A population-based approach to integrated healthcare delivery: a scoping review of clinical care and public health collaboration

Background A population-based approach to healthcare goes beyond the traditional biomedical model and addresses the importance of cross-sectoral collaboration in promoting health of communities. By establishing partnerships across primary care (PC) and public health (PH) sectors in particular, healthcare organizations can address local health needs of populations and improve health outcomes. The purpose of this study was to map a series of interventions from the empirical literature that facilitate PC-PH collaboration and develop a resource for healthcare organizations to self-evaluate their clinical practices and identify opportunities for collaboration with PH. Methods A scoping review was designed and studies from relevant peer-reviewed literature and reports between 1990 and 2017 were included if they met the following criteria: empirical study methodology (quantitative, qualitative, or mixed methods), based in US, Canada, Western Europe, Australia or New Zealand, describing an intervention involving PC-PH collaboration, and reporting on structures, processes, outcomes or markers of a PC-PH collaboration intervention. Results Out of 2962 reviewed articles, 45 studies with interventions leading to collaboration were classified into the following four synergy groups developed by Lasker’s Committee on Medicine and Public Health: Coordinating healthcare services (n = 13); Applying a population perspective to clinical practice (n = 21); Identifying and addressing community health problems (n = 19), and Strengthening health promotion and health protection (n = 21). Furthermore, select empirical examples of interventions and their key features were highlighted to illustrate various approaches to implementing collaboration interventions in the field. Conclusions The findings of our review can be utilized by a range of organizations in healthcare settings across the included countries. Furthermore, we developed a self-evaluation tool that can serve as a resource for clinical practices to identify opportunities for cross-sectoral collaboration and develop a range of interventions to address unmet health needs in communities; however, the generalizability of the findings depends on the evaluations conducted in individual studies in our review. From a health equity perspective, our findings also highlight interventions from the empirical literature that address inequities in care by targeting underserved, high-risk populations groups. Further research is needed to develop outcome measures for successful collaboration and determine which interventions are sustainable in the long term. Electronic supplementary material The online version of this article (10.1186/s12889-019-7002-z) contains supplementary material, which is available to authorized users.


Background
The population health approach describes a shift in our healthcare system from a narrow model of acute care targeted at the individual patient, to one that focuses on the health and overall wellness of the broader population it serves [1]. In doing so, this approach highlights that clinical care, such as primary care, is only one of a wide range of 'institutions' that impact health [2]. Additionally, public health efforts, with a greater focus on health promotion and chronic disease prevention, can complement clinical care in order to provide populations with a comprehensive set of promotive, preventive and curative health services, thereby promoting overall population health [1].
Increasingly, healthcare systems around the world are facing persistent pressures that result in poor performance and growing inequities in care [3]. These include a rising burden of illness attributable to major chronic diseases, as well as increasing costs and complexity of healthcare delivery [4,5]. To address these growing pressures and achieve a shift from our traditional biomedical model of healthcare to one that prioritizes wellness of populations [6], various authors have highlighted cross-sectoral collaboration as a key feature of the population health approach [7][8][9][10]. In particular, the US Institute of Medicine (IOM) [11] and others [12][13][14][15] have noted the role of collaboration between primary care (PC) and public health (PH) sectors in achieving lasting improvements in health outcomes. In their report on PC and PH integration, the IOM point out that pressing health needs within populations such as management of non-communicable diseases, maternal and child health, and cancer prevention fall within the scope of both PC and PH, yet these sectors are largely functioning independently of each other.
Importantly, various jurisdictions around the world have developed integration efforts involving PC and PH. Across Europe, for example, as part of the Health 2020 European Policy Framework, several countries are implementing collaborative models of healthcare and strengthening PH capacity in clinical settings [16]. Furthermore, with the introduction of Clinical Commissioning Groups (CCGs) in the UK in 2013, healthcare underwent a significant shift towards integrated, population-based delivery [17,18] Additionally, in Canada, a number of provinces are developing initiatives in line with community-based priorities and population needs-based funding models to promote health of local communities [19,20]. However, even in these jurisdictions, the attempts at collaboration are, at best, fragmented across the respective health systems [3].
There are, however, some concerns that merging clinical care and PH resources will lead to the "tyranny of the urgent"the idea that demand for acute PC services will take precedence over PH needs for resources and investments that only show results in the long term [21]. Elsewhere, PH academics and practitioners have pointed out that combining their sector's activities with PC and community care will limit the scope of PH, if due consideration is not given to different functions and expertise of all sectors involved [22]. For these reasons, we need to understand what such a transformation involving clinical care and PH would look like practically.
Previous reviews on PC and PH collaboration have highlighted different aspects of partnership while also advocating for greater interaction. The IOM report outlines core principles necessary for effective integration [11], Levesque et al. make note of overlapping functions and activities of both domains [23], and Martin-Misener et al. identify facilitators and barriers of PC-PH collaboration at systemic, organizational and interpersonal levels [24]. Furthermore, an Institute for Healthcare Improvement project on clinical-community linkages identifies high-level foundational steps involved in collaboration [25]. Significantly, they conclude that there are various conceptual models and principles in the literature, but these alone are not sufficient for healthcare practices seeking to design collaboration efforts for a defined geographic population. Similarly, these reviews have not found any significant efforts to produce practical tools for frontline clinicians, PC practices, and PH units to achieve collaboration.
We conducted a scoping review to identify specific interventions adopted by healthcare organizations internationally that resulted in collaboration between clinical care and PH. By analyzing empirical examples of integration, we aim to categorize these interventions into a resource healthcare practices can utilize to self-evaluate their practices and identify opportunities for collaboration with PH.
Given the strong interest across jurisdictions to address inequities in care, we further hope our findings can be utilized by healthcare practices to match health services to unique population health needs and thereby target underserved, high-risk populations groups.

