The Role of Virtual Consulting in Developing Environmentally Sustainable Health Care: Systematic Literature Review

Introduction

Background

The risks of climate change are now widely accepted. Current projections suggest that we could see increases of 3 to 4°C in global temperatures by 2100, with catastrophic effects [1,2]. Extreme weather events including heat waves and flooding confirm the urgency of the challenge. Given the relationship between health and the environment, many of the effects are being felt in people’s health, both directly (eg, heat-related deaths and rises in vector-borne infections [3]) and indirectly (eg, disrupted food systems [4], reduced access to health care facilities owing to extreme weather events, and impacts on transport systems [5]). Urgent action needs to be taken to reduce carbon emissions and radically reduce the impact on health and well-being [1,6-10].

Health systems are major contributors to climate change because of the large quantities of greenhouse gases (GHGs) they emit (eg, owing to patterns of energy use, waste disposal, and complex supply chains [11-16]). If the global health sector were a country, some estimates put it as the fifth-largest source of GHG emissions on the planet [8]. In the United Kingdom, the National Health Service (NHS) is responsible for approximately 4% of all carbon emissions and approximately a fifth of public sector emissions [8]. This not only pollutes the environment but also leads to increased rates of illness (eg, cardiovascular disease and asthma [17-20]), with knock-on effects on costs [21]. Challenges are compounded by the rise in acute and chronic diseases, increased patient load, declining workforce, and system fragmentation [22].

This concern with environmental sustainability is part of a wide-ranging agenda to address climate change; progress has been made in the last decade, including via intergovernmental agreements and actions aimed at reducing global emissions (eg, the Paris Agreement [2]). More urgently needs to be done [23].

In the United Kingdom, the Climate Change Act of 2008 introduced public and private sector obligations to meet carbon reduction targets [24,25]. In health care, this led to the NHS Net Zero strategy [8], with targets set across the United Kingdom to achieve net zero health services by 2045 (or earlier where possible) and strategic roadmaps and actions for local organizations. To achieve this, multiple solutions are needed to reconfigure the delivery of care and the wider supply chain [26]. Technology offers a potential route for change. This includes the use of virtual consulting, which involves synchronous use of telephone or video platforms when consulting, either between clinicians and patients or across facilities (eg, from primary to secondary care [27]). Previously a novel service development [28,29], the shock of the COVID-19 pandemic enabled rapid and large-scale changes to the way services were delivered [30-32], with telephone and video technology allowing clinicians and patients to connect remotely and ensure physical distancing and infection control. There is a well-established evidence base supporting telephone consulting [33]. More recent research has demonstrated high levels of feasibility and acceptability of virtual consulting, with reductions in travel contributing to reduced emissions [11,34-36].

A recent systematic review of the evidence on the carbon footprint of telemedicine [35] showed that telephone and video consulting offer potential benefits in terms of reducing emissions and that these are largely attributable to patient and staff travel. The review was framed in terms of low-carbon alternatives to standard care and showed significant reductions (between 0.70 and 372 kg of carbon dioxide equivalent [CO₂e] per consultation) associated with reduced transport-associated emissions. This was a significant step forward in this small (but rapidly growing) area of research. However, the review paid limited attention to the adoption, routinization, and spread of video consulting (eg, [28,29,32,36]) and the ways in which wider organizational and system incentives and initiatives (eg, specific policies, operational resources, and pathway redesign) shape potential for environmental gains. Papers in the review tended to focus uncritically on patient travel, with little consideration of other potential mitigating factors limiting, or indeed contributing to, carbon emissions. Hence, although the published evidence to date has highlighted that a shift to virtual consulting could generate a positive impact on NHS-related carbon emissions, the net effect of such changes has not yet been established.

Objectives

More intensive action on the environmental impact is required to enable health systems to meet long-term carbon footprint goals. Important learning remains to be gained on the role of remote consulting in meeting these goals. This paper therefore asked, What is the impact of virtual consulting on environmental sustainability in health care? and What can we learn from current evaluations that can inform future reductions in carbon emissions?

