How this instrument works
The Wells score for pulmonary embolism adds seven clinical findings, weighted by how strongly each predicted PE in the derivation data: clinical signs of DVT (3 points), PE being the leading diagnosis or equally likely as anything else (3 points), heart rate over 100 (1.5), immobilization or surgery in the past four weeks (1.5), a prior DVT or PE (1.5), hemoptysis (1), and active malignancy (1). Three of the seven criteria carry half points, so the running total can land on 1.5, 4, 6.5, or any other half-integer, across a possible range of 0 to 12.5.
Wells and colleagues published the model in Thrombosis and Haemostasis in 2000, built from patients worked up for suspected PE and paired with a rapid bedside D-dimer assay, SimpliRED. The goal was a pretest-probability estimate simple enough to calculate at the bedside without a calculator or lookup table — a running tally a clinician could keep in their head while examining a patient.
Two genuinely different ways of reading the total are both in active clinical use today, and this instrument reports both. The original 2000 paper split results into three tiers: low probability below 2, moderate from 2 through 6, and high above 6 — stated as inequalities rather than clean integer bands, since half-point totals like 1.5 or 6.5 fall between whole numbers. A separate, later line of work, the Christopher Study published by van Belle and colleagues in JAMA in 2006, popularized a simpler two-way split built around imaging strategy: 4 or below is read as 'PE unlikely,' anything above 4 as 'PE likely,' driving whether a clinician orders a D-dimer first or goes straight to CT pulmonary angiography.
- Toggle Clinical signs/symptoms of DVT — leg swelling and tenderness on exam — worth 3 points.
- Toggle whether PE is the #1 diagnosis, or equally likely as any alternative — also worth 3 points.
- Toggle Heart rate over 100, worth 1.5 points.
- Toggle Immobilization 3+ days or surgery in the past 4 weeks, worth 1.5 points.
- Toggle Previous DVT or PE, worth 1.5 points.
- Toggle Hemoptysis and Malignancy, each worth 1 point, then read the total and both probability schemes.
Worked example — 7.5 points, high probability by both schemes
Clinical DVT signs are present (3), PE is judged the leading diagnosis (3), and the heart rate reads over 100 (1.5), with nothing else positive: 3 + 3 + 1.5 = 7.5. Under the original three-tier scheme that's high probability, since 7.5 sits above the 6-point line; under the two-tier scheme it's also comfortably 'PE likely,' well past the 4-point threshold.
A different mix reaches nearly the same total without either 3-point item: a heart rate over 100 (1.5), recent immobilization (1.5), a prior DVT or PE (1.5), hemoptysis (1), and malignancy (1) sum to 1.5 + 1.5 + 1.5 + 1 + 1 = 6.5. That's just over the 6-point line into the high tier by the original scheme, and past 4 into 'PE likely' by the two-tier split — even though no single finding here carries the maximum individual weight.
Questions
What does a Wells score of 7.5 mean for PE probability?
By the original 2000 three-tier scheme, anything above 6 is read as high probability, so 7.5 sits well inside that band. By the newer two-tier split from the 2006 Christopher Study, anything above 4 is 'PE likely,' which 7.5 also clears comfortably. Both readings point the same direction here — toward imaging rather than a D-dimer as the next step.
What are the two ways to interpret a Wells score?
The original derivation paper (Wells et al., 2000) split results into three tiers — low below 2, moderate 2 through 6, high above 6. A later, separate study, the Christopher Study (van Belle et al., 2006), validated a simpler two-way split: 4 or below as 'PE unlikely,' above 4 as 'PE likely,' built specifically to decide whether a D-dimer test alone is enough or imaging is warranted next. Both schemes are legitimately used in practice today, and clinicians often quote whichever one matches the decision they're making.
Why can a Wells score include half points like 1.5 or 6.5?
Three of the seven criteria — heart rate over 100, recent immobilization or surgery, and a prior DVT or PE — are each individually weighted at 1.5 points rather than a whole number. Add any combination of half-point and whole-point items together and the running total can land on a half-integer, which is why the scale's true range is 0 to 12.5 rather than a clean 0 to 12.
What counts as 'PE is the top diagnosis' on the Wells score?
It's the clinician's own judgment call: does pulmonary embolism look at least as likely as any other explanation for the patient's symptoms, given the full picture? It carries 3 of the possible 12.5 points, the same weight as clinical DVT signs, and it's also the one criterion that depends on clinical impression rather than a fixed objective finding — which is exactly the item the Revised Geneva score was built to eliminate.
Does a low Wells score rule out pulmonary embolism on its own?
No. A low score lowers the pretest probability enough that a negative D-dimer test is generally considered sufficient to stop the workup, but the score alone, without a D-dimer or imaging, isn't treated as a rule-out. It's a probability estimate meant to guide which test comes next, not a substitute for that test.
How does the Wells score differ from the Revised Geneva score?
Both estimate pretest probability for PE from a short list of findings, but the Wells score includes one subjective item — is PE the most likely diagnosis — that depends on the examining clinician's overall impression. The Revised Geneva score, covered separately on this site, was deliberately built from only objective, checklist-style findings, so two different clinicians examining the same patient are less likely to land on different Geneva totals than different Wells totals.
References
- Wells et al. 2000, Thromb Haemost — Wells score derivation (PubMed)
- van Belle et al. 2006, JAMA — Christopher Study, PE-likely/unlikely dichotomy (PubMed)
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.