I get V=ir, but I'm not seeing any good answers to the question of non-linearity of humans' resistance.

I will note that 0.2A is enough to stop your heart, and 2A will almost definitely kill you (with enough damage that you're probably not coming back).

Consequently I'm not sure how a 15A (USA) circuit could be considered safe even in the relative sense.

Voltage determines the current that actually flows during a shock. The amperage ratings you are citing are the maximum amperage that can be drawn before the breaker pops. This protects from fire, it does nothing to prevent death from shocks, in any country. You will be burnt to a crisp before you trip the current-limiting breaker, in any country.

High voltages are dangerous because they can overcome the resistance of human skin, which then enables current to flow.

Right, so assuming everyone's got dry hands, or comparably dry hands, then you'd expect to see lower fatality rates in the USA (and other countries that use ~110V) compared to countries that use 230-240V.

Does that hypothesis actually align with the data?

I'm struggling to find good figures that spell this out on a region by region basis.

The data are often muddied, f.e. by including lightning strikes, ladders or trucks hitting overhead high-voltage power lines, self-harm / suicide attempts, etc.

For example, some Australian data for a 2-year period [1]

  > 55 deaths involving electrical injury - 50 male, 5 female
  > 8 were lightning strikes
  > 17 were intentional self-harm
  > 18 were were due to exposure to other specified electric current

Either way that's about 26 / year out of a population of ~25M ~= 1 per million.

In the USA I'm finding:

  > Electrocution is the fifth leading cause of occupational injury death
  > 41% of cases involved overhead power lines

Alternatively [3] for USA:

  > Electrocution is sixth among causes of workplace deaths
  > Electrical hazards cause more than 300 deaths

300 out of a population of ~330M ~= 1 per million

[1] https://www.aihw.gov.au/reports/injury/electrical-injuries-d...

[2] https://injuryprevention.bmj.com/content/8/4/306

[3] https://www.electrocuted.com/safety/statistics/

Addenda - my numbers are wrong.

I was including lightning strikes in the AU numbers, but the US numbers were electrical (lines, domestic, construction, etc) only.

So fatality ratios skew higher in the USA than AU, at least. The numbers are quite low in both cases, so the ostensible safety benefits of 110V at 15-20A may not be terribly compelling in practice.