Next steps

It makes sense to start by piloting creation of computable knowledge and decision support in areas where progress will reduce clinical burden and enhance safety.

The development of computer-driven decision support will place clinicians in the driving seat of ever-improving, high quality learning healthcare systems. Its wider availability will underpin patient empowerment and self-care. Implementation of computable knowledge and decision support will also be vital if policy makers are to succeed in devolving as much healthcare and resources as possible from secondary to primary care.

The infrastructure required for developing such decision-support – translation of knowledge into computable formats – also opens the door to a further major development. Once computers can process healthcare knowledge and connect it to patient data, a truly learning healthcare system becomes much easier. Machine learning - searching for patterns in system-wide records of practice and patient experience – becomes possible. It will provide a new source of knowledge and insights that will sit alongside traditional scientific learning. It will further improve the excellence of healthcare delivery and enable more personalised patient care, drawn from analysis of so many different patient experiences. It will also help to improve the quality of decision support because it will be possible to examine which types of support are associated with the best outcomes.

We encourage healthcare policy makers, managers and practitioners to continue exploring and discussing how they can make the most of the opportunities outlined in this White Paper while also overcoming the challenges it describes. There are already promising signs that these challenges are increasingly recognised, for example NHS England has identified decision support as a priority for the new Transformation Directorate. NHS Scotland already has a national decision support strategy and service. NICE is actively exploring ways to make its guideline content computable. Health Education England provides guideline-based decision-support to the frontline, both as a point of care tool and a learning resource, and is investigating cultural, behavioural and motivational factors influencing uptake. For the good of the whole UK, indeed the world, these efforts must be coordinated and open to sharing good practice and lessons learned, both within Britain and beyond.

We therefore recommend:

1. Collaboration: the creation of an active pan-UK cross-sector stakeholder group with support at central and devolved government level to develop and promote a shared agenda for healthcare decision support across the four nations. This should include all the relevant clinical, research and informatics organisations, from the public and private sectors.
2. Piloting of next-generation decision support to tackle particular healthcare issues. This should focus on a selection of urgent, specific areas where clinicians can most benefit. This could be because of the volume of knowledge that is frequently updated, where the relevant content from practice guidelines lends itself to becoming computable, where decision support would reduce errors, to enhance safety and quality or where existing decision support is not good enough. Focus in these areas should avoid more controversial areas for decision support where there are doubts about its safety and assurance.
   
   Pilot projects might, for example, include new or improved decision support for:
   
   
   • Ordering laboratory tests in response to patients’ symptoms;
   • Optimising the management of long-term conditions;
   • Antibiotic prescribing;
   • Urgent referral guidelines in primary care

Making a shift to computable knowledge and decision support offers great hope (see Appendix for a table of benefits for each of the main stakeholders in health and care). As one NHS clinician explained: “We have an excellent NHS, but it relies on, as always, people within that system caring a lot and then going above and beyond, and doing a lot of manual work, so they become very stressed themselves trying to keep that system going.”

A “clinical satnav” – a truly digitally enabled system with a detailed library of medical knowledge properly flowing through it like our car's map - would make practitioners’ lives easier. It would make the system itself healthier and more sustainable and would ultimately result in a better standard of care for us all.

Appendix