Intervention for Diagnosis of Deep Vein Thrombosis in Acute Stroke Patients: A Hospital-Based Study

Background: Immobile stroke patients are at high risk of deep vein thrombosis (DVT). Demographic studies suggest a low incidence of DVT in Asian patients, but that might be underestimated. Objective: Intervention by in-hospital case management for diagnosis of DVT in patients with acute stroke. Patients and Methods: Intervention was defined as: recommendation of D-dimer test for patients who are immobile by day 4 after stroke onset and compression ultrasonography if the level of D-dimer is >= 500 ng/ml. Treating physicians were notified by case managers before they failed to do so for qualified patients. Data of patients hospitalized 12 months before and 8 months after the intervention, including basic demographics, Glasgow Coma Scale score, National Institute of Health Stroke Scale (NIHSS) score, laboratory results, and examination reports, was retrieved from electronic medical records for analysis by code searches for acute stroke. Results: A total of 2523 patients were identified. 1528 were before and 995 after intervention. More patients after intervention had D-dimer test and ultrasound examination than that before intervention (22.1% vs 8.6%, P<0.001 and 15.1% vs 1.2%, P<0.001, respectively). Ultrasound diagnosis of DVT was significantly more after than before intervention (2.0% vs 0.3%, P<0.001). DVT was 55.7 per 1000 in patients with a NIHSS score >= 18. Male sex (Odds ratio 0.33, 95% confidence intervals: 0.11-0.98), NIHSS score (Odds ratio 1.05, 95% confidence intervals: 1.00-1.09), and intervention (Odds ratio 5.39, 95% confidence intervals: 1.88-15.44) were independent predictors of ultrasound diagnosis of DVT. Conclusions: Intervention by in-hospital case management may be an effective strategy for improvement of under-diagnosis of DVT in acute stroke patients.