The productive workflow
In my last blog, I highlighted the importance of speech to text to support clinicians recording data in an electronic format in the EPR. I also described the role that AI could play in meeting the aspirations of the NHS Long Term Plan.
The considerable progress that has been made in AI is due to a combination of increased computing power combined with the huge quantities of data being generated. Without the collection, storage and availability of all this data, these major developments in AI would not have been possible.
AI and speech recognition
There are two ways that AI driven speech recognition is making a difference. First, AI uses acoustic (sound) and language learning programs (algorithms) to interpret voice. Secondly, AI supports the mathematical backbone to speech recognition.
It is the sophistication of this backbone that determines the accuracy and quality of speech recognition and the subsequent ability to automate some of the clinical and administrative tasks that burden clinicians today.
How AI is making a difference in documentation
The key workflow task that AI-driven speech recognition has addressed today centres around the documentation process; admission notes, ward round notes, procedure notes, discharge summaries, outpatient letters etc. The simple fact is that we speak three times faster than we type. However, the speed, ease and freedom of movement that speech recognition provides is only the beginning of AI’s impact on clinical documentation.
Through computer-assisted clinical documentation AI has the potential to bring knowledge automation to the point of care. By searching clinical terms in all documents through the patient care pathway, AI supports the clinician by providing advice and guidance within the workflow and detects possible missing diagnoses that have not previously been recorded or considered, prompting the clinician to amend the notes.
The next step – coding
If clinical coding can be an automated as a bi-product of this AI powered speech journey then it will also help clinicians meet the requirement for SNOMED CT coding by April 2020.
AI powered speech recognition then not only improves the quality of the clinical documentation, but will have an impact on coding and appropriate reimbursement.