| dc.description.abstract | Introduction: Scaling-up health interventions is a complex process with growing recognition of the need to make adaptations to interventions to address diverse contexts and populations encountered across multiple sites. However these adaptations remain poorly understood with little documentation of what these adaptations are in practice, along with lack of specific guidance on what actions to use to make these adaptations when scaling-up. This leaves implementers with a lack of clarity as to why adaptations may be needed across sites when scaling-up, with minimal guidance on how to undertake the challenging process of adapting health interventions. As a result this may lead to ad-hoc or reactive approaches to adaptations with potential impact on intervention fidelity and effectiveness.
This research set out to uncover (i) what adaptations are made and why, when scaling up health interventions in practice and (ii) what are the actions that can be used to achieve these adaptations when scaling-up health interventions. Furthermore, the specific action of local decision making (LDM) was explored in more detail to discover (iii) how does LDM work (i.e. by what mechanisms, and in what contexts) as an action to achieve adaptations during scale-up of health interventions across diverse contexts.
Methods: These research questions were addressed using realist review methodology. Specifically, Pawson’s five stages of realist review of: (Stage-I) clarifying the review and developing initial theory, (Stage-II) developing a search strategy, (Stage-III) screening and appraisal, (Stage-IV) data extraction and (Stage-V) data synthesis and analysis, were utilised iteratively throughout the research process. As a final stage (Stage-VI), key informant interviews with those with experience of scaling-up health interventions were conducted and used to refine the theory for how and why LDM works for adapting health interventions when scaling-up. There were three distinct phases of this research.
Phase One - A background search was completed to develop an Initial Programme Theory (IPT) framework and focus the research question. The IPT was developed through analysis of previous adaptation, fidelity and scale-up guidance and frameworks, along with the wider implementation literature.
Phase Two - A systematic search was completed to identify examples of scale-up in practice where adaptations of health intervention occurred where the actions used to achieve adaptations were reported. The databases searched were; PubMed, Cinahl, Scopus, Global Indicus Medicus, Web of Science, EMBASE, and Psycinfo. Grey literature searches were also completed in Social Care Institute for Excellence (SCIE), Grey Lit and Open Grey. Further to the database searching, the references of identified articles and the ExpandNet bibliography were also hand-searched for more case examples. Through the case examples identified in this search, (i) details on the adaptations that were made (i.e. the reasons and type) and (ii) the actions that were used for these adaptations during scale-up were identified - in fulfilment of the first two research questions - and subsequently led to further refinement of the review question. For extracting the type and reasons for adaptations FRAME guidance (1) for adaptation reporting was used. For extraction of actions the IPT framework developed in phase one was used, with the actions headings refined as the review progressed.
Phase Three - The specific action of LDM was explored. IPTs relating to LDM for adaptations to health interventions during scale-up were developed and refined. Screening of case examples for relevance and rigour took place, with additional searching to identify more documents and provide further contextual information on these case examples. Data was extracted in the form of context-mechanism-outcome configurations (CMOCs). These were synthesised for demi-regularities and programme theories were refined. In addition, further searching for relevant substantive theory and interviews with key informants with experience in scale-up were utilised to refine programme theory and aid the development of an Middle Range Theory (MRT). This led to development of programme theories and an MRT on (iii) how and why LDM works for adaptations of health interventions during scale-up across diverse contexts, in fulfilment of the third research question.
Results
Phase One – An IPT framework was developed. The research question and search strategy for the subsequent phase were developed.