Methods

This review forms part of a wider study on Remote by Default Primary Care [37] and builds on a program of work on virtual consulting undertaken before and during the COVID-19 pandemic [28,31,32,34,36]. The PRISMA (Preferred Reporting Item for Systematic Reviews and Meta-Analyses) guidelines [38] were consulted throughout to guide the review. PRISMA-P (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols) was used to draft the protocol for this review (unpublished).

Eligibility Criteria

Inclusion and exclusion criteria are summarized in Textbox 1.

Information Sources and Search Strategy

Our search strategy was developed with the help of a research librarian and used a mix of keywords and Medical Subject Headings—Major Headings. We searched MEDLINE, PubMed, and Scopus in November 2021 and updated the search again in August 2022. Search terms were identified from relevant published literature using 2 categories of keywords covering environmental sustainability (eg, “carbon footprint” and “environmental impact”) and virtual consulting (eg, “telemedicine” and “remote consultation”). Filters were applied to limit the results to published peer-reviewed articles and the English language only. A full list of the search terms can be found in Multimedia Appendix 1.

Using the same search terms and eligibility criteria, as outlined in the paragraph above and in Textbox 1, we identified additional studies via other methods. One author (AB) maintains an EndNote library of studies on environmental issues in health care, which was searched for any additional papers relating to telemedicine. This database was put together through historical searches on environmental impact and sustainability in health care using proximity searching (within 5 words) to bring search terms on environmental impact and sustainability closer together. We used citation tracking to identify further papers focusing on virtual consulting and sustainable health care. Finally, we included 1 organizational document reporting research on the environmental impact of telemedicine, identified via Google Scholar.

Selection Process

The search results were imported into EndNote and duplicates were removed. Each record was initially independently and manually reviewed by title and abstract by 2 authors (MPS and MA). Where there was doubt about whether articles met the inclusion criteria based on the title and abstract, the full-text version was obtained and screened by 3 authors (AB, MPS, and MA), and inclusion or exclusion was verified by the project lead (SES). At each stage, articles were eliminated if they did not meet the inclusion criteria. In 1 case [39] where it was unclear whether the paper met the inclusion criteria given that it was published as a report, the paper was discussed, consensus was reached that it did meet the criteria, and it was included.

Data Collection Process and Synthesis

At a descriptive level, we first extracted data on study location or setting, study timing, clinical focus, definition of telehealth, service model and type of technology, research design, and key findings. Second, as the majority of papers focused on carbon footprinting, we extracted specific data on the approach to carbon footprinting, methodology used, and any reductions in carbon emissions achieved in the service or project. We elected not to convert the carbon footprint data to the same units or to extrapolate the carbon savings per individual consultation in each study but to present and work with the measurements presented in each paper. Each paper worked with different units, different types and setup of virtual consultations, different health care services, and facilities that were located at different distances from patients, and we felt that converting the studies to the same units would be misleading and detract from our analysis of the various methods that the authors used to determine carbon footprints in their papers. Finally, we extracted data on how each paper conceptualized environmental sustainability and the opportunities and challenges perceived to be provided by virtual consulting services. This process was undertaken independently by 2 authors (MA and MPS) before being checked by SES. All data were extracted into a Microsoft Excel spreadsheet and are presented in tables.

We used the Planning and Evaluation of Remote Consultation Services (PERCS) framework [31], which includes a specific focus on a wider system and consideration of climate emergency (Figure 1), to guide synthesis. Sensitized by PERCS, we worked inductively to surface the opportunities and challenges for environmental sustainability presented by virtual consulting services. We then worked deductively to examine if and how any of the 8 domains were identified as relevant to developing sustainable virtual consulting services. One author (MPS) piloted and refined this process on one paper, and we discussed the findings as a team. We then extended the process across the rest of our data set. Using content analysis, we compiled a descriptive overview of the opportunities and challenges identified in the papers and then synthesized and interpreted our findings using the PERCS framework [31].

‎

Quality Assessment

To our knowledge, there is no checklist currently available to guide the assessment of carbon footprinting studies. Our understanding of the quality of the articles was informed by using the Mixed Methods Appraisal Tool for mixed methods studies [40], which is designed for the appraisal stage of reviews that include qualitative, quantitative, and mixed methods studies. Given that Mixed Methods Appraisal Tool (like other currently available critical appraisal tools) does not apply specifically to articles reporting carbon footprinting, we elected not to exclude articles on the basis of quality appraisal of other study methods but took this into account in the interpretation of their findings.

