Participatory implementation research in the field of migrant health: Sustainable changes and ripple effects over time

Abstract Objective This study aimed to explore whether positive impacts were sustained and unanticipated ripple effects had occurred four years after the implementation of interventions to improve cross‐cultural communication in primary care. Background Sustaining the implementation of change using complex interventions is challenging. The EU‐funded “RESTORE” study implemented guidelines and training on cross‐cultural communication in five Primary Care sites in Europe, combining implementation theory (Normalisation Process Theory) with participatory methodology (participatory learning and action—PLA). There were positive impacts on knowledge, skills and clinical routines. Design, setting and participants Four of the five original sites (England, Ireland, Greece, The Netherlands) were available for this qualitative follow‐up study. The study population (N = 44) was primary healthcare staff and migrants, most of whom had participated in RESTORE. Intervention; main outcome measures PLA‐style focus groups and interviews explored routine practice during consultations with migrants. Etic cards based on the effects of RESTORE stimulated the discussion. Deductive framework analysis was performed in each country followed by comparative data analysis and synthesis. Results Changes in knowledge, attitudes and behaviour with regard to consultations with migrants were sustained and migrants felt empowered by their participation in RESTORE. There were ongoing concerns about macro level factors, like the political climate and financial policies, negatively affecting migrant healthcare. Conclusion There were sustained effects in clinical settings, and additional unanticipated positive ripple effects, due in part, from the participatory approach employed.


| INTRODUC TI ON
There are complex relationships between research, policy and practice in primary care, as well as increasing attention to translational gaps between them. 1 To address these gaps, a number of strategies have been proposed for researchers. These include greater use of theoretical approaches in research focused on implementation 2 and the use of participatory methods. 3 Implementation science has grown rapidly in recent years, and a number of implementation theories are in use. Each offers a specific lens on the implementation process. For example, diffusion of innovation 4 focuses on the introduction and spread of innovation in a clinical setting, while normalization process theory (NPT) has an extended focus from introduction of new practices through to embedding and sustaining them to the point that they are considered routine, that is normalized. 5 Implementation theories, however, are rarely used prospectively, and it is not always clear how to operationalize them. 6 Further, they are not designed to support individuals and groups through the development of action plans to shape implementation work in primary care. 7 Participatory methodologies, on the other hand, do just that. 8 All share a focus on including stakeholders affected by the issue under consideration and in a position to act on the findings to develop action plans (see www.ICPHR.org).
Participatory health research and specific approaches such as participatory learning and action (PLA) can enhance public and patient involvement (PPI) and support implementation processes in primary care. 7,[9][10][11][12] There is also increasing recognition of the importance of the role patients/public may play in the broader, macro-level context that shapes organizational capacity and willingness to take action to support implementation of changes in health-care settings. 1,6,13 In this paper, we provide results of a follow-up study of an earlier investigation that combined NPT with PLA in primary care implementation research.
NPT is a contemporary sociological theory, developed by studying the implementation of health-care innovations. It focuses on the work that stakeholders (eg clinicians, managers, patients) must do to introduce, integrate and embed a new way of working in daily routines until it is sustained in routine practice. 5 NPT describes four elements of implementation work: sense making, engagement, enactment and appraisal (see Table 1). These have been used successfully as a conceptual framework to enhance understanding of levers and barriers to implementation for a variety of interventions. 14 PLA is a practical approach to investigate problems among diverse stakeholder groups where asymmetries of power may exist. 16 As such, it provides a valuable approach for meaningful, rather than tokenistic PPI in research. This is particularly the case with groups such as migrants who traditionally are underrepresented in PPI and in decision making in primary care. 17 It has the capability to engage participants in a collegial, inclusive and active processes. This enables their authentic perspectives to emerge clearly in research outcomes. 18,19 The approach requires a specific PLA mode of engagement, which promotes values of reciprocity, mutual respect, co-operation and dialogue within and across diverse stakeholder groups. 20 PLA techniques (see Table 2) are inclusive and user-friendly. These can be incorporated into interviews and focus groups in primary care research. 19,[21][22][23] The EU-funded project RESTORE (REsearch into implementation STrategies to support patients of different ORigins and language background in a variety of European primary care settings; summarized in Table 3) combined NPT and PLA to investigate and support implementation of guidelines and training initiatives to improve communication in primary care with migrants who are not fluent in the language of their host country. 7,11,12,24,25 There was evidence of positive changes to attitudes, knowledge and behaviour in practice settings 12 (see Table 3).
The use of PLA in RESTORE emerged as a key facilitator for the NPT implementation process. 7,11,12 PLA enabled participants with different levels of knowledge and power-doctors, practice assistants and migrants-to work together in a democratic manner. 7,11,12 This led to a shared feeling that the guidelines and training initiatives made sense (NPT construct 1), a shared sense of responsibility and engagement (NPT construct 2) to implement guidelines and training initiatives and shared work to enact them in daily practice (NPT con- These matters have not yet been explored in relation to RESTORE and warrant investigation.

