Speech-to-Text Dictation Technologies: 101

Understanding the basics of speech to text technology in healthcare
By
Arjun
Narayan
April 25, 2024
Table of contents
Get healthcare innovation updates straight to your inbox
Subscribe
Thank you for subscribing!
Something went wrong. Please try again.

Over the last few years, clinician burnout has rapidly increased as the stresses and pressures brought on by COVID-19 have caused many doctors to reconsider their professions. A major contributing factor to this burnout is the increased amount of administrative work that physicians and other medical professionals must perform (estimated to be ~16 hours / week – we’d go insane as well). However, the development of novel speech-to-text technologies that could ease the level of physician administrative work has the potential to dramatically transform the medical profession.

Speech-to-Text Technologies

Speech to text technologies refers to the transcription of a user’s words into written text. Early applications of these tools include household names such as Siri and Alexa, but modern technology has since evolved in complexity, allowing for broader use across industries. Speech to text technologies have particularly potent applications in the medical field, where the increased use of dictation to facilitate administrative tasks can alleviate clinician burnout, ease the burden of administrative tasks, and ultimately improve patient outcomes. Today, companies such as Nuance and Ambience have created AI systems that have transformed provider workflows, eliminating time spent on administrative duties and allowing for more successful patient outcomes.

Medical Applications of Transcription

Note-Taking

A key element to successful patient outcomes is a positive relationship where the clinician is able to focus completely on the patient without having to worry about transcription or other administrative processes. While this might have once necessitated having another staff member such as a clinical or administrative assistant in the room, the rise of speech to text technologies has now transformed the way doctors can interact with patients. Clinicians can now have a direct, face-to-face conversation with patients as they examine them while speech-to-text technology transcribes their commentary as notes.

The notes that the doctor records can also play an important role in care management. After patient appointments, the notes can serve not only as a record of patient history and behavior, but can also be used to notify doctors or patients about follow-up appointments, prescriptions, and other key information essential for proper care. As more of these administrative functions are electronically handled, the application of this technology into clinician systems and workflows can facilitate a more integrated, holistic care system that further eliminates the need for increased administrative work.

EHR Integrations

The notes that doctors take from patient interactions serve another important purpose: they can be the basis of reimbursements from insurers. As such, the integration of speech to text solutions with EHR systems serves an important function in easing the physician administrative burden. While improving patient note-taking can simplify part of the administrative process, if an administrative assistant or physician needs to parse through notes to clarify what happened during a patient’s visit, the application of this novel technology is rendered ineffective. However, if the notes from a patient’s visit integrate cleanly into all health records and serves as a single source of truth for both a provider and payor, the entire health system is better off and more efficient.

AI’s Influence

The popularity of speech to text medical technologies has coincided with the increased power and capabilities of revolutionary artificial intelligence (AI) models. The increased application of AI in speech to text technologies has the potential to make provider’s jobs more patient-centric and less focused on burdensome, tedious administrative tasks. Companies are increasingly using non-confidential PHI to train models to develop advanced behaviors and perform more complex tasks, clinicians can outsource an increased amount of their non-patient care to technology, leaving more time for direct interactions with patients.

Challenges of AI Transcription

Speech to text technologies also come with their own unique challenges that must be addressed. A study of AI scribes in Kaiser Permanente hospitals in Northern California found that while the technology scored an average of 48/50 in 10 quality metrics, doctors still had to do some extra work to ensure note accuracy. Additionally, transcription solutions may sometimes miss essential details crucial to patient well-being, resulting in extra time and effort on a doctor’s part to reconcile missing information. These inaccuracies can also pose a challenge as pertains to insurance billing. Given that the notes can serve as the basis for insurance reimbursement, any inaccuracies or missing information can affect providers’ ability to be paid for treatments they provide.

These technologies may sometimes cost thousands of dollars for providers, ****meaning that smaller provider groups and doctors might be priced out of solutions that could save them money and time and allow for better patient outcomes. As the industry begins to evolve, it’s important to recognize that different solutions may fit different providers and provider groups, based on size, scale, and features that may be more relevant to some providers than others.

Conclusion

Speech to text technologies already pervade every step of our lives. Now, with the AI revolution in full swing, we must figure out how to best integrate these technical solutions within the healthcare industry in order to improve patient outcomes and physician burnout. While transcription solutions clearly have their challenges and are being improved upon daily, they can save physicians time and money, leading to improved health outcomes and greater provider satisfaction. Ultimately, time and technical experimentation will truly tell whether these solutions make for a better patient experience.

FAQs about Speech-to-Text Technologies

Do speech to text technologies eliminate the need for further note-taking / provider data entry?

No, as these technologies are still being developed, doctors will still need to manually enter information about items such as treatments and prescriptions that may not have been discussed with patients or recorded already. The goal of these technologies is to allow the clinical experience to be focused on the patient; however, EHRs and other processes in the healthcare ecosystem must keep up so that practices do not waste time with these workflows and continue to evolve.

What are some other barriers to implementation of speech to text technologies?

Barriers include difficulty or lack of willingness to implement transcription technologies, either from a technical perspective or due to cost. Ultimately, the technology must be right for the physician or practice; if it does not save them time, no matter the technical capabilities, it will not be worth it.

What are further applications of speech to text technologies?

These technologies could facilitate scheduling, calls with pharmacies, or even remainder phone calls to patients about medications or appointments. These technologies should ultimately be implemented in as many effective ways possible as long as the human nature of the patient interaction is preserved while doctors’ administrative burden is reduced.