3.1
This chapter examines investigations by the Victorian Department of Health (department) and Chief Health Officer (CHO) into the possible cancer cluster on the Bellarine Peninsula.
3.2
The chapter first considers recommended approaches to investigating suspected cancer clusters and the difficulties inherent in such investigations. It then details the approaches taken by the department and the CHO in responding to the community’s concerns about the incidence of cancer and autoimmune diseases on the Bellarine Peninsula. It also discusses the responses of the authorities in relation to the concerns raised about the historical use of pesticides. Finally, it outlines some of the potential gaps in the responses and investigations to date, which are further explored in chapters 4 and 5.
Cancer cluster assessments
Process for investigating suspected cancer clusters
3.3
In Australia, reports of suspected cancer clusters are assessed by the relevant state/territory agency (usually the health department) according to its own procedures. The National Health and Medical Research Council recommends that, in determining whether a cluster is present, the agency’s first investigative step should include both epidemiology and hazard assessment and that:
… the detection of significant epidemiological and exposure concerns would indicate the need for further assessment.
3.4
In other words, an agency should assess both the incidence of cancer in a given population (epidemiological concerns) and whether there is a plausible cause of cancer in the given population (exposure concerns). The findings of this first step will determine whether any further investigations are necessary. The National Health and Medical Research Council Statement on Cancer Clusters states that: ‘most reports of a suspected cancer cluster can be resolved when they are first reported’.
3.5
The first step should collect detailed information about the individual cases that make up the suspected cancer cluster, both in relation to the diagnosis and type of cancer for each individual and in relation to each individual’s possible exposure to specific environmental or occupational hazards. This has been called ‘establishing the facts’ and partly relies on information provided directly to the agency by members of the concerned community. This step gives the investigating agency the key information about the geographic area of concern; the number of individuals, their diagnoses and other health data; and any possible chemical exposures and the time frame for these. These facts assist the health agency to assess the two elements of epidemiology and exposure.
3.6
In his submission, Professor Bernard Stewart, a leading expert in cancer clusters, carcinogens and cancer data, confirmed that an initial investigation of a suspected cancer cluster should consider the two elements identified above, increased incidence and plausible causation, and, importantly, that equal attention should be paid to both elements. At the hearing on 20 November 2020, Professor Stewart emphasised the importance of the second element:
Even if the incidence of cancer is greater than expected, that's not the problem. It doesn't establish the causal relationship …
The crucial question is not, 'What is the cancer incidence in the affected population?' but, 'Is there a plausible cause of cancer that has affected this population?'
3.7
Professor Stewart also confirmed the importance of determining the details of the individual cancer cases in question, for example, the types of cancer that are of concern to the community. He explained that true cancer clusters caused by a carcinogen will consist of either a single type of cancer or a group of cancers that are known to be related. He concluded:
… no authority in the world recognises as requiring thorough investigation cancer clusters that involve multiple tumour types, because every specific known carcinogen causes a particular group of cancers.
3.8
If epidemiological or exposure concerns are not resolved at this first step, the National Health and Medical Research Council recommends a staged approach, with each later step being taken only if required by the findings of the previous step. These further steps are cluster assessment, further assessment (if a cluster is present), and monitoring. The NHRMC also recommends reporting of suspected cancer cluster assessments. This staged approach reflects international best practice on how to approach reports of suspected cancer clusters.
3.9
If required, the further assessment may include more focused epidemiological and statistical analysis to determine whether a true cluster exists; and a more in-depth analysis of the cancers of concern and any occupational or environmental exposures that might represent a plausible causation. In rare cases, an aetiological study may be undertaken to determine whether a true cluster can be explained by any of the possible exposures.
3.10
Good communication with concerned individuals and communities at every stage is a fundamental element of Australian and international best practice for investigating suspected cancer clusters. The UK Guidance for investigating non-infectious disease clusters from potential environmental causes explains the reasons for this:
In addition to having epidemiological and statistical investigations, it is important to understand the social dimensions of a cluster: the community’s perception of risk, potential legal ramifications and the role or influence of the media. Addressing communication activities at each stage of the cluster investigation and developing and maintaining community relationships and trust will help the credibility and understanding of the investigation
3.11
Professor Stewart submitted that communications should be clear about the limitations of suspected cancer cluster investigations with regard to identifying a plausible cause, highlighting the fact that ‘there is complete agreement that cluster investigations do not provide insight regarding cancer causation’. His submission states that:
Communication is fundamental to the perception of cancer clusters and the related failure to understand what (apart from smoking and asbestos) causes cancer. Health authorities can declare what is proven to cause cancer but can’t list what is proven not to.
Difficulties with assessing suspected cancer clusters
3.12
Reports of suspected cancer clusters are notoriously difficult to investigate. The US Centers for Disease Control and Prevention website says:
The complex nature of cancer makes it inherently challenging to identify, interpret, and address cancer clusters.
3.13
Inquiry participants also highlighted the difficulties associated with assessing a suspected cancer cluster. For example, Distinguished Professor of Epidemiology Lin Fritschi told the committee:
Cancer cluster investigations are extraordinarily difficult. They end up being, I think, very unsatisfactory for everybody involved.
3.14
Submitters to the inquiry and expert witnesses identified a number of known difficulties with suspected cancer cluster investigations, including:
the latency period, or time lag, of at least five but up to 45 years between exposure to a possible cause and the appearance of cancer;
identifying the cohort of interest, and the related issue of population mobility, which means it may not be possible to locate the entire cohort;
the question of identifying and quantifying historical exposures for each person in the cohort;
the limitations of statistical analysis of a small population in the area of interest;
the possibility of alternative explanations (for example, genetic or lifestyle risk factors) for each particular cancer or alternative exposures for a particular individual, (for example, any exposures that occurred away from a person’s home address);
the likelihood that the investigation will fail to establish either a true statistical cluster or a cause for the cluster; and
the difficulty of assessing the value of further investigations, given that these are resource-intensive and unlikely to provide definitive answers and may cause further distress to the community.
