Over the last decade in many of our health systems, we have witnessed an ever-increasing political and financial investment in the development of approaches to standardise healthcare. In the context of evidence-based practice, care pathways, guidelines and protocols have become common mechanisms for standardising service delivery and clinical decision-making. However, as Allen et al.1 point out in their report of the findings of a systematic review about the effectiveness of integrated care pathways (ICP), technologies such as these have an equivocal evidence base of effectiveness.
The review by Allen et al. was concerned with not only ascertaining whether ICP might be effective, but also by drawing on realistic evaluation,2 to consider in what circumstances might they be effective, for whom and in what contexts. While adopting a traditional systematic review approach to conducting the review, and therefore only including randomised controlled studies (RCTs) of effectiveness, their narrative summary raises some interesting observations about the development and implementation of care pathways in healthcare practice. Adopting a realistic evaluation approach to writing the narrative enabled further information about outcomes to be elicited. However, as it is difficult to capture the complex factors that influence care from the results of an RCT, the value of using traditional systematic review methods for these types of topics may require further consideration. Nevertheless, some of these issues raised by Allen and colleagues, including use of study designs appropriate to capture wider influences on decision-making, were addressed in two recently completed studies that evaluated the implementation and use of protocol-based care in the context of the UK's National Health Service3-6 and funded by the National Institute for Health Research's Service Delivery and Organisation Programme (http://www.sdo.nihr.ac.uk). While the primary aim was to understand nurses, midwives and health visitors contribution to protocol-based care, our studies also included the perceptions and practices of doctors and other allied professionals as well as the views of patients whose care had been informed by a tool to standardise care.
We used the term protocol-based care to encompass a range of standardised care approaches including ICPs, guidelines and algorithms.3,7 Therefore, our research was concerned with standardised care approaches in its broadest sense - which included ICPs, but also other sorts of tools. However whatever the tool, as Allen et al. rightly point out these sorts of standardisation mechanisms are complex interventions; there are various component parts that are difficult to isolate; therefore, it is difficult to tell whether the sum is greater than its parts.5 However, through qualitative realistic evaluation multi-method case study research we set out to establish how certain causal mechanisms (particular type of protocol-based care, and practices they prompt) operate in particular contexts (e.g. in particular clinical settings or circumstances) to create certain impacts or outcomes. If we pick up on some of the issues that Allen and colleagues raise in their summary of evidence, we can start to explain how recent empirical findings from studies that included the use of ICPs begins to shed some light on issues that emerged from a systematic review of the literature.4
One of the questions that Allen et al. raise is about the active ingredients of ICPs and how they interact in the reality of practice and/or when they are used for particular purposes. In a very broad sense we found that the use and impact of standardised care approaches is varied, and largely context and professionally specific (see Rycroft-Malone et al.4 for detailed description). As Allen et al.1 revealed ICPs have the potential to direct, coordinate and organise care. In our studies, commonly standardised care approaches were used as checklists and references. They prompted what needed to be done next, how, or (more commonly) as a check that everything had been done; in other words, the ICP informed care in retrospect. We observed examples of nurses and midwives referring to available standardised care approaches during interactions with patients; however, more commonly they would be referred to after a procedure or at the end of a shift. There was a concern raised by clinical staff that using these tools as checklists, while directing care, could also lead to a 'tick box mentality', and our observation of practice did uncover evidence of this.
Our decision-making ethnography revealed that while standardised care approaches are often developed to guide decision-making in fact interview data and observations showed that nurses and midwives continued to use their clinical judgement even when referring to or using them.6 As such they were believed to support, rather than remove the need for clinical judgement. The role of experience, the environment and philosophy of care, also emerged as importantly linked to making appropriate judgements and decisions. There was an expectation that senior nurses, because they were experienced, should already be aware of the information contained in the protocol-based care tools. For example, midwives at one of our study sites firmly believed that the ICP supported, but did not inform clinical judgement as 'they would practice like this anyway'. As such, it was felt that standardised care approaches were particularly useful and relevant for students, new or newly qualified staff, including doctors. Additionally, many of the standardised care approaches available to practitioners (including ICPs) only covered specific aspects of patient care and did not address issues such as information giving or communication, which patients raised as concerns when they were interviewed. Therefore, their utility for decision-making may be limited by the fact that they only have the potential to cover discrete aspects of the decisions that need to be made about the patient's journey, particularly those with comorbidities.
Standardised care approaches had facilitated and supported the extension of nurses' and midwives' roles, which had led to the development of new service initiatives such as nurse-led clinics and midwifery-led Birth Centres. Specifically in the sites where nurses and midwives were practising autonomously they perceived they were able to deliver more streamlined care because on a patient-by-patient basis they did not have to refer to, or follow up with doctors. In this sense the increasing use of such tools has had an underplayed, yet powerful effect on professional identity, boundaries and workload. An important issue which needs to be considered further is the impact of ICPs implemented for use among one specific group of clinicians on working relationships with other clinical colleagues. One unexpected outcome of our study was the detrimental effect the introduction of an ICP had on working relationships in some sites, resulting in groups of professionals defending their practice and the role of the ICP when care was criticised by colleagues. Allen et al.1 state that ICPs can be effective in improving documentation of treatment goals and communication - we would caution against this, as it was not our experience across all of our study sites. This demonstrates that what may work (i.e. be effective) in one context, may not work in another.
In terms of organisational impacts, standardisation of practice had resulted in a perceived standardisation of resources and so cost containment in some cases. For example, determining the sort of medicines, dressings and tests to be used within tools such as care pathways and protocols, had the potential to impact on what was used, how much and when (assuming that such tools were being used of course). In some cases there was a perception that this had been an outcome, and in one site actual reduction in costs had been noted. The potential to re-distribute roles and tasks between doctors and nurses with the introduction of extended role supported by standardised care approaches had resulted in a reduced workload for doctors in some cases. In another site, the introduction of a patient pathway was perceived to have impacted on containing, and in some cases reducing, patients' length of stay.
Finally, some important questions about development and implementation are raised by Allen and colleagues.1 Our study revealed that in most cases, while there had been considerable and active investment in the development of standardised care approaches, less attention had been given to their implementation. This had resulted in patchy and questionable sustainability of use. Such findings also highlight some critical implications for evaluation: if a pathway has not been fully or properly implemented; what is it that one would be evaluating?
Our full report gives details of what we found worked, for whom, how and in what circumstances.4,5 Essentially however, the impact of the use of standardised care approaches was determined by a number of interrelated factors including who was using them, how, for what reasons and in what contexts. These factors are difficult to disentangle. If, as our studies found, standardised care approaches can work differently, in different contexts with different people; what is the most appropriate way to make sense of, and unpack these complexities and heterogeneity? While ICPs may support a linear approach to a patient journey, human behaviour and the countless other complex factors that influence decisions about care suggest we may not always achieve the outcome(s) anticipated.
1Professor of Health Services & Implementation Research, Bangor University, Bangor and
2Professor of Evidence-Based Midwifery Practice, King's College, London, UK
References