Dear Editor,

I most certainly agree with the “reviewer’ comments” cited by the authors in the text of their paper; “The paper is timely and can be potentially useful for health policy makers” (p. 1032). However, “timely” in this case is a bit of an oddity. In response to a recent critique of progress that I had written (Betts, 2005) a rather noted philosopher, who is kind enough to review scholarly work for me from time to time, asked me; “I would think that this language [of Progress] is pretty hollow now. Do your colleagues really believe it?” The point is, that serious thinkers have not “believed” in Progress since – well perhaps since Nietzsche (1990) suggested that;” “Mankind surely does not represent an evolution toward a better or stronger or higher level, as progress is now understood. This ‘progress’ is merely a modern idea, which is to say, a false idea.” (p. 4), over a century ago. Though, certainly since Ivan Illich gave the term iatrogenesis a whole new meaning. And yet virtually all of our social and political institutions, particularly those of health and welfare, are founded, perhaps even premised on the idea of Progress.

One wonders then, why it has taken so long for many of us, who generally consider ourselves to be health practitioners of one kind or another, to catch on to this serious problem of Progress? Could it be that we have subscribed so heavily to a techno-rational science model, again particularly in the health and welfare industry, that all of our, so highly advertised, critical faculties have become stuck in a paradigm of delusion? Indeed, in a recent article by Roy (2005), a chemist by the way, aptly titled in my view; Scientism and Technology As Religions he claims that; “Science-and-technology is the most powerful force under human control; hence, scientific fundamentalism is the most dangerous.” (p. 836). Of course Roy isn’t the first to suggest that techno- rationalism (or scientism) may well constitute a religion of sorts, in fact, once again a century ago, Nietzsche claimed something similar, as have many others. In fact, more recently, Schaler (2002) has proposed that health itself had indeed become a religion, with “the medical profession as a priestly caste.” (p. 67).

Those of us who do the work of healthcare (be it diagnosis, intervention, education, promotion and what not) must now begin some serious thinking, or at any rate begin to pay attention to some of the serious (anti-modern if you will) thinking that has characterized the 20th century. Perhaps we might begin with a question of the profoundest sort that has already, and again recently, been proposed by Joseph Heath (2004); “Is it possible to emulate many of the attractive features of western societies, while avoiding the social pathologies?” (p. 665-6). While I do agree with the authors’ claim that; “Public health needs to be more passionate about health issues associated with human progress and adopt a health promotion stance.” (p. 1033), we must also be careful for at least two reasons. The first concerns what both Fitzgerald (1994) and Fitzpatrick (2001) refer to as The Tyranny of Health, while the second is perhaps best summed up by Ulrich Beck regarding his theory of reflexive modernity (or second modernity); “Due to the close link between institutions that depend on one another to exist, it can be assumed that processes of change and transformation in particular parts of the structure will trigger problems in other parts (the side-effects of side-effects)…” (Beck & Lau, 2005, p. 533). Indeed, side effects are the order of the day, and nothing happens without them happening. The problem then, for those who claim to be engaged in the fixing of problems, is which side effects are good and which are bad. Or put differently, which are acceptable and which are not. Moreover, these are socio-political and socio-cultural issues of value, and they are anything but scientific. Finally, in the spirit of the authors’ well put claim that; “public health has to become more assertive and politically aware…” (p. 1033), with, perhaps, a continuing attention to Foucault, let us begin a plea for the progress of serious thinking about health, rather than maintaining the modern course of unhealthy Progress.

References

Beck, U. & Lau, C. (2005). Second modernity as a research agenda: Theoretical and empirical explorations in the ‘meta-change’ of modern society. British Journal of Sociology, 56(4), 525-557.

Betts, C. E. (2005). Progress, epistemology and human health and welfare: What nurses need to know and why. Nursing Philosophy, 6, 174–188

Fitzgerald F. T. (1994). The tyranny of health. N Engl J Med, 331, 196-198.

Fitzpatrick M. (2001). The tyranny of health: Doctors and the regulation of life style. Routledge: New York.

Heath, J. (2004). Liberalization, modernization, westernization. Philosophy and Social Criticism, 30(5-6), 665-690.

Nietzsche F. (1990) The Anti-Christ (tr. R.J. Hollingdale). Penguin, New York.

Roy, R. (2005). Scientism and technology as religions. Zygon, 40(4), 835-844.

Schaler, J. A. (2002). Moral Hygiene. Society, May/June, 63-69.

