1/
One day everyone will recognize #selectionbias due to a #collider and the world will be a better place.

This time observational studies found a higher risk of omicron reinfection after a 3rd dose of #COVID19 vaccine. As usual, alarms went off.

Can you see the obvious bias? 2/
Those who receive a booster and get infected are, on average, more susceptible to infection than those who don't receive a booster and get infected.

So no surprise than those who receive a booster and get infected are more likely to get reinfected.

Led by @susanamcorella...

1/
Our findings on a fourth dose (2nd booster) of the Pfizer-BioNTech #COVID19 vaccine are now published.

Compared with 3 doses only, a fourth dose had 68% effectiveness against COVID-19 hospitalization during the Omicron era in persons over 60 years of age.

Interestingly...

https://twitter.com/NEJM/status/1514351552042979340

2/
... this is yet another example of the need for good #observational studies that emulate a #TargetTrial.

Would it be better to have a real randomized trial? Yes

Do we have a randomized trial? No

Will we have a randomized trial? Perhaps, but too late for a timely decision.

1/
We emulated a target trial of two #COVID19 mRNA vaccines in the largest healthcare system in the US.

Both vaccines were similarly effective, with Moderna slightly better than Pfizer-BioNTech.

But that isn't the most important conclusion of our study.

https://twitter.com/barbradickerman/status/1466173899712962561?s=20

2/
Spring of 2020: #COVID19 vaccines are developed.

October 2020: Results from randomized trial are announced.
businesswire.com/news/home/2021…

~6 months from development to evaluation of effectiveness.

Utterly impressive. Unprecedented.

Kudos to the pharmaceutical industry.

Now...

1/
Vaccine safety: We compared excess adverse events after #COVID19 vaccination (Pfizer-BioNTech) and after documented #SARSCoV2 infection.

nejm.org/doi/full/10.10…

Take-home message: Low excess risk of adverse events after vaccination, higher after infection.

Some thoughts👇 2/
Preferring #SARSCoV2 infection over vaccination has become even harder. (Remember: infection also increases the risk of severe disease/death)

This is a good illustration of how #randomized trials and #observational studies complement each other for better #causalinference...

1/

Many countries are vaccinating their elderly. Can we relax control measures now?

No.

Even with 50% of elderly vaccinated, uncontrolled #SARSCOV2 transmission may overrun the healthcare system.

We explain @AMJPublicHealth today, led by @_gmales
ajph.aphapublications.org/doi/10.2105/AJ… 2/

Data from Madrid, Spring 2020:

Critical care requirements peaked at 5 times the usual capacity.

Hospitals managed to increase ICU capacity by 3-fold.

Heroic but, sadly, insufficient.

The healthcare system collapsed. Not everybody who needed critical care received it.

@ProfMattFox 1/
The odds ratio from a case-control study is an unbiased estimator of the

a. odds ratio in the underlying cohort when we sample controls among non-cases

b. rate ratio in the underlying cohort when we use with incidence density sampling

No rare outcome assumption required. @ProfMattFox 2/
Because the odds ratio is approximately equal to the risk ratio when the outcome is rare, the odds ratio from a case-control study approximates the risk ratio in the underlying cohort when we sample controls among non-cases and the outcome is rare.

But...

1/
We recently confirmed the effectiveness of the Pfizer-BioNTech vaccine outside of randomized trials @NEJM.
nejm.org/doi/full/10.10…

Studies like ours are being used to promote a vaccine passport to travel in the US, UK, and European Union.

A few clarifications are in order. 2/
Before we start, a disclaimer:

Vaccine passports involve complex ethical, economic, and political considerations.

Here I talk exclusively about scientific issues. The goal is that those making decisions have a better understanding of what we do and do not know as of today.

1/
We've just confirmed the effectiveness of the Pfizer-BioNTech vaccine outside of randomized trials.

Details @NEJM: nejm.org/doi/full/10.10…

Yes, great news, but let's talk about methodological issues that arise when using #observational data to estimate vaccine effectiveness. 2/
A critical concern in observational studies of vaccine effectiveness is #confounding:

Suppose that people who get vaccinated have, on average, a lower risk of infection/disease than those who don't get vaccinated.

Then, even if the vaccine were useless, it'd look beneficial.

1/ Estimating the infection fatality risk (IFR) of #SARSCoV2 is hard.

Our estimates from Spain's #ENECOVID (just published):
Men: 1.1% to 1.4%
Women: 0.58% to 0.77%

After age 80
Men: 12% to 16%
Women: 4.6% to 6.5%

Why is the #IFR hard to estimate?
bmj.com/content/371/bm… 2/ The IFR in a population is the ratio of

number of deaths from SARS-CoV-2 infection (numerator)

and

number of individuals infected by SARS-CoV-2 (denominator)

during a prespecified period.

Both numerator and denominator are hard to quantify.

1/ Madrid es la única gran ciudad europea con más del 100% de sus camas de UCI en hospitales públicos ocupadas por enfermos con #COVID19

DURANTE DOS MESES.

Mantener este nivel de ocupación de UCIs es jugar con fuego.

https://twitter.com/covid19_m/status/1326253729864151044?s=20

2/ Porque:

Se reasignan a UCI camas que se deberían usar para otras enfermedades.

Se agota al personal de UCI que debe atender a más pacientes.

Si la epidemia resurge, no queda capacidad de respuesta.

Las UCIs son nuestra última línea defensa.

1/
Ayer el 36% de UCIs de Madrid estaban ocupadas por enfermos con #COVID19, según cifras oficiales.
elpais.com/sociedad/2020-…

Incorrecto.

