Hospital Saves With Speech Recognition Tool

CIO — With more than 800,000 patient visits a year, Beth Israel Deaconess Medical Center (BIDMC) needed to efficiently document all the information those visits generate without busting the budget or placing extra work in the hands of the clinicians. With the trend toward electronic health records (EHRs) escalating, CIO John Halamka decided the cure was speech recognition technology and computer-aided medical transcription.

BIDMC selected such a solution in 2002, implementing it in phases between 2003 and 2005. Three thousand BIDMC clinicians now dictate and record patient visits into a handheld device that saves and sends a voice file directly to a voice recognition server. "In near real time, the voice files are processed into text and inserted into the [patient's] electronic health record," says Halamka. After the file is transferred, the hospital has "correctionists" review the notes for accuracy. The clinicians then review these corrections and sign off on them. In a traditional transcription process, clinicians would take notes by hand or voice recorder and then send the information to a transcriptionist who would type them into the hospital computer system. "The speech recognition software turns clinicians' dictations into accurate, fully formatted draft documents that are quickly reviewed and edited, often doubling productivity," says Halamka. The solution from Nuance, called eScription, cut records turnaround time from five days to less than an hour and has saved the medical center over $5 million since 2003.

To implement it, an IT project team at BIDMC worked with the vendor on user support and project and system management. They also spent time with the medical records staff to get them up to speed on the new system, says Halamka.

Some hospitals have held back on implementing electronic medical records despite the improvements they offer for the quality of health care. Privacy is one concern. But they also can significantly change physicians' workflow, says Forrester Research Director Eric Brown. By eliminating the typing step, voice recognition provides the benefits of the electronic health records and avoids extra work.  "Without addressing the usability issues for the physician, it's hard to imagine the U.S. reaching its goals of widespread electronic medical record adoption," says Brown. Halamka chose the eScription solution because there is little impact on physician workflow. "The improved turnaround time of completed medical records is the only operational indication that the eScription platform is present," he says. Also, since eScription is based on an ASP model, it required little capital investment.

Halamka's advice to other companies is to embrace the change toward voice recognition software. "It saves money and it is easy to maintain," he says.