Artificial intelligence can save lives in healthcare.
Foto: Agência Brasil
The routine in large hospitals and health centers is anything but easy. Even for a skilled medical team, managing care and accurately monitoring hundreds of hospitalized patients represents a superhuman challenge. Recognizing this problem – which in the United States, for example, contributes to an annual average of 400,000 deaths due to medical errors – scientists are focusing on the use of artificial intelligence in healthcare. In the article “Illuminating the dark spaces of healthcare with ambient intelligence ,” published in the journal Nature in September, researchers from Stanford University analyzed how “ambient intelligence” can save lives.
Alerting doctors and visitors when they fail to sanitize their hands before entering a patient's room, and discreetly monitoring frail elderly people for behavioral clues of impending health crises are some applications for AI sensor systems that could yield better results in the field.
Read more: In Folha, Viana recalls the discovery of the Rosetta Stone.
There is no linear path in research, says Karen Uhlenbeck.
Carolina Araujo wins the Ramanujan Prize 2020
“We have the ability to build technologies into the physical spaces where healthcare is delivered to help reduce the rate of fatal errors that occur today due to the large volume of patients and the complexity of their care,” said Arnold Milstein, professor of medicine and director of Stanford’s Center for Clinical Research Excellence (CERC) and one of the authors of the paper, in an interview with the university .
An experiment conducted at Stanford used AI-controlled infrared depth sensors to discern whether people had clean hands before entering a room. Attached to the door, a tablet alternated its screen color from green to red if there was any failure in hand hygiene. The choice of a visual alert instead of an audible one was based on advice from doctors. “Hospitals are already full of buzzing and beeping. A visual cue would probably be more effective and less irritating,” explained Milstein.
Experts are testing whether alert systems like this can reduce the number of ICU patients who contract nosocomial infections, potentially fatal diseases contracted by patients due to a lack of adherence to infection prevention protocols by people circulating in hospitals.
Although these tools improve the effectiveness of treatments, they raise some legal, regulatory, and privacy concerns. Computer scientist Fei-Fei Li, co-author of the study, indicates that these concerns should be identified and addressed publicly, gaining the trust of patients, providers, and the agencies and institutions that pay for healthcare costs.
“Technology to protect the health of clinically vulnerable populations is inherently human-centered. Researchers must listen to all stakeholders to create systems that complement the efforts of nurses, doctors, and other caregivers, as well as the patients themselves,” he pointed out.
Source: Stanford Engineering .