If ALSTDI is going to be a leader in multitherapy (ducktails), y'all are going to have to learn how to deal with that kind of data. The principles involved are the foundation of modern communications engineering. Of course if you hire a communications engineer, the geek may not be able to look at medical experiment design and understand that it is the same problem as transmitting data over a noisy communications channel. If that's how it goes, fire the cuss and go find yourself another one who "gets it".

The miracle of introducing relevant variables into an experiment is that you can get so much more information out of the experiment with same number of (for instance) mice. Preferably more than one kind of mouse.

The secret is that you know what you put into the experiment. It's the "secret decoder ring" that enables you to extract maximum information out of experimental results which to the naked eye look scrambled. Instead of striving for certainty of one output variable, you get lots of different output variables which are not so certain, and in the aggregate their information content is far greater than the "if we do the same thing enough times" sort of experiment. This is basic information theory the roots of which go back to heat engine design in the mid-1800's, and is the reason why a DSL modem sounds like white noise on a phone line.

To explain it another way, consider a finite impulse response filter, or a correlator which is a different name for the same thing. Suppose you have a controlled experiment with multiple variables and the output data of the experiment looks like a scrambled mess. It really is a scrambled mess: you intentionally scrambled it because you knew you could get more information out of the experiment that way. ......Now rather than applying one simple test to this data (e.g., "is the average greater than x?) you apply as many filters as you like, up to the number of variables (mathematical degrees of freedom) you know something about, and you choose the filter coefficients based on what you know about what you put into the experiment in the first place. Now you've got gobs of data from one experiment.

The mice are a lot more expensive than good experimental design. And with multiple variables you need good experimental design (of which data analysis is a part). So by golly, please master the stuff that mere Pharma lab drones have no comprehension of, and reinvent their world for them.

--Dave J.


[EDIT] Everything I have just said may seem counter-intuitive. Here's the evidence of the power of understanding variability.

If you have a communications channel with noise in it (as all do), you can try achieve certainty of information by transmitting the same signal over and over again. You can never achieve certainty that way, all you can do is keep streteching the probability that you have the right data from the signal. The fallacy is that the more you repeat the same data the less channel bandwidth you're using to transmit the information.

It is our everyday experience that we can download huge computer files at incredible speeds over an ordinary phone line. This is possible only because the signal is encoded to look like noise. Listen to a DSL modem and there it is, sounding like a waterfall. And if the download succeeds at all, all noise has been eliminated from it-- the signal has been faithfully reproduced with certainty despite the noise on the phone line.

Another example is the newfangled digital television. There is a very thin margin between having an absolutely clean picture and no picture at all, and the only reason that margin is greater than zero is because of overhead requirements of encoding it.


--Dave J.

The more I think back on all the stuff that's been said over the last several years about the difficulty of dealing with multiple variables, together with what we see in how clinical trials (and even mouse trials) are actually designed, the more I realize that the medical research industry has no tradition of scientific experimental design. The experiments are designed around the greatest fallacy there is about how to get information through a noisy channel-- which is to keep repeating the same signal until you finally think you have "statistical significance" for just one parameter about which certainty can never be achieved. The problem with this fallacy is that it's easy to understand that if you've got a signal mixed in with noise, redundancy is a way to improve signal to noise ratio. To put it another way, the technique works. There are simple statistical tools for it and since it works people keep on doing it.

The problem is that of all the ways of improving the ratio of information to noise, basic redundancy is the most inefficient. In the medical research industry, evidently without any knowledge of experimental design other than with redundancy the goal of which is to achieve "statistical significance", this inefficiency dooms any effort to learn from multiple variable trials.

In another thread, we are discussing an extreme example of this-- a clinical trial involving 600 hundred ALS patients who are forfeiting their lives in the tragic quest to answer only one question-- does ceftriaxone offer any improvement over placebo? [Actually the question is slightly more nuanced than that, but the basic proposition is that multiple experimental variables must be excluded from the experiment.] There is real doubt over whether this question can be answered even with 600 patients-- such is the tragedy of getting sucked into the mistake of believing that redundancy is how to get information through noise. It's like using DSL for Morse code communication. No self-respecting aggie would ever design a field plot experiment this way.

