When Will We Feel Safe to Cruise Again

[HappyInVan]

1 minute ago, nocl said:

If you actually understand what my concerns were.

 

 

Well, the primary concern of the general population is that they receive the vaccine without all the maybe, if, and buts. 

 

For populations in low risk areas, we're doing fine with masks and distancing. I'm willing to wait for vaccines with better parameters and metrics. Something to inspire confidence.

 

For the rest, good luck to you. Hope that life returns to normal soon!

 

 

[nocl]

7 minutes ago, HappyInVan said:

 

Well, the primary concern of the general population is that they receive the vaccine without all the maybe, if, and buts. 

 

For populations in low risk areas, we're doing fine with masks and distancing. I'm willing to wait for vaccines with better parameters and metrics. Something to inspire confidence.

 

For the rest, good luck to you. Hope that life returns to normal soon!

 

 

AS far as the general population they generally do not know any more than drug  prescribed by their physician.  Most don't read the package insert or have any clue what it means.  They have no idea the details of clinical trials, the way the trial design impacts the data collected, the nature of indications, the scope of the label, etc.  With this vaccine there is more visibility of the process than ever before.  However, visibility is not the same as education about the process.  So there are all kinds of mis-information, some intentional by various people, a magnification of concerns due to the lack of knowledge about the process. 

 

Making drugs is a lot like making sausage, quite a mess while it is being made, but looks good when finished.

 

What is going on with the vaccines, is about the same if you were going to buy an old house and brought in a contractor to tell you every possible thing that might go wrong with the house, without actually inspecting the house.  Different groups on the Internet are very happy to tell people every possible thing that might go wrong with the vaccines, without actually having access to any of the data.

Edited November 2, 2020 by nocl

[HappyInVan]

14 minutes ago, nocl said:

Different groups on the Internet are very happy to tell people every possible thing that might go wrong with the vaccines, without actually having access to any of the data.

 

As we can see from this thread, there are numbers of people willing to endorse the vaccine without access to the data. or understanding the limitations.

 

Anyway, I'm sure that the public won't be thrilled to learn that millions of lives will be put at risk because of a study on just 150 (or 32) cases (out of a population of 44k).

 

 

Edited November 2, 2020 by HappyInVan

[nocl]

2 minutes ago, HappyInVan said:

 

As we can see from this thread, there are numbers of people willing to endorse the vaccine without access to the data. or understanding the limitations.

 

Anyway, I'm sure that the public won't be thrilled to learn that millions of lives will be put at risk because of a study on just 150 cases (out of a population of 44k).

 

 

That is exactly where you are wrong.  There will not be millions of lives placed at risk because of the study of 150 cases.  That is normal for this kind of study.  What you seem to be missing is that in ethical clinical trials (which all of them run in the US are most other countries are) are design to get the information needed, while minimizing risk to those volunteers that are participating in the trial.  150 cases is sufficient to test the drug.  Waiting for more will not change the results, only increase the risk for the volunteers and the delay the potential for the vaccine to provide medical benefit to meet an un met medical need.

 

There are also people on this thread that are willing to criticize the vaccine and the process with out truly understanding the process or have any experience with it.  That is why this work and the analysis of the data is done by people that do under stand and do have experience in drug development and approval.

 

Those that do have experience and understanding of the process are waiting for the data.

[HappyInVan]

Just now, nocl said:

150 cases is sufficient to test the drug.  Waiting for more will not change the results, only increase the risk for the volunteers and the delay the potential for the vaccine to provide medical benefit to meet an un met medical need.

 

 

I have confidence in the Law of Large Numbers.

 

https://en.wikipedia.org/wiki/Law_of_large_numbers

 

 

[nocl]

Just now, HappyInVan said:

 

I have confidence in the Law of Large Numbers.

 

https://en.wikipedia.org/wiki/Law_of_large_numbers

 

 

And clearly demonstrate that you truly do not understand the drug development process, nor the numbers/statistic involved.

[HappyInVan]

4 minutes ago, nocl said:

And clearly demonstrate that you truly do not understand the drug development process, nor the numbers/statistic involved.

 

 

Explain to us why 150 was chosen, and not 500 (or 15)? 

Edited November 2, 2020 by HappyInVan

[nocl]

1 hour ago, HappyInVan said:

 

 

Explain to us why 150 was chosen, and not 500? 

Do you want to want for 3-6 months longer just to get higher numbers that will not be much more statistically significant. That would mean based upon current numbers world wide between 600,000 and 1.2 million additional deaths before approval.

