Article Text


COPD care bundles, IT systems, service analysis and beyond
P147 The Wells PE Score – an Effective Pre-Test Probability Tool?
  1. J Karalis,
  2. J Curtin Norfolk
  1. orwich University Hospital, Norwich, United Kingdom


Introduction and Objectives British Thoracic Society guidelines advise all patients with possible PE should have pre-test clinical probability assessed and documented. The Norfolk and Norwich University Hospital (NNUH) uses the Wells PE score. Validation studies have shown it to be a sensitive tool with a positive predictive value of: low risk (5%), medium risk (25%), and high risk (70%) respectively.

We wanted to assess the correlation of the Wells PE score with PE positivity rates on CT Pulmonary Angiogram (CTPA) in our patient population. We anticipated similar positive predictive rates as those seen in the original validation studies.

Methods We retrospectively collected data from CTPA requests submitted between October 2009 to September 2010. The Wells PE score was used to risk stratify patients into low (score < 2), moderate (score 2–6), or high risk (score > 6). These were correlated with the outcome of the CTPAs. We also used the alternate dichotomised Wells PE score of PE Unlikely (score ≤ 4) and PE likely group (score > 4) and correlated with results.

Results 1160 CTPA requests were identified, 66 were excluded, (they lacked a wells PE score). 1094 CTPA requests were analysed.

Conclusion We found poor correlation between the pre-test probability score and confirmed PE;

Results show a positive predictive value of low risk (21.4%), medium risk (22.1%) and high risk (35.3%). The dichotomised score showed PE Unlikely (19.1%) and PE Likely (27.4%). Analysis of the Wells PE variables revealed that they lacked a high specificity and high sensitivity, the subjective variable demonstrated the lowest specificity.

The Wells PE score is not an effective pre-test probability tool for discriminating low, moderate and high risk groups in the patient population at NNUH


  1. Wells PS, Anderson DR, Rodger M, et al. Derivation of a simple clinical model to categorise patients probability of pulmonary embolism: Increasing the models utility with the SimpliRED D-dimer. Thrombosis and Haemastasis. 2000; 83 (3): 416–420.

View this table:
Abstract P147 Table 1

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.