CrowdSpeak: power calculations for Charcot

More about our crowdfunding initiative: the importance of power. #MSResearch #CrowdSpeak #CrowdacureMS

"What have we learnt this week about infectious mononucleosis (IM), Epstein-Barr virus (EBV) and MS? 

  1. People who develop IM due to EBV infection you are more than twice as likely to develop MS than people don't have a history of IM. 
  2. People who don't get infected with EBV are protected from getting MS. 
  3. If you have been tested EBV negative and go onto develop MS you alway get infected with EBV prior to developing MS. 
  4. People infected with EBV shed EBV in their saliva intermittently that may vary with the season. 
  5. EBV is transmitted via saliva, which is why it referred in lay terms as the 'Kissing Disease' and more uncommonly by other body fluids. 
  6. Peak age of onset of IM is 3-4 years earlier in females (13-14 years of age) compared to males (16-18 years of age). 

"As part of The Charcot Project we are proposing using viral shedding of EBV as a readout for antiviral drugs targeting EBV. Before we can do this we need to know how many pwMS shed EBV in their virus serially over several weeks. Somebody made the comment that we don't need to do this because we have this data already from publications on normal people. I countered that if we relied on data from normal people we get things wrong in terms of power calculations of our study. In addition, a large number of pwMS we will study will be on disease-modifying therapies (DMTs), in particular interferon-beta (IFNbeta). IFNbeta is an antiviral drug and may suppress EBV viral shedding. Finally, will the peer-reviewers' of our grant accept old data on EBV shedding from normal people? May be but having contemporary data from pwMS, who we will be studying, using assays run in our own lab will be so much better and increase our chances of getting the next phase of the Charcot project funded."

"What are power calculations?"



"The power simply refers to the probability that when we do a study we correctly reject the so called null hypothesis, in our case that a certain dose of famciclovir is ineffective in suppressing EBV shedding in the saliva and accept the alternative, or true hypothesis, that a particular dose is effective at suppressing EBV shedding. As the power increases, there is decreasing chances of a so called type II error (false negative) or getting a negative result by chance. A similar concept is the so called type I error, also referred to as the false positive rate; getting a positive result by chance. Both a false negative and positive study result can be catastrophic in science and for pwMS. A false negative result means that an effective drug is deemed ineffective. In comparison a false positive result is we assume a drug is effective when it is actually ineffective. The latter is why the FDA and EMA typically require two or more positive phase 3 trials to limit the chance of false positive results. Power calculations are therefore necessary to try and limit the chances of getting false negative and positive results and to make studies efficient. Why spend extra money making studies larger than they have to be? There are ethical issues as well; why recruit and expose people to a drug when the study won't give a definitive result? Power calculations are used to calculate the minimum sample size required so that one can be reasonably confident to detect a true effect of a given size. 

Example: Let's say 40% of pwMS shed EBV in their saliva at any one time, but only half of these pwMS (20%) will continue to shed virus in weeks 3 and/or 4. We are trying to test whether FamV at dose x suppresses EBV shedding by 90% at both weeks 3 & 4, compared to placebo. How many pwMS do we need to recruit into this study to have 95% confidence of the result with a type 1, or false positive, error rate of less than 5%? As half  of pwMS on placebo will simply stop shedding EBV by chance (40% to 20%) in the FamV arm EBV shedding will have to be reduced to less than equal to 2% (10% of 20% = 2%) of subjects, for us to reject the null hypothesis and accept that that dose of FamV is effective as defined by the power of the study using our assumptions. Please note we are proposing that our assumptions will be based on hard data. 

  • Estimated reduction of EBV shedding in placebo-treated pwMS = 50% (40% to 20%)
  • Proposed reduction of EBV shedding in FamV treated pwMS (compared to placebo) = 90% (20% to 2%)
  • Controls per subject treated with FamV = 1
  • Alpha = 5% (chance of being a false positive result by chance)
  • Power = 90% (chance of being a real result - true positive, or 10% chance of being a false negative result)
  • Using the above assumptions and making some statistical corrections for small numbers in the Fam-V treated group we will need to recruit at least 73 subjects per group (144 pwMS). 

We simply can't do these sorts of calculations without have our own data on EBV shedding in the saliva of pwMS. The data in the literature is simply not adequate for our purposes. This is why we are asking you to help us raise the money to do the lab work for this study ASAP. I sincerely hope you can help. If everyone who visited this site donated £1 each time they visited we would raise the money we need in 1 or 2 days. I would like to thank all our supporters who have already donated so kindly; things are going well. Thank you."

Click here to find out more!

"If this work inspires you in any way please send the link to anyone you know who has a personal link with MS. I can't tell you how important it is for us not give up on the viral hypothesis of MS."



CoI: Team G will be recipients of a grant from Crowdacure to perform this research

Labels: , , , ,