St. James Mercy Health System Achieves Quality Improvement, Cost Savings with Speech Recognition

Speech recognition technology continues to be under-utilized in the health care industry, as most organizations use the software in only one or two departments. But one organization is bucking this trend by taking an entirely different -- and more universal -- approach. St. James Mercy Health System (Hornell, NY), an integrated delivery network in western New York and longtime MEDITECH customer, is utilizing speech recognition technology throughout its affiliated physician practices. They chose Peter Cohen and Associates (PCA) of Natick, MA, MEDITECH's Preferred Speech Recognition Solutions Integrator, to implement a speech recognition solution in physicians' offices. PCA's unique approach, which includes Dragon® NaturallySpeaking® software combined with their own MEDITECH-compatible applications and necessary services, enables St. James Mercy to both improve patient care and save money.

Surprisingly, the push for speech recognition at St. James Mercy primarily came from its practice physicians. The physicians were using the MEDITECH/LSS Physician Practice Management (PPM) suite, but they were sending their report dictations to be transcribed by the hospital's backlogged transcription department. The physicians were waiting hours, sometimes days, for their transcriptions. In addition, providers were directly paying for this service. Speech recognition offered solutions to both of these problems -- immediate report availability and cost savings. Now the physicians use speech recognition technology to convert their dictations directly into data that is immediately accessible in the MEDITECH system.

"Slow turnaround time delays patient care," says Denise Becher, applications analyst at St. James Mercy. "The report timeliness gained through the immediate accessibility that PCA's speech recognition solution provides improves quality, and the physicians are very pleased with the cost savings, too."

Becher provides an example of the cost savings that physicians have experienced. One physician super user saved $2,000 in transcription expenses during July and August. Although $2,000 is a substantial cost savings for any provider, it is an artificially low figure, because the summer months were slower than normal for the practice due to summer vacations. Actual cost savings are almost always higher.

"This physician reached 100% usage of the software in only two weeks. It's time well spent," Becher notes.

The investment needed to implement PCA's solution was relatively low compared to transcription costs. Peter Cohen, president and founder of PCA, says, "The physicians realized return on investment within 60 to 90 days of going LIVE and utilizing the technology."

In order for health care organizations to maximize speech recognition's potential, Cohen advises them to prepare early and to designate someone for physicians to turn to as a resource.

"Having an individual dedicated to the success of the implementation makes all the difference," he says, "and St. James Mercy really exemplifies this. Denise was committed to the project from the very beginning."


A Solution Designed with Physicians in Mind
St. James Mercy implemented PCA's Direct Dictation Model, a real-time, front-end solution which allows physicians to dictate directly into a microphone. The practice physicians wear microphone headsets plugged into their office PC. This model requires that every PC used for direct dictation have a copy of Dragon NaturallySpeaking software.

A typical day begins with the physician opening up his or her copy of Dragon NaturallySpeaking. After treating a patient in an exam room, the physician goes back to the office to dictate the history and physical, notes, and more. The dictated information is available immediately in PPM's Provider Encounter feature, appearing seamless to the user. The physician is then ready to see the next patient.

Physicians do not have to be computer literate to use the software. According to Cohen, the most important factors in using speech recognition proficiently are the manner and speed with which one speaks.

"Mumbling, slurring, and other habits can affect accuracy, as well as speaking too fast or too slowly," he says, adding that every user develops their own equilibrium between maximizing accuracy and minimizing misrecognitions.

PCA offers software to use in conjunction with Dragon NaturallySpeaking for even greater efficiency. St. James Mercy uses PCA SAVE (TM) (Specialized Additional Vocabulary Extensions) and PCAMEDIMACS. PCA Save is a collection of specialty vocabularies (e.g., Emergency Department, Radiology, and Pathology), that save physicians time by increasing their accuracy through preloaded words and phrases that are likely to be dictated. PCA MEDIMACS(TM) includes voice-enabled command and control for the MEDITECH HCIS, as well as various report templates. These components expedite the process by providing quick navigation through MEDITECH screens, required fields, and dialog boxes.

"Voice commands are essential," says Cohen, "because they dramatically reduce the time it takes to enter a report. The screens flash by when physicians use them."


Incentives and Adoption Strategies
St. James Mercy identified a couple of key considerations for achieving adoption of speech recognition.

"Users have to have an incentive," Becher says. "They need to make an initial time investment in order to train the software to understand them. The more physicians use the software, the higher the accuracy rate. Although more time is initially invested in the dictation process, the data's immediate availability makes up for the investment.

"All in all, the challenges are not difficult to overcome. Physicians find the software well worth it, because it provides real-time documentation and saves on expense."

St. James Mercy and PCA are still training physicians at affiliated practices to use speech recognition. Becher says they are getting used to the new process and learning to use the templates. Currently, the practices are the only environments in which physicians use the software, but St. James Mercy plans to train their hospitalists in the acute care setting.

Although most of St. James Mercy's house staff is still using hospital transcriptionists, speech recognition has eliminated the influx of reports from the physician practices, opening up resources for other providers and reducing delays in processing their dictations. Add to that the immediate availability of information, improved patient care, and cost savings, and it makes a compelling case for the technology. "In the next five years," predicts Becher, "everyone will be using speech recognition."