The battle against AKI and the role of technologyFeature of the week
As one of the first health organisations to address the growing issue of acute kidney injury, or AKI, the NHS launched the Think Kidneys Programme to help improve patient outcomes while also reducing the rising costs associated with the complications of the medical condition. Acute kidney injury has been a high priority focus of the NHS for the last four years, and the development and implementation of the programme is meant to not only raise awareness among patients who are susceptible to AKI but also provide clinicians education to direct personalised healthcare. Nearly 40,000 incidents of AKI take place in the UK alone each year, making initiatives like Think Kidneys imperative to improving patient experiences and outcomes on a broad scale.
In 2015, the Royal Free Trust of London approached Google-owned DeepMind Health in an effort to push the Think Kidneys programme forward another step. The technology firm is a leader in artificial intelligence and machine learning, but the partnership with the NHS did not initially tap either of those strengths. Instead, DeepMind Health was tasked with creating a mobile application, named Streams, to help in the early detection, prevention, and timely treatment of AKI in Royal Free Trust patients. Because AKI has a plethora of underlying causes, but little to no warning signs of a problem beneath the surface, patients are at high risk of enduring life-altering medical conditions when AKI is left untreated. Through real-time mobile alerts, data viewing and task management, the Streams mobile application in the hands of healthcare providers had the potential to affect change among the patient population in a drastic, positive way.
The problem with streams
Most individuals in the UK were unaware of the partnership between DeepMind Health and the Royal Free Trust until it was announced that a breach of confidentiality took place early in the development of the Streams app. The trust handed over 1.6 million patient records to DeepMind for the purpose of improving the functionality of the application, but it failed to get consent from the patients whose data was shared across organisational lines. Because of DeepMind’s association with Google, the world’s largest advertiser, many were concerned that the identifiable data shared with the technology firm could ultimately be used for purposes other than the development and implementation of Streams. Without holding a public comment session or receiving explicit consent from patients involved in the data transfer, the Information Commissioner Office found that the Royal Free Trust had broken the law as it relates to the privacy of sensitive information.
One of the interesting aspects of the Streams debacle is that most patients do not have a problem sharing their medical history to benefit research efforts with various companies. Most feel strongly that if their specific experience with a condition like AKI or other health conditions can lead to the improvement of health outcomes for others, there is no harm in sharing that information. However, in the case of the Royal Free Trust and DeepMind Health, the transmittal of data included patient-specific information, including identifying information that was ultimately found to have no relevance to the development of the Streams application.
The future of medicine and AI
Both the Royal Free Trust and DeepMind were apologetic about the transfer of data that sparked such controversy in the healthcare realm, but some proponents of patient rights are unsatisfied with the response. A representative from a medical negligence team of solicitors in the UK shares that while mobile applications like Streams stand to benefit tens of thousands of individuals relying on the NHS for quality healthcare, these technology-based tools cannot be developed or implemented in a successful way when patients’ privacy rights are violated. Instead of working to include the patient population in the discussion of advancing technology in the diagnosis and treatment of AKI, the Royal Free Trust and DeepMind Health may have set back progress toward a viable solution in reducing patient harm due to delayed diagnosis.
The Streams project did not involve the use of artificial intelligence or machine learning in its applications within the Royal Free Trust, but questions abound as to how the sensitive patient data, now sitting idle on DeepMind’s servers, can or will be used in the future. Information regarding the success or failure of Streams in the Royal Free Trust won’t be unveiled until the end of 2017, meaning the glaring breach of confidentiality is not now and may not be in the future, justified. Even if there are steps toward progress with faster detection of AKI in the Royal Free Trust thanks to Streams, the patient population cannot be expected to be comfortable with their medical information being shared between public and private entities. Looking ahead, it will be interesting to see how applications like Streams and companies like DeepMind Health are able to use the data they lawfully collect for the benefit of patients and providers, without sidestepping privacy issues on a grand scale.
The editorial unit