146 replies to this topic

[AgeVivo]

smithx, you are using strong words like "useless" and "of course" and others, so that whatever is different from what you are proposing sounds obviously inadequate. I could take your proposal and do the same -- I recognize I slightly started to follow that style in my last email and it is not a good idea.In the next posts, if you agree, Let's try to think of the two possibilities without using the "appearance" of words and with an open mind on the 2 solutions. When I try to do it I think that large openness is a large garantee, in addition to be best for advocacy; the risk you point is a not so big risk as your words try to suggest it seems to me; imagine you proposed the other solution and I liked the isolated experiments solution; I would say being independant is best not to be influenced; you would say it is not possible to be independant especially with so much interest and so long studies, there would of course be leaks everywjere so what's the point, then let's do fully otransparent studybecause anything else is sneaky; I would say better slight leaks than terrible leaks; you would say that between two similar 'leaks' better the one with greater advocacy.
please continue both sides/make a shared 2 column tables perhaps (googledoc) so that we are doing smthg together

myself I have the impression that the hair of my mice is longer than what I had with my previous mice (look at the pictures in my thread) and I am planning to ask the c60 researchers, next time I meet them, about how complex it would be to test droppings (they have analysis machines and use them in their daily work).

[smithx]

Do you disagree that one of the major topics of discussion will be guessing which mice have the c60 and which have the placebo? If we put the study subjects in touch with each other, the experiment will certainly not be able to be considered a true "double blinded" study.

The whole idea of testing droppings really doesn't seem to make sense in that it's a whole new set of issues, and if anyone was really cheating they could relatively easily substitute some other droppings from some friends' mice.

If we are really trying to get information which is good enough to help people make a decision about whether to take this stuff themselves, we need to get data which is reliable. Adding all these other levels of complication just reduces the quality of any data we get to the point that it's not going to be good enough to base any decisions on, and that, to me, would make the whole exercise basically worthless.

[niner]

AgeVivo, what about my proposal to set up google docs for each pet owner that we can also look at? Are you opposed to that because it's "sneaky"? It wouldn't be sneaky if we tell them that this is our communication method, and we will of course be looking at them in order to gather important data. If we weren't, what would be the point of making them keep a log? If we require them to send us samples of rat poop, I think you can kiss this whole shebang goodbye. That is a profound level of complication. They can buy C60 on the open market, and if they know how to use google, they'll be able to figure that out. These are pet owners, not the rat facility at the NIA. They care a lot about their animals, and if they are typical pet owners they will want to have the active, not the placebo.

[AgeVivo]

Hi Niner

Concerning google docs: I wouldn't say "Sneaky" because I don't like such terms (it was an example in my last post indicating that such words are not constructive to find solutions together). The question is who will have access to it? Who are you designing with "we"? Because, for the pet owner, volunteering to something fun and ending up working for a few persons (who may have strong beliefs on how to do this or that although they do not do it themselves) might not be the best; participating with anyone in the open world (with a much more varied analysis) is not the same; and in terms of advocacy, it is in very far from the solution currently considered.

If we were to do a googledoc I would prefer (*) that the logs are done on LongeCity in a way such that only paid LongeCIty members can read the logs: it would at least bring money to LongeCity. That would be good for LongeCity's budget (**), but considerably reduced advocacy compared to the opened part of LongeCity, and also considerably reduced possibility of new persons to come and discuss.

(*) Unless perhaps I misunderstand the googledoc solution you have in mind? Do not hesitate to elaborate. Thank you.
(**) a difficult tradeoff, I recognize...

Edited by AgeVivo, 20 September 2012 - 08:21 PM.

[AgeVivo]

In fact, I was also counting on the first participants' blogs to lead other pets owner to join the experiment.

I realize that if logs are not easily accessible, that may not work as well; and doesn't help/could ruin the experiment.

Edited by AgeVivo, 20 September 2012 - 08:47 PM.

