TECHNOLOGY FOR CHANGE

Speech Recognition May Speed EMR Adoption

Ericka Chickowski | Date: 08-28-09 | Comments

With a five-year adoption mandate of electronic health records hanging over health organizations, they are faced with many hurdles�but speech recognition may very well be the critical enablement technology they have been looking for.

Table of Contents:

Speech Recognition May Speed EMR Adoption

As health care organizations strive to improve the way they handle patient records in the digital realm, many are struggling with the ultimate electronic medical records bugaboo. That is, how the heck do they get the doctors to actually use EMRs?

�The big challenge is getting the doctors to use the electronic health records,� says Dr. John Halamka, CIO at Beth Israel Deaconess Medical Center, who cites American Hospital Association numbers that show only about 17 percent of physicians in the United States currently use EMRs.

A study by the New England Journal of Medicine of 3,000 hospitals in April shows even worse adoption rates�only 1.5 percent of all non-federal U.S. health care facilities use a comprehensive EMR system, and just 8 percent have EMRs installed in at least one unit.

Federal lawmakers are trying to promote better adoption of EMRs in order to develop the infrastructure necessary to reap the benefits of pervasive EMR use. With the passage of American Recovery and Reinvestment Act (ARRA) of 2009, the government plans to disperse $36 billion in aid to promote EMR investments nationwide.

But even though the average 500-bed hospital will expect to rake in about $6 million of those funds if they implement EMRs by 2011, that will hardly cover the expense of a full deployment, according to PriceWaterhouseCoopers consultants.

�PricewaterhouseCoopers� analysis shows that the stimulus incentives to comply with the new requirements for purchasing, deploying, and maintaining interoperable EHRs [electronic health records] do not come near to compensating the overall costs,� a recent report by PWC reads.

Unfortunately, though, health care organizations are caught in quite a pickle because if they don�t comply by 2015, they face severe decreases in Medicare and Medicaid funding as stipulated by ARRA.

That five-year adoption mandate is quite the daunting prospect considering that the IT departments tasked with supporting EMRs are bleeding budget dollars by the bucketful in the current downturn. Not only that, but organizations cannot simply make due with a cockeyed, checkbox compliance approach to deployment due to the mandate by ARRA that organizations employ "meaningful use" of EMRs.

Clearly, organizations need to find a way to not only install EMR systems but to also promote physician adoption in as economical a way as possible. While there certainly are a number of back-end IT systems that need to be squared away to support these efforts, Halamka believes that the key piece of the puzzle lies in the front end. In other words, how physicians enter their information into the EMR system. If the system is too unwieldy, it is doomed to collect cobwebs. If it�s too expensive to maintain, it�ll sink the organization.

Which is why he believes speech recognition very well may be the critical enablement technology to help organizations meet the ambitious 2015 EMR deadline.

According to Halamka, many hospitals and other health care organizations face the prospect of either having doctors record patient information manually via computer terminals or through dictation. Many doctors despise having to enter information for each interaction through mouse and keyboard because it cuts into their practice time and because many old-school doctors are simply not comfortable with computers. Physicians would much prefer to dictate their thoughts on patients, but there�s a catch: Traditional medical transcription is extremely costly.

�The challenge is this: The cost of transcription is high,� Halamka says. �The more people you tell to dictate notes, the bigger your transcription budget. That is made worse by the fact that sometimes the doctors� dictations ramble compared to if you were typing a report where you�d focus and put it in an orderly template.�

At Beth Israel, Halamka has tackled the problem by focusing on delivering a way to give physicians an entry method they can live with while eliminating the costs of transcription. The solution is the deployment of speech recognition and medical note automation software from Burlington, Mass.-based Nuance Software.

The server-based program enables the doctors to either call in reports to the system or dictate with a handheld recorder and upload to the system without Halamka�s team having to maintain server-based programs. The software takes disorganized speech and, based on keywords spoken by the doctors, automatically organizes it into specific templates. And it does all of this with about 97 percent to 98 percent accuracy.

According to Halamka, the initial implementation cost $500,000. But since the organization started phased roll-uts five years ago, Beth Israel has saved more than $5 million in transcription costs alone.

Next: Doctors Love Ease of Use >>