Bad science - Chris Busby and his articles on Fallujah

In two blog posts (Part 1 here and Part 2 here) we have discussed the issue with Chris Busby's studies on health effects in Fallujah, focusing on the sex ratio. In the present post there will be a summary of all errors and strange issues found so far by us in NPYP, and by others, in the two articles.

If you find any errors in our reasoning, or find further issues with the Busby articles that we have missed, then please leave a comment. Unlike Busby we are keen on getting it right.

It should be noted that when we write "Busby" we mean "Busby and co-authors". The co-authors share the responsibility for the articles, and the errors within them, but we suggest that they find more serious collaborators in their future efforts.

Furthermore it should be noted that we are also concerned about the reports of health effects in Fallujah and other parts of Iraq after the wars. Our scrutiny of Busby's articles are in no way reflecting our personal opinions about possible reasons for the health effects (short story: we do not know). Our message is that the serious flaws in the articles, and the alarmistic statements by Busby in media and elsewhere regarding the findings, should not be taken seriously by those who are concerned about the population in Fallujah. The research was probably initiated with good intent, but the results clearly show that Busby once again is more keen on promoting his own hypothesies rather than finding out what is going on.

For the independet network Nuclear Power Yes Please:

Mattias Lantz, Johan Kreuger, Nils Rudqvist, Michael Karnerfors, Liisa Petrykowska and Johan Kihlberg

 

1. Comments on the first Fallujah article by Busby and co-authors

Chris Busby, Malak Hamdan, and Entesar Ariabi, Cancer, Infant Mortality and Birth Sex-Ratio in Fallujah, Iraq 2005–2009, Int. J. Environ. Res. Public Health 2010, 7(7), 2828-2837 (direct link to the article here)

List of comments

This list will be extended if more things come up. Posts will also be corrected or omitted if they are shown to be in error.

1.1. Invalid argument (1/3)

On page 2829 (Section 1. Introduction) Busby writes:

The method is described fully with a sample questionnaire in Busby 2006 [7] where breast cancer rates in the town of Burnham on Sea, Somerset were reported. The study was later investigated by the official South West Cancer Intelligence Service and was shown to have given an accurate result for the breast cancer incidence rates.

The following link (COMARE  Statement on Green Audit Occasional Paper 2002/5) gives harsh criticism of the study that Busby refers to, not only about the way the questionnaire was used, but also about the interpretation of the results. It turns out that the only thing Busby was correct about in that study is an increase in breast cancer, but the reason for that had nothing to do with radioactivity. Instead it was a result of breast cancer screening programmes:

The excesses of breast cancer, by contrast, which were carefully mapped throughout the county, were shown to occur as a result of the breast cancer screening programme which calls up women for screening every three years according to their place of residence. The excesses were the additional undiagnosed cases unearthed at particular time points by the screening procedure.

1.2. Weak argument (1/1)

On page 2830 (Section 2.3 Strengths and Weaknesses) the strengths and weaknesses of the study are discussed. There is a claim that the leakage of cases due to deaths and population migration, causes the results to show minimum rates. This could very well be the case, but it is not necessarily true, the opposite could also be the case if there is a larger fraction of healthy people than affected people who leave the area.

1.3. Sex ratio fallacy (1/5)

On page 2831 (Section 2.5 Sex-Ratio) Busby writes

The population data in 5-year age groups was used to examine the sex ratio in 5-year birth cohorts.

It became evident at the BSRRW meeting in August 2010 (see information here and in links therein) the sex ratio was derived not from the number of born children, but from the number of children who were available at the time of the survey.

1.4. Sex ratio fallacy (2/5)

Busby has called the deviating sex ratio the most significant finding of the survey. As shown here, variations in sex ratio may be as large as in Fallujah, or even larger, without any obvious reason. The reason for the low sex ratio in Fallujah is very likely due to low statistical power.

1.5. Sex ratio fallacy (3/5)

On page 2832 (Section 3. Results and Discussion) Busby refers to very old studies (45 year or older) of effects on sex ratio, claiming that radioactivity causes mutagenic stress which in turn reduced the number of born boys. He ignores later studies, including later studies from the same people that he refers to, that show that there is no such effect due to radioactivity.