| Aims and objectives
The aim of the present study was to describe the impact of the NPT-

| Study population
The study population was primary care staff and migrants (community representatives working with NGOs/patients using primary care services) as well as other stakeholders (eg professional interpreters in Ireland) who participated in the general practices involved in the RESTORE project that were also involved in this study. At some sites, new staff members also participated.
Participants for the follow-up study in each country were recruited through purposive sampling, striving for representatives of each stakeholder group involved in RESTORE, via the RESTORE principal investigators (PIs) at each primary care site, using a combination of email and letters.

| Data generation and analysis
Data were primarily generated using focus groups. Individual interviews were conducted to facilitate the involvement of those who could not attend the focus groups. Additional data were obtained based on the needs of participants: observation of clinical practice in the Netherlands because practice staff were unavailable for interviews, and email submission in Ireland for a participant who was abroad but wanted to take part. In England, given the interest of the local team, an analysis of minutes of meetings of a local policy group on cross-cultural communication between spring 2015 and end 2018, as well as related documents, was conducted (see Table 4 data generation methods).
A PLA etic card sort technique was designed for focus groups and interviews. This is an interactive method for facilitating and recording brainstorming around topics, which draws on relevant knowledge from previous research. 26 For this study, 23 cards (see Table 5 Data analysis took place in pairs (RESTORE PI and student) in each country, in the language of the country, following the principles of deductive framework analysis. 27 Specifically, the a priori

TA B L E 3 Information on RESTORE project (2011-2015)
RESTORE was an EU-funded qualitative case study project, which investigated and supported the implementation of guidelines and training initiatives that were designed to support communication between migrants and their primary care providers in five countries (Austria, England, Greece, Ireland and the Netherlands) 24,25 RESTORE was innovative in its combined use of PLA and NPT to guide methodology and provide a theoretical implementation framework. 24 Throughout this process, the PLA approach facilitated health-care providers to work collaboratively with migrants; to select and adapt a guideline or training initiative for their local setting; and to introduce it into their practice setting. 11,12 There were multiple impacts across settings. These included changes in knowledge (eg new knowledge and skills from completed training), attitudes (eg more tolerant and positive attitude towards migrant service users among receptionist staff) and behaviour (eg more effective communication in consultations between general practitioners and practice nurses and migrants with low literacy; increased flexibility in accommodating migrants' appointments among all staff). Impact on clinical practice routines was strongest in England and the Netherlands. 12 Lack of resources for interpreting services in primary care and the impact of economic austerity reduced the impact in Ireland and Greece 12

| Study participants
There were 40 participants in this study who also participated in the RESTORE study (65% of the sample in the original study (N = 63).
Other participants in the RESTORE study had moved away or were otherwise unavailable. In addition, four new practice members in the 22. Researchers changed their research approach: their willingness to think about and share ideas with others and admit gaps changed Section 5. Other data 23. Other spontaneously offered thoughts not related to any of the above topics UK participated. Table 6 provides a breakdown per country and information on socio-demographic and professional backgrounds as well as a comparison with the participants in the original study.

