Background

Equity of access to healthcare – and specifically, to primary healthcare – is a core value in public policy and a feature of highly-performing health systems in many high-income countries [1,2,3]. Although reforms to strengthen primary healthcare often invoke equity as a principal goal, they rarely succeed in adequately reaching vulnerable populations whose needs tend to be more complex than those of the general population [4,5,6,7,8,9]. Indeed, many innovations to improve access to care tend to favour the wealthiest or most educated segments of the population – as famously captured by the Inverse Care Law [10]. Achieving healthcare equity, therefore, requires organizations to implement pro-vulnerable innovations, tailored to reach and meet vulnerable populations’ specific needs [11].

Vulnerable populations are groups or individuals who are more susceptible to harm because they lack the personal, material, and social resources to successfully cope with the challenges they face and to counter potential harm [12]. These populations are at higher risk for poor health status and problematic access to healthcare [4, 13]. Vulnerable populations include: Aboriginal or Indigenous communities, refugees, visible ethnic minorities, individuals living in poverty, people experiencing homelessness, people with disabilities, people with limited social support, those with complex health conditions, certain age groups; and many other marginalized and underserved populations [4, 13,14,15,16,17]. Inequitable access to healthcare stems from gaps between vulnerable populations’ abilities to access care, and healthcare organizations’ accessibility (see Fig. 1) [18]. To close these gaps, healthcare organizations must adapt their accessibility to vulnerable populations’ abilities. However, designing pro-vulnerable innovations is challenging for decision-makers, providers, and other health system stakeholders, as they may lack knowledge of the range of possible innovations to address vulnerable populations’ specific needs.

Fig. 1
figure 1

Patient-Centred Accessibility Framework [18]

Full size image

This challenge of designing pro-vulnerable innovations was highlighted in our “Innovative Models Promoting Access-to-Care Transformation” (IMPACT) research program [19]. IMPACT brought together local health system decision-makers, providers, primary healthcare researchers, community members, and other stakeholders in six regions in Canada and Australia. In each region, a local multi-stakeholder partnership designed and piloted an organizational innovation aiming to improve access to primary healthcare for vulnerable populations. A foundational premise of the IMPACT program was that innovations should aim to adapt primary care organizations and service delivery arrangements to meet vulnerable populations’ needs, rather than placing the onus solely on vulnerable populations to improve their abilities to access care. To inspire local health system partners, the research team was to provide a menu of existing pro-vulnerable organizational innovations to improve access to primary care.

Substantial research has been conducted on interventions to improve the delivery of care – most notably the taxonomy produced by the Effective Practice and Organisation of Care (EPOC) Group Cochrane Review Group [20]. However, they provide an overview of general interventions that, for the reasons highlighted above, may not address the specific needs of vulnerable populations. Early in the IMPACT program, members of the team conducted a scoping review of studies that described organizational innovations that improved vulnerable populations’ access to primary healthcare – significantly reducing unmet need for care, use of hospital emergency rooms or hospital admissions [17]. This scoping review [17] presented complex interventions, mapped onto the EPOC taxonomy, in which various components of interventions appeared across different interventions. While decision-makers appreciated having results specific to vulnerable populations, they found transposing the results to their contexts challenging and, therefore, of limited use to inform the design of organizational pro-vulnerable innovations.

To better support decision-makers, we aimed to develop a typology of pro-vulnerable organizational innovation components – akin to “building blocks” that could be combined in different ways into new complex innovations or added to existing organizational processes. This typology is intended to provide a comprehensive range of components of organizational innovations to be considered by health service decision-makers as options to address the primary healthcare access needs of vulnerable populations.

Methods

Design: typology

We developed a typology of components of organizational innovations. A typology is a description and categorization of complex organizational forms [21] developed using qualitative (or quantitative) analysis [22]. The goal of a typology is to divide a whole phenomenon – in our case, organizational innovations to improve access to primary healthcare for vulnerable populations – into distinct but related categories [22, 23]. Typologies have been used in primary healthcare with the intent of guiding organizational change and can provide a “menu” of items to inform the design of interventions [24,25,26,27,28,29]. In a “typology” – as opposed to a “taxonomy” – items are not ordered hierarchically and are not entirely mutually exclusive.

