I've had a policy of not consuming things to change how my brain works for most of my life, but perhaps I should relax it? Specifically, I'm worried about driving while I'm sleepy. Playing for contra dances can mean getting home between 11:30 and 2am, which is a lot of drowsy driving. I normally try to stay alert with cold air, talking, and occasional breaks to run around, plus making sure multiple people are always awake at once, but perhaps I should be considering drugs?

The most common drug people take for alertness is caffeine, which also has the advantage of being legal. But does it actually improve driving performance or just make you feel more alert?

It turns out this is a question that lots of people care about, which means there are studies. Some looking turns up The Effects of Coffee and Napping on Nighttime Highway Driving: A Randomized Trial (2006, pdf requires copy-pasting the url). The study seems reasonably well done: they had people drive both during the day and late at night, while on coffee, decaf, or a rest-area nap. Issues:

• The participants were all young men and weren't consuming other stimulants (no regular caffeine consumption) during the study. This weakens the study for most people, but happens to be my situation.
• The nap condition was a little unreasonable, in that they added an electric heating system to the car so it could be a constant 66 degrees, and the people had blankets. But I'm not as interested in napping.
• The numbers for caffeine content are strange. They say they gave people 125mL (half a cup) with 200mg of caffeine (15mg in the decaf case). That's about four times the caffeine normally in a half-cup. I'm going to go by their 200mg number.
• The main flaw is the size: only 12 people, each tested in each condition.

They conclude that "drinking coffee or napping at night statistically significantly reduces driving impairment without altering subsequent sleep," but the core of their data is small enough that we can just look at it. Here are the number of times a driver drifted out of their lane, per driving session, by participant:

	Daytime	Coffee (half cup, 200mg Caffeine)	Decaf (half cup, 15mg Caffeine)	Nap (1/2 hr)
	0	1	11	6
	1	3	2	1
	0	0	3	3
	1	0	13	6
	0	1	2	1
	0	0	2	0
	0	13	17	8
	0	2	0	0
	0	0	11	0
	0	0	11	0
	0	0	1	0
	0	0	0	0
total	2	20	73	25

The daytime driving is clearly the best, but driving on coffee seems much closer to daytime than to the placebo of decaffinated coffee. It does sound like I would drive better late at night if I were to take caffeine, and a crash would be bad enough that this should overcome my resistance to taking drugs.

So, how should I do this? While people usually use coffee I don't like the idea of getting caffeine through something I enjoy because I might consume it when not necessary. It looks like pills are somewhat unpleasant while also being really cheap (~$0.05 per 200mg pill). Caffeine also takes some time to reach full effect, so I should probably take one before we start cleaning up so that it's active about when I start driving.

Is there anything I'm not thinking about? A reason that the benefits don't outweigh the potential harms in this case? I'm a little worried about risk compensation but suspect it's not big enough to outweigh the gains from less sleepy driving.

Update 2013-11-24: Last night I tried this for the first time. At 11:15pm I took a 200mg caffeine pill. Driving home, from 11:35 to 12:35, I think I might have been less sleepy, but I was still sleepy. It's hard to tell; I should get some fake pills so I can do a blind test. After unloading sound gear and getting into bed around 1pm, however, I definitely noticed an effect. I would normally fall asleep almost immediately, within just a minute or so, but instead it took me about half an hour. I noticed my mind jumping around all over the place and had multiple tunes stuck in my head at once. Unlike normally when I can't fall asleep, my body felt very still and restful, falling into a single static position, just my mind felt awake.

Update 2014-02-13: I've now tried caffeine about three more times. After my first try, I made sure to take it earlier in the evening, targeting 4hr before I wanted to fall asleep. It's still a little harder to fall asleep, but not that bad. Nowhere near as hard as the first time.