Learning from error: rethinking critical incidents to make paediatric prescribing safer

Abstract

Prescribing errors in children are a major patient safety threat, which existing improvement efforts are struggling to tackle. This thesis contends that medical errors are a consequence of the complexity of modern healthcare and that to address them, much more needs to be known about their underlying causes. Using a large dataset of reported errors in secondary care, this thesis presents a programme of research that investigated how and why prescribing errors in children happen. Based on critical realist epistemology and a theoretical orientation to complexity science, the research comprised five studies. Study I reviewed existing literature to identify paediatric-specific causes of errors such as individualised dosing and different medication formulations, but also found a lack of primary explanatory evidence. Study 2 synthesised evidence in children and adults to develop a detailed framework of errors’ wide-ranging, interrelated contributing factors. Study 3 presented a quantitative analysis of reported errors, identifying areas of high-risk practice and offering a pragmatic method by which incident data can guide quality improvement. Study 4 involved an in-depth analysis of reported intravenous fluid prescribing errors, describing their types and contributing factors, and demonstrating the potential of incident data to support causal analysis. Study 5 brought together the findings of previous studies to form a preliminary theory of prescribing error in children. Using a behavioural approach supported by the COM-B framework, it then analysed reported error data to develop a behavioural model of prescribing error in children and offer rich descriptions of how errors happen. This research contributes to knowledge by showing how a wide range of highly- interrelated individual, social and contextual factors lead to prescribing errors in children. Beyond these, it also found that a host of child-specific causes further complicate paediatric prescribing and undermine its safety. Causes of errors were complex and interdependent, meaning that multi-faceted solutions that address a range of underlying causes in tandem are required. This thesis carries this knowledge forward into direct implications for improvements in prescribing safety. It also proposes ways that critical incidents could be used more effectively to support learning from error in healthcare. Finally, in light of its approach and Findings, it suggests how research, quality improvement, and education could better address the complex, seemingly-intractable problem of medical error.