Coronavirus 2019 disease (COVID-19) is having an enormous impact on the practice of dermatology, with a considerable reduction in face-to-face consultations in favor of teledermatology. Once the SARS-CoV-2 outbreak was reclassified as a pandemic on March 11, 2020, almost the whole of Hospital Universitario La Paz (HULP), Madrid, Spain was transformed in order to prioritize care for patients with the disease. The transformation entailed a reordering of care positions, thus making treatment for patients with dermatologic conditions a genuine challenge.

Within the framework of an active national care alert, which led to a mandatory lockdown for the whole population, our activity was restricted to the management of dermatologic emergencies and interdepartmental consultations. In this context, the United States Centers for Disease Control and Prevention and other public bodies stated that telemedicine should be considered part of the health system response to COVID-19.1–3

Teledermatology is defined as the clinical evaluation of skin lesions by dermatologists using telemedicine techniques. Teledermatology can be asynchronous (store-and-forward), synchronous (real-time), or hybrid (an asynchronous phase and a synchronous phase).4 It is generally indirect, between primary care and dermatology, with or without associated dermoscopic images.5 The direct-to-patient store-and-forward modality involves a direct digital consultation between the patient and the dermatologist based on images captured using the patient’s smartphone. This modality has the disadvantage that both the image quality and the information necessary for appropriate history taking are poor.6,7

The advantages of teledermatology have been demonstrated in multiple studies and comprise the ability to reduce the waiting list, act as a triage system, and enable remote access to a dermatologist.6,5 Spain is among the countries that publishes most articles on teledermatology, which is well established in more than 25% of all Spanish public dermatology departments.8,9

Platforms such as WhatsApp and email were used during lockdown, although sending images via insecure platforms is not recommended.10–13 Therefore, we performed a pilot study to evaluate the application of direct hybrid teledermatology for the care of new patients. We examined appointments scheduled during the period running from the declaration of the state of emergency to the initial easing of restrictions, with the entry of phase 0 in the Autonomous Community of Madrid.

The study was approved by the Clinical Research Ethics Committee of HULP on April 20, 2020.

From the Admissions Department of HULP, we telephoned 2316 patients, all of whom had been scheduled for appointments as new patients at HULP and the PSCs. Patients were gradually called from the date the study was approved by the local ethics committee on April 20, 2020 until May 4, 2020. A total of 1497 (64.6%) patients agreed to participate. In the email inviting patients to download the app, we provided the telephone number of the Dermatology Office, as well as the email of the app helpdesk. Patients who did not wish to participate or could not participate were offered the option of making an appointment that would become active once the face-to-face appointment schedule was opened.

The reasons for not participating in the study among the remaining 819 patients were not having a smartphone (9%, n = 73), not having email (24%, n = 193), not having visible lesions on the skin (23%, n = 185), or simply not wishing to participate (45%, n = 368).

Of the 1497 who agreed to participate, 896 patients completed the registration process, although only 452 completed a remote consultation with their dermatologist (30%). Of these, 374 attached an image (Fig. 1).

Figure 1.

Patient recruitment flowchart.

(0.42MB).

Diagnosis

As shown in Fig. 2, the diagnostic categories were as follows: melanocytic lesions, inflammation, infection, cancer-precancer, pediatric disease, and other. The cancer-precancer category included 34 patients, of whom 18 (4.8%; 95% CI, 2.6%-7%) were considered to have cancer.

Figure 2.

Percentage of patients in each diagnostic category.

(0.23MB).

Table 1 summarizes the most frequent diagnoses. Fig. 3 provides a few examples.

Table 1.

Most Common Diagnoses Using Teledermatology in the Study Sample.

Diagnosis	No.	Percentage (95% CI)
Nevus	74	20 (15.9%-24%)
Acne	36	9.6 (6.6%-12.5%)
Eczema	29	7.7 (5%-10.4%)
Seborrheic keratosis	26	6.9 (4.3%-45%)
Actinic keratosis	14	3.7 (1.7%-5.6%)
Basal cell carcinoma	10	2.6 (0.9%-4.2%)

Figure 3.

