IBM corporation’s Watson machine learning software is famous for its performance on Jeopardy! but it is really making a name for itself in the medical field these days. AHMC is using CareDiscovery based on Watson’s design to address the issue with sepsis in its hospitals. Because of the ability of machine learning or artificial intelligence (AI) to collect and process large amounts of data, systems like Watson can find trends that humans can’t see.
The Dangers of Sepsis
Sepsis is a condition that develops when the body is fighting off different types of infection. The chemicals the body produces to kill the infection build up and cause inflammation. If this becomes widespread, it can lead to septic shock and death. Sepsis is often seen in hospitals among people with weakened immune systems, especially in elderly patients and people who have autoimmune diseases. Some people recover completely but others have lingering kidney damage.
Hospitals have protocols to treat sepsis, which includes massive doses of antibiotics, oxygen and sometimes treatment in the ICU, which is very expensive. Companies like AHMC are very interested in discovering how to control or eliminate sepsis. The list of actions to be followed includes early identification of patients with sepsis, starting IV fluids and antibiotics, and close observation in the ICU.
AHMC developed and implemented a sepsis protocol, but they couldn’t tell how effective it was until they began using CareDiscovery. This machine learning-based system collects data while employees work, which allows the hospital to see what actions are effective and what aren’t. The data indicated that the most important part of the sepsis protocol was early identification of the reaction, so the company focused on forming an early intervention team to identify cases and begin treatment as soon as possible.
AHMC developed a screening questionnaire with a symptom checklist. Staff checked off symptoms when patients came into the emergency room and this data was entered into the computer system. Posters reminding staff of the parts of the sepsis protocol were also distributed and put up. This led not only to a decrease in deaths from sepsis, but it also allowed the hospital to figure out which patients were likely to come down with the condition. That meant the hospital could catch cases before they became so severe and that result came directly from AI.
Machine learning is often credited with being the driving force behind social media and other apps, but it can save lives, too. The future is bright for data collection.