A few months back we’d seen this documentary on Netflix (I THINK) on the effects of alcohol on health. Like you would expect from a well-made documentary (rather than a polemic), the results were inconclusive. There were a few mildly positive effects, some negative effects, some indicators on how alcohol can harm your health, etc.
However, the one thing I remember from that documentary is about alcohol’s effect on sleep – that drinking makes you sleep worse (contrary to popular imagination where you can easily pass out if you drink a lot). And I have now managed to validate that for myself using data.
The more perceptive of you might know that I log my life. I have a spreadsheet where every day I record some vital statistics (sleep and meal times, anxiety, quality of work, etc. etc.). For the last three months I’ve also had an Apple Watch, which makes its own recordings of its vital statistics.
Until this morning these two data sets had been disjoint – until I noticed an interesting pattern in my average sleeping heart rate. And then I decided to join them and do some analysis. A time series to start:
And then sometimes a time series doesn’t do justice to patterns – absent the three recent big spikes it’s hard to see from this graph if alcohol has an impact on sleep heart rate. This is where a boxplot can help.
That said, like everything else in the world, it is not binary. Go back to the time series and see – I’ve had alcohol fairly often in this time period but my heart rate hasn’t spiked as much on all days. This is where quantity of alcohol comes in.
On 26th, a few colleagues had come home. We all drank Talisker. I had far more than I normally have. And so my heart rate spiked (79). And then on June 1st, I took my team out to Arbor. Pretty much for the first time in 2022 I was drinking beer. I drank a fair bit. 84.
And then on Saturday I went for a colleague’s birthday party. There were only cocktails. I drank lots of rum and coke (I almost never drink rum). 89.
My usual drinking, if you see, doesn’t impact my health that much. But big drinking is big problem, especially if it’s a kind of alcohol I don’t normally drink.
Now, in the interest of experimentation, one of these days I need to have lots of wine and see how I sleep!
PS: FWIW Sleeping heart rate is uncorrelated with how much coffee I have
PS2: Another time I wrote about alcohol
PS3: Maybe in my daily log I need to convert the alcohol column from binary to numeric (and record the number of units of alcohol I drink)