There is a growing number of patients living with 2 or more chronic conditions in Canada and worldwide [1,2]. These individuals not only manage multiple chronic conditions [1] but also experience additional biopsychosocial challenges [3,4]. Living with multiple chronic conditions can introduce additional complexity into this patient population’s care management experience, such as high symptom and treatment burden [5,6]. In addition, this patient population can have complicated and contradictory treatments or medication regimens [7] that require care from multiple health professionals and services spanning various health and social care settings [8]. Owing to the multitude of challenges this patient population faces in the current disease-specific care model [9], this patient population is often referred to as patients with complex care needs or complex patients. Navigating a fragmented care system can leave patients with complex care needs feeling overwhelmed and dissatisfied with their care [10,11]. An agile and coordinated care system can better address this population’s challenges.

Integrated care models that involve primary care are well positioned to support people living with chronic conditions as most chronic disease–related care is delivered in this setting [12]. The definition of integrated care is a coordinated, collaborative, multidisciplinary, and person-centered care delivery system [13]. In recent years, integrated care approaches have been shown to support improved care coordination within primary care [14,15]. However, implementing integrated primary care models is an ongoing challenge within health systems.

The Rainbow Model of Integrated Care (RMIC) offers a conceptual way of understanding integrated care, suggesting that integration can occur at different levels (clinical, professional, organizational, and system) and is enabled through functional and normative mechanisms [13]. This validated model is helpful in understanding the different types and degrees of integration that a care model or intervention aims to achieve [16,17]. Each level of integration (clinical, professional, organizational, and system) may require different strategies, tools, and mechanisms to support integration efforts, such as care coordination across health professionals for effective care continuity and coordination of services [18]. Information and communications technologies (ICTs) can enable these types of integration efforts regarding care continuity or shared decision-making by offering their functionalities, such as sharing patient data across professionals and organizations [19,20]. In the health care setting, ICT is used as an umbrella term that includes different types of health technologies [21]. Therefore, to operationalize this term for this study, we have used the World Health Organization ICT definition in the context of health systems, which is as follows: health services and information delivered through the combined use of the internet and other electronic communication technologies [22,23]. ICT-enabled integrated care can be helpful for patients with complex care needs in terms of care coordination and as a decision support tool as this patient population may require multiple support services from multiple health professionals and organizations [18].

Textbox 1 shows the RMIC levels and how ICTs can support each component.

Another way to conceptualize how technology can support patients with chronic conditions is the eHealth Enhanced Chronic Care Model (eCCM). This model builds on the foundational Chronic Care Model introduced by Wagner [32] that has been used to guide chronic disease management for patients with complex care needs in primary care and other settings [33]. The eCCM version offers a salient way to identify how ICTs and ICT-generated information can aid in improving patients’ health outcomes by increasing patients’ and providers’ knowledge of chronic disease management [21]. Textbox 2 shows the key components of the eCCM.

Taken together, the RMIC and eCCM offer useful suggestions on how ICT may be of use to enable integrated primary care delivery for patients with complex care needs. Although these frameworks and emerging literature in the field of digital health and care integration provide some evidence of how technology may be of service to patients with complex care needs, there has not been a comprehensive systematic review of the literature focused on how ICT supports integrated primary care models to identify current trends or gaps in research. Therefore, the purpose of this knowledge synthesis was to scope the current literature on the types of ICTs that are being used to support patients with complex conditions in integrated primary care settings while also identifying potential areas for continued research.

A scoping review approach was used to systematically map and identify relevant literature at the crossroads of integrated primary care models, ICTs, and patients with complex conditions. Scoping reviews capture a breadth of literature, which can effectively identify emerging literature as well as evidence gaps on a topic of interest [34]. Given that ICT-based integrated primary care is comparatively a newer model of care, a scoping review is an ideal approach for this study. To conduct a rigorous review, we followed the six stages of the scoping review framework by Arksey and O’Malley [34] enhanced by the work by Levac et al [35]: (1) identifying the research question; (2) identifying relevant literature; (3) selecting the studies; (4) charting the data; (5) collating, summarizing, and reporting the articles; and (6) disseminating knowledge. Each of these 6 stages is described in detail in the following sections. For further details, the study protocol has been published elsewhere [36].

