In a memorable quotation from the 1999 Hollister Lecture at Northwestern University, Illinois, Sir Cyril Chantler said “Medicine used to be simple, ineffective and relatively safe. Now it is complex, effective but potentially dangerous”.¹ For 50 years the dangers have been recognised, initially as diseases directly related to medications² and subsequently as hospital induced complications in which an early study showed that 20% of patients admitted to hospital were damaged by the care process. This was regarded as the price of medical progress.³ Neither the medical profession nor departments of health were prepared to address the issues. However, by the 1970s in the USA the insurance industry was crumbling under the weight of a growing number of claims of alleged medical negligence. To reveal the size of the problem the Californian Medical Association commissioned a group of doctors to develop a method of case record review. They showed that about 1% of patients admitted to hospital suffered an iatrogenic event as a result of negligence.⁴ Ten years later the Commissioner of Health of New York State asked Harvard Medical School to re-examine the issues. The resulting study showed that defects in medical treatment caused disabling injuries in 3.7% of patients admitted to hospital.⁵ The insurance industry had temporarily stabilised and neither study made much impact.

It took the media to bring the problem to the surface—especially after Betsy Lehman, a Boston Globe reporter, died in 1995 from an overdose of cytotoxic drugs.⁶ At the end of 1999 the Institute of Medicine published the document “To Err is Human”⁷ in which it was asserted that 48 000–98 000 Americans died in hospitals every year as a result of preventable errors, followed in 2000 by “Crossing the Quality Chasm”.⁸ This led to a huge flurry of activity. Within 2 years 15 states had passed laws to make the reporting of adverse events mandatory; five states and the District of Columbia requested voluntary reporting systems.

Meanwhile, clinicians in Australia were quietly ahead of the game. In 1988, in Adelaide, Runciman had initiated incident monitoring and the following year the Australian Patient Safety Foundation (APSF) was established.⁹ Subsequently, a case record review study based on the Harvard model,¹⁰ but addressing quality of health care rather than negligent practice, showed that about one in six patients admitted to hospital were subject to adverse events of which half were deemed preventable. This prompted strong reactions from government, healthcare professionals, and the public at large. As a result, the Australian Incident Monitoring System (AIMS) was set up with APSF collecting and analysing the data generated.¹¹ Many publications aimed at improving anaesthesia and intensive care followed.

In England and Wales action was triggered by several high profile cases of serious wrongdoing by individual doctors and by the Bristol scandal in paediatric cardiac surgery.¹² The Chief Medical Officer was determined that clinicians should face the problem.^13,14 In July 2001 the National Patient Safety Agency was established as a Special Health Authority. It aimed to develop a national reporting system for critical incidents. This has led to several national initiatives including measures to minimise risks of medications and of hospital acquired infection.¹⁵

But is incident reporting enough? Studies show that doctors remain unwilling to report. In 2003 Don Berwick, CE of the Institute for Healthcare Improvement, wrote “… so far I see no evidence that health care in the United States is becoming safer—the healthcare workforce remains blind to the enemy”.¹⁶ This may be unfair to the VA Patient Safety Programme,¹⁷ but certainly the efforts to make hospital care safer are, at best, patchy.

How well do they do in Japan? In this issue of QSHC Nakajima and co-workers¹⁸ describe a sophisticated system of voluntary anonymous computerised reporting in a 1000 bed hospital. It has an impressive management structure, facilitated feedback, educational seminars three times a year (with monitored attendance), ward rounds by peers, and ward checks by area clinical risk managers. 200 reports a month are received and the authors report examples of corrective action. Certainly the healthcare workforce in Osaka University Hospital must be well aware of safety problems. Yet, of the incidents reported, only 10% came from doctors. In their discussion Nakajima et al suggest that their system may miss episodes of “poor medical practice”. This suggestion is supported by a report that slipped quietly into the medical literature. The research unit of the Royal College of Physicians of London set up a collaborative study in which experienced physicians examined the case records of 200 patients who died after being admitted to large general hospitals as emergencies.¹⁹ The reviewers agreed that, in 10% of cases, erroneous management had contributed significantly to the fatal outcomes. Almost certainly, none of these cases were reported as incidents.

There is therefore a need to integrate other methods of recognising substandard hospital care with incident reporting. Wolff et al²⁰ have shown how this may be achieved to good effect in a small community based hospital in Australia, and the recently described comparative methodology from Aquitaine provides additional interesting ideas.²¹ The problem of patient safety is universal and much more needs to be done to engage the whole workforce at a national level.