Significant results do not require significant proof. This is a common misunderstanding. All hypothesis would require the same amount of proof, regardless of a person’s biases going into it.
In the kind of first principles sense you are thinking of, sure, proof is proof. But in the real world the burden of proof for an epidemiological headline like this absolutely is much higher, in the practical sense of "should I believe this bullshit or is there a mistake somewhere?"