Health News Digest

October 8, 2012

Network Discovery In Health Care: Exploring Both Meanings Of This Term

Network discovery

By  Michael D. Shaw

The notions of network discovery and network mapping apply to the field of health care in two ways, and depending on your personal perspective, one of them will probably seem more obvious than the other. The networks described can either refer to how humans interact, or how computers are connected. Clearly, there is overlap here.

Visionary Valdis Krebs speaks of—and provides software for—social network analysis. For the purposes of this article, “social network” refers to the actual connections of people, and not the websites that might facilitate this. In a widely-cited paper entitled “Building Smart Communities through Network Weaving,” Krebs and June Holley, founder of the Appalachian Center for Economic Networks, begin with this challenge:

Communities are built on connections. Better connections usually provide better opportunities. But, what are better connections, and how do they lead to more effective and productive communities? How do we build connected communities that create, and take advantage of, opportunities in their region or marketplace? How does success emerge from the complex interactions within communities?

The answer, they say, lies in improved connectivity, and this is created “through an iterative process of knowing the network and knitting the network.” A first step here is network mapping/network discovery.

For a bit more insight, I looked at a paper entitled “NET GAINS: A Handbook for Network Builders Seeking Social Change,” written by Peter Plastrik and Madeleine Taylor. The authors define a network as “[A] set of nodes and links, of things that are connected to each other. Picture, for instance, networks of roads between cities, or computers emailing to each other, or living cells joined as tissue in the body.”

They identify and discuss five Network Effects:

  • Rapid Growth—Members benefit from adding new links
  • Rapid Diffusion —As more nodes are added, resources are diffused more widely through its links
  • Reach—Networks provide short pathways between individuals separated by geographic or social distance
  • Resilience—Network nodes quickly reorganize around disruptions or bottlenecks
  • Adaptive Capacity—Networks are nimble, able to assemble or disassemble capacities with relative ease

One interesting example of how social network discovery can work in a health care environment was presented a few years ago by June Holley, applied to the huge problem of Hospital Acquired Infections (HAIs). HAIs cost around $45 billion, and claim 100,000 lives per year in the US alone.

As she put it…

Infection transmission in a hospital can be understood as a result of social networks that exist among patients as well as those that exist between staff and patients: Transmission of infection occurs as patients touch each other, share cigarettes or drinking cups, or engage in other interactive behavior. By identifying these social networks, we can generate highly targeted strategies for lessening infection transmission.

More recently, social network analysis has been utilized to develop enhanced vaccination strategies for health care workers.

We now turn to computer network discovery. I recently spoke with Michael Markulec, President and CEO of Somerset, NJ based Lumeta Corporation. Spun-off from Bell Labs in 2000, the company’s flagship product is IPsonar, touted as the most widely deployed network discovery solution for large enterprise and government.

Markulec reminds us of the proliferation of devices that exist across computer networks today. Network operations must get a handle on this. He told me that “Organizations only know about 80–85% of what’s really on their network. What we do is provide a very accurate inventory of where all those devices are, how they’re interconnected, and how to better protect the core data.”

Beyond network discovery, IPsonar also provides Host Discovery, Leak Discovery, Service Discovery, Enhanced Perimeter Discovery, and Device Discovery. Device profiling is a big deal for Lumeta. “We maintain a library of over 26,000 devices,” he told me.

Markulec says that most people take their computer networks for granted, but they must consider how many different types of devices are connected. In a health care setting, you might have medical equipment, HVAC, diagnostic units, and physical security components—in addition to the storing of patient data. What good is having an MRI machine on site, he asks, if it gets infected with a computer virus?

It seems clear that as we improve our networks—in both senses of the word—health care will improve as well.