The problem is that ad-based revenue has an incentive to:

1) Publish whenever there is a single study.

2) Be overly dramatic in either direction.

What we need is for a Bayesian-oriented medical doctor to spend a bunch of time keeping up to date on the literature, especially meta-analysis of multiple randomized controlled trials and then give that information to us with appropriate context. Luckily, this exists:

A YouTube channel called Healthcare Triage: https://m.youtube.com/#/user/thehealthcaretriage

EDIT (thanks mmorris!) https://www.youtube.com/user/thehealthcaretriage

A psychiatry (among other things) blogger called Slate Star Codex: http://slatestarcodex.com/2014/06/16/things-that-sometimes-h...

The former depends on donations through Patreon: https://www.patreon.com/healthcaretriage?ty=h