Follow-up on Anti-Vax

So, apparently my impression that anti-vax was a right wing cause (anti-government mandates/one-world government or whatever) was unusual and most people think it is a left-wing cause (anti big-pharma/pro natural remedies etc). Turns out that neither is the case and that there are people on both the left and the right (at least in the US) who are anti-vax. Actually, it seems that there is a slight maximum of concern about vaccines near the center of the political spectrum, as shown on this graph from the linked article:

 

Transmission of Infectious Parasites Slows with Rising Temperatures, Researchers Find

So says a commentary in Nature on an article in Biology Letters*. This result is from a study with rodent malaria rather than human malaria. If this extends to human malaria it could be another factor explaining why we don't find a positive relationship between temperature and malaria.** Our assumption has been that interventions have been successful in overcoming the tendency to more malaria in warmer climates. But maybe that tendency is not so strong in the first place.

* Link to letter in Biology Letters from Nature is broken and I can't find it on their website.
** My coauthor, Simon Hay, is quoted in the Nature piece.

Pascual et al. Respond

Mercedes Pascual and colleagues have written a response to our paper in PLoS ONE. Our paper is largely a response to the Pascual and colleagues criticism of our 2002 papers in Nature, Emerging Infectious Diseases, and Trends in Parasitology.

I won't comment on it further at this stage until my colleagues have had time to read it.

Why Our Malaria and Climate Change Research is Important

This morning I was interviewed by Clement Paligaru on the Breakfast Show on Radio Australia. This is Australia's international broadcaster. One of the questions that he asked me was why it was important to know that malaria was declining despite climate change rather than thinking that malaria incidence was increasing due to climate change. What I came up with was that if we think that the disease will just become more widespread in the face of climate change regardless of what we do we might be apathetic about doing the things which have reduced the incidence of the disease. I'm not sure how convincing that is. But the rise in cases in the 1990s was probably due to resistance to chloroquine, an older antimalarial drug, and not climate change. That is something that is important to know.

Malaria Cases in Kericho, Kenya

In my IPCC AR5 chapter writing group we have been told to think of "iconic figures" we could use in the chapter. In our paper on malaria and climate change in East Africa, I think this is the iconic figure:



It just shows monthly malaria cases at the tea estate hospital in Kericho, Kenya. The increase in malaria in the 90s was linked by some researchers to climate change. But in this decade, malaria cases have collapsed at this location. In the meantime we find that there does seem to be a significant increase in temperature especially when the most recent years are included in the analysis. The figure is simple and really easy to understand.

Temperature and Malaria Trends in Highland East Africa

Our article: "Temperature and Malaria Trends in Highland East Africa" is published today in PLoS ONE. I've noted some of the key points already in a series of posts as the article developed. The key points are:

• In 2002 we (Hay et al., 2002) could find little evidence of significant increase in temperature or precipitation in the East African highlands in a period of a surge in malaria cases. It, therefore, seemed most likely that the increased malaria was mainly due to other causes such as increased drug resistance.


• This was a very controversial finding, which has been discussed by various other researchers who generally find that there has been a significant increase in temperature in these areas and that, therefore the increase in malaria was likely due to this.


• We have now tested the more recent versions of the UEA CRU database as well as the temperature series from Kericho in Kenya prepared by Omumbo et al. (2011) using a new robust test for trends in series.


• Using the new test, we find that there is no significant trend in the data we originally tested but that there is a significant increase in mean temperature in the newer versions of the CRU database for the same pre 1995 period. This change in data explains the results of several of our critics.


• When the post 1995 data is included in our tests the results show an even more significant increase in temperature.


• We do not find a significant increase in temperature in the Kericho data for the period up to 1995 but there is a significant trend when post 1995 data is included.


• We conclude that there is now clear evidence of increased temperatures in highland East Africa especially in the last 15 years.
• The twist is that malaria incidence has now declined. So it's still not clear if climate change was the main cause of the surge in malaria and despite recent warming malaria has radically reduced and, therefore, other factors appear to be more important than climate in malaria incidence.
  

Press release version.

Two Post-Doc Positions Available to Work with Simon Hay/MAP at Oxford

My coauthor Simon Hay is looking to hire two post-docs to join their team at Oxford for up to four years. One post-doc will work on "Defining the population at risk and burden of disease of Plasmodium vivax malaria" and the other on "Global dengue occurrence – present and future". The positions are funded by the Wellcome Trust.

Episode 3

Evolutionary biology, HIV, and the Central African World War... This one is even more out there... There is more about the series on Wikipedia.