Results

Study Selection

The database search identified 241 results (Figure 2), with 18 titles relevant to virtual consulting and environmental sustainability in health care. Other methods identified 1431 results with 24 relevant titles. As 3 were duplicates, these were excluded. We reviewed the remaining 39 full-text papers, excluding a further 16 based on relevance or format (Figure 2).

This process provided a total of 23 papers for the review (refer to Table 1 for an overview).

‎

Study Characteristics

The studies reviewed were based on data collected from 1996 to 2021. Eight studies were conducted in the United Kingdom, 7 in the United States, and 2 in Spain, with the remainder in France, Ireland, Sweden, Canada, and Portugal. One study examined the impact of video consulting at a national level across different health care specialties [58]. Two other studies attempted to extrapolate from the analysis of the reduction in carbon emissions in 1 telemedicine service to the national level [39,62]. When combined, these studies provided a broader understanding of the potential reach of virtual consulting in informing the agenda for sustainable health care, including the potential for reduced pollution.

All 23 papers described virtual consulting in terms of telemedicine, teleconsultations, or virtual clinics (Table 1), capturing the use of telephone or videoconferencing equipment (eg, Attend Anywhere) to connect patients with clinicians or primary care clinics with higher levels of care. In 1 instance, videoconferencing equipment was used alongside a companion device (laryngoscope [48]). Two studies used videoconferencing software combined with digital imaging [61,62]. In 1 instance, both telephone and videoconferencing were used to deliver virtual services [42], and in 4 other studies, patients were given the option of either telephone or audiovisual/video consulting [43,46,53,58].

All studies focused on the environmental sustainability potential of telemedicine through the carbon savings achieved by a reduction in travel to face-to-face appointments. The key terms used reflected this, with issues of environmentally sustainable health care framed in terms of reduction in “carbon emissions.” A handful of papers included additional data on other “greenhouse gas emissions” (eg, [61]), whereas approximately half of the studies (11/23, 48%) chose to display these data as a CO₂e figure (eg, [42,44,47,48,50,53-56,59,62]). None of the papers included in the review provided data on the current footprint of providing their service but instead focused on the impact of virtual consulting on carbon emissions or comparison of emissions via in-person and virtual consulting.

The clinical use of video consulting varied. Some studies focused on the use of video consulting across a range of acute and community clinical settings [39,49,53,55,58,59,61]. The remaining papers focused on local or regional services tied to specific specialties, including urology [45,46,50,56]; neurology and neurosurgery [42,62]; orthopedics [47]; cancer [41,54,57,63]; ear, nose, and throat [48], renal [44], and specialist rehabilitation services [52]; and pediatrics [43,60]. No study focused solely on virtual consulting services in primary or community settings. Most papers (22/23, 96%) looked at telemedicine use within specific medical specialties (eg, neurology and oncology) or at cross-specialty referrals. Across all papers, 12 (52%) looked at telemedicine use in patients who already had a diagnosis, whereas the remainder (n=11, 48%) looked at telemedicine use for patient diagnosis (eg, assessment and imaging done remotely and then communicated to specialists via telemedicine consultations) or first-contact consultations.

Evidence on Carbon Savings

The language, measurements, and technical processes related to calculating and presenting data on carbon emissions varied across studies. Approximately half of the studies (11/23, 48%) reported estimated carbon emissions in kilograms of CO₂e, a common unit used to describe all GHG effects as the equivalent global warming potential of carbon dioxide [42,44,47,48,50,53-56,58,62]. Other studies described a simple measure of CO₂ emissions [39,41,43,45,46,52,57,58,60,63]. Three studies calculated the volumes of other air pollutants or GHGs, such as nitrous oxide and sulfur dioxide [49,55,61]. Papers reported carbon savings using different units, for example, metric tons, tons, kilograms, or pounds, and for different sample sizes of telemedicine consultations and across varying time periods. Only 7 papers reported carbon savings per individual patient consultation [44,47,48,50,59,61,62].

