Disclaimer Disclaimer Disclaimer Disclaimer Home           Home Home           Home About  site   About  site About  site   About  site About  me    About  me About  me    About  me Current reading   Current reading Current reading   Current reading CONTACT   CONTACT CONTACT   CONTACT Links                  Links Links                  Links Book shelves      Book shelves Book shelves      Book shelves
Blog Archive    Blog Archive Blog Archive    Blog Archive Recent posts  Recent posts Recent posts  Recent posts
I was recently asked to review a new book for an academic journal. The book, “The Invisible  Rainbow”, is quite interesting and well written but does require you to believe in a conspiracy  to hide the real detrimental effects of non-ionising radiation that was started from its  invention; so yeah, there is that… (I will put the link to the actual review up here once  published).  So part of the evidence that shows that the introduction of electricity in society has resulted  in the increase in the numbers of newly diagnosed cases of a variety of different diseases  included in the book builds on work by Samual Milham, who batched these “Diseases of  Civilization”; principally cancer, diabetes, and heart disease, but including some others as  well. Dr Milham described his hypothesis in a book ‘Dirty Electricity. Electrification and the  diseases of civilization’, which I believe is considered a classic in certain groups, as well in a  number of scientific papers (main link, link2, link 3). Together with Igor Burstyn, also a  contributor to this blog, I wrote a refutation regarding one of these papers (link), but the  epidemiological principles in that letter relate to the other papers dealing with electricity as  the causal factor for the ‘diseases of civilization’ as well. Anyway, this blog is not aimed at  discrediting that hypothesis per se, and I do have some sympathy for the difficulty of trying to  investigate such a generic hypothesis in the absence of available funding to obtain the most  appropriate data and do it properly (as a result, one is forced to rely on publicly available  data), but I would like to discuss my interpretation of those data here.  Hopefully, someone “in the know” will read this and will use the comments section to explain  what is wrong with my inferences, and why therefore electricity is the cause of these  ‘diseases of civilization’. I am not trying to be cynical or cause some angry never-ending spat  here or, for example, on twitter; I am genuinely interested in the whys and hows, and I think  so would quite a lot of other people….  The data I am using is provided in several tables in “The Invisible Rainbow”, which I copied  across (and so can you if you buy the book; as said, if you find this topic of interest it is an  interesting book to have), and describes the level of household electrification in all US states  in the years 1931 and 1940, together with the state-level mortality rate (per 100,000 people)  for cancer, diabetes and heart disease. I have recreated the figure in the book related to the  cancer mortality rate below (for reference, this is Figure 6, page 252, but without names of  the states added): The interpretation of this figure, and similar figures for rural diabetes and heart disease  mortality is that you can see a clear and positive correlation between the rate of rural cancer  (which, incidentally, should be the Y-axis and not the X-axis) and the percentage of  electrification (which, vice versa, should have been the X-axis), which shows that more  exposure to electricity leads to a higher incidence of cancer in rural areas. Now there is an  immediate an obvious problem with this interpretation, and that is that it typically takes a  long time – decades – between exposure to a carcinogen and a detectable cancer to have  grown (called the lag), so the 1940 electrification is not the relevant exposure here. This  problem would be less of an issue, if at all, for rural heart disease and diabetes…  Regardless, let’s ignore that for now and look at this in a bit more detail (I will only look at  cancer in this blog; results for the outcomes are pretty much comparable).  If we overlay the 1931 and 1940 data (and switch the axes around), we get the following  figure:     Showing that from 1931 to 1940 household electrification has increased somewhat, as you  would expect, and overall cancer incidence has increased as well. It also shows that this  increase in the cancer rate was higher for areas with higher household electrification. So all  in all that seems pretty conclusive, and is indeed interpreted as such.   This is however the comparison of two cross-sectional datasets, which is generally not  considered very strong evidence. We can do better and look at the difference from 1931 to  1940. After all, if electricity indeed causes cancer then a change in the electrification rate  should also result in a change in the cancer incidence rate (again, ignoring the lag). With the  data provided in the book this is easily done. Surprisingly, at least to me, in some state the percentage of household electrification went  down from 1931 to 1940. That seems unlikely to me, but let’s just assume the numbers are  correct and have a look at only the states where electrification increased, to see if this  confirms the hypothesis. So that interesting. In the 9-year time period the percent household electrification, where  this increased, ranged from 0.8% to just over 16% and the corresponding cancer mortality rate  in these states increased from about 3 per 100,000 people to 52 per 100,000. Importantly for  our story though, there is, if anything, a negative correlation between the increase in  electrification and rural cancer mortality rate: the higher the increase in electrification, the  lower the increase in cancer risk (on average). In other words, these data really do not  suggest an association between electrification and rural cancer rates.  We can do the same for those strange states where household electrification went backwards,  and see pretty much the same picture, but in mirror format. Again, there is very little  evidence of a correlation; the apparent decrease is only there because of the 2 (only 2!)  states where both the electrification rate and the cancer mortality rate went down. If it  wasn’t for those two states – New Jersey and especially Massachusetts – there would be no  correlation at all.  From these two figures we can pretty much conclude that there is hardly a correlation, but definitely no evidence of a (causal) association between the two. In fact, it is possible to show this even clearer when we take the average difference in cancer rates of those states where the household electrification rate was within +/- 1%, which can be interpreted as the result of any other changes over that decade that are not electrification (for example, improvements in medical diagnostics, improvements in medical techniques, faster/easier transport to hospitals, but also maybe, for example, some other environmental exposure (we don’t know)), and subtract this from the measured difference in rural cancer mortality rates. What’s left then may be associated to electricity. So I did this, and the result is shown in the figure below:   The figure has 4 sections, with states where both the electrification rate and the rural cancer  rate increased in the top-right and conversely where they both declined in the bottom-left  (both after subtracting 24 (per 100,000 population)) from reported rates as described above. The grey line shows the remaining correlation, and I am sure you can also appreciate that  that pretty much shows a horizontal line; from which we can conclude that there is no  correlation, let alone a (causal) association between rural cancer mortality and household  electrification (in fact, there is a small positive correlation, indicating that the cancer  mortality rate increases on average by 0.03% per 10% additional electrification; in other words  not relevant).   * Now, just in case someone (not you, obviously…)  were to point out that I may just have picker cancer because that was the only one where there was no association, because for example because of the lag, these are the figures for rural diabetes morality and rural heart disease mortality, respectively: If we are to conclude anything from the above figures, I think we can pretty much say that  heart disease mortality is not related to household electrification either, while the more a  state electrifies households, the more the diabetes mortality rate decreases.    * The above all fits with the generally accepted alternative hypothesis, that  improvements in medical diagnosis and registration have resulted in increased  identification and correct registration of new cases of various diseases and hence an apparent increase in disease risk, but not necessarily mortality from those  diseases. In a typical constructivist approach I suggest that the impact of new  developments and technologies is best shown by the fact that the largest  decrease in the cancer rate was observed for Massachusetts, known for its  science and technology, as well as already having an >90% household  electrification rate in 1930-1940, while using a similar approach I suggest that  the benefits of medical therapies, which are related to electrification, is best  shown by the decrease diabetes morality risk with increased electrification; as  shown in the figure above.    Of course, there is also room for other (environmental, and other, most importantly lifestyle)  exposures to play a role in this, and contribute to increased incidence rates of cancers,  diabetes and heart disease observed in these data and continuing in national statistics.  However, it is very unlikely electrification is one of those….
Serving the community with a smile.......... ................in a Public Healthy kind of way
Back Back Back Back
Electrification and the diseases of other causes