Comparison of Patient Self-Assessment of Clinical Risk of Pulmonary Embolism With Physician Assessment Using the Wells Score: A Prospective Observational Study

Emergency departments (ED) are facing greater patient volumes, causing challenges in triage. ED physicians employ validated clinical decision rules to risk stratify patients, including the Wells score for pulmonary embolism (PE). The current investigation explores if patients can generate a Wells score comparable to the score determined by their physician.

This was a prospective observational study comparing patient- and physician-assigned Wells scores. Patients were identified if they were adults who presented to the University of Virginia (UVA) ED with chest pain. Patients were excluded if there was no associated physician Wells score. Patients completed a questionnaire containing the components of the Wells score prior to seeing a physician, and also ranked their likelihood of a PE from 1 to 5. After the patient was seen by a physician, the physician assigned a Wells score. The primary outcome was the concordance of physician- and patient-generated Wells scores when analyzed by group as high, intermediate, or low risk.

After excluding patients without physician scores from the analysis, 89 subjects were included. Median patient score was 1.84 (IQR 0-3); median physician score was 1.38 (IQR: 0-1.5). For each one-point increase in patient likelihood score ranked 1-5, the physician Wells score increased by 0.73 (p<0.01). A chi-squared test was completed using three risk categories: low (<1.5), moderate (1.5-6) and high (>6). When scored by the patient as low risk (N=60, 67%), 60% of physicians (N=54) gave the patient a low score, and 7% (N=6) rated moderate; no physicians gave a high score. When scored by the patient as moderate-risk (N=24, 27%), 6% (N=5) of physicians gave a high score, 7%(N=6) gave a moderate score, and 15%(N=13) gave a low score. When scored by the patient as high-risk (N=5, 5%), 3% (N=3) of physicians gave a moderate score, 2% (N=2) gave a low score, and no physician assigned a high score (p=0.005). The question “is patient heart rate >100 bpm” was answered discordantly the most often, with 29% of patients answering differently from physicians.

Patient-generated Wells scores demonstrate meaningful agreement with physician assessment, particularly among low-risk patients. Patient-generated risk stratification may represent a feasible early triage tool with applications in ED workflows and telemedicine evaluation of chest pain.