Anne Kinderlerer: Can digital solutions really fix outpatient care?

As part of our joint work with NHSE on the future of outpatient care we have been exploring how better digital clinical systems might support outpatient transformation, to better manage risk across patient pathways and direct resources to those in greatest need. We have asked ourselves what we mean by “outpatient care.” An answer... is to provide specialist input to a patient, to make a diagnosis, to advise or organise treatment at the point at which the patient needs it.

A recent report from the Health Foundation (1) shows that clinicians already think that the electronic patient record (EPR) saves them time but are frustrated by the fact that existing systems are clunky and difficult for clinicians and patients to navigate. Patients told us how the current system relies on them to identify a problem, recognise its significance, and then manage to access primary care. Primary care clinicians, often without access to diagnostics, need to decide on which specialty, priority and communicate all relevant information to allow triage. Our existing systems of triage, where secondary care clinicians try and predict who is most likely to benefit from specialist care and reject those who they deem will not, increases the difficulty of managing risk across the pathway and can make primary care clinicians feel undervalued. (2)

The performance of two week wait (2WW) cancer diagnostic pathways illustrates how this system fails to manage risk. Of the 2m+ patients referred a year by 2WW, only 7% are diagnosed with cancer and only 50% of the cancers diagnosed are diagnosed on 2WW pathway. (3) I wonder if the other 50% wait longer because the system is so poor at identifying patients with cancer.

Reimagining the patient pathway

The diagram below attempts to illustrate areas in which existing digital technologies could support better care for those patients requiring access to planned specialist clinical care – highlighted in yellow boxes. I have tried to visualise how digital might support a model of care which operates as a continuum where the patient is able to access the right support; signposting to self-management, generalist or specialist care at the time when they most need it.

When patients access the system, digital tools could extend and integrate the clinical decision support systems that GPs already use to prioritise, so they can better signpost to proven self-management, direct patients to community services or direct access diagnostics.

For some conditions, this could allow us to test whether digitally supported self-referral allows faster triage and review for those at greatest risk. Making it easy to share more data, such as images, longitudinal data and patient questionnaires will improve the accuracy of triage. Making the outcome of those triage conversations between clinicians visible to patients would allow advice and guidance to reach the patient directly and better support shared decision making. Booking systems should allow patients to choose when they are seen and increase their ability to choose whether time to review or distance from services is more important to them.

Patient initiated follow up (PIFU) that, in principle allows patients to request follow up when they need it, is in many areas still an aspiration. Patients must fight with complicated booking systems, and lack of capacity prevents them being seen when they need it.

We can imagine a system in which patients are able to use patient portals to report flares, or record patient reported outcome measures (PROMS), or ask questions of clinicians. In a world where secondary care clinicians have time to respond and to organise the most appropriate follow-up, there would be less need for routine appointments that exist only to prevent patients getting lost in the system. Data from the patient portal, and linked data across the system, can flag risks so that a patient’s follow up can be expedited, even if they are unable to use digital systems. Opening the possibility that digital transformation could directly address health inequality rather than exacerbating it.

As we improve the use of data from primary care records and its integration with secondary care datasets, we enhance our potential to target interventions to patients at high risk of complications from long term conditions. Rapid cycle randomised control trials will enable us to determine what approach and to which service makes the most difference to under-served populations. (4)

The diagram also illustrates where AI could support better care depicted by green boxes – AI systems that allow appropriate further blood tests and risk stratification into secondary care further improve outcomes. (5) Coding of reason for referral (and collection of data from patient answers to clinical decision support) and coding of diagnosis and outcomes in the clinic and afterwards will allow machine learning algorithms to design better pathways. Making this possible without burdening clinicians with further administration will require imaginative use of generative AI (such as large language models) to summarise consultations and results, and to pull structured data out of unstructured clinic letters.

Final thoughts

Transforming outpatients involves reimagining how we work across systems. Changing the way we work is only possible with digital clinical systems that are designed to make tasks easier to complete, to highlight risks to patients in a way that makes it easy for us to respond, and to make it possible to see a patient who needs review at the ‘click of a button’.

I share below some references that have helped to shape my thinking. As ever, I would welcome hearing from you if you were working on this area already and are interested in sharing your work with us at the RCP or discussing it with me further.

References

1. https://www.health.org.uk/publications/long-reads/which-technologies-offer-the-biggest-opportunities-to-save-time-in-the-nhs?

2. Is the GP’s professional opinion no longer valued? Rammya Mathew: BMJ 2023;383: p2559

3. Cancer detection via primary care urgent referral and association with practice characteristics: a retrospective cross-sectional study in England from 2009/2010 to 2018/2019 Thomas Round, et al. British Journal of General Practice 2021; 71 (712): e826-e835

4. Creating a Learning Health System through Rapid-Cycle, Randomized Testing Leora I. Horwitz, M.D., M.H.S., Masha Kuznetsova, M.P.H., and Simon A. Jones, Ph.D. N Engl J Med 2019; 381:1175-1179

5. Intelligent liver function testing (iLFT): A trial of automated diagnosis and staging of liver disease in primary care, John F. Dillon, et al Journal of Hepatology, 2019 71, 699-706, https://www.dundee.ac.uk/stories/intelligent-liver-function-testing-ilft