Predictive model cuts back on unnecessary imaging requests for headaches in the ED
New data could improve the process of determining which patients would benefit from neuroimaging in emergency settings, potentially reducing unnecessary requisitions.
Headaches are a common complaint among patients presenting to the emergency department, with millions of cases documented annually in the United States. Though the majority end up being discharged without a concerning diagnosis, there are instances when severe headaches indicate a life-threatening occurrence. As such, the workups for severe headaches are thorough and often involve imaging.
A team of experts recently sought to develop a prediction model capable of guiding providers toward more efficient use of neuroimaging for headache patients. They detailed their efforts in a new paper published by Emergency Radiology.
“Headache is common at emergency services and neuroimaging can help in timely diagnosis of life-threatening pathologies,” Roger Figueroa-Paz, with the Clinical Research Center of Valle del Lili Hospital in Colombia, and colleagues noted. “We evaluated clinical indicators associated with abnormal neuroimaging in patients with acute headache, aiming to develop a scoring system with reliable diagnostic performance.”
The team incorporated data from every patient who presented to a regional hospital ED in Cali, Colombia, from 2011 to 2019. All patients with non-traumatic headaches who underwent neuroimaging were included. The team included demographic and clinical data in the model's training, also taking note of patients’ symptoms, imaging findings and diagnoses to determine whether certain factors increased the likelihood of abnormal exams.
Three models were developed, but “model 1” yielded the greatest predictive value, with an AUC of 0.757 when using a cutoff score of 0.179. Of the 626 patients included, 15.5% had abnormal neuroimaging findings. Factors found to be associated with increased risk of abnormal findings were age (older than 40), motor deficits, visual deficits and gait disturbances.
The team suggested their model is more accurate than using “red flag” signs alone. Using their cutoff point, their model could result in fewer unnecessary imaging requests in headache workups, they added.
“Our straightforward scale incorporates clinical factors associated with abnormal neuroimaging, with the aim of improving diagnostic performance and predictive capacity to distinguish patients who require neuroimaging," the group concluded.
Learn more about the findings here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.