BUILD MODEL WITH ASIA PACIFIC REGION IN MIND- MODELING INR CONTROL IN A COST-EFFECTIVENESS MODEL FOR STROKE PREVENTION IN ATRIAL FIBRILLATION PATIENTS

Author(s)

Chen J1, Wan Y2, Chen F2, Lu ZG3
1Merck & Co, Inc., North Wales, PA, USA, 2Nanjing Medical University, Nanjing, China, 3Brigham and Women’s Hospital, Boston, MA, USA

OBJECTIVES: Patients, physician behaviors, and health systems differ across regions. Models without features to account for heterogeneity can’t be readily adapted in AP region. Warfarin, the standard care for stroke prevention in atrial fibrillation patients, requires routine monitoring of International Normalized Ratio (INR). Below or above the target range (2.0-3.0) increases ischemic or hemorrhagic events, respectively. Time in therapeutic range (TTR) determines therapeutic effectiveness. INR monitoring and increased hemorrhagic risk created barriers for broad effective warfarin use in AP region. The mean TTR in US was 66.7%. In contrast, the estimate was 36% in China. All existing anticoagulant models, mostly Markov-based, have no explicit INR modeling.  To allow for effectiveness projections for warfarin in AP settings, our model incorporated an INR module. METHODS: We adopted a microsimulation approach to accommodate the patient heterogeneity and the strong association between a patient’s history and future outcomes. An INR control module was built in three steps: a) estimated event rates within target INR range, b) probabilistically chose an INR range for a patient at each cycle based on the simulated population’s TTR distribution, and c) projected event rates at the chosen INR range by incorporating the relative risks of INR control effect on events. Inputs were derived from the RE-LY trial, published literature and expert opinion.  RESULTS: We performed internal validation by projecting 2-year clinical events using the RE-LY-like TTR distribution (22.2% INR<2.0, 64.4% 2.0≤INR≤3.0 13.4% INR>3.0), achieving events rates similar to those observed in RE-LY. Applying a China-like TTR distribution (59% INR<2.0, 36% 2.0≤INR≤3.0, 5% INR>3.0) increased ischemic event rates while reducing hemorrhagic event rates.  CONCLUSIONS: Explicitly modeling INR control produced realistic projections which allowed adaptation of our anticoagulant model to AP settings. Proactively incorporating flexible model structure to accommodate heterogeneity was an efficient approach to improve model generalizability.

Conference/Value in Health Info

2014-09, ISPOR Asia Pacific 2014, Beijing, China

Value in Health, Vol. 17, No. 7 (November 2014)

Code

PRM17

Topic

Methodological & Statistical Research

Topic Subcategory

Modeling and simulation

Disease

Cardiovascular Disorders

Explore Related HEOR by Topic


Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now

×