Reviews of community participation in in rural health suggest a hierarchy of participation types. This does not include what Kenny et al. (2013) [1] term “non-participation,” which involves being serviced or manipulated as users or research subjects with no real influence on what is done or how it is done [3]. Actual community participation sits on a continuum ranging from consultation (“tokenism” according to Kenny et al.) [1] to ownership (“power”). Tokenism includes informing, consulting and “placation”, while power includes partnership, delegated power, and ownership.

Preston et al. (2010) suggest that limited research at that time had focused on the contributions of community participation to health service design and health outcomes [2]. Bath and Wakerman (2015) asserted that evidence for the impacts of community participation were stronger in communities which were more marginalized and distanced from “mainstream” primary care systems [3]. They used the example of Indigenous communities, where the need for community-specific service models has long been recognized. Perhaps the evidence for impact of community participation is stronger in these settings because participation is essentially mandated, there are more or less formal structures of leadership and community representation, and research into service models in those communities consequently is more likely to focus on community participation.

More broadly, however, Kenny et al (2013) note the difficulty in continuing to find new leaders for community participation initiatives, reflecting persistent challenges in smaller rural communities of volunteer fatigue and systemic exclusion of community members who sit outside the dominant social structures [1]. There was also a tension between formal structures of community representation (such as local government) and structures that might be preferred for rural health initiatives. In small rural settings, the “local” government may not be local at all (headquartered elsewhere) or may not represent a broad range of community interests.

Our review extends the concept of community participation beyond simply “communal” models (where people are brought together to collectively represent the community), and we reject the dichotomy between “community” and “health service providers” in cases where those providers are also members of the community [18]. We define LALA as:

A local actor is any individual or group who lives in the location where the innovation is taking place.

A local action is any non-passive role played by a local actor in the innovation process.

Wakerman (2009) [6] and Wakerman and Humphreys (2011) [5] focused their reviews of rural health innovation on primary care and referenced both “community readiness” and “public participation” in their proposed models of enablers of rural health innovation, but these were not discussed in the text, and the process by which the models emerged from the research is unclear.

Wakerman and Humphreys (2011) [5] describe four types of service innovations—discrete (relating to a specific primary care task), integrated (involving multiple tasks or services), comprehensive (typically across a population sub-group) and outreach (providing services at a distance). They claimed that discrete and integrated models are more appropriate for larger communities, while comprehensive and outreach models are more appropriate for smaller communities. Asthana and Halliday’s (2004) [4] review encompassed more than primary care, but the typology of innovations they proposed was similar, with various forms of integrated care and outreach. They also suggested that smaller rural contexts would benefit from novel approaches to “intermediate care” to facilitate transitions to and from (distant) specialist services, and substitution of one type of professional with another (such as nurse practitioners, or physician-pharmacists).

Beyond service models, innovation may include public health interventions, clinical interventions, uses of eHealth, information distribution to build health literacy, and patient or user engagement [4]. Recent attention has largely focused on eHealth [19] where there has been some discussion about the extent to which eHealth local actors can influence tech companies and central health departments.

The rural health innovation reviews support the common idea that rural health and care systems need to do more with less—“innovation on a shoestring” according to Mathieu et al. (2020) [20]. While small budgets may be viewed as a hinderance to rural health innovation, the approaches to innovation that they dictate—small steps and “optionality”—have been promoted as a more robust approach than large projects which carry higher risk of failure or “lock-in” to unsuitable models [21]. “Optionality,” in particular, may be important as it allows lessons learned or resources gathered for one purpose to be used in other ways (Petrie et al. [21] give the example of teleconferencing facilities being used for non-health purposes). Similarly, Hodge et al (2016) [22], postulated that distance from the agencies which typically drive health innovation (provincial health departments, universities) can both limit local influence and provide opportunities for increased local action as local actors “fly under the radar.”