The guiding research question for this scoping review was as follows: What ICTs are used in the delivery of integrated primary care to patients with complex care needs? The subquestions were as follows: (1) Which technologies are being used for this patient population? What are the functionalities and characteristics of these technologies? (2) How are these technologies being used in the integrated primary care model in terms of care integration? (3) What implementation factors are being reported in the selected studies that describe ICT-enabled integrated primary care models?

These questions were selected as there has been no systematic documentation in the current literature of what sort of ICTs are available to support an integrated primary care model or how ICTs are enabling the process of care integration in primary care settings. In addition, the third subquestion was posed given the importance of implementation strategies in the successful use of ICT-based interventions in the integrated primary care model [37]. Given the interconnectedness of the primary research questions and the 3 subquestions, a single search strategy and review were determined to be sufficient, and addressing all 3 questions in a single study was expected to lead to a richer and more meaningful analysis and findings.

To operationalize the research questions, we identified and defined 4 major concepts to inform our search strategy. The 4 concepts and their key definitions are outlined in Textbox 3. Multimedia Appendix 1 shows the keywords associated with each concept for each category.

Relevant studies were identified by searching the following electronic databases of published literature: Ovid MEDLINE, Ovid Embase, EBSCO CINAHL, and Ovid PsycINFO. The MEDLINE search strategy was developed first and peer reviewed using the Peer Review of Electronic Search Strategies tool. This search strategy was then translated to the remaining 3 databases. The searches were limited to articles published between January 2000 and December 2021. Articles published before 2000 were excluded as a preliminary search showed that there were very few articles published on ICTs during this time. The MEDLINE search strategy was included in the published protocol [36]. In addition, the search strategy for PsycINFO can be found in Multimedia Appendix 2. The initial search of all databases was completed in the summer of 2019 and then updated in December 2020 and again in December 2021.

The search results were imported into the knowledge synthesis software Covidence. The inclusion and exclusion criteria were imported into the software as well. The inclusion and exclusion criteria were decided a priori and were used for screening citations first, for the title and abstract review, and then again for the full-text review. Both levels of review were conducted in duplicate by 2 reviewers.

To be included in the review, articles were required to report on an intervention that (1) had an ICT-enabled health care model, (2) was based on an integrated health care model or team-based care, (3) included adult patients with complex care needs, and (4) took place in a model that included primary care. If the intervention had a target population of (1) individuals aged <18 years (given the focus of this review on adults with complex care needs) or individuals with (2) cancer or (3) mental disorders, it was excluded as cancer care and mental health provision have unique care pathways that may not be translatable into integrated primary care settings [39,40]. Any published article, including quantitative studies, qualitative studies, mixed or multimethods research, both comparative (eg, randomized, controlled, cohort, or quasi-experimental) and noncomparative (eg, survey and narrative audit) methods, educational materials, and reports, could be included in the review if it met the aforementioned inclusion criteria. These inclusion and exclusion criteria were published in the study protocol [36].

To ensure consistency between reviewers, a series of training exercises and discussions were held before commencing title and abstract screening. At first, all 6 members of the research team screened a random sample of 30 articles to assess interrater agreement. Interrater agreement for study inclusion was calculated using percentage of agreement; when it reached >70% across the team, we proceeded to the next stage. If a lower agreement was observed, the team had a discussion about eligibility criteria, and another pilot test was conducted. In total, 6 rounds of pilot tests were required for the title and abstract screening of a random sample of a total of 110 citations. After the initial agreement between the 6 screeners was established, they were divided into groups of 2 for an efficient screening process. The pairs worked together to screen titles and abstracts for inclusion. Any disagreements within pairs were resolved through weekly group discussions.

