- The importance of guideline development.
- The gap between knowledge and practice.
- Evidence for implementation strategies.
The past 25 years have seen widespread development of evidence-based practice in the form of systematic collation, interpretation and dissemination of research information. Despite the propagation of written guidance, the impact on practice has often been less than hoped. While some of this likely represents inertia in local units, much relates to the innate difficulty of translating the written word into reliable clinical processes; published guidance simply gathers dust on the shelf. Much effort has been concentrated on designing strategies to overcome these barriers to implementation, using systems to overcome human imperfection – making the right way, the easy way. These initiatives include attempts at behavioural change as well as the use of physical tools. In this review, the authors describe some of these methods, look at the evidence for their effectiveness and describe management systems for effecting change reliably.
Integrated care pathways (ICPs) incorporate clinical guidance into a multiprofessional medical record, using an event-sequenced preformatted document. Some guidance is written as a conventional guideline, some is made implicit in the form of checklists. ICPs cover an entire patient journey such as antenatal care and it could be said that the handheld maternity record is an early form of an ICP. Newer examples include elective gynaecological surgery and antenatal care for women with diabetes. Care bundles (CBs) are similar but define the care delivered at a particular point in the pathway, for instance elective caesarean birth, or the first hour of suspected sepsis (Appendix S1).
These tools were pioneered in the US in critical care settings to reduce the variation in the standard of care provided to patients with complex clinical problems. Their use was endorsed by the NHS Modernisation Agency and they are used in over 100 UK critical care units. The care of women with high-risk pregnancies is particularly suited to their use, as care is often multifaceted and multiprofessional (midwifery, obstetric, anaesthetic and medical professions).
Both tools consist of all elements of multiprofessional care, derived from national guidance whenever possible. The content is agreed by the local team, usually led by a steering group representative of all healthcare professions. While ICPs and CBs guide care to reduce inappropriate variation in practice, they should avoid being overly prescriptive, seeking to limit any reduction in clinical initiative while optimising patient choice.
ICPs and CBs are promoted as supporting good care in several ways:
The RCOG has provided guidance on obtaining consent since 2004. This includes guidance on the standards for the information required for valid consent in a number of common gynaecological and obstetric procedures. Preformatted consent integrates this advice within the Department of Health Consent Form 2 so that essential information about the procedure forms a part of the preprinted consent document (Appendix S2). Flexibility can be built-in with free-text boxes. These documents save time, reduce mundane tasks, enhance legibility for patients, and probably decrease errors of omission. As with ICPs, trainees can use the documents as an on-the-job learning tool.
These practical tools are a grouping of the equipment required for dealing with a clinical scenario, usually a rapid emergency, often together with guidance on a laminated card. Equipment bundles overcome the difficulty of rapidly assembling equipment for infrequent emergencies that might fall upon staff new to the unit. They act to save time and remove pressure on staff in an already stressed situation. The bundles are streamlined to include only relevant equipment in manageable numbers, usually grouped in transparent pouches – where possible with preprepared medications – all to enhance ease of use. Examples of these include postpartum haemorrhage and eclampsia boxes (Box 1; Figure 2), and cardiac-arrest caesarean packs in accident and emergency departments.
These tools aim to improve complex decision making. Examples include observation charts incorporating the patient at risk (PAR) score, used to identify those at risk of clinical deterioration. Within obstetrics, recent examples include the cardiotocography (CTG) interpretation sticker (Figure 3), thromboembolism assessment tools (Appendix S3) and the modified early warning score (MEWS) observation charts. Some groups have evaluated tools to aid women in their own decision making.
Bedside visual representations of tasks seek to prompt key actions particularly for rapid emergencies. Intrauterine fetal resuscitation (IUFR) for suspected severe fetal hypoxia requires five steps. A laminated chart outlining the critical steps for IUFR is attached to CTG monitors (Figure 4), where it can be immediately viewed in the event of acute deterioration in the fetal heart rate pattern. Again, they serve to remove pressure on staff in stressful situations, obviating any difficulty in recalling critical steps. Laminated icon charts also include illustrations of blood loss to improve the clinical estimation during PPH.
Eight plastic zip bags with contents grouped according to task, all held in a workman's toolbox labelled eclampsia. The bags contain the following:
Separate from any generic team cultures that can encourage positive working, there are some specific techniques for frontline staff that systemise team communication. Examples include surgical safety briefings and labour ward structured rounds.
