This paper is the first to investigate and report factors influencing implementation of virtual care interventions in an Australian primary care context using the CFIR framework, with several key findings. Firstly, no included papers utilised implementation frameworks to structure their evaluation of the implementation of virtual care. Secondly, there was a notable imbalance of patients perspectives in the included studies. Third, CFIR helped identify a cluster of common influencing factors including: innovation relative advantage, information technology infrastructure and capability, however, it was necessary to develop new constructs to capture the unique factors influencing the implementation of virtual care interventions.
While multiple theories and framework exist to guide the identification of barriers and enablers to implementation (such as the ones included in this study), no paper referred to, or included such frameworks into the method of the study. This is surprising as a recent scoping review which sought to understand the use of implementation theories, models and frameworks across digital health interventions found 156 studies published between 2000 and 2022 that used them for either data collection, data analysis or identification of factors influencing implementation [9]. Further, another scoping review that analysed the use of CFIR in telehealth implementation located 64 studies [27]. The narrow context of this review (Australian Primary care settings) potentially contributed to this finding. It is encouraging that more broadly, implementation science is being used to evaluate virtual care interventions, but evidently, further uptake needs to be encouraged in Australian Primary care settings.
In addition to a lack of implementation theory use, there was a distinct lack of perspectives from patients, with only two of fourteen included studies reporting such perspectives within general primary care and chronic kidney disease contexts [17, 23]. The absence of patient perspectives could be due to difficulties with patient recruitment, specifically in a primary care setting, which have been lamented by numerous studies [28–30]. Several factors are believed to influence the difficulties with recruiting patients including: a lack of awareness of the study, lack of interest in research (in general), perceived burden for the patient (time), interference with usual care (recruitment during appointments) and practice characteristics [29]. Other concerns centre around a reliance on the General Practitioner to recruit and the factors that influence their motivation (attitude to research in general, incentives, relationships with academics) and capability (lack of time and resources) [28]. One study that iteratively modified recruitment strategies to determine the most effective methods found that strategies that removed reliance on practice level involvement had the greatest impact on improving recruitment of patients [30]. In future, studies should seek to include patient perspectives and when recruiting, adopt a holistic approach (one that includes strategies at individual, outer setting and inner setting levels) to ensure adequate representation.
Despite a lack of implementation theory use and suboptimal patient inclusion, the study revealed a cluster of common factors influencing implementation of virtual care in primary care contexts, including Innovation Relative Advantage, Information Technology Infrastructure, Capability, Accessibility, and Suitability. Innovation Relative Advantage was an influencing factor in 14 papers and is defined as ‘the innovation is better than other available innovations or current practice’[7]. The advantages of virtual care solutions over usual systems of care in primary care, and how these are perceived and experienced by patients and clinicians are an important factor to successful virtual care implementation. Clinician’s perceived virtual care advantages often described improvement to patient care and their workflow, such as increased efficiency of care [17, 18, 22], access to a comprehensive suite of patient information [16, 20] and accessibility for disadvantaged patient groups [16, 21, 25, 26] which assisted in facilitating implementation. This is in line with international literature, where reduced transmission of contagious diseases and continuous monitoring of patients have been cited as advantages [31–33]. However, such perceived advantages are constrained by clinician’s concerns surrounding clinical liability, and the validity and quality of care administered and received using virtual care [15, 18]. Such issues could be addressed through the provision of data demonstrating positive patient outcomes and satisfaction to clinicians when attempting to implement virtual care [34]. Patients must also see the advantages of virtual care over traditional modalities to encourage engagement. Such advantages include the convenience, no transport costs, and no lost time off work [19, 26]. Despite the clear benefits, they are not always enough to convince patients that virtual care is the same or superior to face to face consultations [35]. Efforts to ensure that advantages are clearly communicated to both clinicians and patients when introducing virtual care should be made, whilst also making allowances for those who prefer face to face consultations [36].
