AVT: Why communication matters as much as regulation

A communications plan is a critical safety tool for ambient voice technology (AVT), writes Natalie Chishick, founder of Chishick Communications

The roll out of AVT has moved faster than almost any other digital tool in the NHS in recent years.

There is an expectation for the NHS to scale AI ‘at pace’. In some organisations – particularly in primary care – it is being embedded into day-to-day clinical practice and helping increase face-to-face patient time despite relentless pressure on the workforce.

Yet, nationally, there is growing recognition that AI tools, including AVT, need clearer oversight.

The Medicines and Healthcare Regulatory Agency (MHRA) is working towards a more comprehensive regulatory approach for AI-enabled technologies.

In certain places, AVT is starting to become business-as-usual. In others – particularly in acute trusts – the focus is squarely on frontline digitisation and electronic patient record rollouts. As a result, AVT is exposing gaps between policy intent and operational reality.

Capability vs reality

On the one hand, there are huge benefits to be had from extensive rollouts. A recent report by the Modality Partnership, the NHS’s largest GP super partnership, found that the AVT tool their staff had been using halved the time GPs spend on clinical documentation.

Likewise in the US, where the technology has been embedded for longer, research from Kaiser Permanente found that 50% of doctors didn’t amend anything in the AVT-written notes six months after implementation.

However, recent anecdotal examples from the NHS have shown blurred boundaries.

Clinicians are discovering that some ambient voice tools are having hallucinations or generating content that edges into clinical decision support without the necessary MHRA regulation.

While government expects progress at pace, the workforce doesn’t feel effectively supported in practice

It’s not possible to determine whether the errors arise due to user behaviour or technology issues, or both, but they are critical factors that need to be addressed.

As one chief information officer pointed out: “You are on your own if you try something and it goes wrong.”

While the government expects progress at pace, the workforce doesn’t necessarily feel effectively supported in practice.

With regulation and governance still evolving, communication needs to be considered as a critical tool for success and patient safety.

Incident communications

Most organisations rolling out ambient voice have focused, rightly, on procurement, information governance and clinical safety sign-off.

However, what happens if there is an issue? How is it identified and escalated? How are staff briefed while the facts are still emerging? What is said to patients if the clinical record is affected? Have plans been agreed with the supplier on their role and the necessary support?

A practical incident communications plan doesn’t need to be overcomplicated or time-consuming. In many instances, it can be an addendum to an existing preparedness plan, if one is already in place.

But without a communications plan, it will take longer to identify, action and resolve certain issues effectively and appropriately. It can be the difference between managing an internal issue and risking patient harm.

And with NHS staff under so much pressure on the frontline, as well as tackling competing digital priorities and having some tools already in use, it’s understandable to think that these factors may not have been considered.

Model for success

Suppliers can and are playing a crucial role in supporting their NHS customers on this journey. For example, by explaining capabilities, limitations and licensing conditions in plain language as part of default training and support rather than treating it as a one-off message or only using it in marketing assets.

This, combined with the robust processes and protocols that AI and AVT suppliers go through to develop their solutions, will help mitigate critical incidents.

In turn, healthcare organisations need to reinforce these messages to staff consistently – not just once a tool has been switched on, but as it becomes business as usual, and as upgrades are made.

A clear internal plan of action – with roles, responsibilities, methods to communicate, levels of escalation and protocols – will also give staff the right resources to know when and how to act.

As AI in healthcare becomes more visible to patients and the public, it’s important to have messages ready to explain both benefits and risks

Externally, transparency and communications matter just as much.

As the use of AI in healthcare becomes more visible to patients and the public, it’s important for organisations to have messages ready to explain both benefits and risks.

And as regulatory expectations advance, boards will be expected to demonstrate oversight. Knowing what tools are in use, how they are governed, and how incidents would be handled in crisis will no longer be a nice-to-have.

Communication sits at the centre of all of this – it links governance to behaviour, policy to practice, and incident response to trust.

Shared responsibility

AI tools such as AVT are clearly part of the solution to unsustainable workloads and growing demand. The problem is that readiness and adoption are varied across the NHS and, in some instances, have run ahead of shared understanding.

Trusts and suppliers that recognise this and work together to set clear expectations, rehearse incident scenarios and communicate openly, will be far better placed as adoption scales and scrutiny increases.

Now is the time to consider communication an integral part of clinical risk management rather than a nice-to-have tool for issues/crisis management.

Communication is one of the few tools we have available to us that can help mitigate the impact on patients as NHS staff embrace these vital new technologies.