Enhancing safety in nuclear medicine in Australia: a multi-method investigation of risks, incidents and work-processes

Nuclear medicine is an important part of modern healthcare in which radiopharmaceuticals are administered to patients for the purpose of diagnosis or treatment. Different types of errors can compromise patient safety in nuclear medicine, however maladministrations are of greatest concern because patients are exposed to unintended ionising radiation with the subsequent potential for organ damage. Although Australia has a statutory incident reporting system, there is no contemporary national perspective on the incidence, causes or complications of maladministrations. As well, there is a paucity of applicable maladministration data from other countries. Therefore, an analysis of contemporary Australian incident data, allied with an approach which defines and measures safety in nuclear medicine in broader ways is warranted.! ! Since maladministrations are rare incidents, it is unsurprising that existing studies have been descriptive in nature. However, alternative statistical methods may be more suited to study maladministrations. One such method, known as control charts, has been employed for decades in industry and medicine as a quality process tool to graphically display and monitor temporal variations in key incident data. Control charts have advantages over descriptive studies in analysing rare incidents and by identifying 'unnatural' variations in data can lead to novel quality improvement strategies in nuclear medicine. ! ! Understanding how safety is maintained in dynamic, busy and interruption-laden clinical environments requires research beyond incident reports. Work observation studies provide insights into how care is delivered as well as contextual factors, such as interruptions and disruptions to workflow, which may jeopardise safety. Such investigations can also illustrate the way in which staff adapt to these dynamic environments and could complement existing quality and safety initiatives in nuclear medicine which emphasise technical training and procedural compliance as means to avoid maladministrations. No observation studies of workflows in nuclear medicine have been previously conducted.This thesis addressed each of these three challenges in order to understand safety in nuclear medicine, address existing gaps in knowledge and stimulate a more comprehensive suite of quality improvement strategies in the future. The program of research commenced with a detailed analysis of the causes, consequences and incidence of maladministrations from 2007-2011 using the Australian Radiation Incident Register (ARIR), as well as an evaluation of the strengths and weaknesses associated with incident reports and the incident reporting system in Australia. There were 149 maladministrations, with an incidence of 5.8 per 100,000 procedures. Nearly half (48%) were caused by failures in radiopharmaceutical preparation and 67% occurred in the work domain of nuclear medicine technologists. Ninety-eight percent of maladministrations occurred in a diagnostic context and led to a mean effective whole body radiation dose of 7.9mSv. There was significant heterogeneity of maladministration notification rates across Australian States and Territories (0-12.2 per 100,000 procedures; p<0.05), as well as evidence of incident underreporting (odds ratio=5.9). The ARIR could be improved by attention to latent causes of maladministrations, identifying barriers to notification, implementing uniform prescriptive notification criteria in all Australian States and Territories and better integrating with Medicare data. Second, application of control charts from 2007-2012 investigated factors associated with 'special cause variation' (signifying greater than expected) in monthly maladministration rates. Special cause variation occurred in only three of 72 months, but accounted for a disproportionately large number of maladministrations (21%; 42 of 197 patients). Most of these incidents (n=27) were due to maladministration 'clusters' in which multiple patients were affected either by errors in the bulk manufacture and preparation of radiopharmaceuticals or equipment failure. Control charts reinforce the idea that radiopharmaceutical preparatory processes are vulnerable, especially when occurring on a commercial or bulk manufacturing basis. The ability to actively monitor safety data in nuclear medicine is appealing and could foster engagement with key stakeholders.Third, an 100-hour work observation study of 11 nuclear medicine technologists at a major Sydney public hospital was undertaken from October to December 2015. The proportions of time spent in eight categories of work tasks, location of task, interruption rate and type and multitasking (tasks conducted in parallel) were recorded and specific safety-oriented strategies used by technologists were catalogued. Technologists completed 5227 tasks and experienced 569 interruptions (mean=4.5 times per hour). Interruptions during radiopharmaceutical preparation occurred at a mean rate of 4.4 times per hour. Some interruptions were initiated by other technologists to convey important information and/or to render assistance. Technologists employed a variety of verbal and non-verbal strategies in all work areas (notably in the hot-lab) to minimise the impact of interruptions and optimise the safe conduct of procedures. Although most were due to individual choices, some strategies reflected overt or subliminal departmental policy. These type of strategies may highlight process or organisational deficiencies not readily apparent from ARIR incident reports.The thesis findings show that the ARIR has a central role in characterising maladministrations, but refinements are needed. Control charts can broaden how safety in nuclear medicine is measured, provide new insights on vulnerable work processes and permit more active monitoring of incident data and stakeholder engagement. By illustrating how work is conducted rather than imagined, work observation studies in nuclear medicine offer an understanding of 'real-world' safety and vulnerabilities in nuclear medicine, thus informing quality improvement in complementary ways to incident reports. The research findings are not only pertinent to nuclear medicine in other countries, but can act as a template for promoting safety and refining quality more broadly in medicine.