most deprived));14 (ii) tumour descriptors – histology type (ductal, lobular, other), tumour size (mm), grade (low, intermediate, high), nodal status, oestrogen and progesterone receptor status, HER-2 receptor status, phenotype (basal (triple negative), luminal A, luminal B, ER-/PR-/HER-2 + , other), vascular invasion, tumour location (ICDO-3 classification), whether unilateral or synchronous bilateral tumour, and number of cancer foci; and (iii) other descriptors – diagnostic year, referral type (symptomatic, BreastScreen, other) and annual surgeon case load (≤10, 11–30, 31–100, 101+). Traditional Kaplan–Meier disease-specific survival from BC was calculated, regarding as a BC death those deaths where BC was recorded as the direct or antecedent cause on the death certifi© 2014 Royal Australasian College of Surgeons
547
cate.4,15,16 Disease-specific survival using this definition is a valid epidemiological option for population-based studies when there is an adequate standard of death certificate recording. This option has been shown to give similar results to relative survival in populationbased studies of BC.17–19 Disease-specific survival may be preferred where patients in clinical studies may not be entirely representative of the population and may have risks of causes of death that do not equate with population norms. Traditional disease-specific survival also was used in an earlier study of the Australian component of the BQA database and its use for the NZ study was designed to produce comparative data.1 Survival times were calculated from date of diagnosis to 31 December 2010 or time of death, whichever occurred first. Hazard ratios for BC death (95% confidence limits) were calculated by patient and cancer descriptor for the follow-up period, using Cox proportional hazards regression, and by applying the same censoring criteria as for the Kaplan–Meier analyses.15,16 Survival analyses were repeated using only cases with complete data for all variables (i.e. ‘complete case analyses’) and by using competing risk regression.20 Results were very similar to those presented from the traditional Kaplan–Meier and proportional hazards regression methods and are not presented in this report. In particular, hazard ratios for variables with higher numbers of missing values (due to not being core items throughout the study period) were similar in the complete case analysis and full analysis. Imputation of missing values was not undertaken due to similarities of the results from the full and complete case analyses.
Results Survival outcomes (unadjusted predictors) Overall BC survival was 90% at 5 years from diagnosis and 84% at 10 years. Differences by patient and cancer characteristic, referral type and surgeon case load are shown in Table 1. Hazard ratios (95% confidence limits) indicated the following differences: • By demographic characteristic: ○ Age at diagnosis: lower hazard ratios for 40–69-year-olds than the reference age (