All 23 studies focused on estimating the carbon savings achieved by reducing the amount of road or air travel to in-person appointments (Table 2). Only 1 study included carbon savings associated with an in-person outpatient visit, for example, clinic lighting and heating [47].

Across the 23 studies, the estimated total carbon emissions saved varied greatly because of the focus on anything from a specific service or specialty to regional or national health provision (Table 1). The quantity of carbon savings estimated across studies largely depended on the distance traveled by patients and the mode of transport typically used to get to in-person appointments (eg, cars, buses, trains, or planes). Eight studies conducted surveys or asked patients about the distance they traveled and their preferred mode of travel during consultations [44-47,50,54,56,59]. The remaining studies obtained data to determine the distance traveled from patients’ home addresses listed in patient records, national data, or by determining distances between referral health care facilities. Studies that did not specifically ask patients about their preferred mode of transport typically assumed that car travel would be used and used varied sources to approximate car type, size, fuel type, and driving conditions as the basis of their calculations. For instance, Blenkinsop et al [42] modeled scenarios on cars using petrol only, diesel only, and proportions of petrol and diesel based on the current UK figures for cars sold, excluding 12 consultations that would have involved journeys via air or ferry. Vidal-Alaball et al [61] modeled their data using equal numbers of petrol and diesel cars. Miah et al [56] calculated the difference in savings if people relied on personal car versus public transport (underground train). Three studies also included air travel [41,48,62] in their calculations of transport emissions. Only 1 study included potential carbon savings from reduced staff travel [44].

In studies that used data to calculate commuting distances, specific tools (eg, GPS, Google Maps, and Esri ArcGIS) were used to calculate the approximate distance from the patient’s postcode or between health care referral facilities and assumed that the shortest or quickest route was taken.

The methods used to calculate the carbon footprint varied across studies. Four studies used preexisting web-based carbon footprint calculators [45,46,54,56], the most common being Carbon Footprint Ltd. Other studies manually calculated the carbon savings by multiplying the determined distance traveled by the emission factors for the various modes of travel obtained from emission conversion factor databases, such as the Environmental Protection Agency database in the United States or the Department for Environment, Food and Rural Affairs in the United Kingdom.

Two studies extrapolated their findings to a wider area: Midlands and Lancashire NHS Commissioning Support Unit [39] estimated that savings of 533,535 kg of CO₂ could be achieved if 15% of follow-up consultations in the NHS Midlands (a central United Kingdom region) were made virtual, and Whetten et al [62] estimated that up to 213,279,000 kg of CO₂e could be saved if neuroemergency consultations were held virtually across the United States before making the decision to transfer a patient to a trauma center.

Five studies [42,50,52,55,62] calculated the carbon footprint associated with telemedicine consultation (Table 2). The estimated savings ranged from 0.002 kg of CO₂e per hour of consultation (for telephone calls used to deliver virtual care to patients with complex epilepsy during the COVID-19 pandemic) [42] to an upper limit of 8.43 kg of CO₂e for a 1-hour videoconferencing appointment including energy consumption and embodied emissions of the equipment (eg, design, manufacturing, disposal, and recycling of the equipment) [52].

Four of these studies modeled the carbon footprint of the telemedicine element using a single type of technology [50,52,55,62], with 1 expanding this analysis to make comparisons if patients and doctors used different technologies such as mobile phones, laptops, or personal computers [42]. Although 2 studies referred to additional equipment required for virtual consulting (ie, imaging equipment) [61,62], there was an otherwise limited assessment across papers on the carbon footprint of additional equipment that might be required (eg, headset and software).

Wider Impact and Development of Virtual Consulting

All reviewed papers concluded, to varying degrees, that virtual consulting significantly reduced carbon emissions. In a handful of papers, authors were able to suggest the point at which virtual consulting offered a greener choice (eg, “at a distance of a few kilometers when the alternative is transport by car” [52]). This led to the overarching conclusion that future health services must be built on sustainable and low-carbon systems and work models that include forms of virtual consulting. In other words, if ambitious targets are to be met [68], then changes to health care practices will be needed at many levels. Although critical and a much-needed ambition in light of the major contribution of health systems to emissions, all papers focused largely, and in most cases (18/23, 78%) exclusively, on patient travel. This clearly represents one of the key areas where reductions in emissions can be achieved (Table 2), particularly in remote and rural communities where distances from providers, emissions, and allied transport costs (eg, via helicopter [62]) are significant. However, patient travel is only one part of the picture, with analysis typically limited in terms of other changes that might be needed, how they might be achieved, and the implications for carbon emissions.