[niner]

The people who would have access would be the pet owner who is making the entries and one or more Longecity members who are analyzing the data. I don't see the logs as a general access document.

I think we should decide what we want to get out of this exercise, because it can't be all things to all people. I would recommend that we do not try to use it as a fundraising venture. We might be able to summarize data and post that periodically, though that might have the effect of letting people know who has the active and who has the placebo, so maybe that wouldn't work. If it comes down to quality science versus advocacy, I think that the quality science would be the best advocacy. A bit of advocacy at the cost of a ruined experiment, it seems to me, doesn't advance the cause as much.

[AgeVivo]

The people who would have access would be the pet owner who is making the entries and one or more Longecity members who are analyzing the data. I don't see the logs as a general access document.

I see:
- essentially Smithx's point but with a googledoc; between a googledoc and Smithx's professional solution it seems to me the later is more adapted
- not open science no crowd sharing. no associated advocacy, no details viewed and exchanged by many. Closer to traditional hidden science (with a science logbook and a supervisor).

Sincerely I think you should trust open science a little more.

it can't be all things to all people

To me, it seems it can:
- more quality due to open science(*)
- more advocacy due to open science
- more attractiveness to LongeCity due to advocacy and quality

[niner]

Sincerely I think you should trust open science a little more.

- more quality due to open science(*)

I'm all for open science where it's appropriate, but NO ONE would run a trial this way. I think that in this case, being excessively "open" gives you less quality, not more. Of course, no one would use multiple uncontrolled sites, with different species, ages, and quality of care either. You can certainly run the exercise however you'd like, but I think that you should trust open science a little less...

[smithx]

To niner's point, we have two choices here:

1) Try to actually replicate Baati and get reliable data which people can use to help them decide what they should do about C60.

2) Use this as a fun, community-building project, which will not get any useful data, but which may make people excited and increase pageviews on the site.

I was thinking we were trying to do 1. If so, we need to be fairly rigorous and actually try to do a double-blind study.

But maybe what people here really want to do is 2. In that case, it seems disingenuous to claim to be doing any kind of science, so if that's what people here really want to do, just come out and say so and I will stop suggesting ways to make sure we get good data.

[Mind]

To smithx's point: I obviously don't see the same downfall through openness, but maybe I have some extra bias or trustworthiness in people. Maybe I am not adequately empathizing with the typical pet owner, in that they would want the best for their pet, and would do anything (including wreck the experiment) in order to give their pet a longer life). As I mentioned earlier, I place some faith in community peer pressure in order to maintain some rigor, but maybe that would not be enough.

We have 100 free samples of C60oo and can mix up another 100 of placebo. This is a good amount of samples to work with.

We have 2 people already who are willing to conduct a blinded experiment with multiple mice/rats in one environment/home. If we can find a couple more, maybe we can get a sample size up around 50 or more. Perhaps these could be our attempt at a more controlled experiment. They would not post openly about animal status. They would keep a private log (google doc) accessible to a couple of Longecity members (perhaps Smithx? being one), but of course not AgeVivo, myself, or anyone involved in making or mailing the samples.

For the other approximately 50 C60oo samples, we could blaze some new ground in open community science. The knowledge gained about how the project works (psychology, logistics, etc.) might be as valuable as any data that is collected.

Edited by Mind, 21 September 2012 - 05:33 PM.

[smithx]

I think Mind's idea is a good one: if we can do a study with 50 subjects, we would expect to see significant results if Baati's results are replicable.

For an open study, I suggest that we not pretend it's double blind when it clearly won't be. So for the other 50 I propose that we do an open-label study in which everyone knows what the mice are getting.

The advantage to this is that people who want to give their mice C60 would know they were doing that, and would not have an incentive to cheat as they might if they were in a blinded study and suspected their mice were not getting it. People who committed to testing only olive oil would be more likely to follow through on that because that's what they agreed to in the first place, and people who just want to serve as controls would also be more likely to follow through on that.

