The Cancer Council NSW Research Division

The Cancer Council NSW Research Division Cancer Epidemiology Research Unit, (CERU) W’loo   30 people Causes, Health Service research, Descriptive / Analytical Epi Centre for Health Research and Psychooncology, (CHeRP) Newcastle   15 people Psych needs, behavioural science, prevention The NSW Cancer, Lifestyle and the Evaluation of Risk (CLEAR) initiative Freddy Sitas Karen Canfell Research Division, The Cancer Council NSW, 9334 1860 freddys@nswcc.org.au, karenc@nswcc.org.au The team Collaborators:      Clinical Collaborators: StV, RWH, Liverpool, Mater (Newc) … Study Team:    Dianne O’Connell Karen Canfell Emily Banks Robyn Ward Michael Barton  Karen Canfell (Sydney Rotary Research Fellow & study manager) Alicia Delgado (Project Coordinator) Leslie McCawley (Research Assistant) Claire East, Victoria Young, Renate Thielbeer (Patient recruitment) Advisors: Valerie Beral, Bruce Armstrong CLEAR A cancer biobanking project Possible applications    Genetic and environmental risk factors for cancer in NSW. Risk profiles for targeted screening / diagnosis Genetic / biomarker and lifestyle factors that improve survival Excellent postgraduate training in biomarker epidemiology A long term resource (not another 3-y project) Builds capacity in clinical epidemiology and laboratory science Program: Idea: 2003 Scientific review / Funding 2004 Questionnaire 2004+ (Focus groups) Ethics 2005 Karen Canfell, D.Phil (Sydney Rotary Res Fellow) ‘Launch’ 17 May 2006 First Collaborators meeting  June – Dec 06 TCCN Help Line and Call Back Squad Biobanking facility Path lab network Hospital ‘Recruitment’  StV, RWH, Liv, …. Recruit 5000 cases and 5000 controls. Expert Biomarker Advisory Cttee (Ethics) Risks in male migrants Compared to Aus. Born. Supramaniam et al 2006 Colorectal Liver Head and Neck Stomach H1 Large fence H3 H2 Cases Controls In reality: 26 Cancer treatment centres Complex referral patterns Case control study Spouse control design Cancer Lifestyle Evaluation And Risk (CLEAR) Study Male Cases Female Spouses Female Cases Male Spouses Spouse / partner controls In THIS CONTEXT (all cancer types)      Valid Efficient Cheaper High response rate THIS DESIGN minimises overmatching on ‘lifestyle’ Used by Jiang et al (20,000 oesophageal cancers and 105,000 spouses from 1m deaths in China) Variant (Familial controls, Hopper, Kiemeney)  Familial aggregation, genetic.. The CLEAR study Patients <18 mo and their partners fill questionnaire Provide a blood sample (EDTA, SST -80°C) & Consent to access path specimen & Consent to link to CCR, Death & ISC What does all this involve? Access to clinics and eligible patients  (18 years +, diagnosed in the last 18 months) Very little interference with normal work Team of 5 interviewers to approach patients and partners Subjects self-complete consent forms and questionnaire and go to lab for bloods. Cumulative number of cancer patients (18-74 years). CCR NSW 0.5% 1% Power to detect RR=2, α=0.05, 1-β=0.9, 3:1 for exposure prevalence of: 10% Initial questions / hypotheses Longitudinal measurement of performance status (Barton) Risk profiles for targeted screening / diagnosis of pancreatic cancer (Biankin) Infections and cancer (Hazel Mitchell, …) Effect of known risk factors in a local setting     Migration Oral health…(Westmead) Tattoos, piercings Examine distribution of subtle effects, BMI Screening behaviour Hormones (Banks) Platform for genetics, proteomics / pharmacogenomics … Why do we need large biobanks with lifestyle information ‘Classical’ tumour banks generally…     Do not collect lifestyle information Do not collect clinical information (Staging) Concern about representivity / bias Do not have tissues from ‘controls’ Great concern about machine driven research (fishing expeditions) Great need for laboratory science to be integrated into epidemiology Great need to develop screening tools for early diagnosis / prognosis Even greater need for structured, hypothesis-driven research program Public / ethical concern ‘about all these tests’ Benefits Ready made LARGE data / biobank for PhD in:     Genetics Biomarkers Risk factors Infection Opportunities MPH / PHD Epi. To work on systematic review to support cervical modelling project (12 month contract) Pearl Bethel Allan PhD Fellowship into cancer causes and markers (3y). Thank you! Are we up to scratch? Comparisons in the management of lung cancer in New South Wales between 1996 and 2002. Leonardo Simonella, Dianne O’Connell, Shalini Vinod I, Danielle Miller II, Niloofar Esmaili III, Michael Hensley III, David Goldsbury, Rajah Supramaniam, Leslie McCawley, Bruce Armstrong IV The Cancer Council NSW I Cancer Therapy Centre, Liverpool Health Service II NHMRC Clinical Trials Centre, University of Sydney III Hunter Medical Research Institute, Newcastle IV Sydney Cancer Centre, RPA Acknowledgments • Advisory group: - Shalini