Implementation of evidence-based health promotion and disease prevention interventions: theoretical and practical implications of the concept of transferability for decision-making and the transfer process

The relevance of incorporating evidence of the effectiveness of health promotion and prevention interventions into policy and practice has been acknowledged for decades [1, 2]. In 2015, the German Act to Strengthen Health Promotion and Prevention (Prevention Act [PrävG]) entered into force, assigning far-reaching new responsibilities to the statutory health insurances (Gesetzliche Krankenversicherung [GKV]) for the prevention and reduction of disease risks, the promotion of self-determined health behaviour, and the reduction of social and gender-related health inequalities. The act requires basing interventions on the best available scientific evidence and evaluating their effectiveness in local settings [3, 4]. Evidence-based health promotion and disease prevention are regarded as relevant to reduce supply deficits, such as the waste of limited financial resources [5‐7], the disorientation and demotivation of actors and target populations in the absence of intervention effectiveness, and the underestimation of the success of policy measures caused by the implementation of interventions with weak or unknown effectiveness [5]. Furthermore, interventions should not lead to undesirable side effects in predominantly healthy people. The Robert Koch Institute (RKI) [7] demands the selection of health promotion and disease prevention interventions that can actually improve the health of the target population: They must be suitable and effective in the practical implementation on site.

This requirement assumes that the selected interventions are transferable to the target context. Transferability addresses the extent to which an intervention whose effectiveness was established in a primary context is effective in a target context [8‐11]. Interventions in health promotion and disease prevention are more or less complex and context dependent. Thus, complexity may refer to the intervention itself (such as components, flexibility, and outcomes), to the intervention context (such as population behaviour or organisational levels targeted), and to the interaction between the two [12‐14]. That means, even if an intervention is based on sound scientific evidence, “the intervention itself may be applicable, but it may generate other effects than those seen in the primary intervention” [15, p. 1]. The PrävG focuses on health promotion and disease-prevention interventions in the everyday living environments of the people, i.e., in specific settings such as the community, the school, or the workplace [3, 4]. This focus underscores the importance of anticipating and evaluating transferability of interventions to these specific contexts (see Infobox 1 for an example).

Two essential questions need to be considered: First, with regard to the primary context, what kind of evidence and information is available for decision-makers in order to anticipate transferability? Second, how can transferability be assessed by decision-makers before and during the transfer process? In recent years, few structured approaches have been developed. For example, the first question is addressed by Munthe-Kaas et al. [16], who developed the TRANSFER approach to assist review authors in considering transferability to the context specified in the systematic review in collaboration with stakeholders. Cambon et al. [10] developed a checklist tool (ASTAIRE) with transferability criteria for health promotion interventions. It addresses both questions in that it can be used either to report evidence or to assess transferability to a target context. However, it does not provide guidance for the transfer process itself. Burchett et al. [17] found in their methodological study of the usability and usefulness of assessment tools and frameworks that checklist tools (including transferability tools) have been shown to be insufficient because they fail to address mechanisms on how the intervention works and its potential interactions with the context; these authors express the need to incorporate information on assessment criteria with information on mechanisms of action. Other approaches focus on implementation or evaluation, without systematically considering the concept of transferability and transferability criteria ([13, 18‐20]; see Table 1).

To fill this conceptual gap, Schloemer and Schröder-Bäck [21] developed a theoretical and methodological approach for the assessment of transferability in a qualitative systematic review. In this review, articles on transferability of health interventions were systematically searched, ranked for their quality, and further analysed with a thematic synthesis. This form of an interpretive synthesis shares similarities with approaches from grounded theory and meta-ethnography [22]. Two interrelated models were built through the structured, inductive method on the basis of an identification and systemisation of criteria for transferability: The Population–Intervention–Environment–Transfer Models of Transferability (PIET-T) explain the theoretical concept of transferability and provide guidance to integrate the use of transferability criteria in transfer processes. Based on these models, we aim to discuss theoretical and practical implications of the concept of transferability for health promotion and disease prevention. In the following, we introduce the theoretical understanding of transferability and propose an approach for the integration of information about characteristics of primary and target contexts in transfer processes with consideration of potential practical challenges.

A transfer decision should always be based on the need for an intervention [7, 8, 33]. This requires information from the target context. The need can result from the prevalence of a health problem, its public health burden, health priorities and goals, priorities of decision-makers, and the perceived needs of the target population and local actors [8, 33]. This also includes the consideration of existing synergistic or antagonistic interventions, which might influence the need for and transferability of a new intervention [8, 9, 10]. The German RKI [7] calls for the selection of an effective intervention based on the needs determined on site for the target population. Barriers are also reported here, since the relevant data are predominantly available only at the federal or state level. Attempts are therefore increasingly being made to create small-scale health reports that depict the specific health problems in individual municipalities or city districts [7].