There is abundant literature describing both roles and activities associated with innovation [17]. Thune and Mina (2016) describe four broad roles in health service innovation processes: idea generation; idea development; selection; and implementation and dissemination [23]. Our synthesis of these and other “roles” frameworks [24] (see horizontal axis in Table 1, with definitions below) leads to explication of five roles which could be assumed by local actors.

Moullin et al. (2019) have promote the Exploration, Preparation, Implementation, Sustainment (EPIS) framework for understanding implementation of evidence-based practices in health and medicine (vertical axis, Table 1) [25].

Aarons et al. (2011) description of the EPIS framework defines the exploration phase as developing “awareness of either an issue that needs attention or of an improved approach to an organizational challenge” [26]^p6. The preparation phase involves the research, policy making, and consultation which leads to the decision to test (or adopt directly) a particular innovation idea or set of ideas. Implementation is the active task of putting the idea into practice through experimentation, physical construction, producing or procuring software or technology, formally adopting new guidelines and so on. Sustainment encompasses the securing of resources, establishing policies and protocols, and dissemination (through education and training or knowledge mobilization) that allows the initiative to become “standard practice.” It is often the case that academic attention to the innovation process “departs” at the point of implementation, but our review at least aims to allow for the possibility of investigation of sustainment of an initiative either in concept (plans for sustainability) or in practice (longer term evaluation).

The phases of the innovation process are not necessarily linear, and certainly overlap temporally and conceptually (for example, research and testing can continue throughout and beyond the implementation phase). All the roles (horizontal axis of Table 1) can be undertaken during all the phases in the EPIS framework. As with Thune and Mina’s (2016) review of hospitals as innovators, our framework suggests that local actors CAN engage in all roles and in all phases of the innovation process, but there has not yet been a synthesis of empirical evidence of the extent to which they do so engage, and the impacts of engagement on the outcomes of health service innovation in small rural contexts [23].

Results are summarised in Tables 2–5 which partition the 32 included papers according to “local actors” “levels of action” [1] which we have classified within the ownership-consultation continuum as being “completely local” (7 papers), “local initiated” (12 papers), “local active” (7 papers), and “local passive” (6 papers). Completely local papers (Table 2) had local actors as the only participants and through all phases. These seven papers were also the only ones where local actors served as evaluators. In Tables 2–5 we have added “a” to each of the 15 health professionals’ papers to indicate whether there was reference to any other level of community participation.

Local initiated papers (Table 3) covered projects that emerged locally, but had some external actors invited to participate in specific phases or perform specific roles.

Local active cases (Table 4) were typically initiated by external actors, but local actors were invited or appointed to active leadership and developer roles.

Local passive cases (Table 5) had less active local actor participation but did include at least some advisory functions which were acknowledged as influencing the course of the project.

The review included ten articles relating to public health interventions (for example an examination of a model for preventing chronic disease [29]), seven articles relating to clinical interventions (such as a palliative care model [30]) and seven relating to service models (e.g., an emergency care model [31]). There were also four articles relating to eHealth initiatives, three to knowledge or capacity building (such as an investigation of the establishment of a partnership for research in Indigenous communities in Australia [32]) and an initiative to improve patient journey management [33].

The tables do not include the theories of innovation used in the literature, as only seven papers explicitly cited any theoretical framework at all. These focused on approaches to engagement or participation of local actors (university-community engagement models [34], “principles for community governence” [35]^p614 and community based or socially accountable practice [35, 36]). One study used a decolonisation framework to direct their research on the impact of an Indigenous youth suicide prevention program in Alaska [37].

Four of the seven innovation-related theoretical frameworks were concerned with change management (theory of change [38, 39], “Kotter’s model of change” [41]^p2 and an improvements model [40]. Another study employed a diffusion of innovation theory in evaluating a telepsychology initiative [41], and one investigated the long-term outcomes of four primary care service models in the US using a “6 domains of sustainability” framework [44]^p1613.

