IBM Launches Health Care Analytics Software to Reduce Hospital Readmissions

IBM's new Content and Predictive Analytics software uses technology similar to that of the Watson supercomputer to help doctors spot patient trends and improve care.

IBM has introduced Content and Predictive Analytics for Healthcare, an application that provides content analytics like that of IBM's Watson, the supercomputer made famous on "Jeopardy."

Big Blue has developed the application to help the health care industry reduce hospital admissions, IBM reports. The company announced the new software on Oct. 25.

IBM bills its Content and Predictive Analytics for Healthcare application as "ready for Watson," which means it lacks the supercomputer's question-and-answer capabilities, but uses a similar understanding of speech.

"It is not Watson, but it does leverage the same natural language technology as Watson," Craig Rhinehart, IBM's director of enterprise content management strategy and market development, told eWEEK. "Instead of putting in a question-and-answer machine, we're running content analysis over natural language processing and then doing analysis."

Watson helps doctors know what questions to ask patients, and Content and Predictive Analytics will be able to show trends and patterns in unstructured data.

The new software will be used in the future to find patterns in health care information to allow Watson to ask better questions, Rhinehart explained. "The ability to visualize past, present and future scenarios gives you a much broader context to form questions from," he said.

In September, IBM announced an agreement with health insurer WellPoint to jointly develop commercial applications for Watson.

Content and Predictive Analytics will allow health care organizations to pull clinical data from large amounts of patient information to understand patient histories and predict future outcomes.

Seton Healthcare Family, a Texas health system, will be the first health care organization to use the software. By employing the application, Seton hopes to reduce readmissions, curb mortality rates and improve care through access to clinical and operational information in unstructured data, the health system reports.

With more than 80 percent of a health care organizations' data unstructured, the industry is looking for ways to analyze the data. Unstructured data includes physician notes, registration forms and discharge documents, IBM reports.

Content and Predictive Analytics will allow physicians to better understand patterns and abnormalities.

In addition to potential improvement in patient care, health care organizations will have a new means of reimbursement by using the analytics tool, according to IBM. It also provides built-in support for medical terminology.

Features of the software include searching, exploring, mining, monitoring and reporting.

The new application works with IBM's Health Integration Framework, which involves integrating data warehouses, business intelligence, master data management and advanced case management.

In addition, it combines IBM's Content Analytics, an application that analyzes text in medical records to support physician decision making, with SPSS Modeler Professional, a product that uncovers trends in structured data.

Content analytics is the historical analysis, such as a note that a patient had started smoking in the past or knowing what causes congestive heart failure. Predictive analytics forecasts future outcomes.

Health care organizations have used content and predictive analytics separately, but by combining them, Seton will be able to get a complete picture of which treatments can keep patients out of the hospital, according to Rhinehart.

"If you could only act retrospectively on the information, that's only half the answer," he said. "You have to be able to look forward."

By combining content and predictive analytics together, organizations will be able to analyze structured data, Rhinehart noted.

In a study, IBM could identify the 40 percent of respondents who were smokers through unstructured physician notes.

"The critical capability is to analyze the structured and unstructured together," Rhinehart said. "They smoked cigars, and at least 40 percent of the time you can find that in the unstructured text, but the EMR might still show they're a nonsmoker," he noted.

Only 1 in 5 hospital readmissions are nonpreventable according to a study in The New England Journal of Medicine, and the Centers for Medicare & Medicaid Services (CMS) within the Department of Health & Human Services (HHS) plans to penalize hospitals starting in October 2012 for high readmission rates.

Meanwhile, Southeast Texas Medical Associates is also using IBM analytics software to curb hospital readmissions and claims to have reduced them 22 percent.