Tracking illness to stop an outbreak

I chanced upon this article, which I thought is a good reflection on utilising emerging technology in public health.

In what has been called the “Age of Big Data,” when corporations are finding new ways to mine information to boost profits, University of Iowa professor Alberto Segre and a team of colleagues are channeling their work in data science to achieve something greater.

Segre is among the leaders of an interdisciplinary group of UI experts known as CompEpi, short for computational epidemiology, which conducts sophisticated data-driven research that probes how diseases spread.

The team, which includes six core faculty members and additional contributing professors and students, has put together studies as varied as tracking flu outbreaks using Twitter, to determining who should be vaccinated first among hospital staff to prevent the spread of an illness.

With studies showing that the failure of health care workers to perform proper hand hygiene is one of the leading causes of health care associated infection, of particular interest to the research group has been how technology can be used to monitor and improve hand washing practices in hospitals.

“There are all these different and great computational problems — real data that have real impact in peoples’ lives,” said Segre, chairman of the Department of Computer Science. “In the case of hospitals, health care-associated infections are a huge source of mortality and increasing health care costs, not to mention suffering. If we can reduce that, I feel like as a computer scientist I’ve done more than shave an epsilon off of Walmart’s trucking schedule to save another million dollars. So there’s this really nice social dimension to this.”

Since CompEpi’s formation about five years ago, the group has conducted a series of studies inside UI Hospitals and Clinics, beginning with hiring graduate students to observe and record workers’ movements and interactions to determine how health care-associated infections such as MRSA are spread in a hospital.

From there, the researchers used medical record system log-ins — each instance a hospital worker signed into the computer system — to track the workers’ movements throughout the hospital. That project yielded more than 2 million pieces of data, which was a vast improvement over the 6,500 samples collected by the grad students in the observational study, and the researchers used the information to build a contact network model to show how an infection could spread from person to person.

The scientists, wanting even more precise information on hospital workers’ movement, next teamed with UI engineers to build wearable computers — a re-purposed pager case that houses a processor and radio that broadcasts a worker’s location every 13 seconds. Additionally, they fixed instruments to hand soap dispenser pumps that measure frequency with which workers were washing.

Future studies will use even smaller tracking devices that will clip behind a doctor’s badge, as well as equip workers with wrist sensors that will measure their hand movements to gauge how effectively they’re washing.

“One of our overarching goals is to develop computational approaches to help understand why and in what situations health care workers do not practice appropriate hand hygiene and to use our findings to help model and understand other behaviors in order to make hospitals safer,” said Phil Polgreen, an associate professor in the Department of Internal Medicine and one of CompEpi’s founders, in an email.

Geb Thomas, an associate professor in the Department of Mechanical and Industrial Engineering, has overseen the development of the electronic hardware used in the hand washing studies.

“It’s been a lot of fun for the engineering students to bring their skills to bear on practical problems in the hospital,” Thomas said. “Almost all of the students who have worked for me have been excited about how important the problem is, and they’ve all gained a lot of practical experience in terms of how to actually make something that’s going to work reliably during the experimental time.”

Other research projects undertaken by CompEpi in recent years include:

• Researchers developed a program that uses the Twitter stream to monitor the spread of influenza by analyzing keywords such as “flu,” “sick,” “sniffles” and “fever,” then uses the geotagged locations of the tweets to map cases of illness. Although the Center for Disease Control collects data to measure flu activity, it typically takes about two weeks for the agency to compile results, Segre said. Using Twitter, however, measurements can be made in real time.

• The group developed a tool for the Iowa Department of Public Health that helps determine where to locate surveillance sites for influenza to ensure maximum coverage for the population.

• CompEpi members developed an app, which is available on the iTunes Store, designed for health care workers to manually track hand hygiene using iPads or iPods. It’s been downloaded thousands of times and is used in hospitals throughout the nation.

Ted Herman, a professor in the Department of Computer Science and one of CompEpi’s founding faculty members, said the collaboration with medical experts and those in other disciplines has allowed them to apply their respective skills in new contexts.

“We try to bring in people from different backgrounds and different departments because some of these problems are things which can’t be solved in just one way,” Herman said. “You need talents from different areas.”

Segre says projects like this are just the beginning of how data science can be applied to health-related research.

“These things are early attempts to understand how data science can impact quality of care and patient outcomes,” Segre said. “And I think the conversation is changing.”

Now the idea for secondary use of data captured in EMRs is not new but the application of Big Data analytics in real-time will lead to some new innovate applications.

The next question is, how do we spread the knowledge so the average Joe working in healthcare IT can apply this technology in an affordable manner.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s