Interactive tool
Calculate the IMPROVE score
Seven variables collected at admission. The usual decision threshold for in-hospital prophylaxis is ≥ 2 points.
How to use it
Instructions
The seven variables
| Factor | Points |
|---|---|
| Previous venous thromboembolism | 3 |
| Known thrombophilia | 2 |
| Current lower-limb paralysis | 2 |
| Current cancer | 2 |
| Immobilisation ≥ 7 days | 1 |
| Admission to intensive or coronary care unit | 1 |
| Age > 60 years | 1 |
Notes
- Immobilisation: confined to bed or chair for at least 7 days, including the days immediately before admission. The definition is stricter than Padua's (3 days).
- Current cancer: active malignancy, excluding non-melanoma skin cancer.
- Thrombophilia: a previously documented congenital or acquired abnormality.
- Lower-limb paralysis: current paresis or plegia, not a long-standing stable deficit that does not change current mobility.
- The maximum score is 12; in clinical practice values above 5 are rare.
Balancing against bleeding risk
The same IMPROVE cohort produced a second tool, the IMPROVE Bleeding Risk Score, which should be used in parallel: a score ≥ 7 identifies patients at high risk of major in-hospital bleeding, in whom pharmacological prophylaxis should be reconsidered in favour of mechanical prophylaxis. The full table is given in the guidelines section.
Clinical references
Guidelines and interpretation
| Score | Estimated risk | Interpretation |
|---|---|---|
| 0 | 0.4% | Low: pharmacological prophylaxis not routinely indicated |
| 1 | 0.6% | Low |
| 2 | 1.0% | Decision threshold: in-hospital prophylaxis indicated |
| 3 | 1.7% | Increased |
| 4 | 2.9% | High: in-hospital prophylaxis; consider extending it after discharge, taking bleeding risk into account |
| ≥ 5 | 7.2% | High |
IMPROVE Bleeding Risk Score
To be calculated at admission together with the thrombotic score. A total of ≥ 7 points defines high bleeding risk.
| Factor | Points |
|---|---|
| Active gastroduodenal ulcer | 4.5 |
| Bleeding in the 3 months before admission | 4.0 |
| Platelets < 50×109/L | 4.0 |
| Age ≥ 85 years | 3.5 |
| Hepatic failure (INR > 1.5) | 2.5 |
| Severe renal failure (GFR < 30 ml/min/1.73 m²) | 2.5 |
| Admission to intensive or coronary care unit | 2.5 |
| Central venous catheter | 2.0 |
| Rheumatic disease | 2.0 |
| Current cancer | 2.0 |
| Age 40 – 84 years | 1.5 |
| Moderate renal failure (GFR 30 – 59) | 1.0 |
| Male sex | 1.0 |
The four possible combinations
| Thrombotic risk | Bleeding risk | Management |
|---|---|---|
| Low (< 2) | Low (< 7) | No pharmacological prophylaxis; early mobilisation |
| Low (< 2) | High (≥ 7) | No pharmacological prophylaxis; mechanical prophylaxis if appropriate |
| High (≥ 2) | Low (< 7) | Pharmacological prophylaxis for the whole hospital stay |
| High (≥ 2) | High (≥ 7) | Mechanical prophylaxis; reassess daily whether to switch to pharmacological prophylaxis as bleeding risk falls |
Extended prophylaxis after discharge
This is where IMPROVE shows its added value over Padua. In patients with IMPROVE ≥ 2 — and even more so with IMPROVEDD ≥ 2, which adds D-dimer — thromboembolic risk persists in the weeks after discharge. The APEX trial showed a benefit of extended prophylaxis with betrixaban precisely in this subgroup, with an acceptable safety profile. The decision remains individual and requires a low bleeding risk.
Padua is more sensitive (it classifies about 40% of medical patients as high risk) and is preferable when the aim is not to miss candidates for in-hospital prophylaxis. IMPROVE is more specific (about 20%) and is the reference tool when considering extended prophylaxis, where the bleeding risk of prolonged treatment carries more weight. The two models should not be added together, nor used interchangeably to justify a decision already taken.
References
Sources
- Spyropoulos AC. et al., Chest 2011 — Predictive and associative models to identify hospitalized medical patients at risk for VTE (IMPROVE registry).
- Decousus H. et al., Chest 2011 — Factors at admission associated with bleeding risk in medical patients: findings from the IMPROVE investigators.
- Rosenberg D. et al., J Am Heart Assoc 2014 — External validation of the risk assessment model of the IMPROVE study.
- Cohen AT. et al., N Engl J Med 2016 — Extended thromboprophylaxis with betrixaban in acutely ill medical patients (APEX trial).
Quick-reference tool for healthcare professionals. It does not replace the original guidelines, the summaries of product characteristics or the specialist's clinical judgement. All calculations run in the browser: no patient data is transmitted to or stored on any server.