A trial involving 600 ALS patients investing their very lives into the project should be designed to produce a heretofore unimaginable wealth of information about the efficacy of a vast slew of therapeutics, in combination. With 600 patients, a clinical trial should be producing quite a few dramatic success stories which banish forever the concept of ALS as an "untreatable disease", and which lead directly to so-called "proven treatments".

* * * * * *

So, here's the problem and let's solve it.

The problem is that only multitherapy makes sense, but the industry doesn't know how to do the experimental design needed to develop multitherapy.

So how do we get that expertise into the industry, presumably first at ALSTDI since ALSTDI is proud of being "the leader"? It doesn't look like the answer comes from within the medical research community since that's where the problem itself is.

I didn't learn experimental design from medical research, although I've seen a few isolated examples of superb experimental design from individual researchers who "got it". Most of it's just common sense use of high school math if you think carefully about it, but they never told anyone this in the University as far as I can tell. Ag research is very good at it, it maps nicely to medical research, and the aggies (not being the whizzbang geeks of the University) like to keep it comprehensible. It's easy to learn the fundamentals of experimental design from aggies. The best statistics textbook I ever had was a UC Davis aggie publication. It put mathematics department and engineering department type textbooks to shame.

A communications engineer will be a whizzbang geek who can't connect the dots between a medical research problem and a communications channel problem. Forget I suggested that. You'll have to fire 10 of 'em before you luck onto one who "gets it" and by that time how much more research money has been spent doing the wrong things, and how many more lives have come to an end?

So, hire a good aggie as a teacher and tell the medical researchers they're gonna get real humble and learn fast. That's my suggestion for getting with the multitherapy program that everyone finally acknowledges is necessary.

--Dave J.

Rob & Dr. Vieira & forum denizens,

The single most useful thing that can come out of an in vivo (preferably in human) trial is a success story. The patient either stops progressing for a long time, shows measurable improvement (for more than 3 months in a human subject to weed out placebo effect), or fails to develop the disease (in the case of genetic, and dosing prior to development of symptoms).

This is unlikely to happen without ducktail therapy. ALSTDI's webinar admits to this. Even Rothstein admits to it.

If you've got ten mice, the way to get the maximum information on therapeutic efficacy out of those ten mice is to give each one a different (but sensibly designed) ducktail. The same is true of human subjects despite their high "noise level", or perhaps better said because of their high noise level. This ought to be obvious to anyone involved in ALS experiment design, but if it's not, it can be made obvious with a bit of high school arithmetic.

In a biologicals lab, there will always be times when you have questions about some specific "drug" where its therapeutic efficacy is not the "biggie". Much in vitro research is of this type, and some mouse research will be of this type. But there really is a "biggie" in ALS research, and that is how to actually do something unequivocably useful for a human patient. Since to learn how to do this requires taking best shots at it, there simply isn't any excuse for monotherapy human trials. The best shots are multitherapies. And, since so much of what guides what should be tried in humans is information from animal models, which are expensive tools, good stewardship of resources would dictate that the primary use of the animal models would be in multitherapy screening.

Right now ALSTDI has '846, a drug which helped a whole lot of mice a little bit. My understanding is that this is a murine-only molecule which is an analog of a molecule which is already being used in humans for different purposes. So the mouse data on 846 presents two problems-- it isn't what would be used in humans, and it is of low efficacy even in the mouse.

ALSTDI researchers (and many forum denizens and yes even Dr. Rothstein) can think of a whole slew of sensible things to co-trial with 846 in order to target as much of the degenerative process in the mouse as one can figure out how to target. So when it comes to 846, one hot setup is to take 846 back to the mouse with ducktails, a different ducktail in every mouse. Not big bunches of mice, maybe 10 or 20. You aren't trying to find out if you can extend life by 10%, you're trying to demonstrate something which it only takes one mouse to demonstrate when you finally demonstrate it. Just one "hit" in one mouse. That changes everything. You replicate it-- in principle, only one more mouse is needed, not another 10 or 20 or 100. Now you tell the world "we have a ducktail that's an effective therapeutic in the mouse. Anybody else got anything like that? We think this ducktail should be put into humans, substituting the analogous for-humans molecule for the 846. Who's interested?"