In Germany, there is a heterogeneous picture of actors in the field of health promotion and disease prevention. More coordination of measures is recommended, particularly with regard to limited financial resources. Therefore, overarching health goals are formulated, for example in the PrävG [3, 7]. This offers the potential to anticipate and evaluate transferability of interventions to reach these goals in target contexts. The involvement of coordination players (such as networks, partners, and local actors) and the needs and views of the target population in a context may be necessary at an early stage and in all phases of the process [15, 24, 25]. This may facilitate a better understanding of the health problem and potential solutions. Depending on the conditions in the target context, it might be important to clarify who is involved in the determination of specific goals and relevant outcomes for the target population and the subsequent selection of and decision for an intervention (for example, cooperation between researchers, policy makers, local actors, and representatives of the target population) [21]. This also raises the question of the level of participation of relevant stakeholders [7, 34]. These aspects described above are important criteria for the transferability of an intervention, as they might be essential for stakeholders’ acceptance of and involvement in the intervention and, thus, for its sustainability [15, 21, 26, 34]. A helpful instrument from the beginning might be the Criteria for Good Practice in Health Promotion addressing social determinants [35].

The identification of one or more effective interventions with regard to the identified needs and goals is in itself a comprehensive step and may be dependent on many factors, such as available data, resources, knowledge, and experiences of responsible persons. In Germany, the decision to implement health promotion and disease prevention interventions is mostly in the hands of actors from practice or politics. These actors are often not yet familiar with the concept of evidence-based practice [32]. Furthermore, given the variety of prevention and health promotion programmes and interventions available, it can be difficult for a community or organisation to make a choice [7]. A list of relevant databases to use for the search for interventions is available from the BZgA [32].

From the perspective of transferability, two main criteria of the evidence base are important for understanding whether an intervention is appropriate to influence the health problem of the target population: the quality and the usefulness of the primary evidence. Ideally, information on the quality of the evidence is already available in the databases. Based on the consensus statement of the American Society of Prevention Research [29, 36], the BZgA defines standards of evidence for Germany. Accordingly, an intervention is considered to be sufficiently evidence-based if it meets the criteria of efficacy and effectiveness, in particular with regard to (a) a detailed description of the intervention and its pathways in a certain population and a certain context, (b) its effectiveness and safety under everyday conditions, and (c) insights gained from process evaluations for a successful implementation [32]. In addition to collecting information on the quality of the evidence, an equally important and pragmatic approach is assessing the usefulness of the evidence for the target context (see Infobox 2) [21]. If not all criteria of quality and usefulness are met, an intervention can still be transferable. Depending on conditions in the target context, it should be weighed how much loss of quality and usefulness of the evidence can be accepted. In some cases, it may be ethically inacceptable to wait until high-quality evidence from a similar context becomes available [11]—for example, if only the evidence of efficacy is available for an intervention, but the usefulness for the target context is acceptable. On the other hand, if no high-quality evidence is available, a well-described intervention from a practical project evaluated in a similar context might be useful. The selection of one or more potentially suitable interventions builds the basis for the next steps.

After an analysis of the need for and identification of an effective intervention, the initial assessment of transferability should compare relevant criteria in both contexts with regard to the population (P) and environment (E) to anticipate conditions that are likely to influence the outcome(s) of the intervention (I) in the target context for decision-making and the planning of the intervention transfer (T) [21] (see Fig. 2). In other words, the assessment should assist decision-makers in systematically analysing what information is available and how to use this knowledge for the transfer decision, intervention implementation, and evaluation of transferability. An explanation and order of criteria use as well as guiding questions for an initial assessment and process planning are available elsewhere [21].

For the transfer decision, it is relevant to discuss whether the relationships between the intervention and the outcome(s) are expected to differ for the population (or subgroups) in the primary and target contexts [21, 23]. The focus should lie on consideration of the conditions under which the intervention is expected to be successful or not in the target context: Is it potentially transferable as intended, is it potentially transferable with adaptations, or is it not transferable at all?

The effort for the anticipation of transferability and decision-making can depend on various aspects, such as the existing data; the complexity of the intervention; and the number, networking, values, and expertise of decision-makers and interventionists or those implementing/delivering the intervention. Furthermore, the number of criteria relevant to analyse might differ with regard to the health problem, the intervention, the compared contexts, and the transfer level. Since basic information from the target context and the primary evidence is ideally already available from the first steps, a possible pragmatic approach is to select criteria that are relevant to assess for decision-making on transferability. Therefore, it may be helpful to note why these criteria are important to consider. Subsequently, it can be determined and documented whether the primary and target contexts differ with regard to the selected criteria, and whether and why this difference is relevant for reaching the expected outcome(s) in the target context. In other words, an emphasis should be placed on criteria that may potentially promote or hinder transferability of the intervention. The aim is to differentiate and select potentially influencing criteria for the further process [21].

For this step, it may be necessary to identify and document what information about criteria is missing and whether it is important to collect additional data for decision-making, as the relevance of detail and validity of information from both contexts may differ among criteria. There may be criteria relevant for the target context that cannot be compared. These should nevertheless be taken into account because data could be collected in the course of the further process, if necessary. In general, conclusions from reviewing evidence for transferability may be limited due to poor description of an intervention [31], a lack of adequate process and contextual information [11, 27], or unknown suitability of an intervention for different contexts [27, 31]. It may be an option to contact intervention developers for collecting information and for advice on the further process, as knowledge exchange between actors in the target context and intervention experts may foster mutual learning and transferability [15, 24, 25, 31, 37].

The validity of information from the target context can be enhanced by gathering data—for example, from epidemiologic and demographic studies in the target context—and by conducting qualitative research with stakeholders [11, 28]. As previously discussed, participatory approaches might be promising, such as in the form of discussions with stakeholders in workshops about the criteria. This may help to uncover facilitators and barriers to the transferability of the intervention, to assess their relevance, and to further specify the relevant potential influences on the outcome(s). Local actors have practical knowledge of the target context. A joint analysis of relevant facilitators and barriers to transferability provides a structured approach to identify the need for adaptations of the intervention or target context. Addressing barriers and needs of professionals before implementing an intervention and enabling adaptations to improve its usability may enhance effectiveness of the intervention [38]. For example, it may be possible to overcome identified barriers with regard to the environment through suitable measures, such as providing resources for intervention delivery. Discussion and identification of an intervention’s core elements and a context-relevant form of the intervention may enable formulation of potentially transferable elements as well as elements that should be adapted in order to facilitate implementation while considering intervention fidelity, and are relevant to planning the evaluation of the intervention’s effects [24, 26, 37]. Such discussions can also reveal whether the intervention is not suitable for the target context. Then the information gathered can be used as a basis for identifying an alternative intervention or developing a new intervention. In the event of subsequent structured adaptations of the intervention, the expanded Framework for Reporting Adaptations and Modifications to Evidence-Based Interventions (FRAME) may be a useful additional tool [39].

When a decision to transfer the intervention has been made, the data from the initial assessment of relevant criteria is a basis for planning the intervention transfer, including implementation preparation, the implementation itself, and evaluation. The transfer criteria (T) are intended to help identify important aspects for transferability with regard to the steps of implementation and evaluation in the target context under consideration of strategies and processes in the primary context. These aspects might be incorporated in an implementation guide or material. The data from the previous steps builds a basis to specify hypotheses and/or questions for the evaluation. Furthermore, transferability criteria relevant for successful implementation and intervention effectiveness can be operationalised to be included in the evaluation of the process (e.g., questions for qualitative analysis) and outcome(s) (e.g., potential moderating factors for quantitative analysis). Piloting the intervention helps to refine hypotheses and implementation and evaluation plans, to pilot adaptations, and to identify the need for further adaptations during the transfer process. Finally, transferability can be explained by measuring whether and under which conditions the intervention is effective in the target context [21].

The choice of evaluation design in the target context can depend on various factors, e.g., ethical, political, or financial reasons [23]. An important aspect is the level of transfer. For health promotion and disease prevention, various authors suggest using phase models as an orientation. The underlying assumption is a sequential fashion of efficacy research, effectiveness research under real-world conditions, and dissemination on a larger scale [2, 23, 30, 31, 40]. Considerations about the evaluation design should ideally be based on the existing primary evidence as well as on changes to the intervention through adaptation in the target context [23, 32]. The BZgA proposes a model with recommendations for Germany: For the transfer of a well-described but hitherto not evaluated intervention from a practical project, an evaluation with a pilot study is recommended; for the transfer of an intervention with plausible prerequisites for effectiveness (“BZgA promising practice”), an effectiveness study with a process evaluation is recommended; and for the transfer of an intervention with causal proof of effectiveness, monitoring during transfer and dissemination is recommended (“BZgA best evidence”). If an intervention with causal proof needs to be adapted, at least a process evaluation should be carried out; in the case of larger adaptations, a new effect evaluation might be needed [32].

Bödeker et al. [23] provide an overview and classification of experimental, quasi-experimental, and nonexperimental study designs for evaluating intervention effects in health promotion and disease prevention, with examples. A structured documentation and ongoing evaluation of the details of implementation processes (by quantitative and qualitative approaches) may elicit facilitators and barriers in the target context and aid in understanding issues of transferability [26]. Depending on the selected outcome(s), the evaluation of transferability goes beyond measuring intervention effectiveness, for example, by investigating return on investment or stakeholder satisfaction. Measurements of the public health impact include reach (participation rate and the representativeness of participants), effectiveness (impact of an intervention on specified outcome criteria, negative outcomes, and intended results), adoption (percentage and representativeness of organisations or settings that conduct the intervention), implementation (intervention fidelity, or the quality and consistency of delivery), and maintenance of the intervention (individual-level long-term results; setting-level institutionalisation of the intervention [RE-AIM]). [2, 20, 38].

Depending on the study design, questions might be answered with regard to the intervention and implementation: Were the hypotheses on transferability confirmed? (For example, was the intervention effective with adaptations?). What are the criteria that contributed to the success of the implementation? What are the lessons learned? (For example, which facilitators and barriers have been identified? Have barriers been removed? How do stakeholders explain the success or failure of the intervention and/or implementation?). Reflections can include a comparison with the primary context. The knowledge reported on transferability can form an information basis for further transfers of the intervention. Reporting statements might be used in addition to transferability criteria, such as the Template for Intervention Description and Replication (TIDieR) checklist and guide [41].