Telemedicine—or the delivery of health care at a distance—can improve health care access and quality while reducing health care utilization and costs [1]. For instance, telemedicine can improve access to specialists for patients in rural areas or in underresourced care settings by increasing the convenience and availability of health care (eg, extended hours, decreased wait times) [2-4]. Studies have also demonstrated that telemedicine may achieve comparable clinical outcomes to in-person care across a variety of conditions, such as stroke care [5], heart failure [6,7], hepatitis C [8], and diabetes [8]. Telemedicine can also reduce the utilization of in-person care and reduce health care costs [3,4,9]. Furthermore, some telemedicine types have demonstrated cost-effectiveness [10,11]. Studies have shown that the impact of telemedicine on health care delivery and patient outcomes varies across telemedicine types. Some forms of telemedicine, such as telestroke [12-14], have a strong evidence base while other forms of telemedicine, such as electronic visits (e-visits), are understudied.

E-visits involve asynchronous communication between clinicians and patients through a secure web-based platform, such as a patient portal. Generally, patients answer questions about their medical history and symptoms through a structured questionnaire and upload photos (if relevant). The data are then reviewed by a clinician, who develops a diagnosis and treatment plan. Although e-visits often involve clinicians who the patient is familiar with, they can also involve a third-party clinician through direct-to-consumer telemedicine. E-visits offer notable benefits, such as allowing patients and clinicians to communicate at a convenient time (ie, eliminating scheduling barriers) and improving documentation of patient-clinician communication (eg, patients can review clinician’s instructions) [15,16]. However, e-visits have a number of implementation barriers, such as those regarding workflow integration (eg, having dedicated clinician time to respond to messages) [16-18], those regarding lack of reimbursement [15,16,19,20], and concerns about the quality of communication from patients (eg, failing to submit sufficient information for diagnosis) [21]. Health care organizations have also raised concerns about the quality of care provided through e-visits, such as the potential for inappropriate antibiotic prescribing and difficulties with providing care without being able to see the patient face-to-face [22-29]. Despite these challenges, studies have reported positive effects of e-visits on health care delivery, such as lower costs [24], comparable follow-up rates to those of in-person care (ie, a proxy for diagnostic accuracy) [30], and comparable or improved patient outcomes (eg, lower uric acid level for patients with gout) [31].

Because of COVID-19, e-visits are being implemented more frequently and are now reimbursable by Medicare and other payers in the United States [32,33]. As the nation moves forward in telemedicine implementation and health care systems decide whether to integrate telemedicine into future care delivery, it is critical to determine the costs and benefits of telemedicine models, such as e-visits. To date, a systematic review has not been conducted to assess the state of the evidence regarding e-visits. To address this gap, the objective of this review is to summarize the findings of studies examining the association of e-visits with clinical outcomes, quality of care, access to care, utilization, and costs.

We conducted a systematic review based on PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) guidelines [34].

After reviewing 1859 studies, a total of 19 studies met our inclusion criteria. Figure 1 shows our study selection process.

This was a systematic review that assessed the impact of e-visits on clinical outcomes and health care delivery. To our knowledge, this is the first systematic review to assess the state of evidence for asynchronous e-visits. Most studies found that e-visits were associated with lower treatment costs and comparable clinical outcomes to in-person visits. Studies reported mixed effects on health care utilization, and no studies evaluated the impact of e-visits on health care access, suggesting future research is needed in this area. We provide implications for research and practice below.

Our review found that e-visits may be an adequate substitute for in-person care for chronic disease management. Studies showed that e-visits were effective for the treatment of diabetes, hypertension, and gout. Our findings are consistent with those of previous studies [52] that suggest that other forms of asynchronous communication with providers (eg, secure messaging without a formal e-visit) can improve chronic disease management. Further study, however, is needed to determine the ideal conditions for e-visit implementation. For example, many chronic conditions co-occur with other conditions (eg, diabetes and hypertension), and current studies were not designed to determine whether e-visits are effective for complex patients, such as patients with multiple chronic conditions and older adults who are frail. Additionally, it is important to note that not all types of medical issues may be appropriate for asynchronous management and health care systems may need to implement safeguards to ensure that the right patient is using an e-visit [53]. For example, if patients send a message about an urgent condition, e-visit technology could be harnessed to flag certain keywords (eg, chest pain, breathing difficulties) and display automated pop-up alerts to patients to instruct them to seek care in-person at the office or emergency department [54]. Some health care systems have also implemented guidelines, such as requiring at least one in-person visit prior to an e-visit to ensure that a patient has an established relationship with a clinician [55]. Additional research is needed to determine how e-visits should be implemented (eg, how to prevent inappropriate usage and for which patients does it work best).

