28 November 2016
An exciting new era in patient care powered by electronic health records and artificial intelligence
Electronic health records and artificial intelligence herald new era of patient care
News about Google Deep Mind and IBM Watson applying artificial intelligence (AI) to healthcare is exciting for two reasons.
First the practical application of AI to improve patient care by monitoring vital signs in real time holds huge potential for saving lives. Second, the power of AI to advance prevention of illness in individual patients, combine information in new and innovative ways for research to prevent and treat illness in populations, as well as plan services and improve patients' experience of care is staggering. Concerns about data protection and consent are paramount and are being addressed with confidence.
Watson is being introduced in Alder Hey Hospital to improve patient experience and Deep Mind will use patient care data in the Royal Free Hospital to support
doctors and nurses to provide more timely patient care in real time. This project will provide alerts to doctors and nurses in real time using clinical data from pathology results and vital signs (pulse, temp, respiration, blood pressure, fluid balance). It is the start of decision support in health care happening in real time as a patient's condition evolves.
The Royal Free has decided to focus on acute kidney injury because it is common, has high mortality and a very high economic cost – more than the cost of breast cancer or lung and skin cancer combined. It is frequently missed and early identification is the key to prevention and effective treatment. The next test is likely to be sepsis, another condition with a similar pattern of high mortality, easily missed in the early stages, and very treatable if recognised early. As with all artificial intelligence applications, as the available data increases over time, the precision of condition recognition increases, far beyond what is likely without the assistance.
The Alder Hey project with Watson is focussing on the patient and patient experience. The expectation is that they will be able to access this pre-admission to hospital through a digital application on a tablet or smartphone, such as a mobile app. It could be used to drive vital research projects by proactively matching suitable patients to clinical studies, monitoring admission patterns to help with bed planning or to help management of chronic illnesses through educational applications which could alert patients and their doctors when their symptoms reach the point at which they should seek medical help, or even automatically make an appointment for them.
From the PRSB's standpoint, what is interesting about the Deep Mind project, is that it will almost certainly be associated with the development of open standards so that the clinical data can be very precisely and accurately shared interoperably – transferred from one computer (the EHR system) to another (eg smart phone, laptop, tablet). INTEROPen is an open collaboration of individuals, industry, standards organisations and health and care providers, who have agreed to work together to accelerate the development of open standards for interoperability in health and social care.
The emerging standard specifically developed to support interoperability, and a big focus of INTEROPen is FHIR (Fast Healthcare Interoperability Resources). PRSB is a significant player in this development, ensuring that the very technical standards are unambiguously consistent with good clinical practice, meet the needs of patients and care professionals and are consistent with the already published PRSB standards. What this will do is allow access to any health care data in electronic records made available in a web format. "Michael's story" illustrates what this could mean for a patient with complex health care needs.
Why get excited about all of this one may ask. The answer is that it is the beginning of a massive step forward for health care. Once EHR's are built on open interoperable standards, the data can be accessible anywhere, at any time, to any authorised person – on any computing platform. The data can be aggregated from EHRs (in England probably via NHS Digital) and will be directly available for service management, policy making and research. Clinical research published using data based on the standards will be able to include the actual data, so that the decision support will not only be absolutely up to date, but also immediately available,
and greatly augmented by the artificial intelligence algorithms generating new knowledge and new solutions. Clinical guidelines such as those developed and published by NICE will be embedded in EHRs and could be instantly updated in this new world. The 100,000 Genomes Project will have access to the huge amount of clinical data to maximise the benefit of the genome sequencing, biobank will have less of a struggle gathering medical data to accompany tissue samples for research. Medication management for people taking multiple medications will benefit from the analysis of vast quantities of real-time information about the effects of medicines in real life – something that cannot be achieved with conventional randomised controlled trials.
Acutely ill patients and people like Michael with complex care needs will not only benefit from better care coordination and delivery, but also care provided with up to the minute evidence-based decision support; and research published with the available interoperable data accessible will form the basis of up-to-the-minute, evidence-based care guidelines.