1.6. Sex ratio fallacy (4/5) / Invalid references (1/4)

References 10, 11 and 12 are about 50 years old and are published in German and French, in journals or conference proceedings that are not so easy to find. If the results from these references were significant there would surely be valid references from the same groups available in English.

1.7. Ignoring other possible explanations (1/4)

On page 2832 (Section 3. Results and Discussion) Busby writes, after referring to six studies (references 10-15) about affected sex ratios in relation to radioactivity:

Thus the evidence suggests that exposure to ionising radiation at low doses and specifically exposure to Uranium may cause a reduction in the sex ratio.

Busby ignores all other possible reasons for variations in sex ratio, and is keen on pushing for radioactivity being the only option. He does not mention a single one of the possibilities given on the Wikipedia entry about human sex ratio (here).

1.8. Ignoring other possible explanation (2/4)

We do not know for sure that depleted (or enriched, as suggested by Busby) uranium weapons were used in the battle of Fallujah. If they were not, then they can not explain the health effects.

1.9. Ignoring other possible explanation (3/4)

If uranium-based weapons were used in Fallujah, we have no evidence of exposure levels that would be likely to cause the observed health effects.

1.10. Ignoring other possible explanation (4/4)

If uranium-based weapons turn out to be the cause, it is quite remarkable that Busby does not even mention the known chemotoxic properties of uranium. There is no mentioning of any health effects in the kidneys, which is an established health effect from uranium. He focuses entirely on eventual genotoxic effects due to radiation, in order to promote his own theories in the matter.

1.11. Invalid argument (2/3)

Busby disqualifies any conclusions about the sex ratio in the age group 5-9 years due to leakage (for instance here), but sees no problem in using it in the age group 0-4 years and claim that the 18 percent decrease is significant. Admittedly, the amount of leakage may be lower in the age group 0-5 years due to the shorter time span, but to ignore the uncertainty the way Busby does and call the results significant is irresponsible.

1.12. Sex ratio fallacy (5/5)

On page 2832 (Section 3. Results and Discussion) Busby writes:

Lejeune et al. (1960) [11,12] examined the offspring of fathers who had been treated with pelvic irradiation; at high doses there was an increase in the sex-ratio, but this reversed in the low doses (around 200 mSv).

We have not access to the original articles, but when parts of the work is referred to in other works by the same authors (for instance here) the effect is not so obvious. Furthermore, it is clear from their work that Busby over-simplifies the whole issue about male foetuses being more sensitive to radioactivity. In several studies there is a trend of increased sex ratio if the father has been exposed to ionizing radiation, and a decrease in the sex ratio if the mother has been exposed. Busby states repeatedly that it is only the male that is sensitive due to the fact of having only one X-chromosome while the female have two.

1.13. Invalid references (2/4)

On page 2832 (Section 3. Results and Discussion) Busby writes:

Schull et al. 1966 [13] found a reduction in the sex ratio in A-Bomb survivor fathers (mothers “unexposed”) for children born 1956–1962 a reversal of an earlier finding by Schull and Neel 1958 [14] of a positive effect in the 1948–1955 births. It should be noted that there were external and internal irradiation effects in these groups, with the internal effects predominating in the later years.

But reference [13] actually says:

The suggestion of an effect of exposure on sex ratio in the earlier data is not borne out by the present findings. One can argue either that a small early effect has disappeared or that the original observation had no biological significance.

1.14. Invalid references (3/4)

On page 2832 (Section 3. Results and Discussion) Busby writes:

Yoshimoto et al. 1991 [15] found an overall reduction in the sex ratio for A-Bomb survivors for children born 1946–1984.

But reference [15] actually says:

Each of these studies has failed to show a significant genetic effect of exposure to A-bomb radiation with the single exception of an early report (12) of changes in the sex ratio of the children of exposed survivors, a report not confirmed by later studies (13).

Furthermore it says:

Given the importance of the sex of an individual in determining mortality rates, particularly in the early years of life, we have examined the ratio of male births to female births among the F1 children to determine whether a sex bias might exist as a result of the addition of the F1-Ext sample and the use of the DS86 doses. Because of the procedure for the sample selection, for children born during 1946-58 the proportion of males among the F1 children has to be examined in the children of proximally exposed parents whereas for the F1-Ext sample the data are for the group of children born to parents exposed to a dose of 0.10 Gy or more total kerma as calculated by the T65D dose schedule. In these samples there is no significant change in the proportion of males among the F1 children in relation to parental dose (Table 3), and thus no evidence that the mortality surveillance sample is biased toward one sex or the other.

