AU & NZ
SG
MY
US
IE
AI Note Taking
Digital audio capture is nothing new, nor is transcript generation, but the ability to use artificial intelligence to enhance these recordings through note summarisation is a relatively new element to the technology, developed in recent years.
AI scribe software is quickly populating the healthcare solution market because of the time-saving it can offer in the data capture and note taking, which in turn allows clinicians to focus on the patient or client they are meeting.
What is natural language processing technology?
These note-taking or scribe solutions work by using Natural Language Processing (NLP) algorithms, but what does that actually mean?
NLP is made up of three components: computational linguistics, machine learning, and deep learning. Together these make up the Natural Language Processing algorithms and each part is crucial to the generation of the transcribed text.
Computational linguistics is the oldest of the three models and is the complex interpretation segment of the solution. It analyses the human language detected – regardless of language – to understand the information within and the semantics or context inferred. Search engine prompts, speech recognition, and even virtual assistants like Alexa or Siri utilise these computational linguistics are part of their operating protocols.
Machine learning is the next part, which is very simply explained as repetitive exposure to subject matter. The machine aka the software is exposed to sample data to improve its understanding through pattern recognition and correlation. This is particularly important for the elements of semantics and context within human speech, with elements like sarcasm or anger to consider, as well as colloquialisms or metaphors.
Finally, deep learning. This is the part that we really associate in our minds with artificial intelligence. This is computers ‘thinking’, and making decisions or responses based on the prior pattern recognition and deciphering of input information – the audio recording.
There is a risk, however, in that generic AI note taking solutions aren’t suitable to be an AI healthcare assistant. Clinicians need dedicated or bespoke medical AI scribe software to accurately handle documentation in healthcare, else they could incorrectly record and report conversations and cause problems – even harm – for patients or clients through inappropriate care and treatment. There’s also the safety of patient data to consider. It’s imperative to implement medically-approved solutions that guarantee the proper safeguards for the highly personal and highly private information gathered.
What is Ambient Voice Technology?
Ambient Voice Technology, also referred to as Ambient Listening Technology, is the next stage in the evolution of AI scribe software. Whereas scribe solutions typically act as a catch-all for detectable background sound, solutions using Ambient Voice Technology offer more precise capture of audio.
This is important for medical AI scribe software because accuracy is crucial in delivering safe and appropriate care. Ambient Voice Technology is more capable of identifying different voices and correctly parsing sentences based on the subject matter.
Solutions like Access Evo SmartNotes can automatically capture conversations and summarise them for clinicians in the form of high-quality clinical notes. This is done in a fraction of the time usually assigned to note taking and updates to patient records.
Allocation of clinician time
According to the Health Innovation Network, the NHS handles more than 1.75 million patient appointments daily – which means there’s a large quantity of patient notes and large spells of time required to accurately recount and document them.
“Clinician burnout and workload is a national emergency impacting on patients and so bold moves have to be made. AVT could be the single most important technology we can deploy to improve this but if rapid progress is to be made, it must be shaped nationally rather than at ICB level.”
Health Innovation Network – Ambient Voice Technology discussion - 2023
Testimony from multiple sources points at roughly 30 minutes spent transcribing and uploading notes for every 60 minutes spent on clinical duties. This is very obviously unsustainable and an inefficient use of clinical time, and in some cases healthcare professionals are coming in at weekends or staying after hours to catch up, which in turn eats into their free time that is supposed to be for their work/life balance and destressing.
A big part of why this time is so substantial is because of ‘functional creep’. This phrase refers to the steady increase of responsibilities that fall upon a worker. In this case, we’re referring to clinicians doing more and more than their core remit of diagnosis and treatment. That’s not to say that clinicians shouldn’t take notes, and shouldn’t ensure accurate patient records, and that they shouldn’t assist with other duties elsewhere in a hospital or similar healthcare setting, but all of these necessities or things that benefit from their input ultimately lumber them with an ever-growing workload that eventually leads to burnout.
This burnout is hurting staff retention within the NHS, and the inability to retain staff means new recruitment, new training, and greater delays/waiting lists whilst there’s reduced employment. All of these things increase costs at a time when budgets are extremely tight.
The solution is solutions: using healthcare software to alleviate the burden of functional creep upon the workload and to unleash greater clinical productivity.
The Benefits of Ambient Voice Technology
Ambient Voice Technology delivers intelligent document processing in healthcare. Increased efficiency in the clinician’s workflow, greater accuracy, better note recall, and easier collaboration are all possible with the majority of the AI scribe software, but Ambient listening technology supports documentation in healthcare in a way that AI note taking does not.
We mentioned before the issues of distinguishing between voices and only recording what’s immediate and specifically required. Ambient Voice Technology ensures this greater relevancy and is effectively AI note taking 2.0.
In an article by Great Ormond Street Hospital about their AI tool pilot, they report trying to trick their solution with accents, loud background noise, and irrelevant information about the X-Men comics. Despite giving incorrect information and using loud background noise, the AI assistant was able to correctly understand the information shared and draft clear clinical letters. GOSH also references the testimony of a young patient’s mother, with the pair recently having an appointment with the technology.
