I’ve worked in the healthcare industry for many years, and the ways consumer trends shape health IT design have always interested me. Working for a company that represents both industries is an exciting place to be, but has also underlined the fact that there are fundamental differences between the two types of technologies. And, when it comes to clinical speech recognition, accuracy and personal health information security are extremely important.
Health IT goes to medical school, too
There is no doubt that speech recognition and Clinical Language Understanding (CLU) technology is incredible—it’s smart, self-improving, and reliable, but in addition to those fundamental capabilities, healthcare requires a more precise means of capturing and documenting the way a physician speaks—the words and jargon specific to clinical specialties —about the patients they care for in clinics or hospitals. That is why Nuance research and development teams created clinical speech recognition solutions that recognize medical vocabularies and use clinical ontologies to understand what clinicians say, how they speak, and the complex environments they work in. The result is a statistically tailored speech engine designed specifically for clinicians so that they have the most accurate and efficient way to document patient care whether they are in a loud emergency room packed with patients following a disaster or sitting shoulder-to-shoulder with patients dictating clinical notes during a visit.
“Humorous” or “Humerus”?
Medical dictation is a key component of patient care. When clinical information is incomplete or inaccurate, this problem lands squarely on the shoulders of clinicians, creating extra work, delays, and reviews amounting to the tune of 12.1 hours a week, according to a new study on the impact of clinical documentation. Physicians and care teams need to have accurate information in order to provide the best treatment and care pathways for their patients and that’s why we build technology that is continually updated with new medical terminology and rules to generate, on average, 38 percent fewer errors than non-medical speech solutions.
What does this all mean to doctors? A physician who dictates an injury to the humerus bone into a non-clinical speech solution might see it transcribed as “humorous” – and for a clinician seeing 40 patients a day that’s nothing to laugh about. A prescription of “Cardizem CM 120 mg one p.o.q. day” will result in something like “Carter’s MCM 120 milligrams 1 pm okay.” One additional speech recognition error per patient can take 30 to 45 seconds to fix, which adds reviews and corrections, translating to a doctor who sees 1-2 fewer patients per day, or works a longer day. Highly accurate clinical speech recognition technology allows clinicians to spend their much-needed time with patients, and has a significant impact on clinical workflow and provider revenue. That’s why we want to help providers get documentation right from the start.
Securely supporting physicians anywhere, on any device
Many failed HIT implementations and vendors have shown the importance of supporting clinical documentation and healthcare workflows with the appropriate technology, security, privacy policies, and procedures. And this is something we take very seriously, which is why we design technology specifically for healthcare professionals, including cloud-based offerings with end-to-end security, to support them as they care for their patients —whether in a hospital, clinic, or office environment, or roaming between care areas. The ability to take our deep knowledge of consumer speech trends and preferences, combined with the many benefits of clinical speech recognition technology, means physicians have anywhere, anytime convenience of seamless connectivity, and a system that truly understands them.