So, yes, let's do two studies: one traditional rigorous double-blind study which will give us absolutely reliable data, and one completely open study, which will hide nothing from anyone.

If both studies give us the same results, we will have a good case to be made for the reliability of those results. If not, I would tend to rely more on the rigorous double blind study's results.

[AgeVivo]

I suggest that we not pretend it's double blind when it clearly won't be

?

Smithx, I have tried many times to express that a double-blind, multi-centered, open-science approach is not deemed to fail. So far I have felt not listened. I see pros and cons in the 2 solutions, where you appearently only see pros on one side and cons on the other?

I see potential troubles in dividing the small N further by 2. So far I am not sure we are heading to the (/a) right way. It you could, *you*, this time, list pros and cons of the different solutions, I hope I would be reassured. Because I could be wrong, but I prefer to be reassured that we are at least well balanced in our choices. Thank you for your comprehension.

[niner]

I suggest that we not pretend it's double blind when it clearly won't be

Smithx, I have tried many times to express that a double-blind, multi-centered, open-science approach is not deemed to fail. So far I have felt not listened. I see pros and cons in the 2 solutions, where you appearently only see pros on one side and cons on the other?

I see potential troubles in dividing the small N further by 2.

I don't think that smithx and I see only pros on one side and cons on the other, and we are certainly listening to you. I just think that you are underestimating the power of human nature to be irrational in order to protect a loved one. Maybe not even irrational, just caring more about their pet than they do about the whole experiment.

I'd really like to bring up the issue of the "small" N. The reason that most trials need a large N is because the effects they are measuring are not very different from statistical fluctuations. In the case of Baati, the effect is enormous. There is no need for a large N, because the effect is so much larger than the statistical fluctuations in normal rat lifespan. I think it's a fine idea to have the N be larger than 6, just because it makes people feel better, but N = 50 is still way more than we need to demonstrate a 90% life extension convincingly.

How many pet owners do we have at this point?

[AgeVivo]

Thank you for reassuring me. Then, if you don't mind, while I try to contact each person who said was interested during this week-end and in the next few days (about a dozen, but very few came back to me later), you could list the pros and cons of the different solutions as objectively as possible and write them (eg in a table, that can evolve) so that we choose together based on a correct view of things.

Edited by AgeVivo, 22 September 2012 - 01:33 PM.

[smithx]

Here is a list of the advantages and disadvantages and disadvantages of randomized, placebo controlled, double blind studies:
http://en.wikipedia....rial#Advantages

Double blind placebo controlled studies remove any bias and also remove the ability to fudge data, either intentionally or subconsciously.

My point about the "open" version with public logs is this:

- Public logs will enable speculation as to who has the placebo and who has the compound under test.
- This will compromise one aspect of a "double blinded" study, in that participants may be able to guess who has which type of sample.
- Some people randomly assigned to one group or another would have had a preference other than what they got assigned to.
- This could produce dissatisfaction leading to either dropping out of the study or cheating.
- A further aspect of a blinded study is that no one concerned with the study should see the results until the study is completed. Open logs will also clearly invalidate that part of the requirement as well.
- Given all these points, it's clear that an open study can't be called "double blind", therefore my proposal that if we have an "open" arm of the study, we make it an "open label" study and assign people to groups they agree to be in: olive oil, control, and C60.
- The fact that people have volunteered for a given role will tend to make them more committed to that role and less likely to cheat.

All that being said, open label studies are not considered to be able to provide as reliable a set of results as double blind placebo controlled studies, hence the desire to have a rigorous study and not to put everyone in the "open" category. My strong preference, as noted many times above, would be to do the entire study as a placebo controlled double blind study, but if people really want to do an "open" study as well, and we will have enough participants, then it makes sense to have two arms, one rigorous and the other open.