Across the studies, there was limited consideration of wider factors influencing emissions, including wider services and pathways that intersect with and support virtual consulting; the use of (and emissions from) technology enabling health care design and delivery; and wider organizational, infrastructural, and policy-level factors shaping these emissions (all of which involve some form of human or technological activity, with implications for emissions). Consideration of these wider factors was rare and tended to focus on the service or clinic level. For instance, Blenkinsop et al [42] considered emissions at the level of the clinic, focusing on electricity consumption to deliver virtual clinics. They also investigated a sample of patient records and treating clinicians to identify adverse clinical outcomes (n=1). Holmner et al [52] conducted a life cycle inventory to evaluate the carbon reduction potential of telemedicine activities beyond a reduction in travel-related emissions, taking into account end-point devices (eg, computers, monitors, cameras, and local area network components) and video codecs used to compress and decompress digital video signals, and found a 40 to 70 times decrease in carbon emissions compared with in-person visits.

None of the studies (Table 1) assessed the carbon footprint of the entire clinical pathway in which telemedicine consultation was provided as an alternative to in-person visits. This is an important point as virtual consultations do not sit in isolation but are embedded in the health care experienced by patients (and the providers they interact with) and the wider health care system [34]. For instance, although emission reductions were significant from virtual consultations alone, there is potential for further consultations, either virtually or in person, beyond this specific episode of care. There is some evidence that virtual consultations are suitable or appropriate for certain types of conditions or consultations and bring questions about risk and quality for others [69,70], with concerns over potential misses and adverse events. Such instances, although rare [42], potentially bring a need for additional care, at best in the form of a further—likely in-person—consultation and at worst with an admission or more intensive management. Such scenarios were rarely, if ever, considered and potentially negate at least some of the emissions saved from virtual appointments.

Other system-wide considerations [31] were also not included in the papers. For instance, all 23 papers assumed that patients did not need support or care at home (eg, via family) to participate in a virtual consultation (digital inclusion) and that patients who were offered a virtual consultation did not seek other health care or support elsewhere in the system. With the exception of those studies that looked at the carbon savings of virtual consultations between levels of care, for example, clinics and tertiary hospitals [48,55,59,62,63], all the papers assumed that patients remained in their homes to conduct a virtual consultation and did not travel elsewhere (eg, for Wi-Fi access or support from a caregiver). All these scenarios potentially result in increased emissions (although likely small in comparison with reduced emissions through travel). None of these scenarios were considered in the papers included in this review.

Discussion

Summary of Main Findings

This review has shown the potential of virtual consulting for helping to address the impact of the climate emergency in health care by contributing to reducing carbon emissions. Most published studies focus primarily (if not, in some cases, exclusively) on reducing patient travel as the main means of achieving this. Our use of the PERCS framework, combined with the systematic search and review of published literature, has enabled us to do the following. First, we have identified the current evidence base on carbon emissions relating to virtual consulting and identified the varied, but consistently sizable, emissions savings enabled by the use of this service model. This ranged from minimal to significant quantities of carbon saved depending on the distance that would have been traveled to the health care facility by patients (and in some cases staff), the method of transport used, the methodology used to calculate the carbon footprint, and the assumptions underpinning it. Second, we have shown that although patient travel is a major consideration, the extent of its contribution varies, for example, between an inner-city primary care clinic (with average travel of <10 km and typically via public transport) and a specialist hospital (serving a regional or wider population with average travel of >100 km by car). Third, we have highlighted that virtual consulting is not a stand-alone interaction but one part of the wider health care system that is rarely considered in papers when thinking about or calculating carbon emissions. Although a focus on patient and staff travel was critical and contributed significantly (albeit variably) to emissions, an exclusive focus on this (episodic sustainability) presented a skewed picture of progress toward system-wide sustainability. Fourth, and related, we have pointed to a range of other factors that are relevant in terms of potential emissions saved and cost in terms of emissions produced via design, provision, and the use of virtual consulting services.