3.15
Investigations of suspected cancer clusters are very unlikely to completely explain any individual case of cancer. To this effect, Professor Fritschi told the committee:
… there aren't ways to say, 'This cancer in this person was caused by this exposure to this chemical'—we don't know how to do that, and that's unfortunate.
Government responses to community concerns
3.16
The Victorian CHO has a statutory function under the Public Health and Wellbeing Act 2008 (Vic) to provide expert clinical and scientific advice and leadership on issues impacting public health. The department advised that it not uncommonly examines concerns about potential non-communicable disease clusters in the community, particularly if the concern is related to an environmental hazard or contaminant.
3.17
As the municipal authority for the Bellarine Peninsula, the City of Greater Geelong (City) advised that its role has been to respond to community inquiries and support the department in the dissemination of information. The City informed the committee that it was not in a position to investigate or comment on whether or not a cancer cluster exists, as this is not within its expertise or jurisdiction. Responses by the City to community concerns about chemicals used in the Bellarine Peninsula region are considered in chapter 5.
3.18
In January 2019, the Victorian Government, through the CHO and the department, began to investigate a suspected cancer cluster on the Bellarine Peninsula. This action was prompted by media reports of community concerns about the incidence of cancer on the Bellarine Peninsula and a possible link to the historical use of dieldrin, an agricultural chemical. The Victorian Government’s responses to these concerns were as follows:
the initial investigation by the CHO (January 2019);
the Community Open House event in Barwon Heads (February 2019);
convening the Potential Cancer Cluster Expert Advisory Group (March 2019); and
commissioning the epidemiological analysis by Cancer Council Victoria and its review by the Potential Cancer Cluster Expert Advisory Group (October 2019).
3.19
During the inquiry, the committee asked Cancer Council Victoria about the feasibility of conducting further epidemiological analysis of cancer rates going back to the 1980s. The committee asked the CHO to commission this additional analysis from Cancer Council Victoria. This additional analysis, Bellarine Peninsula cancer incidence report: Update, was completed in March 2021 and reviewed by the Potential Cancer Cluster Expert Advisory Group in April 2021. The report update and the expert review were both made public via the department’s website and via the City’s website in April 2021.
3.20
Each of these responses will be considered in turn.
Chief Health Officer’s initial investigation of cancer rates on the Bellarine Peninsula: January 2019
3.21
The committee heard that the department would normally become aware of a suspected cancer cluster through a report from one or more individuals. This is consistent with other jurisdictions.
3.22
However, in the case of the suspected cancer cluster on the Bellarine Peninsula, no direct approaches were made to the department. The CHO explained that the department had instead responded proactively to community concerns reported in the media, telling the committee:
… we did see the concerns that were expressed through media and therefore we tried to respond to them with the information that we had through those media stories, bearing in mind that's really all we had. We were trying to piece together where the concerns were focused in terms of the types of cancers, the period of time that it seemed to be concerned with and the geographical area or the particular setting that concerns were expressed about, and that did change over time.
3.23
Based on that information, in early January 2019 the CHO undertook an initial investigation into cancer incidence rates on the Bellarine Peninsula for each of the cancers identified in media reports, and into the carcinogenicity of dieldrin, which was the chemical reported by the media to be of concern to the community.
Methodology and approach
3.24
The CHO’s investigation sourced epidemiological cancer incidence data from the Australian Cancer Atlas (Atlas). The CHO’s report explained that the Atlas includes comprehensive coverage of all cancers diagnosed from 2010 to 2014. The Atlas reports cancer incidence by geographical area using Statistical Areas Level 2 (SA2s) relative to the Australian population, using standardised incidence ratios (SIRs).
3.25
SIRs are calculated by dividing the number of observed cases in that area by the number of expected cases, based on the average age- and sex-specific incidence rates for the Australian population. In plain language, the SIR indicates whether people living in that area have a greater or lesser risk of being diagnosed with cancer when compared with the Australian average risk.
3.26
The CHO’s report noted that, with no generally agreed geographical boundaries for the area known as the Bellarine Peninsula, the investigation included the broad region to ensure no area was missed. It also noted that SA2s are the appropriate sized area to ‘capture’ sufficient cases of cancer to avoid the random fluctuation in rates that can appear when looking at smaller areas.
3.27
The CHO explained to the committee that using the Atlas data had several other benefits:
… it was publicly available, it could be checked by the community if they had an interest in doing so and we could do an initial investigation that was quick and could provide some kind of answer with all the caveats around what the limitations of that investigation were, which I put upfront in the foreword of that report. I did it because I could provide an early indication and say that we would continue to investigate.
3.28
The CHO used the standard incidence data for all geographic areas in the Bellarine Peninsula to analyse incidence rates in that region for the following categories of cancer:
each of the four specific cancer types mentioned in media reports (multiple myeloma, leukaemia, non-Hodgkin’s lymphoma and brain cancer);
each of the two cancer types possibly linked to dieldrin (breast cancer and liver cancer); and
3.29
The CHO’s investigation also included a toxicological review of scientific evidence to assess whether dieldrin was a plausible cause for the reported cancer types.
Key findings
3.30
Published in January 2019, the CHO’s investigation reported there was:
no evidence of a higher rate of cancer overall in any geographical areas of the Bellarine Peninsula than elsewhere in Australia;
no higher number of the specific cancers of interest (breast, liver, non-Hodgkin lymphoma, multiple myeloma, brain cancers and leukaemia) than would be expected (based on the average cancer rates in Australia).
3.31
The CHO’s report on this investigation also acknowledged that ‘[c]ancer in young people is unusual and particularly distressing’.