Methods
We conducted a scoping review using an updated version of Arksey and O'Malley's [26] scoping study methodology developed by Levac et al. [27] A scoping review is a form of knowledge synthesis that utilizes narrative integration to systematically chart relevant evidence in research related to a broadly defined area and map the results according to key issues or concepts [26,27].

Search strategy
MEDLINE, EMBASE and PsycINFO electronic databases were searched from 1990 to 2017 for articles containing key words or MESH terms related to "Public Health", "Primary care", "Population health approach", "Collaboration", and "Integration", along with all related terms. This was an iterative process and the search strategy was refined several times by combining different key words using the Boolean operators "AND" and "OR". Our combined searches yielded 2375 articles (after duplicates were removed). After screening titles and abstracts for relevance to PC and PH collaboration or integration, 366 articles were retained.

Study selection and inclusion criteria
We restricted our search to studies in English, published between 1990 and 2017, and published in a peerreviewed journal with any of the following study designs: randomized controlled trials, cluster-randomized controlled trials, cluster-controlled studies, observational cohort studies, case reports and series, project evaluations, and review articles. Furthermore, studies were included in our scoping review if the reported collaboration was based in Canada, United States, Western Europe, Australia and New Zealand to allow comparability of health systems in terms of training, range of clinical governance and funding mechanisms [24]. Finally, similar to the approach by Martin-Misener et al. [24], we limited our review to studies that reported on structures, processes, outcomes or markers associated with PC-PH collaboration.
A web search was conducted to review the grey literature and we retrieved relevant reports published through websites of various associations and research networks. The reference lists of identified articles were reviewed for additional sources. Overall, the study selection criteria and reference lists yielded a total of 45 articles for final analysis.

Data extraction and synthesis
Data were extracted from the final studies using a structured abstraction process (forms available on request from the corresponding author, M.S.). Standardized forms were created and entered into Excel by one author (M.S.) and data allocations were independently reviewed by two other authors (A.D.B., and R.U.) and verified for completeness. Discrepancies and methodological quality of selected studies were discussed and resolved through consultations among three authors (M.S., A.D.B., and R.U.).
Applying a narrative approach [27], extracted data included year, location and context of where collaboration occurred, research methods of authors/organization, objectives and purpose of collaboration, specific intervention leading to collaboration, and description of intervention (structures, processes, outcomes or markers of collaboration). Furthermore, we utilized a series of determinants for collaboration developed by Roz Lasker and the Committee on Medicine and Public Health [28] to guide this data extraction process and the categorization of studies included in our review. Specifically, the classification process involved an adapted version of Lasker's models of Medicine and Public Health Collaborations, which the authors reported as 'Synergies'. Lasker defines these synergies as combinations of resources and skills utilized by professionals in medicine and public health, along with other partners in the community, to allow effective cross-sectoral collaborations [28].

