You say it’s “obvious” but it’s the opposite of my intuition. Can you explain the logic a bit more?
I would have thought you’d give a risky new drug to the least healthy, most likely to die without it population, to raise the chance of survival from low to unknown.
Because you may not be raising it at all. The side effects could do more harm than the intended effects do good. So you need to do a large very controlled study with healthy people so you can more easily find side effects. Second, many side effects can be tolerated well by a healthy person but could be very dangerous in an at risk person. A drug that causes hypertension might not be a problem for a healthy person, but could kill someone with a heart condition. If you skipped this step, you could put people at risk of heart attacks and make it harder to detect the drugs effect on the heart vs already increased morbidity in those heart condition patients.
You are also evaluating the intended effects (as opposed to side effects), so you can see if it is actually doing what we want it to do, and how well. We could find out that the vaccine is very effective at preventing covid infections. But we could find out that it does not reduce the rate of covid infections, but does reduce the severity, or we could find out it does not reduce anything for reasons we don’t yet understand.
Once that is done, we will have a good understand of A. the level of benefit and B. the level of harm. From there we can make an educated decision to proceed or not. If we do, then it is approved and given to the general population, including those at risk. It is at this stage that the effects on at risk groups can be studied, during the ongoing monitoring of the approved drug.
Hopefully this answers your question and shows why this is the best way to do it.
There is always a balance of risks, and this is why there are ethics review board.
My intuition on this has become that the more certainty there is that you are to die without treatment, the easier it is to try relatively risky new treatments on you (with your consent). But say having stage IV pancreatic cancer is very different than having a comorbidity for a disease you haven't contracted yet.
Additionally with something like a vaccine, you are very interested in what it will do to the general population, not just the most vulnerable, because it should be rolled out to the general population. Anything screening program, you have to pay attention to the unlikely side effects since the denominator will be large.
Finally, you do these trials in phases because you need to find out if it is safe for humans first, then if it is effective.
Even among the 65+ survival is liable to be 90% even if infected and stand a good chance of avoiding the virus entirely even if they must wait a a few more months. Thus the probability of survival waiting isn't low to start with.
The proper calculation would be probability of infection * probability of death in the intervening months before we collect data on the efficacy of various options vs expected outcome if they just take the oxford vaccine now.
The users chance of survival is already quite high in the intervening period and while the benefits of the oxford vaccine are highly hypothetical they are also largely contingent on vaccinating everyone. People are liable to only see IF it works a 30% - 60% effectiveness, lower for some with worse immune systems. The real benefit is decreasing the spread of the virus until it dies out not merely the individuals benefit which will never happen if it turns out it doesn't work.
Worse it may be most apt to kill or harm those who are also most vulnerable to covid.
At worst the patient trades a very good chance of survival for a substantial downside with no upside and a future vaccination with a different vaccine they ought to have waited for.
There are a few factors -- one is that the major goal is to measure the side effects and safety of a particular new drug, in relative isolation... someone who is already sick is more likely to be on other drugs, thereby greatly increasing the risk of drug interactions.
Another is that when someone is already sick, they are more likely to not do well with any particular new drug (pretty much nothing is 100% effective for 100% of people), so you'd have to have some way to trying to discern whether the fact they got worse or even died is related to the new drug vs. just a natural progression. This can get complicated to measure.
And specifically for this particular thread, this is a vaccine, not a cure -- so in any case, when we say "least healthy" in this context, we're not talking about giving it to people who are already diagnosed with the virus, but rather to those who are _at most risk of complications_ from the virus, _if they were to contact it_. So it would be giving a drug to people who _don't_ immediately "need" it to survive, but who are still most at risk from any side effects.
You don't want people to drop dead from your experimental vaccine . As a general rule, the people least likely to drop dead from your experimental vaccine are young, fit, healthy people. Start with them. If they drop dead, well shoot. If they nearly drop dead, lucky you didn't start with the people who are one bad day away from the grave already.
In addition to what others said, if you know or suspect the procedure is extremely risky, and no other good options exist, you _do_ start with the least healthy.
As an example, the patient who got the first heart transplant:
“As a result of heart attacks in 1965, approximately only one third of his heart was still functioning. In late October, he went into a diabetic coma, but regained consciousness. […] He was also suffering from kidney and liver failure.”
It’s a vaccine, not a cure once you have it. So the people you would be testing it on wouldn’t be infected yet. Therefore you don’t first test it on unhealthy people when the vaccine might have unrelated nasty side effects.
You have a built in assumption in your statement that you haven't acknowledged: that the drug helps rather than harms. A new drug may help or it may harm.
If the drug turns out to harm people, a healthy person may be able to deal with it and recover. While a person that is barely hanging on to life may be pushed over the edge. There's no data to support a new drug raises the survival rate. It may in fact lower it from low to zero. During testing you generally want to kill as few people as possible.
Once more data has been collected on the drug, then the balance starts to shift toward giving it to at risk groups.
If you did start experiments with those most vulnerable you exposing them to far more risks.
Drug trials always start from learning more about safety of treatment (as opposed to efficacy) and once deemed statistically safe, you move to test for efficacy.
I assume you're thinking about treatments who could give a chance to those who have little hope/time for surviving. There's probably a chance they can benefit from taking experimental drugs but if we know little about it (because they are early in trial phases) we less confident this treatment will be useful and worst, we maybe not learn much from why it failed
Presumably the risks of an unknown response outweigh an as yet unknown benefit when testing is starting. These risks would be amplified for those with already compromised systems.
This is preventative medecine, as far as I know, most vaccines cannot be given to someone who already has the disease.
So it makes sense to target healthy people first, to test out the vaccine side effects, make sure there aren't many or bad ones, and then test out if the vaccine works. And then once you know it has low side effects and works, you can risk testing it on more compromised individuals.
One of the goals of pre trials is to establish that negative effects, if there are any, aren't caused by other factors other than the drug itself.
It's already hard to prove it on healthy candidates, it's pretty much impossible on candidates whose immune system is already weakened by previous conditions.
> I would have thought you’d give a risky new drug to the least healthy
Yes, but at this stage there is no drug. You're in the process of making the drug. And you don't want to test your beta versions on sensitive subjects.
Once nothing obviously bad happens, you can expand the test circle.
The high-risk population introduces statistical complications evaluating the results - were bad outcomes caused by the treatment under investigation or by the preexisting conditions?