1.15. Invalid argument (3/3)

On page 2834 (Section 3. Results and Discussion) Busby writes:

It is clear that the 0–4 population, born in 2004–2008, after the fighting, is significantly 30% smaller than the 5–9, 10–14 and 15–19 populations. This could be a result of lower fertility or early foetal losses in this cohort. It has been pointed out by a referee that it might also in principle be a result of the deaths of men in the 2004 fighting but this does not seem to be supported by the sex ratios in the men and women aged 25 and over.

Busby thus responds that the reasoning from the referee is not supported by the observed sex ratios in men and women aged 25 years or older. This is very odd of Busby, if we can not draw any conclusions from the 5-9 year age group, then we can certainly not do it from any of the higher age groups.

1.16. Invalid references (4/4)

On page 2835 (Section 3. Results and Discussion) Busby writes:

Tondel et al. have reported increased cancer risk in Northern Sweden peaking less than 5 years after the Chernobyl contamination and significantly associated with the levels of Caesium-137 fallout in municipalities [20].

Reference 20 is to the work by Tondel and co-workers, implying an early onset of cancer in Sweden after the Chernobyl accident. The study has been severely criticized for a number of reasons (here is some of the criticism by Holm et al., in Swedish, here is the response by Tondel, and here is the response to that by Holm et al.) and can not be seen as valid reference in order to strengthen the reasoning. Busby repeatedly refers to this study in support for his various theories about health effects due to radioactivity. It is notable that this is the only article that supports his reasoning, and that the validity of this article can be easily questioned.

Statements by Busby in relation to the article

Transcript of Busby's talk about Fallujah in Stockholm, 10 August 2010, here.

Busby's talk about Fallujah at the Human Rights Council, Geneva, 22 September 2010, here.

Links related to the article

 

2. Comments on the second Fallujah article by Busby and co-authors

Samira Alaani, Muhammed Tafash, Christopher Busby, Malak Hamdan,and Eleonore Blaurock-Busch, Uranium and other contaminants in hair from the parents of children with congenital anomalies in Fallujah, IraqConflict and Health 2011, 5:15 (direct link to the article here)

 

List of comments

This list will be extended if more things come up. Posts will also be corrected or omitted if they are shown to be in error.

2.1. Flaw in the method (1/4) - cause and effect?

The article shows (high?) levels of (enriched?) uranium in the hair from parents to children born with congenital malformations. The authors draw the conclusion that the parents have been exposed to (enriched?) uranium from weapons used during the 2004 battle of Fallujah, and that this is the cause for the malformations of their children. But the article fails to show a causal relationship, high levels of uranium does not necessarily lead to the observed health problems. It can be on the list of suspected reasons, but the article gives no evidence for that this is the reason.

2.2. Flaw in the method (2/4) - no control group

In order to prove a causal relationship, the article needs to show, among other things, that the health status of the children with health problems have parents with high levels of uranium while there are healthy children whose parents do not have high levels of uranium. But there is no comparison with any kind of control group, and therefore the article does not show any causal relation between uranium and health effects.

 

We appreciate the difficulties in performing the survey and obtaining the measured samples. But there is no excuse for the alarmistic statements and press releases that followed the publication of this article when there is no control group to compare with.

2.3. Flaw in the method (3/4) - cause and effect?

Table 3 on page 7 shows the levels of uranium and other elements for individual cases. The levels of uranium differ quite significantly even in the subgroups. For instance, the group of children reported with heart problems have parents whose hair concentration of uranium vary between 0.02 mg/kg and 0.40 mg/kg. This is a factor of 20 difference. No explanation is even attempted on why such large differences should lead to the same health effect. And again, there is no control group to compare with.

2.4. Invalid references (1/3)

The first study has limitations in accuracy due to the used method for the survey (knocking on doors), and the sex ratio is based on the number of children available at the time of the survey instead of the number of born children. The latter issue was brought up by people in the audience during Busby's talk at the BSRRW meeting in Stockholm, 10 August 2010 (transcript here), so Busby is well aware of it. In spite of this, the second sentence in the article says:
In addition to the increased cancer and rates and infant deaths, the epidemiological study [1] showed that there was a sudden significant drop in the sex ratio (an indicator of genetic stress) in the cohort born in 2005, one year after the battles which occurred in the city, suggesting that the cause of all these effects is related to the time of the US led invasion of the city in 2004.

