Research undertaken prior to the COVID‑19 pandemic highlighted the challenges in the analysis of peer-reviewed articles on telemedicine. The previous search conducted by the OECD was forced to limit the inclusion criteria to only systematic reviews given the high number of records returned when searching the term “telemedicine” in scientific publication databases. Since the COVID‑19 pandemic, there has been a significant increase in the use of telemedicine to ensure continued access to health services and to limit the spread of the disease. Health services are now using an increasingly mixed approach to delivering health care, including both in-person and telemedicine services. It is therefore critical to understand the use and impact of telemedicine as a tool for the delivery of patient care. To address this gap in knowledge, a literature review has been undertaken to obtain insights from a wider range of peer-reviewed and grey literature on telemedicine, focusing on evidence published since January 2020.
The COVID-19 Pandemic and the Future of Telemedicine
Annex C. Systematic review of the literature
Background and motivation
Definitions
Telemedicine can be defined as “the use of telecommunication systems to deliver health care at a distance” (see Box 1.1). Telemedicine can be split into three categories, which can be combined as appropriate:
Remote monitoring: This is the use of mobile devices and platforms to conduct routine medical tests, communicate the results to health care professionals in real-time, and potentially launch pre‑programmed automated responses.
Store and forward applications: These systems are used for clinical data that are less time‑sensitive and for which a delay between transmission and response is acceptable (e.g. they have been widely used in dermatology and in regions with poor connectivity that precludes real-time transmissions).
Interactive (real-time) telemedicine: This involves direct and synchronous communication between health care professionals (e.g. in health care facilities or dedicated telemedicine centres) and patients (e.g. at home or in health facilities).
This definition of telemedicine excludes:
Applications that do not involve any sharing of data or interaction between the patient and a health care provider/professional.
Physician education and provider-to-provider communications.
Objectives
The aim of this literature review is to capture high quality evidence of the use of telemedicine in OECD countries since the onset of the COVID‑19 pandemic, and to clarify the impact of the pandemic on the use of telemedicine. The search has included all OECD countries.
The main objective of this systematic review is to assess the impact of telemedicine services on the following five dimensions related to health care system performance (see Chapter 3):
Assessment of user (e.g. patients, health care workers) satisfaction with telemedicine services.
Evaluation of the impact of telemedicine services on patient safety.
Utilisation rates by socio‑economic status and demographics.
Utilisation rates by type of telemedicine (video, remote monitoring, asynchronous).
Cost analysis (e.g. economic evaluation) and pricing of telemedicine services.
Review protocol
Study design
A systematic review was performed, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
Sources and search Strategy
A systematic search of the literature published between January 2020 and May 2022 was performed on PubMed/Medline, the Health Management and Policy Database from the Health Information Management Consortium (HMIC), and the Cochrane Library, using a combination of controlled terms or free text depending on the database functionality (e.g. “Telemedicine (MESH term)” or “Telemedicine” free text). The keywords “Telemedicine” and “Teleconsultation” were each searched to maximise results towards ensuring the search picked up all relevant publications in the set timeframe. The OECD countries were used as keywords in the search. A search of grey literature within the same time parameters was undertaken to identify possible additional studies that meet the inclusion criteria. Searches were conducted on the websites of relevant stakeholder organisations (OECD and World Health Organization (including WHO regional offices and WHO ICTRP Search Portal)), relevant national Departments of Health, and conference proceedings of several related conferences.
Inclusion and exclusion criteria
To be included in the systematic review, individual studies had to adhere to the following criteria:
Studies will be included if they (1) had an intervention consisting of a telemedicine approach as defined above (alone or part of a complex intervention) and/or (2) evaluated the impact in quality of care, in an OECD country.
Types of studies considered will include systematic reviews and/or meta‑analysis, randomised control trials, cohort studies, case‑controlled studies, cross-sectional studies and surveys, or case studies.
Published between 1 January 2020 and 31 May 2022.
Published in the English language.
The exclusion criteria for the systematic review included:
Feasibility studies.
Non-systematic reviews.
Commentary and editorial papers.
Publications for which the full manuscript was unavailable.
