Those coffee studies

Two new studies came out recently on the benefits of coffee. The usual caveats apply, of course (see this pretty good CNN summary), but a few things make these two articles particularly strong as observational evidence:

  1. The European study looked at liver function in a subset and found evidence for better function in coffee-drinkers, lending some biological plausibility.
  2. The Multiethnic study found a stronger beneficial effect among younger participants - this is consistent with survivor bias diluting the effect in populations with a range of participants.
  3. The European study did a formal bias analysis which suggested that unmeasured confounding would have to be on the order of another smoking-level determinant to explain away the observed association.
  4. The simple fact that between the two of them, they span multiple countries and multiple races and ethnicities, with little evidence of heterogeneity.

Get the papers here and here [firewall]