Punishing medical errors won't improve hospital safety or quality

Australian public hospitals will soon be penalised for serious errors, with the aim of improving the quality and safety of health services. The punitive measure was a core proposal in the pricing framework for Australian public hospital services presented to the Health Council of COAG (Council of Australian Governments) in March 2017.

Withdrawing funding for serious errors, known a “sentinel events”, was advised by the Independent Hospital Pricing Authority (IHPA) and will come into force in July 2017. Australia has eight nationally agreed sentinel events.

These include: “procedures involving the wrong patient or body part resulting in death or major permanent loss of function” and an “infant [being] discharged to the wrong family”. Such events are clear and unambiguous errors, which can be prevented.

But cases of maternal death associated with pregnancy and childbirth, also included in the list, are typically unrelated to deficiency of care. Penalising unfortunate events such as these is unlikely to improve the overall safety of the hospital.

Even where the event is a clear, avoidable error, there is little if any evidence their frequency reflects the overall quality of care provided by a hospital. IPHA’s recommendations are drawn from opinion rather than empirical evidence. While the idea may appear economically rational – we wouldn’t pay for faulty goods, for instance – it does not withstand scrutiny.

Policies on gross medical errors

The term “sentinel event” is generally considered to be a preventable event in health care that causes significant harm or death to a patient. It’s the kind of event that should not be seen in a modern health-care institution. For this reason sentinel events are sometimes referred to as “never events”.

In practice, very few errors actually fit the sentinel category. Most hospitals would have fewer than one such event per year. Only around 100 sentinel events are reported from well over 5 million hospital admissions across Australia each year.

The concept of sentinel events originated in the United States in 1996, when the Joint Commission – the body responsible for hospital accreditation – released its first formal policy. The purpose was to probe the cause of the mistake well beyond the individual; to explore the training of those involved, work pressures and the level of supervision offered.

This type of review would look particularly at the barriers and systems that should be in place to prevent such occurrences. Sentinel events would be pooled from multiple institutions and appropriate recommendations made to apply across the health system. Pooling information is important because sentinel events are so rare a single hospital is unlikely to see enough for common features to emerge.

For instance, a pooling of patients who took their own lives in US hospitals identified a high proportion who hanged themselves from door hooks or curtain rails. When the frequency of such events became evident it led to these fixtures being replaced with breakable plastic models.

These types of review programs grew through the 2000s and now exist in one form or another in Australia, New Zealand, the United Kingdom, Canada and US.

Do sentinel events reflect overall hospital quality?

There is no evidence sentinel events reflect the overall quality of care provided by a hospital. First, whether these events are in fact wholly, partially or not at all preventable is contested. While they are awful and we wish they were prevented, the idea that a perfect system could exist in human services with finite resources is unrealistic.

Second, understanding and addressing why one dramatic and unexpected event occurs does not make the whole system safer. That would be equivalent to saying because we know why the Titanic sank, all ships in the future would be safe.

Safety requires a comprehensive approach that addresses all domains of health care, including effectiveness, patient-centredness, timeliness, efficiency and equity of access.

Why we shouldn’t punish hospitals

Most people accept we should receive a refund if we pay for a product or service that does not deliver as advertised; the idea being that mistakes should not be rewarded. It makes sense public hospitals are not funded when a sentinel event occurs for that patient’s episode.

But given the constant demand for health funding, it is somewhat surprising there was broad support for this idea from the state and territory governments. The cynic would argue there are so few events (0.002% of admissions) that it does not affect the bottom line.

So what does not paying hospitals do? Perhaps it sends a message about preventing unacceptable practice. But paying for performance remains a widely debated idea, which has not achieved the hoped for gains in health effectiveness and safety.

It may have some negative unintended consequences. A punitive environment may make health-care organisations and professionals less likely to report errors and more likely to practise in a defensive manner. This might mean gaming the system to select the patient population that generates the greatest financial benefit – essentially avoiding sicker, and therefore riskier, patients.

Setting standards for the quality and safety of health care is a community responsibility and one a mature society accepts. Improving safety and health outcomes will be achieved most effectively if incentives are provided for excellent care, rather than punishing rare instances where care may have lapsed.

Joseph Ibrahim works for Monash University in an academic role involving teaching and research. He receives funding from Victorian Managed Insurance Authority, Department of Health and Human Services (Victoria) and Commonwealth Department of Health. He is affiliated with the Victorian Institute of Forensic Medicine and is the consulting editor for the Future Leaders Communique, Clinical Communique and is the editor in chief of the Residential Aged Care Communique.

John McNeil does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond the academic appointment above.