The economics of patient safety and incident reporting systems: translational evidence

Abstract

The aim of this thesis is to expand current knowledge of the economics of patient safety, with a focus on the costs and consequences of incident reporting systems. Intellectual enquiry into these closely aligned topics is demonstrably inadequate for policy development, which was the core motivation for conducting the research. To address these inadequacies, a systematic review of papers focusing on the economics of patient safety is presented, comprising 17 high-quality economic evaluations of patient safety that were published between 2007 and 2018. The findings indicate that few digital interventions for patient safety have been evaluated using methods from health economics, which explains why a rigorous conceptualisation of incident reporting systems and the economic issues surrounding them is required. A process map of the national incident reporting system in England was then developed and validated using semi-structured interviews, followed by a large survey conducted with 168 NHS staff participants, whose numerical responses facilitated the design of the resultant conceptual model. This model was considered valid based on the analysis of respondent scores and was therefore used to perform a time-driven activity-based costing study. A total of 103 NHS staff members with experience of incident reporting and management then estimated the time spent on each activity. The deterministic estimate of £956 million per annum of economic activity, needed to operate incident reporting systems, is a novel finding that exceeds previous low-quality estimates given in the literature. However, the economic costs of these systems do not encompass the consequences (and benefits) they may have for the English NHS. For this reason, an interrupted time series analysis using ten years of registry data from the Myocardial Infarction National Audit Project (or MINAP) was used to evaluate the impact of feedback reports produced by these systems. Using the most robust design, which incorporated a Scottish control group, no effect of feedback reports on bleeding rates in heart attack patients was detected. This finding, together with those described in each chapter, represents an important advance in the application of methods that can address economic questions of intrinsic importance to patient safety. Several methodological challenges emerge from this thesis, and these can be addressed by the research community to estimate the cost-effectiveness of incident reporting systems, along with other complex interventions for patient safety.