 

As well as potentially exposing another 350 volunteers to a potentially fatal illness

 

To put it simply the number has been determined to be statistically significant for the clinical trials for a 50% efficacy rate once all of the factors including estimated drop out are taken into account.  That statistical plan has been approved by multiple regulatory authorities including, but not limited to the FDA, the EMEA(Europe) , the MHW (JAPAN), (do not know if HPB in Canada has the vaccines going through their regulatory process, but if they are and I seem to recall that they have placed orders). As well as approved by the experts in the Data Monitoring and Safety Boards over seeing each trial.

 

So one could wait for much larger numbers and you might get some variation in efficacy, but the safety data collection does not change, and while waiting a lot of additional people die.

 

I have been trying to see if anyone has actually published the statistical plan for one of the COVID clinical trials but so far have not been able to get access to one.

 

Here is one paper talking about design and how to determine sample size

 

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3148614/

 

 

Edited November 3, 2020 by nocl

[cheone]

1 hour ago, HappyInVan said:

 

As we can see from this thread, there are numbers of people willing to endorse the vaccine without access to the data. or understanding the limitations.

 

Anyway, I'm sure that the public won't be thrilled to learn that millions of lives will be put at risk because of a study on just 150 (or 32) cases (out of a population of 44k).

 

 

Pfizer had 41,000 volunteers, Moderna had 31,000 volunteers and Johnson and Johnson had 62,000 volunteers.  These 3 Companies will have 100s of million doses available in USA in December.   I don't know where you get 150 cases.  Is that in Canada?

[Paula_MacFan]

9 minutes ago, cheone said:

Pfizer had 41,000 volunteers, Moderna had 31,000 volunteers and Johnson and Johnson had 62,000 volunteers.  These 3 Companies will have 100s of million doses available in USA in December.   I don't know where you get 150 cases.  Is that in Canada?

 

Wait -- I thought you just said we were getting the vaccine the third week of this month?  Now it's December?  Maybe you really just don't know.

[nocl]

1 hour ago, cheone said:

Pfizer had 41,000 volunteers, Moderna had 31,000 volunteers and Johnson and Johnson had 62,000 volunteers.  These 3 Companies will have 100s of million doses available in USA in December.   I don't know where you get 150 cases.  Is that in Canada?

The end point for the efficacy calculations for the Pfizer trial is once 150 volunteers have developed symptoms.  The trial will then be unblinded and the efficacy calculations performed to see if the 50% efficacy lower limit has been reached.  There are also interim looks at 32, 64, 96, and 128 cases.

[Roberto256]

4 hours ago, Steelers36 said:

Another aspect of Warp Speed was the use of logistics experts in the military to get vaccines out to the States and have them in designated centers available to put into use. 

 

Was?    How's that going so far?

 

In the words of Steven Wright ... if you're driving your car at the speed of light, and you put your

headlights on, does it do any good?

 

[Steelers36]

3 minutes ago, Roberto256 said:

 

Was?    How's that going so far?

 

In the words of Steven Wright ... if you're driving your car at the speed of light, and you put your

headlights on, does it do any good?

 

WTH are you talking about?  It is premature to be sending out vaccines as they are still being tested.

[HappyInVan]

3 hours ago, nocl said:

Do you want to want for 3-6 months longer just to get higher numbers that will not be much more statistically significant. That would mean based upon current numbers world wide between 600,000 and 1.2 million additional deaths before approval.

 

As well as potentially exposing another 350 volunteers to a potentially fatal illness

 

 

 

Here's a certainty. That science works best with more data than with little.

 

The fact that vaccine development has proceeded at warp speed should ring alarm bells.

 

As I have discussed earlier, it's not a surprise that Moderna can have a vaccine candidate just a month after Wuhan was quarantined. Western pharma had channelled many billions of investments each year into their projects.

 

The companies with new technologies merely adapted their delivery system to the covid payload. The hardest part for pharma is the testing regime on humans, because of the competitive pressure.

 

The first vaccine that receives approval will be deployed. Possibly ending other projects because their participants could withdraw in order to get vaccinated with the approved candidate.

 

That said, I have no idea what is an 'optimal' case size in this situation. However, we do need to recognize that any review point for unbinding (32/150/500/5000 cases) is necessarily arbitrary.

 

Statistically, the confidence interval will be very wide for a tiny sample at one SD. For example, to be confident of an actual 50% success rate you may need to generate 67% in that sample.