Evidently, communicating and demonstrating how virtual care interventions provide benefit to patient care and clinician’s workflow is important to support successful implementation and engagement amongst clinicians. However, it is essential that this perceived relative advantage can be actualised by the virtual care solution in practice without flaws/mistakes/inconveniences, lest it fail to be accepted as an appropriate addition or replacement to existing systems of care [33]. Information Technology (IT) Infrastructure was an influencing factor in 12 papers and is defined as ‘Technological systems for tele-communication, electronic documentation, and data storage, management, reporting, and analysis support functional performance of the Inner Setting’. Functional IT infrastructure for patients and health care professionals has been readily identified in the literature as a key factor to successful implementation of virtual care interventions [37–39]. However, this remains a prominent barrier to virtual care implementation in Australian primary care settings, with issues concerning internet connectivity and coverage [13, 14, 22], lack of appropriate technological infrastructure [16, 19] and audio and visual technical difficulties [16, 17] being frequently reported. In contrast, sufficient infrastructure, equipment, and regular network updates, bolstered virtual care implementation, particularly in rural and remote areas [14, 21, 26]. Further, virtual care innovations’ inability to allow clinicians and patients to access data remotely was a barrier to implementation, as this hindered efficiency and continuity of care [18]. Pre-implementation, it is essential that organisations work to ensure that there is working and appropriate infrastructure in place to support the use of virtual care modalities [40]. Thorough testing of equipment and maintenance of internet access infrastructure will help ensure a smooth transition. In addition, ensuring selection of virtual care products that are preferably interoperable with a wide range of practice management suites (including mHealth apps, videoconferencing) that are user friendly and fit for purpose should also support successful implementation [41].
In addition to ensuring there is adequate and functional IT, clinicians and patients need the requisite skills to engage with virtual care. Capability was an influencing factor in 12 papers and is defined as ‘the individual(s) has interpersonal competence, knowledge, and skills to fulfill the role’. The review found that clinician’s previous experience with various modes of virtual care supported successful implementation due to increased confidence and the ability to problem-solve [21, 22, 24, 26]. Such findings mirror international literature, where the capability and self-efficacy CFIR constructs are consistently identified as factors influencing virtual care implementation, as health care professional’s positive self-beliefs regarding their capability to use virtual care solutions with their relevant clinical population facilitates implementation[31–33]. Patient’s digital competency, and certain demographic’s capability of engaging with virtual care interventions was also perceived as a barrier to virtual care implementation [14, 18, 19]. Hardie et al. identified that patient age and language were determining factors of patient’s telehealth capability, in line with international findings that identify geriatric and CALD (Culturally and Linguistically Diverse) populations as often requiring additional assistance and education to engage with virtual care [42]. When mandatory training for a mental-health focused intervention was provided, mental health clinicians found it helpful in delivering the service [13]. However, it was recognised across several settings that further training was needed to educate both clinicians and patients on the use and clinical utility of virtual care innovations to facilitate proper implementation [13, 14, 22]. A literature review by Edirippulige and Armfield (2017) concluded that as the introduction of telehealth necessitates a change in clinical practice, this must be supplemented with appropriate education and training within the workforce [43]. ECHO COVID, a Canadian COVID-19-specific telehealth education program improved clinician telehealth self-efficacy when implemented in a primary care setting, demonstrating that education and training may be a method of increasing adoption [44]. Training can be operationalised through the use of existing frameworks that support competency development and provide a structure to ensure comprehensive education is provided such as the one posed by Hilty and colleagues [45]. The type of training received also impacts telehealth adoption among patients and endeavours should be made to accommodate their varying cognitive capabilities, as failure to do so may result in decreased engagement [17, 46]. A holistic approach to training program development that includes both face to face and hybrid modalities should be used to ensure maximum benefit for those patients regularly involved with using telehealth (e.g. those in rural areas and limited access).