3.32
The CHO’s report found no evidence linking dieldrin to any of the cancers mentioned in media reports:
… the hazard of concern (dieldrin) has not been identified as an agent that results in the cancers cited in the media.
Limitations
3.33
The CHO’s report identified two limitations of this investigation. First, to ensure statistical validity for the small population for each geographic area, the Atlas combines data from several years, which may mask some differences across areas or across years. Second, this data is based on the address at the time of diagnosis for each individual and so may not accurately reflect the cohort of interest (for example, if a person uses a non-residential address or has only recently moved to the area).
3.34
The CHO’s report acknowledged that the use of SA2 data limited the study’s findings to some extent. SA2 data combines available data from smaller geographic areas known as Statistical Areas Level 1 (SA1s). However, the report also noted that, even at SA2 level, the numbers of people diagnosed with cancer each year are often very small and need to be grouped together with data from other years in order to conduct reliable calculations.
3.35
The CHO also told the committee that his initial investigation, using the Atlas data, could not be precisely targeted to the specific concerns of the community at that time. The CHO explained that this was because no detailed demographic, health or exposure information about the individual cases of concern had been provided to the department. In an answer to a question on notice, the CHO advised the committee:
I, and the Department, sought to obtain information from other sources to further the investigation of cancer rates on the Bellarine Peninsula, with limited success. There was a reluctance from some individuals and other parties to provide the Department with information. Both Gordon Legal and Mr Ross Harrison were approached, but no information was forthcoming.
3.36
In an answer to a question on notice, the CHO advised that this investigation had been based on the information available to the department at the time, stating that ‘the information from media reports was the best information available’.
Issues raised by submitters in relation to the epidemiological studies
3.37
Most issues raised by submitters about the CHO’s investigation related to its use of Atlas data. There were concerns that this data set:
did not capture the relevant geographic or demographic cohort;
did not capture all relevant populations, such as holidaymakers; and
did not capture the relevant time period, as it only included cancer diagnoses from 2010 to 2014.
Failure to capture relevant geographic or demographic cohort
3.38
Community group Discovery 3227 submitted that the CHO’s investigation had used a ‘flawed methodology’, including the Atlas data’s use of SA2s, and therefore did not accurately determine cancer rates for Barwon Heads specifically.
3.39
Some submitters were concerned that, since the Atlas data reports age-standardised cancer incidence rates and not age-specific rates, the CHO’s investigation did not consider cancer incidence specifically in young people on the Bellarine Peninsula, despite community concern about a number of younger people diagnosed with cancer.
3.40
For example, one submitter stated:
Concerns around a potential cancer cluster on the Bellarine Peninsula are not based around a representative age sample of the wider population. Concerns centre on a specific cohort of age, typically being 24-34 years old at the time of diagnosis. Across Australia, this age cohort experiences a very small incidence of cancer, as per the Australian Cancer Atlas.
Failure to capture population variability and mobility
3.41
The Bellarine Peninsula is a popular holiday destination and its population almost triples on weekends and in the warmer months. Discovery 3227 suggested that the CHO’s use of Atlas data was therefore inadequate, as the Atlas uses location data based on a person’s primary residence and does not capture part-time populations.
3.42
Discovery 3227 also commented that the CHO’s investigation failed to account for population mobility before or after the time period covered by the Atlas data (from 2010 to 2014). Discovery 3227 told the committee:
The Cancer Atlas fails to capture those residents exposed to mosquito OP chemicals that develop cancer at a later date, as their new postcode will be the registry area for the illness.
3.43
On the other hand, another submitter suggested that part-time residents were less likely to have developed cancers connected with any chemical exposures in the Barwon Heads area, stating:
The argument by the accusers that disease data for people who are in the region for holidays periods should be included only weakens their argument, since the doses associated with such transient exposures will be much less than for those living in the area.
Failure to capture relevant time period
3.44
The Atlas uses data from the Australian Cancer Database to report either a five-year incidence rate (for each of the most common cancers and for ‘all cancers’) or a ten-year incidence rate (for each of the less common cancers). Some submitters considered that the five-year time span for the cancers investigated by the CHO was not sufficient for investigating the specific cases of concern to the community. At the hearing on 1 May 2020, Discovery 3227 stated:
This time span fails the core chemical exposure period of 1980 to 2000 in Barwon Heads. Even with a latency of five, 10 or 15 years for cancer to develop, that is 30 years of lost immune and cancer data being ignored, which is of the utmost importance to establishing whether a cancer cluster exists.
Issues raised by submitters in relation to chemical exposures
3.45
Some submitters were not satisfied with the findings of the CHO’s investigation in relation to possible chemical exposures. Gordon Legal argued that some in the community had been exposed to unsafe levels of dieldrin and other organochlorine pesticides (OCPs) in the environment between 1997 and 2007.
3.46
Most individual submitters and Discovery 3227 were concerned not about OCPs such as dieldrin but about organophosphorus pesticides (OPs) and synthetic pyrethroids. Discovery 3227 believes that the CHO insufficiently investigated the carcinogenicity of certain chemicals alleged to be used by the City for mosquito control, such as organophosphate insecticides containing temephos, malathion or fenthion.
Exposures to organochlorine pesticides (OCPs)
3.47
Until the 1980s, the OCPs dieldrin and DDT (dichlorodiphenyltrichloroethane) were widely used as insecticides, both for agricultural purposes, such as crop pest control and building pest management, and for domestic purposes, such as garden pest control. For example, as mentioned in chapter 2, dieldrin was used by farmers on the Bellarine Peninsula.
3.48
OCPs were phased out of use in Australia during the 1980s because of concerns about toxicity, short-term and long-term health impacts, and environmental persistence; in some cases, OCPs have taken decades to break down. The sale and use of dieldrin was banned in Victoria in 1987. Since 1997 OCPs have been internationally recognised as ‘persistent organic pollutants’.
