Step 1: diagnosis and management using still images; step 2: diagnosis and management using videoconferencing; and step 3: experts’ diagnosis or management comparison using still images or videoconferencing. ED indicates emergency department.
Duong TA, Cordoliani F, Julliard C, Bourrat E, Regnier S, de Pontual L, Leroy C, Gallula S, Aparicio C, Legendre N, Bagot M, Guibal F. Emergency Department Diagnosis and Management of Skin Diseases With Real-Time Teledermatologic Expertise. JAMA Dermatol. 2014;150(7):743-747. doi:10.1001/jamadermatol.2013.7792
Skin conditions are a common reason for patients to consult emergency department (ED) physicians.
To evaluate real-time teledermatologic expertise with the use of mobile telephones for the diagnosis and management of skin conditions in patients seen in the ED.
Design, Setting, and Participants
This observational study of adults who consecutively consulted in the ED for a dermatologic condition was conducted under routine conditions in the ED from May 1, 2008, through June 30, 2010.
Main Outcomes and Measures
Diagnosis agreement and management concordance.
One hundred eleven patients were enrolled in the study. Eighty-three patients (74.8%) were evaluated using videoconferencing. Dermatologic remote expertise invalidated, enlarged, or clarified ED physicians’ diagnosis and management in 75 of 110 cases (68.2%). Videoconferencing improved the diagnostic performance in 57 of 83 cases (68.7%) (P < 10−4). Management concordance was moderate between ED physicians and dermatologists for specialist consultation within 24 hours (κ, 0.49; 95% CI, 0.14-0.84) and immediate hospitalization (κ, 0.49; 95% CI, 0.41-0.57). Patients were significantly more often discharged by dermatologists (46.8% vs 39.1%) (P < 10−4).
Conclusions and Relevance
Compared with standard hardware, new-generation mobile devices reduce the cost of videoconferencing, increase the versatility of teledermatology, and decrease general practitioner investment time.
Skin conditions are the most frequent reason for patients to consult primary health care professionals, reaching up to 15% of general practitioners.1 In the emergency department (ED), skin conditions account for 9.2% to 40% of the consultations for children2,3 and up to 10% for adults.4,5 In a 2009 study in Singapore,4 among 4061 patients presenting with a skin condition of 157 527 total patients, the most common diagnoses were herpes zoster infection (20.8%) and dermatitis or eczema (11.6%).
Comparisons of dermatologic diagnostic performances between primary care physicians and dermatologists reveal a significant difference, with 93% correct diagnoses for dermatologists vs 50% for nondermatologists.6 The development of tools to provide access to specialists should improve quality of care and cost of dermatologic conditions. Teledermatology applications include live interactive (ie, videoconferencing) and/or store and forward (ie, still images) technologies.7 Diagnostic concordance with an in-person visit is slightly higher with live interactive systems (71%) than store-and-forward technologies (66%).7 Implementing store-and-forward technologies has already been successfully experimented with in the ED, and implementation of videoconferencing technologies using new-generation mobile devices should be considered a relevant and cost-effective option.8 The aim of our study was to evaluate real-time teledermatologic expertise performed with mobile telephones in the diagnosis and management of skin conditions seen in the ED.
This observational study was conducted under routine conditions from 9 am to 5 pm in the ED of 4 French hospitals (Hôpital Jean Verdier, Hôpital Louis Mourier, Hôpital Saint Louis-Lariboisière, and Hôpital Saint Antoine). Patients were adults (age ≥18 years) who consecutively consulted in the ED on dermatologic conditions. The study was divided into 3 steps (Figure). The first step recorded the dermatologists’ and ED physicians’ diagnosis and management using still digital clinical images taken and sent via the same mobile telephone. The second step recorded the dermatologists’ and ED physicians’ diagnosis and management using videoconferencing via the same mobile telephone used to take and send the still digital clinical images. The third step was the evaluation of the diagnosis and management concordance between ED physicians and dermatologists using either still digital clinical images or videoconferencing technologies. Three board-certified dermatologists (F.C., C.J., and E.B.) performed this last step.
All patients gave written informed consent to participate in this research. This study was approved by the CPP Paris VII local ethics committee.