“It was great to see the tool in practice – during our consultation with Dr Kusters we were able to ask all our questions without any interruption thanks to the tool picking up our discussion.
"We’ve always had a great relationship with our clinicians, but this time we felt like full focus was on us. It was really positive. Due to Bea’s medical needs, we attend many appointments at different hospitals and can see this making a difference.”
AVT is able to give clinicians more face time with patients by reducing the time spent on the administration surrounding note taking. Many doctors have reported how they often face their computer to type as they listen, to ensure greater record accuracy. This isn’t an awful practice, but it’s not very engaging with the individual in need of care, which can feel alienating.
Greater clinician focus encourages greater patient engagement. As the above GOSH anecdote shows, the patient and her mother felt more positive about the interaction featuring the Ambient Voice Technology. It wasn’t intrusive and it allowed the doctor to be more involved. Body language and eye contact matter in human interactions. The resulting positivity felt helps breed trust, helps deliver greater adherence to treatment plans, and leads to quicker responses to reporting any chances in health and wellbeing which could become worse and more costly over time.
“I don’t think I behaved differently because it was being recorded. It was nice to feel the doctor had the time to speak to me. I found it easier to remember what was said.”
Patient Representative – Health Innovation Network
Stanford Health Care in the United States also had similar trials to GOSH. Their own preliminary surveys found that roughly 96% of their physicians found the AVT solutions easy to use, 78% found it improved clinical note taking, and 66% found it saved them time.
Niraj Sehgal MD, the Chief Medical Officer at Stanford Health Care, said: "This could have a transformative impact on the way we provide clinical care. AI tools will never replace the clinician, but it might replace parts of their workflow.
"As people get more comfortable using AI-powered technology, it creates a fertile ground to continue adapting other tools for the workplace that support providers, freeing them to better care for patients.”
Challenges of implementing new software
The benefits of AI note taking and Ambient Voice Technology specifically are very clear to see – but that doesn’t mean they are without their own challenges and risks.
There are concerns around artificial intelligence and securing patient data. Solutions must be properly vetted by healthcare providers during the research process to ensure that there are proper protocols and operating guidelines to maintain the same high levels of personal privacy as currently exists in the NHS and other healthcare organisations.
Another standout issue is interoperability. Standalone solutions can be functional, but integration matters. That interoperability (the communication and cooperation) between the Ambient Voice Technology and things like electronic patient records that allows for greater time saving for clinicians. Without integration, clinicians are still copying and pasting summaries into records and formatting the text. That’s laborious, and integrated solutions can be programmed to automatically fill out these records to turn admin time into clinical time.
Data might be more space-efficient than paper-based records, but large volumes of data can be problematic. Large-scale storage isn’t cheap, and there are energy costs associated in powering the hard drives or servers that these files reside within, and space to house them, and security coverage both in terms of digital and real-world protection. Typed notes are some of the most space efficient files around, but again we refer to the earlier point about how the NHS handles more than 1.75 million patient appointments daily.
Even with very small file sizes, the accumulation of these appointments for a surgery or a hospital over time will add up. This then poses a question about deleting full-scale notes after a certain time period and keeping a brief summary, or offering the patient the choice of preservation, or keeping them anyway and handling the burden.
Then there’s the adoption of this document management software for healthcare. It’s a well-known fact that some staff members are hesitant to adopt new solutions, be that due to age, familiarity with other solutions, habit, or a lack of digital literacy. Work must be done to convince people, be that through education or enforcement, and that can have side effects in terms of cost, time management, and staff morale. There’s a delicate balance to the work environment and new elements always risk disturbing it.
The Ambient Voice Technology, for all its perks, may be a victim of its own success too. Stanford reported that approximately 66% of clinicians found they saved time – so why did a third not find savings? Well, if notes are more accurate there’s a good chance of more information being captured. Regardless of typing them up, the clinician still needs to read through to proof the copy, and that increased reading time could just sub in for the time they were initially typing. There’s also the risk of longer notes missing the key points raised by a patient in their consultation.
Document automation for healthcare
The conclusion then must be that Ambient Voice Technology is a very strong software solution that has removed a lot of the hindrance elements of AI note taking as a broad tool. The increased accuracy of audio capture and distinction, as well as the language processing, mean it’s an incredibly useful support function for clinicians.
The important thing is to remember the distinction of pursuing specific healthcare-friendly ambient voice technology (or ambient listening technology) rather than generic note-taking software. Healthcare information is too valuable to permit mistakes in the transcription process, which is why solutions like Access Rio Smart Notes are the way forward.
In an article by Forbes magazine, published in January 2024 on the subject of ambient technology, they highlighted the importance of integrating this technology with EPR/EHR solutions to enhance provider efficiency, but they also pointed out that while there might be cases where clinicians don’t save time, there will always be a benefit to the patient in terms of the engagement they get from the healthcare professional and the access to the notes of the meeting and their records in general.
Even if time isn’t saved, the patient experience is improving, and that will lead to cost savings elsewhere within an organisation that will come back round to benefit the clinician in due course.
