More than 40 organizations have launched ACT for Better Diagnosis, an effort to improve diagnostic processes and their resulting accuracy.
The organizations—representing providers, specialty societies, federal agencies, patients, policymakers and other groups—believe that as many as 80,000 patient deaths annually can be attributed to inaccurate or delayed diagnosis, and that the use of information technology can play a key role in reducing mortality.
“This is a campaign to mobilize and motivate public stakeholders, policy makers, informaticists and data vendors,” says David Newman-Toker, MD, a professor of neurology, ophthalmology and otolaryngology at Johns Hopkins Medicine.
For example, Johns Hopkins is using real-world data from patients for simulation training of clinicians, supported with clinical decision support, to quickly diagnosis patients who don’t exhibit stroke but could be at high risk, and refer the patient to a specialist via a telemedicine consultation.
Data from health information exchanges also can be used to track adverse outcomes for patients who have been misdiagnosed, such as a patient that was diagnosed with a viral infection and came back a few days later exhibiting signs of sepsis, which may result in refresher courses on current clinical best practices.
For now, the simulations are not ready for prime time use in a clinical setting, according to Newman-Toker. “This is the beginning of a journey to raise awareness and release new data and scientific work to improve diagnosis.”
One of the best diagnostic tools is an eye exam, which can reveal a wide range of medical issues that could include anemia, kidney disease, glaucoma, dehydration, allergies and infection, says Newman-Toker. But he emphasizes taking advantage of big data to improve the quality and safety of care. “We need to improve diagnosis streams with big data and statistical analytics to get closer to overall performance in real-world events.”
For instance, intensive care units have streams of data coming from patients and the data can be analyzed to identify patients who will die of sepsis within 48 hours absent a medical intervention. “This is a good use of machine learning technology to treat these patients before they get sicker.”
More information on the Society to Improve Diagnosis in Medicine is available here.