Ayer el 95% de UCIs de Madrid estaban ocupadas por COVID-19 (112% en hospitales públicos).

https://twitter.com/covid19_m/status/1308494103370436609

¿Por qué la discrepancia? 2/
Porque las cifras oficiales cuentan como UCI cualquier cama donde se puede instalar un respirador:
quirófanos, salas reanimación postquirúrgica, unidad coronaria, UCI pediátrica...

Y no cuentan que el 70-75% de UCIs de verdad suelen están ocupadas en periodos no pandémicos.

1/
Look at the shape of these curves.

New York and Madrid had similar epidemics until they spectacularly diverged.

In March, both cities were caught by surprise and shut down because of #COVID19.

In September, the situation is under control in NY and alarming in Madrid.

Why? 2/
Let’s start with the similarities: two big, dense cities with a large network of public transit and lots of visitors.

An explosive outbreak of #SARSCOV2 overwhelmed their contact tracing system and their hospitals. A lockdown was required to reduce the public health disaster.

1/ Five months ago I asked about a #stratifiedlockdown to handle #COVID19.

The idea was to restrict lockdowns to people over age 50 or with preexisting conditions while the rest of society lives a relatively normal life.

Time to revisit this approach.

https://twitter.com/_MiguelHernan/status/1239227279512829953?s=20

2/ No country has explicitly adopted a #stratifiedlockdown, but many have implicitly defaulted into some version of it.

That is, governments haven't ordered older people and their cohabitants to stay home, but they do recommend those in vulnerable groups to be extremely careful.

BREAKING: Risk of #COVID19 hospitalization in 77,590 persons with #HIV by antiretroviral type:
TDF/FTC: 10.5
TAF/FTC: 20.3
ABC/3TC: 23.4
Other: 20.0
per 10,000 (Febr-April 2020)

WANTED: Randomized trials of TDF/FTC (Tenofovir/Emtricitabine)

doi.org/10.7326/M20-36… That is, individuals on TDF/FTC had about half the risk of #COVID19 hospitalization than those on TAF/FTC or ABC/3TC.

Rate ratio 0.53 (95% CI 0.29, 0.95)
journals.lww.com/epidem/Citatio…
Any reasonable person should be concerned about confounding, so we did the following 3 things
👇

Germany: 0.4% mortality of #COVID19 cases (27,300 diagnoses, 115 deaths).

Compare with 4% France, 5% UK, 6% Spain.

Germany, either share your secret weapon with the rest of the world or fix the data errors.

Good EU policy requires good data across EU.
worldometers.info/coronavirus/ Let's see some possible explanations for Germany's surprisingly low #COVID19 mortality:

1) Epidemic started later in Germany so fewer cases have resolved.

Unlikely because the epidemic started around the same time in Germany as in, say, France.

...

Europe and US in total or partial lockdown to mitigate #COVID19.

Soon more hundreds of millions will be confined to their homes. Huge personal sacrifices. Staggering economic losses.

Ok, we accept the price for the common good. But how does this end? Here are some strategies: 1) Develop a vaccine and vaccinate.

If a sufficiently high % of the population is vaccinated and becomes somewhat immune, this “herd immunity” lowers the probability of transmission to a susceptible person until the epidemic dies out.

Ideal but unlikely to happen soon enough.

#Causalinference that talks the talk and walks the walk.

Claim: "Continuing #breastcancer screening past age 75 doesn't reduce 8-year breast cancer mortality."

Emulation of a #TargetTrial led by @xabieradrian with Medicare data
doi.org/10.7326/M18-11…

Let the discussion start. @xabieradrian @AnnalsofIM @HarvardEpi @harvard_data @HarvardBiostats @HarvardChanSPH @CMSGov @CMSgovPress @MonganInstitute @MassGeneralNews 2)

Because there's so much talk about #causalinference around here.

Computer scientists, economists, statisticians... talk a lot about the merits of #DeepLearning, instrumental variables, or whatever their preferred methodology is.

Everybody: This is your chance to shine.

1/
Suppose you want to extend causal inferences from a randomized trial to a target population.

Is that #transportability or #generalizability?

Issa Dahabreh and I propose an answer in this brief commentary in the European Journal of Epidemiology:
ncbi.nlm.nih.gov/pubmed/31218483 2/
For those interested in methods for extending inferences from randomized trials to a target population:

Take a look at our tutorial
arxiv.org/pdf/1805.00550…
(soon to appear in Statistics in Medicine)

You will find identification conditions AND three estimation approaches.

This week the US Supreme Court refused to remedy an extremely serious constitutional violation.

All because a majority of the Court's Justices don’t know how to handle quantitative information.

Bad times for democracy when numerically illiterate people are in charge.

THREAD👇 Background:

Each state is divided into districts and each district sends a representative to Congress.

States decide how to draw district lines.

The state's ruling party is tempted to redistrict to maximize its % of seats in Congress. This is known as partisan gerrymandering.

Randomization never ensures zero #confounding bias. It provides probabilistic bounds on confounding.

Therefore, by bad luck, the effect estimates from some perfectly conducted randomized #trials are substantially confounded. But we don't know which ones!

An eye-opening example: In Denmark, 860 individuals were randomly allocated to either "intervention" or "control":
• No intervention was implemented
• Individuals were unaware of their allocation
• Mortality was higher in the intervention group with p=0.003

Keep this in mind when evaluating a trial.

Today Mendelian Randomization (MR) is usually implemented as a form of instrumental variable (IV) estimation.

Aware that valid #IVestimation requires strong assumptions, MR advocates often retreat to the position that MR numerical estimates need not be taken seriously...

https://twitter.com/mendel_random/status/1001025528101527552

Their position: The goal of MR is to "test causality". MR studies aren't designed to yield a valid numerical IV estimate of causal effect. MR studies are designed to answer a yes/no question: Is the causal null hypothesis true?

But retreating to *null testing* is problematic...