Big data technologies offer multiple opportunities to meet the so-called Quadruple Aims of health, namely: improved patient experience, better health outcomes, improved staff experience and lower care costs. When combined with artificial intelligence (AI), big data can support decision-making, improve operational efficiency and empower people to take better care of their own health, with personalised insights and tips. But for these solutions to be effective, they need to be integrated into the current workflows of healthcare professionals and the habits of patients. “These technologies will only make a difference at the point of care, if combined with a deep knowledge of clinical realities. That is why healthcare professionals and patients are at the heart of the BigMedilytics approach,” says project coordinator Supriyo Chatterjea, from Philips, the project host. The BigMedilytics Blueprint for big data technologies was informed by 12 in-hospital pilots, in collaboration with 11 European hospitals. The Blueprint is available as an interactive website which maps the technical and non-technical aspects of integration across various use cases. It has already been used by the Emergency Department at the University Hospital, Frankfurt, for a pilot to optimise patient processing. The team’s work has also led to the modification of Spanish healthcare data protection legislation. Additionally the project’s materials have been included in the training curriculum for future healthcare managers in the Netherlands.
Learning from hospital pilots
The 12 project pilots, designed to inform specific aspects of the BigMedilytics Blueprint, covered three medical fields: population health, oncology and the industrialisation of healthcare. One of the pilots focused on comorbidities, and demonstrated how data for 5 million patients monitored over a 5-year period could identify patients at risk of secondary care admission, information that is valuable to general practitioners when referring patients to hospitals for follow-up treatment. Similarly, a prostate cancer pilot demonstrated the importance of combining urology, radiology, pathology and even financial data sources. AI analysis can then be applied to make short- and long-term forecasts of treatment outcomes, weighing up likely benefits against possible side effects. Another pilot, using real-time locating systems (RTLSs) to find mobile hospital assets (such as infusion pumps), eliminated the 30-minute average search time spent by nurses per shift, looking for equipment. The same RTLS technology allowed the hospital, for the first time, to estimate use levels for mobile assets, valuable for future investment plans. “The COVID pandemic has provided the perfect business case for the BigMedilytics solutions. Telehealth enables patients to be monitored remotely when they can’t visit hospitals regularly, while AI and big data techniques allow clinicians to analyse data to inform treatment decisions,” adds Chatterjea.
The non-technical challenges
Technically, the IT landscapes of hospitals/health systems are fragmented, making it very difficult to scale big data across Europe. Other challenges include ensuring data quality, with input errors corrupting algorithms. “From a non-technical perspective, we need a code of conduct with updated and consistent guidelines. Currently, EU Member States do not hold a common position on the legal definition of personal data, limiting the re-use of healthcare data, also compounded by GDPR restrictions,” explains Chatterjea. Getting care providers to use big data also presents an obstacle, with several of the pilots demonstrating the importance of change management programmes to overcome barriers around lack of training or trust. Consortium members, and other healthcare providers, will continue to use the Blueprint, so the impact of BigMedilytics is expected to be felt long into the future.
BigMedilytics, artificial intelligence, big data, diagnostics, healthcare, prostate cancer, treatment, hospitals, COVID, patient, GDPR