Foreword: the growing need for advanced digital healthcare
As healthcare faces mounting pressures, there is a growing need for efficient and cost-effective digital services, including digitally enhanced self-care. Integration of Machine Learning (ML) techniques has an important role to play, bringing opportunities to derive new insights using the vast quantities of medical data generated every day.
Recent accelerated developments in the ML domain provide additional opportunities to build automated science-driven predictive solution systems for high-level personalisation and preventive healthcare. However, potential clinical and commercial benefits must be evaluated objectively alongside implementation challenges. ML-enabled health solutions present unique issues due to their complexity, the need for explainability, and the iterative data-driven nature of their development.
Here, we outline a three-step process for developing an ML-enabled digital health solution to support patient/consumer engagement and maximise health outcomes. Such evidence-based solutions constitute ‘digital medicine’ which measures and intervenes to improve human health.
The process is based on combining multidisciplinary principles and expertise. This enables the build of a multi-channel data foundation and ensures a deep understanding of intended use, including desired clinical benefits and associated risks. We discuss the need to demonstrate the effectiveness, safety, and security of a system to maximise its clinical and commercial potential, as well as the verification and validation needed to obtain FDA clearance or approval if it is a medical device that requires premarket authorisation and CE marking. Finally, we explore commercial matters related to digital healthcare solutions, which is one of the most critical concerns of innovators at present. We consider how commercial uncertainty can be tempered with a good understanding of existing payment mechanisms, and alternative options.