Data collection and analysis
Two independent researchers (NO, KL) performed the title, abstract, and full-text screening based on the inclusion criteria. Conflicts were resolved by consensus at each stage of the process.
Data extraction and management
Data from the included studies was downloaded directly from the publication databases by a member of the research team (NO) and uploaded into Covidence, a software for systematic review management. The data extracted for each study included: title, authors, year of publication, characteristics of the intervention, and summary of main outcomes.
Assessment of methodological quality of included reviews
The quality of included publications was not assessed. However, an independent researcher (ALN) reviewed the full list of included publications following full-text screening to ensure that only academic studies and grey literature that met pre‑specified criteria of rigour, including the quality of included primary studies, were listed in the results tables.
Data synthesis
The data were synthesised according to the following areas:
User satisfaction with telemedicine services;
Impact of telemedicine services on patient safety;
Utilisation rates by socio‑economic status and demographics;
Utilisation rates by type of telemedicine and;
Cost analysis and pricing of telemedicine services.
The focus of data presentation is descriptive, with tabular and graphical presentations where appropriate. Due to heterogeneity of populations, interventions, and outcomes (PICOs) across the included studies, no attempt was made to compare telemedicine interventions across reviews or across review populations.
Results
Figure C.1 summarises the review selection process in a PRIMSA diagram. The searches yielded 3 942 results after duplicates were removed, of which 2 646 were not directly relevant to the outcomes of the review and were excluded at the title and abstract screening stage. Full texts were reviewed for 1 296 studies to assess eligibility, and 272 were eventually included in this review. The grey literature search yielded 5 additional relevant results.
Figure C.1. PRISMA diagram for literature review
Summary of specialty areas
A total of 240 reviews addressed specific specialty or disease areas, while 32 reviews either did not address a specific specialty or addressed more than two areas. Reviews that assessed two diseases or specialty areas were included as double listings in Table C.1 below resulting in a total of 283 entries. The included studies undertook research across 28 medical specialties/specialty groups, with 32 (11.3%) of papers not specifying the medical specialty or covering more than two specialties (Table C.1). The most common specialties in the included reviews were cardiology (9.5%), mental health, psychology, and psychiatry (9.5%), and surgery (7.4%). The least common were dentistry (0.3%), haematology (0.3%), and neonatal care (0.1%).
Table C.1. Summary of specialty areas in included studies
Note: Reviews that assess two diseases or specialty areas are included as double listings, resulting in a total of 283 entries.
Brief summary of findings
The included studies most commonly reported on user satisfaction with telemedicine (142 studies), which includes patient, carer, and clinician satisfaction. However, many studies reported on the cost of telemedicine (87 studies), the impact of telemedicine on patient safety (45 studies), telemedicine use by socio‑economic and demographic group (37 studies), and telemedicine use by type of technology (22 studies). Narrative results of these five areas are presented below.
User satisfaction with telemedicine services
Patient and carer satisfaction with telemedicine
Of the 142 studies included in the systematic review, 110 (77.5%) focused only on patients and 21 (14.8%) examined both patients and clinicians. Ten (7.0%) studies included in the systematic review did not specify the user group in their results. Overall, there were positive findings in relation to patient and carer satisfaction with telemedicine, with 90 (68.7%) of the 131 studies investigating patient and carer satisfaction reporting either positive satisfaction, usability, or acceptability. Two (1.5%) studies reported poor patient and carer satisfaction with telemedicine, and 23 (17.6%) studies found that when telemedicine was compared with a control group, there was no difference in patient and carer satisfaction between options. Twelve (9.2%) studies had mixed results, while four (3.1%) had unclear findings due to inconsistent reporting of patient and carer satisfaction, or unclear results.
