Introduction

Innovations in digital health, such as the introduction of smart infusion pumps, have the potential to improve patient safety, even though the evidence base remains weak.1 Equally, however, new risks can be introduced, which might contribute to adverse events and patient harm.2 3 The Healthcare Safety Investigation Branch (HSIB), which carries out independent investigations into patient safety concerns in the National Health Service (NHS) in England, published in December 2020 its findings from a national investigation into the procurement, usability and adoption of smart infusion pumps.4 The report includes safety observations (suggested actions for wider learning and improvement) that suggest the use of (clinical) safety cases in order to demonstrate that patient safety risks have been addressed rigorously and proactively.

Safety cases are a common regulatory instrument used in UK safety-critical industries, as well as other countries such as Norway, Australia and New Zealand.5 Previous attempts at importing safety management practices from other industries to healthcare have not always delivered the anticipated benefits, for example, the application of failure mode and effects analysis6 or the adoption of incident reporting systems.7 It is important to understand both the principles underlying such approaches and the context within which they were developed, as well as the unique cultural and institutional context that is specific to healthcare.8 9

In this paper, we review the thinking that underpins the use of safety cases across different safety-critical industries, and then reflect on their potential use for assuring the safety of digital health innovations. This builds in part on a previous review,10 but also considers the recent debate about the need for an evidence base for safety case adoption. We focus on digital health innovations, because safety cases are likely to be particularly relevant for software-based systems including, more …