In 2000, an article published in this journal described a method of reviewing serious incidents based on peer group discussion, a model then in its sixth year (Reference RoseRose, 2000). This paper describes how the model has changed in response to the National Health Service (NHS) developments since then.

The original purpose of the review had been to expand the range of people involved in identifying lessons from ‘bad clinical outcomes’. Staff of all grades and from every mental health discipline were included, as well as other professionals such as the police, the fire brigade, medical ethicists, health and safety experts, and lawyers. The reviews took place three times a year, lasted 3 hours, used external facilitators, and resulted in a report being forwarded to the organisation's operational board. Progress on implementing recommendations was fed back in subsequent meetings. The reviews proved popular, partly because of the sense of local ownership the process fostered, and partly because they were an opportunity to discuss difficult cases from a wide range of perspectives.

Since 2000, three particular changes in the NHS have influenced how the reviews have developed: the expansion of formal management procedures for dealing with ‘bad outcome’ incidents; the appearance of new ways of delivering the clinical service; and a greater interest in analysing the failure of systems rather than individuals. Although these influences have shaped the content of the peer reviews, the basic structure has remained unchanged.

How the peer reviews have changed since 2000

Compared to 10 years ago, serious incidents are now much more likely to be the subject of a formal management-led review, frequently with external input. The peer group review has therefore moved from often being the sole form of inquiry, to usually complementing a prior management-led one. As a result, the findings of these inquiries are now routinely presented at the peer review meetings. This is usually done by members of the inquiry team, in addition to a presentation by the relevant professionals involved in the case. The fact that this arrangement works reflects the gradual build-up of trust in the process over time, helped by tough adherence to a ‘no blame’ culture. As a result, the interests of good patient care are served by an extremely inclusive review process, which combines peer group frontline expertise with managerial inquiry skills. Clinicians and other professionals benefit through regular participation in case-centred discussions about managing risk, while management benefits through enabling frontline staff to be part of the process of identifying and learning lessons from bad outcomes.

The complexity of psychiatric care delivery has increased in recent years, with the development of a great many more specialist teams working alongside each other, especially in the community. One of the unintended consequences of this more compartmentalised way of meeting needs is that service users often seem to get into trouble at team boundaries (Reference ApplebyAppleby, 2006). The peer review process, because it draws its participants from across and beyond the organisation, has therefore become a key tool for taking a cross-boundary system-wide approach. This has been as important within and between clinical departments as it has been between agencies.

During the 1990s, as mental health services moved increasingly into the community, they became more politicised. One of the consequences was the development of a serious incident culture, fed by the media and politicians, where blame was legalistically apportioned (Reference SalterSalter, 2003; Neal, 2004). However, this culture has been increasingly challenged and seen as producing little new knowledge at considerable cost. Salter (Reference Salter2003) writes of inquiries as being essentially unhelpful because they all reach similar conclusions, the implication being that serious incident inquiries imposed from above may not be sufficient in themselves at getting people to learn from bad outcomes. More inclusive ways need to be found which will recruit as many staff as possible in order to achieve better ownership of reviews and their recommendations. But this was not easy in the highly politicised serious incident climate of the 1990s.

Arguably, this politicisation has abated a little in recent years, or at least been deflected onto other areas of medicine. This has coincided with a slight shift in thinking about bad outcomes, and the realisation that the causes of untoward events in large organisations often have complex interrelationships and need to be reviewed in this light (Reference Andersen and FagerhaugAndersen & Fagerhaug, 2000; Reference Vincent, Stanhope and Taylor-AdamsVincent et al, 2000; Neal, 2004). The peer group review has from the start taken a more inclusive systems approach to understanding bad outcomes, and recently this has become even more evident. Importantly, it has also become the norm in local management-led inquiries. This is exceptionally important, for a peer group review system will find it hard to survive in an organisation that tends to blame individuals.

One final change over the years, in response to internal demand rather that outside circumstances, has been to use the review as a platform for presenting general information on risk. Areas covered have included nationally publicised inquiries, local audits, the National Confidential Inquiry reports (Reference ApplebyAppleby, 2006), Driver and Vehicle Licensing Agency and gun law briefings, evidence-based reviews, and presentations from the local Multi Agency Public Protection Arrangements, Multi Agency Risk Management, and the NHS counter fraud and security management service. Where possible, these presentations are linked to local incidents or circumstances.

Before and after 2000

Incidents reviewed

Table 1 lists the incidents reviewed during the years 1994-2007. The range of incidents appears to have broadened post-2000, and for the first time included sexual assaults and sexually inappropriate behaviour, driving and firearms incidents, the stalking of staff, illicit drug use on in-patient units, and system failures such as boundary problems between teams, and temporary ward closures due to escalating challenging behaviour. The reason such incidents were not covered before 2000 may reflect a different perception of review priorities, which in the 1990s were dominated by a newly introduced government target for reducing suicides. The inclusion of sexual incidents after 2000 may reflect a growing awareness of the difficulties of looking after acutely ill in-patients on mixed-gender wards, where those with challenging behaviour may be nursed alongside people who are both vulnerable and exploitable. It is also possible that higher thresholds for admission combined with bed losses have concentrated behaviour problems in a way that makes sexual incidents more likely. However, since early data collection on the frequency and severity of such occurrences is unreliable, it is difficult to be certain about this. The concentration of challenging behaviour effect may also have contributed to the more alarming occurrences of hostage taking, police use of CS gas, and ward closures due to escalating behaviour problems. The serious incidents of illicit drug use on hospital property appeared to reflect a growing problem in the local community since the late 1990s.

Table 1. Oxford serious incident reviews 1994–2007

Incidents reviewed	1994-2000	2000-2007
Suicide	27	29
Homicide	1	2
‘Near miss’ of suicide or homicide	11	3
Physical assault (in hospital)	14	9
Sexual assault (in hospital)	-	8
Inappropriate sexual relationship (in hospital)	-	3
Hostage taking (in hospital)	-	2
Use of CS gas (in hospital)	-	1
Illicit drug use (in hospital)	-	3
Offending while on leave	-	1
Psychiatric system difficulty	-	3
Firearms issue	-	4
Driving issue	-	1
Stalking of staff	-	2
Arson on hospital property	7	1
Hypothermia while AWOL	1	-
Cardiac arrest during ECT	1	-
Patient found unconscious in disused part of hospital	1	-
Total number of incidents	63	72

Constraints and difficulties

Not all bad clinical outcomes can be put down to system problems. Individuals can make serious mistakes, or be negligent, and need to be held accountable. In these situations, particularly if there appear to be no wider lessons to be learned, an incident may be less suitable for a peer group approach, or perhaps may need to be looked at anonymously. For the aim is not to have a witch-hunt, but to look at bad outcomes in an atmosphere that fosters trust, openness, and self-reflection, but does not shy away from asking penetrating questions.

Conclusion

A long-established peer group serious incident review has adjusted to recent NHS changes, with its importance, if anything, growing over time. This is partly because it enables as many clinical staff as possible to contribute to the organisation's thinking about risk management, and partly because it serves as an effective educational tool. This latter role will become of ever greater importance as the education agenda of Modernising Medical Careers takes effect, with its emphasis on self-reflectivity and learning within the clinical context. In fact, one of the challenges for all health systems trying to improve clinical care is to design effective structures and feedback loops for learning from bad outcomes, and for these to become deeply embedded within an organisation's learning culture. It is hard to envisage peer reviews not playing a key part of this learning and governance philosophy in the years ahead.

Declaration of interest

None.