Drivers behind patient and carer satisfaction with telemedicine
The primary drivers of patient and carer satisfaction with telemedicine were the ease of use, access, and convenience that it provides to patients. This was particularly so for the management of chronic conditions, reducing the need of patients to attend multiple appointments in-person while in the safety and security of their home. The ability of telemedicine to reach people who may not engage in in-person care, and how it can enhance the clinician-patient relationship, were also discussed. Conversely, a common barrier identified was the heterogeneity of patients and their changing care and service needs. As a result, not all patients would be suitable candidates for telemedicine, especially given their changing needs over time. Technology infrastructure, particularly user interface, internet connection and software, were highlighted as areas for improvement. One of the two studies found via the grey literature review echoed findings that patient and carer satisfaction was associated with the technical performance of telemedicine systems (Wherton and Greenhalgh, 2020[1]). Health inequities related to telemedicine require further investigation in the future, as one study highlighted higher satisfaction in higher income households.
Clinician satisfaction with telemedicine
Only one (0.8%) study focused solely on clinicians while, as mentioned above, 21 (15.7%) studies examined both patient and clinician satisfaction with telemedicine. Clinician satisfaction follows a similar pattern to patients, with 16 (72.2%) studies reporting positive satisfaction. Mixed results were found in 3 (13.6%) studies as varying satisfaction levels were reported among clinicians, while 1 (4.5%) study found no significant difference in clinician satisfaction. Two studies (9.1%) had unclear results due to inadequate data collection. The grey literature search led to the discovery of an additional paper reporting clinician satisfaction based on a systematic review where 89% of included studies showed moderate to high levels of physician satisfaction with telemedicine (Hoff and Lee, 2022[2]).
Drivers behind clinician satisfaction with telemedicine
Themes identified in clinician satisfaction with telemedicine use were like those identified for patients, such as the need for individualised services and the ability to effectively manage chronic conditions remotely. Factors that were associated with greater satisfaction in clinicians included telemedicine call quality and comfort using telemedicine. Concerns that were unique to clinicians included the fear that telemedicine would result in compromised safety and quality of care, although these perceptions improved over time.
Impact of telemedicine services on patient safety
The patient safety impact of telemedicine services was reported in 45 studies identified through the systematic review. Of these, 20 (44.4%) assessed the adverse events or errors that occurred using telemedicine interventions, 4 (8.8%) assessed hospital readmission and/or deaths, 3 (6.6%) assessed the safety of technology, 3 (6.6%) assessed patient/clinician perceptions of safety, and 2 (4.4%) other safety outcomes not specified. Thirteen (28.8%) papers did not provide clear information on how patient safety was assessed.
Of the 28 original research papers which reported on patient safety, the majority (71.4%) reported positive safety outcomes or outcomes comparable with equivalent in-person treatment. Four (14.2%) of the studies reported negative outcomes or poorer outcomes compared to in-person care. Of these, two studies (from Canada and Germany) related to medication and drug errors (Amkreutz et al., 2020[3]; McGillion et al., 2021[4]) and two studies (from Germany and the United States) were related to technological issues with the potential for patient harm (e.g. poor or missing audio, interrupted sessions, problems with data storage) (Willems et al., 2021[5]; Rametta et al., 2020[6]).
Safety and quality of telehealth interventions was also assessed using subjective reporting by either clinicians or patients. Three systematic reviews reported that the quality was satisfactory, two reported that patients found virtual clinics safe and of good quality (Vinade Chagas et al., 2021[7]; Edison et al., 2020[8]) and one reported that 81% of clinical staff considered the quality of telemedicine good or excellent (Appleton et al., 2021[9]).
Notably, one systematic review found that the impact of electronic visits (involving asynchronous communication between clinicians and patients) on quality of care varied across conditions. It was equivalent or better for chronic conditions, but variable quality was observed in infection management (Nguyen et al., 2021[10]).
Utilisation rates by socio‑economic status and demographic group
Telemedicine use by socio‑economic status and demographic group was reported in 37 studies. Of these, two were systematic reviews and 35 were original research studies. Twenty-six (74.2%) were undertaken in the United States, and the remaining nine (25.7%) were undertaken in Australia, Chile, Colombia, Finland, France, Germany, Italy, and the Netherlands. Given the heterogeneity of patient cohorts across the studies, further research is required to determine whether the usage trends seen in the results are consistent across OECD countries, medical specialties, and patient groups.
Findings by specific socio‑economic and demographic variable
Twenty-five (24.2%) studies reported on utilisation by age. However, based on the heterogeneity of patient cohorts, it is not possible to draw meaningful conclusions on whether telemedicine is used more often by specific age groups.
