A four-stage process for intervention description and guide development of a practice-based intervention: refining the Namaste Care intervention implementation specification for people with advanced dementia prior to a feasibility cluster randomised trial

Palliative and end-of-life care interventions in care homes for people living with and dying from dementia will always be multi-faceted and context sensitive. This requires interventions to be carefully developed, tested and implemented [1‐4]. However, experience shows that innovations can be recommended, adapted and implemented without this measured approach, with the flawed implementation of the Liverpool Care Pathway a cautionary tale for those working in palliative care and beyond [5]. An example of an innovative intervention that has had rapid uptake in care homes is Namaste Care, a multi-component approach to care for people with advanced dementia. Interventions in this field are important, as care for people with advanced dementia is usually provided in long term care settings, and these are likely to become the main place of death in the future [6]. Developed as a response to a lack of active care being offered to people with advanced dementia it has a philosophy based on person centred, holistic care [7, 8]. However, early findings on how and why it does (or does not) work are only just beginning to emerge [9].

Practitioner engagement and attitude and ‘fit’ of an intervention are known to have a major effect on adoption of innovation [10], and Namaste Care appears to have such an intuitive ‘fit’ with practitioners. Implementing evidence-based practice in nursing care homes is complex, with issues such as being on ‘common ground’, connecting with practice, and reconciling new practice with other priorities affecting change [4, 11]. Namaste Care resonates with practitioners because of its context sensitive, innovative, and effective approach to care for an overlooked resident group [12‐14]. Evidence from small scale, qualitative or uncontrolled studies indicates an effect on symptoms such as agitation [15, 16] and behavioural symptoms [17]. Qualitative studies identify that staff recognise positive features of the intervention such as providing sanctuary, connections and community, calmness and vision [9, 18‐20]. Problems implementing and sustaining the programme do, however, exist. Adjusting to the routines of Namaste Care can be difficult, and workforce turnover and management disruption endemic in long-term care can be barriers to both implementation and sustainability of the intervention [9, 21]. It is likely that the label ‘Namaste Care’ is applied to a wide variety of activity, and implemented in different ways [22]. The requirement for robust evaluation of effectiveness has been recognised, as there are no controlled, comparative trials of this intervention [9].

The challenge for any study of Namaste Care is that the intervention already exists in practice, albeit without sufficient evidence of effect. This is not a novel problem, health and social care practitioners are adept at identifying areas of care need and devising and implementing potential solutions that have little underpinning empirical evidence [23]. Healthcare practices, without evidence of effect, have been categorised in three ways: those that are known not to work, those where the evidence of effect is uncertain, and those in development or implemented without evidence [24]. Whilst the field of de-implementation is developing in order to assist the reduction or cessation of use of interventions known not to work, be unproven, or harmful [25], there is less attention paid to how best to test complex interventions that have already been organically implemented in some areas of practice, but where robust evidence is absent.

A particular challenge in a situation where a broadly defined intervention has already started to be implemented in practice is that of intervention description. A clearly specified intervention is required for a number of purposes including training, understanding fidelity, ascribing outcomes to the intervention, future replication, cost effective and appropriate implementation [26]. The Medical Research Council guidance on developing and testing a complex intervention focuses on intervention development (identifying the evidence base, identifying or developing theory, and modelling process and outcomes) and acknowledges that a common failing is inadequate description of the intervention [1]. The guidance requires a full description of the intervention, and an understanding of its components, so that it can be delivered during the evaluations, allowing for (and understanding) any flexibility and variation, and so that others can implement it outside the study. Understanding the components of an intervention is also important in understanding how the intervention works: what are the ‘active ingredients’ of an intervention and how do they exert their effect [27]?

Implementation scientists also focus on the importance of intervention description. It is recognised that an intervention can have interacting components: ‘core components’ (the essential and indispensable elements of the intervention) and an ‘adaptable periphery’ (adaptable elements, structures, and systems related to the intervention and organisation into which it is being implemented) [2, 28]. Intervention over specification should be avoided, to enable variation to fit different contexts, recognising the impossibility of describing every component of a complex intervention [29]. However, compared to knowledge on how to evaluate and implement interventions, there is relatively little guidance on how to develop and describe an intervention in a way that might maximise likely effectiveness [30, 31]. There is a gap in knowledge for those testing effectiveness of practitioner developed and implemented interventions. In these situations the intervention may have been differently understood, frequently adapted, and may differ from the original intent of those initiating the intervention [22]. Its theoretical underpinnings may be absent or not clearly articulated. It is unlikely that it has been carefully specified or adapted for a particular culture or context.