If we have 50 C60 mice and 50 olive oil (and I hope for 50 controls) in each arm, we will have a much larger sample size that Baati in each, so the entire exercise will be very useful nonetheless.

[AgeVivo]

Thank you SmithX for the details. I have noted those cons, I have also pros and cons for the 2 types of studies. Sorry to ask again, but could you describe cons of the close study? and pros of the open study I prefer to ask you (and/or niner), so that we here all have more balanced talks. Then I'll try to do a summary table of pros and cons that can evolve (if you want to do it, great)

[AgeVivo]

While you are working on a list of cons of the closed-type-of-study and pros of the opened-type-of-study, and while I am contacting likely participants,
here is the take home message of the C60 LongeCity article done with s123 and revised with Caliban:

http://www.longecity...lates/c60health
The study requires repetition in an independent laboratory. This case has illustrated the importance of open and transparent science for progress in life extension research.

[niner]

The study requires repetition in an independent laboratory. This case has illustrated the importance of open and transparent science for progress in life extension research.

It's absolutely correct that the study needs to be replicated in an independent lab. However, the next sentence doesn't follow from the first. Baati et al. were as open and transparent as they could be; they hid nothing, and even provided samples of their C60-oo. I think that's the exact right level of openness and transparency. It doesn't mean that compromising a trial with open logs is a good thing.

I like the idea of having two tracks- an open label study for the pet owners, and a double blind study done with control animals at the same site as the C60 animals.

[AgeVivo]

I largely agree.
One piece of different view, but that is already expressed in the posts above: I still think that the transparency of public logbooks is not deemed to be incompatible with the "treatment or placebo" approach, and is actually a very desirable thing to do (unknown content till the end, despite convictions by some that may reveal to be wrong; very exciting ) (unless the placebo is white or blue, of course). I agree there are dangers: participants sending pictures of their solutions even if told not to do so, etc. As a result, in terms of wording, where you say "open-label" for now I would be less precise and I am stil thinking "open-science" (which may be either be treatment-or-placebo, or open-label).

I also see cons of the closed-type of study but I would prefer that it is said by smithx or you in order to avoid complex discussions. I suppose that the logs would be made public at the very end and can't be modified in-between.
=> Waiting for cons of the closed-type-of-study and pros of the opened-type-of-study, from smithx and/or you, before discussing further.

Edited by AgeVivo, 22 September 2012 - 10:54 PM.

[AgeVivo]

I have given the empty bottles and the full bottles cointaining C60 (packaged as I received them, without looking at the content; the researcher told me he will have to open the package to view the color: nice packing job carbon60), and the olive oil. This week he will adjust his placebo to look like the full bottles.

When the researcher will feel ready with the placebo content (e.g. one or 2 weeks) AND when we will feel ready with the randomization procedure (e.g. one or 2 weeks?), I will tell him to number the placebos and treatments randomly, this will be done under a video by some person who is independent, someone with whom I discussed about the experiment, who is not into biology or life extension, and who will soon change countries.

Details, that should be seen on the video // draft version:

1. Just before the randomization procedure
- the labels will be removed (by washing the bottles, them removing the paper label then drying the label with a dishcloth => the treatment bottles will be put on the left, and the placebo bottles will be put on the right.
- The "left" and "right" sections should be very far apart, with some clear separators that also give a lot of room for the "center", where a keyboard, a computer screen, a computer mouse, two sheets of papers and a paper-pen and a glass, water-resistant pen should be visible. Each sheet of paper should contain the same, namely a table of two columns and at least 20 rows: the first colum contains "1 2 3 ... 20" and the right column is empty. The computer screen should show an excel file with the 2 columns, similarly.

2. Then for each number from 1 to 20, the researcher will roll a dice.
a)- If the dice shows 1,2 or 3, he will assign the number to a bottle on the left. If the dice shows 4,5 or 6, he will assign the number to a bottle on the right. The number will be assigned with water-resistant inc. At this stage the bottle with remain in its section ("left" or "right")
b)- He will fill the two papers, by writing on the empty column, at the line of the number, whether it is "placebo" or "treatment". He will fill it also on the excel file, similarly, as save the excel file. The bottle will then be put in the center.