Twenty-one studies (20.3%) reported utilisation by sex/gender. In this review, the terms sex and gender are used interchangeably as there was no consistency among the studies. Twelve of the studies (57.4%) found that telemedicine was more often used by women, while four (19%) reported telemedicine more often used by men, three (14.2%) reported no difference in utilisation or mixed results, and two studies (9.5%) were unclear or did not make a comparison.
Seventeen studies (16.5%) reported utilisation by race/ethnicity. Again, the terms race and ethnicity are used interchangeably as there was no consistency among the studies. While there were six studies (35.2%) which reported unclear findings or did not make a comparison in usage between ethnic groups, eight studies (47%) found that White patients were more likely to use telemedicine and three studies (17.6%) found no significant difference in telemedicine usage. No studies found greater telemedicine use among ethnic minority groups.
Thirteen studies reported utilisation by location. Of these, three studies (23%), all from the United States, measured differences in use between national geographic regions, three (23%) measured differences based on the patients’ distance to medical facilities, and seven (53.8%) measured differences in use between populations living in urban and rural areas. Of these seven studies, six (85.7%) reported greater use of telemedicine among urban dwelling patients and one study reported unclear findings.
Beyond the most explored socio‑economic and demographic variables in the systematic review results, nine studies (8.7%) reported results on use of telemedicine by type of health insurance, six (5.8%) by educational attainment level, six (5.8%) by socio‑economic status (SES), four (3.8%) by first-language of patients, and two (1.9%) by employment status. One study discovered through the grey literature search also explored telemedicine use by income level and found no significant difference in use (Havasy, 2020[11]).
Several of the included studies noted that through the COVID‑19 pandemic, the use of telemedicine increased among men, ethnic minorities, older people, and those from a lower SES.
Utilisation rates by type of telemedicine
Telemedicine use by type of technology was the area of focus least reported in the literature. Of the 22 studies which presented results on telemedicine use by type of technology, 12 (54.5%) were systematic reviews, one (4.5%) was a systematic review and meta‑analysis, and nine (40.9%) were original research studies. Six (66.6%) of the original research studies were undertaken in the United States, and the remaining three (44.3%) were undertaken in Australia, Canada, and France.
Of the 22 studies reporting on type of telemedicine, 21 described teleconsultations and seven discussed remote monitoring. Seven (33.3%) of the 21 studies reporting on teleconsultations, reported the use of telephone and video technology in delivering telemedicine only, while 10 (47.62%) reported the use of telephone and video technology as well as other technologies, including remote monitoring technologies, virtual/augmented reality, online programmes, smartphone applications, Internet/Internet of things (IoT)/cloud computing, and social media/network. One (4.76%) of the studies reported on the broader use of technologies only and three (14.2%) of the studies did not specify the specific technology used to deliver teleconsultations.
Of the three studies that described the type of tele‑transmission used to facilitate teleconsultation, all reported the use of a combination of synchronous and asynchronous telemedicine, with synchronous being the primary form of contact. For example, one study (from the United States) noted that, of all telemedicine visits, 179 were synchronous video and audio, 15 were synchronous audio‑only, and 4 were asynchronous interprofessional consults (Moss et al., 2021[12]).
Cost analysis and pricing of telemedicine services
The cost of telemedicine services was reported in 87 studies identified through the systematic review. Of these, 45 were original research papers where 31 reported costs only and 14 reported cost-effectiveness. Twenty-five (80.6%) of the 31 papers which only calculated costs reported that telemedicine care reduced costs as compared to usual in-person care (including costs incurred by patients which were measured in four of the studies), 3 (9.6%) reported no significant difference in costs,1 (3.2%) reported increased costs using telemedicine, and two (6.4%) were unclear in their results. On cost-effectiveness, 12 (85.7%) of the 14 studies reported that telemedicine was more cost-effective than usual in-person care, 1 (7.1%) study found no difference in cost-effectiveness, and 1 (7.1%) study was unclear in the reporting of results.
