Wells Score Calculator for DVT (Deep-Vein Thrombosis)

Background.

The Wells score for deep-vein thrombosis is a bedside pretest-probability aid for clinicians assessing a patient with a suspected first episode of lower-limb DVT. It is not a diagnosis, it does not replace clinical judgement, and on its own it neither confirms nor excludes a clot. Its only job is to decide what happens next: a patient in the lower band is normally taken to a D-dimer, and a patient in the higher band is normally taken straight to compression ultrasound. A result of "DVT unlikely" excludes DVT only when it is paired with a negative D-dimer — never by itself.

This calculator implements two published models rather than blending them, because they are genuinely different instruments. The original 1997 model of Wells and colleagues has nine criteria and reports three bands: low, moderate and high. The modified 2003 model adds a tenth criterion, previously documented DVT, and reports a two-level answer: DVT likely or DVT unlikely, split at a score of 2. The three-level bands were measured on the nine-item model, so applying them to a ten-item total would be off by up to a point. The "Published model" selector therefore chooses the item set and the banding together, and the previous-DVT criterion is simply ignored when you select the 1997 model.

Nine of the ten criteria score plus one point each. Only one criterion moves the score downwards: if an alternative diagnosis is at least as likely as DVT, the score drops by two. That single negative item is what allows the total to fall below zero, and it is the reason a patient with three positive findings and a convincing alternative explanation lands on a score of 1 rather than 3. The calculator reports the number of positive criteria and the alternative-diagnosis deduction as separate values so the total can be audited rather than trusted blindly.

Every weight in this score is an integer, so there is no rounding anywhere in the calculation and no ambiguity at a boundary. A score of 1 is unlikely and a score of 2 is likely, with nothing in between. That matters more here than in most calculators, because a single tick box moves a patient between two different diagnostic pathways.

The numbers behind the bands come from the derivation cohort, not from an individual. In the 1997 study of 593 symptomatic outpatients, DVT was confirmed in 10 of 329 patients in the low band (3 percent), 32 of 193 in the moderate band (17 percent), and 53 of 71 in the high band (75 percent). Those are the proportions observed in that population. They are not a probability for the person in front of you, and they will differ in a population with a different underlying prevalence — a primary-care list, an oncology ward and an emergency department do not share a baseline.

The score has boundaries of validity that matter more than the arithmetic. It was derived and validated in symptomatic outpatients with a suspected first lower-limb DVT. It is not validated for upper-extremity DVT, it is not validated in pregnancy, it performs poorly for suspected recurrent DVT in the same leg where post-thrombotic changes mimic every physical sign, and it is not intended for patients already receiving therapeutic anticoagulation. If your patient falls into one of those groups, the number this page produces is not the number the literature validated, and the page says so beside the result rather than burying it.

The criteria themselves also require examination, not inference. "Calf swelling more than 3 cm larger than the asymptomatic side" means a measured circumference difference taken at a fixed landmark below the tibial tuberosity, not an impression that one leg looks bigger. "Localised tenderness along the distribution of the deep venous system" is a specific finding, not general leg pain. A score assembled from guesses produces a confident-looking integer with nothing behind it, which is the most dangerous failure mode a scoring page has.

What is wells score calculator (dvt)?

The Wells score for DVT is a clinical prediction rule that converts a short list of history and examination findings into a pretest probability of deep-vein thrombosis. It was developed by Philip Wells and colleagues in 1997 and modified by the same group in 2003. Each criterion is a yes/no clinical finding; nine of them add a point and one — an alternative diagnosis being at least as likely as DVT — subtracts two.

The score exists because unselected imaging is expensive and slow, and because D-dimer is sensitive but not specific. Combining a structured pretest probability with a D-dimer lets a clinician safely discharge a large fraction of patients without any imaging at all, and lets the remainder go straight to ultrasound. The score is the front half of that pathway; it is not the pathway.

Two scoring schemes circulate under the same name. The two-level scheme, split at a total of 2, is the one used by the American Society of Hematology's 2018 venous thromboembolism guideline and by NICE guideline NG158 in the United Kingdom. The three-level scheme — low, moderate, high — is the original 1997 formulation and is the one whose published DVT prevalences (3, 17 and 75 percent) are quoted most often. This page keeps them separate so that each banding is only ever applied to the item set it was derived on.

What the score is not: it is not a diagnostic test, it is not a treatment decision, it is not validated for the leg you have already imaged, and it carries no information about whether the clot, if present, is proximal or distal. Those questions belong to imaging and to the clinician, not to an integer.

How to use this calculator.

