Bayesian hypothesis testing: which treatment should we choose to reduce mortality rate?
Using a Bayesian network model of a typical problem in medical trials, this provides a different and more infortmative analysis than you get out of a traditional null hypothesis p-value test ... https://www.youtube.com/watch?v=R9QS1n3DrOA
This 2 minute video presents a simple graphical explanation of why it is som important to get an accurate estimate of the true proportion of unvaccinated in a population in order to determine the efficacy and safetyof any vaccine (or indeed any type of medical intervention)
...
https://www.youtube.com/watch?v=LxGSMdE_TN0
Using a graphical explanation of Bayes’ Theorem we show that that if you test positive there is only a 1 in 6 chance you have the virus.
...
https://www.youtube.com/watch?v=jDecehzaVQU
3-minute trailer from Presentation to UK Doctors For Patients 14 September 2022. Full 30 minute presentation is here: https://www.bitchute.com/video/xeZO0vOyOSEa/
...
https://www.youtube.com/watch?v=ahFmHri28rk
Another misclassification problem hiding possible adverse outcomes inn pregnancy?
Instead of comparing bad outcomes (like miscariage and stillbriths) for vaccinated v unvaccinated women UKHSA grouped all unvaccinated with those vaccinated prior to (but not during) pregnancy into a single "no doses in pregnancy" category. So, in comparing pregnancy outcomes this "no doses in pregnancy" category was used as a surrogate for “unvaccinated”. Laura Kenny was fully vaccinated before August 2021(before she was pregnant) and then suffered 2 'bad outcomes' (miscarriage, Nov 2021 and ectopic pregnancy, Jan 2022). This means that Laura Kenny's bad outcomes would have been classified as part of the "no doses in pregnancy" category. The For full details:
See: https://wherearethenumbers.substack.com/p/laura-kenny-and-the-vaccinated-olympic
...
https://www.youtube.com/watch?v=fiFjhLyDlZE
Whta happened after I was invited to give a talk on Bayesian networks to a big NHS conference of Health and Care Analytics.
Full details here: https://wherearethenumbers.substack.com/p/blasphemers-begone
...
https://www.youtube.com/watch?v=ULa1P5b6SQs
In determining the efficacy of a medical intervention (such as a drug or vaccine) to stop a particular disease or virus it is typical to assume that the treatment needs time to work before a person is classified as 'treated'. For example, a person vaccinated against a virus may be classified as 'unvaccinated' until 2 weeks after getting the vaccination. This simple animation with a hypothetical example shows that, with such a classification, a placebo (i.e. no effect) vaccination can be shown to be highly effective.
...
https://www.youtube.com/watch?v=a4LcPF87YLE
If both handling explosives and playing cards are equally likelty to result in a positive test, most people assume the positive test result is not probabtive in support of the hypothesis that the suspect handled explosives. This video shows using simple Bayesian networks why this may not be correct
...
https://www.youtube.com/watch?v=0oflOdl1TIg