A qualitative typology is generally structured around a conceptual framework that helps classify emerging categories [22, 23]. In the IMPACT research program, access to primary healthcare for vulnerable populations was conceptualized primarily based on the Patient-Centred Accessibility Framework [18]. This framework posits that access to healthcare results from the interaction, at different stages, between organizational dimensions of healthcare and patients’ abilities (Fig. 1). We focused on the dimensions of accessibility on the organizational side of the framework (i.e., Approachability, Acceptability, Availability & Accommodation, Affordability, and Appropriateness).

Inspired by Greenhalgh et al. [30], we defined organizational innovations as: “a novel set of organizational behaviours, routines, and ways of working that are directed at [a common objective] and that are implemented by planned and coordinated actions.” Organizational innovations to improve access to primary healthcare for vulnerable populations were identified from two complementary sources of data: a) a scoping review of the peer-reviewed literature [17] and b) an environmental scan [16]. Both are described briefly below and have been described in detail elsewhere [16, 17]. The scoping review and scan were conducted as part of the IMPACT research program [19].

Data source: a) scoping review

A scoping review was conducted to explore the breadth of available evidence on organizational innovations in primary healthcare [17]. The search focused on academic, peer-reviewed literature and was conducted in three of the largest and most relevant databases for studies related to primary healthcare (Medline, Embase, and CINAHL). The search was performed by a specialized librarian (see an example of the search strategy in Additional file 1). In addition to the database searches, four primary care experts from the IMPACT team (including JH) were asked to share their personal primary care reference files, from which citation tracking was performed to identify additional relevant studies. The search was limited to articles published between January 2000 and March 2014, a period corresponding to an international commitment to strengthening primary healthcare, up to the beginning of the IMPACT program. One researcher (VK) scanned 8694 titles and abstracts for relevance, then assessed 1760 potentially relevant studies for eligibility. For the typology, we selected any quantitative, qualitative, or mixed methods studies carried out in Organization for Economic Cooperation and Development (OECD) countries and published in English or French that met all four of the following eligibility criteria:

  1. (1)

    involved at least one organization at the primary healthcare level in the health system;

  2. (2)

    was organizational (not directed at the system as a whole or only to the population);

  3. (3)

    had an explicit objective to improve access to care;

  4. (4)

    was directed to a vulnerable population.

The 129 eligible full texts were read by three team members (MAS, JH, SD) to select 90 articles where innovations were described in detail.

Data source: b) environmental scan

The environmental scan was conducted after the review. It was designed to capture organizational innovations that had not been published in the academic literature [16]. Briefly, a 5-min online survey was disseminated using a social network approach over 6 weeks between July 10 and August 21, 2014. Primary healthcare informants known to the research team were sent a link to the online survey by email and, in turn, were asked to share the survey link within their social networks. The survey was also promoted on social media through 248 posts on Twitter linked to findings of interest and emerging findings from the survey. Participants were invited to identify a program, service, approach, or model of care that they considered innovative in helping vulnerable populations access primary healthcare. They were encouraged to provide links to any available description, such as websites or documents. The definition of an organizational innovation was left to the discretion of the respondents to ease the response burden. We received 744 survey responses. After screening innovations for eligibility and redundancy, 240 unique innovations were retained for the typology.

Typological analysis

We conducted a typological analysis [22, 23]: first, a) of the selected peer-reviewed articles (n = 90) and, subsequently, b) of the selected survey responses (n = 240). Each article (n = 90) was carefully read independently by two individuals (MAS, SD), who focused on the ‘types’ of organizational innovations that addressed accessibility and highlighting all passages describing the innovation. The highlighted passages were used to draft a detailed description of each innovation in an Excel document. For each article, we also extracted information about the setting, target population, and vulnerabilities addressed (e.g., frail elderly, homeless, Aboriginal, low income). Reading the detailed description of each innovation, we used a predominantly inductive approach, grouping similar components recurring across organizational innovations. The unit of analysis was a distinct component that could either be a stand-alone intervention (navigation & information), or a combination of components consistently occurring together (e.g., case management, advanced access). We applied standard labels to components where possible (e.g., community health worker) and descriptive labels to others (e.g., proactive identification of need, cultural adaptation).