Omumbo et al. Revisit Temperature at Kericho, Kenya

Last year I wrote a series of blogposts about the controversy over malaria and climate change in highland East Africa, and most specifically at Kericho, Kenya. Our previous research had shown that there was no signficant trend in temperature in various locations across Eastern Africa and so the increase in malaria morbidity was unlikely to be due to climate change. Our work was challenged and more recent papers claimed that there were in fact trends. We found that among other things newer versions of the CRU database did show trends in temperature. Now Omumbo et al. have analysed a cleaned up version of the data from the weather station in Kericho.



We were reluctant to use this data previously because of the apparent breaks in the series. The graph shows an apparent increase in temperature of around 3/4 of a degree over 30 years. So I think it is settled now that, at least when more recent data is included, temperature has increased at Kericho. The twist is that in the years since our 2002 publications the incidence of malaria has dramatically reduced.

Changes to CRU Database

I recently wrote about the controversy regarding climate change and malaria. I've now applied a couple of different models to both the data we used in our 2002 Nature paper and data from the latest version of the CRU database. Our Nature paper used an older version of the CRU database. One of the models I used was the basic structural model also used by Chaves and Koenraadt. The other one is a test for trends in time series that is robust irregardless of the true model underlying the data.

My estimates of the basic structural model came out nothing like those of Chaves and Koenraadt casting doubt on their results but on reflection I don't think this is a particularly useful model for addressing this question. I don't really want to get into that here but would be happy to answer questions. The trend test finds though that:

1. There is no significant increase in temperature in the data we used in our 2002 paper in line with our conclusions there.

2. But that there is a significant increase in temperature for the new CRU dataset, especially for Kericho in Kenya. It's easy to see the differences between the old and the new series.

3. Adding post 1995 data increases the effect further.

The key point is that CRU have changed their data. Chaves and Koenraadt appear at this stage to be correct that there was an increase in temperature at Kericho. But it appears that that is due to the changes in the dataset and not due to the specific modeling reasons that they cite in their paper.

Exploratory Analysis of Chaves and Koenraadt

A couple of days ago I mentioned the Chaves and Koenraadt paper on malaria. I've looked over it a bit since but would really have to repeat the analysis myself in order to come up with anything conclusive as the time series analysis is so poorly documented. To recap, here is the temperature time series from Kericho in Kenya that they use:



One of the models they fit is the basic structural model. This model consists of a stochastic trend a seasonal component and residual noise. The stochastic trend is a local linear trend model:



Beta is a simple random walk with random error term Xi. It acts as the slope of the I(2) stochastic trend Mu which also has an additional random error term Eta. According to the paper this is what Mu looks like (I think):



The trend shows that temperature increased by about 0.08 C over the period when we remove the seasonal and noise components. Unfortunately, the authors do not provide a confidence interval in the chart and so it is hard to tell if this is a significant increase or not. But they do provide estimates of the standard deviations of eta and xi, which are 0.29 and 0.12 respectively.* Both of these are larger than the entire increase in temperature shown by the trend, suggesting that the increase is insignificant.


* They actually give the variances in the paper and these are the square roots of the variance.

New Controversy on Malaria and Climate Change

For discussion of the 2011 paper in PLoS ONE see this new blog post.



Some of my coauthors on our work on malaria and climate change have an article (with others) in the latest issue of Nature. Their main point is that even if climate change has had an effect on the prevalence of malaria in the last century, that effect is swamped by everything else that has been going on. Also that the current distribution of malaria endemicity is no guide to future trends. Both these points seem pretty sensible to me but Joe Romm is outraged. He describes the authors of the paper as "sloppy" because he thinks they exaggerate the degree to which the IPCC support the "Malaria is increasing due to climate change hypothesis". This is a rather indirect criticism. He thinks the IPCC underplayed the threat and accuses them of saying the IPCC overplayed the threat. That's the best he's got against their paper... It's typical of Romm to trash a paper for extraneous reasons if it doesn't fit the global warming is always bad everywhere party line.

In passing, he notes the recent Chaves and Koenraadt paper which he quotes Science Daily as saying debunks our 2002 paper in Nature. I hadn't read it up till now, though citation tracking meant that I was aware of it. I need to do some background research before I can comment in detail on the paper. In the meantime, here is their monthly data for Kericho, Kenya:



Note that this isn't for 1966-2002 in fact despite the article claiming that that is the data they analysed. And here is the data we actually used for 1966 to 1995:



The series are not identical. In both cases the variance seems to go down a little and there does seem to be some possible increase in mean annual temperature. The question is whether it is truly statistically significant. Our answer was no, Chaves and Koenraadt suggest yes. If so, it is a very small change. People are still arguing about whether there is a significant trend in global mean temperature, where there is a much clearer trend...