Objective: To determine the diagnostic accuracy of an algorithm structured in four clinical scenarios to discriminate benign primary headaches from serious secondary non-traumatic headaches (NTH) in the emergency department (ED).
Background: NTH is usually a benign symptom but can occasionally result in serious outcome making the disposition of patients with NTH difficult in the ED.
Design and methods: Consecutive adults patients referring to 8 EDs of the Emilia-Romagna region in Italy for NTH as the chief complaint were recruited in the study for a 30-day period. ED physicians attributed to each patient one of the four clinical scenarios (1, 2 and 3 identifying serious secondary headaches and scenario 4 identifying benign primary headaches) or an undetermined scenario when none of the four scenarios applied. Reference standards of the study were the head CT scan and a follow-up telephone interview after three months by the ED admission.
Results: The test was administered to 256 out of 302 (85%) eligible patients. The analysis (scenario 1,2,3 vs scenario 4) was based on 180 patients who completed the follow-up showing a sensitivity of 100% (95% confidence interval, 81% to 100%) and a specificity of 64% (56% to 71%). The likelihood ratio for a positive test was 2.67 (2.15 to 3.31) and the likelihood ratio for a negative test was 0.04 (0.003 to 0.64).
Conclusions: An algorithm based on four clinical scenarios can be administered to the majority of patients presenting to the ED with the chief complaint of NTH. The algorithm showed a good accuracy in identifying patients with non-life threatening causes of headache and could be used as a risk stratification tool to improve clinical decision- making. Further studies are required to validate this diagnostic algorithm.