The main aim of this paper is to reflect on how findings from the process evaluation have implications for the PARIHS framework and its development as framework that represents the implementation of evidence into practice. In discussing these issues, we draw on Helfich et al.’s
 critique of PARIHS and the three opportunities to refine the PARIHS framework they identify: being clearer about the interrelationships and dynamics between elements/sub elements would eventually help to identify more generalisable patterns; the need for a more explicit definition for successful implementation; and drawing on other conceptual frameworks and models to further elaborate on core PARIHS elements.
Interrelationships and dynamics
Consistent with the underlying tenant of PARIHS, the findings present a multi-faceted and dynamic story of implementation. Reflecting on the proposition that successful implementation is a function of evidence, context, and facilitation, the element of evidence requires further scrutiny. Previous research indicates that where there is strong research with clinical consensus it is more likely to be used in practice
. In this study, the evidence for shortening fasting times is scientifically robust and was generally acceptable, but these qualities were not sufficient to outweigh other factors. Changing an organization’s systems and processes to enable individualized fasting times requires more than robust and believable evidence. In this study, the presence of good quality evidence was a constant, however it proved to be a necessary (for example, to guide the development of new local policy/guidance), but not a sufficient condition for changing current practice and routines.
Nilson et al.
 discuss the potential role of habit theory in changing healthcare practitioners’ daily practice. They suggest that those who develop habitual behaviors are less likely to act on, or may avoid new information that challenges current practice—particularly in contexts that remain stable. These authors suggest that breaking ‘bad’ habits could be achieved by changing something in the context or by removing a person from that context. This is similar to the idea of creating a dissonance or awakening that current practice is not necessarily appropriate practice, which is a feature of other literatures including practice development
. In this study, the information in recommendations would not have been new to most practitioners; however, one explanation is that their practices were habituated towards maintaining traditional (non evidence-informed) ways of working, which then becomes their stable (and familiar) context. Equally, it could be suggested that organizations and systems become habituated, such that in the case of fasting for example, 12-hour fasts become so embedded and institutionalized that this standard becomes the acceptable norm. The link between behavior and context is made explicit by habit theory, and fits well with the relationship between evidence and context in PARIHS.
Helrich et al.
 suggest that the high-low continua within PARIHS could encourage a tendency towards linear relationships between elements. In this study, we had difficulty mapping the findings onto the high-low continua, which could be illustrative of two things. First, that the ideal position of the elements may vary from project to project, such that in some initiatives for example, it is not always necessary to have ‘high’ evidence and ‘high’ context alongside ‘high’ (appropriate) facilitation, which is the current theory of PARIHS. Second, because each implementation project will have a particular dynamic and multiple interconnections that may vary throughout its lifetime, it is not possible to plot this on a high-low continuum. Using the high-low continuum may be more helpful in providing a visual representation at diagnosis (i.e., a snapshot), but less useful in evaluating the process of implementation because this does not capture dynamism and patterns of interactions over time.
Findings from this study show that some factors were more important than others in providing the conditions for, and influencing the effect of implementation interventions, which included inter-professional team working (including communication), decision-making authority (mediated by inter-professional tensions), and organizational buy-in, issues evident in others’ work [e.g.,
[34, 35]. The interventions, including facilitation components, did not overcome the challenges presented by these factors to a sufficient level to affect fasting outcomes even though facilitators reported working with individuals and teams on a variety of activities. Therefore the main interactions in this study were between individuals and teams and context. Currently individuals are not explicitly part of the PARIHS framework but are embedded implicitly within evidence (individuals interact with evidence), context (individuals are part of context), and facilitation (facilitators work with individuals and teams). A case for making individuals more explicit within PARIHS is made below.
Successful Implementation (SI) has not been explicitly defined in previous PARIHS publications. Helfrich et al.
 suggest that successful implementation should take a logic model approach to linking the implementation strategy to outcomes, including the realization of an implementation plan, the achievement and maintenance of the targeted evidence-based practice, and the achievement and maintenance of patient or organizational outcomes. A logic model can (although does not have to) encourage a linear and deductive approach to the identification of inputs, processes and outcomes, which does not fit well with the underlying premise of PARIHS, which acknowledges dynamism and the potential for inductive explanation
. For this project, successful implementation was defined in broad terms as the use of the recommendations in practice with associated impact on practice and patient outcomes. A more helpful definition, which acknowledges implementation as a process might be: an orchestrated (active, planned) effort to make evidence-based changes by organizations, teams, and individuals that result in sustained improvements to care, patient outcomes, and service delivery, which are driven by and embedded in organizational strategy. This definition includes the need to pay attention to planning, the process, and evaluation of implementation activity in an iterative rather than staged approach. It is a definition that could apply equally to one-off implementation projects, such as this guideline implementation study, as to initiatives or programs that intend to create the conditions for sustained use of evidence and improvements in practice, such as the Collaborations for Leadership in Applied Health Research and Care in England and the Department for Veterans Affairs in the United States.
