Climate change in cardiothoracic intensive care.

Abstract

A recent article in this journal investigated the link between individual cardiac anaesthetists and mortality as an outcome and concluded that publication of individual anaesthetist mortality was unwarranted 1. The accompanying editorial described this as a fundamental question and called for further outcome research into cardiac surgery outcomes 2. Cardiac surgeons in the UK and the US have of course become familiar with this kind of disclosure of individual and unit mortality 3-5. The choice of mortality as an outcome is widely questioned and is considered by many as too blunt an instrument to measure the quality of a service. What is indisputable is that it is easy to measure and definitive. Needless to say, it is also an outcome that is very important to the patients and their families. Critical care is a crucial, although not necessarily essential, part of the pathway for cardiac surgical patients and management therein is a major contributor to outcome. Cardiac surgeons have, correctly in my opinion, identified the current system in the UK as a flawed system of outcome audit – the entire public responsibility for a death (and likewise for a successful outcome) lies with a named surgeon. Many decisions are made in critical care in both the acute postoperative phase and in the long-term recovery period that impact on outcome. This has led to many an animated discussion at the head of the cardiac operating table and at the foot of the intensive care bed. The public disclosure of unit-level cardiothoracic intensive care outcomes may go some way to rebalancing these discussions. In this context, the article published in this edition of Anaesthesia by Shahin et al. is to be welcomed 6. The Intensive Care National Audit and Research Centre (ICNARC) have a track record which is internationally recognised as the most comprehensive national critical care outcome audit in the world. Currently, all UK NHS general intensive care units submit their outcome data to ICNARC for comparative analysis. Cardiothoracic critical care has evolved somewhat separately to general critical care, and, until relatively recently, many surgeons were effectively directing cardiothoracic critical care. Possibly due to the historical narrative linking surgery directly with mortality, little attention has been directed towards comparative national critical care outcome audit for cardiothoracic ICU. The Association for Cardiothoracic Anaesthesia and Critical Care (ACTACC) – note the significance of the recent name change – whose members now deliver most of the cardiothoracic ICU services in the UK, entered into a collaboration with ICNARC around five years ago to address this gap. The intention was to ensure development of an audit that was fit for purpose for cardiothoracic ICU and to recruit UK cardiothoracic centres to join up. The project described here by Shahin and colleagues and the ever increasing number of units that are submitting data suggest that these objectives are close to realisation. The first cardiothoracic units to join the ICNARC case mix programme were compared with general intensive care units. This is useful for quality indicators such as methicillin-resistant Staphylococcus Aureus (MRSA) acquisition and catheter-related sepsis. However, it is not really suitable for auditing mortality – it is difficult to compare a unit that receives elective cardiac surgical patients with a unit that receives a high proportion of critically ill septic emergency patients. Enough cardiac units are now submitting case-mix data to ICNARC to enable meaningful comparison between these units. This is published online on the ICNARC webpage, with a similar sub-speciality audit for neurocritical care 7. But how does the cardiothoracic surgical patient differ from the general critical care patient – can we draw out some broad differences? Obviously, there are different kinds of cardiac surgical patients and the elective majority will have a short stay similar to a non-cardiac surgical patient. However, there are also a significant number of higher risk elderly patients undergoing complex cardiac surgery, with an increasing proportion of urgent cases. The combined insults of pre-operative morbidity, cardiopulmonary bypass and ischaemia may be significant, and the final common pathway of end-organ dysfunction and failure in these patients may have much in common with the emergency patient admitted to a general intensive care unit with sepsis and circulatory failure. Unlike septic patients with acute circulatory failure, sick cardiac patients will be optimised in theatre before ICU admission, most likely using transoesophageal echocardiography to guide fluid and vasoactive drug infusion therapy 8, 9. In interpreting recent information, the implication is that there has possibly been an over-emphasis on the technical aspects of the intra-operative period and much greater emphasis should be brought to bear on the patient and their fitness to undergo surgery 1. Improving a patient's pre-operative condition, may be seen to reduce the ICU mortality risk for a patient by improving the ICU 24-hour admission data. As the authors correctly identify, there are a number of issues around using pathophysiological data collected during the first 24 h after surgery to predict the risk of death in a patient group following cardiac surgery. The Acute Physiology and Chronic Health Evaluation (APACHE-2) score has been demonstrated to