Dear Editor,

The Journal of Epidemiology and Community Health has recently published two papers by Knox[1,2] which conclude that “childhood cancers are strongly determined by prenatal or early postnatal exposures to oil based combustion gases especially from engine exhausts. 1,3-butadiene, a known carcinogen, may be directly causal”. This is a strong conclusion for a study whose novel analysis uses only the birth and death addresses of childhood cancer cases and which uses atmospheric emissions data from 2001 as a surrogate for exposure that may have occurred as early as 1937 in the first report, or as early as 1955 in the second report.

Furthermore, Knox notes that the basic method of the study, and all the results, depend upon a premise of a “migration equilibrium” among the general child population. Strong claims are also made about the studies in two accompanying editorials in the journal.[3,4] One editorial acknowledges the lack of a proper control group, but still claims that the papers contain “striking evidence relating atmospheric contaminants, and mostly emissions derived from engine exhausts, with increased risk for children of dying from leukaemia and other cancers”.[3] A second editorial describes the work as “unique in its capacity to combine in a meaningful way a series of routinely collected data on births, deaths, residences, sources of exposure, and specific exposures”.[4] However, this letter will demonstrate that a migration equilibrium cannot be assumed and that Knox’s estimate of relative risk largely reflects the relative likelihood of families living in the inner and outer exposure zones to move house. Consequently the studies provide no meaningful results and the findings should be disregarded.

The two reports by Knox aim to identify specific causal agents for childhood cancer using methodology that links the birth and death addresses of fatal child cancers (pre 1980) to current emissions hotspots for specific chemicals. Knox has developed a methodology based on the proximity of migrant children to pollution sources at birth and death which he notes was initially dictated by the absence of a suitable set of non-cancer controls in the study from which the case material was extracted. For a particular definition of exposure (for example, living within 1.0 km of an emissions hotspot), Knox estimates the relative risk among the subset of children whose birth and death addresses differ in respect of exposure by calculating the ratio of outward migrations (birth address exposed and death address unexposed) to inward migrations. This letter focuses on this estimator of relative risk because it has some plausibility and the approach has been likened to a case-crossover study design.[4] However, Knox also states that the ratio of births to deaths is an adequate estimator of relative risk for short distances around emissions hotspots although this clearly gives results that are inconsistent with the migration ratio unless all families move house.

Knox acknowledges that the validity of his approach depends upon the premise of a short term migration equilibrium among the general child population and the second report[2] claims that this has been validated in the earlier report.[1] However, Knox presents no data to test the validity of this critical premise in either report and he only presents a calculation of the scale of demographic movement that would be necessary to explain the migration ratios that he has observed. In the case of 1,3-butadiene, Knox calculates that there would have to have been a net general migration rate among children sufficient to have depopulated all the exposed zones over the study period. However, this argument is patently wrong as it overlooks the fact that the numbers of children in the exposed areas will also be replenished by children born in these areas. Knox’s premise of migration equilibrium is not only untested, it is also readily apparent that is not valid as it completely ignores the contribution of births (and deaths to a much lesser extent in the case of children) to the maintenance of a population equilibrium. For example, consider the housing stock at the foot of the housing ladder that is rented or purchased by many young couples before they start a family. Many of these couples will move to larger properties in better residential areas after they start their families. Other families with young children are unlikely to move into the properties and replace the children that have left. However, the area will not become depopulated of children as couples without children will move into the properties and start their families as the cycle repeats. In contrast, many children living in larger properties in better residential areas will grow up there and move out as adults, and their parents are likely to be replaced at some point by families with young children.

Kogevinas and Pearce[4] correctly identify that the findings of Knox’s studies will be invalid if there is not a short term migration equilibrium among the general child population or if there is a specific tendency for families to move away from environmental hotspots after the diagnosis of cancer in the child. However, they incorrectly state that Knox attempted to address these two assumptions by restricting his analysis to short migrations of less than 1.0 km. Knox did not perform such an analysis in either report and most of Knox’s analyses are restricted to migrations that were more than 1.0 km in distance. However, there is sufficient information in tables 1, 2 and 5 of Knox’s first report[1] to perform the analysis suggested by Kogevinas and Pearce for 7 of the 8 exposures that had the highest migration ratios in the analysis of migrations greater than 1.0 km. The short distance migration ratio for benzo(a)pyrene (BAP) could not be calculated because Knox gives different numbers of hotspots in tables 1 and 5 for BAP (the lower limit for BAP hotspots given in table 2 of Knox’s first report is also higher than the figure given in the National Atmospheric Emissions Inventory (NAEI) emissions maps[5]). However, the other 7 migration ratios for short migrations (less than 1.0 km) range from 1.43 (1,3-butadiene, 33 outward and 23 inward migrations) to 0.67 (benzene, 10 outward and 15 inward migrations) with a median of 1.02 (PM10, 44 outward and 43 inward migrations). Hence, the migration ratios for migrations of less than 1.0 km provide no evidence of an association between exposure and childhood cancer.

