The authors argue for using better predictive models to screen drugs at an early stage, because too many drugs are failing at costly clinical trials.
In my opinion we already know the best predictive model to use, but we choose not to use it. When Jenner developed the first vaccine back around 1800 he didn't use any petri dishes or mathematical models: he just found some nearby children and tested his theory directly.
Now you may find this ethically repulsive but from a consequentialist point of view it was the right thing to do at the time. This line of research saved hundreds of millions of lives.
I say we have gone too far in the other direction. The problem with failed costly clinical trials is not the failure per se but the cost: in money spent, time, reputation, commercial return, legal consequences, etc. And the hidden cost of how long the whole process takes even if successful.
Take the field of monoclonal antibodies. We now have almost off-the-shelf technology to develop drugs that specifically block certain biological mechanisms. But it still takes 10-15 expensive years to progress through different lab and animal models, then healthy humans, before clinical trials on actual patients.
Instead we should reduce the cost of human trials: have a network of terminally ill volunteers to test mabs at an early stage. Fail fast to move fast. Use the money saved to offer defined compensation when things go wrong.
The drugs that we really want to find are the ones with an obvious effect on real patients. This process would find those drugs much faster, and save lives overall.
I get where you're coming from, but I think your idea moves too far in the opposite direction.
Vaccines are somewhat different as "drugs". They are typically inactivated virus or bacteria (or fragments of them), so the safety risk is somewhat muted in the sense you aren't introducing something completely new into the body. The risk is still there, but not as much as a completely new molecule.
If you were to test new drugs on terminally ill patients and the patient died, was it the drug? Or did they die form their disease? Sometimes you can, often you can't. When you do clinical trials you're trying to eliminate any many variables as possible by running the trial on a relatively homogeneous population. That's why they do phase 1 with healthy volunteers.
The other challenge is that when human testing goes wrong, it can go horribly wrong. Check out the clinical trial of the FAAH Bial compound where people died. Also, TGN1412.
If you treat a terminally ill patient and they die, then you can still learn something. Maybe there was no effect at all on the target illness, e.g. the cancer spread and grew anyway. Maybe there was a large effect but the side effects were too severe, in which case you can investigate strategies for mitigation.
TGN1412 is a good example. This drug turned out to provoke the immune system in an extreme way. What if it had been tested first on cancer patients and, even with the life-threatening side effects, it worked as a cancer immunotherapy? Then we would be able to work on dosing or mitigation. Instead, the company folded, the research was abandoned and we'll probably never know.
And further, it's a cold calculation but instead of having healthy lives ruined you would have terminally ill patients die slightly sooner.
Frankly I think this is a ridiculous, inhumane, and fundamentally ignorant point of view.
Ridiculous - The idea of testing drugs with the hope that they magically treat something completely unexpected and without any theoretical framework supporting it.
Inhumane - Terminally ill people could hardly consent en masse to such a program. I suppose there is a utopic idea here, but you might as well wish everyone were healthy forever.
Fundamentally ignorant - The previous comment explained why terminally ill patients with many different medical problems would make poor subjects for a study that ostensibly informs medicine applied to generally healthy people.
"Ridiculous": the whole point of the OP is that we don't have good predictive models for what treatments will be effective. We have hunches and theories but it is currently too costly to go through the whole process to test them all. One approach would be develop better models, but I'm suggesting another approach would be to reduce the cost of testing them in the real target.
"Inhumane": well that's the ethical question. Some countries compell vaccination because of the societal benefit. I'm not even proposing compulsion.
"Fundamentally ignorant": in the end medicine is needed to treat sick people, not healthy people. The OP makes the point that sometimes a biased, representative sample is better than an unbiased, unrepresentative sample.
Only works for treatments that go after the terminally ill, who already have massive issues.
What if it is a vaccine against cancer (such as the HPV one)?
You think a sane parent would offer their healthy child for a drug trial? You'd need the fucking Nazis to accomplish this (they did, see Mengele et al).
"ethically repulsive but from a consequentialist point of view" or just plain crazy.
In the early days, most drug discovery (while amazing for its time) covered comparatively simpler diseases with arguably less complex physiological and pharmacological models i.e. low hanging fruit. Many of these diseases (e.g. Type 2 Diabetes) are still improperly addressed to this day. But they are not as lucrative for today's pharmaceutical company as they used to be in simpler times.
Governments and insurers are too astute to pay excessively for incremental improvements in mass or lifestyle disease categories. It makes more business sense for Big Pharma to target complex, hard-to-cure diseases that affect a small population set. While the success rate for such molecules is much lower, it's also much harder for payers to apply their usual mass-oriented drug economics to them.