Subsequently, the emerging typology components were tested by coding the survey responses (n = 240) from the scan. Two individuals (MAS, LR) independently coded the responses in an Excel document, then met to resolve discrepancies and discuss possible additions and clarifications to the typology components.

Our initial analysis yielded 48 unique innovation components. Most of the innovations reported in the studies were complex interventions that involved 1–14 components, with an average of six components each. Almost all components were established in the first 40 published studies; analysis of additional published studies and the environmental scan led to refinements. The scoping review and scan data coding were reviewed based on this initial typology and adjusted for consistency and to ensure all relevant components had been captured. We then reviewed the innovations coded to each component of the typology to write a short general description of the component and to select illustrative examples of the component.

Two individuals (MAS, JH) then reviewed the initial typology components, descriptions, and examples and mapped them to the organizational dimensions of the Patient-Centred Accessibility Framework [18]. The classifications were based on principal and secondary dimensions of accessibility addressed by the component. Discrepancies in classification were resolved through discussion and checked by the rest of the team.

Typology field-testing and refinement

The scoping review and environmental scan were conducted in 2014. From 2014 to 2018, the initial typology was presented to academic audiences and field-tested with local partners (e.g., decision-makers, patients, health professionals, researchers) to help them design pro-vulnerable organizational innovations to improve access to primary healthcare. When these initial components were presented to local partners and peers at conferences, they affirmed that the initial 48 components were useful for expanding the options to be considered when designing an innovation. However, they also perceived redundancy between components or lack of direct relevance to accessibility. Therefore, we excluded components that were not organizational components per se, but rather resourcing mechanisms: student health professionals, financial incentives, organizational networks, and provider education. We also excluded components that are, in fact, care attributes, such as transparency, advocacy, patient-centred care, and empowerment. Finally, we excluded a few components that applied more to the content or quality of care, such as quality improvement initiatives or self-management education. Components of a similar nature were further collapsed into 18 components of organizational innovations.

Results

Organizational innovation characteristics

The 90 selected peer-reviewed studies were set in: the USA (n = 60) [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90], Canada (n = 9) [91,92,93,94,95,96,97,98,99], Australia (n = 8) [100,101,102,103,104,105,106,107], UK (n = 6) [108,109,110,111,112,113], New Zealand (n = 2) [114, 115], Israel (n = 2) [116, 117], Italy (n = 1) [118], Mexico (n = 1) [119], and Germany (n = 1) [120]. For the environmental scan survey, 45.0% of responses originated from Canada, 40.8% from Australia, 9.4% from others countries (e.g., Ireland, UK, USA, Netherlands, Italy, Israel, Switzerland, Cameroon, India, Indonesia, Sudan) and 4.8% were missing country information [16].

The organizational innovations identified in our data targeted a wide variety of vulnerable populations. Targeted populations typically combined various vulnerabilities, most commonly: low-income, chronic illnesses, Indigenous populations, homeless, migrant or refugee status, ethnic minorities, uninsured or underinsured, marginalized groups (drug users, recently incarcerated), persons with mental illness, the frail elderly, at risk youth, and frequent users of emergency departments.

Typology

The final typology was comprised of 18 components of organizational innovations, presented with examples [42,43,44, 54,55,56,57,58,59,60,61, 78,79,80,81,82,83, 94,95,96, 105, 110, 113, 119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136] in Table 1. The innovation components are organized by the principal accessibility dimension they address. The 18 components principally addressed the dimensions of Availability & Accommodation (7/18), Approachability (6/18), and Acceptability (3/18). Only one component addressed each of Affordability and Appropriateness as principal dimensions, although these were addressed as secondary dimensions as part of other innovation components.