Results
Of the 2962 studies identified by the search strategy, 45 met our inclusion criteria and were selected for the final analysis, as illustrated in a PRISMA flowchart outlining the search and screening process involved in selection of articles for the scoping review ( Fig. 1).
The final 45 studies that described interventions leading to PC-PH collaborations were classified based on an adapted version of Lasker's models of Medicine and Public Health Collaborations, which the authors reported as 'Synergies [28].
We focused on the following four synergies: 1. Coordinating healthcare services (Table 1); 2. Applying a population perspective to clinical practice ( Table 2); 3. Identifying and addressing community health problems (Table 3); 4. Strengthening health promotion and health protection (Table 4).
Importantly, these combinations encompass all health professionals and organizations across both health sectors, and also involve various areas of PC and PH including clinical practice, health policy, education and research [28].

Synergy 1: coordinating healthcare services
It has been well documented in the literature that a core strategy for promoting cross-sectoral collaboration between clinical care and PH is the coordination of healthcare services. Thirteen papers included an intervention leading to PC-PH collaboration that was categorized under Synergy 1.
Interventions     2. PC and PH professionals involved in these forms of public policy advocacy can draw upon various non-clinical tools/resources such as scientific/technical expertise, lobbying and public relations skills, and influence with policymakers and the public in order to influence regulations that promote conditions more conducive to safety and well-being of populations C. Launch "Healthy Communities" Initiatives 1. Initiatives can be targeted at particular health problems or needs specific to groups within the community 2. PC-PH Initiatives can also be developed on a larger, community-wide scale through multiple projects designed to promote health of local populations 3. These projects go beyond categorical health promotion activities by developing a broad-based process to tackle multiple community health issues, and evaluative mechanisms to determine outcomes and benefits of institutional investments in community health status improvement b. PC pediatricians can conduct environmental health assessments as part of their standard care practices for patients with asthma, obesity or repeated injuries (assesses for presence of structural hazards in the home that can contribute to development of asthma and lead poisoning); c. To tackle childhood obesity, healthcare professionals can advocate for communitybased initiatives to promote walkability and cycling, and increasing availability of fresh fruits and healthier food choices in stores close to school areas; d. PC/PH professionals can inform patients/ families about community resources for mental health, and should also advocate the need for comprehensive care models for mental health and substance use in children and adolescents; e. PC pediatricians can also advocate reducing children's exposure to violence and collaborate in policy/legislative circles to promote public awareness around impact of violence exposure on children's development [67]. C. Launch "Healthy Communities" Initiatives (n = 7)community-wide projects that bring the public, private, and non-profit sectors together to develop solutions to community health problems [14,17,28,41,45,68,69]. 1. Gosling et al.
-PC-PH Initiatives can be developed on a community-wide scale, as demonstrated by the "Healthy Liverpool Program" in the UK which implemented a series of projects to promote city-wide population health: a. "Smoke-Free Liverpool" is a project that takes a multi-pronged approach by advocating for local smoke-free legislation, encouraging voluntary smoke-free workplace policies, and providing comprehensive smoking cessation services to city residents; b. Projects also addressed underlying causes of health problems. E.g. PC practices in identified low Vitamin D levels in a Somali population and designed health promotion activities and screen-and-treat programs targeted at these groups; c. Projects such as the "Healthy Homes Initiative" tackle broader determinants of community health; this initiative involves collaboration between PC, environmental health, and housing services to mobilize community resources and provide at-risk populations with safe and affordable housing options; d. Low-income patient groups facing financial hardship can be referred to the Citizens Advice Bureau by GPs as part of the "Advice on Prescription" project to offer individuals with advice on welfare, benefits, debt and housing [14].

Pickens et al. -
a. Parkland Health System's Community-Oriented PC (COPC) model in Dallas TX is organized around six key elements: [1] assessment of community needs/assets, [2] community prioritization of healthcare issues, [3] collaboration with community groups, [4] provision of primary healthcare, [5] evaluation, and [6] financing; b. Particularly noteworthy is their development of the Community Health Improvement Measurement and Evaluation System (CHIMES) to document areas where community health investment has produced savings (E.g. fewer hospitalizations, fewer days lost from work, less school absenteeism etc.) [41].
Two additional areas for collaboration noted in Lasker's report are 'improving access to care for the uninsured' and 'collaborating around policy, training, and research [28]. As highlighted through Lasker's models of Medicine and Public Health Collaborations, the former can be implemented through various interventions such as establishing free clinics and referral networks, increasing clinical staffing at public health facilities, and transferring uninsured patients to mainstream medical settings. The latter model can be achieved through notable interventions that include engaging in crosssectoral education and training, as well as conducting cross-sectoral research. Although the studies in our review did not address these areas consistently enough to warrant their inclusion under separate synergies within our findings, these interventions can be of value for practices developing integration initiatives.