 

At 150 cases, the confidence interval will be smaller. At 500 cases, even smaller.

 

The practical implication is that if a successful candidate only scores 66% in the tiny sample. Due to chance. Do you continue with the trial or abort?

 

On the other hand, a candidate may generate exactly 67% in the tiny sample. Do you apply for EUA, or continue on to a larger sample size?

 

I'm happy to leave the review process to the Advisory Committee of experts. Hopefully, they will publicize a full report. Read the fine print. That's why political interference would muddle the waters and reduce public confidence.

 

As a strategist, I would prefer to continue with the successful best practices. Testing and tracing, distancing and masks, lockdown and quarantines. Until we have the data from several EUA vaccines. Then, we can make an informed choice for the general population.

 

[nocl]

35 minutes ago, HappyInVan said:

 

Here's a certainty. That science works best with more data than with little.

 

The fact that vaccine development has proceeded at warp speed should ring alarm bells.

 

As I have discussed earlier, it's not a surprise that Moderna can have a vaccine candidate just a month after Wuhan was quarantined. Western pharma had channelled many billions of investments each year into their projects.

 

The companies with new technologies merely adapted their delivery system to the covid payload. The hardest part for pharma is the testing regime on humans, because of the competitive pressure.

 

The first vaccine that receives approval will be deployed. Possibly ending other projects because their participants could withdraw in order to get vaccinated with the approved candidate.

 

That said, I have no idea what is an 'optimal' case size in this situation. However, we do need to recognize that any review point for unbinding (32/150/500/5000 cases) is necessarily arbitrary.

 

Statistically, the confidence interval will be very wide for a tiny sample at one SD. For example, to be confident of an actual 50% success rate you may need to generate 67% in that sample.

 

At 150 cases, the confidence interval will be smaller. At 500 cases, even smaller.

 

The practical implication is that if a successful candidate only scores 66% in the tiny sample. Due to chance. Do you continue with the trial or abort?

 

On the other hand, a candidate may generate exactly 67% in the tiny sample. Do you apply for EUA, or continue on to a larger sample size?

 

I'm happy to leave the review process to the Advisory Committee of experts. Hopefully, they will publicize a full report. Read the fine print. That's why political interference would muddle the waters and reduce public confidence.

 

As a strategist, I would prefer to continue with the successful best practices. Testing and tracing, distancing and masks, lockdown and quarantines. Until we have the data from several EUA vaccines. Then, we can make an informed choice for the general population.

 

That said, I have no idea what is an 'optimal' case size in this situation. However, we do need to recognize that any review point for unbinding (32/150/500/5000 cases) is necessarily arbitrary.

 

is absolutely not arbitrary.  For example at the 32 measure point. the 26 and 6 numbers that they said would be needed at that look would be a 77% efficacy rate.  That measurement at that point be sufficient to generate the appropriate power and P factor that the end result would reach the minimum 50% efficacy when the fully 150 i reached.  The 64 interim readout would be successful at an efficacy below 77% (probably around 72% but without the stat plan do not have all of their assumptions) and so on with the 96 interim readout and the 128 interim readout until you reach the 50% being the minimum number needed at 150.

 

If at the final read out at 150 any efficacy over 50% would be successful, at 32 anything below 77% would not be sufficient. The numbers are pretty rigid because they are driven by the calculations for that trial.  Anything below 77% at 32 would fall below the required confidence calculation.

 

You may be a strategist, but you are neither a clinical statistician, nor an expert in clinical trials. So continue on with your criticism of the process and certainly you would want to avoid the vaccine. 

 

The numbers are the same if they were working on a vaccine and had years to complete it.  The difference might be that they would only enroll a much smaller number in the trial instead of 30,000.  The size of the trial is largely driven by the  incidence of the illness in the population and the speed by which one wants to get to the number of cases necessary for determining efficacy.  The calculations are well developed and well accepted by the professionals involved with such work.

 

As far as this question "The first vaccine that receives approval will be deployed. Possibly ending other projects because their participants could withdraw in order to get vaccinated with the approved candidate." The other trials will continue after they may have better results.  There will not be enough drug until well after the first round trials are over, plus most of the volunteers have committed to the trial timeframe and most will have the ethics to stick to that commitment.

 

For a person who likes to state that you only listen to scientists you certainly seem willing to make up your own reasoning when the experts are saying something that disagrees with your view. The experts certainly have not complained about the clinical trial design as far as end point and sample size.

 

So I guess you I only listen to scientists really means when they agree with you.

