• 04 MAR 09
    • 0

    #1030: Outcome of the offical Hazelwood School cancer cluster investigation

    Note in the below media release where it is claimed that “Detailed EMR measurements were carried out using new equipment to give a better idea of exposures in multiple locations in classrooms and offices over a period of time”.

    As I have previously pointed out some time ago the Department of Health and Community Services (DHHS) EMF survey was conducted only AFTER extensive electrical work was carried out at the school to rectify a major ground current leakage problem. This problem would have likely resulted in excessive magnetic fields in the building but would have been eliminated before the DHHS survey. Thus any assurance of low EMF levels is only relevant to after the repair work, NOT BEFORE.

    Don

    *********************************************************************************

    Tasmanian Government media release

    http://www.media.tas.gov.au/release.php?id=26111
    Hazelwood School Cancer Investigations

    The Director of Public Health, Dr Roscoe Taylor, today presented to staff and parents the findings from investigations into reported cancer cases involving past and present staff at Hazelwood School in Hobart.

    The cancer investigations commenced in August 2008 and followed on from earlier investigations into concerns at the school in relation to Electro-magnetic Radiation (EMR) exposure from sources such as the nearby overhead power lines.

    The investigation was overseen by a steering group comprising a number of staff from the school, the Education Department and the Public & Environmental Health Service.

    The Tasmanian Cancer Registry was asked to analyse all cancers arising in staff who have worked at the school at any time over the period from 1978 “” 2006.

    Among the 263 staff there have been 19 cases of cancer over the 29 years, compared with an expected 11.4 cases determined from the Tasmanian average, allowing for age and sex.

    “Given the low numbers involved, this moderate excess barely reached statistical significance, but if the 63 employees who had worked at the school for less than one month were excluded (some of them had been there for just one or two days as relief staff), the figure dropped from 19 to 13 cancer cases compared with an expected 9.4 and this difference was not statistically significant,” Dr Taylor said.

    The Public & Environmental Health Service also looked for causal factors and environmental hazards. Detailed EMR measurements were carried out using new equipment to give a better idea of exposures in multiple locations in classrooms and offices over a period of time.

    The levels were then compared with other schools both near and far from powerlines. An effort was also made to assess whether there may have been any major changes in sources of EMR exposure at the school over previous years.

    The test results indicate that the level of EMR exposure in Hazelwood classrooms and offices was low and comparable with other schools whether or not they were near overhead powerlines.

    “Taking into account the lack of identifiable hazards, as well as the very small population we are dealing with, the most probable explanation for the numbers of cancers appears to be chance variation,” said Dr Taylor.

    “This was a rigorous and thorough investigation and while we probably can”t provide staff with the absolute assurances and certainty they would have been looking for, I think the results of the analysis mean we can be fairly confident that there has been no serious threat to public or occupational health at the school.

    “We were fortunate to have the assistance of Associate Professor Alison Venn of the Menzies Research Institute, who oversaw the statistical analysis as Director of the Tasmanian Cancer Registry.

    “I would also like to acknowledge the assistance of the staff and colleagues on the working group, and the cooperation of the Department of Education in gathering the necessary baseline information about staff,” Dr Taylor said.

    Frequently Asked Questions

    Why was the period 1978 “” 2006 chosen for this investigation?

    The Cancer Registry only began collecting data from 1978 onwards and only the data up to the end of 2006 is available at this point in time. As with all cancer registries in Australia there is always a delay of several years in finalising a validated dataset.

    Isn”t nineteen cases of cancer out of 263 people more than you would expect?

    Cancers are more common than most people realise. Cancer also occurs in younger people, although the risk increases sharply in older age. With our ageing population in Tasmania, on average one in three people can expect to experience a cancer during their lifetime. The risk of diagnosis of breast cancer for women before the age of 85 is now around 1 in 8 women.

    In this study, 8% of the past and present staff population during the study period have experienced a cancer diagnosis. The average age at the time they were diagnosed was 51.5 years. As one would expect, a number of the diagnoses have occurred after a staff member has retired or left the school.