Discussion
Strengthening respective functions of clinical care and public health sectors through collaboration Each of the examples of interventions identified in the literature feature a conceptually novel aspect of healthcare deliverycross-sectoral collaboration through coordination of a range of services traditionally provided independently of each other. By linking clinical services to those of other sectors such as PH, as highlighted in synergy 1 (coordinating healthcare services), healthcare practices can enhance patient follow-up and improve health outcomes [11], reduce duplication of services, and achieve economies of scale by centralizing services across PC and PH sites [11,28,30].
Several interventions found in our study largely focus on the clinical end of the PC-PH collaboration spectrum. However, as collection and analysis of population-based information are core functions of PH, it is also important to touch upon the benefits to this sector. Synergy 2 (applying a population perspective to clinical practice) offers strategies that can be useful and cost-effective for PH agencies collaborating with PC. For example, integration can facilitate translation of epidemiologic findings developed from population health tools into the clinical practice setting. Additionally, these integration efforts can ensure that patients identified by PH screening programs receive appropriate referral and follow-up services [48,49]. Furthermore, integration efforts under Synergy 3 (identifying and addressing community health problems) also target opportunities present in clinical care to advance broader communitybased goals of PH.
Significantly, as shown consistently across several interventions and most notably under intervention 3B (using clinical encounters and sharing data to build community-wide databases), collaboration between PC and PH sectors has been greatly enhanced through recent advancements in health information technology (IT) [14,31,43,44,46,[58][59][60][61][62] and the growth of population health informatics [70]. Widespread adoption of EHR systems across hospitals in countries such as the US, along with greater involvement of hospitals and academic health centres in health IT exchange initiatives has facilitated community needs assessments, design of population-level health interventions, and tracking of health outcomes in specific patient populations [43,44,71].
Another important finding is that interventions across our review findings are not mutually exclusive and often overlap in the empirical studies identified. In particular, interventions under Synergy 4 (strengthening health promotion and health protection) often incorporate strategies from other models of collaboration which, importantly, highlights the mutually reinforcing nature of the synergies and the value in employing a range of interventions to achieve integration. Cross-sectoral efforts targeting health promotion and protection are becoming increasingly important in the current healthcare environment with rising burden of illness and disability due to health problems involving violence, substance abuse, chronic disease and environmental hazards, among others [3,28]. Efforts at addressing such health problems associated with a variety of social, behavioral and environmental risk factors will have to go beyond the traditional biomedical model and employ a populationbased approach to promote the health of communities.

