Cognitive computing will bring increased intelligence to EHRs

The typical electronic health record (EHR) system holds more information than any one person could analyze, as do all the medical journals and medical data repositories that exist.

But physicians could soon leverage all of that information to make better decisions, according to leading health IT experts, as EHRs and other healthcare software systems begin to incorporate cognitive computing.

Cognitive computing, a branch of artificial intelligence, harnesses self-learning systems, data mining, natural language processing and other technologies to analyze information, identify patterns and draw conclusions – just as the human mind does, only on a vastly larger scale and speed.

The result will be computers that act more like virtual assistants than data-entry systems.

 “Cognitive computing will impact how we deliver care, it will impact clinical workflows and it will impact other spaces within the physician business as well,” said Todd Evenson, MBA, chief operating officer at the Medical Group Management Association (MGMA).

 “It’s able to identify outputs that a physician didn’t necessarily think of,” said Ian E. Hoffberg, applied innovation manager, health information systems, for the Healthcare Information and Management Systems Society (HIMSS).

Health IT experts offered a vision of how cognitive computing will work in a typical physician’s office:

Task-automation software, known as a “bot,” could answer calls and take information, directing callers to the right person or to a computer system where the caller can upload health data to be analyzed; the system could alert staff to callers who need immediate attention based on the analysis.

In the exam room, physicians will interact with their EHRs on a new level, dictating notes while receiving back in real time information guiding them through the exam and treatment options based on all available information that the computer can access.

“Cognitive computing can take you beyond providing advice based on a set of knowledge. This is where computers start to see relationships among data that you and I don’t have the capacity to shift through. That’s the real next step,” said Rod Piechowski, MA, senior director of health information systems at HIMSS.

Structured vs. unstructured data

Experts noted that EHRs already incorporate some reporting and analytics tools. But EHRs generally use structured data, that is, highly formatted information entered into set data fields. Examples of structured data include blood pressure and weight, which clinicians enter into fields specified by their systems.

Cognitive computing, however, can search and analyze unstructured data, such as clinician notes and medical journal reports, to look for patterns and trends that can impact diagnoses and treatment plans.

One drawback to cognitive computing, however, is the length of time it will take to arrive in the office. Health IT experts said major vendors, startup tech firms and medical institutions are working with technology, but said it will be years before cognitive computing is widely available in EHRs and other systems used in physician practices.

But when it arrives, Evenson said he expects it will allow the physician to do more of the work they want to do, and less administration.

“Doctors feel the EHR made their jobs harder and took them away from direct contact with patients,” he said. “But cognitive computing should be an opportunity to give clinicians that time back and allow them to focus on the human interaction. It will let them work on the highest level of work.”