Table 5 of Knox’s first report[1] also provides information about the proportions of children born in the inner (within 1.0 km of a hotspot) and outer exposure regions that migrated (any distance). For all the exposures in Knox’s table 5, the proportions of children that migrated were much higher for the inner exposure regions than the outer regions. The largest difference is seen for 1,3-butadiene (inner migration rate = 80.1%, outer migration rate = 59.3%) and the smallest for BAP (inner migration rate = 74.1%, outer migration rate = 65.1%). These differences are strongly related to the magnitude of Knox’s estimate of relative risk for exposure (based on migrant children with discordant birth and death addresses). Knox obtained the highest estimate of relative risk for 1,3- butadiene (3.73) and the lowest estimate of relative risk for BAP (1.92). Furthermore, if the inner and outer migration rates are used to calculate the odds ratios for exposure when comparing migrant and non migrant children, then these are similar in magnitude to Knox’s estimate of relative risk for an exposure. For example, 1,3-butadiene has an odds ratio for migration of 2.73 and BAP has an odds ratio for migration of 1.53. However, there is absolutely no reason to expect any relationship between a child moving house and their exposure. Knox’s estimate of relative risk is largely a measure of the relative likelihood of a family to move house in the inner and outer exposure zones and almost certainly reflects differences between the type of housing to be found within the inner and outer exposure regions. This is readily apparent when one looks at the NAEI emissions maps.[5] Most of the exposures with high outward/inward migration ratios have exposure hotspots that are mainly situated alongside busy roads and/or in inner city areas. The housing in these areas is often the sort of housing that families with young children will try to move away from i.e. small terraces and flats. Nitrogen oxides (NOx), 1,3-butadiene, benzene and PM10 are examples of such emissions and the emissions map for NOx (http://www.naei.org.uk/images/mapping_2002/6_large.png) is typical of this group of emissions. In contrast, the locations of hot spots for emissions with a high inward/outward migration ratio are very different. A good example is the emissions map for nickel (http://www.naei.org.uk/images/mapping_2002/16_large.png) with many of the hotspots on the edges of urban areas situated close to housing areas that are attractive to families. This also explains the surprising fact that Knox obtained similar results in analyses of the different tumour classes. Even if some of these agents are the cause of childhood cancer, it is biologically implausible that they will affect all tumour classes.

It has been demonstrated that the outward/inward migration ratios are not measuring the carcinogenic effect of substances or the non-specific emissions from certain types of site as claimed by Knox. However, it would also be very surprising if the 2001 modelled NAEI emissions data are a good surrogate for exposure that occurred between 21 and 64 years earlier (21 to 46 years in the second report). Knox focuses on 1,3- butadiene, but this is one of the compounds for which the 2001 NAEI emissions data are least likely to be representative of exposures so many years earlier. Stationary source emissions for 2002 have now replaced the information for 2001 on the NAEI website, but the emissions spreadsheet for 2002 shows that there were 200 tonnes of emissions from 1134 stationary sources. Over 80% of the emissions occurred at just two locations (97% from 8 locations). Most of the 1134 sites (96%) are motorway service stations and supermarket filling stations that would not have existed when the majority of the children in these studies were born. Of the total of 3.5 ktonnes of 1,3-butadiene estimated to have been emitted in 2002, 2.6 ktonnes were due to road transport activity. Motorways and major roads are clearly shown as hotspots on the NAEI emissions map. However, none of the motorways existed when the first children in the studies were born and many did not exist when the last children died in 1980 (although proximity to a motorway is also one of the individual exposures studied by Knox).

