Thread by @DrEricDing

Analogy: “Drunk driving car accidents don’t kill people—metal and concrete kills people, & maybe their internal bleeding”

Wanna know what causes heart attacks in a 41 year old Congressman-elect w/ no risk factors? #COVID19 blood clots & #COVID19 oxygen deprivation. (HT @atrupar)

2) 40% of all COVID related deaths are due to Covid heart complications and direct heart damage.

American Heart Association: heart.org/en/news/2020/0…

3) To be clear, causation is multi factorial, but there is often an underlying key cause. That’s why death certificates can have UNDERLYING CAUSE OF DEATH codes and MULTIPLE CAUSE OF DEATH codes per deceased. COVID is often the underlying cause. Nosology is a whole field itself.

4) Also there are 4 classes of causes in epidemiology:

—Necessary and sufficient (rare)

—Necessary but not sufficient (HPV virus required for cervical cancer)

—Not necessary but sufficient (smoking for lung cancer)

—Not necessary nor sufficient (MAIN TYPE—almost all causes)

5) the first type of “necessary and sufficient” is extremely rare — usually huge extreme things — eg like big asteroid ☄️ collision for extinction level planetary 🌍 wide instantaneous event. Hence its rare.

6) The last type “not necessary nor sufficient” is most common because most things have multiple causes, and each cause typically nudges up/down the statistical risk of the outcome by a few % points usually. Think drugs, diet, texting while driving, drunk driving, or COVID.

7) Basically, there are never any certainty with most causes, even like smoking and cancer, but increases risk a lot, and often the key factor that drive a causal chain. That’s usually the underlying cause, even if not necessary nor sufficient by itself. Epidemiology is fun eh?