Elaborating on PARIHS elements
Helfrich et al. encourage better elaboration of PARIHS’ core elements. To date, the role, behavior and attributes of individuals have been implicit within the PARIHS framework. Findings from this study and evidence from others’ research and conceptualisations of evidence-based change and theory show the crucial role that individuals plays in the relative success of evidence-based change
[6, 7, 27, 37–39]. Over the last decade or so there has been a shift away from a focus on individuals (in the context of the evidence-based practice movements) to one that recognizes the role that context plays in implementation. Arguably this shift has resulted in an inattention to the study of individual factors, specifically the interplay between actors and the contexts in which they work and how that interplay influences change processes and impacts. Evidence suggests that many individual level factors including beliefs, attitudes, motivations, values, skills, competence, behavior, and characteristics may be influential
[40, 41]. These are consistent with findings in the current study where individual’s risk taking behavior, emotional response, skills and experience, enthusiasm, commitment, and decision-making authority were important factors in the intervention’s implementation and impact.
This is a timely opportunity to consider the inclusion of individuals as an explicit additional element to PARIHS. Individuals are currently implicitly embedded within PARIHS in that facilitators work with individuals, contexts include individuals, and individuals interact with evidence, however, the significance of individuals within implementation is perhaps currently under-represented. In other theories (e.g., Rogers Diffusions) and frameworks (e.g.,
[6, 8, 27], the individual is acknowledged as a core component. Therefore we suggest that individuals should be represented explicitly in the PARIHS framework so that successful implementation related to how individuals (at an individual, team, and organization level) interact with evidence, context, and how these interactions are facilitated towards successful processes and outcomes. Our findings do not suggest that evidential factors be displaced by individual factors; rather, the observable pattern of interaction between evidence, context, and facilitation was influenced by individual (patients and practitioners) and collective behavior such as over-estimating risk, caution, and team functioning. Therefore, it could be argued that individual’s behavior, intentions and actions should be part of a framework that seeks to explain successful implementation.
A new representation of the PARIHS framework, including an in-depth consideration of the implications of including a new component related to individuals’ or actors’ characteristics, behavior, actions, and how this impacts on the development of this middle range theory, will be the subject of a future publication. However, drawing on the components of other frameworks, theories, and evidence [e.g.,
[6–8, 27, 37, 40–42], we propose that the concept of the individual might incorporate: capability, capacity, motivation (including recognising a need for change), resilience, acceptability, feelings, knowledge and beliefs (including self efficacy) about the intervention/evidence, position and fit within the organization/social system, and approach to decision-making (e.g., experimentation, use of information). It is important to state that making individuals more explicit within the PARIHS framework does not mean they should be separated from the other elements, particularly context. The strength of the framework is in representing the elements’ interconnectivity, which should be preserved in this new representation. As such, this addition could be summarized in the following updated working proposition:
The successful implementation of evidence into practice is a planned facilitated process involving an interplay between individuals, evidence, and context to promote evidence-informed practice.
The additional element has the potential to strengthen the framework’s usefulness for planning and evaluating implementation efforts and is a reflection of current evidence and theory, particularly discussions about the social processes involved in knowledge mobilisation work
The process evaluation was designed to capture data across intervention sites, rather than conduct in-depth case studies or ethnographies in a few sites. In-depth data collection within purposively sampled sites may have provided more illuminating evidence about intervention implementation, particularly fidelity. We also did not have the capacity to undertake any observational work, and therefore have been reliant on self-reported data to reach the conclusions reported in this paper. Additionally, there are some voices missing from this account, including those of surgeons (who could have participated, but did not consent to) and senior operational managers (who were not included in our sampling strategy).
Although the study design and interventions were prospectively designed based on the core elements of PARIHS, our evaluation of the framework has been retrospective. The addition of the concept of the individual needs further consideration, elaboration and clarification.