In 2005, I wrote about the new UK National Patient Safety Agency (NPSA).1 Table 1 lists some NPSA initiatives relevant to children introduced since then. However, patient safety is about much more than a government agency – it is about a huge cultural change in how doctors practise. Two case studies highlight some of the issues.

In October 2007 the US Department of Justice fined British Petroleum US$50 million relating to the 2005 Texas City refinery explosion which killed 15 people and injured 170. In November 2006 I spent a week with the oil company Shell, an attachment organised by the National School of Government. I wanted exposure to an industrial environment where, like paediatrics, safety is crucial and where things can go wrong very quickly. I retain three memorable images from my arrival at Shell, none of which depended on the “rocket science” of the high-tech petroleum industry. The car from the airport could park at Shell only by reversing into the parking space – accidents happen when drivers reverse out in the evening darkness after work. On entering the building and climbing the stairs, I was politely told I had to hold the hand rail. This behaviour would not come as second nature on a North Sea oil rig if not insisted upon ashore. And my hot drink had to be covered with a lid if I wanted to carry it around the office with me.

Here seemed to be an organisation imbued with a deeply entrenched safety culture, running through every aspect of its work. In their view, there were two ways of doing something – “the Shell way” and “the wrong way”. Some say that such an approach stifles initiative and creativity, but the correct place for these virtues lies in well conducted, controlled research of new ideas, not in allowing every member of an organisation to do their own thing. Almost 80% of Shell employees are external contractors and yet this same culture was enforced upon them too – one breach of the examples given above results in a polite reminder, two breaches a written warning, three breaches and the contract is terminated. This may appear draconian but the risks of getting it wrong are immense – as the Texas City explosion exemplifies. I doubt there are randomised controlled trials on holding hand rails, but absence of evidence is not evidence of absence.2 The guidance is part of embedding a safety culture as second nature. Can we in the health sector truly say we comply in the same way with patient safety alerts? While the examples from Shell relate to protecting their employees, the public and the environment (ie, traditional “health & safety”), and while the National Health Service (NHS) also has initiatives relating to these groups, this article is principally about the safety of patients.

Contrast the Shell culture with the following debate about hand washing.3 A hospital informed medical staff that since effective hand-washing reduces cross-infection, it seems reasonable to expect that all staff including doctors will comply with simple measures which promote the effectiveness of hand washing. The hospital noted that doctors comply very poorly with hand hygiene in audits (rates of compliance of 30–50% compared to 50–97% for nurses) and that rolling up cuffs and taking off watches are actions that facilitate effective hand washing. Ties should be absent or tucked in, and hand and wrist jewellery avoided.

A senior clinician alleged that the hospital’s assertions about personal hygiene were insulting and profoundly lacking in evidence, a view supported by some others. The hospital found it surprising that medical staff were so ready to challenge the professional advice of microbiology consultants who agreed these were sensible, prudent measures. The hospital required that staff confirm that they would comply with the code. It was left for a paediatrician to observe that the issue of shirt sleeves is common sense – if you don’t roll them up you won’t wash your hands as well. Our own children are taught to do that from a very early age – certainly in infant school!

Ignac Semelweis demonstrated in the 1840s that if all visitors to his antenatal wards washed their hands in a dilute solution of carbolic acid, the incidence of puerperal sepsis fell dramatically. Twenty years later, Joseph Lister began to clean wounds and dress them using a solution of carbolic acid. He was able to announce at a British Medical Association meeting in 1867 that his wards at the Glasgow Royal Infirmary had remained clear of sepsis for 9 months. Although his methods initially met with indifference and hostility, doctors gradually began to support his antiseptic techniques. Plus ça change in the health sector?

In a 4-year prospective Saudi study following a hand washing educational intervention, out of 1175 opportunities for hand washing, compliance among nurses was 97.5% while the lowest compliance of 37.6% was observed among doctors (RR 2.591, p<0.001).4 Yearly assessments following the educational program also revealed significant increases in hand washing compliance among nurses (p<0.001), while no statistically significant changes in the hand washing behaviour among doctors was observed during the study period.

A study from a Hong Kong neonatal unit showed that a decline in infection rates followed a hand hygiene educational intervention.5 The observational study was repeated 6 months after the completion of the intervention program: the health care-associated infection rate decreased from 11.3 to 6.2 per 1000 patient-days.

SETTING SAFETY STANDARDS FOR DOCTORS IN THE UK

Recent clinical failures highlighted by the Redfern Report on organ retention from children, the Bristol inquiry into paediatric cardiac surgery and the Shipman case (GP Harold Shipman was convicted of killing 15 patients in 2000 - he is believed to have killed at least 200 more) have eroded public confidence in the UK NHS.6^–8 Dame Janet Smith, the chair of the Shipman Inquiry, looked at the role of the NHS in ensuring patient safety and concluded that plans for a new system of revalidation (regularly checking doctors’ competency to practise) were not tough enough. Dame Janet said the lack of objective tests of doctors’ efforts meant the proposed revalidation procedure would not be able to fulfil its role evaluating a doctor’s fitness to practise. Revalidation is now upon us in the UK,9 although the final format remains under debate.

