Radical change in health and care

Given the pace of change in the technology sector, Dr Philip Scott discusses how technology is radically changing the way the NHS provides services

The NHS is a beloved institution and its staff always feature as the most trusted professions – so why does it need radical change? Since its inception in 1948, the NHS has faced ever-increasing demand. Innovations in healthcare mean that previously untreatable conditions can now be cured or prevented and now we are experiencing the ‘demographic time-bomb’ of an ageing and unfit population. The NHS Long Term Plan calls out urgent public health issues that add to the drain on resources: obesity, alcohol abuse, smoking, antimicrobial resistance and air pollution. On top of all this, we have changes in the workforce with declining numbers of experienced general practitioners, challenges in recruiting and retaining staff, increasing numbers of people with multiple long-term health conditions and no sustainable financial model for social care. And then we have a pandemic and the gigantic project to get routine services back to normal! How can digital technology support radical change for the better?

Supporting the triple aim
The high-level goal of quality improvement in health and care services is often characterised as the 'triple aim': (1) improving population health, (2) reducing per capita costs and (3) improving the patient experience of care. Digital technology has the potential to help in each of these.

Reducing costs is one of the goals of large-scale pathology and radiology networks – achieving economies of scale by centralising services and sharing diagnostic test workload. Likewise, the increased use of 111 (online or phone) to divert people away from unnecessary GP or emergency department visits should have financial benefit (though that is not yet proven). Electronic health records have been used by OpenSAFELY to achieve rapid low-cost clinical research in the pandemic. SMS text reminders are widely used to reduce healthcare appointments being wasted by non-attendance.

Improving population health is the aim of online guidance on nhs.uk and personal health apps for long-term conditions like diabetes or dementia. Other apps exist to help people with repeat medications keep to their prescriptions, and innovative clinical decision support helps doctors to reduce overuse of antibiotics.

Better patient experience is the purpose of the NHS App. It enables patients to access their primary care record, book appointments, request repeat prescriptions and give proof of their Covid vaccination status. An innovative technology that was very helpful in the pandemic was Oximetry@home, which enabled ‘virtual wards’ of people to be supported by hospital clinical teams while remaining in their own home.

Innovation with artificial intelligence
There are high expectations, not always realistic, of what artificial intelligence (AI) can do for healthcare. Numerous AI innovations have come to nothing, but in fairness that is inevitable for ground-breaking new technology. One aspect of the problem is that there is no single definition or taxonomy of 'artificial intelligence' that is universally accepted. Broadly speaking, AI includes rule-based systems, where human knowledge and expertise is hard-coded into software; natural language processing (NLP), where software extracts meaning from or generates text or speech; machine learning, where analytic techniques determine patterns of association in source data; and intelligent automation, where software or devices act autonomously to trigger or stop some action(s) based on some kind of monitoring or input data.

Examples of rule-based AI in healthcare are risk calculators, drug interaction alerts and symptom checkers in patient-facing apps. NLP is put into practice in applications such as voice recognition from digital dictation, chatbots and voice-activated searches like finding the patient’s latest diagnostic reports. Machine learning has been used to build predictive models, such as identifying deteriorating patients from vital signs observations. Another application is diagnostic image analysis, for instance aiming to provide early detection of cancer. In the future, machine learning may support adaptive clinical decision support, where standard treatment guidelines are tailored for the individual patient based on their specific history and medications. Intelligent automation is implemented in so-called ‘robotic process automation’ (RPA), which is often automated transcription from one software application to another to obviate human re-keying of data (when proper data interfacing is not an option). Other forms of automation are smart infusion pumps, predicting demand (for staff, supplies or beds) and waiting list prioritisation.

Professional leadership and citizen co-design
The Topol review addressed the overall workforce need for appropriate skills in the digital delivery of healthcare and the new NHSX data strategy highlights the need for greater capacity in analytics and data science. BCS is working with partner bodies in the Federation for Informatics Professionals to lead professionalization across the whole informatics workforce, and across government departments to support the emergent Digital, Data and Technology Profession (DDAT).

Patient and public involvement is crucial. Patients often still have poor experience of healthcare communication, but there are outstanding exemplars such as the Connected Health Cities programme that demonstrated what can be done in extensive patient and public involvement to build trust in data and technology.

Citizen co-design is vital for two intertwined reasons. Firstly, general health outcomes are largely determined by social factors. Secondly, there is a 'digital divide' in the population that is also related to literacy, housing insecurity, cultural factors, economic status generally and broadband cost specifically. A recent BBC report demonstrated the impact of this in the UK pandemic response. If social determinants are not sufficiently addressed, the digital divide will worsen and the population that most needs improved health will miss out on the benefits of modern technology.

Conclusions
Digital technology has already done much for healthcare and has the potential to do far more. However, we are right to be cautious. Technology is not magic. Startup costs are often high and the evidence of effectiveness is usually weak. By all means let us innovate, but we must also evaluate. It is often difficult to know whether a healthcare quality improvement initiative is working or not. Electronic health records are helping to make it practicable to conduct rapid A/B testing and stop ineffective programmes early, as demonstrated by New York University Langone Health.

The NHS and social care still have a long way to go in laying solid technology foundations like adequate infrastructure and universally adopted data standards. Health and care are the poor relations in comparison to the level of investment in IT infrastructure in other sectors. Until we get these basics right, we will struggle to realise the possibilities of more advanced capabilities and achieve the radical system change that we need.

Dr Philip Scott is chair of BCS Health & Care and Reader in Health Informatics at the University of Portsmouth’s School of Computing.