Thread by @Neurophysik

T1: Thread on data analysis & representation emerging from cognitive bias associated with research studies involving serious malpractice in #altmed #naturopathy #ayurveda #AYUSH research, i.e. no placebo controlled, no blinding, incorrect statistics, no error bars on mean data.

T2: From the recent article published at @AltNews Science on dengue and papaya extract pill called Caripill altnews.in/is-caripill-th… we evaluated the two clinical trial studies in India given by pubmed

T3: The first study had serious misrepresentation of data : while there is no way to find out the raw data & conduct statistical analysis, I replotted the figure using Mean +_ standard error of mean given in the results section. ncbi.nlm.nih.gov/m/pubmed/27739…

T4: Even though elevation of platelets does not mean a recover from dengue: From the link above, the study suggested that Caripill normalised platelets from day 4 as compared to control (no drug).

T5: when I used the data and replotted the same numbers using my trusted analytical software. The error bars overlapped each other on each day
This suggests that even though the mean was higher in Caripill group, a large proportion of subjects recovered in control group as well.

T6: Also, that a large number of people did not recover in the Caripill group at day 4, as with the control group. The average platelet count also dropped at day 5-6 with Caripill to come close to the control group data.

T7: More importantly, the control group was not a placebo group where a sham pill should have been given. And the patients on Caripill knew what they were taking through the consulting clinic. This hugely changes drug expectations.

T8: This study wasn't conducted by some junior medical officer. The last author (most important one) is the professor and HOD of medicine from SP medical college & Mahatma Gandhi medical college, Rajasthan.
So beware of sham conclusions & look out for details in methods+results!!