Potential ‘top down’ and ‘bottom up’ problems also exist. First, trial interventions can be challenging to incorporate in to day-to-day practice [32‐34]. In the care home situation there are particular issues with conducting research including factors such as time constraints, staff turnover and low education levels [4, 35, 36]. In specifying this intervention for research purposes, it was important that it remained relevant to practice, and did not take on features known to affect implementation. Second, interventions developed from practice do not always reflect the intervention encountered in practice. For example, the aim of the Liverpool Care Pathway was to take excellent hospice care principles and embed them in acute hospital practice. However, the intervention as specified (the paperwork developed), did not reflect the knowledge, skills and attitudes required for its safe and appropriate use [37].

The aim of this paper is to present a four-stage model to refine an existing Namaste Care intervention and develop an intervention description, guide and training package to support a feasibility trial of the Namaste Care intervention. The four stages include collation of existing materials, exploring comprehensibility with staff who do not have experience of the intervention, using nominal group techniques to refine and prioritise the intervention and its format, and refining with our patient and public involvement panel.

The overall aim of the study is to establish the feasibility of conducting a cluster randomised controlled trial of Namaste Care in a nursing care home context in the UK [38]. This is a phased research study involving the development of programme theories of how the Namaste Care intervention achieves particular outcomes and in which circumstances; developing an evidence-based Namaste Care intervention description and training package; and a feasibility cluster randomised controlled trial with embedded process and economic evaluations. Phase one (programme theory development) involved a realist review process [39]. This paper reports on phase two work as an exemplar of a method of developing and refining an intervention that has some existing practice presence, using SQUIRE 2.0 as the basis for reporting [40]. The research team included nurse academics, a research practitioner who had implemented Namaste Care, the trial manager, and patient and public involvement (PPI) representatives.

We planned four iterative stages to this phase of the study, with co-design of the intervention description with nursing care home staff and family carers central to the methods chosen (see Table 1).

Lancaster University Faculty of Health and Medicine Research Ethics Committee granted approval for this phase of the study (17 Nov 2016/FHMREC16028).

The four-stage process for describing and developing an existing practice-based intervention prior to further testing and implementation appears to have utility. We were able to describe succinctly the Namaste Care intervention in a 16-page A4 booklet in a way acceptable to the nursing care home context. This was supplemented by four A4 infographics summarising the main elements of the intervention in an easy to read and user-friendly format. The guide is colour coded (to match the infographics) and uses flow charts and graphics to facilitate the reader’s understanding of and engagement with, the materials. Training materials follow the same style and format. The guide specifies the boundaries of the intervention, and guides implementation, whilst retaining the flexibility both inherent in Namaste Care, and required in a pragmatic feasibility trial.

Intervention development is central to the Medical Research Council guidance on studying complex interventions—researchers are advised to consider whether they are clear about what they are trying to do, that the theoretical basis of the intervention has been used systematically to develop the intervention, and that it can be described fully [26, 54]. The Medical Research Council guidance is frequently used to optimise intervention development, but other frameworks such as intervention mapping, MOST (Multiphase Optimisation Strategy), the six steps in quality intervention development (6SQuID), and intervention modelling are also available [30, 55‐58]. Although they use staged approaches which have similar features to the approach reported in our study (e.g. working with key stakeholders, involvement of patients and the public), these typically still are only used in novel intervention development [59]. The four-stage process used in this study to describe the intervention for research use may have utility for other researchers faced with similar challenges. These four stages are conceptually congruent with many frameworks for intervention development or implementation. For example, the Knowledge to Action Framework emphasises that resources should be produced in a collaborative fashion with end users and other interested parties [60], and this involvement was a key feature of the four step process described here. We propose that this four-stage process could be integrated as an additional component to existing frameworks for intervention development or implementation where there is a requirement for an existing intervention to be described, developed or refined. This generic process is presented in Fig. 2.

This four-stage process could, for example, be implemented in the development element of the Medical Research Council guidance for complex interventions [1], or the optimisation stage of MOST [57].

The four-stage process described here may have utility for researchers testing the effect of existing interventions, or where they need to adapt an existing intervention in a culturally or context specific way. Careful development and specification of an intuitively helpful intervention both enables an understanding of fidelity within the subsequent trial, but also facilitates future implementation, or indeed de-implementation. Future research could test these steps with other interventions, and report on its utility and development both in process evaluations of trials, in implementation studies, and in conjunction with other frameworks.