A systematic review has shown that one in every 150 patients admitted to a hospital dies as a consequence of an adverse event. Two-thirds of deaths were associated with surgical care, and about half were avoidable. The Safer Patients Initiative theatre project focused on the development of a safety culture as well as reliable processes to reduce surgical site infection and adverse cardiac events. This has been adopted by the NHS Patient Safety First campaign, supported by checklists. The checklist has now been adapted for maternity theatre. Briefings are a whole-team exercise that prioritise list order, review the peri-operative care bundle for each patient, and plan contingencies for foreseeable complications. Departmental information such as equipment issues can also be communicated. Briefings also give an opportunity to enhance team working with staff introducing themselves and their roles, and being encouraged to question and comment. This approach seeks to break down hierarchy and improve communication, imprinting a safety culture that encourages every member of the team to speak freely if they recognise that an error has been or is about to be made. An airline pilot has described the events that led to the death of his wife during routine minor ENT surgery. He highlighted the human factors that contributed to her avoidable death, including crucial hesitation in communication by junior members of the healthcare team. He contrasted medical safety systems and team cultures with those in the air flight industry. Debriefs at the end of the list are focussed on quality improvement by collating glitches in theatre working, and in ensuring they are documented and acted upon. It is also an opportunity for giving praise for good team work.
Information exchange at shift end is a potential weakness in continuity of care, with crucial information being omitted, understated or misunderstood. Similar in principle to safety briefings, structured rounds are a formalised system for communicating care. Formatted handover sheets can be added to reduce errors. The RCOG Governance team has proposed the SHARING handover sheet (staff, high risk, awaiting theatre, recovery ward, inductions, NICU [neonatal intensive care unit], gynaecology), created in the Norfolk & Norwich University Hospitals NHS Trust. NICU consultants and anaesthetists can also attend the meeting and contribute greatly to timing of deliveries and other aspects of perinatal care. They too provide an opportunity to inculcate a positive team ethic; an open culture for discussing management plans.
The challenge for healthcare teams is to identify system problems, to design a change to improve the quality of provision, and to implement that change successfully while encouraging the clinical team to embrace it. Clinical leadership is a key component to the success of all three steps.
The science of improvement was developed for American industry by W. Edwards Deming during WWII. He then used the principles to help kick-start Japanese manufacturing in the postwar era. He is widely regarded as having transformed Japanese manufacturing, and the quality improvement systems that currently exist in companies such as Toyota and Honda are still largely based on his work.
The improvements in quality and productivity that followed did not rely on chance but rather they resulted from a system of implementation. Although caring for patients is not the same as building a car, the improvement systems developed by Deming can be applied in clinical settings. The aim of his methods was to accelerate and sustain improvement through systematic implementation that embeds reliable processes. Each step of the implementation process is designed, assessed and adapted. Deming's model combines two essential parts.
The first part asks four questions that enable change to be made:
The second part is the plan-do-study-act (PDSA) cycle that is used to test a system change within just one small area of the hospital – planning it, trying it, observing the impact of the change, revising the plan according to what is learned. Using this cycle of change process within a small section of an organisation, small changes can be rapidly tested and reviewed, to be continued or discarded depending on the effect. In this way, effective processes are developed in one area using numerous test cycles before being spread throughout the organisation.
All of this is underpinned by continuous measurement of compliance with key aspects of the care bundle (process measures). Deming termed this ‘statistical process control’. This monitoring system provides rapid feedback, which both informs the implementation and helps the clinical team to see the benefits of their efforts quickly. Ultimately, it is as important to measure clinical outcomes to be sure that the change results in meaningful improvement, as it is to make the change.
Many of these principles have been incorporated in the Improving Healthcare Initiative (IHI) that set out to save 100 000 lives in the USA through effective clinical leadership, structured team behaviours, practical tools, and run charts for rapid feedback. The IHI runs a website with examples of safety programmes including practical resources.
Temperature control during the peri-operative period reduces postoperative morbidity such as surgical site infection. This is supported by NICE guidance. If this is accepted as a clinical standard, then teams must not simply have knowledge of this. Rather than just stating at a departmental audit meeting that patients should not be hypothermic on admission to recovery, the focus should be on the theatre team being empowered to develop triggers to measure and control patient temperature, regardless of who was looking after them in theatre. The minimum aim being that 95% of patients have a core temperature of greater than 36°C on arrival at the recovery department. Data on patient temperature on arrival at the recovery department is then fed back to the relevant teams as run charts, initially weekly and later monthly (Figure 5).
Supporters of these methods write with much enthusiasm about their value, a personal quality that is often needed when introducing new systems into clinical units. More formal evidence of effectiveness was initially sparse and conflicting, but is growing steadily.