Phase Two - A total of n=22 case examples reporting on actions for adaptations made during scale-up were identified as a result of the systematic search in phase two. These were primarily based in Low and Middle Income Countries (LMICs) (n=19), with the remainder in High Income Countries (HICs) (n=3). The focus of the interventions ranged from: sexual and reproductive health (n=9); HIV (n=6); maternal and child health (n=4), two of which included vaccination programmes; non-communicable diseases (NCDs) (n=2); and mental health (n=1). Adaptations were primarily reported for the reason of increasing cultural acceptability (n=15), notably often to increase cultural acceptability in the wider population and context and not just for specific minority groups within a population. Adaptations were also commonly reported for resource availability (n=14), which occurred due to resource shortages at varying levels of the system i.e. at wider socio-political, organisational or recipient levels and were often unavoidable in nature. Proactive adaptations which were embedded into the scale-up design were reported in the majority of case examples (n=18). In addition to this, reactive adaptations also occurred in a number of case studies (n=11), these adaptations, although not embedded or planned from the outset, were often reported to serve needs that were identified as the intervention scaled-up. Ad-hoc unplanned adaptations (n=2) were also reported, primarily occurring at frontline provider level with a loss to fidelity noted in these case examples. Adaptations were reported at different levels of the system for example at national level (n=5); provincial, regional or district levels (n=6); organisational levels e.g. health facility or school (n=14); for specific cohorts within a population (e.g. based on gender, religion) (n=6) and based on individual recipient needs (e.g. literacy) (n=1). Fidelity considerations of adaptations and assessments were rarely documented. Adaptations were also often poorly reported. This research reflected and made suggestions for advancement of the FRAME adaptation reporting guidance, in particular for use in LMIC settings, for example suggesting further headings to differentiate resource limitations as a reason for adaptation, and the addition of further guidance on consideration of impact of adaptation on fidelity.
The actions used when scaling-up for adaptations to health interventions were identified with a resulting eight action headings emerging: (i) identification of intervention theory, core elements, components, functions or minimum standards; (ii) providing guidance to sites on intervention theory or components, on how to implement and adapt the specific intervention and/or sharing tried and tested examples of adaptations from other sites; (iii) LDM; (iv) using an adaptive scale-up design; (v) creating peer learning opportunities; (vi) ongoing assessment of the wider political, socioeconomic, cultural and organisational context to inform adaptations; (vii) the use of data to assess progress and inform or evaluate adaptations and (viii) the use of QI methods. These actions were often used in tandem. 16 of the 22 case examples reported using LDM where autonomy was given to local stakeholders to make decisions on adaptations for local fit within their setting. This occurred at varying levels within the system for example at provincial, regional, district, community, health facility levels, by frontline providers or the recipients themselves. The review was refined at this point to focus on developing theory on how and why LDM works for adaptations of health interventions during scale-up across diverse contexts.
Phase Three - Eight programme theories and an MRT were developed from the case examples which used LDM (n=16) on how and why LDM works for adaptations to health interventions when scaling-up These theories were further refined through interviews with key informants with experience in scale-up (n=6) and use of substantive theory. Multiple contexts such as: where the intervention does not align with the social norms, values and beliefs of the wider community or the local decision maker; where there is reliance on external funding and poor understanding of the intervention components from funders; and in resource limited settings where unavoidable adaptations may occur were all identified to impact how LDM works. These circumstances may lead to lead to intra- and person-role conflict for the local decision maker. Mechanisms at the level of the local decision maker of; perception of compatibility, awareness (i.e. of intervention theory), sensemaking, problem-solving, fear of repercussions, feeling valued and respected, were all found to contribute to how LDM works. The use of complementary actions such as provision of accessible guidance and support, for example on intervention components and on how to implement and adapt the specific intervention, can support local sites. Other complementary actions such as peer learning through sharing tried and tested examples from other scale-up sites, and using a data driven approach to inform and evaluate adaptations were suggested to support LDM by reducing the time needed for adaptations in later phases and contributing to buy-in from staff and the target population for the intervention.
Conclusion: Scale-up is complex, with the additional layer of making adaptations during scale-up further adding to this complexity. This research evidenced that adaptations are occurring in practice during scale-up of health interventions and has provided some guidance to implementers on actions that can be used for undertaking these adaptations. Implementers need to examine in advance what contextual challenges may occur in their setting (for example resource shortages, cultural conservatism etc.) that may lead to future adaptations and their potential impact on fidelity. This may assist implementers in considering whether adaptations are unavoidable, and in selecting what actions to embed into the scale-up plan. LDM is a promising approach for adaptations to health interventions during scale-up, however it occurs in complex systems, within a wider cultural and socioeconomic context, with varying levels of capacity, and often in tandem with numerous other approaches. While LDM may support those who know the setting to problem solve based on local knowledge leading to beneficial adaptations within their setting, it may also place local decision makers in a position of intra and person-role conflict leading to adaptations that impact the fidelity and effectiveness of the intervention. Therefore implementers need to consider the complexity of how LDM may work across the diverse contexts they encounter during scale-up. | en |
