Witches, Prince George of Cambridge, other babies and the full moon

Posted on February 5, 2015


This blog took a little bit longer to sort, because it actually involved some data preparation and statistical analyses (plus Christmas, new year and my birthday), but I hope it is worth the wait.

stork_w_baby_4401Something I got, for reasons I cannot remember, interested in – out of curiosity rather than professional reasons – is whether it is busiest at maternity wards during a full moon or not. As I said, I cannot for the life of me remember how I got into that discussion. It may have been, as it very often is, Daily Mail-infused since they discussed a while back whether the full moon sent Kate into labour. More specially, based on the arrival date of his royal highness only, they conclude that “Old Wives’ Tale proven in the timing of Royal baby delivery” (article). So if it is now proven, you ask, why do we need to look into this a bit deeper? I know, waste of time, but why not…

The Daily Mail even describes the mechanism by which this would work, or in other words, they accidentally generated a hypothesis! Apparently:

“…the moon’s gravitational pull can influence embryonic fluid around a child in the womb in much the same way as it affects the tide.”
Christmas Photographs Of Prince George
They didn’t come up with that themselves, others have posed this before them, and according to someone from the Royal College of Midwives Midwives usually do say “I’m on call. It’s a full moon. I’ll be busy tonight”.” She referred to a study about 20 years ago at a hospital near the Thames, which is a tidal river, which showed there were more births during the full moon. I am not sure what the proximity to the Thames has to do with anything, but the Daily Mail couldn’t be bothered finding the actual study and neither will I. However, expanding on the hypothesis it was explained that ‘The idea is that because the baby is surrounded by water, the time of the full moon and the high tide causes the waters to break.’

But being good scientists, the Daily Mail also explored alternative explanations, and points out that “Some believe an electrical storm above Kensington Palace may also have had an effect, with thunder and lightning and isolated showers in West London at 6am yesterday. “ This, according to a midwife of 20 years, this is because There’s a pressure around the baby in the uterus. There’s change from a high atmosphere to a suddenly low one when a storm comes and the difference in pressure is what the uterus is trying to equal, so your waters can break.” I’ll leave the alternative hypothesis for you to ponder about, and will focus for now on the full moon as the cause (I may come back to the storms if I stumble on the right data for that). Aside from the various websites that discuss whether the full moon has an effect on births (howstuffworks, babycenter, thebump, babble, just to name a few) a quick PubMed search results in about 15 relevant papers, so work has been done on it. The outcomes are, let’s say, mixed. Although an early study of about 6 million French births showed a correlation (link), the most recent two papers (2012 and 2013) looked at data from India and from Germany and neither found a significant increase in the number of births during full moon (link1, link2). A couple of years earlier there was a letter to the editor in the American Journal of Obstetrics & Gynecology by a Dr Schaffir (link), who seems to be mildly irritated about the fact that, after “at least 6 decades’ worth of studies” this is still a topic for research. It seems, based on this letter that it is because nurses, but not the patients themselves, believe in this correlation; and a likely explanation of this is that “like so many mythodologies [it] is passed on through oral tradition to an audience eager to impose order on a system that in fact operates randomly”.

So that’s that than. The Daily Mail may have been wrong!

But is it……..


We are in a good position to have a closer look at this. For one of my studies I used a database of births in the northwest of the UK between 2004 and 2009. So for this study I added them all up for each day of the year, and linked that to data on the lunar cycle for that same period. It is quite straightforward to do this, using the website www.timeanddate.com and enables linkage to the daily phases of the moon, the altitude of the moon in the sky, and the distance of the moon to the earth (in km) and the percentage illumination. To make it more scientifically correct, it provides these numbers specifically for Manchester as well; the centroid of the catchment area of the births. So in summary, we have a database of daily number of births (a total of 259,275, and a daily average of 145) and can now link this to the various characteristics of the moon for a period of 1826 days (or 5 years).

If we plot the daily number of births (the black bars), the distance of the moon to the earth on each day (blue) and the illumination of the moon on each day (red) in a graph this is what you get:


 That’s a bit unclear so here is the same figure for the first 100 days, just to get an idea of the data:


So now that we have an idea of the data, let’s see if it is associated to either the distance of the moon to the Manchester or, although this is not part of the original hypothesis, illumination of the moon and its altitude.

We just run a normal linear model with the number of births as the dependent variable and either distance, altitude or illumination as the explanatory variable (for the purists amongst us, the number of births are counts, so a Poisson or negative binomial model should be used. I did this and it doesn’t make any difference to the results in this case. These numbers are just easier to interpret….).

For illumination we get a positive estimate, so the more illuminated the moon is the more births occur, but the P-value is 0.51. For altitude we find a negative estimate, so the higher in the sky the moon is the lower number of births, but the P-value is 0.72. In other words, both have absolutely no effect. Distance of the moon to Manchester, the focus of our hypothesis, also has no effect (P=0.45). The estimate itself is negative though, which is what we would have hypothesized.

So in other words, there is nothing there. That’s a bummer.

So let’s have a look at whether there is a full moon or not. Maybe there is something else with this full moon that causes a higher likelihood of birth; something not just related to the distance of the moon. As it turns out, when there is a full moon, on average there are 3-4 more births on that day compared to other days. The P-value for this is 0.11, so although it is close, with so many days in our dataset this kind of indicates there is probably nothing serious there. Besides, the actual effect is pretty small so no need to get too excited over this.

There is however, something called a “Supermoon” or “Super Full Moon”. I did not know this, but a supermoon happens when a full moon occurs at the same time as the Moon’s closest approach to earth, and when this happens the moon may seem bigger and brighter (link). Note that the website helpfully points out that the moon itself did not actually change size….that’s good to know.

Anyway, that seems to fit with our hypothesis, so let’s have a look at this. During our time period of 1826 days there were 62 full moons of which only 8 were supermoons. That’s not a lot, but let’s split those out. When we run that model we find that when there is a normal full moon on average there are only 2 more births that day (P=0.43), but during a super full moon there are on average 15 more births! The P-value for this is 0.02, so if you are a sucker for P<0.05 this is statistically significant too!

So there you go. The old wives were right after all! Well kind of…only during these special days of a supermoon it seems. It’s only based on 8 events in 5 years, so not very robust…but it seems we may have just improved on the hypothesis, and we may have even explained why previous studies did not find anything.

Is this useful? Well it is more interesting than useful I guess, but at the same time midwifery departments now only have to get additional staff in during 8 days rather than 62 days in a 5-year period. So if someone who controls NHS budgets, say Jeremy Hunt, happens to read this – I just saved you loads of money! Maybe you could use this to get a couple of extra nurses in, or buy some extra beds…


1182175500_8474-3-the-wicker-manAs an interesting side note, apparently “This supermoon coincides with true Beltane, the Spring fertility festival, when traditionally Witches, Wiccans, Druids and Pagans of all kind celebrate the season with joyous rites under the Moon and stars.” I could not work out whether this is any supermoon, just the ones in spring, or just one particular one, but “joyous rites” sounds good. Also, having linked the supermoon to the fertility festival seems particularly appropriate; it seems they were right after all, those druids…

Posted in: General