Toronto researchers are exploring the use of artificial intelligence to autonomously order imaging and other exams in the emergency department to help expedite patient triage.
Scientists detailed their investigation Wednesday in JAMA Network Open, noting that increased wait times and lengths of stay can result in poor patient outcomes in the ED. Machine learning-based medical directives (MLMDs) could help tackle this challenge, steering visitors to vital testing more quickly.
Physician Devin Singh, MBBS, and colleagues developed an AI model, training it to predict the need for forearm X-rays, electrocardiograms, and ultrasounds of the abdomen or testicles. Testing it on retrospective data from the Hospital for Sick Children in Toronto unearthed promising results. Implementing machine learning could potentially streamline care for more than 22% of all patient visits while making tests results available earlier by 165 minutes per affected patient.
“We expect that as more time passes and we have more training data, MLMD systems could streamline care for a progressively larger number of patients in the ED,” Singh, lead for clinical AI and machine learning in pediatric emergency medicine at the hospital, and co-authors wrote March 16. “Identifying testing needs for patients at the start of their ED visit can also assist with administrative planning for key stakeholders, such as diagnostic imaging departments, that often face staffing challenges when responding to surges in ED imaging requests,” they added later.
Singh et al. utilized data from patients ages 0-18 who presented at the ED between 2018-2019. Altogether, the final sample covered more than 77,000 visits, with subjects an average age of 5 years old. Their model maintained a high area under the receiver operator curve (0.89-0.99) and positive predictive value (0.77-0.94) across the differing imaging tests. It also performed with little to no sex-based bias, the authors added.
Next steps will include collecting race and ethnicity data to analyze how they may influence the model. Singh and colleagues also plan to investigate other human factors and how computer-provider interaction will impact results. They hope this will eventually lead to clinical trials.
“This prospective work will be important in not only building confidence in patients and [healthcare providers] in the use of autonomously acting models in healthcare, but will also be essential to inform legal, ethical and regulatory bodies on associated policy development,” they noted. “Such work will be required before model acceptance at a scale beyond the context of research studies.”
Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic.