Background
-
Demystify some of the jargon through defining the nature and role of frameworks, models and theories
-
Describe and compare commonly used theoretical approaches, and how they are applied in practice
-
Suggest how to select a theoretical approach based on purpose, scope and context of the individual project
Case study part 1 (experience of author AM): I am a physician and I plan to implement a delirium prevention program. I have some implementation experience and know that it won’t be easy. I have heard about implementation science, so I hope there may be tools to help me.I understand a bit about Knowledge to Action (KTA) to guide my planning. I have strong evidence of effectiveness and cost-effectiveness [knowledge creation & synthesis], and there are established clinical practice guidelines [knowledge tools/products]. There is an effective model to implement delirium prevention developed in the USA (http://www.hospitalelderlifeprogram.org), but it used skilled geriatric nurses and large numbers of trained volunteers, which is not feasible in my hospital. None of the strategies in the guidelines are “hard” but they just don’t seem to get done consistently. I need to find out from staff and patients why this is the case, and then try to find ways to support them. Perhaps they need more education or reminders, or maybe we can reallocate the tasks to make it easier? Or are there strategies I am not familiar with? Whatever I do, I want to measure better care in some way to keep my boss happy and the staff interested. And my previous projects have tended to fizzle out over time… KTA gives me part of a plan but I need some more tools to know how to take the next steps.
Main text
Defining frameworks, models and theories
-
A framework lists the basic structure and components underlying a system or concept. Examples of typical frameworks are the Consolidated Framework for Implementation Research (CFIR) [6], the Theoretical Domains Framework (TDF) [7, 8], RE-AIM [8‐10] and Promoting Action on Research Implementation in Health Services (PARIHS) [9, 10].
-
A model is a simplified representation of a system or concept with specified assumptions. An example of a model is the Knowledge to Action (KTA) cycle [11].
-
A theory may be explanatory or predictive, and underpins hypotheses and assumptions about how implementation activities should occur. An example of a theory is the Normalization Process Theory (NPT) [12].
Knowledge to Action [11] | |
Purpose (as described by authors) A framework to conceptualise the process of knowledge translation which integrates the roles of knowledge creation and knowledge application. Provide conceptual clarity by offering a framework to elucidate the key elements of the knowledge translation process | |
Brief description: This approach provides an overview to help guide and understand how knowledge is created and synthesised, and tools (like clinical guidelines) are developed, then how these tools are applied in clinical settings through tailoring and adaptation, implementation, monitoring and sustaining. Assumes that action plans will be realised (underpinned by assumption that actions are rational). Takes a systems approach – recognises that knowledge producers and users are situated within a larger social system | |
How developed: Developed by reviewing literature of > 30 planned action theories, identified common elements. Added to planned action model a knowledge creation process and labelled the combined models the knowledge to action cycle. | |
Changes/developments over time: No | |
Ease of use: clear and easy to understand, intuitive. No specific guidance on how to do each step of the action cycle but provides some guidance on important elements to consider. | |
Additional resources: no specific resources currently available on how to action each step of cycle | |
Purpose (as described by authors): An integrative theoretical framework, developed for cross-disciplinary implementation and other behaviour change research to assess implementation and other behavioural problems and to inform intervention design. | |
Brief description: provides a holistic list of factors that influence behaviour – application of TDF can give researcher confidence that factors influencing an individual’s behaviour will be identified, which in turn can identify factors that need to be addressed in order for behaviour change to occur (i.e. can be used to inform behaviour change strategy development/selection). Can be used in conjunction with Behaviour Change Wheel to develop and deliver behaviour change strategy | |
How developed: through an expert consensus process and synthesis of 33 theories and 128 key theoretical constructs related to behaviour change. | |
Changes/developments over time: Validity was investigated by behavioural experts sorting theoretical constructs using closed and open sort tasks. Validation study demonstrated good support for the basic structure of the TDF and led to refinements, leading to publication of new iteration of framework in 2012 | |
Ease of use: Quite straightforward to apply, can be time consuming to use for analysis – potential to overwhelm novice researcher given the 14 domains and 84 component constructs. COM-B and Behaviour Change Wheel work together with TDF. | |
Purpose (as described by authors): Originally developed as a framework to guide consistent reporting of evaluations regarding the public health impact of health promotion interventions, thereby providing a framework for determining what programs are worth sustained investment and for identifying those that work in real-world environments. | |
Brief description: Reporting checklist for public health interventions (what patient groups are receiving intervention, have patient outcomes changed, what health professionals/ health professional groups are providing intervention, are they delivering intervention as intended, will the program be sustained in the long term) to evaluate real world impact. Can be used when designing or evaluating a public health intervention. | |
How developed: Through inductive thinking building on results of previous research | |
Ease of use: Easy, interventions can be rated on the five dimensions, providing a score. Some of the reporting points (in particular Reach and Adoption) are not being interpreted and reported as developers intended | |
Additional resources: dedicated website with online tools, examples [33] | |
Consolidated Framework for Implementation Research [6] | |
Purpose (as described by authors): Framework to promote implementation theory development and verification about what works, where and why. | |
Brief description: list of factors (5 domains and 37 constructs) that can influence an implementation project, can be used in planning or in evaluation stages (does not guide how to implement). Research focus in contrast to doing/practitioner focus | |
How developed: Published theories which sought to facilitate translation of research findings into practice in the healthcare sector were reviewed. Team identified constructs that had evidence that they influenced implementation and could be measured. Some constructs were streamlined and combined, whereas other constructs were separated and delineated. | |
Changes/developments over time: No | |
Ease of use: Clear, but may be difficult to digest language if new to area of implementation science | |
Additional resources: dedicated website that provides examples, templates and tools to assist in developing and evaluating implementation projects, collecting and analysing data [28] | |
Conceptual model of evidence-based practice implementation in public service sectors [15] | |
Purpose (as described by authors): A multi-level, four phase model of the implementation process that can be used in public service sectors. | |
Brief description: Conceptual model of factors that can influence implementation in the unique context of public sector services (focus on role of service delivery organisations and the services in which they operate) at each of the 4 implementation stages: Exploration, Adoption/Preparation, Implementation, Sustainment (EPIS). Explicitly recognises that different variables play crucial roles at different points in the implementation process. Does not provide guidance on how to move through different stages of implementation. | |
How developed: based on literature and authors’ experience of public service sectors, funded by the National Institute of Mental Health | |
Changes/developments over time: No | |
Ease of use: Little clarity on how to operationalise different factors, potential to be confusing for those unfamiliar with implementation | |
Additional resources: California Evidence-Based Clearinghouse for Child Welfare have developed webinars regarding use of EPIS framework. Freely available from http://www.cebc4cw.org/implementing-programs/tools/epis/ | |
Conceptual model of implementation research [19] | |
Purpose (as described by authors) a heuristic skeleton model for the study of implementation processes in mental health services, identifying the implications for research and training. | |
Brief description: Guides how implementation research can be organised, how it fits/aligns with evidence-based practices. May be useful for complete novice who needs clarity between clinical interventions, implementation strategies, and working through how to measure clinical and implementation effectiveness. Various theories can be placed upon the model to help explain aspects of the broader phenomena. | |
How developed: drawn from 3 extant frameworks: stage pipeline model, multi-level models of change and models of health service use. | |
Changes/developments over time: No | |
Ease of use: Clear and easy to understand | |
Additional resources: No | |
Implementation effectiveness model [16] | |
Purpose (as described by authors): an integrative model to capture and clarify the multidetermined, multilevel phenomenon of innovation implementation | |
Brief description: A list of constructs that can influence implementation effectiveness, based on the premise that implementation effectiveness is a function of an organisation’s climate for implementing a given innovation and the targeted organisational members’ perceptions of the fit of the innovation to their values. Does not provide specific guidance for how to implement, was not designed specifically for the context of health care. Likely to be most useful for projects with a clear organisational approach. | |
How developed: from authors’ personal experience with reference to literature | |
Changes/developments over time: No | |
Ease of use: Main manuscript very wordy (text-based). Concepts are clear. | |
Additional resources: No | |
Purpose (as described by authors): Organisational or conceptual framework to help explain and predict successful implementation of evidence into practice and to understand the complexities involved. | |
Brief description: Conceptualises how evidence can be successfully implemented in health care settings using the process of facilitation. Underlying premise is that facilitation will enable people to apply evidence in their local setting, which is situated within a broader organisational and societal context. Framework strives to capture the complexities involved in implementation, so most useful in more complex projects. | |
How developed: from authors’ experience working as facilitators and researchers on quality improvement activities and health service research projects. | |
Changes/developments over time: Has had several iterations since first publication in 1998 in response to findings from empirical testing. Revised to integrated or i-PARIHS framework in 2015 [10] | |
Ease of use: Does not operationalise its constructs, so may be difficult for novice to understand and apply, particularly when not being supported by expert facilitator. Facilitator’s toolkit easy to apply to conduct pre- and post-implementation evaluation. For people experienced in implementation, framework provides guidance on all of the things to consider when implementation is complex. | |
Additional resources: Facilitator’s Toolkit in book associated with 2015 iteration of PARIHS guides user through how to assess, facilitate and evaluate [32] | |
Interactive Systems Framework [13] | |
Purpose (as described by authors): Heuristic to help clarify the issues related to how to move what is known about prevention (particularly prevention of youth violence and child maltreatment) into more widespread use. | |
Brief description: Framework regarding translating findings from prevention research to clinical practice. The framework comprises three systems: the Innovation Synthesis and Translation System (which distils information about innovations and translates it into user-friendly formats); the Innovation Support System (which provides training, technical assistance or other support to users in the field); and the Innovation Delivery System (which implements innovations in the world of practice). | |
How developed: Collaborative development of the framework by Division of Violence Protection staff members, university faculty and graduate students, with input from practitioners, researchers, and funders. | |
Changes/developments over time: No | |
Ease of use: Easy to understand, no clear guidance available regarding how to apply framework | |
Additional resources: No | |
Normalization Process Model, Normalization Process Theory (NPT) [12] | |
Purpose (as described by authors): provides a conceptual framework for understanding and evaluating the processes by which new health technologies and other complex interventions are routinely operationalized in everyday work, and sustained in practice. | |