Drivers behind cost reductions and cost-effectiveness results
The primary drivers of cost reductions were time savings for both patient and provider; for the former, time savings included travel time and reduced need to take time off work. This was particularly so for the management of chronic conditions, reducing the need for patients to attend multiple appointments in-person. On the provider side, cost savings were also derived from reduced unplanned hospital visits and admissions, as well as subsequent lower ambulance transportation fees.
Conversely, a common cost barrier identified was the costs required to set up technology to facilitate telemedicine, including virtual consultation platforms, and their longer-term maintenance. As a result, cost-effectiveness analysis is required as cost savings varied substantially between studies, with providers required to consider variables, such as treating a minimum number of patients, or treating patients living a set geographic distance from the health facility, to determine the cost savings from telemedicine use.
Detailed overview of findings
Tables D.1 to D.5 in Annex D provide more detailed information on the studies included in the systematic review.
References
[3] Amkreutz, J. et al. (2020), “Medication safety in a German telemedicine centre: Implementation of a telepharmaceutical expert consultation in addition to existing tele-intensive care unit services.”, Journal of telemedicine and telecare, Vol. 26/1-2, pp. 105-112, https://doi.org/10.1177/1357633X18799796.
[9] Appleton, R. et al. (2021), “Implementation, Adoption, and Perceptions of Telemental Health During the COVID-19 Pandemic: Systematic Review.”, Journal of medical Internet research, Vol. 23/12, pp. e31746-e31746, https://doi.org/10.2196/31746.
[8] Edison, M. et al. (2020), “Understanding virtual urology clinics: a systematic review.”, BJU international, Vol. 126/5, pp. 536-546, https://doi.org/10.1111/bju.15125.
[11] Havasy, R. (2020), Consumer Perspectives on Telehealth, https://www.himssmedia.com/sites/himssmedia.com/files/himss_acceleratehealth_consumertelehealth_fall2020_final.pdf.
[2] Hoff, T. and D. Lee (2022), “Physician Satisfaction With Telehealth: A Systematic Review and Agenda for Future Research”, Quality Management in Health Care, Vol. 31/3, pp. 160-169, https://doi.org/10.1097/QMH.0000000000000359.
[4] McGillion, M. et al. (2021), “Post-discharge after surgery Virtual Care with Remote Automated Monitoring-1 (PVC-RAM-1) technology versus standard care: randomised controlled trial”, BMJ (Clinical research ed.), Vol. 374, pp. n2209-n2209, https://doi.org/10.1136/bmj.n2209.
[12] Moss, H. et al. (2021), “The Impact of COVID-19 on Neuro-Ophthalmology Office Visits and Adoption of Telemedicine Services.”, Journal of neuro-ophthalmology : the official journal of the North American Neuro-Ophthalmology Society, Vol. 41/3, pp. 362-367, https://doi.org/10.1097/WNO.0000000000001356.
[10] Nguyen, O. et al. (2021), “Impact of Asynchronous Electronic Communication-Based Visits on Clinical Outcomes and Health Care Delivery: Systematic Review.”, Journal of medical Internet research, Vol. 23/5, pp. e27531-e27531, https://doi.org/10.2196/27531.
[6] Rametta, S. et al. (2020), “Analyzing 2,589 child neurology telehealth encounters necessitated by the COVID-19 pandemic.”, Neurology, Vol. 95/9, pp. e1257-e1266, https://doi.org/10.1212/WNL.0000000000010010.
[7] Vinade Chagas, M. et al. (2021), “The use of telemedicine in the PICU: A systematic review and meta-analysis.”, PloS one, Vol. 16/5, pp. e0252409-e0252409, https://doi.org/10.1371/journal.pone.0252409.
[1] Wherton, J. and T. Greenhalgh (2020), Evaluation of the Attend Anywhere / Near Me video consulting service in Scotland, 2019-20.
[5] Willems, L. et al. (2021), “Satisfaction with and reliability of in-hospital video-EEG monitoring systems in epilepsy diagnosis - A German multicenter experience.”, Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology, Vol. 132/9, pp. 2317-2322, https://doi.org/10.1016/j.clinph.2021.04.020.