Did he though? He even mentions that customers generally pick whatever (because they don't really care what's the perfectly optimized headache pill they should take) and doctors thought it was neat, but no one really wanted to pay for it. Their solution (e.g. prescribing what they would typically prescribe) was good enough.
It doesn't really sound like there's a problem there.
Tom Blomfield worte about how during YC S11 they had trouble growing at all. He chanced upon a customer who fit their product's user persona to a tee but at the end of a 20m conversation Tom realised the it wasn't really a burning problem for them at all.
And so they pivoted...to GoCardless.
I don't care if the medicine is perfectly optimal in terms of standard deviations above placebo (which I don't trust, since the data is not likely to be reliably comparable); I care if it sufficiently solves my issue with least hassle (side effects/access/cost/works for me personally). Placebo might just be fine, in some cases.
If I am not able to identify what medicine to take, I ask my doctor, who has years of training for that, already knows what Google sources are good, and can look it up in a reference manual.
No, he didn't have a problem.
He proposes that OTC is the target. The reality is, the problem is not very deep here. Without access to prescription medications, you basically have a handful of options that are more of less the same. It's like walking onto a rental car lot. You need to know what category you need, but everything in that category is pretty much the same.
He isn't solving anything for the physician market. His recommendations are very naive and lack the clinical backing to support actual usage. A physician isn't going to buck their go-to's because a random tool suggests something else. This tool needs to provide scientific backing for it's recommendations and why that recommendation is better.
I go to my local grocery store and look at the painkillers they have. There are generic aspirin and Tylenol available for a price almost anyone can afford. There are branded versions which are still reasonably priced. There are reasonable variations such as with caffeine or with decongestant, again priced perfectly fairly. If you want to pay $2 extra for awesome marketing you can. In some stores you might find a 'natural' remedy - essentially a placebo, again priced ok.
Everything on sale is basically safe. Nothing has a business model that relies on addiction. Nothing costs hundreds of dollars. Nothing is adulterated or counterfeit. It's easy to get information about everything available. The vast majority people would be able to choose something to match their own needs (for example, they might be allergic to aspirin, or need to avoid Tylenol because of a kidney condition).
If you bought the worst painkiller on sale in the store, it would be fairly effective and not too expensive. The experience in most stores in large parts of the world would be very similar.
Also, $40k is downright cheap compared to most failed projects to improve medicine. If it were someone else's money, that is.
Also I don't think $40K is cheap for an MVP. He could have used static HTML mock-ups, or even a few Power Point slides to see if his potential customers would have paid for that service.
I've made this very mistake myself but I was lucky enough to have enough runway to start over and talk to customers first then pivot the product to something that they actually need.
I call this the "I have an idea for a startup!" issue. You hear it all the time from family/friends. Where they tell you this great idea for a product they had. This is the wrong approach. What you want to say is something like "There is this really interesting problem that everyone in ecommerce is facing right now"