See also our blog post here.

2.5. Flaw in the method (4/4) - no experimental uncertainties given

In Figure 5 the data points for the derived uranium isotope ratio are seemingly given without error bars. Could it be that the statistical errors are so small that they are not seen, or are they ignored? The text gives no clues on the uncertainties of the measurement technique. Thus the claimed finding of enriched uranium may be invalidated if the error bars overlap with the value for natural uranium (137.88). The very fact that no uncertainties are reported raises the suspicion that Busby once again are ignoring them on purpose (we have seen this before).
Below is our attempt to display the data in a more correct way, based on the information we can find in the section The Uranium isotope ratios, page 7-8.
  • The first picture (upper left) is identical with Busby's Figure 5, but we have added lines showing the level for natural uranium (solid line at 137.88) and the 95% confidence limits given in the text, i.e. values above 144.1 are considered to be depleted uranium, and values below 132.1 are considered to be enriched uranium. As seen all data points with more than 1000 counts of U-235 are within the limits for natural uranium. We could end here, claiming that none of the data supports any theory about anything else than natural uranium.
  • The second picture (upper right) is the same, but we have added the mean value and the standard deviation for the 14 data points with more than 1000 counts. We get the same values as Busby, mean ratio is 135.16 with a standard deviation of 1.45. Busby thus claims that the uranium is enriched to 0.73% instead of its natural composition of 0.72%. Why he disqualifies his own lower limit at 132.1 is truly a mystery to us.
  • The third picture (lower left) shows the same as the first picture, but we have added the statistical error for each data point. We may be in error here, but we have taken the uncertainty by applying the formula for error propagation for the ratio r=U-238/U-235, setting the uncertainty for each value as the square root of the number of counts. Essentially the contribution from U-238 is negligible to the total error, but we have included it. As seen 9 of the 14 data points above 1000 counts overlap with the line at 137.88. Any claim of enriched uranium should by now be seriously questioned.
  • The fourth picture (lower right) shows the same thing as the lower left, but we have included the weighted mean and its errors as red lines. The weighted mean is different from the arithmetic mean because it takes into account the statistical error on each individual data point. As we have seen before, Busby prefers to avoid this calculation, it usually disproves his theories. If our understanding of the data is correct, this once again disproves any statement about enriched uranium. The red box is fully within the 95% confidence limits defined by Busby, and it overlaps with the line at 137.88. We call it natural uranium. If we are in error, it is probably up to Busby to explain his methods better.

Figure 5 in Busby's second article about Fallujah. Upper left: The data as displayed by Busby, but we have added lines for natural uranium (137.88) and the 95% confidence limits where data are considered to be natural uranium. Upper right: We have added the arithmetic mean value and its upper and lower range, 1 standard deviation. Lower left: We have calculated the statistical errors for each data point. Lower right: We have added the weighted mean (red solid line) and the uncertainty on this value.

2.6 Incorrect claim - where are the data?

On page 10 (Section: The Uranium isotope ratios) Busby writes:

...and results of these measurements showed clearly that the Uranium in the soil was not natural. It was not, however, depleted Uranium. It was, in fact, slightly enriched, with ratios varying from 118 to 132. Under the conditions of the extraction we are able to assess the 95% CI limits from the count variance found in relation to the total counts. We are able to say that for defining natural Uranium the Ratios must fall in the range 132.1 < Ratio < 144.1. Values below 132.1 are thus enriched, above 144.1 depleted with p < 0.05.

So there is a claim of soil samples with ratios from 118 to 132, but it is not shown in any other way than this. And there are no indications about the accuracy of the measurement of each soil sample, only on the 95% confidence limits for natural uranium. We should thus trust that the soil samples have slightly enriched uranium just because Busby says so. Frankly, we don't.

2.7. Invalid use of statistics

Busby writes below Table 2 (page 7):

Elements which are found to be more than 2SD from the Literature [43] for Sweden unexposed mean are asterisked.