In the USA, certification is a voluntary process that allows doctors to demonstrate competencies beyond the minimum acceptable standards required for licensing. In 2002, more than 85% of licensed doctors in the USA held a valid certificate. There is some evidence that board certification is associated with better outcomes for patients and fewer disciplinary concerns.10 Although this does not prove causality, these data are difficult to ignore.

Whereas the General Medical Council (GMC) sets generic standards of behaviour, the UK Royal Colleges set clinical standards for specialists by developing an appropriate examination structure, encouraging evidence-based practice and having ultimate responsibility for the continuing professional development of their members. Certification in the USA provides assurance that a specialist has completed an educational programme and passed an objective evaluation of their knowledge, skills and experience. Successful candidates are certified as specialists. In the UK, success in membership examinations, such as the MRCPCH, does not usually confer specialist status, but instead is recognised as a qualification for higher specialist training.

SETTING SAFETY STANDARDS FOR SERVICES IN THE UK

The National Institute of Clinical Excellence (now the National Institute for Health and Clinical Excellence, NICE) was established in 1999 in order to reduce variability in clinical decision making, to improve access to care and to ensure evidence-based practice within the NHS. National service frameworks also aim to reduce variation in care by setting national standards and performance measures for a service. The Healthcare Commission was created in 1999 as an independent watchdog. Its aim is to check whether the local systems in place to monitor and improve the quality of care provided are safe and effective. The work of the Healthcare Commission is based on an annual self-assessment by trusts against a large number of standards. For example, there are standards covering a trust’s child protection work. This is supported by visits and patient feedback. Each trust’s assessment is readily accessible on their website.

The NHS Litigation Authority’s accreditation scheme (formerly the Clinical Negligence Scheme) has significant financial benefits for trusts who can demonstrate compliance. This therefore creates incentives for managers to introduce clinical risk management standards. For example, it includes measures to avoid the mis-identification of patients, blood transfusion errors, medication errors and hospital infections as well as measures on the reporting and learning from adverse events. These are all relevant to patient safety.

APPROACHES LIKELY TO CHANGE PAEDIATRICIANS’ BEHAVIOUR

All patients in the NHS are entitled to high quality care. This should not depend on the geographic accident of where they happen to live. The unacceptable variations that have grown up in recent years must end. Clinical decisions should be based on the best possible evidence of effectiveness, and all staff should be up to date with the latest developments in their field – Department of Health. A first class service: quality in the NHS11

A decade later:

There is variation in clinical practice in the NHS and in the treatments that patients receive. Some of this may be appropriate but patients may also receive treatments which are not very effective or for which there is no evidence, or they may not get treatments which are effective. Patients have the right to expect the same high standard of care and treatment wherever they access NHS services – Department of Health. Report of the High Level Group on Clinical Effectiveness, chaired by Professor Sir John Tooke12

Plus ça change in the health sector?

Interventions designed to change clinician behaviour include monitoring the quality of care provided in an institution (neonatal or paediatric audit), searching for evidence on best practice (research activities), making changes by using the evidence (see box 1) and monitoring the impact of the change on the quality of care provided (completing the audit cycle). Few of the recommendations in box 1 are new, as evidenced by advice from the Medical Protection Societies over the past two decades. Why the resistance to change?

We know from efforts to stop people smoking that while knowledge may be important, it is rarely by itself sufficient to bring about change. A different example, demonstrating the inertia of clinicians, is that reliable evidence of the effectiveness of antenatal corticosteroids in reducing the mortality of preterm infants was available long before antenatal corticosteroids were given routinely.

Practising paediatricians have to swim in an ocean of research evidence. They are expected to stay up to date, be authoritative and practise to a high standard. The transfer of research into practice is not easy and often needs to involve a range of techniques to promote, sell, implement and then maintain the adoption of guidance. The pharmaceutical industry takes these issues seriously and puts much effort and resources into persuading doctors to prescribe new drugs. A range of interventions has been shown to be effective in changing professional behaviour including:

• educational outreach - using a trained person to meet with health professionals (the classic “drug rep” approach) to provide information in a simple format with the intent of changing the health professionals performance,

• reminders – any intervention (eg, pens with brand names), manual or computerised, that prompts the health professional to change their behaviour.

Clinical guidelines are one example of translating research into practice which have attracted significant resources over the last decade. Each NICE guideline costs in the order of £250 000 to develop and £30 million for full nationwide implementation. Guidelines are used to promote evidence-based health care, ensure the highest standards of patient care, reduce variation in practice, and close “the gap between what clinicians do and what scientific evidence supports”. However, an analysis of their impact was circumspect about the effectiveness of clinical guidance.13 A review of 19 studies of passive dissemination of consensus derived recommendations for practice concluded that there was little evidence that passive dissemination alone resulted in behaviour change. Another review based on 23 studies assessing the factors influencing compliance with guideline recommendations found that compliance was lower for recommendations that were more complex and less easy to pilot. Finally, a review of 91 studies evaluated the introduction of guidelines. It concluded that guidelines can change clinical practice but were more likely to be effective if they took account of local circumstances, were disseminated by active educational interventions and were implemented using patient-specific reminders.