3. After the video.
- A password will be written on the papers and on the excel file. The two papers will be put in two closed enveloppes and some zig-zag lines will be put on the sealing, so that if opened at anytime, looking at the sealing later will show it. Pictures of the sealed enveloppes will be taken and uploaded online.
- The excel file will be compressed in zip, and the zip will be password-protected with that password. The video will be encoded in the zip file as well. The zip file will be copied in a USB key that the video-taker will keep. The zip file will be sent to mind and I (we'll need to figure how, esp if size is big) or directly put on LongeCity or elsewhere online (preferably a place that not everyone has access to; in case some stupid person decided to try to decode the password and publish it). This must be done with great care to ensure that only the protected zip file is provided, with the right password.

4. Later, when someone will have to receive "treatment or placebo",
- I will tell the researcher the LongeCIty pseudoname and the number of bottles needed (one bottle per tested rat; one bottle for many mice). The researcher will roll a dice: if 1, 2 or 3, then the person will receive treatment. If 4,5 or 6, the person will receive placebo. In the excel file, he will write on a third column the LongeCity-username of that person, for as many corresponding lines as there are corresponding tubes. Of note: he will randomly chose which lines to fill (e.g. if tubes 1 5 7 9 13 16 18 contain the desired content, and 3 bottles are needed, he may for example choose 5 7 18)
- The researcher will put the bottles in a box and write the numbers of the bottles on a piece of paper. He will give me both. I will send the box to the participant, without having looked at the content and communicated the numbers I received. The participant will then indicate to me (/online?) the tube numbers he received. I will check that I received the same numbers. (/take a picture of the paper I received and post it online?)
- The researcher file will be password protected and put online (again, a place that not everyone has access to).

Of note, if mind and I decide that a participant can have both the treatment or placebo, for different rats, then this procedure will be done twice: the first time for treatment or placebo, the second time for placebo or treatment. I will just indicate it very clearly to the researcher so that instead of rolling the dice the second time he will take the other type of solution.


Warnings/variants/other


- asking for more bottles
If the researcher feels that we are starting to lack treatment or placebo bottles (no quantified limit will ever be fixed in advance, this will be his pure personal judgement and he will not communicate how many tubes are remaining; otherwise one might guess the treatment versus placebo distribution so far) then he will ask another 20 tubes to me and mind, and we will ask it to carbon60. Once we receives the empty bottles and full bottles, just like today, he will do the same (counting 21...40). I imagine that the video is nnot needed each time. If he makes two other papers that he puts in sealed enveloppes, and send one of the enveloppe to the video-taker, that is largely sufficient.

- Treatment-or-Placebo, or Treatment-or-ColoredOil-or-Colored-Water?
The procedure could work as well with more than 2 types of solution. (the dice results would be aggregated in 1,2 then 3,4 then 5,6). This is of course a fundamental question. Eg, should 50% of the bottles contain C60oo, 25% contain colored-oil and 25% colored-water? [I think that we do all those complex things in case the treatment could increase lifespan but much less spectacularly that in Baati et al]


- Other?

[AgeVivo]

Of note, the previous (long...) post is the randomization procedure.

[smithx]

How variable are the number of mice per participant? The only reason I can think of for such a complicated procedure would be if we have totally variable numbers of bottles which need to be sent to each participant.

I think it would be simpler to just randomize a number list and put labels on ziplock bags containing 1, 2, or 3 bottles.

Then there's no thinking required when a new participant is added: just pick the next pack in numerical order and send it to them. No one will know what it contained until after the end of the experiment.

I also suggest that each group should contain equal numbers of control (no treatment, not even colored water), olive oil, and c60 mice.