1. 01Choose the published model. Use "Modified (2003)" for the two-level likely/unlikely answer used by ASH 2018 and NICE NG158, or "Original (1997)" for the three-level low/moderate/high answer.
2. 02Answer each criterion from documented history and a physical examination — not from an impression. Calf circumference must be measured, not estimated.
3. 03Mark "Alternative diagnosis at least as likely as DVT" only when you have a specific competing explanation such as cellulitis, a ruptured Baker's cyst, a calf haematoma or chronic venous insufficiency. This item subtracts two points.
4. 04Read the total, the number of criteria met and the alternative-diagnosis deduction together, so you can see which findings produced the score.
5. 05Check the pretest-probability band against your local diagnostic pathway. The score selects the next test; it does not confirm or exclude DVT.
6. 06Do not use the result in isolation for a patient who is pregnant, has suspected recurrent DVT in the same leg, has suspected upper-extremity DVT, or is already anticoagulated — the score was not validated in those groups.
7. 07Record the model you used alongside the number. A score of 2 means different things under the 1997 and 2003 schemes.

The formula.

The Wells DVT score is a weighted sum of yes/no clinical criteria. Nine criteria carry a weight of +1: active cancer; paralysis, paresis or recent plaster immobilisation of the leg; recently bedridden more than three days or major surgery within twelve weeks; localised tenderness along the deep venous system; entire leg swollen; calf swelling more than 3 cm larger than the asymptomatic side; pitting oedema confined to the symptomatic leg; collateral non-varicose superficial veins; and previously documented DVT. One criterion carries a weight of −2: an alternative diagnosis at least as likely as DVT.

Rounding stage: there is none. Every weight is an integer, the total is an exact integer sum, and the band is decided by comparing that exact integer against the thresholds. No intermediate value is rounded, so the result at a boundary is unambiguous — the calculator cannot produce a 1.9 that displays as 2.

The previously-documented-DVT criterion belongs to the 2003 modification only. When the "Original (1997)" model is selected, that criterion is not counted, because the 1997 three-level bands were measured on a nine-item score. This is why the same patient can produce a total of 2 under the 2003 model and 1 under the 1997 model when previous DVT is the only difference.

Banding under the modified 2003 model is a single cut: a total of 2 or more is "DVT likely", and a total below 2 is "DVT unlikely". Banding under the original 1997 model uses two cuts: 3 or more is high probability, 1 or 2 is moderate probability, and 0 or less is low probability. Because the alternative-diagnosis item subtracts two, the minimum possible total is −2, which sits in the lowest band of both schemes. The maximum is 9 under the 2003 model and 8 under the 1997 model.

Worked through the example on this page: active cancer (+1), recently bedridden more than three days (+1), localised tenderness (+1), calf swelling more than 3 cm (+1) and pitting oedema confined to the symptomatic leg (+1) give five positive criteria and no deduction, so the total is 1 + 1 + 1 + 1 + 1 = 5. Five is at or above 2, so the 2003 model returns "DVT likely". Five is also at or above 3, so the 1997 model returns high probability for the same patient.

A worked example.

Frequently asked questions.

References& sources.

1. [1]Wells PS, Anderson DR, Bormanis J, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet. 1997;350(9094):1795–1798. Original nine-item model and the three-level bands with DVT prevalence of 3%, 17% and 75%. (Abstract retrieved 2026-07-29; full text paywalled.)
2. [2]Wells PS, Anderson DR, Rodger M, et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med. 2003;349(13):1227–1235. Modified ten-item model adding previously documented DVT, dichotomised at a score of 2. (Abstract retrieved 2026-07-29; publisher full text paywalled.)
3. [3]Lim W, Le Gal G, Bates SM, et al. American Society of Hematology 2018 guidelines for management of venous thromboembolism: diagnosis of venous thromboembolism. Blood Adv. 2018;2(22):3226–3256. Independent issuing body; states the two-tier split as Wells < 2 "DVT unlikely" and Wells ≥ 2 "DVT likely". (Retrieved 2026-07-29.)
4. [4]National Institute for Health and Care Excellence. Venous thromboembolic diseases: diagnosis, management and thrombophilia testing. NICE guideline NG158 (published 2020, updated 2023). Recommends the two-level DVT Wells score; a score of 2 or more directs to proximal leg vein ultrasound, below 2 to D-dimer. (Bibliographic reference — nice.org.uk blocks automated retrieval; consulted 2026-07-29.)
5. [5]Chopard R, Albertsen IE, Piazza G. Diagnosis and treatment of lower extremity venous thromboembolism: a review. Table 1, Wells' score for DVT — used solely to verify the ten-item list and the −2 weighting letter by letter. PMC8462022. (Retrieved 2026-07-29.)
6. [6]National Institute for Health and Care Excellence. Venous thromboembolism in adults. Quality standard QS201, quality statement 3: proximal leg vein ultrasound scan for a 'likely' deep vein thrombosis Wells score. (Consulted 2026-07-29 as a second NICE-issued confirmation of the ≥ 2 cut.)