Scott14 has summarised the different historical trends in trying to “get research into practice”:

1. Browsing journals, attending conferences and listening to didactic lectures may have little impact on changing practice.

2. It was hoped that meta-analyses and systematic reviews, decision analyses and practice guidelines developed by authoritative groups would synthesise best evidence from multiple studies and define actions for particular sets of clinical circumstances. However, guidelines and systematic reviews did not gain as much traction as initially hoped in guiding practice, either because clinicians were unaware of them or refused to adhere to them.

3. Case reviews by peers, audit and feedback, and reminder systems resulted in modest effects in optimising care, shifting the proportions of patients receiving optimal care upwards by between 6% and 13%.

4. “Pull” rather than “push” – electronic information systems that were readily accessible and easy to use would allow clinicians to answer their immediate questions in real time at the bedside. Here the clinician was offered the means to find (or “pull”) answers to queries the moment they arose in practice.

5. It had been previously overlooked that it was the desire of clinicians to belong to, identify with and achieve recognition within a group of like-minded people that very much determined thoughts and actions. The social environment of health professionals is governed by norms and customs that can be unspoken and hidden.

6. Empowering patients and their advocates (including the media) to become more aware of care that makes a real difference.

A systematic review15 commissioned by the UK Department of Health concluded that innovations are more likely to be adopted if they have the following attributes:

• Relative advantage – innovations that have a clear, unambiguous advantage in terms of either effectiveness or cost-effectiveness,

• Compatibility – innovations that are compatible with the values and perceived needs of adopters,

• Simplicity – innovations that are perceived as simple to use,

• Trialability – innovations that can be experimented with by intended users on a limited basis,

• Observability – the benefits of an innovation are visible to adopters,

• Re-invention – if a potential adopter can adapt, refine or otherwise modify the innovation to suit his or her own needs.

PRIORITISING SAFETY ISSUES

It is important to know which patient safety issues cluster within a particular care setting since this may give clues to the root of the problem. A statutory requirement placed on the NPSA was to devise and implement a national reporting system for patient safety incidents.1 The commonest reported safety incidents for patients <18 years of age are different from those for adults. Comparing the 33 446 reports from January to December 2006 for children to the 663 658 reports for adults, the relative rates for the eight commonest paediatric issues were: medication problem 19% (children) vs 9% (adults), treatment/procedure problem 14% vs 7%, patient accident 13% vs 41%, access problem 9% vs 8%, documentation problem 8% vs 5%, clinical assessment 7% vs 4%, consent issue 7% vs 4%, and device problem 6% vs 3%. While it is gratifying that children suffer fewer accidents in hospitals or other health facilities, the high rate of problems with medications is particularly worrying.

Patient safety incidents in neonatology accounted for more than 12% of incidents reported for children and young people in the National Reporting and Learning System. The single most frequently reported neonatal safety incident was also medication error. Of the 13 320 reports from April 2006 to March 2007 for neonatal incidents (approximately 4000 from the <18 year olds dataset and 9000 from the obstetrics dataset), the rates were: medication problem 18%, treatment/procedure problem 16%, access problem 16%, infrastructure 8%, clinical assessment 8%, device problem 7%, documentation problem 7% and consent issue 6%.

In March 2007, the NPSA requested that a “Safer Practice in Neonatal Care Group” group be established, led by the RCPCH with other key stakeholders (the charitable sector, British Association of Perinatal Medicine, Neonatal Nurses Association and Royal College of Nursing), to identify areas of neonatology where “care bundles” could be implemented to improve safety. A care bundle is “a group of evidence based interventions related to a care process that, when executed together, result in better outcomes than when implemented individually”. Two potential areas have been identified: bloodstream infections and medication errors. The areas were chosen partly because data from the National Reporting and Learning Scheme identified safety incidents in these areas. The next phase is to identify components of a care bundle from current research evidence. For example, interventions for the care bundle on preventing bloodstream infections might include hand washing or disinfection prior to and after all patient contact, strict asepsis during central vascular device insertion, daily review of the continued need for a central vascular device, and daily review of the continued need for parenteral nutrition.

CHANGING THE SAFETY CULTURE

In a study entitled Good to great, Jim Collins and a team of researchers at Colorado University identified 11 companies which were transformed from a history of 15 years of returns below the stock market average to returns at least three times the market over the succeeding 15 years.16 How was this change achieved? There were two messages: the first is that the leaders were not as important as “getting the right people on the bus”. The second is the Stockdale paradox – you must retain faith that you will prevail in the end and you must also confront the brutal facts of your present reality. Jim Stockdale was the highest ranking US prisoner of war in Vietnam, tortured 20 times over 8 years: “What separates people is not the presence or absence of difficulty, but how they deal with the inevitable difficulties of life”.

So as paediatricians face the challenges of patient safety, and as we wonder whether we would prefer to be led by Shell or by our current masters, look not to politicians or to health service leaders, who will come and go. Look to “the people on the bus” – ourselves. Paediatricians must confront the unsatisfactory safety culture of today, determined to prevail in the future.