Prediction Score for Anticoagulation Control Quality Among Older Adults

Area of Science:

• Geriatric Medicine
• Pharmacology
• Clinical Decision Support

Background:

• Time in therapeutic range (TTR) is crucial for vitamin K antagonist (VKA) effectiveness and safety.
• Optimizing VKA prescribing is essential for managing conditions like atrial fibrillation and venous thromboembolism.

Purpose of the Study:

• To develop and validate a prediction model for TTR in older adults (≥65 years) using VKA for nonvalvular atrial fibrillation and venous thromboembolism.

Main Methods:

• A prediction model was developed using electronic health record data from 1663 patients (training set) and validated on 1181 patients (validation set).
• Predictors were identified through systematic review and clinical knowledge.
• The new model, PROSPER (Pneumonia, Renal dysfunction, Oozing blood, Staying in hospital, Pain medication, no Enhanced anticoagulation services, Rx for antibiotics), was compared to the SAMe-TT2R2 score.

Main Results:

• The simplified PROSPER score, with 7 variables, outperformed SAMe-TT2R2 in predicting TTR ≥70% (AUC 0.67 vs. 0.55).
• PROSPER also demonstrated superior performance in predicting thromboembolic and bleeding outcomes (AUC 0.62 vs. 0.52).

Conclusions:

• The PROSPER score serves as a valuable clinical decision aid for selecting appropriate candidates for VKA therapy.
• It can also function as a research tool to address confounding factors and treatment heterogeneity in anticoagulation quality.