An overview of the effect of computer-assisted management of anticoagulant therapy on the quality of anticoagulation.

Risks and benefits of anticoagulant therapy depend directly of the quality of anticoagulation. We carried out a meta-analysis of published randomized trials to assess the overall effectiveness of computer-assisted prescription systems on the quality of anticoagulation. Randomized controlled trials were identified through electronic searches of the Medline database (1966-1997) and systematic analyses of the references of articles. Two investigators selected relevant papers and summarized data from the studies. The outcome variable was the proportion of days within the target range of anticoagulation. A pooled estimate of the common odds ratio of being in the target range and its confidence interval was obtained by the Mantel-Haenszel method. Nine trials having included 1336 patients were identified. Computer systems were based on a pharmacokinetic-pharmacodynamic model and a bayesian prediction method. Most of them concerned the oral anticoagulant warfarin. The global odds ratio of being in the target range was 1.29 [95% CI: 1.17-1.49], thus meaning that the use of a computer for anticoagulation optimization increased by 29% the proportion of visits where patients were appropriately treated. The proportion of clinical events was too low for allowing a summary analysis, but major hemorrhages tended to be less frequent among patients of the computer groups than among patients of the control groups (2.0 versus 3.9%). Evidence from randomized controlled trials supports the effectiveness of computer-aided anticoagulant prescription. Widespread use of these systems in ambulatory care could increase the benefit/risk ratio of anticoagulant treatment at a low cost.