The measures ordered by the NHS to try to deal with sustained pressure on the emergency care system have highlighted the difficulties of piecemeal approaches. In the second part of their series on why we need to revisit past lessons on whole system methods, James Barlow and Brice Dattée discuss Scotland’s experiences
The target for hospitals to assess, treat and admit or discharge 98 per cent of patients arriving in A&E within four hours (later reduced to 95 per cent) was first introduced in NHS England in 2002. Scotland followed later, launching its national initiative, the Unscheduled Care Collaborative Programme (UCCP), in May 2005. The aim of the UCCP was to help Scottish local health boards (similar to trusts in England) meet the target by December 2007.
This targeted the key relationships between hospitals and other bodies in their local care systems: primary care, social services and ambulance services. The aim was to divert patients from attending A&E where possible and speed up discharge of patients back home.
The rhetoric of the UCCP placed great emphasis on ‘complexity’, system interdependencies, ‘whole system’ change, and the need to empower local care organisations to search for their own solutions to their problems.
The unprecedented cuts in overall NHS finances have destabilised healthcare, while social care for older people is in meltdown
Local teams in each health board designed and implemented actions to address the various patient flows. The UCCP adopted a ‘plan-do-study-act’ (PDSA) model of short term and incremental experiments to help the local teams explore their interdependencies across the care system and monitor the impact of the changes on performance. The national team advised local teams, monitored progress towards the four-hour target, and organised national events where participants shared the results of their local initiatives and experiments. By adopting a whole-system approach, local teams had to jointly discover their interdependencies across different health system levels and agree on new ways of working.
Many of the interventions will be familiar to those working in A&E today: ‘minor streaming’, where minor injuries or illnesses are quickly recognised and treated; ‘see-and-treat’ areas staffed by GPs; improvements to the accuracy and speed of information across a patient’s journey; and other initiatives to facilitate the transfer of patients across departments and manage bed capacity more precisely.
At several hospitals, new non-medical roles – ‘bed busters’ – were created to swiftly clean up and prepare available beds without taking time from nurses. To reduce transport or information delays across departments radiology departments were moved closer to A&E, new procedures were introduced to obtain laboratory results quicker, a new role of ‘flow coordinator’ was created, hospital porter rostering was redesigned and even carpets were removed in one hospital to make new trolleys roll more easily.
The rhetoric of the UCCP placed great emphasis on ‘complexity’
Novel ways of working with part of the care system outside the hospital also had to be found as the various interdependencies were uncovered. Newly created discharge planning teams categorised patients and identified those ready for discharge but still waiting for transportation or social care, in order to prioritise their social care packages.
The initial emphasis of the UCCP was on streamlining the discharge process and working backwards across departmental boundaries, starting with the ‘back door’ of the hospital and the interdependencies with the ambulance service, community hospitals and social care. But it was also essential to reduce unnecessary attendances to A&E in the first place. This required better coordination with out-of-hospital organisations.
Out-of-hospital flow leaders worked with the NHS 24 telephone triage service and GPs to prevent unnecessary attendances at A&E. If possible, referrals of patients to A&E by GPs were timed to ensure a bed would be ready if needed. Other GPs were located in the ‘see and treat’ areas to act as the first point of triage.
Other collaborative experiments to address the out-of-hospital interdependencies included community paramedic schemes with ambulance services, schemes with pharmacists to provide oxygen concentrators out-of-hours, local directories for care services, and regional campaigns to educate the general public about when to attend A&E.
Even carpets were removed in one hospital to make new trolleys roll more easily
The initial reaction of many health professionals when the four-hour target was announced was to see it as another example of central government imposing top-down measures on them. But as the UCCP was established, they began to see the target as a form of permission to try out new ideas without fear of penalty.
As well as some evidence-based interventions that were mandated by the national programme team, there was a great deal of local experimentation and innovation. This involved individuals and teams broadening their understanding of the wider system they were situated in, the interdependencies and their effects on A&E. How easy it was to do this depended partly on the level at which interactions across the care system occurred. The higher the organisational level – between a hospital trust and local GPs or social services rather than between different departments within the hospital – the harder this was. This was because the greater the ‘cognitive distance’ between different teams or individuals, the more difficulty they had in trying to convince others that the A&E problem had anything to do with them. As one manager put it,
“Some clinicians still don’t believe [the need for a joint solution] because their perception is that the problem doesn’t lie with them. The four-hour target is an A&E issue, not a downstream issue. They don’t see how they could possibly affect it.”
When confronted with resistance, flow coordinators and teams tried to legitimise the problem by using quantitative data to demonstrate the interdependencies and need for collaboration. They also used rhetoric to create empathy by stressing how everyone is ‘in it together’ to raise the quality of care for ‘our’ patients.
As the PDSA experiments took place, demonstrable improvements – small wins – could be seen. These helped to reduce the cognitive distance between individuals or teams in different departments or organisations. As mutual performance improved, the weight of evidence from different experiments further reduced cognitive distance. According to one interviewee,
“We got the enthusiastic people who wanted to make it work, who could see the possibilities. Pilots were put in place, and that won the hearts and minds of others who saw it working. Once we started to get improvements and the results all began to mount up, people began to understand what we were talking about, and it was easier and easier to persuade other people. And then, once they started to see a real outcome in terms of their own patient flows, I think they realised it was a good idea.”
All this effort led to fairly rapid improvement: the 98 per cent target was reached about 20 months after the UCCP started. But when performance neared the 98 per cent target, diminishing returns began to set in.