In addition to the above CFIR influencing factors, there were issues that emerged which were not adequately covered by existing constructs, requiring the development of additional ones. Of the eight, two of these introduced constructs, accessibility, and suitability, were identified as key influencing factors. Accessibility, a new sub-construct introduced in the review, was an influencing factor in 11 papers and is defined as ‘the degree to which an innovation as designed can be accessed by everyone’. Accessibility is viewed here as a sub-construct of equity within the innovation domain, which emphasises the importance of access for all. Historically, accessibility is a core tenet of virtual care as such innovations have been utilised as a pathway to care for individuals with physical and geographical factors that impact their ability to access appropriate healthcare [47]. The ability to provide improved access of care for patients, such as those who are immunocompromised, live in rural or remote areas, have a chronic condition or disability, was seen by many clinicians as a major advantage of virtual care [13, 15–17, 19, 21, 24, 26]. Accessibility remains a prevalent facilitator of virtual care implementation in Australian primary healthcare contexts, with a recent study by Clay-Williams et al (2023) including 49 care providers finding Sydney-based clinicians consider patient accessibility a key strength of virtual care [48]. However, the extent of access virtual care can provide is limited to the extent of the surrounding technological conditions the virtual care innovation is implemented in. Clinicians cited poor internet connection or mobile reception in rural areas as a barrier that impacted the value of utilising virtual care [13, 14]. In a Canadian qualitative survey study, rural-practicing clinicians were more likely to cite access to Wi-Fi as a barrier to virtual care usage than urban-practicing clinicians suggesting that there is a greater need to support the establishment of appropriate technological infrastructure in rural primary healthcare services [38]. In addition to the technological infrastructure there also needs to be support through adequate policy that facilitates access to subsidised care. In Australia, the introduction of an Medicare Benefit Schedule (MBS) claim number has greatly assisted with supporting telehealth use [4]. Before the introduction of this claim number, practices were absorbing significant costs associated with telehealth use as there was no facility for reimbursement through the public health system [4, 49]. As a consequence, telehealth was a luxury of the few. Using a lens of equity, it is evident that support from the government is needed to help ensure that those who really need access to virtual care are able to reliably do so, and software providers should make allowances to ensure that those with disabilities can also access virtual care services.
Suitability, a new construct introduced in the review, was an influencing factor in 11 papers, and is defined as ‘the degree to which an innovation is clinically appropriate or useful as delivered to target populations’. Virtual care suitability influenced implementation through patient compatibility with the innovation [16], and clinician’s perceptions concerning the role and utility of virtual care in primary healthcare in the context of providing appropriate clinical care [15]. Suitability posed a barrier to virtual care implementation when there was poor congruence between the patient’s clinical needs and the innovation design [16]. Across several papers, telemonitoring patients with chronic conditions required more extensive in-person support than the innovation could provide and, in some case, resulted in patient populations developing disease passivity or anxiety when exposed to their monitoring data [14, 15]. This suggests the modality was not clinically suitable for certain patients. While determining patient suitability for different virtual care innovations is a complex process [26, 35, 50] the refinement of guidelines for clinicians concerning patient suitability for various virtual care modalities may encourage proper implementation in primary healthcare while improving clinical outcomes and patient experience [51]. In a primary mental health care setting where clinicians considered the clinical appropriateness of patients for participation in a virtually administered mental health care program, clinical risk was averted, and implementation facilitated [13], demonstrating that having a benchmark to ascertain patient compatibility with virtual care may be beneficial. Despite refinements needed in screening patient’s suitability for virtual care, the clinical benefits virtual care contributes to patient care through remote monitoring, and self-management is acknowledged by clinicians as a facilitator of implementation [14, 15].