3.49
In 2016, the World Health Organization's International Agency for Research on Cancer (IARC) classified dieldrin as ‘probably carcinogenic’ to humans. The IARC has explained that this terminology relates only to the strength of the scientific evidence and does not quantify the cancer risk.
3.50
In late 2016 the law firm Gordon Legal began investigations for a possible legal case on behalf of Corinne Beyer in relation to the incidence of cancer in former students and staff at the Drysdale campus of Bellarine Secondary College. In 2019 Gordon Legal issued media releases announcing plans for a test case claiming that soil residues of dieldrin and other OCPs may have caused cancers in three individuals.
3.51
In late 2018 and early 2019, media articles reported that individual community members shared these concerns about a connection between possible historical exposures to OCPs and cases of cancer in the community.
3.52
Gordon Legal disagrees with the CHO’s finding that there was no plausible association between the historical use of dieldrin in the region and local cases of cancer. Gordon Legal submitted that the CHO’s investigation should have calculated historical soil contamination levels by using the known half-life of dieldrin. Gordon Legal also criticised the CHO’s investigation for failing to closely examine possible exposures to 12 other OCPs in addition to dieldrin, or the potential combined effects of such exposures.
3.53
Gordon Legal alleged there was an ‘increased risk to the BSC cohort on account of exposure to a multiplicity of OCPs’. Mr Gordon told the committee that he believed that these exposures to OCPs were likely to have been similar to occupational exposures, through ‘pretty heavy exposure to dust’.
3.54
Mr Gordon explained to the committee that the law considers the question of ‘probable cause’ and that he believed that question was satisfied by the circumstances of the BSC cohort and the three individuals in question. This test case is currently under way in the Victorian Supreme Court.
Responses from the authorities
3.55
As mentioned above, in January 2019 the department reviewed cancer incidence data and the scientific literature, which it reported ‘did not identify dieldrin as a chemical associated with the specific types of cancers cited in the media.’ It further advised that 'some pesticides (like DDT) have previously been associated with non-Hodgkin lymphoma, chronic lymphocytic leukaemia and multiple myeloma' but that 'this association has not been found for dieldrin'.
3.56
The committee heard similar evidence from Professor Stewart, who advised that the IARC evaluation of dieldrin does not associate occupational exposure to dieldrin with non-Hodgkin lymphoma, other types of lymphoma or leukaemia, or any other specific malignancy. Professor Stewart told the committee that scientific scholarship on dieldrin has established that ‘[t]here is limited evidence in humans for the carcinogenicity of dieldrin’.
3.57
Also in January 2019, the City reviewed all previous soil contamination assessments for organochlorine pesticide residues, in particular dieldrin, in soil across the Bellarine Peninsula. Its submission stated that both dieldrin and DDT levels were found to be safe.
Exposures to chemical insecticides used for mosquito control
3.58
Discovery 3227 and some individual submitters believed that the CHO should have more closely investigated the carcinogenicity of specific chemicals that Discovery 3227 alleged had been used by the City for mosquito control. These included OPs such as fenthion, temephos and malathion and synthetic pyrethroids.
3.59
Many individual submitters were worried about whether they had been exposed to OPs through mosquito spraying or fogging, and whether these exposures could have caused their cancers or autoimmune diseases.
3.60
Gordon Legal submitted that the CHO should have considered the possibility of additional ‘background exposure’ to OPs, such as malathion, for individuals living on the Bellarine Peninsula.
3.61
Chapter 4 discusses the pesticides used for mosquito control on the Bellarine Peninsula. Chapter 5 discusses community concerns about exposures to the chemical insecticides used for mosquito control on the Bellarine Peninsula and considers the responses from authorities.
Interagency group: January 2019
3.62
In mid-January 2019, the department established an interagency group to communicate with and respond to the community. The interagency group developed an information sheet with FAQs (frequently asked questions) about the possible cancer cluster, which was published on the City’s website.
3.63
The group also considered soil-testing data provided by a number of agencies, This included results from recent testing for dieldrin levels in the soil at both Bellarine Secondary College and Barwon Heads Primary School and from previous soil testing done in the area, for example, for land redevelopment purposes. The group excluded the possibility of any current soil contamination or risk to public health from organochlorine residues in soil.
Community Open House event: February 2019
3.64
In response to ongoing community concerns, on Monday 25 February 2019 the interagency group held a drop-in Community Open House event in Barwon Heads.
3.65
The event was hosted by the City and was advertised through social media, print media articles and flyers, with an information sheet and promotional material were published on the City’s website. The department liaised with the Barwon Heads Association community group before and after this event, and the association also promoted the event in its newsletter and through its website and social media presence. Forty-five community members attended the event.
3.66
The purpose of the Community Open House event was for the CHO and relevant government agencies to present the findings of the CHO’s recent investigation, to provide information, and to answer questions from community members who were concerned about a suspected cancer cluster possibly caused by environmental exposures to dieldrin.
3.67
Some key community concerns identified by the department at this event were:
‘methods and limitations of the Chief Health Officer’s investigation of cancer rates on the Bellarine Peninsula’;
‘potential spray drift from local golf clubs and maintenance of the Barwon Heads Primary school oval’; and
‘the mosquito spraying program in the area, particularly the chemicals and processes used historically’.
3.68
The 25 government representatives at the Community Open House event provided information about these concerns to community members in person. After the event, the department updated the FAQs published on the City’s website with information responding to these community concerns. The department prepared a formal report of the event which was also published on the City’s website.
3.69
At the end of the event the department solicited feedback from attendees, which is summarised in its report as follows:
The feedback on the Open House event itself indicated that in general, the open house format was useful, friendly, helpful and effective, and concerns and issues were understood but not always able to be answered. It is intended that the Open House and follow-up processes are the beginning of an ongoing engagement that keeps communications open and timely.