This was partly because of the effects of factors that are hard to fix, such as the physical layout of hospitals. But another reason for slower progress was that, once the performance target was achieved, teams put less effort into maintaining it. Interviewees talked of people ‘revelling’ in the glory and beginning to sit back. Diminishing returns meant coordination became harder. Teams outside A&E stopped accounting for the wider interdependencies as they began to focus more on their own performance targets, which were competing for attention. This was reinforced by a lack of continuity of personnel as well as active resistance. Delays in observing the impact of initiatives, because of the inertia and time lags in the system, also played a part. Moreover, in time new Scottish policies and strategic reviews of health board services all helped to destabilise the wider environment, further dissipating the impact of local efforts under the UCCP.
Despite these pressures, performance at or near the target was sustained for about two years after the UCCP ended. Since then, however, performance has deteriorated consistently. Several factors seem to be responsible. In June 2010, the mandatory 98 per cent ‘target’ was converted to a ‘minimum standard’ and lowered to 95 per cent, with weaker sanctions. Staff turnover, including the departure of UCCP managers, also meant knowledge about which local initiatives worked or not was lost. The frequency of monitoring fell, as some changes were believed to be already firmly embedded in mainstream practice. More improvement programmes were started to meet other targets, and managers shifted their attention to the priorities of the moment.
As the UCCP was established, health professionals began to see the target as permission to try out new ideas
The pressures on the NHS are more acute today than at the time of the previous drive to improve A&E waiting times in the early 2000s. Demand on all care services has risen as the population ages: the number of people aged 65 and over grew by two million in the decade up to 2015. The GP system is hugely over-stretched, with no signs of improvement on the horizon. The unprecedented cuts in overall NHS finances have destabilised healthcare, while social care for older people is in meltdown. And the signs are that Brexit may significantly damage levels of staffing in the NHS and social service.
Most health service managers, health professionals and government now accept they need to treat the challenges of unscheduled care on a ‘whole system’ basis. But knowing what to do in theory and doing it in practice are very different matters in the NHS. A problem with the measures signalled last year by the Government is that attention remains on a four-hour target. The out-of-hospital landscape and interdependencies across the system are essentially unchanged, although in time the trend towards integrated care or ‘accountable care organisations’ may help.
Today’s pressures on health and social care make it all the more important for policy makers and NHS managers and leaders to focus on two areas: cohesion within the relevant individual components of the health system, ensuring they are well-designed and properly resourced, and coupling between these components, so the interdependencies are understood and carefully managed. This would provide real foundations for the basic building blocks for a more integrated care system capable of managing demand on A&E services.
The building blocks – each of which in itself poses huge challenges in a cash-strapped NHS and social care system – include the creation of an industrial-scale model of preventative care for older people. This has to include exercise programmes to dramatically reduce falls – otherwise screening for medical conditions is wasted effort. Another building block is services to keep patients with dementia and related conditions out of A&E. This requires diagnosis and response to be delivered outside hospitals, again at industrial scale. Thirdly, there are opportunities for improvement to reduce the number of people with routine problems such as blocked catheters or pressure ulcers who pitch up at A&E, but tackling these requires a systematic response from GP and community care services.
Knowing what to do in theory and doing it in practice are very different matters in the NHS
Does Scotland’s UCCP experience, with its explicit whole systems perspective, hold lessons for improving unscheduled care today? Well, yes. One is that the UCCP highlighted the importance of understanding ‘scale’ in the system: the way in which actions play out over different timescales and at different organisational scales.
Often, the way complexity theory is used for understanding or changing the behaviour of health systems is wrong because it fails to understand scale. Numerous uncoordinated, parallel performance targets and actions are introduced. These may well achieve change performance, but only at a sub-system level and only for a period before tensions in the overall system and unpredicted feedback begins to take effect.
The UCCP – with its emphasis on sharing knowledge and experience across all stakeholders at different system levels – also moved us away from simplistic approaches to managing flows of patients or resources. These may use the language of complexity, but their focus is on designing better systems rather than better adaptive systems, able to learn and evolve as conditions change.
We also learnt from UCCP that a clearer differentiation between targets for system behaviour and targets as performance monitoring needs to be made. This would involve a smaller number of strategic targets which take into account the interactions between subsystems and potentially limit the impact of competing objectives. The challenge for policy makers and managers is to create a portfolio of measures that address:
- the higher-level system behaviour (e.g. ‘improving emergency care’).
- the performance of individual subsystems (e.g. ‘A&E waiting time’).
- the evolving interdependencies in the system (e.g. the interactions between primary and secondary care).
A four-hour target is essentially a performance metric. Performance metrics may well be relevant for driving change but a single metric such as ‘speed’ will not achieve systemic change that takes into account all the interdependencies. While individuals should still be accountable for their performance, what is needed is a few relevant targets, formulated so their stakeholders have a sense of involvement in achieving them and which generate an awareness of the wider interdependencies.
Some limited moves in this direction were, in fact, made in 2011 in the English NHS. These began to embrace the idea of more holistic, multilevel goals, such as reducing rates of unplanned re-attendance at A&E within seven days, measuring time to treatment from a decision-making clinician, and the percentage of people who leave A&E without being seen. However, since then performance has seriously deteriorated. Many factors are involved, including a steep rise in delays in discharging patients from hospital, which prevent beds being freed up for those who need to be admitted and add to pressures on A&E departments.
The inexorable rise in the number of patients spending more than four hours in A&E is not only a matter of political concern or a technical problem in the health system – it can worsen clinical outcomes for individual patients. A&E is often described as the barometer of performance across the NHS. The mounting problems in recent years are not the result of a single cause but a combination of factors that reflect the current pressures on the health and social care systems. There needs to be a system-wide response, but an adaptive one. The legacy of Scotland’s UCCP experience may linger today – Scotland’s A&E system is reported to have weathered the current crisis slightly better than England.
The full paper is published in Organization Science.