 

Edited November 3, 2020 by nocl

[HappyInVan]

7 minutes ago, nocl said:

is absolutely not arbitrary.  For example at the 32 measure point. the 26 and 6 numbers that they said would be needed at that look would be a 77% efficacy rate.

 

 

Actually, you seem to have rushed to your own conclusion. What I meant was that the choice of numbers is arbitrary. It could have been 50/100/200 etc. Of course, the statistical requirement is well known. (Tables are available.) As well as the possibility that confidence is not the same as certainty?

 

 

Edited November 3, 2020 by HappyInVan

[nocl]

1 minute ago, HappyInVan said:

 

Actually, you seem to have rushed to your own conclusion. What I meant was that the choice of numbers is arbitrary. It could have been 50/100/200 etc. Of course, the statistical requirement is well known. As well as the possibility that confidence is not the same as certainty?

 

 

They could have used other number except for one factor time.  The read outs were chosen with the intent of identifying the reasonable points where sufficient data exists to determine efficacy that meets the statistical requirements.  Could they have waited until 64 for the first read out.  Certainly, but they would give up a month or two (which means another 20-30,000 deaths world wide) in exchange for being able to get approval with a few points lower efficacy rate. The trial sizes are calculated with time as being a factor, the interim readouts also take time into account, along with the statistics that drive the necessary power, error and statistical significance calculations.

 

In the end the entire purpose of this vaccine development effort is to save lives.  As several people here on CC that are in the industry and have worked with vaccine development this has been an all hand on deck effort because of both the cost in human lives as time passes, as well as the economic impact this relatively unchecked outbreak is having.  The time the vaccine is available will clearly impact lives and economic loss, even if it is not a silver bullet and it will probably not be.

 

Just as delaying infection has resulted in 30% better survival today than 4 months ago.  The availabilty of a vaccine that meets those targets will also save lives.

[HappyInVan]

5 minutes ago, nocl said:

They could have used other number except for one factor time.  The read outs were chosen with the intent of identifying the reasonable points where sufficient data exists to determine efficacy that meets the statistical requirements.  Could they have waited until 64 for the first read out.  Certainly, but they would give up a month or two (which means another 20-30,000 deaths world wide) in exchange for being able to get approval with a few points lower efficacy rate. The trial sizes are calculated with time as being a factor, the interim readouts also take time into account, along with the statistics that drive the necessary power, error and statistical significance calculations.

 

 

 

I see that some states have a positivity rate of >10%. Some states have no intention of controlling the disease. Good luck to you.

[nocl]

12 minutes ago, HappyInVan said:

 

 

I see that some states have a positivity rate of >10%. Some states have no intention of controlling the disease. Good luck to you.

Which has absolutely to do with the proper design of a clinical trial

[HappyInVan]

31 minutes ago, nocl said:

Which has absolutely to do with the proper design of a clinical trial

 

I meant the real world decisions by governments and pharma.

[caribill]

2 hours ago, nocl said:

Which has absolutely to do with the proper design of a clinical trial

 

OK, I admit I am not an expert (or anything close to one) in vaccine testing, but I thought that past vaccine (polio, for example) trials worked as follows:

a) Some people received the vaccine, some received the placebo.

b) Over a period of time, the number of actual cases of disease that occurred were looked at and it was determined how many that developed the disease had the vaccine and how many had the placebo.

c) If the number of people who had the disease and also had the vaccine was appropriately lower than those with the disease and a placebo, the vaccine was deemed effective. For example, to make up numbers, if 1% of those who had the vaccine had the disease but 10% of those with the placebo had the disease, the vaccine might be considered very effective.

 

If I understand you correctly, the corona virus vaccines being tested are not doing this. They are not doing periodic testing of everyone in the trial to determine how many get Covid-19 and how many do not. They are just looking for those who develop obvious symptoms. Anyone who gets the virus and never shows symptoms would not be detected this way and, unlike a number of past diseases that have had vaccines developed, it is possible that many people catch the virus, never show symptoms, but can spread the virus.

 

Is my understanding correct? If so, I do not understand how true effectiveness can be measured as it will not be determined if the vaccine was effective in reducing the cases of people who never show symptoms.

[Roberto256]

13 hours ago, Steelers36 said:

Another aspect of Warp Speed was the use of logistics experts in the military to get vaccines out to the States and have them in designated centers available to put into use.

 

8 hours ago, Steelers36 said:

WTH are you talking about?  It is premature to be sending out vaccines as they are still being tested.

