But it's incredibly more likely than if left to themselves, people will self-medicate, and not necessarily with due research beforehand.

To give an example: in my country, 56% of people have already medicated with homeopathy (a known scam for decades). Imagine if they could just buy any other medication, which unlike homeopathy, could have strong secondary effects ? Those 56% people who have "illnesses" with very benign symptoms that for most cases cures itself in a couple days without any drugs (homeopathy is used for anything, common cold, headaches, itches, you name it ...), now they're potentially taking much more dangerous meds. So it's clearly not equivalent as 56% of people definitely don't have "faint-at-the-wheel" symptoms today.

The existence of side-effects affects how many people will take the drug. I don’t think the zero side-effect situation illustrates anything any more than giving away your product for free tells you how many people will buy it.