Improving timeliness while improving the quality of emergency department care

See also pp. 303–312 Increased ED waiting times, increased ED length of stay and their association with adverse events, including mortality,1, 2 have refocused government policy on time targets. Time-based performance targets for emergency medicine were highlighted as a potentially useful strategy in the mid 1990s in Victoria with the Emergency Services Enhancement Program, when bonus payments were shown to improve patient access.3 In the UK, reform of the National Health Service (NHS) included implementation of the 4 h standard or target in 2004,4 with Western Australia (WA) and New Zealand following suit in 2009.5, 6 The original intent was that 98% of patients would be seen, admitted, discharged or transferred from ED within 4 h. This was soon revised to 95% in the UK and 85% of presentations in WA. In contrast, the New Zealand Shorter Stays in EDs (SSEDs) target was set at ‘95% in six hours’, following advice from local clinicians and nurses, with the belief that 6 h was ‘long enough for good clinical care but not unjustifiably long’.6 Evaluation of the NHS targets suggests that they have been successful in reducing long waits in the ED. Significantly, UK emergency physicians describe the major benefit of the target as enforcement of a ‘whole of system’ change without focus on the ED performance in isolation.7 However, there are conflicting reports on associated clinical outcomes, with concerns about the potential for a fixation on meeting the time standards to the detriment of the quality of patient care.8 It is notable the new British government responded to these fears by modifying the standard in 2011 to include a suite of eight clinical quality indicators (CQIs) that were designed to focus on timeliness along with the quality of clinical care and the patient experience.9 In Australia, the impact of time targets on improving the quality of care is unclear, with the Stokes report from WA suggesting need for further study and further monitoring.10 Initial research has suggested a possible reduction in mortality since the introduction of the 4 h rule in WA;11 however, methodological issues, including the lack of long-term trends, limit any conclusions at this stage.12 Introduction of mandatory reporting of CQIs potentially results in a more balanced approach to patient care.13 As such, the danger of only achieving faster care should be reduced. CQIs need to encompass measures of clinical outcomes, clinical effectiveness, safety and service experience, as well as timeliness. In doing so, they align with the Institute of Medicine's Quality Care framework of effectiveness, safety, timeliness, patient/family centredness, access and efficiency.14 CQIs should also support a culture of continuous improvement in the quality of care; however, this can only be achieved with ongoing monitoring and regular audit, alongside feedback and implementation strategies – which is not explicit in the introduction of these indicators. In this issue of Emergency Medicine Australasia, Jones et al. report on a comprehensive process to develop a set of evidence-based and clinically and locally relevant quality of care indicators for New Zealand.15 Indicators were initially selected following review of the existing international literature with feedback from the New Zealand emergency healthcare community. This was followed by subsequent validation by a multidisciplinary reference group. The reference group was representative of the wider health system and the validation process was approached from a system's point of view, as opposed to a focus on the ED alone. An understanding of the elements of quality care by the reference group was demonstrated, with initial development of a framework comprising communication, access, appropriateness, timeliness and satisfaction. In spite of this, the resultant suite of indicators does not actually reflect all of these domains of care quality. The authors acknowledge that their focus on time-based indicators is a limitation; however, they outline recognised associations between these indicators and patient outcomes. Furthermore, the authors do not fully explore the likelihood of a limited set of indicators influencing clinicians to undertake perverse actions in other clinical areas in order to meet targets limited to measurement of discrete events. An example of this might be unintentionally encouraging clinicians to make sure head injuries have CT scans quickly (to meet the target) and limiting access and delaying CT scans for patients with abdominal pain (because this is not measured). Other examples might include giving early antibiotics by triage nurses inappropriately, resulting in 100% compliance with the indicator target, but resulting in poorer patient care overall. Examples of these behaviours have been well described in other settings.16 The omission of CQIs that directly measure two of the five domains initially selected by the reference group, namely communication and satisfaction, is a concern, although this is possibly due to current limitations with use of the existing administrative dataset and retrospective audit of medical records. In addition, a comprehensive set of quality indicators would require a major upgrade in IT systems. It should be acknowledged that there is no value in using CQIs in isolation; otherwise there is a danger that measurement and achievement of pre-specified goals becomes the inappropriate primary focus. CQIs should be regarded as the means for driving change and improving performance.17 Accordingly, the next phase for the SSED project is measurement of these indicators to monitor the effects of the time targets on the quality and safety of the care delivered, alongside patient outcomes. However, this is reliant on clear standardised definitions of the CQIs and measurement methods. It is not yet apparent how this could be done in the New Zealand setting. The use of clinical quality registries to measure process and outcomes in some of the high-risk and high-cost areas, such as trauma, cardiac and intensive care patients, does not seem to have been emphasised as a potential future direction for quality measurement and improvement. This is a major focus of the Australian Commission on Safety and Quality in Health Care18 and will enable fully risk-adjusted and standardised comparisons over time and between institutions into the future. Although these clinical areas are limited in scope, the information is more detailed and better validated than from administrative databases. Standardisation of definitions and comparisons of similar groups of patients is essential if benchmarking is to occur between facilities, even though there is little agreement internationally about how to do this. A quality indicator, such as readmission, seems straightforward, but defining planned versus unplanned and defining optimal rates is difficult. For example, does return of a patient with abdominal pain and appendicitis the following day represent bad care when it was one possibility of early and appropriate discharge with undiagnosed abdominal pain? Having no unplanned readmissions might represent bad care, that is, an inappropriately high admission rate. Jones et al. have instigated an important process for continuous quality improvement which, with appropriate benchmarking, will serve to drive better emergency patient care, and the current suite of CQIs will no doubt be reviewed and refined over time. The integration of these indicators into a comprehensive framework that is designed to monitor and improve the quality and safety of ED care is essential, particularly in the context of persisting high demand for ED services, crowding and the imposed time-based targets. Promoting a small number of isolated CQIs without integration into a large comprehensive quality programme that measures structure, process and outcomes across the Institute of Medicine domains of quality is a risk that should be avoided. In addition, it will be important to ensure an appropriate governance structure to implement and refine required improvements to clinical care as they are identified across institutions and emergency care systems. None declared.