One of the better examples of this is the recent introduction of Hep C "curing" drugs. Sponsors are going uncontested in pricing their molecules up to $84,000 per patient year[1,2].
With 3 million patients in the US alone, that's a TAM of $252 billion per year. For one drug category.
Gerhard Domagk tested his few hundred drug candidates in live mice with bacterial infections. The recent attempts started with single bacterial proteins in a tiny dish.
This goes back to the phenotypic screening argument.
Unless your biological model captures most of the factors involved in a biological system, it's not worth much. You might get a great hit, then put it in an animal model and get nothing.
There are a ton of examples of the value of phenotypic screening in past drug discovery. The discovery of benzodiazepines (Valium, etc) was completely random. Hoffman created some new molecules and just injected them into mice. When the mice fell asleep, he followed up on it.
Of course, phenotypic screening is not a panacea. It's expensive and has problems of it's own.
>This goes back to the phenotypic screening argument.
Truth. Domagk would have missed sulfa all together if he had been using an in-vitro strep model.
Prontosil (what Donmagk discovered), and azodye connected to a sulfa molecule, must be enzymatically digested into the individual azo and sulfa molecules before it becomes biologically active.
Sometimes reduction of a problem, looses the problem all together.
Great interview for a great PLOS ONE article! Many thanks!
My take away is that this is what the so-called Systems Biology / in silico Biology ppl have been talking about for the past 10 years or so. Unfortunately the drug discovery industry has its own systematic problems as A) most major biopharma are run by MBA/lawyers/D. Pharm, (this is why M&A / tax inversion / outsourcing have higher priority than basic R&D); B) from A the basic R&D (left along Systems Biology) are on the chopping block for cost cutting; C) and it's hard for Bio Data Hacking companies to survive in the past 10 years or so. Most bioinformatic companies are either out of business or being absorbed. The Entelos the authors mentioned actually went through IPO, relisting, chapter 11 and recapitalization.
So my question would be how these road blocks could be overcome before in silico biology could contribute more to the drug discovery process for a better "compass"?
The article's conclusion would indicate that you would do use in-silico biology to 'discover' existing drugs. Then once computer methods can reliably discover, say both penicillin and tetracycline, you'd push forward.
Funding this, however, would never come from industry. You'd need to convince NIH that spending many years rediscovering old drugs is the magic leap forward we need.
As someone ignorant of the field - has this been done? What is the state of validation in computer biology?
yeah, more or less. it's called drug repositioning for 're-discovering' existing drugs for new indications. several firms are currently focusing on drug repositioning by mining fda databases such as drug adverse events data to mine possible new indications. major pharmas already use the find-new-trick-for-old-drug to find new usage (eg botax). by the way viagra is also such 're-discovered' drug as it was originally targeted for hypertension treatment with poor clinical trial results but with a surprising you-know-what 'side-effect' in some male patients.
any way, the solution might not be so straight forward for in-silico biology as several major bottlenecks still hard to resolved, such as those failed clinical trials data are never revealed and shared, and the most cash-burning process still concentrate on clinical trial stages.
In short, this paper calls for "Do more math on the front-end of the pipeline."
The core issue is that CEOs of biopharmas aren't compensated as such, so if you were to go to one with mathematical techniques that would take 3 to 6 months to run, (Combining research, bioinformatics, and diff EQ) they would look at you like you're crazy because the timeline and their incentivization is skewed toward not failing a trial.
Trials exceed CEO lifetime, and that's one core issue as to why pharma development isn't done this way today.
I think when they were referring to models they weren't necessarily referring to mathematical models, but also experimental models. They more stressed the importance of the predictive models that you are using, which in biomedical research is typically an experimental rather than mathematical/computational model.
From the takeaways portion of the article
"Judging the quality of the 'compass' (i.e., the 'validity' of experimental models) is very difficult. However, the drug industry should put its brightest and best people onto the problem of model validity"
I didn't get that at all from either the interview or the paper.
From the paper, the conclusions seem to be: Decision theory creates a logically grounded explanation as to why high predictive value experiments are being done less often. To resolve this, research needs to be done in measuring the effectiveness of drug discovery models.
Nowhere in the article do they call for more math - the final conclusion of the paper is a call for more experiments validating existing drug models.
We'd like to think that we're a modern civilization, with systems in place, and the right priorities when it comes to life saving research & development.
Think of it this way: do you cringe at the thought that 70 years ago, there were almost no antibiotics available? That if you had a serious infection that you'd probably die? That such an infection would now be trivial to cure?