For the full-text screening stage, 3 rounds of pilot tests were conducted on a random sample of 30 articles. All 6 team members participated in the pilot-testing round. In each round, interrater agreement was calculated using percentage of agreement. In the third round, we reached >70% consensus across the team. Subsequently, the 6 team members were divided into 3 groups consisting of 2 group members each. All 3 pairs screened the full texts of potentially relevant articles to determine inclusion using the inclusion and exclusion criteria. All discrepancies were resolved through weekly group discussions.

All the included studies were reviewed and charted independently by 6 team members. To ensure consistency between team members regarding data extraction, the entire team pilot-tested the data abstraction form on 3 articles that were randomly selected. Full data abstraction began only after full agreement was reached. The full agreement was decided based on percentage of agreement, and it reached >90% agreement in the first round of pilot-testing. After reaching a full team agreement, each team member extracted the study data independently. The data extraction form can be found in the published protocol paper [36]. Subsequently, the studies were divided among the team to be abstracted by a single team member and verified by a second reviewer (FT and AA) to ensure data accuracy. Consistent with a scoping review approach, the methodological quality of the included articles was not appraised [41].

The charted data were summarized to gain a descriptive understanding of the data collected. In addition, to answer the 3 subquestions, we used the qualitative content analysis method (Textbox 4).

The study findings have been disseminated in academic conferences [43] and a webinar [44]. A wide variety of stakeholders, such as clinicians, researchers, policy makers, and patients, some of whom provided feedback and reflections on the findings, participated in both academic conferences and the webinar. This manuscript marks another key knowledge dissemination tool for this study.

A total of 52,286 articles were identified. After removing duplicates, 69.76% (36,474/52,286) of the articles underwent title and abstract screening. A total of 293 articles underwent a full-text review, resulting in 31 (10.6%) that met the eligibility criteria and were included in the final analysis. Figure 1 shows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram for this review.

Table 1 shows the full description of each study. The included articles spanned nearly 16 years (2003-2020), with 68% (21/31) published after 2014 (Table 2). Of the 31 included studies, 15 (48%) were from the United States, with the remaining 52% (16/31) being from Canada, the United Kingdom, Spain, Israel, Taiwan, and India. Of the 31 selected studies, 15 (48%) were quantitative, whereas 5 (16%) used qualitative methods, and 3 (10%) used mixed methods research designs.

Most of the interventions (15/31, 48%) [14,53, 55,57,59-63,66,69-71,77,78] only took place in a primary health care setting. However, 39% (12/31) of the studies linked primary care to acute care [30,46-49,52,54,56,65,68], social services [46-48,64], or rehabilitation centers [51].

Of the 31 identified studies, 17 (55%) described the involvement of primary care physicians [47,49,53-55,57-61,64-66,70, 71,77,78]. A total of 52% (16/31) of the studies described the involvement of registered and specialized nurses, such as geriatric nurses [30,48,49,52,54,55,57,58,60,61,65,70,71,73, 74,78]. In addition, 23% (7/31) of the studies described the involvement of nonclinical members such as care coordinators and care managers [46,47,49-52,68].

Aligning with the RMIC goal of improving the patient experience by pursuing the triple aim of health care, the studies reported outcomes related to the triple aim, such as quality of life (improved health), cost-effectiveness (lower cost), and acceptability of the care delivery system (improved care). Outcome measures included patients’ health outcomes (eg, mental and physical conditions), patient and provider perceived outcomes (eg, satisfaction), and cost-effectiveness. Patient experience measures included satisfaction, perceived empowerment, caregiver strain, loneliness, self-efficacy, acceptability, and usability of the technology [19,30,48-51,68-70,72,78]. Cost-effectiveness included costs associated with hospital use and cost-efficiency [30,47,48,51,52,59-61,64,65,73,78]. We did not observe any pattern in the type of outcome or technology.