Comparison With Wider Literature

Our findings resonate with those from an earlier systematic review [35], which clearly demonstrates that virtual consulting has a valuable role to play in achieving lower-carbon health systems. However, this and other literature in the field [71-73] tends to focus solely on carbon emissions allied to travel, with limited analysis of the design and delivery of virtual consulting, limited critical reflection on the assumptions underpinning calculations, and limited focus on the entire clinical pathway. Although it is clear from our own review and that of Purohit et al [35] that reducing travel through the use of virtual consulting holds the potential for significant emissions savings, consideration of patient, carer, organizational, and system-wide factors (Figure 1) shaping the planning and delivery of virtual consulting is needed.

Previous studies have tended to focus on carbon footprinting of virtual consulting in specific services and settings. Some studies have sought to extend findings about reduced emissions from 1 service to the wider health system. Although welcome, care needs to be taken here: not all health care can be provided virtually, and there needs to be reasoned application of it, acknowledging the nature of different clinical presentations as well as increasing the awareness of the environmental impact of the entire clinical pathway in these decisions. Recent work emphasizes this point, highlighting how despite the crisis context and resourcing, video consulting did not pick up in general practice during the COVID-19 pandemic when it was expected to (suggesting that video consulting in the UK NHS at least is not well accepted outside hospital settings) [74]. Our review adds to the existing evidence by shedding light on these wider influences and the need to consider them in future service design and delivery as well as research and evaluation. This requires a multilevel approach to generating evidence that is attentive to multiple considerations spanning policy to practice.

The potential of virtual consulting across health care settings and specialties has been well-documented over the course of the COVID-19 pandemic. Published evidence from ourselves and others [30,31,34,36] indicates that virtual consulting is feasible, broadly acceptable to staff and patients, safe over the short term, and associated with sizable net emission savings. This latter point is a key finding from the review. However, papers say very little about how to develop virtual consulting services nor the policy, regulation, and professional support needed to progress virtual consulting and environmentally sustainable health care. To our knowledge, little has been done at this intersection. There is considerable attention on virtual consulting and sustainable health care and only a small focus on the intersection between them. One recent example is the development of a national-level video consulting program (the Scottish Near Me service) and the drive for more environmentally sustainable health care. Although developing separately in Scotland, these 2 major initiatives have intersected in ways that appear productive, not only allowing for investment in material and technological infrastructure, staff training, and professional and public engagement to support video consultation services [36] but also enabling a parallel focus on emissions reduction.

The carbon savings presented in the studies reviewed seem impressive. However, it is still only a small part of the solution to reducing health care’s carbon footprint. According to the carbon footprint model of the NHS [8], personal travel accounts for 10% of estimated emissions as follows: 5% from patient travel, 4% from staff travel, and 1% from visitor travel. This is in comparison with the medical supply chain, which contributes to 62% of the emissions. Thus, although telemedicine is one way in which the health system can reduce its environmental impact, a focus on reduced travel and increased use of virtual consulting must not be seen as the only solution. Many other innovations and initiatives (see [26,75-80] for examples) need to be developed, adopted, and spread if health systems are to contribute significantly to net zero ambitions. It is likely that digital technology has a major role to play in this. However, caution is required in relation to the “rebound effect,” whereby a more efficient technology results in higher carbon emissions because of the increased use of that technology [80]. This holds potential for increased use of telemedicine and the devices and platforms that support and interconnect with it, to give rise to higher carbon emissions.