Our review found mixed evidence for the effect of e-visits on quality of care. For example, prior studies [22,24-29] have raised concerns that telemedicine usage can increase inappropriate antibiotic prescribing, but this review found mixed evidence on whether comparable antibiotic prescribing rates were observed between e-visits and in-person visits. This inconsistency may stem from differences in the acute infection that was studied (ie, acute sinusitis, urinary tract infections, ear infections, conjunctivitis). Further research is needed to better understand when e-visits can be used effectively for managing acute infections and what implementation strategies can be used to ensure appropriate antibiotic prescribing (eg, use of clinical decision support). Furthermore, additional research is needed to compare the receipt of low-value care (eg, overutilization of services for sinusitis and urinary tract infections across e-visits and in-person visits). Future studies should examine guidelines, such as the Choosing Wisely guidelines [56,57], to see whether e-visits reduce low-value care compared to in-person visits. One study [24] included in this review found that e-visits were associated with lower rates of diagnostic procedures for sinusitis and urinary tract infections and may offer an advantage in terms of reducing unnecessary health care utilization. Studies in this review did report other quality problems, such as lower use of preventive care services in e-visits compared to in-person visits. This may stem from a lack of practice guidance on how e-visits should be implemented (eg, should providers use e-visits as an opportunity to reinforce messages about preventive care?). Prior studies [16] have noted that lack of practice guidance or standard procedures as barriers to e-visit implementation. These differences in quality of care suggest additional implementation research is needed to test implementation strategies for ensuring quality of care delivered through e-visits is consistent across clinicians.

Our review found that e-visits had mixed effects on health care utilization compared to in-person visits. These findings are consistent with those of reviews on other forms of telemedicine [6,58-62]. Specifically, telemedicine can reduce health care utilization in certain instances (eg, reduce the need for in-person visits) or it can increase utilization (eg, meet an unmet demand for a patient that was not previously accessing in-person care). Variation in health care utilization may stem from other various factors that are unaccounted for. For example, if a patient did not submit sufficient information to be evaluated through an e-visit, this could lead to greater health care utilization compared to an initial in-person visit where information exchange is synchronous. Future research is needed to determine strategies for ensuring that complete information is elicited from the patient (eg, structured symptom questionnaires). Additional research should test strategies for optimizing patient data collection and patient-clinician communication through e-visit platforms.

The findings on favorable cost implications align with those of other works that evaluated the effects that synchronous alternatives for care (eg, telemedicine via teleconferencing software) had on cost outcomes for patients [63-70]. Notably, 2 [40,49] of 3 studies used claims data to estimate health care costs; however, insurers commonly include in their contracts with health care organizations contractual adjustments (ie, slightly lower reimbursement rates in exchange for including the health care organization in-network) [71], suggesting that using claims data may underestimate the true cost savings potential of e-visits. Furthermore, many of these studies did not evaluate cost comprehensively (eg, only examined costs of dermatology visits and not visits to providers outside of that delivery system, analyzing average costs per visit instead of average costs of an episode of illness). Additionally, we did not identify any studies that assessed cost-effectiveness, which considers costs relative to outcomes. Future cost-effectiveness studies may also help health care systems make decisions about whether e-visits are worth the investment.

Additional research is also needed on how e-visits impact health care access and the digital divide. There is growing evidence that patient-level disparities exist across adoption and usage patterns of patient portals [72-76]. Similarly, several studies in this review reported differences in usage based on sex [19,30,39,41,42,48], age [30,42,47], employment status [41,48], and ethnicity [19,41]. Since many patients access e-visits through the patient portal, which has known disparities in uptake [72-76], future studies are needed to test strategies for overcoming disparities in patient portal adoption. Since the Pew Research Center reports 81% of Americans in 2019 owned a smartphone [77], there have been studies that have recently tested whether smartphone access to a patient portal could improve access [78,79]. Future studies should test whether strategies, such as smartphone access, could improve uptake of e-visits.

Some of the included literature cited implementation barriers in their discussion that should be further explored in future studies (eg, lack of integration into workflow). Clinicians may need assistance with adjusting their workflows when being trained on e-visits [80]. Best practices should be researched and disseminated to help alleviate concerns on increased workload [81]. Furthermore, the adoption of payment models by insurers that reimburse for e-visits may be a crucial facilitator of e-visit uptake [20,82,83]. Since the COVID-19 pandemic has spurred the introduction of insurance coverage of e-visits among Medicare and some private payers, case studies have been published that suggest e-visits are being used more frequently during the COVID-19 pandemic. For example, one health care system reported that the use of e-visits increased by 4000% during the COVID-19 pandemic and that the majority of e-visits were used for remotely managing patients with COVID-19 [84-86]. Further work is needed to evaluate if the reimbursement policies have led to higher utilization of e-visits by patients and health care organizations.

This systematic review has several limitations. First, a majority of the studies used an observational design, limiting our ability to draw causal conclusions on the effect of using e-visits on quality of care, access to care, costs or clinical outcomes. Second, we found heterogeneity in how studies measured the impact of e-visits, making it impossible to quantitatively pool study estimates. Third, we excluded non-English studies, which may limit our ability to determine the effect of e-visits outside of English-speaking regions. Fourth, the follow-up periods among the studies varied (eg, 2 weeks vs 1 month), limiting our comparisons of findings between studies. Fifth, a majority of the studies examined e-visits that are hosted by the health care delivery organization rather than by a direct-to-consumer vendor or insurer, limiting the generalizability of our results to e-visit programs sponsored by entities external to the health care delivery organization. Lastly, most e-visits occurred in outpatient ambulatory and academic medical center contexts, limiting our ability to comment on inpatient settings.

Overall, the evidence suggests that e-visits can provide equivalent outcomes to in-person care and reduce health care costs for certain health care conditions. There are still notable quality concerns (eg, inappropriate antibiotic prescribing, underutilization of preventive care) that warrant further study. It is also unknown how e-visits have affected access to care. Furthermore, many studies in the review lacked a rigorous design, making it difficult to draw firm conclusions about the state of evidence regarding e-visits. Future trials should be conducted to test the effectiveness of e-visits and determine what factors drive effective implementation.