These have shown highly promising results. Studies of a gestational diabetes ICP have shown high levels of compliance with process measures. Care-goal sheets were associated with an improvement in clinical outcomes with reduction in catheter-related bloodstream infection rates. Theatre briefings have been associated with a reduction in risk-adjusted mortality in a dose-response manner compared with baseline and with contemporary controls. Perhaps most impressively, in North London, 14 targeted care bundles resulted in a large decrease in hospital standardised mortality rates (HSMRs) for those specific conditions while the HSMRs for non-targeted conditions only decreased slightly in the same time period. Overall, 255 deaths were avoided in 1 year; a 14.5% reduction. In the study period, the hospital had the lowest HSMR for acute Trusts in England. A similar study of over 7000 patients (using preintervention controls) looked at the effect of introducing a surgical checklist into six hospitals already thought to have high quality care. The proportion of patients who suffered a complication fell from approximately 27% to 15%.
In the maternity field, the employment of CTG analysis stickers (based on NICE guidance) was associated with a clinically and statistically significant 50% reduction in HIE and low Apgar scores. And in gynaecology, the introduction of a continence-referral ICP appeared to result in reduced average time for patients to be seen, have surgery, and be discharged. Not all projects have been successful, however. In one study, a MEWS chart appeared to be poor at predicting obstetric sepsis or death. With another example, a shoulder dystocia documentation proforma was associated with improved completeness but reduced accuracy of recorded information.
Evidence from RCTs has shown mixed results. Seventy-six patients randomised to a stroke ICP had worse functional recovery and quality of life compared with 76 patients allocated to conventional care. An ICP for older patients reduced readmission at 30 days but not at 60 days compared with conventional care. Overall, few trials have been conducted and most are of poor quality such that a systematic review of controlled trials has concluded that the available studies lack power to detect meaningful changes in health outcomes. Criticisms include excessive heterogeneity and lack of scientific rigour, which prevents meta-analysis of results. Cost analyses of relevant interventions, have also lacked rigour, and they usually have not undertaken full modelling across the entire health economy.
Unsurprisingly, there is emerging evidence that targeted training is accountable for part of the successful introduction of a new system of care or device.[21, 28] Mandatory departmental update days can play an important role in achieving compliance. These may well need to include leadership and generic teamwork training, particularly for rapid emergencies.
Clinicians using these methods champion their use and point to a test of change of process measures as evidence of their effectiveness but a clearer understanding of the true value for many implementation techniques, including cost-effectiveness, is needed. Random-allocation trials are not ideal for all methods and one challenge is to design research techniques for testing implementation strategies. The emerging discipline of implementation science is set to take on that task.
More computerised systems that analyse clinical data, flag alerts, and support decision making are needed. Early work has already been undertaken on decision support aids, showing greater speed and improved accuracy. Examples could include more individualised ICPs, computer-assisted handovers, real-time analysis of CTGs, automatic checking for drug interactions, time alarms for critical actions, automatic relay of results, and so on. Expansion of mobile data-networks could make these tools easier to access during direct patient interactions, across the hospital, and from more remote locations. Rules on data protection have to be followed, but the rapid growth in accessibility and functionality of systems, allied with a willingness to deploy these technologies, promises an exciting future nevertheless.
Many tools are developed by enthused teams in local units, often outwith traditional centres of innovation such as University Departments. They are usually not subjected to formal evaluation, and are generally not submitted to indexed journals. There is a need, therefore, for newer ways of disseminating these novel experiences and expertise. Institutional websites could assist by allowing project details and tools to be uploaded for universal access, enabling interactive discussion and creating a dynamic health community across the globe.
While these strategies attempt to limit the impact of the human condition on patient care, clinicians will need to understand and exercise their responsibility for ensuring that systems are not followed slavishly. In order to inculcate helpful behaviours, and to eliminate negative ones, training in safety culture and implementation strategies will need to start early in medical, nursing and midwifery training. Understanding how to create a positive team ethic, and to overcome difficulties in escalation, are also important goals, as is the formal development of effective clinical leadership.
Although disease processes can trace a fine line between deterioration and recovery, patients should not be expected to walk a tight-rope of variable compliance with national standards. No longer should the safety and quality of health care be dependent on the knowledge and motivation of individual clinicians. We now have an ever increasing opportunity to create and disseminate improvement systems that promote consistent high-quality care. While reliant on sophisticated leadership for implementation, these systems should seek to engender a truly team-based approach to care delivery. And for the whole team, the rewards include the satisfaction of seeing care delivered properly, and the resulting improvements in safety, clinical outcomes and patient experience that follow.
We thank Dr Jane Thurlow for allowing us to reproduce her IUFR icon chart in Figure 4.