    Can we be sure that all cancers cases at the school were identified?

    The Tasmanian Cancer Registry has a very thorough process for ensuring that all or virtually all cancers are registered and in this investigation a matching process involving multiple steps was employed to make sure that no staff members on the Registry could have been overlooked.

    The investigation was aimed at finding out whether an abnormal pattern of cancer types had occurred (eg whether all the cancers were of one specific type), and whether nor not the school staff population as a group had experienced unusually high cancer rates.

    Is this a cancer cluster?

    In the community, the term “cancer cluster” tends to mean an unusual aggregation (real or perceived) of cancers that are grouped together in time and space and that are reported to a health agency by a concerned group. The cause is often attributed to a location, and/or range of exposures.

    To the epidemiologist or statistician a cluster is defined as a greater than expected number of cases that occurs within a group of people, in a geographic area, or over a period of time, and that appear to be related to a particular cause or exposure.

    Most reports of alleged cancer clusters are not shown to be true clusters. A suspected cancer cluster is more likely to be a true cluster (rather than a chance occurrence) if it involves one or more of the following:

    § A large number of cases of one type of cancer, rather than several different types

    § A rare type of cancer, rather than a common type

    § An increased number of cases of a certain type of cancer in an age group that is not usually affected by that type of cancer.

    These criteria are discussed further below, but the real difficulty we face in determining whether or not the Hazelwood School situation represents a cluster is that the numbers are too small to enable meaningful conclusions.

    There are a number of other indicators suggesting that if there has been a cluster, then it has most probably occurred through random variation.

    If biological principles are applied about “latency” periods between the time of exposure to a cancer causing agent in the environment, or the time from when a cell first starts to become cancerous and the subsequent appearance of a diagnosed cancer years later, the cancer figures overall tend to be diminished and statistically non-significant.

    In addition, if the figures are analysed after excluding the cases with only very short time periods of working at the school (e.g. 6 staff worked less than one month) then the increased number observed overall also becomes statistically non-significant.

    If only the cancers that commenced more than 5 years after first employment are included, and people who worked less than one month are excluded (both of these approaches would fit with what we understand about cancer causation) the number of observed cases is 7, which is the same as the number expected (7) for this group of staff.

    In other words, screening out cases that “” for timing reasons – could not realistically be associated with exposures at the school, points us away from the overall increase in numbers being somehow connected with the school environment

    One problem for cancer epidemiology is that random events are not distributed evenly, and this is particularly true in comparing small population groups with one another. This can mean that in one workplace the perception that there has been an above-average number of cases can be quite correct (as in this case), and it is more likely that that particular workplace will be highlighted and come to our attention as a possible cancer cluster. But our investigations are unable to say whether all the other workplaces have a high, low or average number of cancers.

    The NSW Cancer Council has a useful on-line fact sheet on cancer clusters at http://www.nswcc.org.au/editorial.asp?pageid=2305&fromsearch=yes

    Was there an abnormal pattern of cancer types?

    A range of cancer types has occurred, with breast, colon, lung, skin, cervical, and haemopoietic cancers (including lymphomas) being among those recorded. There is nothing particularly “rare” about any of these cancer types. But the number of cases and size of the population is too small to draw meaningful conclusions regarding any cause for an apparent increase with some of these. (Similarly, some cancer types are “missing” but the population being studied is too small for that to have any meaning.)

    Why hasn”t the investigation looked at children at the school?

    The decision to focus on staff alone for this investigation was made by the working group, because that was the best starting point. The initial concerns that were raised were more in relation to concerns about cancer rates among staff and ex-staff.

    A major difficulty with investigating illness rates among students at the Hazelwood School is that as a group they have a range of underlying serious conditions that would make it very difficult to meaningfully compare their health outcomes with the general population. Some of the genetic conditions in some of the students are also associated with a significantly increased risk of cancers such as leukaemia, early in life. The priority has therefore been to assess and exclude the existence of any local environmental hazards for both staff and students.

    Leave a reply →