From the mHealthNews archive

Mobile devices and NLP technology: 4 selling points

By Benjamin Harris

Remember when doctors walking down halls, talking into tape recorders like Agent Dale Cooper from Twin Peaks? Now they're holding conversations with their mobile devices, taking a page from David Bowman and Frank Poole's interactions with HAL 9000 in 2001: A Space Odyssey.

Mobile devices such as iPhones and Androids offer both opportunities and challenges for physicians. Critical EHR data is accessible almost anywhere and near instantly, and patient notes can be recalled with a few taps on a screen. But minimized screen real estate is at a premium. What data should be shown? And that's to say nothing of the challenges of entering the data. Typing full notes on a mobile phone can be a carpal-tunnel inducing strain.

Jonathon Dreyer, director of mobile solutions marketing at Nuance Healthcare, says these strengths can be improved – and the weaknesses eradicated – with voice recognition, cloud-based applications and natural- or clinical language understanding technology. He says these tools create "better access for the physician," who can have an interactive dialogue with his or her device to access and create medical records on the fly.

Dreyer offers four reasons voice recognition and mobile devices were meant to be together.

1. Speech to text. "There's no question that these mobile devices are great for consumption of information, but when it comes to generation ... they fall flat," says Dreyer, who says he has noticed more people attaching external keyboards to their mobile phones, effectively turning them in to mini laptops. That's contrary to the point of a mobile device, he argues, and "even with a keyboard I can't imagine a physician entering data that way."

Speech-to-text services for mobile devices resolve that problem. With doctors becoming comfortable around dictation, this enables them to enter patient notes or clinical information directly into their device. Dictating to a service that can automatically convert a doctor's speech to text allows them to deliver notes in a conversational style. It frees them from typing on a small keypad and it means that they may include things they'd forget to type.

2. Custom commands and navigation. "You've got so much information in an EMR, especially within the confines of a four-inch screen, it becomes really difficult to present that information," says Dreyer. While software vendors are getting better at choosing what data to display and when, he envisions a better approach. Voice recognition lets physicians "have a free-form and flowing conversation" with their devices, "like they were actually talking to a person at their side."

Dreyer references a study that found 81 percent of doctors own a smartphone, and a majority use their devices to access reference materials at the point of care. "A physician can simply say, 'Show me my patients for the day,' or, 'Show me Mary Smith's info,'" and have that information brought up while they're with the patient, he says. That level of flexibility can make a physician's workflow much more efficient and allow them to devote more time to patient care and less to retrieving and looking through records.

3. Clinical language understanding. Natural or clinical language understanding is a process that can pull relevant medical information out of a narrative conversation and convert it to actionable data that a computer can act upon. A physician "can take speech-recognized text, run it through a CLU engine that extracts clinical data and pull structured information out of a patient narrative," says Dreyer. This allows a physician to concentrate more on getting an accurate narrative from the patient, as opposed to asking routine questions. Having a CLU to sort all of the data "allows physicians to document the patient's whole story," he says. Capturing the entire narrative means the resulting care will be better. "You don't want to force patient narratives in to a template," he says.

4. Future developments. What if a speech recognition and CLU system could translate a doctor's notes in real time, realize a prescription is needed and send that prescription order to a pharmacy? "That scenario is totally feasible," says Dreyer. "For as much as there is unknown (in medicine), there is a lot of known." With actions and protocols centered around so many day-to-day routines, integrating voice recognition and CLU systems with a physician's commands could simplify and streamline workflows.

Nike Epic React Presto 19SS Olive Green Yellow Black White AQ2268-004