How Machine Learning Apps May Benefit the Future of Healthcare
Technology has helped move forward business processes in unimaginable ways, and the latest wave of advancements in software and enterprise tools has many excited about the future of many sectors. Machine learning, also referred to as artificial intelligence, has moved swiftly to the top of the priority list for several technology firms. Through intelligent algorithms that change over time, machine learning engines can sift through millions of data points in just a few moments, helping improve the efficiency of business processes that would otherwise take hours or days to complete. The more data mined by artificial intelligence tools, the more intuitive the responses become, leaving less room for error in the future. While most equate machine learnings with business models and security processes, AI has found its way to healthcare in some ways as well.
The most notable shift toward algorithmic tools within the healthcare system is connected to a subsidiary of Google, DeepMind Health. The company’s mission is to improve the overall patient experience through the use of intuitive technology-based tools that use data and algorithms to notify clinicians, nurses, and doctors of potential medical issues, long before they become a problem for the patient. DeepMind Health has worked directly with the Royal Free Trust of London to design and implement the mobile application, Streams, to bring more advanced technology to the NHS, specifically in the case of acute kidney injury. While DeepMind Health and its Streams mobile app is not explicitly artificial intelligence at work, it shows the clear path down which healthcare as a whole is headed.
Why Focus on AKI?
The partnership between the Royal Free Trust and DeepMind Health was born out of an initiative started by the NHS to help reduce the all too common complications arising from acute kidney injury, or AKI. Within the UK alone, nearly 40,000 fatalities are linked to AKI annually because the medical condition causes severe damage to the body. The onset of AKI has no noticeable symptoms or warning signs, making it difficult for health professionals to identify and ultimately treat in a timely manner. These concerns present the perfect opportunity for a technology-based tool to lend a necessary hand in the detection, prevention, and treatment of AKI within the NHS, reducing mortality rates and improving the overall experience of each patient.
DeepMind Health joined together with the Royal Free Trust in 2016 to create the Streams mobile app, based on patient data shared between the two organisations. The initial focus was placed on AKI given the need for more timely diagnosis of the condition to prevent irreversible damage and the need for ongoing dialysis among patients. The mobile app was intended to provide alerts, task management, clinical notes, and data viewing through handheld devices, potentially decreasing the complications of AKI for countless patients. In similar developments, the use of algorithmic technology tools has shown promise in the early detection of heart conditions, strokes, and other debilitating conditions like diabetes and schizophrenia.
While the Streams project may have had positive implications and real-world use throughout the Royal Free Trust, development of the app ceased abruptly after it was found that more than 1.6 million patient records were delivered to the company, without patient consent. In an ICO investigation that took place shortly after the discovery, the Royal Free Trust was found to be in breach of patient confidentiality and data privacy rules, bringing to light several concerns about the future of advanced healthcare technologies.
Uses of AI in Medicine and Related Concerns
Even though the Streams application does not currently involve artificial intelligence in its delivery of alerts to medical staff, there is a notable link between its function and machine learning in healthcare looking ahead. The problem, as explained by a representative from a medical negligence firm in the UK, is that while these tools are beneficial in reduced the occurrence of delayed diagnosis in cases of AKI and other serious medical conditions, they cannot be sustained by patient privacy and confidentiality are ignored. Individuals have a fundamental right to keep their medical information private, and without a viable method to maintain this right, machine learning and other technology tools in healthcare cannot be successful.
In evaluating the future of healthcare as it relates to technology, many suggest that patients, doctors, and policymakers consider several aspects before embracing machine learning moving forward. The players involved must have an understanding of who controls the patient data, and ultimately, who or what will have access to the data. The only way privacy can be maintained is if these parameters are clearly defined and widely understood by the patient population and the professionals who provide care to them. Additionally, it is necessary to recognise how AI tools are coming to specific conclusions, such as a diagnosis or a certain course of treatment. Without these details, neither patients nor providers are likely to trust the technology or use it with any sort of consistency. Finally, the intended outcomes of using machine learning in healthcare must be known, whether that is to improve the patient experience or to reach organisation-wide performance metrics.
As the world takes broad steps toward a healthcare system focused on personalised medicine, the use of artificial intelligence is highly promising. However, only when all parties involved in the process are on the same page as far as access to information, use of data, and the intended results can machine learning in healthcare truly impact change for the better.