In fifteen cases, the “local actors” identified were health service actors, and the nine made reference to other local participants are marked with an asterisk in the tables. Five of these involved an individual health professional [three nurses [33, 35, 36], pharmacist [52] and a clinical champion [40], while ten involved groups of health professionals and/or service staff, either a steering committee formed by local health services [38, 43, 44] or the “health centre” [27, 30, 41, 45–48]. A further two cases involved research and development units established by local health professionals or a local health service [31, 49]. Only four cases involved the participation of existing community-representative structures, including one where there was a health department within the local (county) government [29]. In eight cases, the innovation project itself established a new (and usually temporary) representative group whose structure was generally not well described.

There was evidence of local actors in all of the roles in our framework. The most common role was as researcher/developer (28 cases) and many of the articles reviewed (n = 18) featured local actors as co-authors. Next most common (n = 25) was the stimulator-initiator role, and the similar role of facilitator (n = 20) was also common. Local ownership (n = 18) was apparent in more than half the articles, while only 13 referred to local roles that could be interpreted as relating to translation or maintenance (sustainability) of the initiative.

Every paper had information about local action in the phases of exploration and preparation (Table 6). There were only six cases where local actors (LAs) were not actively involved in initiating the project. Local actors were active participants in the preparation phase in all but four cases. Those four cases involved local health professionals collecting data to a prescribed template and having no (declared) role in data analysis or interpretation.

It was much less common to have information about LALA at the “back end” of the innovation process. The reported research typically covered development activities (pilot testing, mostly), with no “re-visiting” of cases at later points in time. However, implementation was discussed in 13 articles, with local actors taking full responsibility for long term commercialisation of the initiative in six cases. Project partners (universities in all cases) continued to be involved in five cases [31, 32, 36, 49, 54], while in two cases [50, 51] new external partners were engaged to contribute to the implementation process. In terms of sustainment, there were six articles based on evaluations of past projects by researchers contracted for that task after the project had been commercialised, and one further case [51] where researchers involved in the development also conducted an evaluation.

As expected, given that most articles only reported research to the end of the preparation phase, there was very little evidence of the impact of LALA on the outcomes of innovation projects. Many impacts are difficult to quantify, and difficult to attribute specifically to LALA [47]. Ka’opua et al. (2011) claimed that their process of enrolling women in a breast screening program using church leaders and church services led to increased screening rates, but we do not know if these persisted after the initial trial period [52]. Connor (2009) attributed a change in regional health funding policy (from project specific to “pooled funding”) to local action, but few details were provided [53].

The review may have identified impacts of LALA on intermediate processes. Four articles [29, 32, 40, 49] described how local actors were successful in attracting project extension or implementation funding. Three studies described a process of handing control-ownership over to local actors at the end of the project and suggested that it was only local actors who were in a position to provide for the project’s sustainment [36, 43, 50].

There was also some evidence of impacts of LALA on the nature of the projects. In Nykiforuk et al. (2018), local actors showed the researchers how an urban based walking program could be adapted to rural conditions [54]. In two studies it was local actors who were able to redirect health funding to support the project [35, 37]. In four cases [30, 46, 52, 53], the attachment of a local actor (usually a health professional) to a university research centre was seen as giving credibility to projects that had struggled for support.

There were four notable examples of LALA actively influencing the research and development (preparation) process. In Singer et al. (2015), local agitation led to a complete re-design of the trial of an Aboriginal mental health initiative in Australia [55]. In Andersen and Jansen (2012), local input changed what was intended to be a complex technology-enabled telepsychiatry model into a simple telephone-based initiative [50]. Stewart and Conway’s (2000) plan to engage communities in design and trial of a drink driving prevention program had to be changed when those actors were reluctant to commit as much time as the researchers wanted [56]. In Trout et al. (2018), the research protocol was changed significantly by the input of a local reference group which had concerns about cultural appropriateness [37].