Images sent via the app. A, Pityriasis rosea. B, Scabies. C, Pigmented lesion requiring assessment by dermoscopy. D, Impetigo.

(0.19MB).

One case was classified as a dermatologic emergency and involved an atopic patient who presented with erythroderma. Additional tests (ordered online) were required in 105 patients (28.3%).

Diagnostic Performance

Of the patients who sent an image (n = 374), a diagnosis could be made in 87.1% (95% CI, 83.7%-90.5%); in the dermatologist’s opinion, this was probable in 48.6% (95% CI, 43.5%-53.6%) and definitive in 38.5% (95% CI, 33.5%-43.4%).

Confidence in the diagnosis made was evaluated subjectively by each dermatologist, with a score ≥ 7/10 in 290 cases (77.5%, 95% CI, 73.2%-81.7%). Image quality was considered sufficient in 52.1% (95% CI, 47%-57.1%) of cases.

In contrast, diagnosis was not possible in 12.8% (95% CI, 9.4%-16.1%) of cases (n = 49) for the following reasons: the image was not consistent with the consultation (58.3%), image quality was poor (22.9%), and additional testing was necessary (18.7%).

Diagnostic performance and confidence in the diagnosis differed by area (Fig. 4).

Figure 4.

Diagnostic ability and degree of confidence on a scale of 1 to 10 in the diagnosis made according to the diagnostic categories.

(0.4MB).

Impact on Consultations

Of the 374 patients who provided an image for their consultation, 17% (95% CI, 13.1%-20.8%) (n = 64) were referred for follow-up in primary care. Continued care was necessary in 309 patients; this was considered a priority (under 1 month) in 15% (95% CI, 11.3%-18.6%) (n = 56). The face-to-face visit was delayed 3 months in 50% (n = 190) of patients and up to 6 months in 17% (95% CI, 13.1%-20.8%) (n = 64). Therefore, face-to-face visits were avoided for at least 3 months in 85% (95% CI, 81.2%-88.6%) of patients (n = 318) (Fig. 5).

Figure 5.

Continuity of care. Percentage of discharges and check-up visits at 1, 3, and 6 months.

(0.19MB).

Treatment was prescribed by telephone in 101 patients (27%; 95% CI, 22.5%-31.5%), and 8 patients were included directly on the surgical waiting list.

Patients scored their satisfaction with the online visit on a scale of 1 to 5. The average score obtained by the 16 dermatologists was 4.5/5.

The dermatologists scored their satisfaction according to the items consulted (Fig. 6).

Figure 6.

Likert scale for satisfaction according to the disease. TA indicates totally agree; A, agree; N, neither agree nor disagree; D, disagree; TD, totally disagree.

(0.42MB).

The main objective of the present study was to determine the percentage of face-to-face consultations that could be avoided using teledermatology during lockdown. In the sample studied here, we were able to completely avoid face-to-face consultations with a specialist in 17% of cases and to postpone face-to-face consultations by at least 3 months in an additional 68% of patients. Furthermore, we prioritized care for those patients who needed it, even when required within 24 hours, as occurred in the case of a patient with erythroderma.

According to data from the DIADERM study, the percentage of discharges with check-ups by primary care and check-ups by a dermatologist are 6% and 65%.14 The same percentages for teledermatology are 33% and 42%.8 In both cases, the need for continued care is considerable, albeit to a lesser extent than in our study (82%). Very different data were reported from a recent store-and-forward teledermatology study carried out in Germany that included 1364 patients and where 64% of face-to-face consultations were avoided.15 These data contrast with those of our study, although the difference can be explained by the fact that, in our protocol, we established that only patients with a confirmed diagnosis and trivial disease would be discharged for follow-up in primary care; the remainder were given appointments for check-ups with the same specialist. While this protocol was designed in this way to ensure patient safety (given the lack of experience with teledermatology in our center), it prevents us from drawing reliable conclusions on the efficacy of teledermatology as a screening tool. It is also important to remember that direct store-and-forward teledermatology data are usually inferior to those obtained by indirect store-and-forward teledermatology, where images and the associated use of dermoscopy guarantee better diagnostic performance, thus avoiding up to 58% of face-to-face consultations.6