The intention is to call attention to the measured elements that deviate a lot from the Swedish measurement by Rodushkin and Axelsson [43] that is taken as a reference for comparison. Nothing wrong with making a comparison, but the way the comparison has been done is deceitful, by calling attention to the elements where the Swedish data deviates with more than 2 standard deviations (SD) from the Fallujah data. The standard deviations are from the Swedish data and not from the Fallujah data. In theory we could have a situation where the uncertainty in the Swedish measurement was 100 times better, then the deviation would be 200 standard deviations or more! The correct thing to do is to call attention to the elements where the Fallujah measurements deviate with more than 2 standard deviations. We find that for most elements the uncertainties are very large, sometimes larger than the measured values themselves! Thus the error bars overlap with the Swedish data for almost all elements, contrary to what Busby tries to make us believe with the comment below the table.

Below is a plot of the elements that are claimed to deviate with more than 2 standard deviations:

Data from Table 2 in Busby's second Fallujah articleAs seen the statement is correct only for 3 elements; Vanadium (V), Strontium (Sr) and the second Uranium sample (U2). Please note that the y-axis has logarithmic scale in order to allow all elements to be shown in the same figure.

2.8. Invalid argument (1/2)

On page 3 (Section. Anomalous health effects of Uranium weapons) Busby writes:

It is not our intention here to exhaustively discuss the arguments relating to the genotoxicity and fetotoxicity of Uranium; these have been rehearsed at some length in the literature. However, since Uranium is the only known radioactive heavy metal exposure in Iraq, it must be considered to be a major suspect for the cause of the effects found in Fallujah and also in the rest of Iraq.

The radioactivity of uranium (enriched or not) is relatively low, and the chemotoxic effects should occur well before any radiotoxic effects. As usual Busby ignores this in order to push for his own theories.

2.9. Invalid argument (2/2)

The levels of uranium are not exceptionally high when comparing with literature, for instance the levels in hair are sometimes higher in the study by Rodushkin and Axelsson. Busby discusses this, but claims that inhaled uranium is much more dangerous than uranium ingested from water with high uranium levels. The uranium in drinking water is dissolved to molecular levels while the inhaled uranium may be in particle forms of the order of micrometer size. We could buy this argument if there would be some visible difference in the hair concentrations, particles of micrometer size should end up in the human hair in a very different way than the uranium atoms dissolved in water.

2.10. Nonsense statistics given (1/1)

On page 5 (Section: Results - Initial hair study) Busby writes:
The mean ages of the fathers was 29.6 (SD 6.2) and the mothers 27.3 (SD 6.8).

Who gives population data in this way, with standard deviations, while simultaneously ignoring to give similar information on the measured variables? Only someone who wants to impress with seemingly good statistics while hiding it whenever it does not support the reasoning.

2.11. Invalid references (2/3)

References 26 and 29 are from Busby's fake journal European Journal of Biology an Bioelectromagnetics, where Busby has published several articles and probably reviewed them himself. No matter what, the journal does not exist and there is no way to obtain the articles. Any reasoning that refers to these articles for support should therefore be considered invalid.

2.12. Invalid references (3/3)

References 59 and 61 are from Busby's Green Audit reports. Although they can be obtained from the LLRC web site, they have not been through a peer review process and have therefore limited scientific value.

2.13. Manipulation of cited data

Figure 2 on page 8 displays data from reference 46, showing a reduction in uranium concentration in hair over time. The data in reference 46 were displayed as bar graphs, and any minor error may have occurred in the extraction of the data. This is not of importance, the data points seem to agree with the ones in the original article. More important is the fact that Busby has omitted the data points that do not follow the general trend. Inclusion of these data does not in any way change the general conclusions, it just shows that there are variations in the measurements. All three measurements show that there is a decreasing concentration of uranium in the hair as it gets older. But the fact that Busby has omitted them shows that he is not handling the issue in a serious manner, and it should be reason enough to consider a withdrawal of the article.

Here is a comparison of the plots. Busby's version as shown in the article:

Figure 2 in Busby's article, with data extracted from the article by Rodushkin and Axelsson

Here is our version, with the missing data inserted:

Our version of Figure 2, with the missing data from the original article shown as red dots.

