Terminated by sponsor

Patients on warfarin but with unstable international normalized ratio (INR) will be recruited to a randomized trial comparing dosing based on an algorithm (Fearon algorithm, named after the mathematician Michael Fearon), which uses historical data to estimate patients' sensitivity to warfarin and to dose changes as well as the lag time until a dose adjustment takes effect, or to the investigators standard management Main outcome is improvement in time in therapeutic range.

Patients with very variable INRs resulting in a low proportion of time in therapeutic range (TTR) have a higher risk of both bleeding complications and thromboembolic events. A good TTR is generally considered to be over 60% and "excellent TTR" is above 72% of time in range. The investigators will recruit patients with a TTR below 50%.

The Fearon algorithm provides a more thorough understanding of the individual responses to changes in warfarin dose and with a personalized nomogram for dose adjustments and timing of laboratory tests the investigators have an opportunity to improve the TTR and on a larger scale to reduce clinically important adverse events. Before embarking on a large randomized trial with this revised Fearon algorithm-derived nomogram this pilot study with a randomized, open-label design will be performed to demonstrate proof of principle.

Inclusion Criteria:

• Patients with mechanical heart valve prosthesis managed for the warfarin therapy by the Thrombosis Service at HHS-General Hospital.
• Treated with warfarin for at least 1 year.
• Therapeutic INR range 2.0-3.0 or 2.5-3.5.
• TTR in the lowest quartile

Exclusion Criteria:

• Known poor compliance due to for example alcohol abuse or cognitive impairment
• Refusal to provide written informed consent