3.70
Discovery 3227 submitted that this event was a ‘concerning knee-jerk reaction’ in response to community fears of chemical exposure, and that the department and the City did not address community concerns about historical exposures to chemicals other than dieldrin.
3.71
One submitter mentioned his appreciation for the time and effort taken by the CHO and the department at this event.
3.72
After the event, the Geelong Advertiser reported the CHO, Professor Brett Sutton, as saying:
‘There are some people who I think won’t be reassured, they have further questions. … We want to be an open portal to see if we can explore further.’
Potential Cancer Cluster Expert Advisory Group: March 2019
3.73
On 20 March 2019, the department convened the Potential Cancer Cluster Expert Advisory Group (Expert Advisory Group) to provide the CHO with high quality, impartial advice from an independent panel of scientific experts on cancer cluster investigations.
3.74
The Expert Advisory Group comprised Dr David Hill AO (Cancer Council Victoria, University of Melbourne, University of Newcastle), Dr Tim Driscoll (Professor of Occupational and Environmental Medicine at the Sydney School of Public Health, University of Sydney) and Professor Brian Priestly (Director of the Australian Centre for Human Health Risk Assessment, Monash University).
3.75
The Expert Advisory Group Chair, Professor Hill, explained the group’s role to the committee at the public hearing on 1 May 2020:
… the ongoing role of the [Expert Advisory Group] is to independently review the work done by and for the Chief Health Officer of the [Victorian Department of Health] in assessing evidence about a cluster, comment on its quality, and possibly suggest future work, including whether a full research study would be needed and justified.
3.76
Expert Advisory Group members expressed the view that the CHO’s initial investigation and other responses by the department had been appropriate. For example, Professor Hill stated:
I think the department has followed a fairly logical and standard public health approach to responding to community concerns about a cancer cluster.
3.77
However, the Expert Advisory Group recognised that some community members still held concerns about the possibility of a cancer cluster linked to chemical exposures in the Barwon Heads region. The group recommended to the CHO that an epidemiological analysis of Victorian Cancer Registry data would provide a more targeted response to these concerns. Professor Hill told the committee that:
… the [Expert Advisory Group] advised [the Chief Health Officer] that a more fine-grained analysis on a more limited geographical area would be possible, and this would more directly address community concerns about exposure at the two schools in question.
3.78
In relation to this further analysis, the Expert Advisory Group advised that SA1 (Statistical Areas Level 1) data should be combined, to ensure ‘that any relevant epidemiological trends that might be detectable on the Bellarine Peninsula could be identified’.
Cancer Council Victoria report: October 2019
Methodology and approach
3.79
The Expert Advisory Group explained to the committee that Dr Roger Milne of the Cancer Epidemiology Division of Cancer Council Victoria was therefore asked to undertake a new analysis and to:
… prepare a report on a slightly expanded number of tumours for this more constrained geographical area nearer the schools for the period 2001 to 2016 …
3.80
The Cancer Council Victoria report sourced cancer data from the Victorian Cancer Registry and population data from the Australian Bureau of Statistics (ABS). Analyses were conducted for the period from 2001 to 2016. Dr Milne advised the committee that cancer registry data and the relevant ABS data from pre-2001 are not directly compatible, because the ABS has only been using statistical areas such as SA1s and SA2s since 2001.
3.81
The department and Cancer Council Victoria agreed that analysis of the available data from this time period would address some of the community concerns that had been identified in media reports. These concerns were about a specific cohort of the community, described to the committee as being mostly ‘people in their 20s and 30s who had been diagnosed in the 2000s’.
3.82
The CHO told the committee he had asked Cancer Council Victoria to undertake this analysis by using Victorian Cancer Registry data to examine cancer incidence rates in a smaller geographical area, for a greater number of specific cancers and for younger people specifically.
3.83
The Cancer Council Victoria report analysed the incidence in each of the following categories of cancer: liver cancer, breast cancer, testis cancer, cancer of the brain and central nervous system, Hodgkin lymphoma, non-Hodgkin lymphoma, multiple myeloma, leukaemia, other haematopoietic cancers, and all cancers combined.
3.84
Cancer Council Victoria analysed the incidence of these cancers as diagnosed in residents of a defined area on the Bellarine Peninsula over the period from 2001 to 2016 and compared it with the expected incidence based on the average incidence for Victoria. This area was defined using ABS data at SA1 level, which captures the Barwon Heads area and some land to the west of Barwon Heads. This geographical area does not include Ocean Grove.
3.85
The Cancer Council Victoria report included a specific analysis of cancer rates for young people aged 10 to 34 years during that period of time, to address concerns reported in the media about a number of younger people living in the Barwon Heads region who had been diagnosed with cancer.
Key findings
3.86
The Cancer Council Victoria report found that there was no substantive evidence of increased rates of cancer in Bellarine Peninsula residents between 2001 and 2016. The report was published on 17 October 2019. The report states:
For all cancer groupings considered, the number of observed cases was very similar to the number expected, indicating no excess cancer incidence.
3.87
Professor Milne, who conducted the analysis, told the committee:
… the analysis that we did was well defined and quite straightforward in terms of statistical analyses, and we found no evidence of increased incidence of any of the individual cancers considered, cancers of any type—that is, all cancers combined in all people—and cancers of any type in younger people between the ages of 10 and 34.
Limitations
3.88
The Cancer Council Victoria report notes three limitations to its findings.
the time period covered (2001–2016);
the effects of population mobility; and
the effects of known lifestyle risk factors for cancer.
3.89
As mentioned above, the availability of compatible ABS data limited the report to the period from 2001 to 2016. This meant that any cases of cancer diagnosed before 2001 or after 2016 were not captured by this analysis.
3.90
There is a typical time lag of 10 to 20 years between a possible exposure and a diagnosis of cancer. The ABS data used by Cancer Council Victoria are based on resident populations and therefore do not reflect any population movement into or out of the Barwon Heads region between 2001 and 2016. This means that the Cancer Council Victoria analysis could not capture every potential case of concern, since it could not trace individual changes of address over that period. For example, the results do not include people who moved out of the area since a possible exposure but before contracting cancer.