Finally we show the original plot by Rodushkin and Axelsson from which Busby (and we) extracted the data. It is the third staple in Sample 2, and the third staple in Sample 3, that are missing in Busby's plot.

 

Extract of Figure 4 in the article by Rodushkin and Axelsson (Sci Tot Environment 2000, 262:21-36)

As already stated, the missing data should not in any way change the conclusions drawn by Busby, but we find it very serious that he tries to improve reality wherever he can by omitting variations in the data.

2.14 Weak resoning about uranium content in long hair

On page 8 is Figure 1 that shows the uranium concentration in hair from women with long hair. In contrast to the data from Rodushkin and Axelsson, the decrease is much slower, and for the woman with very long hair (NL1) there is no decrease at all. Busby writes

These authors showed that for Uranium the concentration falls rapidly and regularly by distance from the scalp, presumably as a result of washing the hair.

What Busby doesn't write is that this was a surprise to Rodushkin and Axelsson, who had expected the levels to be constant or even increasing in older hair (hair tends to absorb elements that floats around in air). There are some labile elements that are expected to decrease due to repeated washing of the hair, uranium is not one of them. It could still be the reason, but we only have the results from a single article to compare with.

Furthermore, Busby makes a lot of assumptions based on this, but conveniently neglects the question whether it is relevant at all to compare the hair of adult women in a war torn city in Iraq with the hair from two 12-year old girls and a 3-year old boy in the far north of Sweden (Luleå).

2.15. (there will be more...)

 

Statements by Busby in relation to the article

LLRC press release (here) with statement by Busby

Considering the careful phrasing in the article abstract, the following statement by Busby is quite alarmistic:

What we have found makes it perfectly clear that a new generation of Uranium based weapons exists, is being employed in all modern battlefields and leads to shocking increases in cancer and congenital illness in innocent civilians and soldiers alike. Whether there is slightly enriched Uranium to cover up the use of Uranium weapons, or whether the enrichment is an integral requirement of some new weapons system, what we see is the deployment of a device of indiscriminate effect with terrible and indiscriminate consequences.  It is most likely that this weapon is also being employed by NATO forces in Libya, and we will wait and watch with concern for increases in cancer and birth defects following this latest war.

Busby goes on with the usual conspiracy theories against him. The thought that the paper is not good enough to be considered by The Lancet has of course never struck his mind:

There has been a sustained effort to stop us making this study and then to stop us publishing it. We have been attacked by people writing to our funders so we had no money to pay for the analyses, to the Journals we were sending the study to even before we sent it, to our Universities and Hospitals and Institutions.  It was rejected by The Lancet without even being sent for Peer Review. It was rejected by the International Journal of Environment and Public Health where we published our previous paper, without being sent to a reviewer on the basis of pressure brought on the editor from outside by people who hacked into the computers of the authors and knew about the paper before the journal received it.

The implications of this discovery would be serious, if it was a discovery:

The implications of this discovery are extremely serious. First it means that all the measurements made for Depleted Uranium DU on Gulf War 2 veterans are now useless as indications of exposure to Uranium dust and must be revisited. Second it shows that Uranium is now being used routinely in anti-personnel weapons and no longer only in anti tank weapons. Finally it means that the focus of the NGOs on banning Depleted Uranium (DU) is misplaced and such a ban pointless since DU has not been deployed for a long time as the military rightly (and with some concealed amusement) say. The local civilian effects and global civilian exposures to this material represents a human rights issue of the greatest magnitude.

Malek Hamdan is no better in her phrasing:

This extraordinary discovery of a new uranium weapon should serve as a wake-up call to the entire world. It is as if the military were at war with humanity, secretivelywinning their battles with what is effectively a kind of delayed-action radioactive poison gas.  They cannot keep claiming that these radioactive weapons can discriminate in their effects between military and non-military targets. Because of this, enormous numbers of innocent people have died and will die in the future.Countless parents will watch their children with horror and pity as for several generations children will continue to be born with congenital anomalies as result of the genetic heritable effects induced by this exposure to uranium dust.

Statements by Busby in interview on Russia Today (RT), 25 October 2011

The TV clip and a transcript of the interview are found here. Please read it with all the question marks above in mind.

 

Other links related to the article

Links about DU and enriched U. Article by Emil Schön, Geneva talk, LLRC sites, RT, BBC, The Guardian