Tweet

More from @erlichya

Yaniv Erlich

@erlichya

27 Sep

https://twitter.com/erlichya/status/1438612458353070084?s=20

covid19 in 🇮🇱.

Due to ✡️ holiday season, it is hard to get a reliable estimator of R. Previously, I showed that the avg susceptiblity of ISR population (S) is nicely correlated with R. This estimator is not affected by # test but only by # vaxxed and VE.

https://twitter.com/erlichya/status/1438612458353070084?s=20

As a reminder, S(t), the susceptiblity of day t is 1 - Σf_i(t) •VE_i(t). Where f_i(t) is the % of the population that already took the i-th dose and VE is the efficacy of the dose in day t. Both parameters can be in calculated from MoH public dashboard on a daily basis.
So...

So here is S as a func of time (black) versus smoothed R (red). Both are coorelated.

Now, we don't have good data for R in Sep, BUT due to booster, S is in levels similar to Mar 21 (60%)! At that time, it was enough for R=0.9, so I'm cautionally encouraged that we= heading out😍

Read 4 tweets

Yaniv Erlich

@erlichya

16 Sep

Israel 🇮🇱, booster 💉💉💉, and herd immunity 🐑.
A thread based on the Ministry of Health data with many graphs.
Let's dive!
👇👇👇

Many of you heard about the decline of vaccine effectiveness in Israel in the past few months. The MoH publishes now on a daily basis the rate of infection in unvaxxed, vaxxed (2nd dose), and boosted individuals per 100K, which is a real treasure!
nature.com/articles/d4158…

With this data, we can determine the **crude** VE of each group by dividing the rate of infection in second dosed individuals to unvaxxed. From Jan 2020 to now, this is the crude VE (crude = not adjusted to sex, age, and other demographic variables)
👇👇👇

Read 10 tweets

Yaniv Erlich

@erlichya

15 Sep

שרשור על חסינות האוכלוסיה בישראל והתפשטות התחלואה. איפה היינו ולאן אנחנו הולכים.
יהיה גרפים 📊. ויהיה מעניין!
🧵👇🧐

עם הדאטה בדאשבורד אפשר עכשיו לחשב בכל יום החל מינואר את יעילות הגולמית של מנה 2 לעומת לא מחוסנים. מה אנחנו רואים כאן בעצם? זה פשוט יחס המאומתים בין מחוסנים 2 ללא מחוסנים בכל יום לאחר נירומול לגודל האוכלוסיה. זאת יעילות גולמית כי היא לא מתקוננת למגזר או גיל (אבל לאחר החלקה).

שימו לב לעליה הקטנה ביעילות באזור תחילת מרץ. אני מניח שזה נגרם כתוצאה מהתחסנות האוכלוסיה היותר צעירה בסוף ינואר, ששם החיסון יעיל יותר מאשר באוכלוסיה המבוגרת.
סביב החצי הראשון של חודש אפריל הקסום, למתחסן הממוצע היה יעילות גבוהה מאד. ואז החלה השחיקה שהביאה לנפילה חופשית במאי-יוני.

Read 10 tweets

Yaniv Erlich

@erlichya

7 Sep

Important data from Israel: the relative protection against contracting Covid19 at the 4th wave as a function of vax/prior reinfection status. The data is calibrated to RR=1 for people who received 2nd dose in Jan and no booster.

A few thoughts below on waning protection👇

One important take away is that a prior infection protects you but can wane without a vax. People who contracted covid during the 2nd wave (Sep-Oct 20; dark red) have 1.9x risk to get delta compared to people who contracted Covid during the 3d wave (Jan-Feb 21; light red).

Recall that Alpha (3rd wave) is not closer to Delta (4th wave) than the ancestral strain (2nd wave). So it is reasonable to assume that diff are due to waning immunity.

This means that a recovered person looses ~15% of the protection per month. After 2yrs =~ no protection.

Read 7 tweets

Yaniv Erlich

@erlichya

19 Aug

The Israeli MoH released some results about the effectiveness of the booster to tame infection rate.
Behavior or immunity or both?

A short technical thread with interesting results.
🧵🧵🧵

Israel does not conduct an RCT for the third dose. Thus, the main analysis for booster-VE is comparing the cases in boosted vs 2nd dosed as a proxy for increased protection. But boosted individuals might be ⬆️covid19-worried, leading to lower ⬇️exposure.

To mitigate that, the MoH adjusted for basic demography and calendar day of infection and reported the graph below. Y-axis: Booster vs 2nd dose fold protection as a function of days from booster.
But there is something weird...🧐

Read 8 tweets

Yaniv Erlich

@erlichya

19 Aug

Interesting data about the performance of boosted as a function of time. Y: protection compared to 2 doses. X: days from booster.

There is an increase in protection with time after day 7. This reduces the chance of a behavioral effect (but not completely mitigate it)

These results are more puzzling. From right to left: days, # cases, # severe. Top line 💉💉. Bottom: 💉💉💉 >10days

Why puzzling? Because rate of (boosted severe/)(boosted cases) is ~5%. We saw a similar rate earlier in with only 2nd dose. Don't know how to explain that.

Read 4 tweets

Share this page!

Yaniv Erlich

Try unrolling a thread yourself!

More from @erlichya

Yaniv Erlich

Yaniv Erlich

Yaniv Erlich

Yaniv Erlich

Yaniv Erlich

Yaniv Erlich

Did Thread Reader help you today?

Like this author's thread?