Using analytics to help hospitals avoid inadvertently sickening patients and their caregivers

Post Comment
Big Data Evangelist, IBM

Infectious agents (also known as "germs") are generally invisible. Most of us only know they're present in our environment, or in our bodies, when illness strikes. And that's just too late.

Hospitals are fruitful breeding grounds for germs and the nasty consequences they inflict. Sorry to say, but the place most likely to make you even sicker is often a hospital. After all, they're convergence points for the unfortunate people who have contracted infectious diseases. And they're also places where doctors, nurses, orderlies, administrators, patients and visitors inadvertently introduce other germs, and spread the invisible contagions brought there by patients. If you think every doorknob and countertop in the average hospital has been sterilized, you've never actually been to one.

Using Analytics to Help Hospitals Avoid Inadvertently Sickening Patients and their Caregivers imsis400-046.jpg

Nothing I've just said is controversial. It's well-known that before modern antiseptics were invented more patients died from post-surgical infections than were saved on the operating table. This invisible pattern (unsterilized surgical instruments, garments and hands introducing infections) may be obvious to us now, but it was vehemently denied by most of the medical profession until antiseptics' stunning success proved them utterly wrong.

Nevertheless, the invisible spread of infections in healthcare facilities has continued to run rampant. Even today, healthcare-associated infections (HAIs) remain a serious threat. This recent Healthcare IT News article reports, according to the US Centers for Disease Control (CDC), that around four percent of hospital patients have at least one HAI at any point in time. Of the estimated 722,000 HAIs in U.S. acute care hospitals in 2011, more than 10 percent of the impacted patients died during their hospitalizations. More than 50 percent of all HAIs occurred outside of the hospital's intensive care unit.

And let's not forget about all the doctors, nurses and staff in hospitals who also catch HAIs. That sad truth is among the many factors that has led to the general collapse of the healthcare infrastructure in Ebola-plagued western Africa. As this recent article reports, "What’s happening is the general collapse of the health care system because physicians and nurses and other health workers are staying home. They feel unprotected and unprepared to deal with this—and they are."

Fear is a contagion in its own right. It tends to run rampant when people come up against invisible forces that could strike them down, unaware, long before they realize they've been impacted. Nevertheless, pathogen-caused infections, though they spread invisibly in healthcare environments, can be illuminated through judicious deployment of advanced analytics. Indeed, advanced analytics, which involves applying statistical methods to trustworthy data, has long been used to reveal invisible patterns of all sorts. Consequently, their potential role in HAI identification and risk mitigation should be obvious.

Fortunately, today's medical establishment is avidly embracing analytics for this exact purpose. The cited article discusses analytic software being used by hospitals to spot HAIs so that they can be contained and treated lest they become serious outbreaks. The article describes hospitals tracking HAI patterns in daily operations. These deployments involve such data-driven tools as event processing, data mining, segmentation analysis, cluster analysis, anomaly detection, predictive analytics, prescriptive alerting and reporting on diverse operational and patient data sources.

All of these analytics function as an early-warning system on HAIs, not as a foolproof fix, supplementing the facilities' other infection-control efforts. As one analytic solution provider states in the Healthcare IT News article, "Our system uses data mining technology to mine through all the hospital data and identify trends that could be outbreaks or clusters that have occurred that the hospital may not have previously identified. We really look at this technology as identifying the smoke before the fire."

Though the article doesn't mention the Internet of Things (IoT) in this context, its potential role in HAI control should be obvious.

What if IoT infection sensors were installed throughout hospitals, not only in operating rooms, patients' rooms and other places where physicians and nurses attend to patients, but also in waiting rooms, hallways, lobbies, cafeterias and other places frequented by the general public? What if all of these sensors were feeding pathogen-detection data in real time into a central HAI-control nerve center? And what if this intelligence could be used by facilities administrators to send real-time alerts, lock down specific areas and pinpoint the precise source of potential infections and contagions so that they can be neutralized before they do real damage?

Even if medical science has not produced a vaccine or cure for the contagion (as with the Ebola outbreak), illuminating the source and spread of the dread disease can be enormously effective. Shedding analytic light on an otherwise-invisible contagion can support more effective quarantining, not just in hospitals and clinics but in the society at large.

Evils, even those stemming from biological forces beyond our ken, tend to shrink when we shine the right kind of light on them.