This study aims to explore the feasibility of a novel, patient-specific algorithm for adjusting warfarin doses during chronic anticoagulation therapy. Specifically investigators are interested in determining whether patients can use this algorithm to assume responsibility for managing their own warfarin therapy including making independent decisions about their warfarin dose and when to retest their next international normalized ratio (INR) test based on the result of their current INR result obtained using a point-of-care INR monitor.

Vitamin K antagonists (VKA) such as warfarin are very effective anticoagulants for the prevention of thrombosis. However, warfarin is also a leading cause of emergency room visits and hospitalizations for bleeding complications. A laboratory test of coagulation, the prothrombin time (PT), is used to monitor warfarin's anticoagulant effect on the hemostatic system. For warfarin monitoring the PT is expressed as an international normalized ratio (INR) to reduce the impact of interlaboratory PT variation. For individuals not on warfarin, the INR is about 1.0 while most patients on VKA require an INR between 2.0 and 3.0 to best balance the competing risks of thromboembolism and bleeding.

The INR response to warfarin can fluctuate as a result of changes in health status, interactions with other drugs, changes in dietary vitamin K intake, alcohol consumption, as well as for no apparent reason. In addition, the response to a change in warfarin dose varies between patients as does the time required for the INR response to reach steady state. A substantial proportion of patients treated with warfarin have stable INRs and remain on the same weekly maintenance dose month after month. Patients with variable INRs resulting in a low proportion of time in therapeutic INR range (TTR) have a higher risk of both bleeding complications and thromboembolic events. Good TTR is generally considered to be over 60% and TTR above 72% is considered excellent. One quarter to one third of patients have unstable INR results and require more frequent monitoring and warfarin dosing adjustments. This is demanding for patients and also for health care systems. Many of these patients remain unstable even when drug interactions, irregular dietary vitamin K intake, and missed warfarin doses have been excluded indicating the need for improved approaches to adjusting VKA doses in response to out-of-range INRs.

Interventions aimed at improving INR stability during warfarin therapy have been investigated. Using warfarin dosing algorithms is one intervention associated with improved INR stability and decreased risk for thromboembolic and bleeding complications and lower mortality. Patient self-management (PSM) of warfarin therapy using INRs obtained by patients in their home environment using point-of-care INR monitoring devices has also been shown to improve INR stability and decrease the risk of thromboembolic complications and mortality. Unfortunately, the US healthcare system lags behind other countries in widespread adoption of these evidence-based interventions for improving warfarin therapeutic outcomes.

Potential explanations for the underutilization of warfarin dosing algorithms and PSM include mistrust of algorithms among anticoagulation providers, complexity associated with algorithm use, lack of algorithm validation, and the effort required to train patients in self-management. As most patients receiving warfarin therapy in the US are outside the targeted INR range approximately 40% of the time, there is a critical need for approaches that remove barriers to adopting proven interventions like using warfarin dosing algorithms and PSM.

The Fearon Algorithm Mike Fearon, PhD has developed a patient-individualized warfarin dosing algorithm (the Fearon Algorithm [FA]) to predict INR results using parameters derived from the patient's prior warfarin dosing history. Key FA parameters not available with existing warfarin dosing algorithms include: 1) warfarin dose sensitivity (i.e. the likely change in INR from an increment change in warfarin dose), 2) optimal warfarin tablet strength for a given patient, 3) time required to reach INR steady state, and 4) inherent variability of the INR.

Warfarin dosing algorithms generally suggest a fixed percentage change (e.g. 10% to 20%) to the weekly warfarin maintenance dose in response to out-of-range INR results. While seemingly straightforward, training patients to use such algorithms during PSM has proven challenging. In contrast, the FA includes a personalized dose-effect curve for each patient and predicts the optimum warfarin dose required to center the patient's expected INR in the therapeutic range. Patient warfarin dose sensitivity is used to compute an optimal tablet strength for dose changes; sensitive patients require lower tablet strengths to facilitate small step increments in warfarin dosing. The tablet strength is used to provide the closest approximation to the optimum warfarin dose in mg/week. The FA discourages warfarin dose adjustments for INRs moderately deviating from the therapeutic range and encourages instead earlier repeat INR testing to confirm a suspected trend or to demonstrate stabilization. Earlier repeat INR testing is best accomplished with patient testing using point-of-care INR monitors.

Another unique feature of the FA is assigning likelihood probabilities expressed as the percentage of times the INR will fall in a range of INR values for each patient. This enables clinicians or patients to evaluate whether an extreme INR value requires immediate intervention or is simply to be expected occasionally.

Rationale The FA provides an understanding of the individual dose-response to warfarin. With a personalized approach to warfarin dose adjustments, we may be able to improve TTR and on a larger scale reduce clinically-important adverse events. In addition, the simplicity of the FA may facilitate the use of PSM in more patients in the US.

Despite the introduction of direct oral anticoagulants (DOACs) warfarin will be used by many patients in the US into the foreseeable future. DOACs are contraindicated in patients with mechanical heart valves and in patients with impaired renal function. Also, some patients prefer warfarin over DOACs and others are unable to afford DOACs.

A previous FA pilot study showed promising results but also identified improvement opportunities. In this study, the stability of INRs improved in 7 out of 10 patients during the study period. It was also demonstrated that the management of some patients would improve by switching warfarin tablet strengths to allow better dose fine tuning. The study also showed that an INR retest before the warfarin dose is changed from the optimum dose may reduce INR instability arising from unnecessary warfarin dose changes. A major flaw of this study was the failure of anticoagulation providers to adhere to the algorithm following out of range INRs.

Therefore, another study is necessary to demonstrate the feasibility of the FA. It is also important to demonstrate the ability of both anticoagulation providers and patients to use the FA in order to expand the use of PSM among US anticoagulated patients.

Study Objectives Primary objective: Compare measures of INR control (TTR, percent INRs in range (PIR), and control error distribution) between patients managed by anticoagulation management service (AMS) providers without the FA (standard management phase), patients managed by AMS providers using the FA (FA AMS phase), and patients using the FA to support PSM (FA PSM phase).

Secondary objectives: Compare 1) proportion of INRs associated with a warfarin dose change (DC), 2) inherent INR variability, 3) major bleeding events, 4) clinically relevant non-major bleeding events, 5) any bleeding event, and 6) thromboembolic events (TE) between the standard management, FA AMS, and FA PSM phases as defined previously.

Study Design Open label single-center cross-over study

Inclusion Criteria:

• Patients at least 22 years of age treated with warfarin for any indication for at least 18 months prior to enrollment within University of Utah Health system.
• Demonstrates the willingness and ability to test their own INR using a point-of-care INR monitoring device, and willingness to make independent decisions about warfarin dosing based on INR results using a dosing algorithm.
• Willingness to perform INR tests at least once weekly or more frequently as the algorithm dictates.
• Currently have and willing to maintain internet access for the duration of the study in order to complete online case report forms.
• Individual TTR < 60% over the 12 months prior to study enrollment.

Exclusion Criteria:

• Inherent INR variability >0.4 (such patients are unlikely to benefit from the Fearon Algorithm).
• Goal INR range less than a full INR point (e.g. 2.0-2.5).
• Known poor compliance to warfarin therapy (e.g., failure to take warfarin as instructed clearly documented in electronic medical record and/or return for INR testing as evidenced by repeated reminder communications documented in electronic medical record).
• More than one interruption of warfarin therapy for invasive procedure(s) lasting more than three days in the 18 months prior to study enrollment.
• Non-English speaking.
• Refusal to provide written informed consent.