all the information, none of the junk | biotech • healthcare • life sciences

Into the Swarm: San Diego’s Swarmology Mines Online Conversations for Health Data

Xconomy San Diego — 

Swarmology, a San Diego provider of social media analytics for healthcare clients, is stepping into the light today, disclosing initial funding of $1.2 million and the debut of its website.

The company, founded last December by pharmaceutical and health IT executive Malcolm Bohm, provides targeted, Web-based marketing services for its customers by analyzing online conversations about specific health concerns on Facebook, Twitter, and other social media. Swarmology’s customers already include a large healthcare insurance provider and a couple of smaller pharmaceutical companies, Bohm says.

Before arriving in San Diego last year, Bohm says he was working with electronic health insurance reimbursement records at New Jersey-based SDI Health to help drug developers minimize the costs and risks of their clinical trials. By working with “very rich data,” Bohm says he helped pharmaceuticals optimize the design of their clinical trials—in part by identifying the ideal medical center, and by mapping and recruiting patients for a particular study.

Bohm says the idea for Swarmology occurred after he began wondering if there was a comparable richness in the health and medical information that people share on Facebook and Twitter. “Enormous numbers of people are talking about their own health,” Bohm says. For example, he says Facebook users will update their status with such comments as “My daughter had another asthma attack today” or “I’m taking Sumatriptan [for a migraine] and it’s not working.”

Bohm says a Pew internet Life Research study published earlier this year shows that 70 percent of U.S. adults search online for health-related subjects, and some 43 percent are crowdsourcing their health care through online discussions with fellow patients and healthcare providers. Using data analytics, Bohm says, “You can get this wealth of information across wide swaths of social media. We can capture all the conversations—we call that a swarm—in a virtual community.”

While Swarmology’s technology is proprietary, Bohm says the Centers for Disease Control and the Pentagon’s Defense Advanced Projects Agency are using similar techniques to analyze social media “to try to understand if the online signals are sufficient to predict political unrest in other countries or to track and predict where flu and allergy outbreaks are going to stress regional healthcare systems. In the same way, Google can track and predict contagious outbreaks as people go online to search for information about their symptoms.

Bohm says Swarmology can collect online comments posted on a particular topic—diabetes, for example—by searching for specific terms such as “insulin,” “blood,” “sugar,” and “HbA1c” (a standard lab test that measures blood sugar) as well as less-specific terms such as “diet” and “exercise” that are more contextual. By using semantic analysis and natural language processing, Bohm says Swarmology’s technology is sophisticated enough to identify and include comments like “I’m walking like a zombie today” in a search for online comments about insomnia.

Bohm says Swarmology also has been working on a set of algorithms to identify a variety of characteristics in a given community—such as vibrancy, sentiment, traffic activity, and membership—“to connect the dots and predict where a given conversation is headed.”

Bohm says the technology is capable of identifying any individual, or group of individuals, or all of the individuals in a particular swarm, and also can identify the ideal moment to send a marketing message into the swarm on behalf of a customer. As Bohm likes to put it, “We can send the right message at the right time to the right people.”

For example, in an online discussion about cholesterol and heart disease, Bohm says Swarmology can predict the best opportunity for a health insurance company to send a message to key influencers, or to everyone in the swarm, with an offer for free cholesterol testing that includes a website link so they can register online.

The company’s technology relies partly on the field of “swarm intelligence,” loosely defined as the collective behavior of decentralized, self-organized systems. “The math we have created looks for inherent patterns that similarly exist in natural swarms,” Bohm says. “We additionally apply Artificial Intelligence techniques as an overlay to better understand the direction the swarms will take.”

Since starting Swarmology almost a year ago, Bohm has funded most of the company’s operations. (He says, “I’m my own friend and my own family, and the initial round was me.”)

After completing his education in England, Bohm began his career in the pharmaceutical industry, rising through increasingly senior positions at Astra, Pfizer, Novartis, and other Big Pharmas. At Aspreva Pharmaceuticals, a Vancouver, B.C.-based biotech, Bohm served as the New Jersey-based executive director of global data sciences and reporting, and oversaw the company’s U.S. clinical operations. In 2008, Switzerland’s Galencia Group acquired Aspreva in a $915 million deal and merged it with its Vifor pharmaceutical operations.

By then, Bohm had started Trialytics, a healthcare analytics company later acquired by New Jersey-based SDI, which was itself acquired for an undisclosed amount earlier this year by IMS Health.

Bohm says Southern California’s Tech Coast Angels are providing about two-thirds of Swarmology’s new $1.2 million capital infusion. He plans to use the funding to expand the company’s five-person workforce by one or two employees and perhaps accelerate software development. Swarmology is moving into EvoNexus, the “no strings attached” incubator operated for free by CommNexus, the San Diego nonprofit technology industry group.

(Creative Commons credit: Photo of “Swarm Chandelier” in London’s Victoria and Albert Museum from Heatheronhertravels.com)

By posting a comment, you agree to our terms and conditions.

Comments are closed.