perform poorly as a predictor in cardiac surgery patients in comparison with other models 10. This was confirmed by a UK study, which also demonstrated the ICNARC model was a significantly superior predictor of outcome and performed in line with EuroSCORE-2, the widely used pre-operative risk-prediction scoring system 11. Other studies from Canada and Germany have evaluated alternative postoperative cardiac surgery predictive scores, however their validation is limited by the smaller selection of variables and the fact they are based on data analysis from a single centre 12, 13. The National Institute for Cardiovascular Outcomes Research (NICOR), which produces the risk-adjusted outcome data for cardiovascular surgery, is a well-validated pre-operative dataset which includes a specific score for the surgical patient with categories of severity for example for pulmonary hypertension and left ventricular ejection fraction. As a dataset designed for ICU admission and comparison, the current ICNARC case-mix model currently uses very limited pre-operative morbidity categories and no other pre-operative data. However, it would be very appealing to investigate the inclusion effect of this data on the risk-prediction performance of the model in a large multicentre cardiothoracic database 14. Some very interesting work has recently been published using the Australia and New Zealand Intensive Care Society database and linking it to the corresponding cardiac surgical database. An acute risk-change following cardiac surgery at critical care admission can potentially identify underperforming institutions better than conventional mortality data 15, 16. Indeed, discussions are taking place in the UK regarding just such data linkages between the case-mix programme and other national outcome audits such as NICOR. There are a number of barriers and difficulties to overcome for such data-sharing collaboration, but the improved outcome information and research possibilities for patients and their families and directors of services should hopefully justify the effort involved. The population of modern cardiothoracic critical care consists of much more than just elective cardiac surgical patients. Many units have expanded to answer the increased demand from critically ill cardiology patients, the need for extracorporeal support for the patients at the extreme end of the spectrum of cardiorespiratory failure, and patients with acute major vascular syndromes. These patients are also included in this analysis and some will also be treated in non-cardiothoracic general intensive care units. Changes in the service specifications for commissioning of adult critical care services by NHS England now dictate that all cardiothoracic critical care units will have to submit their data to ICNARC in the near future – as discussed, all general critical care units in England already submit their data. Like general critical care, cardiac critical care units come in all shapes and sizes. Annual case volume varies from around 500 admissions to greater than 2000 admissions per year. Cardiac critical care may be a standalone facility or set within an acute large tertiary hospital. Staffing models across these various forms may also vary significantly – resident out-of-hours rotas are often constructed by necessity using non-intensivists. It is important to understand the differences and the problems encountered by different units and it is important that the performance of each unit is measured and compared. Collecting unit demographics and staffing information should enable comparisons and correlations with outcomes and continue to drive quality improvement. The prospect of significant research opportunities is exciting. Large-scale service reconfigurations are difficult to conduct and may take many years and undergo continuing challenges – a good example are the recently announced changes to congenital cardiac disease services. It is important that service reconfigurations are designed around patient benefit rather than political intention. The key to the patient being at the centre of service changes is good quality reliable data. Examples where ARCTiC could help us to understand and plan better services include the treatment of such conditions as acute aortic disease, out-of-hospital cardiac arrest and acute coronary syndromes. Research questions may include where critically ill acute cardiology patients and out of hospital cardiac arrest patients are best treated for optimum outcome - tertiary interventional cardiac centres or other centres. Clearly, inpatient mortality is a very limited outcome. Mortality at five and ten years should be goals for future research. As critical care matures, attention needs to be directed towards other outcomes, which may be more important to patients and their families. As Shahin et al. correctly identify in their conclusions, morbidity and longer-term functional outcomes are important to collect and integrate into our critical care datasets. In conclusion, this article reveals some of the hard work that goes on behind the scenes as cardiac critical care outcome data emerges from the long shadow cast by the cardiac surgical mortality database and gives us a glimpse of some of the possibilities that this work may present for our future patients. NF is the President of the Association for Cardiothoracic Anaesthesia and Critical Care (ACTACC) and a member of the ACTACC/ICNARC steering group. No other external funding or competing interests declared.