Brief description: NPT is an Action Theory, which means that it is concerned with explaining what people do rather than their attitudes or beliefs. Proposes that for successful sustained use: individuals & groups must work collectively to implement intervention; work of implementation occurs via 4 particular processes; continuous investment carrying forward in space and time required. Can be helpful to understand and evaluate how new health technologies/complex interventions are routinely operationalised sustained in practice. Not designed to guide implementation. | |
How developed: in iterations, based on experiences of authors. Initially, developers mapped the elements of embedding processes and developed the concept of normalization. Next a robust applied theoretical model of Collective Action was produced, and applied to trials, government processes and healthcare systems. The final stage focused on building a middle-range theory that explains how material practices become routinely embedded in their social contexts. | |
Changes/developments over time: Through its focus on being a theory, the authors continually refine and test NPT to ensure its validity. More recently, NPT has been extended towards a more general theory of implementation. | |
Ease of use: easy to apply with use of specifically developed resources | |
Additional resources: dedicated website with toolkit, examples. Interactive toolkit can be used to plan project or analyse data [29]. |
Commonly used theoretical approaches
Case study part 2 (experience of author AM): So it is clear that I will need to adapt principles and protocols from successful programs in the USA to my local context. But how do I know what the context is? Top picks on Google scholar for “context assessment implementation science” seem to be Consolidated Framework for Implementation Research (CFIR) and Promoting Action on Research Implementation in Health Services (PARIHS). Both have nice guides that suggest useful questions to ask. There seems to be quite a lot of overlap, although I am drawn to PARIHS because from my experience I know that someone will need to spend time on the ward and build trust before we start to ask questions and introduce change. I suspect this ‘facilitator’ will be a critical role for the complex intervention because there are several behaviours to change.When I look up “behaviour change implementation science”, the Theoretical Domains Framework (TDF) dominates. Like CFIR and PARIHS, there are a lot of elements, but I can see that they would be helpful for planning or analysing surveys and interviews with patients and staff to clarify what motivates, helps and hinders them. I do feel worried about how I will collect and analyse so much data across the several different groups involved in my project.And once we have a thorough understanding of the context, staff and patients, how will I select strategies?And if it does work, how long will it will take until the “new” becomes “normal” so that I can move on to the next problem? My colleague tells me that Normalization Process Theory (NPT) is a useful way to think about this, and I am impressed with the Normalisation of Complex Interventions-Measure Development (NoMAD) tool I find on their website; I can see how I could adapt it to find out whether staff feel the changes are embedded.So where do I go from here? Do I frame the whole project with KTA, assess context with CFIR, assess the patient and staff views with TDF, adopt facilitation as the central element from PARIHS, and then look at how well it has gone using NPT? Am I being thorough or theoretically sloppy? They all look sensible, but I am not sure how to use any of them and I am worried it is going to add a whole lot of work to a complicated project.
How the theoretical approaches have been used
How do you select the theoretical approach for your implementation project?
Who are you working with?
When in the process are you going to use theory?
Why are you applying a theory?
How will you collect data?
What resources are available?
Other questions to ask to help in choosing
-
Does the theoretical approach have particular ‘face validity’ for the implementation project? For example, people interested in facilitation may recognise PARIHS as particularly relevant, or a project aiming to address motivations for behaviour change may lend itself to TDF.
-
Does the theoretical approach draw your attention to aspects of implementation that you may have otherwise neglected? For example, implementing a large scale public health intervention may have varied success due to challenges meeting the most vulnerable populations, or providing the intervention as intended across a range of sites—features that can be captured effectively by RE-AIM.
-
What theoretical approach(es) have studies in your topic area used? This can be helpful in providing worked examples of how particular theoretical approaches have been applied to add understanding to a project.
How can theoretical approaches be used during your implementation project?
Case study part 3 (experience of author AM): I choose PARIHS because I need to work with a whole range of staff, with different views and roles within this complex intervention, and I can see the need for facilitation. I am mostly interested in something to guide the “doing” and as my colleague and I begin to work with the first ward, we find using a lens of implementation science helps us to understand more about what is and isn’t working, so that instead of getting frustrated we can reflect more objectively and search for flexible solutions. On the first ward we also use the framework for reflection and for guiding a simple evaluation [34], and this provides useful structure for planning and adapting our approach on the next ward where the people and context differ in important ways. This then gives us confidence to start training other facilitators to “read” the local teams and context as we conduct a large funded trial across several hospitals [35], providing evidence of transferability that will be critical for spread to improve outcomes at scale. We are able to capture more consistent data about the different wards we are working on, which help us understand inconsistencies in our results, and identify the most important factors that predict which wards can implement this program successfully to help us target scarce resources. We are also able to collect data about the facilitation process, and how facilitators learn this role. We find as we become more familiar with PARIHS it becomes more useful for planning, doing and evaluating our improvements.But we are also involved in projects that this framework is too complicated to use in. Our experience makes us more confident to look for and try out other theoretical approaches that we think will suit those projects better, and help us achieve – and importantly understand – our outcomes.