In total, 68% (21/31) of the studies evaluated outcomes. A total of 45% (14/31) measured selected clinical outcomes such as quality of life (6/31, 19%) [47-49,70,73,78,80], mental or physical conditions [49,69], and hospital use (10/31, 32%) [30,47,48,52,60,61,64,65,73]. Among the nonclinical outcomes, provider workload [50], patient-provider relationship [49], patient satisfaction [30,48,69,80], cost-effectiveness [51,65,78], and attitude toward technologies [50,51,53] were studied. In contrast, patients’ self-efficacy (3/31, 10%) [48,53,78] and acceptability of the technologies (4/31, 13%) [50,51,53,72] were explored mostly in mixed methods and qualitative studies. The cost-effectiveness (4/31, 13%) [51,65,78] of the implemented technologies was evaluated quantitatively and through mixed methods research.

The implementation of technology has been identified in the digital health literature as a key issue that requires further exploration [81]. Thus, we extracted the factors associated with the successful implementation of technology in the models of care described in the studies. Inductive analysis of the studies revealed implementation factors of technologies that were categorized into the patient, provider, technology, and organizational level. Table 5 shows the implementation factors discussed in each study.

Patient characteristics that were described as important implementation factors in the selected studies included age, gender, type of chronic conditions [46,64], patients’ interest in buying into the technology [14], digital literacy [59,63], and perceived ease of use and usefulness [66]. At the provider level, their willingness to adopt digital technology in their practice [55,59,71,80], collaboration and communication among providers [56,57], and providers’ level of ICT training [64] contributed to the implementation of ICT-enabled integrated primary care. Barriers to implementation arose when technology disrupted provider workflow [55,69,71] and increased their workload [71].

Some implementation factors spanned more than one category. Notably, the patient-provider relationship emerged as an important factor [52,55]. For example, Berry et al [52] signaled that technology-enabled care works better when there is a long-established trusting relationship between the patient and the provider.

In addition, technology- and organizational-related implementation factors emerged. Technology-related factors such as system errors [54,55,63,71], lack of interoperability with existing technologies (such as EMRs) [55,69,71], and technologies’ lack of responsiveness to the patient and providers’ needs and tasks were found to deter implementation in 16% (5/31) of the studies [50,55,69,70,78]. At the organizational level, funding support [65] and organizations’ willingness to adopt a new technology [14] were reported as important organizational-level implementation factors. Moreover, having a coordinator to support IT-related issues [63], having a team structure that involves nurses [59], and having additional care coordination [52] were also reported as important implementation factors for ICT-based integrated primary care.

The primary objective of this scoping review was to provide a comprehensive overview of the types of ICTs that are being used to support patients with complex care needs in the integrated primary care model. The findings show that, in the current ICT-based integrated primary care models, wide-ranging technologies such as telemonitoring devices, mHealth apps, EHRs, and telephones are being used to support patients with complex care needs. These technologies enable information sharing among multiple professionals within and across organizations and between patients and providers. These technologies support care integration at the service or clinical, professional, and organizational levels. However, system-level integration was absent in this study. The findings also show that patient-, provider-, technology-, and organizational-level factors contribute to the implementation of ICTs in the integrated primary care model.

Among the identified studies, we observed a number of ICTs that enabled clinical information management across teams in the integrated primary care models. These ICTs promoted teamwork across professionals by providing access to critical patient information across the teams. This shared patient information system is considered a critical tenet of an integrated primary care model [13,18]. Moreover, ICTs that supported patient self-management were also frequently observed. This finding is unsurprising as a previous study identified that, given the nature of chronic conditions, self-measurement of the body is frequently used to track symptoms [82]. This self-management approach to chronic conditions can empower patients by engaging them with their care plan, which in turn can improve patients’ health outcomes [83]. In contrast, too many self-management responsibilities can overwhelm patients, especially if the patient has a low capacity (ie, financial resources) to shoulder the responsibilities [10]. For example, a number of ICT-based interventions in this review required daily reporting of patients’ vital signs and symptoms [51,59-61,64,69]. For patients with complex care needs, this daily reporting of vital signs may be overwhelming because of the number of conditions and complexity related to each condition. Therefore, it is critical to examine how to effectively use and design ICTs to minimize the self-management–related workload for patients.

