I hope everyone is doing well. With some luck I managed to get the email back for this blogger account. That makes 8 years. A lot happens in 8 years... my last post was about Mitt Romney's 2012 campaign against Barack.
It is a little bit like getting back in touch with a High School friend. So much time has past that you can't help but be curious to see how they have fared.
I analyze data for drug companies, devices, diagnostics, and similar such stuff. I've been meaning to do this on Facebook, but I kind of boycott Facebook because it is used to spread lies and hate messages for the profit of Mark Z. So, ah, no.
So my public service work is to point out that when the FDA approves something it publishes a summary basis of approval or a summary of safety and effectiveness. They have varying levels of details and often results will be included in the labeling materials for a product. The FDA is kind of picky about what makes it into a label, so typically you can look for publications with the main study results for clinical trials.
The work done on clinical trials is regulated. The study sites that participate in studies can be audited and sanctioned if they are not following rules well enough. If you work on trials in the US, the government can show up and audit you. If you happen to be involved in the analysis of data for a pivotal clinical trial for publicly traded company and it looks like there is suspicious trading, the government can ask you to share the names of everyone who knew the results before they become publicly shared. I'll skip the other rules that are in place to make sure what the FDA is getting honest data and honestly reporting the results, but the point is ... the information the FDA is providing goes through a lot more vetting and rigorous protection than most of the stuff that shows up on the web. Especially the dark web and internet chat boards.
In my work, I have worked on many approved products and in all of them the FDA's summary documents replicated the analyses tables that were provided to them. There was never anything going on. The same goes for publications in medical journals of the study results.
The Moderna Vaccine Emergency Use Authorization (EUA) is here.
Skip right to Page 20 - it covers the study disposition. This covers the number of subjects who entered the study, were randomized, and if they completed the study. This includes more than 30,000 subjects. A per-protocol population was defined and included 27,000 subjects. A per-protocol population typically excludes subjects who deviate from the planned study assessments in some significant way.
If you look at Page 29, the results of an interim analysis of the per-protocol population was completed. This is interesting because the confirmed cases was quite low when in both arms (5 cases in 13,934 person years in the vaccine arm and 90 in 13,883 person years). This surprised me as I expected the rates would be higher. When Moderna reports a 94.5% vaccine efficacy they are talking about the reduction in this rate of cases). The rate in person years is dependent upon exposure in the populations and the comparison assumes that people getting randomly exposed is similar in the groups. Note this is a blinded trial, so participants will not know what they received.
The final analysis was 11/13,914 and 185/13,883. If you ever had a statistics class you will know a p-value is hovering around. A p-value says how likely the relative difference is to occur by chance. The p-value for this comparison was <0.0001. That means the 11 vs 185 events would happen by chance is less than that 1 in 10,000. Here is a link to a John Hopkins site that at mortality in COVID infected individuals. Much of this appears higher than 1 in 100 in many countries. Given that many people who are not symptomatic likely do not get tested, the actual mortality must be lower.
Is it safe? There are clearly more events and higher grade events when subjects were solicited for them and more injection site reactions. Serious related events and deaths were similar in both arms, so from 30,000' it looks safe in terms of the most serious results by not without some additional risk of an event. Assessing risk benefit is something you should do with your doctor if you are concerned.
If you ever really want to find out for yourself what a drug, biologic or device looks like, start with fda. gov and work out from there.
Stay safe.
No comments:
Post a Comment