3.91
Finally, Cancer Council Victoria was unable to make any individual adjustments for known cancer risk factors, such as smoking, alcohol consumption or obesity. As mentioned above, detailed information about each individual case of concern was not made available to the department. The data available was mostly limited to the information in media reports.
3.92
At the hearing on 1 May 2020, Professor Milne explained to the committee that the statistical analysis provided by Cancer Council Victoria in its report is only one part of the equation and that other kinds of expertise would be needed to investigate other considerations, such as environmental assessment and testing for possible carcinogens.
Expert Advisory Group view
3.93
The Expert Advisory Group independently reviewed the Cancer Council Victoria report and provided a written opinion to the CHO on 27 October 2019:
The [Expert Advisory Group] has concluded that Professor Milne’s analysis provides no material evidence of excess cancer rates in the area examined between 2001 and 2016 for the specific cancer types that were examined … There was also no material evidence of excess rates suggested by the additional analyses of all those cancers combined, for all ages combined, and for all those cancers combined in 10-34 year-olds.
3.94
The Expert Advisory Group opinion to the CHO remarked on the accuracy and scientific integrity of the Victorian Cancer Registry data. The group advised the CHO that, despite the possible effects of population mobility, the use of cancer registry data is nonetheless considered valuable for detecting cancer incidence rates that are connected to environmental carcinogens.
3.95
At the hearing on 1 May 2020, Professor Hill commented on the methods used by Cancer Council Victoria in its analysis:
… the data used in this report are of a very high quality, the methods of analysis are appropriate and appear well executed, and the conclusions are valid. Based on this, the [Expert Advisory Group] does not see any evidence of a cancer cluster on the Bellarine Peninsula.
3.96
In response to a question on notice, the Expert Advisory Group again confirmed this assessment, stating:
The study adequately addressed the question of whether there was any evidence of ‘clusters’ i.e., unusual variation in the incidence of breast , liver cancer, brain & CNS cancer, non-Hodgkin lymphoma, Hodgkin lymphoma, multiple myeloma, leukaemia and other haematpoietic [sic] malignancies in people at all ages combined and separately in 10-34 year-olds living in the area studied.
3.97
The Cancer Council Victoria report and the Expert Advisory Group’s opinion on the report were both published on the department’s website in October 2019. These two documents were also made publicly available on the City’s website.
Issues raised by submitters
Inadequate methodology
3.98
Community group Discovery 3227 submitted that the new analysis by Cancer Council Victoria was flawed. Discovery 3227 submitted that, since the methodology used by Cancer Council Victoria was limited in its ability to capture population variability and mobility, it was not appropriate for the investigation of a cancer cluster:
… the Cancer registry is incapable of identifying a cancer & immune cluster or spikes due to the fact as previously stated in the Senate Inquiry, assumed static population, inward and outward migration, floating permanent holiday make[r]s with other postcodes …
3.99
Gordon Legal submitted that the use of SIRs for cancer did not adequately address the possibility of a true cancer cluster among former students of Bellarine Secondary College specifically. Mr Gordon told the committee:
I think that the overall incidence of cancer on the Bellarine Peninsula is a very poor guide to whether anything untoward was going [on] at Bellarine Secondary College from 1997 in the first years of its operation.
Time period insufficient to capture cases of concern
3.100
Discovery 3227 also stated its concern that the Cancer Council Victoria report had failed to consider cancer incidence rates from before 2001. Mr Harrison told the committee that Discovery 3227 had collected health data from many community members and that ‘the bulk of the data starts from 1980 onwards into the mid-2000s’.
Cancer Council Victoria report update: March 2021
Further epidemiological analysis
3.101
At the hearing on 1 May 2020, the committee asked Professor Roger Milne, Head of Cancer Epidemiology at Cancer Council Victoria, and the CHO about the feasibility of conducting further epidemiological analysis. The committee asked the CHO about cancer incidence rates for 1985–2001 for the Barwon Heads region, noting that the Cancer Council Victoria report of October 2019 analysed cancer incidence rates for 2001–2016.
3.102
At the hearing on 1 May 2020, the CHO advised that he would seek Professor Milne’s expert advice on the kinds of statistical data available for further epidemiological investigation from 1985 onwards and on the feasibility of such an investigation. The committee received the following advice from Professor Milne of Cancer Council Victoria on 29 May 2020:
I have contacted the Australian Bureau of Statistics (ABS) and confirmed that the population data … could be generated by the ABS on request. The ABS would need to (i) confirm that the population in the geographical area to be assessed can be consistently determined by combining collection districts, as defined at each census (i.e. in 1981, 1986, 1991 and 1996, noting that these collection districts can change between census years), and, assuming it can, (ii) combine data to generate the estimated residential population for each calendar year, by sex and age (in five-year categories).
… it appears (but will need to be confirmed by the ABS once they have collated all the data) it would be methodologically feasible to carry out the analysis using data prior to 2001, and as far back as 1982 …
3.103
In response to a question on notice, the CHO stated:
It is now clear from the issues raised at this Inquiry that there are concerns about the agents used in the mosquito spraying program in the Barwon Heads area between 1980 and the late 2000s. Noting this, it may be of value to some members of the community for the analysis to extend back to 1982. It is, nonetheless, highly unusual to explore evidence for changes in cancer rates simply on the basis of concerns about environmental contaminants rather than specific concerns about cancer rates relevant to the period of alleged exposure.