Vinod - Michael Boyer - Michael Hensley - Geoff Delaney - Bruce McCaughan - Bruce Armstrong - Mathew Peters • NSW CCR for identification of new cases in 2002 Presentation outline • • • • • Background Aim Methods Results Conclusions Background Incidence and mortality Lung Cancer profile in NSW 2004 Male Number of new cases 1953 Female 1070 Total 3023 • Incidence: - ranked 5th - contributes 9% total NSW cancer cases • Mortality: - ranked 1st - contributes 19% of all cancer deaths in NSW Source: Tracey et al. Cancer in NSW: Incidence and Mortality 2004. Cancer Institute NSW, 2006 Age standardised incidence and mortality Age specific incidence and mortality Survival Year 1: male - 35% female - 39% Year 3: male - 15% female - 18% Year 5: male - 11% female - 14% Source: Goumas et al. Lung Cancer in NSW in 1973-1998. TCCN Past patterns of care studies • 1993 Victorian patterns of care study (n=868) – Richardson et al. (2000) – 25% received no treatment – No consistency in treatments provided • 1996 patterns of care study (n=738) – Vinod et al. (2004) – 30% received no treatment. • Nihilistic attitudes among doctors – Jenny et al. (2004), Perez et al. (1998) – Treatment preferences vary between medical specialties – Potentially curable Stage III (NSCLC) – ‘no treatment’ indicated among 8% thoracic surgeons vs. 29% among respiratory physicians – ‘No treatment’ inversely related to number of lung cancer patients seen by lung cancer specialist Aim • Compare lung cancer management for three NSW health areas in 1996 and 2002 Methods Case identification: • Diagnosis in the Hunter, South Western Sydney and Northern Sydney health areas between – Jan 1996 to Dec 1996 – Nov 2001 to Dec 2002 • Recruited from either: – New notifications to the NSW Central Cancer Registry – Clinicians in major treating hospitals – Clinicians in private practice Treatment information and data collection: • Questionnaires addressed: – – – – Initial presentation, diagnosis and staging Surgery Chemotherapy Radiotherapy • Diagnosis and treatment data were collected directly from doctors – Field collection was available to doctors if required Analysis: • Explored variables of interest within 6 months of diagnosis for the three health areas in 1996 and 2002. – – – – Clinical characteristics Doctors consulted Investigations Treatment • Chi-squared test to determine variations in care between 1996 and 2002. • Logistic regression using aggregated data to examine differences in odds of being treated in 1996 and 2002 controlling for potential confounders. Results Health area by year Cases diagnosed 1996 n (%) 2002 n (%) Health Area Hunter Region South Western Sydney Northern Sydney Total cases 212 (29) 256 (35) 198 (35) 195 (34) 270 (37) 738 174 (31) 567 Demographic and clinical characteristics: 1996 Median age Gender (male) (%) ECOG status (0 or 1) (%) 2002 71 63 52 P - value 70 64 53 0.75 0.83 Tumour characteristics: 1996 (%) 2002 (%) P - value 0.52 Small cell Non small cell No pathological confirmation 14 73 13 13 75 11 Tumour characteristics Stage (cont): 1996 (%) 48 52 Limited Extensive 2002 (%) 22 78 P - value Small cell * 0.0004 Non small cell ** I II 20 6 21 9 <0.0001 III IV unknown 39 36 0 23 41 6 * Hunter region had significantly higher % with extensive disease in 2002 compared with 1996 ** Consistently higher % with stage III cases in all health areas in 1996 compared with 2002 Consistently higher % with stage IV cases in all health areas in 2002 compared with 1996 Doctors consulted: 1996 (%) 8 65 No lung cancer specialist seen Respiratory physician 2002 (%) 9 58 P - value 0.55 0.007 Cardiothoracic surgeon Radiation oncologist Medical oncologist 31 57 37 26 45 40 0.06 <0.0001 0.41 Investigations: 1996 (%) Chest x-ray Chest CT Bone Scan 100 92 38 2002 (%) 89 89 35 P - value <0.0001 0.089 0.21 Brain CT scan Bronchoscopy Mediastinoscopy PET Scan 30 48 5 3 29 40 4 18 0.89 0.001 0.42 <0.0001 Treatment (controlling for AHS): Treatment* n % treated 65 ORMH (95% CI) 1.0 (ref.) 1996** 707 2002 567 64 0.98 (0.78 – 1.25) * Treatment within 6 months of diagnosis. ** Data on treatment missing for 31 cases. Logistic Regression: • Odds of having any anti-tumour treatment within 6 months of diagnosis, according to year of diagnosis (2002 vs. 1996) • Models based on aggregated data • Controlled for:  Area health service  Age  Gender  Pathology  Performance status (ECOG)  Stage (applied to regression models restricted to pathology type) Odds of receiving treatment OR* All cases NSCLC only SCLC only 1996 2002 1.00 1.01 95% CI Ref. 0.77 – 1.34 P - value 0.92 1996 2002 1996 2002 1.00 1.27 1.00 0.13 Ref. 0.91 – 1.78 Ref. 0.04 – 0.40 0.15 <0.0001 * Adjusted for: Area health service, Age, Gender, Pathology, Performance status (ECOG), Stage (applied to regression models restricted to pathology type). Discussion • Small variation in patterns of care between 1996 and 2002 • Stage of disease at