@PypAyurved @omlakhani I am not going in debate regarding paid and non paid and I encourage such efforts and hope Ayurvedic medicine feels the pressure and come with more studies. Having said that this article has many flaws and with this article (first link) nothing is proved as under.@nirajvasavada
@PypAyurved @omlakhani @nirajvasavada First sample size is very small and non calculated and 50 and 45 patient in each group gives this article no statical power (1-beta error) so very very high chance of false positive results.
@PypAyurved @omlakhani @nirajvasavada Second ANOVA test is not used to compare two groups, it is used to see effects of two categorical values on third group with continuous value here we are comparing two categoriecal value only day 3 and day 7 RT pcr positive or negative in treatment and placebo arm.
@PypAyurved @omlakhani @nirajvasavada Third Kaplan Meier is used for effect over time for that we need to see mean recovery time in treatment and placebo arm and not day 3 and day 7 rt pcr yes or no.
@PypAyurved @omlakhani @nirajvasavada Fourth both group has all normal blood parameters 99 degree saturation, every thing was fine then why were we treating them? For those patient we need to look at how many patients worsened in each group and then compare, and not rt pcr.
@PypAyurved @omlakhani @nirajvasavada If we want to see effect on viral replication then we need to see viral load. This article has serious methodology flaws. These methodology is not suitable for any study neither aayurved nor modern or allopathy. Highly flawed methodology.
@PypAyurved @omlakhani @nirajvasavada Rest of the studies are either preclinical or animal/fish studies so though may be important we can not approve any drug based on such studies on humans. If such studies were there for allopathic drugs then also I would critisize them just like I had criticised hcq, ramde, toci
@PypAyurved @omlakhani @nirajvasavada This is where two factor annova is used and not in 2*2 tables like day 3 rt pcr +\- vs treatment group yes or no and same way day 7 rt pcr +\- vs treatment group +\- so because of that tables in said manuscript does not make any sense. @nirajvasavada 
@PypAyurved @omlakhani @nirajvasavada Spss can do analysis from what data we input but which data to enter where is real skill.
@PypAyurved @omlakhani @nirajvasavada Plus in table in rct comparing drug you need to mention whether they were on other drugs or not.
@PypAyurved @omlakhani @nirajvasavada Here unfortunately our outcome was categorical (rt pcr +ve or negative) and not continuous and two groups were also categorical. Plus we were comparing two groups and not effect of two groups on third continuous end point.
@PypAyurved @omlakhani @nirajvasavada Okay now let me correct it for the science community if we use chi square here (which should be used) still p value is 0.04. Small effect size hardly significant but good right? But there is a problem cont.. @nirajvasavada
@PypAyurved @omlakhani @nirajvasavada Insmall sample size to decrease false positive rate we need to use Yates correction in chi square test and with Yates correction p value comes to 0.06 and results become non significant. So I would suggest do larger sample size study with good methodology please.@nirajvasavada
@PypAyurved @omlakhani @nirajvasavada Alpha error (false positive) should be accepted as 5% and beta error at 20% (false negative) to calculate sample size.
@PypAyurved @omlakhani @nirajvasavada You can see log rank test compares entire survival experience and whether curves are identical or not. When Kaplan Meier curves are flawed log rank test automatically becomes flapped and understand the word “entire survival experience” that means you need to measure entire time
@PypAyurved @omlakhani @nirajvasavada See here log rank test is used when time to occurrence of event is used rather then just whether even occurred or not, what it means you need to see how many patient become negative on day,1,2,3,4,5,6,7 and not just day 3 and day 7 rt pcr positive or not.
@PypAyurved @omlakhani @nirajvasavada Also conclusion is flawed. All were asymptomatic/mildly symptomatic so we can not conclude recovery without defining recovery. Inflammatory marker data it self says finding was non significant so we can not conclude inflammatory markers were reduced.
@PypAyurved @omlakhani @nirajvasavada One thing more regarding anova using spss (paper said spss version 23 used) you need to have six assumptions. And check first assumption your dependent variable should be continuous and not categorical as used in the study (rt pcr positive or negative). Wrong at first level.
@PypAyurved @omlakhani @nirajvasavada There is more on further analysis odds ratio, there is further thing 32/45 turned rt pcr negative in treatment group at day 3 and 25/50 in placebo group and alas odds ratio is not 0.41 which study mentioned but 0.70 means 30 percent risk reduction and not 60%.
@PypAyurved @omlakhani @nirajvasavada And it is not the end, 95% confidence interval for odds ratio is on both side of 1 that means this finding is not significant. I have used same online calculator. Which they mentioned they used.
@PypAyurved @omlakhani @nirajvasavada For day 7 also odds ratio is 0.6 and not 0.05 which paper mentioned and here also 95% confidence interval on both sides of 1 so again non significant. Taken 100% rt pcr negative in treatment group (all 45) vs 60% in placebo group (30/50). So this is a no n significant finding.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
