Top Tweets for #stat101
🧵#Stat101: The value of simplicity! Simplest statistical test an example👇
➡️Works if randomization is 1:1 (≈equal)
➡️Works for most published RCTs
➡️Quick, useful, reliable "feel of data"
➡️NOT replacement for more conventional statistical tests (e.g., Log-rank or X2 test)
🧵Check this brilliant article by @pocock_stuart on the Simplest Statistical Test !
https://t.co/0b7cTGrr90
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics #biostatistics #math #CancerResearch
🧵#Stat101: How about the landmark "DAPA-HF" Trial @NEJM⤵️
➡️Randomized≈ 4,700 pts: Dapa vs placebo
➡️1ry Outcome: death or HF hospitalization
➡️Dapagliflozin: 386 events. Placebo: 502
🔴HR= 386/502= 0.76
🔴Z= 116/30= 3.87
🔴P-value= 0.0001 (from Z table)
✅HR reported by authors using complex statistics was the SAME: 0.74 (p<0.001).
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics #biostatistics #math #neurotwitter #IDTwitter
🧵#Stat101: Let's apply the 'simplest test' to PROACTIVE Trial @TheLancet !
➡️Randomized≈ 5K pts: Pioglitazone vs placebo
➡️Outcome: composite mortality, MI, ACS, stroke
➡️Pioglitazone: 514 events. Placebo: 572 events
🔴HR= a/b= 514/572= 0.90
🔴Z= (a-b)/Sq(a+b)= 58/33= 1.76
🔴p-value= 0.08 (from Z table)
🧐Simplest method yielded SAME HR published by authors (Log-Rank test)= 0.90.
- HR exactly the same= 0.090
- P-value VERY CLOSE!
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics #biostatistics #math #neurotwitter #IDTwitter
🧵#Stat101: Let's apply the 'simplest statistical test' to the VIGOR Trial @NEJM!
➡️Vigor randomized 8K pts: Rofecoxib vs Naproxen
➡️1ry outcome was major thrombotic CV events
➡️Rofecoxib had 45 events; Naproxin 19 events
➡️Using Simplest Test:
🔴HR = a/b= 45/19 = 2.37
🔴Z= (45-19)/Square(45+19)= 3.25
🔴From Z table👉p-value= 0.001
For real🧐? Compare the results from "simplest test" to the complex statistical tests (e.g., Log Rank) ⤵️
VERY CLOSE!
HR= 2.38 & P<0.01
SEE FIGURE BELOW
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics #biostatistics #math #neurotwitter #IDTwitter
🧵#Stat101: Did u know u can estimate Hazard Ratio (HR) & p-value for most clinical trials in under 15 seconds👉Heard of "Simplest Statistical Test" before? ⬇️
✅Assume RCT with 1:1 randomization⤵️
✅Ignore sample size & F/U duration
✅(a) event rate in control (b) event rate in Tx group
🔴HR = a/b (estimated)
🔴Calculate Z= (a-b)/Square(a+b)
🔴Estimate P-value from Z table below👇
For real🧐? Let's apply these examples to some real & published clinical trials!
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics #biostatistics #math #neurotwitter #IDTwitter
📌 #Stat101: Below 3 trials with "statistically significant" treatment effect based on p-value <0.05. However, they all provide different info:
✅PARADIGM-HF: Overwhelming evidence (<0.001)
✅Champion: strong evidence (<0.01-0.001)
✅Improve-It: some evidence (0.01-0.05)
But is the interpretation of the treatment effect in these trials is different? 👇
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101 Bottomline👇
🛑P-value is not "black & white" but shades of gray!
🛑The smaller the P, the stronger the evidence
🛑Alway report actual p-value not <0.01
🛑Don't obsess about classifying trials into "positive" or "negative", based on p-value of primary outcome
Great reference ➡️
https://t.co/YJtzNqREee
📌 #Stat101: How about P-value of 0.41?
✅Astronaut Trial: Aliskirin vs Placebo in Acute HF
➡️HR 0.92 (95% CI: 0.76-1.12), p=0.41
🛑P=0.41: with such clearly non-significant p-value, we can firmly declare there's "no evidence" that Aliskiren has a treatment effect.
