We included fourteen systematic reviews published between 2004 and 2023 were finally included in this evidence map (5, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22). Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram (23).
Characteristics of included systematic reviews
Table I shows the types of SRs included. Two were costs/economic evaluation reviews, two diagnostic test accuracy reviews, one experiential review, one a combination of experiential and psychometric, one combined diagnostic test accuracy with a costs/economic evaluation, and seven scoping reviews.
Characteristics of the included systematic reviews are shown in table II. The reviews were undertaken in Europe, USA, Australia and Singapore.
The number of electronic bibliographic databases used in the SRs ranged from 1 to 14 (Table III). Five used hand-searching (5, 10, 13, 14, 15). Only three SRs included a statement about prior registration or publication of a protocol (5, 16, 21). All but one review (10) included a statement regarding any conflict of interest, and three reviews reported conflicts of interest (5, 17, 20). Five reviews did not report a funding statement (10, 11, 12, 13, 21), five reviews stated there was no financial support for the review (14, 17, 18, 19, 21, 22), and three reported receiving funding from medical councils or government agencies (5, 15, 16) (Fig. 2).
Dermatological conditions
Eight of the SRs did not specify the dermatological conditions of interest (10, 11, 12, 15, 18, 19, 21, 22). The remaining focused on burns (14, 16), rashes (13, 17), skin lesions (13), psoriasis (5, 17), skin cancer and other associated indications (16, 17), dermatologic complications amongst oncological patients (21), suspected malignant lesions (5), nonmelanoma skin cancer or fast-growth vascular tumour suitable for surgery under local anaesthesia (5), acne (17), wounds (17), atopic dermatitis (17), tinea (17), leprosy (17), circumscribed lesions (13), pigmented and non-pigmented skin lesions (13), chronic inflammatory dermatoses (21), dermatological consultations for suspected COVID-19 (21), and any or unspecified conditions (5, 13, 17). As a single diagnosis acne was addressed most often, it featured specifically in four reviews (Fig. 3).
Overlap of Primary Studies
The overall CCA was 4.9% which suggested a slight overlap of primary studies in the 14 included systematic reviews, and the overlap between pairs of studies ranged between 0.0–43.9%. Of the 91 pairs of systematic reviews, three pairs were categorised as having very high overlap (i.e. 11 and 13 = 43.9%; 13 and 17 = 25.9%; 11 and 17 = 22.2%) and four pairs were categorised as having high overlap (i.e. 20 and 21 = 13.3%; 13 and 16 = 12.4%; 10 and 11 = 10.7%; 15 and 17 = 10/1%). Nine pairs were categorised as having moderate overlap, and 75 pairs were categorised as having none to slight overlap (Table 4).
Quality assessment of the studies included in the systematic reviews
Seven of the 14 SRs reported conducting a quality assessment (5, 13, 15, 16, 19. 21, 22) (Fig. 4). Tools used were the Quality Assessment of Diagnostic Accuracy Studies (QUADAS and QUADAS-2) (13, 15, 16), the Consolidated Health Economic Evaluation Reporting Standards checklist (CHEERS) (5), the rating scheme provided by the Oxford Centre for Evidence-Based Medicine (22), and the abridged version (19). In one instance the review (19) did not describe their findings in detail but commented on a low risk of bias in their included primary studies. Another review (21) did not report the quality assessment in the published article but did append their results on Mendeley Data (https://doi.org/10.17632/xd6ftfpgmc.1), while another did not report any quality assessment results (22).
Among the four reviews that reported quality assessment in detail, three described that at least half of the primary studies were at risk of bias (5, 13, 16). The two reviews using the 14 QUADAS quality assessment made very different observations. In one, the proportion of primary studies that reported at least 10 QUADAS items was only 29% of 78 primary studies (13), in the other one it was 85% of 26 primary studies (15). The systematic review that used the QUADAS-2 reported their findings in detail (16), highlighting that at least half of the 22 included primary studies were at high or unclear risk of bias for participant selection, reference standard, and flow and timing domains, while the majority were at low risk for the index test. In summary they concluded that the- quality of the studies included was of concern. Another systematic review of 11 primary studies (5) reported a wide range of quality scores using the CHEERS checklist (7 to 21 out of a total score of 24). The authors reported that the lower scores were due to a failure of the primary studies to report or discuss economic principles or justify the analytic approach used (5). For the store and forward studies, the most relevant principles that were not included were study duration, appropriate financial conversions, and financial referencing.
