Artificial Intelligence, Real-World Automation and the Safety of Medicines
- PDF / 708,913 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 60 Downloads / 238 Views
CURRENT OPINION
Artificial Intelligence, Real‑World Automation and the Safety of Medicines Andrew Bate1,2 · Steve F. Hobbiger1 Accepted: 8 September 2020 © Springer Nature Switzerland AG 2020
Abstract Despite huge technological advances in the capabilities to capture, store, link and analyse data electronically, there has been some but limited impact on routine pharmacovigilance. We discuss emerging research in the use of artificial intelligence, machine learning and automation across the pharmacovigilance lifecycle including pre-licensure. Reasons are provided on why adoption is challenging and we also provide a perspective on changes needed to accelerate adoption, and thereby improve patient safety. Last, we make clear that while technologies could be superimposed on existing pharmacovigilance processes for incremental improvements, these great societal advances in data and technology also provide us with a timely opportunity to reconsider everything we do in pharmacovigilance operations to maximise the benefit of these advances.
Key Points
1 Introduction
Research on the application of artificial intelligence and machine learning in pharmacovigilance is not new but there has been renewed interest and focus in recent years.
The desire for continual progress in our ability to rapidly and successfully complete work tasks with less and less redundant effort and to maximise appropriate impact transcends generations and scientific fields. Pharmacovigilance (PV) is no different. We have more data than ever and there is enormous excitement in the potential to further scientific knowledge. This coupled with great advances in computational power for collecting, storing and linking data means there is a huge increase in the capacity to innovate in data science. With this focus on data, as with other fields, we see intense discussion of ‘artificial intelligence’ (AI), ‘machine learning’ (ML) and ‘automation’ in PV. Crown suggests [1] there are three aspects of healthcare data that have driven this interest. First, the increase in the volume of data, second, that much of the data is unstructured and third the speed for data refreshes, which presents serious challenges to traditional statistical and epidemiological methods. The interests for PV are clear, particularly when one considers that much of routine PV has not changed greatly for many decades, see e.g. [2–5]. Even the most commonly used approaches for quantitative signal detection based on disproportionality have been around for many years [6, 7]. A challenge, as for many technologies with associated hyperbole [8], is that the terms are used very differently and often vaguely or inaccurately, sometimes seemingly deliberately inappropriately. This lack of harmonisation blurs a
Advances are necessary to ensure capacity to deal with ever-increasing volumes of data and to optimise the value of the data for individual patients. Despite enormous technological advances, pharmacovigilance has not changed enormously since the core principles were introduced in the 196
Data Loading...