Another possibility is to cut to the chase in the mouse race by trialing the for-human molecule based ducktail array, unless there is evidence that the for-human molecule doesn't work in the mouse.

So, with only 10 or 20 mice in the first round, would you get that hit? Maybe and maybe not, but at least you're doing what you know needs to be done if you're ever going to get a hit. As long as the mice are not being used in this way, you don't get the hit that changes everything. Realistically speaking, this is the only path to effective human therapeutics using mouse models, and since you've already got the mouse colonies you could be getting the game-changer hit or hits literally within weeks.

In baseball, nobody ever got to first base without swinging at the ball, if the pitcher is one who pitches strikes. You have to swing the bat, there is no other way. Most of the time you'll miss, sometimes you'll hit foul, sometimes you'll get thrown out on first. But if you ever want to get to first base, you have to swing the bat at the ball. There are lots of other things a batter might do, but without that one essential thing none of the rest of them matter.

Several years ago I was calling for "give it your best shot" trials. Oddly, it scandalized a few folks. In ALS there is no such thing as a monotherapy best shot, you're shooting blanks. They're loud but they don't hit anything. But now the barge has been turned 180 and everyone admits that multitherapy is the only thing that stands a chance (unless you count stem cell surgery, which is based on entirely different principles). If you want to hit the target, ya gotta take best shots.

And those best shots just ain't all that difficult to take. They're actually easier than trying to demo marginal efficacy in a motherapeutic trial.

--Dave J.

MULTITHERAPY ("DUCKTAIL") TRIAGE

In catastrophes, hospitals manage the chaos by the principle of "triage". Those patients who will get better without treatment, ignore them. Those patients who will not get better even with treatment, ignore them. Devote your attention to Group #3, those patients who need medical resources and will respond favorably to them. Of course all this is a judgment call. The only certainty available is that eventually all will die, including the medics. What matters is what one does meanwhile.

Anyone attempting to formulate a therapeutic regime for treating ALS is stuck with an embarrassment of riches, not with the hopelessness often proclaimed. In order to take action on the chaos, you've gotta do triage.

Don't concern yourself with those things that you can't afford, which have appreciable risk, or which you don't think you understand.

Don't bother with things which probably have no benefit: that is a very long list and you can't fit them all into your stomach.

That leaves a much shorter list of things to take seriously: those things which "make sense for you personally".

We have such an abundance of riches in potentially useful therapeutics, that even that triage maneuver will leave more on the list than can fit in your stomach. So, you have to go through it one by one and prioritize. It's all a judgment call: no way of knowing for sure about anything. The cemetery is the only sure thing and you are trying to play that other game where nothing is certain.

--Dave J.

Nemesis, not having experienced the thing myself, I'll let you and a certain other denizen of the forum work that one out if he's game.

--Dave J.

Also you suggest avoidance of Electromagnetic fields. How is that possible when most PALs are ordered MRI's?

This is possibly the most important thread ever posted in this forum so it deserves to be bumped up now and then. I hope that someday the contents of this thread can be reduced to a single coherent essay (beyond what Aketri has attempted to do by occasionally editing and updating his first post).

* * * * * * *

It is easy to get lost in the molecular minutiae of the leading edge of ALS research and lose sight of big pictures. Here's a slightly different "big picture".

1. ALS is not one disease-- in fact nobody knows how many diseases it is. What they share is a clinical picture of deterioration of both upper and lower motor neurons, without otherwise generalized neuropathy although there may be other neuropathy either as an ALS subset or concurrent with ALS as an independent medical condition.

2. Since ALS is conventionally described as a motor neuron disease, it is easy to think of the disease as beginning in the motor neurons. However there is plenty of evidence that in most ALS the motor neurons are diseased because of processes taking place outside the motor neurons which damage them. Probably the most commonly cited such process is gliosis with release of excess glutamate and y'all know the rest of that story. Another process often cited is something gone wrong at the neuromuscular junction which cause the neuron to disconnect from the NMJ and which probably also leads to axonal transport abnormalities. Yet another process cited as being common although possibly not dominant, is immune system attack on the axons.