The most recent smartphones available at the time of the study (Sony-Ericsson K800 Pi and Sharp 770 SH) were used to take and send digital clinical images and for videoconferencing between the ED physician and dermatologist. Digital images and videos were captured with an integrated camera (Sony-Ericsson K800 Pi, Sharp 770SH). Resolution was 5 × 106 pixels for digital images and 600 × 400 pixels per inch for video images. Images were transferred using a third-generation telephone network to another mobile telephone. The 3 dermatologists performed their diagnosis on the mobile telephone screen.
For each patient, an ED physician recorded digital images and wrote one diagnostic hypothesis on a preformatted form before calling a dermatologist. The ED physician then chose one of the following options: immediate hospitalization, immediate specialist consultation, specialized consultation within 24 hours, or patient discharge with treatment. Images were then sent to the dermatologist via mobile telephone. To provide detailed clinical information, the ED physician called the dermatologist with the same mobile telephone. Once their telephone conversation was complete, both wrote their diagnosis and management choices on a preformatted form.
The ED physician then contacted the dermatologist to perform a videoconferencing consultation via their smartphones. Once the videoconference was complete, both wrote their final diagnosis and management proposals on a preformatted form.
Three dermatologists (F.C., C.J., and E.B.) separately evaluated the forms filled out by the ED physicians and dermatologists. They did so without having access to the patients’ digital images and clinical history. They classified the diagnoses between the ED physicians and dermatologists and between digital image teledermatology and videoconferencing. They also determined management concordance between the ED physician and dermatologist.
The respective diagnoses of the ED physicians and dermatologists were then compared using the following classification: (1) different diagnosis, if the dermatologist’s diagnosis was completely different from the ED physician’s diagnosis (eg, shingles instead of contact allergy); (2) identical diagnosis, if the dermatologist’s diagnosis was the same as the ED physician’s diagnosis; (3) enlarged diagnosis, if the dermatologist proposed several diagnoses and the ED physician only one (eg, leg cellulitis or deep venous thrombosis or contact allergy instead of deep venous thrombosis); and (4) clarified diagnosis, if the dermatologist proposed only one diagnosis and the ED physician several (eg, insect bites instead of viral exanthema or drug rash or insect bites).
The primary outcome was diagnosis agreement. The secondary outcome was management concordance.
Quantitative variables are expressed as median (range) and qualitative variables as number (percentage). Diagnosis agreement was compared between steps 1 and 2 using the χ2 test. Management concordance was compared using mean expert scores between steps 1 and 2 using the t test. Both diagnosis agreement and management concordance were assessed by κ statistics.
All tests were 2-tailed. P ≤ .05 indicated statistical significance. Data were analyzed using SAS statistical software, version 9.1 (SAS Institute Inc).
From May 1, 2008, through June 30, 2010, a total of 111 patients were included in the study. Videoconferencing was performed for 83 patients (74.8%) and was fully usable for 68 (61.3%). Table 1 summarizes the final dermatologic diagnoses for the study patients. Conditions diagnosed in the ED were mainly allergic eczema or urticaria (24.3%) or infectious skin disorders (28.0%): erysipelas or cellulitis (4.5%), impetigo or pyodermatitis (6.3%), herpes zoster virus infection (9.0%), viral eruption (2.7%), or other (5.4%). Inflammatory dermatoses represented 12.6% of diagnoses, whereas drug or iatrogenic eruption concerned 9.9% of the cases.
Among the 111 patients, ED physician diagnoses were confirmed by dermatologists in 34 of 110 cases (30.9%), invalidated in 39 of 110 cases (35.5%), enlarged in 16 of 110 cases (14.5%), and clarified in 20 of 110 cases (18.2%). Videoconferencing confirmed the dermatologist’s photographic diagnosis in 44 of 68 cases (64.7%), invalidated it in 10 of 68 cases (14.7%), enlarged it in 3 of 68 cases (4.4%), and clarified it in 11 of 68 cases (16.2%). Videoconferencing improved the diagnostic performance in 57 of 83 cases (68.7%) (P < 10−4). Photographic diagnoses were invalidated in 8 of 83 cases (9.6%) (Table 2).