3.104
In September 2020, the CHO agreed to the committee’s request to ask Cancer Council Victoria to conduct additional analysis. After noting the limitations that would apply to these further investigations, in particular relating to population mobility, the CHO advised:
In response to questions asked at the hearing of this Inquiry on 1 May 2020 by Senators Bilyk and Henderson, I have provided my support in commissioning further epidemiological analysis of Victorian Cancer Registry data by Professor Milne, extending back to 1982, and in re-engaging the Expert Advisory Group to provide advice as to the methodology involved. The further report by Professor Milne will be made public.
3.105
The additional analysis was completed in March 2021 (Cancer Council Victoria report update). The CHO then re-engaged the Expert Advisory Group to peer-review the report update. The Expert Advisory Group provided its review to the CHO in April 2021 (Expert Advisory Group review). On 15 April 2021 the report update and the peer review report were provided to the committee and were also made publicly available online.
Cancer Council Victoria report update (15 March 2021)
Methodology and approach
3.106
The Cancer Council Victoria report update significantly extends the analysis done in October 2019. It uses Victorian Cancer Registry data from 1982 (when records began) through to 2019 (the most recent data available) and calculates incidence rates for one additional cancer type (prostate cancer). The report update provides estimated SIRs for the periods 1982–2000 and 2002–2019 and for the total period 1982–2019 for the Barwon Heads region.
Key findings
3.107
The report update showed no increased incidence of cancer in the Barwon Heads region for ‘all cancers’, for ‘all cancers (10–34 years)’, or for any of the individual cancers identified except for breast cancer.
3.108
The estimated SIR for breast cancer was calculated to be 1.24, with a 95% confidence interval of 1.01–1.50. This means there was an estimated 24% excess incidence of breast cancer in the region for 1982–2019 compared with the Victorian average. The report update explained that this estimated SIR represents ‘an excess incidence [of breast cancer] beyond what might be explained by chance’.
3.109
At the hearing in Barwon Heads, the committee questioned the CHO about the estimated SIR for leukaemia of 1.42 and whether this represented an above-average incidence of leukaemia in Barwon Heads. Professor Tim Driscoll, a member of the Expert Advisory Group, explained that this estimated SIR does not mean there is an increased incidence of leukaemia. He advised the committee that the associated confidence interval of 0.85 to 1.89 indicated that the estimated SIR could be due to chance:
The standard approach with this sort of thing is that, where the confidence interval is fairly wide and crosses 1, like it does for leukaemia, all we conclude is that we're not sure.
3.110
In response to questions asked on notice at the hearing, the CHO confirmed that ‘the analysis by Cancer Council Victoria did not find that the incidence of leukaemia was statistically significant’. Also in response to the questions on notice, the Expert Advisory Group advised that there was no data suggesting that carcinogens were present at levels that may have caused leukaemia, and that the community should be reassured that the Barwon Heads area does not have an above-average incidence of leukaemia.
Limitations
3.111
The report update identifies that this analysis (using cancer registry data and ABS data) is limited by being unable to control for individual lifestyle-related cancer risk factors; reproductive or hormonal factors; or population mobility. These factors could at least in part explain the estimated SIR for breast cancer.
Expert Advisory Group review
3.112
The Expert Advisory Group conducted a scientific peer review of the report update on 9 April 2021. The Expert Advisory Group review found that Cancer Council Victoria adopted appropriate and scientifically conventional methodology for data selection and for calculating the estimated SIRs and confidence intervals for each SIR, and that the analysis in the report update was ‘of a high standard’.
3.113
The Expert Advisory Group noted that the estimated SIRs for breast cancer for 1982–2000 and for 2001–2019 are very similar, and that this suggests that any causative factors remained stable before and during 1982–2019. The Expert Advisory Group found that these results do not suggest that current or former environmental exposures caused the ‘moderately greater’ observed incidence of breast cancer in Barwon Heads. The Expert Advisory Group also noted Census data indicating that the Barwon Heads region is a very high socioeconomic status area, and that high socioeconomic status is a recognised risk factor for breast cancer.
Gaps in investigations to date
Autoimmune disease
3.114
As discussed in chapter 2, some submitters raised concerns about the incidence of autoimmune disease on the Bellarine Peninsula, and there were some media reports that covered these concerns. Some submitters were concerned about rates of inflammatory bowel disease (IBD) in particular. Discovery 3227 also expressed concerns about possible links between autoimmune disease and cancer.
3.115
In response to these community concerns, the department advised that it had received no clinical reports of increased rates of autoimmune disease on the Bellarine Peninsula. It also reviewed the scientific literature and found that IBD rates in the Barwon area are either lower than or comparable to rates in other developed Western countries.
3.116
The department directly consulted specialist clinicians at the local public hospital, University Hospital Geelong, who did not report concern about increases in the incidence of any autoimmune illness. The department also prepared a community information sheet responding to specific community concerns about IBD, which is available on the City’s website.
Insufficient analysis of possible chemical exposures for Barwon Heads residents
3.117
Many submissions suggested that community members were exposed to environmental carcinogens in the Barwon Heads region in the 1980s and 1990s. Submitters referred to exposures to ‘disturbing levels of organochlorine pesticides, of which dieldrin was one’, to dust pollution containing ‘dieldrin and other pesticides’, and to ‘multiple chemicals such as temephos, malathion and others’.
3.118
Submitters suggested that past environmental exposures to one or more chemical pesticides may be the cause of many cases of cancer and autoimmune disease in Barwon Heads.
3.119
Some submitters raised concerns about the carcinogenicity of a number of organophosphate insecticides and synthetic pyrethroid insecticides and about the methods of application historically used for mosquito control by the City and former council entities.
3.120
In 2019, state and local government authorities responded to concerns about possible historical exposures to chemical pesticides, especially dieldrin.
3.121
However, these responses did not allay community concerns about possible historical exposures to the chemicals used for mosquito control. Chapter 5 discusses the ongoing concerns of some in the community about the historical use of chemical insecticides used in mosquito-control programs on the Bellarine Peninsula.