| Attitude and knowledge
Changes in attitude and knowledge from the RESTORE project continued in all settings. This primarily related to a more migrant-friendly attitude and awareness about migrants' needs (Eng, Nl, Ire, Gre15).
There is more awareness, more awareness regarding this population. In both England and Greece, it was suggested that migrants had therefore become more confident in GPs' diagnosis and treatment (Eng,Gre16). In Greece, students and residents who were engaged in RESTORE had become more culturally sensitive professionals (Gre10).
Their whole team had improved in practices towards migrants (Gre13). In the Netherlands, one of the receptionists expressed a negative sentiment against migrants.
Empathy and understanding is low in these patients.

| Organizational changes
There were reports of continued practical changes, including longer appointments and use of speaker phones to enable interpreted consultations in England (Eng14). Similarly, drawing on interview and observational data in the Netherlands, longer consultation slots were still planned in case of language differences. They had implemented and continued to use easy-to-understand patient information and pictograms, which improved greatly the understanding and accessibility of services for migrants.

| PLA enabled ongoing effect
The continued effect of the RESTORE was associated by many participants with the use of PLA to develop and implement action plans.
Reflections on PLA included the following:

| Quality of the training
In the practices in England and the Netherlands, the continuation of RESTORE's effects was seen to be a result of the quality of the training programme. In line with the findings above, participants reflected positively on the participatory approach to co-design training, that is using role play (Eng15), RESTORE's willingness to tackle difficult issues including racism, and the involvement of both administrative and clinical staff. It was also helped by the ongoing commitment of the participating practice to high-quality health care for migrants (Eng14, 15,16).  There was a strong view from one migrant that the broader political context, including Brexit and a hostile Home Office environment towards migrants, was likely to limit any changes resulting from RESTORE (Eng3,4).

| Lack of funding and local context limits further implementation
In the middle of Brexit, things are probably worse than they were.

(Eng;MIG)
In Greece, the changes in the participating primary care practice had been noticed by other practices, but overall the effect of the RESTORE project was felt to be limited to the participating health-care centre, due to the limited spread of the results to the local community and other primary care settings (Gre9,17).
[ It was emphasized, however, that despite the increase in interpreters' number during the past few years, they were still insufficient to meet the increased needs of the refugee population. Some other Greek respondents thought people had not changed at all (Gre3) and felt the whole health-care system was overburdened due to the economic crisis (Gre7).

| Community participants' empowerment
Although not to the same degree everywhere, migrant participants at all sites reported that they felt empowered by their participation in the RESTORE project, or by the attitude of the participating health-care professionals. Some-but not all-of the migrants in the English setting considered that participation in RESTORE had enhanced their reputation, confidence and sense of empowerment.

| New collaborations for research and policy
At all sites, the academic/researcher participants changed their re-

| Comparison with literature
Our results support the conclusion of the original RESTORE project that PLA is a key facilitator for supporting a sustained implementation process. 7,11,12  The influence of contextual factors, such as political climate, on the implementation and sustainability of changes in primary care was also found in the review by Lau et al, 1 who postulated that the 'fit' between the intervention and the context is critical in determining the success of implementation. This is well documented in global health and Indigenous health in Australia and Canada. [29][30][31][32] Implementation research in the field of migrant health should consider these macro-level influences on the process and outcomes. To achieve changes that are really sustainable, funding and manpower, a favourable political climate as well as the ambition to take new ways of working further is required. 33 This highlights a key lesson learned regarding the value of having more senior-level decision makers involved in participatory dialogues in primary care settings. Would the involvement of senior-level decision makers improve the mobilization of resources after successful small-scale pilots? Can such pilots projects generate change by reshaping policy agendas? These issues warrant further research.

| Methodological strength and limitations
This follow-up descriptive study is one of the few to assess the sustainability of the implementation of an intervention in primary care after four years.

| CON CLUS IONS
Implementation research in primary care that uses participatory approaches supports the introduction of new ways of working in routine practice that can be sustained over time. Further, the use of a participatory approach yields additional, unanticipated, positive effects on all participants. Participatory implementation research should be used to investigate and support other innovations for other populations in primary care.

ACK N OWLED G EM ENTS
We want to acknowledge all participating migrants, nurses, practice assistants, doctors, staff and health service planners as well as Mrs Maria Loukogergaki, who as a student contributed to the Greek data collection.

CO N FLI C T O F I NTE R E S T
No conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.