The advanced analytical capabilities of machine learning and cognitive computing are pushing healthcare to a whole new level, as it is set to disrupt the industry. Google, IBM and Microsoft are leading the way when it comes to advanced analytics and data-based healthcare – and things are moving fast.
You may be familiar with the ‘quantified self’ movement (metrics on your everyday activities related to your health), electronic health records, and the benefits of using real-time data analytics to improve patient care. The next step in this is to assist doctors with complex diagnoses and early detection of health problems. Cognitive computing can unlock the full power of all the medical data that healthcare organisations have been collecting, as it reveals underlying and undiscovered patterns and connections in the data that can be used for diagnoses, care pathways and operational decisions.
What is cognitive computing? While there’s no single, universally accepted definition, the Cognitive Computing Consortium suggests some defining characteristics. These include computers that are able to understand complex situations, and are capable of dealing with ambiguity and uncertainty in the data. They also include computers that can deal with conflicting data, as it changes frequently due to changing situations. The computer is able to factor in all these situations, along with their contexts and influences, and offer a best outcome or best answer to the problem.
These are exactly the types of problems faced by diagnosticians, and cognitive computing can help. IBM with its Watson cognitive computer, for example, is working in partnership with hospitals including Columbia University Medical Center, Memorial Sloan Kettering Cancer Center, and The University of Texas MD Anderson Cancer Center.
CEO of IBM Ginni Rometty elaborated on this at the recent World Health Care Congress in Washington, DC: “We’ve been teaching Watson to see things like X-rays and images. It’s great at forming hidden connections. If you could test your hypotheses faster, it could help us make progress in a range of areas – genomic medicine, that’s an obvious area. And it could be used for drug discovery, or for the discovery of alternative uses for existing drugs.”
It’s all about the data
But all this computing power and machine learning is utterly dependent on one thing – data. Without massive datasets drawn from electronic medical records, systems like Watson won’t have enough information to operate effectively. Without accurate and timely data from individual patients, systems won’t be able to properly ‘understand’ the patient’s condition and make useful diagnoses or suggestions for preventive measures, like diet, exercise and medication.
Real-time data coming from monitoring machines and health sensors, along with historical data from electronic health records, could be used to create machine-learning models that predict and prevent health problems. Machine learning (methods that enable computers to learn without being pre-programmed) could also be used to determine how patients might react to certain treatments or drugs, leading to more personalised and hence more effective treatment regimens.
Security and trust
This brings us to perhaps the greatest hurdle to realising the full potential of cognitive computing for healthcare – trust. While many individuals may choose to wear fitness trackers and use apps to monitor their sleep, medication and eating habits, their adoption is far from universal and public distrust of mass health data collection is high.
This perception could be overcome by technology and healthcare companies committing themselves to the highest standards of probity and data security. Governments also need to pass appropriate legislation and regulations, with strong sanctions for non-complying organisations.
Creating an environment where patients and citizens trust that their data is being gathered ethically and stored securely will help the process, paving the way for greater efficiency – and more importantly – better patient outcomes.