The Royal Free London NHS Trust’s decision to gift Google’s DeepMind 1.6 million confidential patient records understandably caused a huge furor when the story broke in the newspapers.

DeepMind is Google’s artificial intelligence company and had been developing Streams, an app aimed at preventing Acute Kidney Injury (AKI) as well as revolutionizing treatment of the condition. The Royal Free NHS Trust provided DeepMind with patient information in order to test the app with the aim that this would help treat a condition that prevents the kidneys from working effectively.

The Royal Free’s desperation to effectively combat AKI has been driven by the sheer amount of patients in the NHS who are suffering from the condition. It is estimated that 30% of all patients in critical care show signs of AKI, and it is the cause of around 40,000 deaths in the UK each year.

It is clear that AKI is a highly serious issue that urgently needs addressing but some patient and members of the public may be dubious as to how an app could potentially save lives. Streams is a means of viewing of all kinds of data, which may include intelligent patient alerts, clinical noting and task management. Despite the media perception, Streams does not currently involve artificial intelligence although DeepMind have made it clear that the company has plans to implement this in the future.

The app will revolutionise the health delivery process. The information gathered by Streams has a wide range of potential uses, from improving the efficiency of the management of many health conditions, including AKI and many others in the future, to improving the effectiveness of frontline healthcare delivery. The data collated can lead to an overhaul of working practices which could boost response times, lead to more effective diagnosis and, most importantly, save lives.

A solicitor from the medical negligence firm Patient Claim Line explains that the wealth of data collated by Streams could be used to monitor the performance of individual clinicians as well as an entire organization or department. This could be vital in the event of an incident occurring, such as a delayed response harming a patient. In this situation, Streams could be used to investigate what happened and to ensure that the management processes that failed are overhauled.

Concerns over the Royal Free handing over confidential records to DeepMind have been widely reported in the media and by members of the public and patients whose data has potentially been shared without their consent. There are numerous data protection issues that have arisen from this data sharing, with many concerned over the potential risk of private companies having access to confidential data and fears that the long-term safety of the individuals affected may be compromised by the fact that this information is sitting on Google servers.

Another issue is the widespread reporting of data breaches and cyber-attacks in the media, with both Google and the NHS targeted in the past. This fact does not inspire confidence that this personal data will remain secure.

Considerations for the future use of artificial intelligence in healthcare

With an ever increasing reliance on automated systems in healthcare, whether artificial intelligence is used to deliver treatment or improve the diagnosis process, AI in the medical field is here to stay. But there are some questions that must be answered as the era of AI-guided healthcare approaches.

One of the major considerations with the use of artificial intelligence must be the safeguarding of patient data. Who will have access to or control over this information? The use of AI would hopefully improve diagnosis of certain conditions, such as identifying groups which may be at greater risk of heart complications. Artificial intelligence would allow more comprehensive analysis of patient data so it will be vital to establish who retains control of this data to avoid it automatically being included in any research.

Will patients be made aware of how a diagnosis has been completed or whether the use of artificial intelligence has led to a recommendation of treatment? And to what extent is it an issue? Well, if AI is unable to explain a decision then both patients and medical practitioners may be reluctant to follow these instructions. Although a physician may have a “gut feeling” which leads them towards a diagnosis, this is based on extensive experience and is backed up by tests on this possible condition. It is highly likely that patients would expect the same process if artificial intelligence is used at any stage.

Artificial intelligence could prove most effective when applied with a specific goal in mind, whether this is improving a nation’s average life expectancy or targeting specific diseases above others. The introduction of AI in healthcare is a unique opportunity which could aid government decision makers in providing more generalized healthcare plans or empowering individual patients with a more bespoke approach.

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 830 other subscribers.

Follow us on Twitter