Providers need new tools to make sense of unstructured data

Most healthcare information is ‘invisible’ to current systems, says Ginni Rometty


Healthcare is being inundated with a deluge of big health data, 80 percent of which are invisible to current computing systems because the data are unstructured, according to Ginni Rometty, chairman and CEO of IBM.

New cognitive systems will enable the industry to “see” unstructured health data that previously has been hidden from sight.

Traditional structured health data, which “fits into rows, columns and tables,” is visible, but the vast majority of big data—consisting of doctors’ notes, wearables, X-rays, social media, the weather, sensors and sound—is unstructured and invisible to healthcare organizations, Rometty said In her keynote address Tuesday at the World Health Care Congress in Washington, DC.

“Today, systems can hold all that data, but they actually don’t know what it is,” she told the audience, “You couldn’t program enough systems in the world to understand that data.” However, with the advent of cognitive computing, clinicians can gain new insights from unstructured data at both the individual and population health levels, empowering clinicians to make more informed patient diagnoses and enabling proactive and preventive interventions.

She sees cognitive systems, such as IBM’s Watson, as a “learned colleague and an informed second opinion that looks at hidden connections” in the data, forming hypotheses and rapidly testing them to achieve better health outcomes.

“Some people shorthand this capability, calling it artificial intelligence. That would be doing it a disservice,” Rometty added. “These new cognitive systems understand, reason and they learn. They never stop learning.”

Rometty says the IBM Watson Health supercomputer uses natural language processing and machine learning to provide actionable insights from large amounts of unstructured data. While electronic health record systems are more than capable of storing both structured and unstructured data, she contends that Watson can help summarize and analyze that data and incorporate—for example—doctor and nurse notes to give a more complete picture of patients and individualized treatment plans.

IBM’s Watson Health Cloud, which is HIPAA compliant, was launched last year to help store, sort and analyze both structured and unstructured data as a service offering. IBM is hoping that healthcare organizations tap into this cloud-based data to take previously disparate data sets—structured and unstructured—and combining them together to create unique insights.

IBM is “betting its future” on healthcare, Rometty says, having spent $4 billion on healthcare data companies, including Phytel, a health management software vendor, and Truven Health Analytics, a $2.6 billion acquisition that was completed last week. IBM also has partnered with Apple, Johnson & Johnson and Medtronic to make it easier for healthcare organizations to store and analyze patient data by leveraging Watson’s cognitive capabilities.

Apple has solidified its partnership with Watson Health Cloud, which will serve as a secure cloud platform and analytics for Apple’s health-tracking software HealthKit and ResearchKit, a new open source software framework expected to revolutionize medical research by turning iPhones into powerful diagnostic tools for gathering health data.

Last month, IBM and the American Sleep Apnea Association announced the launch of a SleepHealth app designed for iPhones and Apple Watches, which will be used as part of a study to help identify connections between sleep habits and health outcomes. The study, which leverages Apple’s ResearchKit framework, involves the collection of data on both healthy and unhealthy sleep behaviors that can be shared with other researchers in an open source format. The back-end data solution is provided via Watson Health for ResearchKit and will enable researchers to combine data collected throuh the SleepHealth app with a plethora of data sources, such as medical literature, treatment guidelines, and claims and clinical data.

Similarly, IBM last week announced a research collaboration with pharmaceutical giant Pfizer to develop remote monitoring solutions aimed at transforming how clinicians deliver care to patients suffering from Parkinson’s disease. By leveraging a system of sensors, mobile devices and machine learning, the companies will provide real-time, around-the-clock disease symptom information to clinicians and researchers, with the goal of better understanding patient disease progression and medication response to help inform treatment decisions and clinical trial design, while also speeding the development of new therapeutic options.

In its latest partnership, IBM and the American Cancer Society announced on Tuesday that they are joining forces to create the first patient advisor for people fighting cancer, powered by Watson’s cognitive computing capabilities. Still in development, the advisor will anticipate the needs of patients with different types of cancers, at different stages of the disease, and at various points in treatment, becoming “smarter” each time it’s used.

“It’s important to understand the needs through the eyes of the patients, with different cancers, at different stages, and different treatment plan,” said Rometty. “And, it’s about guidance—guidance on symptoms, support, wellness activities and education.”

More for you

Loading data for hdm_tax_topic #better-outcomes...