Which technologies offer the greatest opportunities to save time in the NHS?

AI will also play a significant role in the other two technologies ranked highest by respondents as most likely to save them time within the next 5 years: clinical documentation tools and software for analyzing images and test results. One example of the former is AI-driven ambient voice technology (AVT), which uses voice-to-text software to automatically transcribe patient consultations and then natural language processing to turn these transcriptions into summary notes and letters. Asif Bachlani (Consultant Psychiatrist, Priory Group and Associate Non-Executive Director, Kent and Medway NHS and Social Care Partnership Trust) indicated that voice recognition technology could offer ‘a massive reduction in the administration burden for frontline clinical staff’ . Ben Jeeves (Associate Chief Clinical Information Officer, Clinical Safety Officer and Advanced Practice Physiotherapist, Northern Integrated Musculoskeletal Service, Midlands Partnership University NHS Foundation Trust) similarly highlighted the ‘heavy burden’ of admin and envisioned a future where AVT could lead to ‘the full removing the administration processes associated with all the clinical interactions that we have now to give us much better quality in terms of the consultations that we have, how we interact with our patients.

When it comes to software for analyzing images and test results, some of the more ambitious aspirations about their time-saving potential may take a while to materialize. Stephen Harden (consultant radiologist, University Hospital Southampton and vice-president for clinical radiology, Royal College of Radiologists) told us about his experiences with AI-enabled lung cancer screening, where the current technology effectively identifies and measures possible lung nodules. ‘It doesn’t speed me up a lot, but it does to some extent’, he noted, but also stressed that the technology offers ‘real benefit’ in helping to maintain the quality of image interpretation. However, in the longer term, there is considerable interest in the potential for AI to act as a ‘reader’ for images, which could provide a more transformative way to save time and support workforce capacity. As Harden explained, ‘the hope is that when reliable and accurate AI is fully developed and implemented, it will be of truly useful assistance to radiologists and imaging departments’.

In the rapid evidence review, the sample of 500 studies found a potentially promising evidence base for the time-saving possibilities of automated screening and interpretation of test results, X-rays, patient records, voice messages, and so on. Of the 34 studies in this space included in the sample, 85% reported a positive impact on staff time. However, as the review noted, newer technologies tend to have less evidence available about their impact, and studies with positive findings may be more likely to be published than those where new technologies did not work well. In contrast, more established technologies such as EHRs have generated a broader evidence base, which is more likely to contain studies with mixed or negative findings.

Imagining what technologies might free time beyond the immediate future requires a considerable amount of speculation. The exploratory nature of the interview format helped us grapple with the nuances of this question. We asked experts to comment on potential opportunities over the next 10 years and then 20 years and beyond – although many interviewees commented on the inherent difficulty of envisioning these longer-term time horizons.

Several interviewees discussed the potential of AI and other innovations to improve data analytics to support patient care and broader population health. Experts have talked about a more ‘proactive’ model of care, whereby AI can generate intelligence about a patient’s health or identify individuals or groups at risk of developing certain conditions or in vulnerable circumstances with the aim of earlier interventions guide and support patient self-management. Interviewees described how this could not only improve patient health and tackle health inequalities, but also potentially reduce demand and create more capacity in the system:

We already have the data in front of us. This is how we use it. And I think if we start learning how to use that data that’s already available, then we can really make a difference and we can actually save a lot of time if we automate some of that information gathering and that analysis… you identify things early, whether it’s anomalies, whether it’s risks, whatever it is, if you identify it earlier, it will save you time in the long run. Faith Ndebele, Consultant Psychiatrist, Solent NHS Trust and Chair, Digital Psychiatry Special Interest Group, Royal College of Psychiatrists

It would be tremendously powerful, if I could use my voice assistant within an EHR to ask, ‘This patient’s blood pressure is a little low, in the last 24 hours, what could I have done differently to prevent this from happening?’ And then [for the technology] to go through all the blood results, medications, etc. and then come [up] with the relevant information. That would be the most useful thing I could think of. It would save so much time. Joseph Alderman, Registrar of Anesthesia and Intensive Care, University Hospitals Birmingham NHS Foundation Trust and Doctoral Researcher, University of Birmingham

However, interviewees also acknowledged the barriers to realizing these ambitions, including variable data quality and the need to develop greater data analytics capability among the NHS workforce.