 

Yes, it certainly would be.    The tense of your original post was that the military had already gotten the vaccines to the states.

 

Note the use of 'was'.

 

 

[Steelers36]

3 hours ago, Roberto256 said:

 

 

Yes, it certainly would be.    The tense of your original post was that the military had already gotten the vaccines to the states.

 

Note the use of 'was'.

 

 

Ahh, I see your point.  Should have used "is".

[darknightsdespiser]

Anytime! Princess cruise ships are cleanest places I ever go!

[nocl]

7 hours ago, caribill said:

 

OK, I admit I am not an expert (or anything close to one) in vaccine testing, but I thought that past vaccine (polio, for example) trials worked as follows:

a) Some people received the vaccine, some received the placebo.

b) Over a period of time, the number of actual cases of disease that occurred were looked at and it was determined how many that developed the disease had the vaccine and how many had the placebo.

c) If the number of people who had the disease and also had the vaccine was appropriately lower than those with the disease and a placebo, the vaccine was deemed effective. For example, to make up numbers, if 1% of those who had the vaccine had the disease but 10% of those with the placebo had the disease, the vaccine might be considered very effective.

 

If I understand you correctly, the corona virus vaccines being tested are not doing this. They are not doing periodic testing of everyone in the trial to determine how many get Covid-19 and how many do not. They are just looking for those who develop obvious symptoms. Anyone who gets the virus and never shows symptoms would not be detected this way and, unlike a number of past diseases that have had vaccines developed, it is possible that many people catch the virus, never show symptoms, but can spread the virus.

 

Is my understanding correct? If so, I do not understand how true effectiveness can be measured as it will not be determined if the vaccine was effective in reducing the cases of people who never show symptoms.

Yes, that is the weakness of the current trial.  The efficacy numbers for these tests will be for prevention of symptoms.

 

A couple of comments.  I have not researched how polio vaccines(either Salk or Sabin versions) were tested. While the double blind, placebo controlled trials are the gold standard there are other ways that one could test.  For example one could vaccinate a large number of volunteers, without a control group and then compare the infection rate in the test group vs the general population.  It is less accurate, takes more time, and needs more cases than the placebo controlled method, but there are other ways.  

 

Placebo controlled, double blind studies give the best data, in the shortest time, with the fewest cases needed.

 

Yes your description of steps A,B,C,D do describe how a placebo controlled study works.  In your example the efficacy rate would be  ( (10-1)/10) x 100 = 90% efficacy rate.

 

Yes you have captured the issue with these trials.  The study design does not identify any asymptomatic cases.  So unfortunately it is not collecting data on the number of those cases and/or if they are infectious.  It is not a matter that the vaccine might not prevent such cases, the issue is that we do not know and the trial by itself will not tell us.  It could be that the vaccine prevents asymptomatic cases, it could be that there might be asymptomatic cases, but they are not infectious, or they could exist and be infectious.  The study as designed will not tell us.

 

We don't know why they chose this approach.  Covid is somewhat unique in the asymptomatic cases.  In many other disease infection = symptoms and it is not an issue, though there are rare examples of carriers (typhoid Mary for example). Testing for infection would certainly be more difficult and costly.  Since you would periodically have the participants come in and get tested. Considering the tests involve a swab up the nose and is not comfortable, between that and having to come in for the test you would get a much higher drop out rate.  If you were testing weekly for example if you used grouping of 100-1 (assuming most tests negative) you would need to collect over 30,000 samples and run over 3000 analysis per test period. Not cheap.

 

I am surprised that while these trials are running that they have not started another, smaller trial that looks for asymptomatic cases.  It could be that they do not want to introduce delays and potential confusion with the results of the main trial.

 

In my mind the problem with not detecting infection, is that you do not know who may have asymptomatic disease, and as a result you can not test them for secondary impacts (there have been studies showing cardiac damage even with no other symptoms for example) and and virus shedding.

 

As I understand it Oxford did a challenge trial with primates and that in those trials the primates did not develop symptoms, but that there was virus present in the nasal passages. Not sure if that is influencing trial design or not.  It also was not clear from those trials if it was just residual from the virus administration with no replication.

 

The vaccine advisory committee did spend some time discussing the infection vs symptoms issue during their October 22 meeting and did identify it as a weakness and spent some time discussing how to address it.

 

As far as safety one famous vaccine developer once stated that he only relaxed his concern about side effects after the 3 billionth dose was administered.

Edited November 3, 2020 by nocl