Hacker News Reader

Comment by orangecat | Hacker News Reader

It seems to be an underappreciated point that the only alternative to challenge trials is to wait long enough for a sufficient proportion of the vaccine and control groups to be "naturally" exposed to the virus. You can either intentionally expose volunteers to the virus with X% probability, or send them back into the world and wait for X% of them to be infected. You aren't preventing infections by forbidding challenge trials; all you're doing is allowing millions more people to become infected while you wait for results.

View on HN

You're assuming that pulling or not pulling the lever is morally irrelevant, which may not be the case. To take it to the extreme, it's the difference between killing three people or waiting for three people to die so you can study them.

In the former case, you're definitely causing excess infections (some of which might die, directly due to you).

What if it's killing 3 people, vs waiting for 3,000 to die?

What if it's having a small but non-zero probability of killing 1 person, who understands and accepts the risk, vs waiting for thousands to die who really don't want to?

Indeed, and apparently ethics boards find it preferable to wait for hundreds of thousands to die rather than kill one, but it's not entirely unreasonable.

Also, I think the number of total deaths in both branches are about the same, as they're only waiting for 100ish people to get infected on either arm.

It was entirely unreasonable, though.

Also, I think the number of total deaths in both branches are about the same, as they're only waiting for 100ish people to get infected on either arm.

When you (entirely unreasonably) restrict your analysis only to the experiment subjects, it is true. However, the vaccine trials cannot be analyzed in a vacuum: the whole reason why these are even a thing is the raging pandemic that kills thousands of people every single day. To put it in a trolley analogy, the decision is not between high speed and low speed trolley that will end up rolling over the same number of people. Instead, in the other branch, in addition to one slow speed trolley, there are also thousands of extra trolleys that you just aren’t explicitly aware of, but they will roll over their victims nonetheless.

That seems to be torturing the trolly metaphor quite a bit when the original works perfectly fine (actually better) for this situation.

There is more than that though. Sure with a challenge trial you know that the vaccine works much faster. However you still need to vaccinate tens of thousands and then watch them for a few months to ensure that nothing else bad happens. At least one covid vaccine trails has already failed because the vaccine wasn't safe (in the case I know of because you show up HIV+ even months latter which makes those tests worthless - the vaccine might have worked against Covid for all I know). There have been a lot of other vaccines trials over the years that have failed for safety reasons.

> At least one covid vaccine trails has already failed because the vaccine wasn't safe (in the case I know of because you show up HIV+ even months latter which makes those tests worthless - the vaccine might have worked against Covid for all I know).

That may be a good enough to nix development for common cold vaccine, but how can one argue that it’s serious enough problem to cancel vaccine for disease that kills thousands of people every single day? This is another example of how “safety” and “ethics” regulations have little to do, and in fact are counterproductive to achieve safety and ethics in real world.

It was clear at the time that other vaccine candidates were working at least as well: Pfizer/Bio-NTech got its EUA the same week AND didn’t mess up HIV tests. Seems like a pretty straightforward decision to me.

https://www.nature.com/articles/d41586-020-03570-0

What if the 3000 don't die though. At any time covid could hit some sort of natural limit and stop spreading (as happened to SARS before it). At anytime we could discover a miracle cure that means Covid isn't a problem (as antibiotics did for infections). Since you don't know about those unknowns it is hard to say in advance if you killed 3 so 2997 could live, or killed 3 people for nothing.

But it's not you that's killing the test subjects; they're doing it to themselves. (Well, it depends on the experimental setup, but it can be easily arranged that way: you could leave a spray bottle on the table and tell them "This contains the coronavirus. If you spray your face with it, you will help our experiment, but risk death. You are also free to just go home and face no consequences.") I don't understand the argument for moral relevance.

Edited: "This contains the coronavirus. If you spray your face with it, we will give you money enough to pay a month or two's rent and living expenses, but you risk your death. You are also free to just leave."

Taking a top-level view though, we've pulled about a billion different levers.

In the UK we've restricted the activity of healthy individuals for about 11 months now - the latest episode being an approximately 3 month period within which it's illegal to meet anyone else within your own home.

Psychological trials, even trivial things like filling in surveys or taking reaction speed tests are carefully monitored for ethical concerns, lockdown we just went all in on with no cost-benefit analysis.