Management concordance was compared between ED physicians and dermatologists (Table 3). Using only patients’ images and clinical history, dermatologists and ED physicians agreed that there was no need for an immediate specialist consultation (99.1% vs 100%, P = .04); management concordance was high (κ, 0.75; 95% CI, 0.64-0.86). Specialist consultation within 24 hours was significantly more frequently required according to the dermatologists (30.8% of the cases vs 28.2% for the ED physician, P < 10−4); management concordance agreement was moderate (κ, 0.49; 95% CI, 0.14-0.84). The ED physician recommended immediate hospitalization significantly more frequently (8.2% vs 7.2%, P < 10−4); management concordance agreement was moderate (κ, 0.49; 95% CI, 0.41-0.57). Dermatologists significantly more often chose to discharge patients (46.8% vs 39.1%, P < 10−4); management concordance agreement was high (κ, 0.74; 95% CI, 0.67-0.81) (Table 3).
A comparison of dermatologist diagnosis and management between digital images (step 1) and videoconferencing (step 2) revealed a high concordance for diagnostic and specialist consultation within 24 hours (κ, 1.00; 95% CI, 0.98-1.02; and κ, 0.60; 95% CI, 0.50-0.70, respectively). Concordance for immediate hospitalization was moderate (κ, 0.47; 95% CI, 0.22-0.72) and poor for patient discharge (κ, 0.21; 95% CI, 0.09-0.33).
Infectious or allergic dermatologic conditions were the main purpose of the ED visit, in accordance with previously published studies.4,8 Real-time dermatologic consultations with still images and videoconferencing could be performed in most cases. Unreachable network, poor image quality, and patients’ impatience at the clinic were the main causes of videoconferencing failure. If our aim was to evaluate mobile device accuracy and input in patients’ management, some study limitations, such as image quality, may be associated with technologic support.
Easy access to a dermatologist helped ED physicians manage skin problems. In most cases, their diagnosis was invalidated, enlarged, or clarified (as defined in the Methods section) by the dermatologist. A comparison of patient management also revealed that ED physicians would have more often hospitalized patients than dermatologists. In most cases, videoconferencing either confirmed the dermatologist diagnosis on still images or allowed for diagnosis clarification or enlargement. Management concordance between still images and videoconferencing was moderate for patient hospitalization and poor for patient discharge. According to the investigators, videoconferencing significantly contributed to diagnosis because of direct communication between physicians and improvement of the quality of collected information and/or the accuracy of the patient’s status evaluation. Our study confirmed the contribution of videoconferencing vs still images in dermatologic diagnosis. In their literature review, Warshaw et al7 reported a diagnostic concordance with in-person visits of up to 80% for live interactive technologies (ie, videoconferencing) and 66% for store-and-forward technologies (ie, still images). Collection of additional information or patient history may partly explain these differences.8 The impact of live interactive technologies, analyzed in 1500 sessions, resulted in a significant change in diagnosis in 69.9% of the referred patients and in patient management in 97.7% of the cases.9 User-friendly videoconferencing systems require more investments than store-and-forward applications, which require web-based environments with image transfer. Store-and-forward applications provide high-quality images for dermatologist diagnostic confidence independently of a videoconferencing network or Internet connection.10 The main limitation of store-and-forward applications is the additional time investment for the general practitioner, up to 19 minutes per patient and consultation. For example, Berghout et al11 reported an increase in the mean general practitioner consultation time of 3.5 minutes.
