Teaching Machines to Share
A prototype app for tablet devices aimed at helping reduce preventable medical errors in intensive care units is currently being tested at Johns Hopkins. Studies of U.S. hospitals show that more than 800,000 people—that’s 1 in 5—admitted to ICUs suffer some sort of preventable issue, such as blood clots. The problem, ironically, is the technology found in the ICU. Machines do a wonderful job individually, but don’t electronically “talk” to each other in a way that clinicians can easily decipher. Critical information can be missed among the estimated 200 daily tasks that nurses and doctors perform on an ICU patient.
The app, which involved input from 18 different disciplines including engineering and bioethics, improves how machines talk to one another and to people. Why an app? “When we asked clinicians, patients, and families how they envisioned the information being displayed, they didn’t want another big screen display or a display on a wall,” says Alan Ravitz of the Johns Hopkins Applied Physics Lab. “They wanted something they could hold and interact with.” A patient version of the app will also help families track the care that they can provide their loved one, such as brushing teeth or washing hair.
Determining if a body of water is safe just got easier. Johns Hopkins mechanical engineer Marin Kobilarov worked on a team to develop a pollution-hunting underwater robot. Equipped with navigation sensors and monitoring instruments, it keeps tabs on everything from water temperature to blue-green algae content. “We’re getting a 4-D picture of the water, the fourth dimension being time, so we can predict how quality would evolve day to day.” Piloted last year at Lake Pleasant in Phoenix, Arizona, the technology will be further developed and let loose this spring in the Chesapeake Bay.
Johns Hopkins researchers are using a computer algorithm to analyze Twitter posts and gather information about some common mental illnesses. The algorithm gathered tweets from users who publicly mentioned their diagnosis. It also looked for language cues linked to certain disorders, including words associated with anxiety and insomnia. Reviewing more than 8 billion tweets, the researchers have been able to quickly and inexpensively collect new data on post-traumatic stress disorder, depression, bipolar disorder, and seasonal affective disorder. “Using Twitter to get a fix on mental health cases could be very helpful to health practitioners and government officials who need to decide where counseling and other care are needed most,” says Johns Hopkins computer scientist Mark Dredze.
Tiny Spill, Big Problem
When gasoline dribbles from the pump’s nozzle after a fill-up, few among us think twice about the environmental impact of such a tiny spill. But the truth is, those drips add up, seeping into the concrete pads beneath the pumps and then contaminating the soil and groundwater. That’s bad news for people who live nearby and rely on well water, according to researchers at the Johns Hopkins Bloomberg School of Public Health. Roughly 1,500 liters of gasoline are spilled at a typical gas station every 10 years—and that’s the researchers’ conservative estimate. That’s a problem when the environmental and public health impacts of chronic gasoline spills are poorly understood. “Chronic gasoline spills could well become significant public health issues since the gas station industry is currently trending away from small-scale service stations,” says environmental health scientist Patrick N. Breysse.
Better Ebola Protection
After last year’s Ebola outbreak raised major concerns over the spread of the highly infectious disease, the U.S. government challenged innovators from around the world to help health care workers provide better, more timely care, while also containing the virus. A team from Johns Hopkins Center for Bioengineering Innovation & Design and Jhpiego, an international nonprofit health organization affiliated with Johns Hopkins University, is now among three global winners to receive financial support for its prototype of a health care worker suit. The suit was redesigned for quicker, safer removal and to keep workers from coming in contact with Ebola patients’ contagious body fluids during and after treatment. In addition, it is expected to keep the wearer cooler—an important benefit in hot, humid regions such as West Africa. (The cooling technology used in the garment was originally developed by Johns Hopkins cardiologist Harikrishna Tandri for cooling patients in cardiac arrest.) The suit will now undergo intensive testing to ensure readiness for production and deployment in the field.