Healthcare is a significant contributor to greenhouse gas (GHG) emissions. Telemedicine offers a potential solution for reducing this carbon footprint by eliminating patient travel. This intervention also benefits patients with pollution-sensitive conditions. Besides reducing GHGs from patient travel, telehealth is cost-effective and increases access to care. The widespread adoption of telehealth, particularly following CMS payment parity for telehealth services, highlights its potential impact. While some studies show a correlation between telehealth and reduced patient travel, they lack comprehensive assessments of the environmental impact beyond travel or focus narrowly on specific specialties. This study examines the life cycle GHG emissions from telehealth and in-person care at a health system level to demonstrate how telemedicine can improve access to care and decrease a health system's carbon emissions.

Existing research indicates a positive correlation between telehealth and reduced patient miles traveled. However, these studies often overlook the broader environmental footprint of clinical care, focusing solely on travel or specific specialties. The literature shows variability in reported GHG emission reductions from telehealth (0.69-893 kg CO2e reduced per patient visit), owing to differences in research methods, study boundaries, and the types of telehealth analyzed. Transportation is consistently identified as the leading source of GHG emissions from in-person visits. Other contributing factors include PPE, supplies, and electricity/energy costs. One study highlighted electricity consumption as the largest contributor in Greek military hospitals, suggesting decarbonization of the energy sector as a solution. Significant GHG emissions are also associated with the production and transportation of surgical equipment. The existing literature highlights heterogeneity in the factors contributing to excess GHG emissions based on location (rural vs. urban) and specialty (surgical vs. nonsurgical).

This study used a life cycle assessment (LCA) to quantify the environmental emissions of virtual and in-person clinical visits at Stanford Health Care (SHC) from 2019 to 2021. The functional unit was one clinic visit. GHG emissions were measured in kg CO2-equivalents and metric tons of CO2-e. Data on the number of in-person, phone, and video visits were collected by SHC's Digital Health Care Integration team. Patient travel distance was estimated from patient and clinic ZIP codes; distances over 402 km were assumed to be air travel, while shorter distances were assumed to be car travel. A list of typical supplies was used, assuming manufacturing in China and shipment by boat and truck. SHC's engineering team provided data on electricity, gas, steam, and chilled water consumption, which was used to estimate energy used per visit in a 10' x 10' exam room. The LCA software SimaPro 9.3.0 and LCI database Ecoinvent v3.8 were used, with impact assessment using the EPA's TRACI 2.1. To estimate GHGs avoided by virtual visits, the GHG emissions for these visits were estimated as if they had been in-person. Sensitivity analyses explored the impact of various assumptions, including travel modes, energy consumption, electricity sources, and supply lists. Ethical approval was not required as the study used de-identified, aggregated data.

Between 2019 and 2021, SHC saw a 13% increase in total visits but a 36% decrease in GHG emissions from clinic visits (from ~40,600 to ~25,900 metric tons of CO2e). In 2021, the average in-person visit emitted 20 kg CO2e, while phone and video visits emitted 0.02 and 0.04 kg CO2e, respectively. Patient travel dominated GHG emissions for in-person visits (44% air travel, 55% car travel in 2021). Electricity use for virtual visits still resulted in 29,000 kg CO2e in 2021. Telemedicine reduced 2021 GHG emissions by nearly 17,000 metric tons compared to in-person visits. Psychiatry and cancer departments achieved the greatest GHG reductions due to higher virtual visit rates. Sensitivity analyses showed that patient travel mode significantly impacted emissions; modeling all travel as car travel increased in-person visit emissions by 77%. Energy sources had a greater impact on virtual visit emissions than in-person visits. A maximum supply list increased total GHG emissions from all in-person 2021 visits by about 1.1%. Per-visit emissions varied by department and visit type, with primary care and pediatrics emitting the least for in-person visits.

This study confirms that telehealth significantly reduces GHG emissions associated with healthcare delivery. The major driver of GHG savings is the reduction in patient transportation. The study's comprehensive approach expands upon previous research by considering factors beyond transportation, such as supplies, energy use, and waste. The findings highlight the variability in the feasibility and effectiveness of telehealth across different specialties, with some specialties (e.g., psychiatry) being more suitable for virtual visits than others (e.g., orthopedics). Successful virtual visits, where a diagnosis and treatment plan are established without needing a follow-up in-person visit, maximize cost savings, patient satisfaction, and emission reductions. Barriers to telehealth adoption include upfront equipment costs, transition inefficiencies, and potential lower patient satisfaction compared to in-person visits, although satisfaction is improving. The study's findings have policy implications, suggesting the need to expand reimbursement policies to incentivize telehealth adoption and mitigate the environmental impact of healthcare. While telehealth significantly reduces emissions, decarbonization of electric grids and other initiatives are also crucial for achieving a zero-emissions healthcare system.

This study demonstrates that widespread adoption of telemedicine can substantially reduce GHG emissions in healthcare. The significant reductions in GHG emissions achieved by Stanford Health Care through increased utilization of telemedicine highlight its environmental benefits. Future research should focus on optimizing telehealth implementation to maximize its effectiveness and address remaining limitations, such as accurately predicting successful virtual visits and ensuring equitable access for all patient populations. Expanding reimbursement policies and addressing the technological infrastructure needed to support telehealth are crucial next steps to realizing its full environmental and clinical potential.

The study relied on several assumptions, including estimates of patient travel distances and modes of transportation. The exclusion of clinician commute emissions and HVAC emissions from virtual visits represent limitations. The assumption that all virtual visits were successful and did not require follow-up in-person visits might overestimate the GHG reduction. The study's findings may not be generalizable to all healthcare settings, given the specific characteristics of Stanford Health Care and its patient population. The sensitivity analyses addressed some of these assumptions, but the inherent uncertainties associated with LCA modeling should be acknowledged.