[Impact of an education program on patient behaviour favoring prevention of drug-related adverse events: a pilot study in patients receiving oral anticoagulants for thromboembolic venous disease].

Objective: Adverse events related to oral anticoagulants represent a major public health problem. Including patient education as part of the prevention strategy could contribute to improved effectiveness and safer use of drugs. The present study aimed at evaluating the outcomes of a patient education program inspired by recommendations from French Health Authorities (AFSSAPS) and based on an "individual guidance" approach.

Method: The study was conducted in two groups of hospitalized patients treated with oral anticoagulants for thromboembolic disease. Each patient in the first (intervention) group attended an individual teaching session conducted at discharge by a trained pharmacist. Patients in the second (control) group were given usual care. These two groups were compared at inclusion (before intervention) and three months later. The outcomes considered were the acquisition of: 1) knowledge, 2) risk anticipation and compliance behaviours characterized by the stability of INR and the incidence of hemorrhagic episodes during the period of observation.

Results: Fifty-nine patients (average age 65 years) were included (29 in the intervention group and 30 in the control group). Three months after the intervention, the intervention group exhibited 1) better knowledge (higher rate of restitution of treatment-related information--name of the drug, administration plan, targeted range for INR (...), interpretation of INR results (p<0.05), management of a specific scenario where INR declines concomitant to elevation of anticoagulant dose (p<0.05)); 2) higher rates of relevant behaviours (p<0.05)--in the event of a missed dose, anticipating an event with a high risk of bleeding, dealing with signs of overdose--and higher compliance profile (ns) (stability of INR, and number of hemorragic episodes). A multivariate model integrating the potential explanatory variables for frequency of hemorrhagic episodes at 3 months (demographic data, history of thrombotic disease, INR stability, reference group (intervention/control)), showed that the only variable significantly associated with frequency of bleeding events was the reference group of the patient (p=0.05; odds-ratio=4.5, interval of confidence: [1-21]).

Conclusion: Multivariate analysis demonstrated that the probability of developing a hemorrhagic event when taking an oral anticoagulant is on average 4-fold greater in patients given usual care than in patients given individual guidance a pharmacist. A larger randomized trial is currently under way in the Rhône-Alpes region, France, to validate these exploratory results.