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101: A p-value of 0.06 versus 1.2 are not the same!
🛑P=0.06: this shouldn't be interrupted dogmatically as "X has no treatment effect".
➡️Rather we should declare there's insufficient evidence that's X has treatment effect.
➡️X may have a modest effect or may have NO effect. The data is inconclusive.
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101: The smaller the p-value, the stronger the evidence against the null hypothesis.
✅In IMPROVE-It trial (Ezetimibe vs. Placebo)
➡️OR 0.94 (95% CI: 0.89-0.99), p= 0.016
🛑p=0.015 means there is 1 in 50 chance of such difference arising by chance!
🛑This provides "some evidence" of "modest" Tx benefit! But not as overwhelming as in PARADIGM-HF!
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101: The smaller the p-value, the stronger the evidence against the null hypothesis.
✅In Paradigm-HF (Sac/valsartan vs Enalapril)
➡️HR of 0.8 (95% CI: 0.73-0.87), p= 0.0000004
🛑The chance of getting this magnitude of treatment difference by chance is <1 in million ➡️Such small P provides "overwhelming evidence" of Tx difference.
🛑Such proof beyond reasonable doubt means one can confidently assert that Sac/val is superior to enalapril
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101: Below 3 trials with "statistically significant" treatment effect based on p-value <0.05. However, they all provide different info:
✅PARADIGM-HF: Overwhelming evidence (<0.001)
✅Champion: strong evidence (<0.01-0.001)
✅Improve-It: some evidence (0.01-0.05)
But is the interpretation of the treatment effect in these trials is different? 👇
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
📌 #Stat101: Calling trials as "positive" or "negative" based on p-value <0.05 is an abuse👇
✅The smaller the "P", the stronger the evidence against the null hypothesis.
🛑Always report actual p-value (0.004) not (<0.05).
🛑P-value of 0.04 & 0.0001 are not the same
🛑Always use two-sided p-values
🛑P-value is not dichotomous (significant or not)
#CardioTwitter #MedTwitter #MedEd #cardiology #medicine #health #research #CardioTwitter #statistics
#QuizOfTheDay: In #Statistics, what is Collection, presentation, and description of sample data called? #Stat101 #NationalStatisticsDay
A. Inferential Statistics
B. Qualitative Variable
C. Quantitative Variable
D. Descriptive Statistics
https://t.co/3aVb59tW7j
#stat101 especially statistical inference and sampling were thrown under the bus in spring 2020 indeed
look out your window.mp4
Correlation explained in a way everyone can understand!
Includes definitions of latent variables, spurious correlation, and that hilarious moment when investors bought the wrong ZOOM stock...
#DataScience #statistics #rstats #models #STAT101
https://t.co/NC418sVHQu
No! US needs valid and reliable tests for small-n random sampling #STAT101
Wingardium Leviosa!
Mengenang setahun lalu ada politisi yg klaim bahwa statistik adalah #sihirsains dan orang2 yg menolak data2 yg tak sesuai dgn 'ego' mereka.
#stat101
"Modern Statistics: the myths and the magic"
https://t.co/vZ4pb0s74d
It will seem like a 0.22% increase in recovered, but it's an "artificial increase". It's insignificant because in both days 2 and 3, the same number of people recovered.
Diyan papasok ang "Test of Significance", #Stat101 😅 kaya kailangan nirereport ang base palagi.
Last Seen Hashtags on Sotwe
Trends for you
Most Popular Users
Elon Musk 
@elonmusk
241.7M followers
Barack Obama
@barackobama
119M followers
Cristiano Ronaldo
@cristiano
114.4M followers
Donald J. Trump
@realdonaldtrump
111.9M followers
Narendra Modi
@narendramodi
107.2M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.4M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
90M followers
Taylor Swift
@taylorswift13
84M followers
Lady Gaga
@ladygaga
75.4M followers
Virat Kohli
@imvkohli
73.4M followers
Kim Kardashian
@kimkardashian
70.9M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.4M followers
Bill Gates
@billgates
65.2M followers
Selena Gomez
@selenagomez
63.1M followers
The Ellen Show
@theellenshow
62.3M followers
CNN
@cnn
61.8M followers
X
@x
60.7M followers