Table 1 Typology of organizational innovation components to improve access to primary healthcare for vulnerable populationsa
Full size table

Discussion

The integration of the scoping review and environmental scan of innovations along with field-testing resulted in a typology of 18 components of organizational innovations to enhance the accessibility of primary healthcare for vulnerable populations. The typology was based on both published and unpublished innovations. It offers a comprehensive menu of potential components that can help inform the design of innovations and can be combined into complex interventions or added to existing organizational processes to meet the access needs of vulnerable populations. Mapping of the components to the accessibility dimensions of the Patient-Centred Accessibility Framework [18] allows health service designers to match appropriate innovations to identified access needs.

The typology offers a categorization of health service delivery arrangements inspired by the Cochrane taxonomy of EPOC [20], but tailored to the needs of service designers and specific to the domain of access for vulnerable populations. The ultimate goal is to improve healthcare equity through pro-vulnerable innovation design. A few components are similar to those found in the EPOC and are not specific to vulnerable populations, including group visits, expanded hours, advanced access, virtual health services, one-stop-shops, and role expansion. However, they have been demonstrated to be well-suited to address the needs of specific vulnerable populations, although organizations still have to be intentional about a pro-vulnerable focus to achieve healthcare equity. Other components in the typology are specifically designed to address the needs of vulnerable populations and differ from the EPOC taxonomy. Proactive identification of need, proactive appointment making and contact, and outreach pull vulnerable persons into primary healthcare and maintain contact rather than placing the onus on vulnerable populations to perceive their needs and navigate the care-seeking process. Similarly, community health workers, service brokerage, and transportation services bridge the gap between the health system and vulnerable populations.

Despite the intention to make labels and descriptions as mutually exclusive as possible, there was considerable overlap between some components. For instance, navigation and information and proactive identification of need, although implemented as stand-alone innovations, are also functions of community health workers and case managers. Likewise, culturally adapted services or inter-organizational pathways are functions of community health workers or role expansion innovations; these can be implemented as stand-alone interventions or added to existing organizational processes to address an accessibility need.

Using the typology in the design of innovations

This typology is intended to provide “building blocks” that can be combined into complex innovations or added to existing organizational processes to address specific access needs of vulnerable populations. The analysis done by Khanassov et al. [17] emphasizes that interventions are most effective in reducing unmet needs, emergency department visits, and hospitalizations when the intervention components are formally coordinated or integrated with other parts of the health system. Integration and coordination are critical considerations in implementing any of these interventions.

The Patient-Centered Accessibility Framework [18] is helpful in both identifying the access needs of the population and the components that are most appropriate to address those needs. For example, a population that frequently visits the emergency department may require an innovation that targets Approachability or Acceptability (e.g., navigation & information, proactive identification of need, culturally adapted services). Conversely, if the target population is seeking services but is disengaged from ongoing care, then Appropriateness-related components such as proactive appointment-making, case management, or one-stop shops may be better suited.

It is not surprising that most of the innovation components pertain to Approachability and Availability & Accommodation. Several studies have highlighted that vulnerable populations often face difficulties perceiving health needs, navigating the health system to find services, finding time to obtain services, making an appointment, and finding transportation to reach services [18, 137,138,139]. This typology emphasizes the importance of addressing barriers early in the care-seeking trajectory and provides a range of potential solutions to mitigate these barriers.

Although other components addressed Affordability as a secondary dimension, it is surprising that only one component in the typology (defraying costs to patients) related principally to Affordability, especially given the importance of direct and indirect cost as a barrier to care for vulnerable populations [140,141,142,143]. This result is partly due to our focus on identifying organizational innovations that can be implemented by a single organization or sub-system. Macro-level ‘innovations,’ such as universal health insurance and the Affordable Care Act, were excluded from our analysis because they are country-level legislative policies rather than organizational innovations.