Main findings of two systematic reviews addressing cost/economic evaluation
Published in 2016 (5) and 2017 (18) these two reviews had 5.6% overlap. Snoswell et al. (5) concluded that, whilst the evidence was sparse, SF teledermatology can be cost-effective when used as a triage mechanism to reduce the number of conventional face-to-face appointment. They identified three studies supporting increasing cost-effectiveness of SF teledermatology when patients need to travel long distances to access dermatology services.
Fuertes-Guiro and Girrabent-Farres’s review (18) found that a teledermatology consultation requires more time (7.54 minutes extra) than a conventional consultation to make a diagnosis and management plan. In 2017 this represented an additional cost of 29.25 Euros for remote consultation; in addition, SF teledermatology was less costly than LI teledermatology. The authors observe that whilst there are some cost utility and cost effectiveness studies in the literature that indicate that telemedicine can reduce costs, these have not always been attentive to the cost of the dermatologist’s time, i.e. opportunity cost
Main findings of systematic reviews addressing accuracy of telemedicine
One review (15) found that the diagnostic accuracy of mobile phone based teledermatology was inferior to traditional face-to-face dermatology when comparing the clinical diagnosis with histopathology (weighted mean absolute difference 7.2%). Diagnostic concordance, defined as the agreement between teledermatology diagnosis and face-to-face teledermatology diagnosis, was generally good, and higher than the levels previously reported for SF. Only one study addressed management accuracy (matching management with histopathology) but found very high agreement when comparing the management decision based on teledermatology dermatoscopy and clinical images with histological based management. Overall management concordance rates were very good, with a weighted average concordance of 80%. Whilst the review concluded that mobile teledermatology has yet to achieve a level of accuracy to replace conventional dermatology diagnosis, they described how over time mobile phone technology had developed for data capture, transmission, display and storage improving the accessibility and convenience of mobile teledermatology.
The other systematic review (n = 22) addressed accuracy focussed on teledermatology for detecting skin cancer in adults (16). Data from four studies suggests that fewer than 7% of malignant skin lesions were missed by teledermatology. However, the applicability of these findings to the development of clinical services may be limited as participants were largely recruited from secondary or tertiary care clinics rather than the primary care setting where teledermatology is often used to triage which patients require referral to secondary care.
Main findings of experiential systematic reviews of patient satisfaction
Demiris, Speedie and Hicks looked at the quality of evidence about patient satisfaction with teledermatology (10). They identified 13 primary studies which used self-administered questionnaires to measure patient satisfaction and one study which used phone interviews. The psychometric evaluation of the existing instruments was weak: content, construct or reliability testing were not reported in any of the primary studies. Patients accepted teledermatology as a mode of care delivery but had concerns relating to privacy, embarrassment of being photographed, the limited opportunities to express their problems and concerns, completeness of information transmitted, anxiety about the unfamiliar technology, and frustration with technical problems. The authors noted that the definition of satisfaction differed across the primary studies. They suggested that SF and LI need distinct evaluation tools.
Mounessa et al.’ reviewed 40 studies focussing on patient and provider satisfaction with SF and LI teledermatology (19). Dissatisfaction with SF teledermatology was reported in 1 of 24 studies assessing patients and 3 out of 17 studies assessing teledermatology providers; it was noted that eight of these studies assessed both patient and teledermatology provider satisfaction with SF teledermatology. For SF services 96% of patients and 82% of providers were satisfied, and for LI teledermatology 89% of patients and 100% of providers were satisfied. It was noted that two LI teledermatology studies surveyed non-physician providers, and five studies included both patient and teledermatology providers.