That's how people will view our healthcare of today 70 years from now.
Or, we will have sure fire cures for baldness, and sexual dysfunction?
I keep thinking about this new antidepressant called Rexulti.
The patent ran out on Abilify. It went to generic status. Drug company panicked.
They changed the Abilify molecule a tiny bit, and a did <400 patient study. The patients were evaluated over a two month period, using a 10 question happy test. From mean change from baseline, -5.2 people got better by Placebo. -8.4 got better from drug.
In big bold letters, it screamed to doctors, "62 Percent Greater". In small print they said vs. placebo.
Does it work? I don't know, and I doubt they know. Do they have a new, dubious drug they can con doctors/patients into buying--yes!
And yes--if we knew the actually efficacy of a lot of drugs, we would be astonished.
I sincerely hope we have better medications in the future.
Are you arguing that sexual dysfunction is not a worthy condition to treat? I would argue it has severe psychological ramifications and a negative impact on people's lives. As much as dying from cancer? No. But not something you can ignore.
We've made tremendous strides over the past two decades in treating severe diseases. Look at drugs like Imbruvica, Harvoni, Kadcyla. Women with stage 4 breast cancer can now expect to live 3 to 5 years with their disease vs. dying within a year. Tremendous progress. Just because "lesser" drugs are being developed doesn't change that.
And you'll have to explain Rexulti and Abilify to me. Abilify isn't an antidepressant, it's an antipsychotic. My other comment is that treating psychological conditions is incredibly hard. The placebo effect is a huge issue with clinical trials whether or not for-profit companies run them.
And as for "me-too" drugs that you mentioned, they have value to. Jenkins from the FDA actually stated we need more of them, not less.[1] Yes, a new drug might only be a small tweak of an existing drug, but all humans aren't the same. A patient may not be able to take an older version, but the me-too works for them. That has added value.
And as for "conning" doctors or patients in buying drugs they don't need, I don't have a lot of sympathy. Doctors are the final decision makers on what drugs are prescribed. If they don't do their homework and just prescribe something because they like the sales rep, that's on them.
“Choice is a good thing,” Jenkins commented, as later drugs in a class often enhance safety or efficacy for certain individuals. The move away from “me-too” drugs, he said, is “not necessarily good for public health.”
> Yes, a new drug might only be a small tweak of an existing drug, but all humans aren't the same. A patient may not be able to take an older version, but the me-too works for them. That has added value.
Indeed this is true, individual patients will respond differently, and inexplicably, to medications in the same class (e.g., Prozac vs. Zoloft). Having these choices is very useful.
However there are considerations beyond pharmacology that reduce the value of these choices. Problems arise because of non-clinical factors, particularly what rules of insurers allow. Formularies have become quite restrictive in many environments making it difficult to make use of the range of alternatives that are theoretically available.
In the last 10 years several new antidepressants have been developed, approved and marketed. That doesn't mean that patients can easily get these drugs even if patients have failed to benefit with older options. Insurers erect barriers like stiff co-payments, or complicated prior authorization procedures as effective impediments.
Consequently what good are new medications if they can't reach the people who need them? There's a strong disincentive to try to use these drugs and I think it's likely manufacturers would have less reason to develop them if they can't sell them.
So doctors would like to be able to prescribe what would work better, but that's frequently not practical. I'd add that doctors are not so easily "conned" into prescribing an agent just because companies and reps are promoting it. Quite to the contrary, making up one's own mind based on the merits and ignoring the sales pitch is the reality.
In my opinion we already know the best predictive model to use, but we choose not to use it. When Jenner developed the first vaccine back around 1800 he didn't use any petri dishes or mathematical models: he just found some nearby children and tested his theory directly.
Now you may find this ethically repulsive but from a consequentialist point of view it was the right thing to do at the time. This line of research saved hundreds of millions of lives.
I say we have gone too far in the other direction. The problem with failed costly clinical trials is not the failure per se but the cost: in money spent, time, reputation, commercial return, legal consequences, etc. And the hidden cost of how long the whole process takes even if successful.
Take the field of monoclonal antibodies. We now have almost off-the-shelf technology to develop drugs that specifically block certain biological mechanisms. But it still takes 10-15 expensive years to progress through different lab and animal models, then healthy humans, before clinical trials on actual patients.
Instead we should reduce the cost of human trials: have a network of terminally ill volunteers to test mabs at an early stage. Fail fast to move fast. Use the money saved to offer defined compensation when things go wrong.
The drugs that we really want to find are the ones with an obvious effect on real patients. This process would find those drugs much faster, and save lives overall.