I was thinking about how sauna is supposed to reduce all cause morality, according to some random paper.
I asked GPTSearch to find the paper and summarize the data.
We found:
(I didn’t read the paper, the numbers below could be hallucinated, I’m just having fun this morning)
Summarized by GPTSearch:
Data Summary:
- Frequency of Sauna Use:
- 1 session per week
- 2-3 sessions per week
- 4-7 sessions per week
The study found that compared to men who had one sauna session per week, those who had 2-3 sessions per week had a 24% lower risk of all-cause mortality, and those with 4-7 sessions per week had a 40% lower risk.
Duration of Sauna Sessions:
- Less than 11 minutes
- 11-19 minutes
- More than 19 minutes
Regarding session duration, the study reported that compared to sessions lasting less than 11 minutes, those lasting 11-19 minutes did not show a significant reduction in all-cause mortality. However, sessions longer than 19 minutes were associated with a 52% lower risk of sudden cardiac death.
Good.
I then asked GPT4o to create some Python and plot the dose-response curves for the two datasets:
Nice, now, let’s create a model and see what’s up.
Model of Sauna Sessions Per Week
First, let’s look at dose-response for sessions per week.
Table:
| Sauna Sessions per Week | Hazard Ratio | Relative Decrease in Mortality (%) |
|---|---|---|
| 1 | 1.000 | 0.00% |
| 2 | 0.820 | 17.97% |
| 3 | 0.714 | 28.64% |
| 4 | 0.650 | 34.98% |
| 5 | 0.613 | 38.74% |
| 6 | 0.590 | 41.00% |
| 7 | 0.575 | 42.49% |
| 8 | 0.564 | 43.59% |
| 9 | 0.557 | 44.28% |
| 10 | 0.552 | 44.78% |
Not that exciting, 5 sessions seems to be a point of diminishing returns, at least from this data.
Model of Sauna Minutes Per Session
Now, let’s look at minutes per session.
It first tried to fit a quadratic, which was really wrong.
It then used a log-fit. Better I guess, but not awesome. I’m just hacking around and this is fine for now.
Table:
| Minutes in Sauna | Hazard Ratio | Relative Decrease in Mortality (%) |
|---|---|---|
| 1 | 1.065 | 0.00% |
| 2 | 1.060 | 0.52% |
| 3 | 1.054 | 1.06% |
| 4 | 1.048 | 1.64% |
| 5 | 1.041 | 2.25% |
| 6 | 1.034 | 2.91% |
| 7 | 1.026 | 3.63% |
| 8 | 1.017 | 4.39% |
| 9 | 1.008 | 5.21% |
| 10 | 0.999 | 6.00% |
| 11 | 0.989 | 6.84% |
| 12 | 0.978 | 7.74% |
| 13 | 0.967 | 8.69% |
| 14 | 0.955 | 9.69% |
| 15 | 0.942 | 10.65% |
| 16 | 0.929 | 11.64% |
| 17 | 0.915 | 12.71% |
| 18 | 0.900 | 13.84% |
| 19 | 0.885 | 15.06% |
| 20 | 0.869 | 16.35% |
| 21 | 0.852 | 17.72% |
| 22 | 0.835 | 19.17% |
| 23 | 0.817 | 20.70% |
| 24 | 0.798 | 22.32% |
| 25 | 0.779 | 24.03% |
| 26 | 0.759 | 25.82% |
| 27 | 0.738 | 27.69% |
| 28 | 0.717 | 29.65% |
| 29 | 0.695 | 31.70% |
| 30 | 0.672 | 33.83% |
If you stay in the sauna for a long time you will live forever.
Na, 20 mins is fine I guess.
What we need is better data! Controlled experiments, put humans in cages for their entire lives, randomize everything else and dose with specific sauna sessions/minutes. Super easy, right :|
It’s probably bunk.
A sauna is probably a concentrated collection of people that follow rules.
They are able to turn up 5 days a week and sit somewhere super-uncomfortable for 20+ minutes.
They probably also exercise, eat right, and do all the other things people tell them to do to be healthy.
Selection bias. Healthy user bias. And on.
Sauna might do nothing other than collect these people.
I guess I’m one of them :)