Main findings of combination-type review
One systematic review (13) was a combination of a diagnostic test accuracy and cost/economic evaluation. Using 78 primary studies, Warshaw et al. (13) compared the diagnostic accuracy, clinical management, clinical outcomes, and the cost between teledermatology and clinic dermatology (13). The authors reported that clinic dermatology had higher diagnostic accuracy than SF teledermatology (i.e. six studies, 19% better) and LI teledermatology (i.e. 11 studies, 11% better) that teledermatology accuracy rates improved by up to 15% with teledermatoscopy, and that the diagnostic concordance with clinic dermatology of SF teledermatology was good, but better for LI teledermatology. Regarding management accuracy, the overall rates were similar but teledermatology and teledermatoscopy were inferior to clinic dermatology for malignant lesions. Regarding management concordance, rates were moderate to very good for both SF and LI teledermatology. The authors reported that there was insufficient evidence to evaluate the effect of teledermatology on clinical outcomes, and that patient satisfaction and preferences for teledermatology were comparable with clinic dermatology. The time to treatment was significantly shorter and in-person visits to the dermatology clinic were avoided when patients had a teledermatology consultation. The SR reported that teledermatology was cost-effective compared to clinic dermatology onr key considerations such as distance travelled by the patient, the volume of teledermatology, and the costs of clinic dermatology. However, the authors were unable to pool the data for analysis because these cost studies analysed different outcome parameters.
Main findings of scoping reviews
The first scoping review (N = 99 studies, 101 publications) aimed to describe the maturity status of teledermatology evaluation research and to explore the outcome measures (11). It reported that while the number of feasibility studies increased, there was a lack of randomised controlled trials (RCTs), simulation cost studies, and post-implementation studies. Regarding outcome measures, the authors reported diagnostic accuracy as the most common (53 studies). Regarding study design, there were 43 intervention studies with the same patients as controls, 30 studies using an uncontrolled study design, 12 RCTs, seven intervention studies with diff in the primary studies erent patients as controls, and seven observational studies, SF teledermatology was most frequently used in the primary studies (62%), followed by LI teledermatology (30%), and combination of SF and LI teledermatology (2%). No data was available for the remaining studies.
The second scoping review (N = 11) aimed to provide an overview of the use of tertiary teledermatology (12), identifying four categories of tertiary teledermatology use: expertise (i.e. seeking advice from a dermatologist specialised in a specific area), continuing medical education (i.e. learning from other dermatologists), supervision of residents in training programmes, and second opinion from dermatologists. The review identified three modalities of use (i.e. teledermatology consultation application in seven studies, website in two studies, and email list in one study). Regarding the type of teledermatology used, seven primary studies used SF teledermatology, three used a combination of SF and LI teledermatology, but it was unclear what type of teledermatology was used in one study. Next, the authors reported that the outcome measure commonly reported was the effect of teledermatology on learning, followed by development cost, image quality, efficiency improvement, diagnostic validity, diagnostic reliability, diagnostic accuracy, patient satisfaction, and physician satisfaction.
The third scoping review included 24 primary studies and aimed to assess the evidence for the use of telemedicine in acute burn care and outpatient-based management (14). Of the 24 included studies, seven studies evaluated clinical decision-making for acute burn care, eight studies assessed technical feasibility and clinical validation, and nine studies evaluated outpatient care. Wallace et al. (14) also reported that 14 primary studies assessed SF teledermatology, seven assessed LI teledermatology, and three assessed a combination of SF and LI teledermatology. This review found that teledermatology for burn care was rated from comparable to face-to-face assessment and as a tool that could improve clinical decision-making. The authors added that patients were satisfied and benefited from cost-savings in time and travel, but healthcare providers benefited from limited cost-savings only when a large volume of teledermatology use. Regarding methodology, the authors commented that they did not find any RCTs, and of the 24 primary studies in their review only eight studies had controls. The primary studies in this review did not report a priori power calculation and were mainly subjective reports about teledermatology use rather than formal comparisons.
