Story Highlight
– Ambient voice technology (AVT) gains traction in healthcare.
– AVT reduces documentation time for clinicians consistently.
– Limited evidence on AVT’s impact on patient outcomes.
– Diverse AVT market complicates evaluation processes.
– Clear logic model needed for consistent metrics and benefits.
Full Story
Ambient voice technology (AVT), often referred to as “AI scribes,” is being rapidly integrated into healthcare environments across the UK, including various medical settings such as primary care, outpatient services, and emergency departments. This technology offers a compelling proposition: it aims to reduce the time healthcare professionals spend on documentation, allowing them more opportunities to engage directly with patients and alleviating some of the administrative burdens that often contribute to clinician fatigue. However, despite the swift adoption of AVT, substantial evidence regarding its effectiveness remains scarce.
The first phase of a national evaluation, funded by the National Institute for Health Research (NIHR) and led by the NIHR Rapid Service Evaluation Team, has begun to shed light on this topic. This phase primarily involved a thorough review of existing evidence, an analysis of the AVT market landscape, and the creation of a logic model aimed at clarifying how benefits from AVT are anticipated to materialise.
A notable finding from the research thus far highlights AVT’s potential to save considerable time on documentation tasks. Various studies indicate that healthcare professionals can spend less time drafting notes, and in many cases, there appears to be a reduction in overtime spent on administrative responsibilities. General practitioners have reported that AI tools assist in minimising overtime and relieving their administrative workloads.
However, an important question arises from this time saved: does it lead to an increase in patient-facing care, improve continuity of care, enhance the wellbeing of healthcare workers, or increase capacity within the healthcare system? Alternatively, might this time simply be absorbed by an already stretched system, providing short-term relief without fundamentally transforming daily operations?
Currently, most evaluations focus solely on the time saved in documentation without delving into whether those time savings translate into meaningful improvements in outcomes for patients or the healthcare system. Although there is an understanding of the immediate output—minutes or hours saved on documentation—the broader implications, such as the effect on patient experience, safety, workforce retention, and overall system capacity, remain largely unexplored.
Understanding this distinction is critical. In the context of healthcare, time cannot be straightforwardly converted into enhanced productivity, increased financial savings, or improved care quality. It is possible for a clinician who saves time during a consultation to feel relieved, but that does not inherently create additional appointment availability or generate cost savings. Without intentional strategies, the time saved may merely provide a breather rather than increase actual throughput, which, while having its own value, does not necessarily lead to tangible improvements within the healthcare framework.
The AVT market itself is experiencing considerable evolution. Various products are being tailored, configured, and launched in diverse clinical contexts. Some solutions are specifically designed to generate medical notes, while others cater to creating letters, referrals, summaries, or patient-facing communications. These tools may either be tightly integrated with existing electronic health records or function independently. The market sees continuous entry from new suppliers, while established products adapt to cover a broader array of requirements.
This broad spectrum of AVT applications complicates the evaluation process, making it even more critical. Since the tools and settings vary significantly, insights gleaned from one application cannot be directly extrapolated to another.
Moreover, the evaluation criteria for AVT are inconsistent across studies. Many analyses rely on similar high-level metrics, such as “time spent on notes” and “total time in the electronic health record,” but definitions and measurement approaches differ. Importantly, metrics related to clinician wellbeing, job satisfaction, patient experiences, or safety assessments are less frequently included and are often bespoke rather than standardised across studies. Only a limited number of evaluations assess system-wide impacts or associated costs.
This variability creates challenges when attempting to compare results from different environments, tools, or implementations. Even in studies reporting beneficial outcomes, it remains ambiguous whether these results stem from the technology itself, the clinical environment, the implementation strategies employed, or the specific measurement methodologies.
To tackle these challenges, a coherent logic model can provide clarity and a structured approach. Such a model delineates how AVT is expected to function: detailing what inputs are required (technology, training, changes in workflows), describing the processes involved (actual usage by staff), outlining the immediate outputs (e.g., reduced documentation time), and crucially identifying the targeted outcomes those outputs aim to facilitate.
For AVT, understanding this structured approach is vital. Time saved on documentation serves as an intermediate benefit, not an end goal. True benefits—such as enhancing patient experiences, decreasing burnout among healthcare professionals, improving the quality of medical records, enhancing safety, boosting capacity, or facilitating staff retention—are not guaranteed outcomes. Rather, they are contingent upon how the repurposed time is utilised, how workflows evolve, and how the entire system responds to these changes.
Utilising a logic model fosters clearer articulation of AVT’s potential advantages, ensures that diverse forms of value (including staff wellbeing) are acknowledged, and helps manage shared expectations among clinicians, administrators, suppliers, and policymakers. It also allows for the consideration of any unintended consequences that may arise, such as an initial increase in documentation efforts as clinicians verify transcribed notes and actions for accuracy.
Looking ahead, the next phase of this evaluation process will involve a comprehensive mixed-methods approach conducted across multiple sites within the NHS. This research will not only assess whether AVT effectively saves time but will also delve into the implementation processes, how it alters workflows, the experiences of both staff and patients, and whether the anticipated benefits are indeed realised. Furthermore, it will investigate the economic implications and examine the conditions under which AVT could deliver sustainable advantages for the healthcare system.
As AVT continues to gain traction, the objective should not merely be to accelerate its deployment but also to enhance the learning and insights derived from its implementation. A focus on structured logic, uniform evaluation metrics, and an emphasis on meaningful outcomes—beyond mere outputs—will be essential for AVT to fulfil its promise for healthcare professionals, patients, and the healthcare system as a whole.
Our Thoughts
The adoption of Ambient Voice Technology (AVT) in healthcare highlights several areas where UK health and safety legislation could imply better practices to prevent inefficiencies and ensure patient safety.
Firstly, clear risk assessments and evaluations should be conducted before rolling out AVT tools under the Management of Health and Safety at Work Regulations 1999. Each tool’s integration into clinical practice must be evaluated for both potential benefits and unintended consequences, which could impact patient safety and clinician workload.
Secondly, under the Health and Safety at Work Act 1974, there must be ongoing monitoring of the impact of AVT on clinical outcomes, clinician wellbeing, and patient safety. A standardized approach to measuring the efficiency and effectiveness of AVT tools would facilitate better comparisons and evaluations, reducing the risk of relying on inconsistent data.
Lastly, thorough training for clinicians on using AVT effectively, ensuring that it enhances their workflow rather than adding complexity, is essential. By emphasizing a logic model approach, the planning and implementation of AVT could ensure that benefits are accurately measured and understood, reducing the possibility of overlooking critical safety outcomes amidst technological advancements.