Trust and communication
3.122
Discovery 3227 submitted that the community had a ‘complete distrust’ of the department due to its role in the establishment of the Framework for mosquito management in Victoria, which mentions the use of organophosphate and pyrethroid insecticides. It also stated:
We believe to restore community confidence it is vitally important to have a truly independent epidemiological study, free of the Department of Human Services interference, free of Victorian State Government interference …
3.123
Mr Harrison told the committee that Discovery 3227 had shared its soil-testing results with the City and with the department. However, Discovery 3227 submitted that it considered both the department and the City to have conflicts of interest, and that it therefore did not collaborate further with state or local government authorities.
3.124
The department advised the committee that it had written to individuals, to the Barwon Heads Association, and to Discovery 3227, and that there remains an ‘open and ongoing invitation’ for community members to discuss their concerns with the department. The CHO also advised the committee:
It is a matter of great regret to me that Mr Harrison and Discovery 3227 refused to engage with me and the Department so that we could seek to understand and explore the situation together.
3.125
The committee also received evidence from the Expert Advisory Group explaining that the absence of detailed information about the cases of concern meant that it was not possible to precisely identify the optimal time frame for analysis, and that this had also limited other aspects of the investigations to date. Professor Hill told the committee:
… it would be extremely helpful to know what the community's, I guess, evidence of a cluster is. That evidence really needs to be based on the number of patients, the types of cancers they had, their date of diagnosis and their age …
Committee view
3.126
The committee acknowledges the difficulties associated with any investigation of a suspected cancer cluster. The committee noted the detailed expert evidence to this inquiry from epidemiological, toxicological, and cancer cluster experts about the scientific limitations of suspected cancer cluster investigations, both in relation to epidemiological concerns and in relation to exposure concerns. Unfortunately, investigations of suspected cancer clusters rarely fully explain or provide answers about individual cases of cancer.
3.127
The committee acknowledges that the Victorian Department of Health and Chief Health Officer’s work and investigations to date have not alleviated the concerns raised by some of the Barwon Heads residents. For some residents, the investigations to date have not provided clear answers or brought a sense of closure. In some instances, there have been doubts and misunderstanding around the methodology undertaken to do the epidemiological assessment. Discovery 3227 and a group of residents expressed the view that the Victorian Department of Health and Chief Health Officer have failed to address the community concerns about historical exposures to chemicals other than dieldrin.
Responses from the Victorian Department of Health
3.128
The committee is of the view that in early 2019 the Victorian Department of Health and Chief Health Officer acted promptly in responding to the Barwon Heads community’s concerns about a suspected cancer cluster, noting the additional difficulty for the Victorian Department of Health and Chief Health Officer of working with only the limited information available from media reports.
3.129
The committee recognises the willingness of the Chief Health Officer to acknowledge the shortcomings of the first investigation and to undertake additional rounds of more detailed epidemiological analysis in response to the community’s ongoing concerns and in response to this inquiry. In particular, the committee thanks the Chief Health Officer for agreeing to its request to commission further epidemiological analysis from Cancer Council Victoria, despite the prevailing circumstances of the COVID-19 pandemic and its impacts on Victoria in particular at that time.
3.130
The committee notes that the additional epidemiological analysis completed by Cancer Council Victoria in March 2021 extended the analysis to use data from 1982 to 2019. The committee heard that the updated analysis showed no increased incidence of cancer for 'all cancers', for 'all cancers (10–34 years)', of for any of the individual cancers identified except breast cancer. This latest analysis is publicly available online. However, it could be beneficial if the Victorian Department of Health offered community members the opportunity to discuss these results with officials. Research translation and scientific translation are paramount to ensure that people can understand and interpret the data correctly, and that they understand, for example, how standardised incidence ratios (SIRs) are estimated.
Gaps in information and analysis
3.131
As stated at the beginning of the report, the committee does not have scientific expertise in epidemiology or toxicology. As such, the committee will not further comment about the analysis undertaken to date by the leading experts engaged by Cancer Council Victoria. However, the committee has identified several shortcomings in the approach taken by the Victorian Department of Health and Chief Health Officer, which may have contributed to the community not being satisfied with the types and levels of investigations to date.
Autoimmune diseases
3.132
At the hearing in Barwon Heads, the committee noted that many of the residents who came to talk to the committee expressed grave concerns in relation to the rates of autoimmune diseases in the community. The committee acknowledges that the Victorian Department of Health prepared an information sheet responding to specific concerns about IBD. However, the committee is of the view that it may have been an insufficient and inadequate response to the issues raised by community members. The range of autoimmune diseases present in Barwon Heads is far greater than IBD. The committee sees value in the Victorian Department of Health developing communication and information materials about the prevalence of autoimmune diseases and potential causes and risk factors.
Exposure to chemicals
3.133
The committee heard from experts about the importance of undertaking an assessment of both the epidemiological concerns and the exposure concerns when investigating whether or not a cluster is present. After initially investigating the exposure concerns in relation to dieldrin, the Chief Health Officer failed to allay the deep concerns expressed by some in the community about the potential for harm caused by various chemicals that may have been used in the mosquito program.
3.134
The committee is of the view that the Victorian Department of Health should have considered as part of its responses whether specific kinds of chemical exposures might be related to a higher incidence of diseases in the region, including autoimmune diseases. Chapter 5 specifically considers the issues pertaining to the chemical insecticides used in mosquito programs on the Bellarine Peninsula.
Communication issues
3.135
From early on, it appears that very little communication occurred between the concerned residents and the Victorian Department of Health. This may have contributed to the total breakdown in communication that occurred over time between Discovery 3227 and the Victorian Department of Health. This has also likely contributed to the fact that Discovery 3227 and some residents have developed significant distrust towards the authorities. Without restoring a level of trust and information sharing, it is unlikely that any answers or explanations will be satisfactory to the community members who have been affected directly or indirectly by cancer or autoimmune diseases.