It should also be noted that Knox is incorrect in two respects when he states that the workplace exposure standard for 1,3-butadiene is 1 part per billion (ppb), a limit which he says is not designed to prevent childhood cancers. In fact, workplace exposure is limited to a maximum of 10 parts per million and the 1 ppb standard that Knox refers to is the current UK standard for 1,3-butadiene in urban air (a running annual average). This is a standard that is meant to protect all of the population, including children, and it has been recently reviewed by the UK Expert Panel on Air Quality Standards (EPAQS) who concluded that “concentrations of 1,3-butadiene in the ambient atmosphere of 1 ppb constitute so small a risk to the population as to be undetectable by any feasible study”.[6]

In conclusion, it is clearly evident that Knox’s underpinning assumption is not met and hence that the study has no control group and the results are not interpretable. However, even if Knox had included a comparison group of children who did not develop cancer, it is debatable whether the results would have had much value because of the lack of valid exposure information. An analysis restricted to short migrations of less than 1.0 km which is less dependent upon the migration equilibrium assumption,[4] shows no evidence of an association between childhood cancer and birth proximity to Knox’s pollution hotspots. The studies are fatally flawed and there is no basis for Knox’s conclusions and the assertions made in the editorials that “studies such as this clearly indicate the potential importance of specific environmental exposures in the causation of cancer in children”[4] or that “Knox’s findings should be seriously taken into account by decision makers concerned with environmental emissions’ limits and control” [3]. On the contrary, these reports clearly indicate the dangers of conducting epidemiological research without a proper comparison group and sound exposure assessment methodology, and the need to challenge arguments that are presented with no supporting evidence.

Competing interests: The author provides consultancy services to a range of customers including the chemical industry. The author reviewed these publications on behalf of the Lower Olefins sector group of the European Chemical Industry Council (Cefic). However, the opinions expressed in the letter are entirely those of the author.

References:

1. Knox EG. Childhood cancers and atmospheric carcinogens. J Epidemiol Community Health. 2005;59:101-5.

2. Knox EG. Oil combustion and childhood cancers. J Epidemiol Community Health. 2005;59:755-60.

3. Garcia AM, Alvarez-Dardet C. A journal for evidence based policies. J Epidemiol Community Health. 2005;59:716-7.

4. Kogevinas M, Pearce N. Geographically based approaches can identify environmental causes of disease. J Epidemiol Community Health. 2005;59:717-8.

5. National Atmospherics Emissions Inventory. http://www.naei.org.uk/mapping/mapping_2002.php

6. EPAQS. Second report on 1,3-butadiene. London: Department for Environment, Food and Rural Affairs (Defra), 2002.

Dear Editor,

The idea that the diseases of the future will increasingly have their aetiological roots in 'comfort' is at first sight an appealing and powerful one. However, to accept this idea uncritically could in fact be part of the problem. Just to take one example for brevity: the idea that everyone desperately wants to confine themselves to moving around in a motor car. I often have arguments with people who claim that the car is a precious guarantor of 'freedom'. One thing that argues against this is the difference in the effects on the prices of houses when it is proposed to build a road nearby versus what happens when improvements in local public transport take place. In the former case, house prices plummet. In the latter case, they generally tend to shoot upwards. In this way the market gives us a clear signal about preferences if people are really given a choice. The type of 'comfort' disease that arises from our ever diminishing levels of exercise is in part the result of an unholy alliance between the oil and motor vehicle industries, and governments who would rather not deal with unionised skilled labour forces needed to run transport systems. A furtive privatisation of transportation, by cutting services and forcing people into cars, took place long before what was left was officially privatised. A recent ONS report shows that households now spend on average over £50 per week on their motor cars, one of most expensive items in overall living costs, and around 1/3 of the total guaranteed income level for pensioner households. So I do not really think the pervasiveness of the car can be regarded in any simple way as the result of the human attachment to comfort.

Dear Editor,

As a reader, I found the paper[1] interesting, and I hope that the authors could provide extra information. In table 2, there were about one third of participants did not answer their income level in all the three area “SES” strata, However “Occupation” and “Education” may be associated with income level in the same way among all the participants. So the association between “Education”, “Occupation” and income level among people who reported their income, could be applied to predict or gain more information on the income level among people, who had provided information on their “Education”, “Occupation” but not on their income level. (Chi- Square test showed that the proportion of people who did not answer their income were significant associated with area “SES”)

People are likely to choose a place to live according to their life style, therefore people who are able to choose their living area, may choose a place that meets their physical activities needs, and people who have high income but choose a place not convenience for physical activities, may not like physical activities at all. However this possible confounder could be reduced by restricting the analysis to subjects who were unlikely to be able to choose a place to live, for example, those who are young or those who are in low income. These may lead to a large “CI”, but the mean of “OR” could be considered as a “point estimation”.

Reference:

1. Kavanagh, A.M., et al., Urban area disadvantage and physical activity: a multilevel study in Melbourne, Australia. J Epidemiol Community Health, 2005. 59(11): p. 934-940.