Addressing inequities in care
From a health equity perspective, our review findings also provide several empirical examples of interventions that address inequities in care by targeting underserved, high-risk populations groups.
Under intervention 1C (establishing "one-stop" centers), for example, the Swiss-Hungarian Cooperation Programme developed a general practitioner-centered cluster model for community-oriented PC services in the most disadvantaged regions of Hungary to provide low-income patients access to various new services such as health status and risk assessments, lifestyle counselling, and chronic care rehabilitation [38]. Additionally, within intervention 2B (using population-based strategies to "funnel" patients to medical care), Heller et al. describe a community-based mobile PC clinic that aims to increase access to healthcare services for underserved populations in Maryland through community-wide screening and secondary referrals to specialists for patients with complex conditions [48]. Also under the same intervention, we found initiatives targeting TB patients in homeless shelters through implementation of CDC-recommended mass-screening protocols [52], as well as Rural Mobile Health Units designed for elderly, atrisk populations with limited access to PC services [50].
Furthermore, under intervention 3B (using clinical encounters and sharing data to build community-wide databases), one collaboration effort utilized shared EHR data, such as patient demographics, to identify disparities in Hemoglobin A1c levels of diabetic patients which resulted in creation of the Enhanced Diabetes Care Model for certain racial and ethnic groups experiencing barriers to care [43]. A notable example of addressing inequities in care within intervention 4C (launching "Healthy Communities" initiatives) is provided by Gosling et al. where authors highlight an effort by PC practices in Liverpool, UK to identify low Vitamin D levels in a Somali population and design health promotion activities and screen-and-treat programs targeted at these groups [14]. As part of this same city-wide project, and with a focus on the broader determinants of community health, the "Healthy Homes Initiative" involves collaboration between PC, environmental health, and housing services to mobilize community resources and provide at-risk populations with safe and affordable housing options. Lastly, in an example of a collaborative community-oriented PC model in Dallas TX, authors reported significantly better neonatal mortality outcomes, than those for the US overall, for African American and Hispanic populations residing in Texas [41].
Implications for policy and practice: a population-based approach to address community health needs Previous reviews on PC and PH collaboration have highlighted various aspects of developing cross-sectoral partnerships such as core principles for effective integration, facilitators and barriers, and overlapping functions and activities of both domains [11,[23][24][25]. However, these reviews have not found any significant efforts to produce practical tools for healthcare practices to enable implementation of collaboration. Our scoping review identified a range of interventions designed by organizations in health systems around the world that facilitated cross-sectoral integration. We subsequently categorized these interventions into Lasker's synergy groupings which are combinations of resources and skills required to achieve collaboration [28].
Importantly, our review findings have been organized into a self-evaluation tool that can serve as a resource for identifying opportunities for cross-sectoral collaboration and allow practices to focus on pressing health needs facing their communities. More specifically, to facilitate the brainstorming process, clinicians and PH practitioners can utilize the tool to address questions such as the following: 1. What services are currently being provided by our practice? 2. What are the current health and service needs in our community and which of these listed interventions address our identified gaps? 3. Which of these interventions are useful and feasible for adoption in our practice and what are the necessary steps required to pursue collaboration through these identified interventions?
This tool can be found at the end of the article in Additional file 1: Self-evaluation tool for clinical practices to identify opportunities for collaboration with public health. The scoring systems within this tool have been adapted from 'Self-assessment tool for the evaluation of essential public health operations in the WHO European Region 2015' [72] and can be useful for practices in determining the level of progress made with respect to each intervention as well as identifying key areas for improvement for achieving collaboration.
One of the benefits of the population health approach is identifying community health needs and reorienting health care delivery and services to these needs [11,17,38]. Given the strong interest across jurisdictions to match health services to population health needs, the tool can be utilized by healthcare practices and organizations as part of their criteria for selecting community-based projects to move collaboration forward with. Further, by using the list of interventions outlined in our findings, both sectors can also come together to design interventions tailored towards underserved, high-risk population groups.
In providing healthcare organizations with a range of interventions implemented across different jurisdictions, our tool can be utilized to initiate practice-level discussions around implementation of integration efforts that are useful and feasible for all sectors involved. We hope this self-evaluation process will facilitate development of practical action steps required for practices to operationalize collaboration towards a population-based approach to integrated healthcare delivery.

Limitations
Firstly, very few studies applied rigorous evaluation methods or reported on measures of successful collaboration. There is considerable difficulty in establishing acceptable outcomes and investments in community health are often only realized in the long term in other social and economic sectors [21,41]. Further, the outcome measures that were reported by authors were often not robust. This may in part be due to the difficulty in specifying and measuring community care outcomes, and also because most collaborations noted in the literature were in the early stages of their development, making available outcomes limited. There is also inconclusive evidence to suggest which of the identified interventions are particularly exemplary in terms of impact or ease of implementation. Additionally, the interventions outlined in our findings vary in terms of the resources and activities needed for collaboration. For these reasons, we are limited in our ability to compare the effectiveness of adopted interventions and assess which programs are sustainable in the long term.
It is also important to note that the role of the PH sector in our findings is limited in terms of its scope and definition. The listed interventions tend to highlight the clinical aspects of PH and their functions. In doing so, they reveal certain gaps within the current literature on PC-PH collaboration. Examples of additional areas within PH that should also be considered include collaboration with local municipalities and addressing the other social determinants of health such as education. Importantly, the US IOM report on integrating PC and PH highlights that there is a growing evidence lending support to the value of undertaking such activities to address the full scope of PH functions. It is worth mentioning, though, that Synergy 4 (strengthening health promotion and health protection) does take a step towards bridging this gap by incorporating a wider range of PH functions such as health promotion initiatives through a focus on education, as well as creation of community-wide projects to involve public, private and non-profit sectors in developing solutions to community health issues.