November 26 , 2007

Are Your Doctors Ready for Speech Recognition?
By Christopher Rehm, MD
For The Record
Vol. 19 No. 24 P. 6

The emergence of speech recognition (SR) technology has changed the way many healthcare facilities approach medical transcription, but the limitations of today’s products still leave some frustrated; yet others are achieving phenomenal results—even doubling productivity while significantly reducing costs.

To achieve maximum benefits, SR must be strategically deployed. A phased rollout with the right physicians is critical—especially with back-end SR, which many mistakenly assume never requires any changes by clinicians.

While back-end SR does not involve the physician in the editing process as front-end does, some physicians may have to adjust their dictation practices or ad hoc individual formatting requirements. For this reason, you should carefully choose which physicians should be the first in your organization to make the switch.

Back-end SR has achieved far greater market penetration than front-end, at least partly because doctors’ participation and time commitment does not change. However, those physicians willing to adapt their behavior will likely realize the greatest gains. Once a small number of select physicians has demonstrated the effectiveness of SR technology, others are more apt to be open to change.

According to industry experts, SR will not completely replace traditional transcription in the foreseeable future, and the SR saturation point is likely to be roughly 60% without requiring any change in physician behavior.

But how can you identify which of your physicians are good candidates? First, identify high-volume dictators and focus on those with a history of good dictation. Typically, whether in a large hospital or small practice, 20% of the doctors are responsible for 80% of the dictation files. It is that 20% who have the most to gain from SR technology.

Since they tend to monopolize resources, these physicians have the most potential for efficiency gains. Disregard those who generally resist change or shun new technology. There are plenty of doctors who readily embrace anything that can make them more productive.

The back-end SR process eliminates the need for a medical transcriptionist (MT) to type out each word of dictation. Instead, SR software will provide the MT with a verbatim script. At this point, the MT becomes a medical language editor, which not only boosts productivity and enhances turnaround time but also improves the quality and accuracy of the completed document.

It’s always good practice to assign MTs to specific physicians. Whether employing SR technology or traditional medical transcription, an MT is fastest when he or she is accustomed to the physician’s dictation style, speech patterns, and habits.

Achieving the Best Results
Physicians must be willing to use dictation technology that produces the highest quality of voice recordings. A low-quality voice file will result in a large number of errors that must be corrected by the editing MT, limiting productivity and turnaround time.

Doctors must also comply with established protocols for proper dictation. For example, the facility must require the physician to enter dictation IDs, patient information, and job/work type information at the time of dictation.

They must also eliminate background noise that could hinder the quality of the voice file or confuse the SR software. This includes background conversations, noisy medical equipment, and eating or drinking while dictating.

It is important that physicians dictate clearly and consistently. Although SR technology is capable of learning and adapting to a particular speech pattern, enunciation will improve the accuracy of the computerized transcription and minimize editing. Further, doctors should dictate in an organized manner according to established protocols and templates rather than making special requests of the MT or editor.

While the majority of physician dictations will not require punctuation and formatting instructions such as bullet points, some unique speech patterns demand that level of detail to see maximum productivity. The completed document will be much closer to verbatim than the physician may be used to receiving, especially if he or she typically relies on an MT to correct slang terms or make other minor adjustments for clarity.

Ensure a smooth implementation by investing the necessary time and resources to educate physicians on dictation best practices before deploying SR technology. The benefits of SR are real, but they require careful planning to achieve. By approaching the transition with a sharp focus on how SR could impact clinical documentation workflow, healthcare organizations will have increased the potential for results that exceed expectations.

— Christopher Rehm, MD, is chief medical officer at Spheris.