Nevertheless, our diagnostic performance was high: we were able to make a diagnosis for 87.1% of patients with a considerable degree of confidence. These results are similar to those reported by Sondermann et al.,15 who were able to diagnose 90.3% of patients remotely. Consistent with data form other series,6,16 we found that confidence in the diagnosis was higher for the categories infection, cancer-precancer, and inflammation and lower in melanocytic lesions and pediatric diseases.

The most frequent diagnoses were melanocytic nevus, acne, eczema, actinic keratosis, seborrheic keratosis, and basal cell carcinoma. These diagnoses are illustrative of the most common main diagnoses in dermatology in the PSCs and are similar to those reported elsewhere, such as in DIADERM, where the most common diagnoses were actinic keratosis (8.2%), basal cell carcinoma (8.1%), and melanocytic nevus (7.5%).6,8,15

Image quality was considered adequate for diagnosis in 52.1% of cases, which is consistent with data reported from a recent study on teledermatology based on WhatsApp, where 54.3% of dermatologists considered image quality sufficient for diagnosis. However, this is lower than the 81% recorded for indirect store-and-forward teledermatology.6 O’Connor et al.16 reported the results of a pediatric study that compared the diagnostic accuracy of teledermatology in 2 groups, i.e., with and without previous instructions on imaging. The authors reported no significant differences between the groups.

The results of the survey completed by dermatologists from HULP showed that the overall level of satisfaction was considered good by 68.7%. It is worthy of mention that 81.25% agreed/totally agreed with a high degree of confidence in the diagnosis, 74% agreed/totally agreed that teledermatology had a positive impact on the running of their department, and 87.2% would apply this approach to new areas of therapy.

As for patient satisfaction, the average score given to the 16 dermatologists was 4.5/5, that is, slightly higher than in a recent study of 243 teledermatology consultations, where the average score was 4.38/5.17 This finding supports publications that show a high index of patient satisfaction with direct-to-consumer telemedicine systems.18,19

Online technology is well implemented in Spain,20 as seen in the high number of smartphone users in our study. Nevertheless, only 1 in every 4 patients recruited finally sent an image as part of their consultation; of these, 44% were educated to university level. We did not perform a structured analysis of the reasons why this high percentage of patients did not manage to complete a visit, although according to the data gathered from the telephone calls to the dermatology office and the app helpdesk, the main reasons were technological difficulties, lack of trust in the security of the process, and resolution of the skin disease. We believe that these data may have improved if we had used a system that opened the patient’s camera directly by means of an SMS, without the need to download an app.

Another limitation of our study is that we did not analyze diagnostic agreement in the application of teledermatology. Therefore, we cannot guarantee the diagnostic reliability of the approach. However, Sondermann et al.15 performed a random assessment of diagnostic agreement in 100 patients and found it to be 97%.

Very few studies report the diagnostic performance of direct store-and-forward teledermatology,15,17,21 and we were unable to find studies performed during the COVID-19 lockdown.

In conclusion, teledermatology is an effective tool that enables the practice of dermatology to continue during a pandemic in a way that is satisfactory for both patients and physicians. Implementation of this form of teledermatology could complement face-to-face care activity, especially in patients undergoing follow-up. However, the development of an efficient teledermatology system necessarily implies integration of digital imaging into medical information systems; therefore, we must ensure the necessary infrastructure for secure storage and transmission of images, without unnecessarily increasing the workload of clinicians.

The authors declare that they have no conflicts of interest.