Edited by smithx, 26 September 2012 - 09:18 PM.

[AgeVivo]

How variable are the number of mice per participant?

I don't know yet and I proposed this complex solid procedure in order to yet be able to start already, without complains that the procedure isn't good enough (which I thought was your main point at first) => are you opposed against this complex procedure? (complex but it adds no constrains to the number of animals and experimenters)

I think it would be simpler to just randomize a number list and put labels on ziplock bags containing 1, 2, or 3 bottles.

Why not if it works: For now we have 10 C60 bottles and 10 colored oil bottles: how would you split it so that we can give it to the first participants (starting with Maureen Thomas) and so that no one can guess who receives what.

Please look at Maureen Thomas' thread: http://www.longecity...tters-an-intro/ and note that she has rats of many different ages. One question is whether we should only restrain to starting at age 10 months or not? Because for now the main restriction is to have experimenters really start (eg actually creating a logbook) so I would be in favor of adding another starting age, like 18 months old, which is an opportunistic additional group than consumes some bottles but not additional experimenters.

Thank you for leting me know what you think. Also, for the hidden-type-of-experiment, let me know if I should try to list pros and cons or if you think you can do it.

Best regards
AgeVivo

[smithx]

I was thinking of just randomizing a list using excel and then labeling with random labels, but if you have to give different quantities of bottles to each person, that wouldn't work.

With someone like Maureen Thomas who has many different animals at different ages, I would hope that every animal would be part of the study. Otherwise, I'd be concerned as to whether there could be mistakes made when feeding them every day. Yes she has a card file which would let each animal be identified by its markings, but is it realistic to think that she'd refer to the card file every day, or that she'd never mistake one rat for another? I think that errors are very likely to creep in with situations like this: either all the animals should be fed or none of them, per household/participant. When they die, such a catalog would be fine to identify which animal had died, but for daily feeding I wouldn't trust it.

As to the benefits of double-blind placebo-controlled studies, they are the gold standard, and the type of study universally believed to be the most reliable. The Wikipedia page I linked to a while ago lists pros and cons, but I don't think going through such an exercise is necessary: this type of study is the best, without question.

I thought we had agreed to use 50 bottles for an open-label test with open logs (which wouldn't require any randomization), and the other 50 for a classic double-blind placebo-controlled study. Did we not agree on that? What's the objection?

For the double-blind placebo-controlled wing of the study, the best randomization method I have come up with is to line up all the participants in advance so we know how many bottles each participant will be getting. Ideally there should be at least two participants with each number of animals. After this is done, it's a simple matter of randomizing the packs of bottles using a random number list in excel.

Perhaps niner can provide some feedback on your proposed randomization method or suggest an alternative.

[niner]

Because the animals are of different ages, species, and physical conditions, and they are being cared for under varied conditions and regimens, the only way we are going to demonstrate an effect is for the animals to live a LOT longer than their normal lifespan. The placebos are essentially meaningless in this case, and only serve to reduce the number of animals that would have a chance to show the effect. Therefore, I really think that the pet owners should be in an open label experiment, in which case open logs would not be particularly harmful. I'll ask again, since it's important to how we conduct this study, how many pet owners and animals do we have so far?

Do we have anyone who could perform the placebo controlled study competently? For that to be done properly, the animals should be the same species and starting age, and each arm should have an equal number of each gender. Of course, both groups should be under the care of the same person. To make the dosing easy, there should be placebo cages and active cages; cages and bottles could be labeled with the same code, making it fairly difficult to screw up. I would have no problem with Longecity funding the cost of animals, cages and food for the placebo controlled experiment.