The findings of this study show that a number of ICTs transmit patient data to providers for remote disease monitoring. However, most of these ICTs lack a 2-way communication loop. As a result, providers may receive an overwhelming amount of patient data, but engaging in a productive interaction may be difficult for both the patient and provider without a 2-way communication loop. This is problematic as the quality of the eCCM relies strongly on productive interaction and communication between providers and patients [21]. Therefore, an increased effort to promote meaningful and productive interaction between patients and providers can improve the effectiveness of ICT-based interventions designed for patients with complex care needs [21].

This scoping review revealed several gaps in the current literature. System-level integration and the normative aspect of integration were absent in the sample of studies in this review. Undertaking an ICT-enabled system integration can be challenging because of interoperability issues, provider unwillingness, lack of coalition among stakeholders, lack of data privacy, and security issues [84,85]. For future researchers, it would be important to explore the broader system-level changes that are required for the successful adoption and implementation of ICTs in the integrated primary care model.

In addition, normative aspects of integration, such as culture, shared vision, and regulatory factors, need to be studied further for the long-term sustainability of ICT-based integrated primary care. Evaluation frameworks used in implementation science [86], such as the Consolidated Framework for Implementation Research [87] and the nonadoption, abandonment, scale-up, spread, and sustainability framework by Greenhalgh et al [76], can be useful to explore the normative factors that can contribute to the implementation of ICT-based interventions in the integrated primary care setting.

Furthermore, organizational integration was also minimally described in the studies. Organizational-level integration may be difficult because of a lack of shared vision in terms of policy, regulations, and culture across organizations [18]. This lack of organizational integration may negatively affect patients with complex care needs as they may have to report similar information to multiple providers. For example, although ICTs can be helpful in creating an eCommunity that enables access to patient data for community staff and social service organizations, this aspect of the eCCM is currently underused. However, creating an eCommunity where patients’ clinical data are readily available to formal or informal caregivers can reduce the patients’ need to repeat information to multiple providers, thereby reducing the burden on patients and caregivers [28]. We also identified that the definitions of the components of the eCCM are broad, which made the categorization of the ICTs difficult and interpretive. This criticism of the eCCM has been identified in previous work [88].

In terms of measured outcomes, 45% (14/31) of the studies evaluated the impact of ICTs at the patient level through a clinical lens, such as quality of life. Going beyond clinical outcomes, such as health status or symptoms, and considering the impact of ICTs on patients’ relational factors, such as patient-provider relationships [89] or potential disruption of technology in patients’ lives [5], may be important factors to consider for future researchers.

To categorize the study data, we used 2 established conceptual frameworks, the eCCM and RMIC, which helped identify literature gaps. For example, by categorizing the role of ICTs in the RMIC, we identified that there is little information available about ICT-enabled system-level and normative integration. In addition, the use of these established frameworks will be helpful for future researchers from multiple contexts and jurisdictions to interpret and apply the findings to their own contexts. The use of the scoping review method also allowed us to provide a comprehensive overview of the literature on the topic regardless of methodology, quality of the publications, and measured outcomes [41]. This will be helpful for researchers to gain an understanding of the current landscape of ICTs supporting patients with complex care needs in integrated primary care settings.

This scoping review has a few notable limitations. Owing to time and resource constraints, we were unable to conduct a review of gray literature, potentially missing relevant publications. However, the peer-reviewed sample was extracted from a very large pool of potential articles, which adds to our confidence that the peer-reviewed literature was thoroughly covered. Second, we have observed that the term “complex patient” is defined in various ways in the current literature. We may have missed unexpected characterizations of a patient with complex care needs that are not captured in the current literature. In addition, we did not involve patients or the public during the conceptualization of the research question. However, we tackled this issue by disseminating our research findings among different audiences (conferences and webinars) that were accessible to the public.

This scoping review identified that multiple types of ICTs play important functional roles in supporting patients with complex care needs in integrated primary care settings. ICTs play a critical role in coordinating multiple clinical services across multidisciplinary teams and supporting the self-management of patients with complex care needs remotely. However, to use the full potential of ICTs, further studies are required to understand how to achieve organizational- and system-level integration.