3. Complicating such interpretations is the possibility (some would say probability) that this "attack from outside the neuron" is actually a result of immune activation by degenerative processes happening within the neuron which provoke immune response. The processes that provoke the immune response might not be all that big a deal by themselves, but the immune response then triggers an accelerated degenerative cascade.

4. In any individual patient, in most cases it will not be possible to identify specifically what degenerative mechanisms are at work. The picture is not clear even in the case of SOD-1 familial ALS which has been researched in great detail using a Frankenmouse engineered with the human SOD-1 defective gene.

5. Therefore from a clinical (or extraclinical) therapy standpoint, the therapeutic approach needs to assume that any degenerative mechanism believed to be common in ALS is present in this particular patient, unless for some reason it can be ruled out. This means that the therapeutic regime will include features of immune modulation (especially anti-inflammatory), and will also treat the various elements of the glutamate-induced degenerative cascade. The therapeutic regime need not be limited to those two battlefields, but as a minimum it should include them.

6. Good arguments could be made that a comprehensive therapeutic regime should also include features of accelerating healing processes, and also normalizing problems at the neuromuscular junction. Addressing the former is possible now, but the NMJ problem poorly understood and even disputed as relevant (ALS includes upper motor neuron disease) so it's not obvious what one should do to normalize the possibly problematic NMJ.


* * * * * * *

The good news is that therapeutic agents for most of this are already available at low cost and mostly or entirely without a prescription, at least in the USA. The bad news is that it's a big pile of pills, and that no such pile of pills has ever been "clinically trialed" and thanks to the way that semi-socialized capitalist medicine works, no such pile of pills ever will be "clinically trialed". Even when FDA-approved ALS drugs eventually begin to trickle onto the market, for that tiny fraction of ALS patients who actually have access to such drugs, the problem will still be there that no one pill is likely to fix the whole thing and the pile of pills needed to do it will never be clinically trialed. The lab rat in the mirror is as good as it gets today. And, after we're all dead and today's teenagers finally become the new crop of oldsters who develop motor neuron disease, they will be facing the same problem. Either no access to FDA-approved treatments, or having that access and finding out that the approved treatments offer nothing that resembles a cure. The only doctor of any use will be the lab rat in the mirror.

THEREFORE: "the quest for the cure for ALS" needs to focus on general principles of treatment, and on how to get the knowledge of how to apply those principles into the hands of ALS patients. The present emphasis on developing FDA-approved drugs for Rx'ing by MD's is useless for ALS patients living today and will be useless for the overwhelming majority of ALS patients 20 years from now.

Eric, it is abandoning, "that's how we have always done it" mentality. Between the FDA, and the inability of people to think outside the box, we could still be a decade, or decades away from a cure. Forget all the fancy medical jargon that is thrown around, show me some results. As many a person has said, if somebody did find the cure, it would be years before it became available to the public. It's way past time to try something new, a new approach, a new set of eyes, and less government intervention.

David

After decades of research, is “anything” a given, including things, that we have been told did not work? It appears to me, that we have decades of research, and not one single thing is a given. I will grant you, that a number of medical advances, have been made for other medical conditions, but what does that have to do with ALS? You mentioned the complexity of ALS, how do we know that it is complex, again you are making an assumption, that we do not know is the case. It’s like all these people, tearing their hair out, trying to fix a machine, and the janitor walks by and says, hey guys, did you know it wasn’t plugged in? Also reminds me of the Rubik’s Cube, where kids could solve the Rubik’s Cube, but their parents and/or adults couldn’t do it. So why could a kid do it, and an adult couldn’t? The adult started with the mindset, this is impossible. The kid on the other hand, had no preconceived notions, and they mastered the Cube. Just like you said, the “complexity of our issue.” We need the equivalent of a kid looking at this problem, or people in the future, will be having this same exchange in this forum, decades into the future.

David

Is such way of work proven in any other human disease?

Cristina