The fourth scoping review included in our evidence map review aimed to identify the use and current state of teledermatology across the world with regard to the geographical distribution of published studies, treated indications, research questions, and its reliability in diagnosis and therapy compared to classic face-to-face consultations (17). Based on 204 primary studies included in this review, Trettel et al. (17) reported that the most common category of research questions posed by them was validity, concordance or feasibility (n = 154), followed by effectiveness (i.e. comparison of teledermatology with face-to-face consultations; n = 33), costs, cost-effectiveness or cost–benefits of teledermatology (n = 24), quality of life (n = 4), and safety issues (n = 1). Regarding the comparison of teledermatology with face-to-face consultations, 138 studies reported that teledermatology was feasible, reliable, or effective under certain conditions, 34 studies found teledermatology to be superior to face-to-face consultations, 25 studies reported outcomes to be equivalent, and 15 studies reported outcomes to be inferior to face-to-face consultations. This scoping review included primary studies from a diverse range of clinical areas using teledermatology. Out of 204 primary studies, 127 studies reported either ‘various skin diseases’ or did not specify them, 52 studies focused on skin cancer and associated diagnoses, 11 studies focused on wounds, 7 studies were on psoriasis, 4 studies were on atopic dermatitis, and single studies addressed acne, leprosy, rash, or tinea. Lastly, regarding the application of teledermatology, 105 primary studies were unspecified general evaluations, 59 studies were about patient management (e.g. referral from primary care physician to dermatologist) and triage, 23 studies were about the diagnosis or consultation of patients in remote locations, 17 studies were about the monitoring and consultation of patients in the nursing home or home care setting, and one study was about emergency diagnosis.
The fifth scoping review aimed to summarise teledermatology studies performed during the COVID-19 pandemic in 2020 (20). Elsner (20) reported that two of the seven included studies were surveys among dermatologists showing that more than 80% offered teledermatology. The five remaining studies were retrospective cohort studies of low quality. Three of them investigated teledermatology in acne and inflammatory skin diseases, one the care of oncological patients with dermatological complications, and one teleconsultation in suspected COVID-19 cases. In all studies, teledermatology largely reduced the number of personal consultations. The review concludes that teledermatology could at least partly compensate for the limitations of in person dermatological care during the COVID-19 pandemic.
The sixth scoping review included 27 primary studies and aimed to analyse the global utilisation of teledermatology for patient care during the COVID-19 pandemic (21). Out of 27 primary studies, 10 studies were about SF teledermatology, 6 studies were about LI teledermatology, 8 studies were about the combination of SF and LI teledermatology, and 3 studies did not specify the type of teledermatology used. Loh et al (21) reported that teledermatology was useful in assessing and managing common ambulatory dermatoses. However, the authors highlighted concerns raised in the primary studies about low-quality images used in SF and LI teledermatology that reduced the accuracy of clinical assessments. Within the COVID-19 pandemic, the authors reported that teledermatology decreased unnecessary face-to-face consultations, which reduced the risk of infections and the use of personal protective supplies. The authors also reported that teledermatology was used for the diagnosis of cutaneous manifestations of COVID-19 infection and the follow-up of onco-dermatology patients.
The final scoping review included 15 primary studies and aimed to identify the satisfaction levels of patients and providers of synchronous teledermatology during the COVID-19 pandemic, including the likelihood of patients and providers use of teledermatology in the future (22). Most studies reported that patients were willing to continue using synchronous teledermatology. Regarding satisfaction levels, Miller and Jones (22) reported that patients were satisfied with the patient–provider relationship and increased access to care. It was also noted that patients were generally satisfied with the technical quality and sound quality of their teledermatology consultation sessions. However, patients were reportedly not satisfied with the physical examination or quality compared with face-to-face care. As for the teledermatology providers, the authors reported that they were generally dissatisfied with the video or image quality and the quality of the teledermatology visit compared with face-to-face care. Despite these areas of dissatisfaction, it was noted that both the patients and providers were satisfied with visits meeting patient needs. The authors also observed that most questions asked when assessing satisfaction levels focused on quality of care and technical aspects of teledermatology, rather than access to care, overall satisfaction, and the patient-provider relationship.