Doctors Still Don’t Love Smartphone Dictation

SOURCE: InformationWeek

Digital dictation on smartphones is starting to pick up, but nowhere near as much as a recent mHIMSS blog posting suggests. The main obstacles include physician workflow, speech recognition accuracy, and doctors’ reluctance to edit transcription. But other factors, such as the ubiquity of mobile devices, their increased processing power, and the growth of cloud-based services, are expected to change the equation.

The mHIMSS blog said, “With more and more doctors looking to capture notes on their smartphones, the speech recognition market is booming.”

The market does appear to be growing, as more healthcare providers take advantage of speech recognition to reduce transcription costs. Nuance, Philips, 3M, and M*Modal are all putting out products, the blog noted, and Nuance has launched mobile dictation products over the past couple of years. Yet it is unclear how much of the market growth is related to physician use of digital dictation on mobile devices.

Rosemarie Nelson, a health IT consultant who works with physician groups across the country, said in an interview with InformationWeek Healthcare that she has seen an “uptick in adoption” of digital dictation on smartphones, but not a big surge in utilization. Because of hospitals’ security concerns about personal mobile devices, she said, she’d expect there to be less digital dictation on smartphones in inpatient care than in ambulatory care.

To the extent that doctors are doing more digital dictation, Nelson attributes some of the increase to “the natural adoption curve” of smartphones. In addition, she said, as more doctors have begun using electronic health records (EHRs), many of them have started dictating so they don’t have to take notes in point-and-click templates.

“Also, the smartphones are slicker, faster, better, and everybody’s got one. So that makes it easy to dictate, vs. the old PDA days when you were carrying a tool belt around with all these devices,” she said.

Nuance, which has the largest market share in speech recognition with its Dragon application, has been promoting digital dictation on smartphones since it launched its mobile version about two years ago. Yet Jonathon Dreyer, senior manager of mobile solutions marketing in Nuance’s healthcare division, could not tell InformationWeek Healthcare how many physicians are using the applications or even how much the market has grown.

Nuance offers two platforms for digital dictation on iPhones and Android devices. One of them allows physicians to dictate and transmit their voice files to a cloud-based transcription service that uses the Dragon software. The resultant text might be transmitted back to an EHR or some other documentation solution for the doctor to review later.

Nuance also has a third-party development platform that enables developers to add speech recognition to mobile applications for clinicians. After the speech is transcribed in the cloud, the text is sent back to the device. Dreyer said Nuance has 220 development partners, and a few dozen already have applications on the market.

Digital dictation was available on PDAs 10 years ago, but never caught on with physicians. One reason, Dreyer pointed out, is that physicians didn’t want to carry around PDAs and cellphones with them, whereas smartphones combine the two. Also, PDAs had a shorter battery life, poorer wireless connections, and less power than the current generation of smartphones.

The ability to use the cloud to do speech recognition has also opened up some new possibilities, he noted. For example, he said, a cloud-based service can run data-intensive applications such as speech recognition and image viewing simultaneously. Also, the cloud lets clinicians access their digital profiles–the precise waveforms of their speech–from any device or platform they happen to be using. In Dragon, that profile includes vocabularies for 60 different specialties as well as customized “macros” that speed up dictation, he noted.

However, while Dragon’s accuracy rate is 99% on desktop computers, it’s only 95% to 99% accurate on mobile devices, Dreyer said. Based on user feedback, he added, that seems to be related to the work environment in which smartphones are used.

As for doctors’ reluctance to edit speech-recognized text, he said that has always been the case for some clinicians. But they have the same problem, he pointed out, with conventional transcription, which also contains errors.