Staphylococcus aureus bacteraemia (SAB) is common, and it causes serious morbidity and mortality. In Australia, it is estimated that there are over 6000 episodes per year, most of which are health care-associated.1 SAB is also a major problem internationally, and this problem is compounded by antibiotic resistance. Methicillin-resistant S. aureus (MRSA) infections are common in most countries, and MRSA bacteraemia has reached epidemic proportions in some areas of the United States.2,3

The study by Turnidge and colleagues on behalf of the Australia New Zealand Cooperative on Outcomes in Staphylococcal Sepsis in this issue of the Journal4 highlights again how frequently these infections occur, their association with health care (particularly with intravascular catheters), and the high mortality rate associated with them (over 20% by 30 days). It also illustrates our rapidly growing problem with community MRSA strains, which caused over 6% of all the SABs in the period June 2007 to May 2008, during which Turnidge et al collected data.

Currently, we have only very patchy measurements of this problem in Australia in comparison with New Zealand, Denmark and the United Kingdom,2 where all SABs are measured. Despite probably causing more deaths in Australia than occur on our roads, there is very little investigation into the causes and the preventable factors of individual episodes of bacteraemia. Most of these infections are preventable. Numerous interventions (such as compliance with hand hygiene and improved care of intravenous catheters) have been shown to lead to major reductions in rates of bacteraemia.2,5,6 The current National Hand Hygiene Initiative7 should also achieve a significant reduction. The Australian Commission on Safety and Quality in Health Care has noted how significant these infections are, and the importance of surveillance.2 It has also taken this matter to the Australian Health Ministers’ Conference, which then endorsed the reporting of SABs in each jurisdiction and nationally.

It is time that all health care facilities started collecting and acting on information about SABs. However, this probably won’t happen effectively unless all SABs (based on pathology laboratory results) are made “reportable”. After mandatory reporting was introduced in the UK, the number of documented cases increased by 50%.2,8

It is important that we do more than just collect these data. We need to ensure that, in each hospital, every case is looked at to try to determine why it occurred. Then health care professionals need to intervene at their own local level to make sure that appropriate policies are adhered to (or, where necessary, changed). Nearly all people with SABs are either in hospital already, or will be admitted. Thus, it should be possible to see them at or near the time of their infection. This will usually mean that either a physician or an infection control practitioner would look at every case to determine whether it was the result of a health care-associated procedure, and also look into other issues, such as the likely cause (eg, intravascular catheter) and outcomes.2 This would then also allow timely feedback to the teams responsible for the patients’ current and previous medical care about potential preventable factors.9

Many episodes of SAB, even though they are health care-associated, have their onset in the community. Thus, all cases of SAB will need to be looked at, and not just those with their onset in hospitals.1,2,9 While this will seem arduous to some, at the Canberra and Austin Hospitals, it takes between 30 and 60 minutes to see each affected patient and review each case.2,9,10 Even our largest hospitals are unlikely to have more than a couple of hundred SAB episodes per year. As these episodes have a mortality of more than 20%, this is not an undue task to expect hospitals to perform. They need to devote resources to doing this, and make it part of the core duties of any infection control team. It is important that the data be analysed, reported and acted on in a timely fashion, and this process must involve those at the hospital executive level. The success and benefits of doing this at a local level are highlighted in the study reported by Dendle and colleagues in this issue of the Journal.10

The study by Turnidge et al shows how data can be analysed and reported efficiently at a national (and international) level, with relatively few extra resources being supplied, through the use of a web-based tool.4 Such a system allows summaries to be sent back to the individual hospitals so that they can look at their own data in a timely fashion. It also allows some external auditing of the data to ensure they are accurate. Importantly, it will allow hospitals of similar types to benchmark themselves nationally. This is important because we already know that hospitals in Perth have much lower rates of health care-acquired MRSA than those in the eastern states,11 and we should learn from them. Most Australian states do not have sufficiently large populations to make many comparisons internally, especially of their largest hospitals, where most cases of SAB occur. By using national comparisons, institutions with higher rates of infections should be able to learn from those that have lower rates of infections.

There are always arguments that looking at these data and having them publicly available does not help. However, the experience in the UK suggests otherwise. Amid controversy, SAB rates have been published on the web for many years, and this type of benchmarking, when coupled with funded infection control interventions, has led to a 50% reduction in the number of MRSA bacteraemias.2,8 This has not only resulted in substantial monetary savings, but, more importantly, has saved many hundreds of lives each year.

Internationally, we are seeing rising levels of SAB. This is principally the result of complications of health care, and most of these infections are preventable. It is time we measured these life-threatening episodes much more accurately, examined each case, and intervened to stop further cases occurring.