I've always had concerns about the CHADS matrix and now I've looked more closely at HASBLED.
This is designed to assess one's risk of bleeding but nowhere does it refer to INR except to add just 1 point if your INR is labile (ie not stable). By definition therefore, if you don't take anticoagulants (same as an INR outside therapeutic range) only 1 point separates you from another who is properly anti-coagulated and consistently in therapeutic range. Hardly worth bothering according to this.
But statistically, it most certainly is worth being anticoagulated. Unprotected, the youngest and otherwise healthiest AF sufferer is 5 times more likely to have a stroke than someone with a similar profile in the general population who does not have the condition. If you're not young or healthy then it can be much higher.
The seriously misleading CHADS matrix tells us to wait for a first stroke before we take anticoagulation seriously and the equally misleading HASBLED tells us that the benefit of taking anticoagulants is only very marginal. Furthermore, HASBLED takes no account of the fact that bleed risk only significantly increases at an INR level well above therapeutic range. Stroke risk, on the other hand, significantly increases immediately below therapeutic range. There is not a clear balance between the two and the tipping point for bleed is unlikely to be reached by someone with well managed anti-coagulation. The tipping point for stroke however is ever-present unless properly anticoagulated.
The sooner the AF community wakes up, pops brain in gear and discards these utterly discredited matrices, the better.