It seems bizarre to me that in that context we should be worrying about the ethics of a challenge study with a few hundred or few thousand volunteers.

Loads of people I know have _already_ volunteered themselves for it in the outside world with no compensation and with no benefit to science because they're not able to mentally handle indefinite restrictions.

This is being reported as if the challenge trial is giving people already-approved vaccines (also note the first stage of the trial is not even doing that step).

As such, the benefit of deliberately infecting people over waiting for natural exposure appears to me to be zero.

Interesting point of view, but where would you set the threshold: a challenge trial for a cancer drug by injecting a carcinogen to cause tumor growth? What about viral diseases that are much more devastating, e.g. a challenge study for an ebola vaccine?

There are no ethical issues when the subjects are fully willing volunteers who have had the risks properly explained and emphasized to them. As some have mentioned, if people can volunteer to risk their own life for the army, then why not for this? (The experiment must, of course, fully quarantine the volunteers who are given an infectious disease (cancer is not infectious, ebola is) until they're recovered (or dead); otherwise they'd be risking more than just the volunteers' lives.)

There are ethical issues. You’re just not aware of them.

Do you compensate the participants for their time? If not, can you find enough people to run the trial? If you do compensate them are they truly consenting freely? Or do they just need the money?

Consent is a very complicated issue and you’re hand waving all the nuances away.

I do think it's a silly argument that paying someone to do something means they're not freely consenting to do the thing. As long as it's not your fault that they're in a bad economic situation, all you're doing is offering them an option that they, by accepting it, believe to be an improvement over their current situation. People who don't like sweatshops should be removing all barriers to the setting up of more sweatshops, so that they will compete for workers and eventually have to improve working conditions or raise wages.

However, in this case, I suspect offering a monetary incentive wouldn't be necessary. Everyone knows about the terrible disease; just make a patriotic appeal and I expect you'd have plenty of unpaid volunteers.

Accounting for financial aspects is not silly. People who are desperately poor will do desperate things for money that they would consent to otherwise. Taking advantage of that is unethical.

And I’m not sure if you’d have a ton of unpaid volunteers or at least enough to power a trial. It’s not a simple “lets get a shot” it’s having to travel to a clinic several times over a period of time. It’s a serious commitment of time, that’s why trials are normally compensated.

Otherwise your entire pool of patients are just upper class mid-white collar career who can work from home and skip work in the middle of the day.

> Accounting for financial aspects is not silly. People who are desperately poor will do desperate things for money that they would consent to otherwise. Taking advantage of that is unethical.

Why? The desperately poor person seems to think the money is worth it. Who are you to tell them it's bad for them to take it? The very stigma against "exploiting" such a person makes it less likely that other people will make them less-bad offers.

> And I’m not sure if you’d have a ton of unpaid volunteers or at least enough to power a trial.

In this thread, we have someone saying "I have signed up" and someone replying "thanks for taking a risk to help out the human race". It is patriotic, it is (or can easily be made) prestigious. (To make it visible, give out a hat, a button, a jacket, whatever.) I expect some guys would do it because it seems exciting or meaningful, others because they think it'll help them impress the ladies.

When, specifically (i.e., as of what date?), do you think the risks of COVID were well-understood enough to be properly explained and emphasizes?

Well, to be responsible, all you have to do is not understate the risks. In fact, it's probably best to err on the side of making it seem scary. If you said "N people have died, M people are known to have gotten it and recovered, so you should not assume your survival chances are better than M / (N+M)", with N being at least, say, 100 for sample-sizing purposes, I think that would be reasonable—so one could have said that probably back in February 2020. That would have yielded overestimates—probably 1-10% IIRC—but you could probably find some people willing to risk even those odds for the sake of their country or the world (or bragging rights).

I think an obvious first-pass at a threshold would be a global pandemic with a clearly established non-trivial mortality rate and in which it is clear to epidemiologists that -- unless a vaccine is developed -- essentially everyone is going to get the virus at some point.

It's much more murky territory for more containable viruses such as ebola. But a deadly pandemic that spreads across the globe in weeks? Seems like it would be hard to argue against challenge trials with informed consent.

I think the strongest argument against it is that setting any such bar only makes it possible to lower the bar in the future.

Reply on news.ycombinator.com