The technological advances made in the last decade have encouraged the implementation of mobile telephone devices in teledermatology because of their practicality, high-resolution images, and instant image upload.12- 15 Mobile phone reliability has been successfully evaluated in a pilot study13 for dermatologist resident supervision: both senior and resident dermatologists used handheld devices for image acquisition and interpretation (n = 10). Image acquisition and transmission took a mean of 5 to 15 minutes for 3 pictures, diagnosis concordance was 80%, and limitations were small screen size and image quality in one case. The quality of images taken with a mobile telephone-integrated camera has also been studied in wound management or skin cancer screening.12,15 In a preliminary study evaluating a mobile telephone with a 640- to 480-pixel resolution camera, Braun et al12 compared wound diagnostic accuracy and management of 2 dermatologists to in-person evaluation (n = 61). Physicians were comfortable making a diagnosis in 82% of the cases (n = 50) and judged image quality as good or very good in 59% (n = 36) and 20% (n = 12), respectively. For cancer screening, comparison of mobile telephone diagnosis with in-person diagnosis also revealed a high diagnosis concordance of 82% (95% CI, 0.73-0.89), with a κ coefficient indicating a good agreement (κ, 0.62), and a management concordance of 81% with a moderate agreement (κ, 0.57).15
Implementation of still image diagnosis application in the ED was previously described in a 1-year study of 60 patients.8 Patients’ history and clinical images were sent to a secure e-mail address and then automatically forwarded to the dermatologist. The teledermatologist gave advice on diagnosis and management, and patients were seen within 2 weeks in a follow-up in-person visit. The most frequent diagnoses were dermatitis or eczema, urticaria, and drug eruption. A rapid answer to ED physicians was given within 2 hours. Eighty-three percent of the patients (n = 50) had face-to-face consultations. Diagnosis agreement between telediagnosis and ED diagnosis was high (71.2%; κ, 0.42), This diagnosis agreement may have been artificially high because some patients were referred with their diagnosis. Diagnosis agreement between telediagnosis and in-person final clinical diagnosis was 98% (κ, 0.93), with 96% complete agreement for clinical management.
General practitioner access to teledermatologic advice reduces the number of second referrals to a dermatologist by 20.7% and improves the primary care management quality of dermatologic conditions.16 An increase in general practitioner efficiency and knowledge and a decrease in inappropriate prescription and management and/or physical referral to a dermatologist also contribute to cost reduction of management of dermatologic conditions. Cost comparison between general practitioner access to teledermatology and conventional care revealed a reduction of 18%.17 In a model-based approach, cost savings were also achieved if the distance to the dermatologist was 75 km or more or if 37% or more of referrals to a dermatologist could be avoided.18
Our study confirms the relevance of implementing teledermatology services in EDs. Dermatologic remote advice invalidated, enlarged, or clarified ED physician diagnosis and management in 75 of 110 cases (68.2%). Cases also displayed a high diagnostic concordance between store-and-forward and videoconferencing, with a better diagnostic contribution of the latter, when mobile devices were used to capture or review patients. Our study demonstrates the convenience of using mobile devices for teledermatology in the ED and suggests their use to establish direct communication between physicians after image transfer. The limitations of this study were the absence of data collection or evaluation of the 2 procedures for their mean duration and dermatologists’ confidence in their diagnoses.
Compared with standard hardware, new-generation mobile devices reduce the cost of videoconferencing, increase the versatility of teledermatology, and decrease general practitioner time involvement with teledermatology. However, to guarantee the quality of teledermatologic consultations, the use of mobile devices requires a functional high-speed telecommunication network, a secure protocol for image transfer, a reliable storage system, and some minimal organization.
Accepted for Publication: July 18, 2013.
Corresponding Author: Fabien Guibal, MD, MSc, MBA, Department of Dermatology, Hôpital Saint Louis-Lariboisière, 1 Avenue Claude Vellefaux, 75010 Paris, France (email@example.com).
Published Online: May 7, 2014. doi:10.1001/jamadermatol.2013.7792.
Author Contributions: Drs Duong and Cordoliani had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Guibal.
Acquisition of data: Cordoliani, Julliard, Bourrat, Regnier, de Pontual, Leroy, Gallula, Aparicio, Legendre, Guibal.
Analysis and interpretation of data: Duong, Bagot, Guibal.
Drafting of the manuscript: Duong, Gallula, Guibal.
Critical revision of the manuscript for important intellectual content: Duong, Cordoliani, Julliard, Bourrat, Regnier, de Pontual, Leroy, Aparicio, Legendre, Bagot, Guibal.
Statistical analysis: Duong, Guibal.
Obtained funding: Guibal.
Administrative, technical, or material support: Gallula, Guibal.
Study supervision: Cordoliani, Bagot, Guibal.
Conflict of Interest Disclosures: Dr Duong received a grant from Basilea Pharm for the first year of her PhD program.
Funding/Support: This study received a research grant from the Société Française de Dermatologie (Dr Guibal).
Role of the Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Mobile telephones and communication costs were provided by the Société Française de Radiotéléphonie. Emmanuelle Boutmy-Deslandes, PhD, Department of Biostatistics and Public Health, Hôpital Saint Louis-Lariboisière, AP-HP, Paris, France, contributed to the statistical analysis, and was compensated for her contibution.