Carbon footprinting is a useful tool but remains a relatively new field, with limited mainstream knowledge about the process and evolving quality assessment. A carbon footprint is only a best estimate of the direct and indirect emissions produced by a certain product or scenario and usually includes the 7 GHGs covered in the Kyoto Protocol [81], expressed as CO₂e. The GHG Protocol outlines 3 scopes that aid in the reporting of all emissions associated with an organization or activity: scope 1, direct emissions from the organization; scope 2, indirect emissions associated with the organization’s electricity use; and scope 3, all other indirect emissions including supply chain emissions [82]. When calculating the emissions from driving a vehicle, one needs to consider the direct emissions from burning the fuel, which are different depending on whether the vehicle runs on petrol, diesel, or electricity; the scope 3 emissions embedded in a vehicle, namely, well-to-tank emissions (the extraction of raw oil, transport to the oil refinery, production of fuel, and transport of fuel); and the GHG emissions associated with manufacturing and distribution. This process requires significant expertise. To estimate emissions, some of the studies reviewed inserted numbers into a web-based carbon calculator: this process typically does not take into account scope 3 emissions, frequently uses different methods to produce carbon footprint values (some of which are poorly explained or reported in the studies using them), are usually only valid in the country in which the calculator was produced, and thus commonly results in some uncertainty over the accuracy and reproducibility of reported carbon savings.

The distances used to calculate the carbon footprint in the studies were generally based on estimates from postcode data, with the assumption that there were normal driving conditions and that patients drove the most direct route to the hospital. This, in addition to other assumptions (eg, type of vehicle or fuel), introduces a level of unavoidable uncertainty. This means that carbon footprints remain an estimate of environmental impact.

Carbon footprint calculations consider the environmental impact from an emissions perspective and do not take into account other toxins released or waste produced from a process. In the case of telemedicine, switching to virtual consultations will require electronic technology, which, when it has completed its life cycle, will likely end up in landfills (eg, the “electronic graveyard” [83]). Other environmental impacts of virtual versus face-to-face appointments were not included in any of the studies reviewed. This remains a gap in the literature.

Strengths and Limitations

This paper has shone a much-needed light on the potential of virtual consulting as one means by which health systems can reduce carbon emissions and contribute to addressing climate change. Our use of the PERCS framework [31] was helpful in widening the considerations of environmental sustainability beyond carbon emissions through patient travel alone. However, we were limited in the range of papers reviewed, most of which focused on secondary care (or tertiary or quaternary) services and included virtual consulting services provided to outpatients and those with a preexisting diagnosis. This limits the potential transferability of findings (eg, the extent of carbon emissions), particularly with regard to primary care settings. This may have been compounded by our original search strategy, which focused on 3 databases. It is possible that a broader search across other databases would have identified additional papers. We were limited by the quality of the papers, which used varying carbon methodologies, and the lack of critical appraisal tools available to assess the quality of the academic literature on carbon modeling.

Conclusions

Environmentally sustainable health care aims to facilitate health systems that are able “to anticipate, respond to, cope with, recover from and adapt to climate-related shocks and stresses, while minimising negative impacts on the environment and leveraging opportunities to restore and improve it” (page 26) [10]. The goal is to protect the health and well-being of current and future generations [23,24,52]. Virtual consultations in routine care provision are one means of reducing the environmental impact of health care and can provide significant carbon savings. It is widely accepted that this is due largely (but not exclusively) to the removal of the need for patients to travel to health care facilities. Further work is now urgently needed to fully appreciate the extent of carbon emissions and the potential for reductions (as well as other environmental impacts, eg, electronic waste) across the entire clinical pathway, rather than reductions focused on virtual consultations alone. This research should be conducted across all levels of health care provision, including primary, secondary, and tertiary levels. Those evaluating and funding virtual consulting services need to better appreciate the potential for reductions in emissions and weigh this up in relation to the benefits and potential risks (eg, adverse events and missed diagnoses) associated with providing or scaling up such services. If we are to see a rapid and pronounced change, then we urge both funders and evaluators to include carbon footprint modeling routinely into their design and methods and to fully engage with the transdisciplinary endeavor of reviewing, monitoring, and addressing issues of environmental sustainability.

We encourage all those performing carbon modeling of health care services to apply consistent methodology according to published standards (eg, the GHG Protocol). Policy makers and planners would do well to consider the potential of virtual consultations to help mitigate the effects of climate change, in the context of the wider system changes needed to support the development, adoption, and spread of these services in ways that enable reductions in emissions across clinical pathways while simultaneously enabling access and digital inclusion. This requires investment in climate literacy and skills training. Time is short. We urge funders, evaluators, policy makers, and planners to coordinate; to act with a sense of urgency to address environmental impact in health systems; and, in line with the Paris Agreement, to work toward limiting the mean global temperature rise to well below 2 °C.