As for randomization, if we have one or several people who are willing to run the placebo controlled experiment, with a minimum of perhaps 10-12 rats each, they should get an equal number of placebos and actives, which could be randomized using the Excel method that smithx suggested prior to shipping them to the site. I don't really see a need for a lot of complexity in the procedure. To randomize sample for 20 animals, all we need is twenty labels that have the numbers one through twenty on them. We then computationally shuffle the labels so they are in random order, put the first half of them on the placebo bottles, and the second half on the active bottles. The shuffled list is then the key that must be carefully preserved; the first ten number on the shuffled list are placebo, the second active.

As an aside, I really don't want to wait four years for a result. It would be really great if we could set up some experiments with shorter lived species. Senescence Accelerated Mice come to mind here, albeit with the understanding that their physiology is messed up, but at least they are mammals. I think that some work in lower animals would be worthwhile. We probably should try to partner with some academic labs for some of this.

[AgeVivo]

Summary for us and for those who have not read everything up to here and who might not understand our hesitations. I think it answers the current questions too.

1. Number of animals, use of controls or not
Baati et al found a strong life extension with C60oo... based on only 6 rats treated.

One might think that the impact is so strong that a few animals is sufficient to doublecheck the experiment, without any control. I am doing n=3 in mice. If so perhaps we only need to do it with n=3 in rats?

However the answer is yes... and no, as can be guessed through niner's implicit double view above. We in fact still want to have a good number of animals, and controls as well. Indeed, *if* what Baati et al reported is not biased, depending on the conditions, the animals etc we might find the same life extensions as them, or more or less. Including less. Therefore the desire to repeat with more animals, more conditions than Baati et al, and still with controls to keep some minimal refinement in estimating the impact.

The controls can be colored-olive-oil-without-C60, or even water (that last control can not be done in a blind way)

2. Open-type or hidden-type of experiment
Will pet owners do the experiment correctly if they display their results online? if they don't? if they have both controls and treatments?

Science:
- Pet owners are clearly told that they won't know if what they are giving is treatment or placebo. Will they not resist and buy C60 online and give it to their pets without saying it? Or will they discover what they are giving and give a special care accordingly? It seems unlikely to me but some are concerned about it, and those persons think that having some pet owners not show their experiment online (so unlike what I do with my 3 mice) reduces such risks.
- I and others think that blogging about the exeriment all along (like I do with my 3 mice) is the right way to go, because it will generate more motivation and more feedback to keep the experiment till the end (what I perceive as the biggest risk of the experiment). The ones who are in favor of the hidden-type of experiment may argue the contrary.

Advocacy:
- the open-experiment is of course more exciting
- the advocates of the hidden-experiment think that there is more science in doing the hidden-type and therefore more advocacy too.

So both views exist and it is likely that both types of experiments will coexist. Probably the best is to do it depending on what each experimenter prefers.

3. Randomization: one number per groups of bottles or one number per bottle?

This is probably more a secondary question. For now we have 10 bottles of C60oo, carbon60 told us that we could have up to 50. The one -number-per-bottle method works in all cases but was described in a complex way. The one-number-per-bottle-group is expressed in a way that sounds more simple but requires to start with more than 10 bottles even if we don't reach many pet owners.

As for the number of participants, for now we have about 15 persons who said they are interested with on average approx 4 pets per person, so a total of n=60 (or 30 or 20 depending on the use of controls), but there is only 1/2/3 person(s) who stepped in so far by creating a corresponding longecity logbook (depending on whether I and logic count or not).

Edited by AgeVivo, 30 September 2012 - 10:05 PM.

[niner]

Thanks for an excellent summary, AgeVivo.

- I and others think that blogging about the exeriment all along (like I do with my 3 mice) is the right way to go, because it will generate more motivation and more feedback to keep the experiment till the end (what I perceive as the biggest risk of the experiment). The ones who are in favor of the hidden-type of experiment may argue the contrary.

I think this is the best argument for an open experiment. I agree that there is a risk of people dropping out as the experiment drags on, and something that keeps them engaged would help to prevent that.