In addition to the abovementioned constructs, the review also resulted in the development of new constructs including privacy, governance, unintended consequences, preference, choice and trist. Privacy is defined as ‘ the degree to which the innovation can be delivered in a way that preserves the privacy of the innovation deliverers and recipients’ and appeared in 4 papers. In the context of increasing data hacking and privacy breaches, this is likely to become a more prominent factor to consider in the development and implementation of virtual care [52]. However, it presented itself as an issue here when either the clinician or patient were engaging in consultations without sufficient privacy – e.g. while the patient was at the supermarket. Governance was also an issue, appearing in 4 papers and defined as ‘the degree to which an innovation can or is being supported by local, inner setting policies and procedures’. There were numerous instances, particularly in relation to the provision of devices to patients, where tracking their location, ensuring they were properly maintained and returned proved difficult [13]. To support the judicious use and proper maintenance, it is recommended that all equipment used for telehealth is kept on an equipment register with processes in place to ensuring the performance of routinely scheduled maintenance and its’ safe return from patients when they no longer require technology use, as well as that used in clinic to support virtual care operation[53]. Unintended consequences was mentioned in 1 paper and defined as ‘the degree to which the innovation results in unintended consequences (negative or positive) that could be relevant at patient, provider or system levels’. While this construct was not widely apparent, it is possible that it may emerge as a more prominent issue as virtual care matures. In the context of this review, it emerged as a problem when patients were misusing technology given to them, resulting in high internet usage costs to the practices, as well as difficulties with off boarding patients, where they had grown an attachment to the equipment and saw returning it as a symbolic way of acknowledging disease progression [14]. Studies are beginning to report unintended consequences more widely, with one study identifying multidimensional issues created by telehealth including the disruptive nature of integrating telehealth, potential to fragment care, impact to power relations between patient and provider, impacts to supply chains and mal-adaptive government policy borne out of urgency [54]. A more comprehensive view is provided by Gogia and colleagues which provides a raft of possible solutions to address the many unintended consequences of implementing virtual care [55]. Preference was mentioned in 6 papers and defined as ‘the degree to which an innovation considers the preferences of innovation deliverers and recipients’. Preference emerged as an issue when exploring the design and development of virtual models of care[13, 16, 18, 19, 21, 26]. Choice appearing in 3 papers and defined as ‘ the degree to which an innovation facilitates the ability of the recipients/deliverers to choose’ is a conceptually similar descendant to preference, related more to the operational nature of virtual care solutions [56]. For example, allowing patients to choose whether they have a face-to-face appointment versus telehealth is different to the idea of ensuring that varied preferences are included in the development of an intervention (e.g. an appointment booking application that provides multiple options). Given the importance of clinician and patient acceptability of virtual care interventions, it is important that both of these constructs are considered when developing virtual models of care. The final new construct was Trust, mentioned in 6 papers and defined as ‘the degree to which innovation deliverers or recipients believe that the care provided will not be compromised when providing the innovation’. Trust was an issue where clinicians felt as though the care they were providing would not be equivalent or superior to face to face care, or when they had difficulties trusting data that was collected by the patient or provided to them by a system [14, 16, 18, 20, 21, 26]. In the broader body of evidence, this extends to the patient who may increase their levels of trust in clinicians through the enhanced availability of care[57]. Trust is theorised by Ramachandran in an umbrella review that included 79 studies to be mediated by “patient-related, provider-related, technology-related and organisational factors, such as patient sociodemographic, provider communication skills, technology design and organisational technology implementation”[58]. Thus, it is likely that trust is an important factor in virtual care implementation and strategies to improve trust at all levels should be deployed.
This study has provided an overview of the key factors to consider when implementing virtual care interventions in a primary healthcare setting, and identified additional constructs that may have continued importance once tested in varied contexts. A limitation of the study is that it was conducted using a set search strategy at a singular time point. Though every effort was made to include all key terms related to virtual care, there may have been relevant studies which did not contain such terms, or were published after the search was conducted, and were therefore not included in this review as a result. As such, this review provides a review of the literature concerning virtual care in an Australian primary healthcare context as recent as December 2022. In addition to a lack of implementation theory use, there was a distinct lack of perspectives from patients, with most papers focused on influencing factors as experienced by clinicians. Finally, as this is a scoping review, there is a chance some studies were missed as grey literature was not searched for, and thus we cannot be certain that all relevant literature was included.