diagnosis varied between 1996 and 2002  Ascertainment bias or later presentation of disease? • No significant difference in odds of receiving treatment in 2002 and 1996 for all lung cancer cases  Exception for SCLC Conclusions Are we up to scratch? • Patterns of care for lung cancer appear to remain consistent between 1996 and 2002 for most patients. • 1/3 of patients did not receive any anti-tumour therapy. • Patient decision or nihilism amongst doctors? Questions Cancer in Aboriginal People in New South Wales Rajah Supramaniam Hari Grindley Lisa Jackson Pulver Dianne O’Connell Thanks to: NSW Cancer Registry, David Smith, Freddy Sitas Supramaniam R, Grindley H, Jackson Pulver ANZJPH 2006 30(5)453-456 Overview • • • • • Background information Methods and data sources Results Future study directions APOCC Study Introduction • NSW total population 6.3M • 135,000 Aboriginal people or 28% of total Australian Aboriginal population • 2.1% of NSW population Source: Australian Bureau of Statistics Background Reasons for doing this study • Cancer is the second leading cause of death in Aboriginal people • Double burden of disease – Aboriginal people experiencing burdens from infectious diseases, infant mortality plus – chronic diseases such as cardiovascular diseases, diabetes and cancer Zhao and Dempsey Med J Aust. 2006 May 15;184(10):490-4. Reasons for doing this study (cont) • Most Australian cancer data for Aboriginal people is from the less populated jurisdictions – Issues of identification • Public Health 101: First step in addressing a problem is to identify and quantify the problem Sources of data • NSW Central Cancer Registry (CCR) • Population based and running since 1972 • Populations from Australian Bureau of statistics (ABS) Methods • Minimum of 1 case or death per year either expected or observed • Calculated Standardised Mortality Ratios (SIR) using age-specific rates for NSW nonAboriginal population • Poisson confidence intervals (Ulm) due to small numbers Identification issues We used ABS death data from 1994 to 2002 to update the Aboriginal status field in CCR Aboriginal Males Mortality in NSW 1994 to 2002 *Compared to non-Indigenous population (=100) Aboriginal Males Mortality in NSW 1994 to 2002 (Adjusted) *Compared to non-Indigenous population (=100) Aboriginal Females Mortality in NSW 1994 to 2002 *Compared to non-Indigenous population (=100) Aboriginal Females Mortality in NSW 1994 to 2002 (Adjusted) *Compared to non-Indigenous population (=100) Limitations of Incidence data Aboriginal Males Incidence in NSW 1994 to 2002 *Compared to non-Indigenous population (=100) Aboriginal Males Incidence in NSW 1994 to 2002 (Adjusted) *Compared to non-Indigenous population (=100) Aboriginal Females Incidence in NSW 1994 to 2002 *Compared to non-Indigenous population (=100) Aboriginal Females Incidence in NSW 1994 to 2002 (Adjusted) *Compared to non-Indigenous population (=100) Further research directions • How to improve Aboriginal identification in data collections using already collected data • Methods for prospectively increasing the levels of Aboriginal identification • The determinants of diagnosis and survival of Aboriginal people with cancer • Explore the patterns of care of Aboriginal people with cancer Aboriginal Patterns of Cancer Care (APOCC) • NHMRC Health Services Research Grant – 5 years • Investigators from TCCN, University of NSW, University of Sydney and The Sax Institute (CRIAH) APOCC Research Questions 1. Are Aboriginal people being diagnosed with cancer at later stages than non-Aboriginal people? 2. What are the barriers to Aboriginal people being diagnosed at an earlier stage of cancer? 3. Are Aboriginal people with cancer receiving less or a lower quality of cancer treatment than are non-Aboriginal people? Research questions (cont’d) 4. What are the barriers to Aboriginal people receiving cancer treatment that, if not optimal, is at least as good as that received by nonAboriginal people? What evidence-based steps can be taken to increase early stage diagnosis and improve the quality of cancer treatment in Aboriginal people? 5. Planned Approach (Years 0-2) Phase 0: • Focus groups to scope issues relating to cancer for Aboriginal people and develop interview research instrument Phase 1: • Descriptive analysis and comparison of patterns of hospital inpatient care for Aboriginal and nonAboriginal people with cancer in NSW Phase 2: • Pilot treatment survey of medical records in Aboriginal Medical Services; interviews with Aboriginal cancer survivors and Aboriginal Health Workers Planned approach (Years 3-5) Phase 3: • Population-based patterns of care and pathways to diagnosis studies in NSW Phase 4: • Dissemination of key findings and consultation with key stakeholders • Devise appropriate interventions and implementation plans to improve cancer care in NSW for Aboriginal people Thank You rajahs@nswcc.org.au