As for the number of participants, for now we have about 15 persons who said they are interested with on average approx 4 pets per person, so a total of n=60 (or 30 or 20 depending on the use of controls), but there is only 1/2/3 person(s) who stepped in so far by creating a corresponding longecity logbook (depending on whether I and logic count or not).

I think this should be telling us something: A lot of people would be willing to feed a drug to their animals in the hope of getting them to live longer, but everyone has busy lives and entering a lot of stuff in an online log (particularly pictures or video) on a regular basis is really asking a lot of them. I think we'll find that most will not want to keep up a log if it takes much time or is at all difficult.

We could have the best of both worlds if we give all the pet owners actives, and run a controlled experiment at one or two sites. Perhaps AgeVivo and Logic could run the controlled experiments? It's a big commitment, so I would understand if either of you can't do it.

[Logic]

I am keen to have controls; placebo, and C60oo rats.
Cost is a problem for me.
Also I feel it may be important to get use rats from the same litter?

[AgeVivo]

I think that 2 places is very limited but my guess is that once we start with a few pet owners then the number of pet owners will gradually increase.

On the colorant side :
- the researcher told me he now has found a way to color the oil in a way that really looks like C60oo [still it will be clear that no pet owner should post pictures of the solutions online nor comments on the colors]
- He is waiting for some other colorant (ordered but not delivered yet) to try to minimize the possibilities of biological actions of the colorants while still looking the same.

Things are moving slowly but correctly. Slowly due to the questions required to think correctly about such an experiment and due to the natural 'business' of everyone. But really correctly it seems (crossing my fingers). I will continue my round around pet owners this week-end.

[Mind]

So much discussion. I have to wrap this up through soft-tyranny because I am trying to write a proposal to the Board for a $500 grant.

We will have two branches of this study:

a.) A closed double blind study for pet owners and volunteers that are willing to host multiple animals in their homes. They will, at a minimum, dose the animals with C60 & placebo olive oil. Having a third group of controls who get neither C60 or placebo would be a bonus and be encouraged. So far we have Logic and Maureen Thomas (and maybe AgeVivo) who have multiple animals who could do this type of experiment. Most of the other people who have expressed interest have (or will procure) 3 to 4 rats/mice. So far it has been suggested that these experimenters get placebo only, or C60 only. I agree with this since it will be more simple for the experimenters and will reduce the chance of error. These experimenters will keep a private log under supervision of 2 Longecity appointed supervisors.

b.) An open study where the experimenters know whether they are getting C60oo or colored olive oil. These experimenters will follow the same protocols as the closed branch of the study but will post openly in in the Longecity forum about the progress of their experiment. They will essentially be supervised by "the community",

All experimenters will dose according AgeVivo's protocol - no food the night before, and then fed C60 or placebo on bread the next day. NO GAVAGE.

All experimenters will sign a liability waiver before starting the experiment.

All experimenters will be given suggested standard care tips for mice/rat/rodent husbandry.

The experiment will involve many different species of rats and mice at many different ages. This is not ideal, but we will deal with it.

After starting out with up to a dozen experimenters and perhaps 20 to 40 animals, we will be adding more, gradually over coming months.

Also, a minor correction, Carbon60 has offered 100 samples of C60oo. Therefore with treatment & placebo we can potentially have 200 animals in the experiment, and maybe a few extra since Logic and a couple others will be having several animals, including controls (no C60 or olive oil placebo) in the same environment.

We have to accept that we are blazing new ground. There will be logistical difficulties, errors, disappointment, and drop outs. Even so, I am confident that we CAN get some meaningful data on C60 as a potential life extending substance. We will have to work together, encourage, and monitor the experimenters. We all know that there are many "holes" to be poked in our experiment by seasoned scientists and researchers and we have to accept that RIGHT NOW. We will duly note the peculiar arrangements and potential statistical abnormalities in our final summary. "Outsiders" will then have the proper perspective when evaluating/accepting/rejecting the data.

The randomization procedure has not been completely settled, but we are close.