Example of how the typology has been used

As mentioned, the typology was field-tested for relevance and usability in our local partnerships between 2014 and 2018 as they designed and piloted innovations to address local access priorities for vulnerable populations. In the local partnership in Quebec (Canada), a preliminary version of the typology was used to inform discussions about the design of an innovation for patients from disadvantaged neighbourhoods who face barriers connecting with a regular primary care provider. The innovation was a combination of several components: trained volunteer community members (inspired by community health workers) reached out to patients from disadvantaged neighbourhoods who were on the provincial centralized waiting list to find a primary care provider (existing primary healthcare brokerage) to screen for potential access barriers (proactive identification of need). These volunteers provided support by telephone before and after their first visit with their new primary care provider, including discussing the importance of attending the first visit, offering information about the clinic, giving visit preparation materials, and providing general information about the health system (navigation & information). Stakeholders in the local partnership perceived the typology as a useful tool to expand the menu of innovations and to reflect on how components could be added to existing organizational processes. Furthermore, we used the typology and the Patient-Centred Access Framework [18] to describe each local partnership’s innovations. This tool allowed the IMPACT research team, which spanned six sites in Canada and Australia with different contexts, languages, and terminology, to clarify their aims, improve mutual understanding across sites, and compare their innovations.

Implementing components of an innovation

The selection of components as part of a complex innovation also depends on the resources and implementation control available to innovation designers. The decision to implement components such as group visits, drop-in services, facilitated appointment making, expanded hours, culturally adapted services, case management, or advanced access are generally within the control of a single primary healthcare organization and can be resourced through the reorganization of existing resources. In contrast, components such as outreach of primary healthcare services, transportation, and navigation & information services require investment of new resources that may be possible for a single organization with a strong commitment and sufficient resources. Other components, such as role expansion and virtual health services, may require changes at a higher level since they involve changes beyond the reach of a single organization. For instance, role expansion – such as nurses working at the top of their scope of practice or empowering front office staff to refer patients to social workers – may require collaboration with professional associations to modify regulations governing professional practice. Similarly, implementing a community health worker may require collaboration with other organizations to set up certified training programs or to secure funding to cover salaries. Similarly, one-stop shops or inter-organizational pathways are based on collaborations between various organizations outside of primary healthcare and require substantial support and political will from local or regional health authorities for implementation.

Strengths and limitations

Because we achieved saturation with the peer-reviewed articles from the scoping review and descriptions of innovations from the environmental scan, we are reasonably confident that this typology reflects the most common innovation components to improve access for vulnerable populations. We recognize that we have excluded innovations outside the study period. Yet, as 2000 to 2014 represents a period of renewal and reform for primary healthcare marked by an effervescence of innovations, we believe that we have captured the most common components to improve access to primary care. Since 2014, we have continued to test the comprehensiveness of our typology by informally comparing it to innovations described in more recent peer-reviewed studies (although we did not use the systematic approach applied to the literature of the study period). In subsequent presentations to international audiences at conferences, we heard of additional examples of innovations, such as organizational arrangements between hospitals and farmers’ markets, but were able to locate them within typology components. We have not found any new components emerging from more recent literature or examples of innovations to add to our typology. An additional strength of this study is that from 2014 to 2018, the typology was field-tested for usability and relevance with local partners to design pro-vulnerable innovations. We are therefore reasonably confident that, although a new scan or scoping review would add detail and examples to the typology, it would not fundamentally change the components.

We also recognize that our scoping review data, from which we developed our initial typology components, described innovations in only nine OECD countries (Canada, USA, UK, Australia, Germany, New Zealand, Italy, Israel, and Mexico). We minimized the effect of publication bias by doing an environmental scan to identify unpublished innovations, but this method is susceptible to selection bias. Most responses came from Australia and Canada, and running the survey during the Northern Hemisphere’s summer may have limited responses from Europe and the USA. We also have no way of assessing the response rate or comprehensiveness of responses gathered using the social media approach. Although we did have a small number of respondents from low-income countries in our environmental scan, our approach may have excluded some innovative initiatives in low- and middle-income countries (e.g., community-based insurance plans [144], identification of accredited clinics [145], and subsidized payments for primary healthcare services [146]. Another example is the well-known hub-and-spoke models in India [147], which combine role expansion with inter-organizational pathways to provide low-cost, high-quality services to underserved populations [148].

A final limitation is that the typology labels reflect available examples and the judgment and language of the analysts. Labels such as proactive identification of need are clumsy but they circumscribe a unique set of examples.

Conclusions

This typology is unique as it presents components of innovations that can be put in place by primary healthcare organizations or other health system stakeholders to improve access to primary healthcare for vulnerable populations. Further research on the effectiveness of combining different components may help inform efforts to improve access for vulnerable populations.