Venous thromboembolism (VTE) including deep vein thrombosis (DVT) and pulmonary embolism (PE) is a frequent disease and the third most common cause of cardiovascular death in the world after myocardial infarction and stroke. Anticoagulant therapy drastically reduces the risk of early VTE recurrence and death, but it exposes patients to a substantial risk of bleeding. Hence, determining the optimal duration of anticoagulant treatment for VTE is a major public health issue.

When major transient risk factors for VTE are identified (major surgery, immobilization...), patients generally do not need to extend anticoagulation beyond 3 months, whereas for VTE diagnosed in the context of cancer, therapeutic anticoagulation is required for as long as the cancer is considered "active".

However, in more than 50% of cases, venous thromboembolic disease occurs spontaneously, i.e. without any significant clinically detectable circumstance (known as unprovoked venous thromboembolic disease). In such patients, the risk of recurrence is high (35% recurrence rate at 5 years, with a 10% risk of death per recurrence). Scientific societies therefore recommend continuing anticoagulant treatment "indefinitely" (i.e. without programming a stop date or long-term treatment). However, this practice exposes these patients to an ongoing, non-negligible increase in the risk of bleeding, which could ultimately exceed the risk of recurrence of venous thrombo-embolic disease.

Optimizing anticoagulant therapy beyond the first three to six months of treatment is therefore a crucial and challenging issue, which could improve the long-term prognosis of patients with unprovoked thromboembolic venous disease.

Based on the quantitative and qualitative approaches implemented in MORPHEUS project granted by European Commission (HORIZON-HLTH-2022-TOOL-11-01 call), the investigators have combined predictive personalized medicine, through the use of risk biomarkers, with a patient-centered model of medicine, which, while based on an understanding of the patient's experience, leading to develop Time-Dependent Multicomponent risk prediction scores and socIo-anthropological scales (TDMI) integrated in a shared decision-making process regarding anticoagulant treatment duration in patients with a first episode of unprovoked VTE.

The aim of this study is to demonstrate that this strategy, based on a medical decision-making process shared between patients and physicians and including TDMI, reduces the risk of recurrence of thromboembolic venous disease (fatal or non-fatal), the risk of bleeding and all-cause mortality, and is associated with greater patient satisfaction after a first episode of unprovoked thromboembolic venous disease.

Inclusion Criteria:

• Patient > or = 18 years,
• Patient with a first episode of symptomatic unprovoked pulmonary embolism (PE) and/or proximal deep vein thrombosis (DVT) treated for 3 to 6 uninterrupted months with full dose anticoagulant therapy,
• Signed informed consent.

Exclusion Criteria:

• Unable or refusal to give informed consent,
• Isolated distal DVT,
• Isolated sub-segmental PE
• Previous unprovoked VTE
• Known CTEPH
• Indication for anticoagulation other than DVT or PE (e.g.; atrial fibrillation, mechanic valves…),
• Interruption of anticoagulation for 14 days or more before the inclusion,
• Active cancer of less than 24 months,
• Current pregnancy,
• Life expectancy <18 months (e.g.; patients with an end-stage chronic disease)
• Not affiliated to national insurance, social security (only for France)