Hospital, where you go to get better, can have the opposite effect and high on the list of hazards is infection acquired while there. Progress has occurred but more needs to be done. IT opens up great possibilities for scaling mountains of data that could improve patient welfare and save wasted money.
Hospital-acquired (or healthcare-associated) infections (HAI) are often assumed to be an unavoidable consequence of placing sick or injured people together in hospital wards, or even single rooms, where they are cared for by busy healthcare workers who can transfer bugs from one to another.
According to a recent WHO review, around one in 20 hospital inpatients and up to one in three intensive care unit (ICU) patients in high-income countries will develop at least one HAI, which at worst can be fatal or at very least distressing for the patient.
Many HAIs are caused by bacteria that are resistant to most commonly-used antibiotics – so-called multi-resistant organisms (MRO) – which are more likely to colonise patients who have recently had a course of antibiotics, as about half of all inpatients have, during their hospital stay or before admission. Some MRO-colonised patients will develop a serious HAI, but others remain unrecognised and a potential source of spread of the MRO to other patients.
At least half of all HAIs (and MRO acquisitions) are judged to be preventable but it often requires a special program to motivate healthcare professionals to observe well-established, effective infection prevention measures consistently.
For example, ICU patients are most at risk from blood stream infection (BSIs) associated with catheters (central lines) inserted into large veins to administer fluids and medications. In 2004, a well planned preventive program in more that 100 ICUs in Michigan led to a 66% reduction in CLABSI rates. Similar programs in ICUs in New South Wales and elsewhere in Australia achieved reductions in CLABSI rates of 50% or more.
Measuring HAI rates, comparing them between similar hospitals and public reporting of the results usually leads to improvement. The only publicly available national HAI data in Australia are Staphylococcus aureus BSI (SaBSI) rates, which have fallen significantly since they were first published on the MyHospital website in 2011. National hand hygiene compliance rates, also published since 2011, have improved as well. Despite some controversy about the reliability of these data, the improvements are real and significant .
But there is no cause for complacency. In high-income countries with national HAI surveillance programs, all BSIs represent only about 10% of HAIs; catheter-associated urinary tract infections (30-40%), surgical site infections (20-25%) and pneumonia (10-20 %) are far more common.
Australia is one of few high-income countries without a national HAI surveillance program; we have no overall national HAI data. Some jurisdictions and individual hospitals have HAI surveillance programs, but they differ in quality, scope and methods; even if results were shared, they could not be used for benchmarking, which is a powerful motivator for improvement. As a result there is significant duplication of effort and unnecessary cost, for limited benefit.
For many hospital administrators and Health Departments there is little incentive for HAI surveillance. Most HAIs are not life threatening and they often occur after discharge; so hospitals and treating doctors are either unconcerned about or unaware of them. The community and patients themselves bear the costs – financial and otherwise – of extra GP visits, medications, lost income, inconvenience, delayed convalescence and stress.
Apart from the ethical imperative to minimise preventable harm and unnecessary costs, surveillance has other benefits. It can help us understand the epidemiology of HAIs, including unrecognised, but potentially modifiable, patient risk factors and geographic, demographic or organisational differences; it could identify the emergence or inter-hospital spread of MROs, monitor the effectiveness of preventive measures and inform HAI policy development.
A well-designed national HAI surveillance program would be cost-effective, but health systems focus on politically-sensitive short-term performance measures – annual budgets, national emergency access targets (NEAT) and surgical waiting lists rather than equally, if not more, important individual patient outcomes.
Unfortunately, conventional manual surveillance is time-consuming, resource-intensive and therefore expensive. The use of routinely collected administrative data, such as ISD-10 codes and specialist registries – some of which collect HAI data – is an attractive alternative. However, although there are vast amounts of underutilised administrative data, most are neither sensitive nor specific enough to be used, alone, for HAI surveillance .
On the other hand, the belated, but now rapidly progressing, rollout of electronic medical records in Australian public hospitals provides an ideal opportunity to take advantage of emerging electronic surveillance data-management systems and well-designed computer algorithms, which could reduce data collection time by up to 65% and provide more consistent, accurate data for benchmarking .
A well-designed HAI surveillance, feedback and reporting system would not be cheap, but would have enormous benefits in preventing avoidable suffering and deaths, reducing healthcare costs and giving healthcare staff more time to care for patients.
Lyn Gilbert has had a long career as an infectious disease physician working in major hospitals. Her second PhD (in progress) concerns the ethics and politics of infection control.