Evaluating a telemedicine system to assist in the management of dermatology referrals

Abstract

Background: Teledermatology systems fall into two categories: live video or store-and-forward. In the former, video-conferencing equipment is used to connect a patient with a remote consultant. This method has been evaluated as an aid to dermatology, but it is expensive both in terms of capital and running costs. Video consultations are generally longer than conventional ones and harder to schedule. Some authors have considered store-and-forward as an alternative to live video: instead of a consultation, specialists could make a rapid inspection of a transmitted still image.

Objectives: A study was conducted to evaluate the role of telemedicine in the dermatology outpatients department of a district general hospital (Whittington Hospital NHS Trust, London, U.K.).

Methods: One hundred and ninety-four patients were seen by one of two consultant dermatologists. A nurse used a video camera to store digital images of each patient's problem and compiled a history from the GP's referral letter. The images were reviewed 13 months later by both dermatologists; they recorded a provisional diagnosis and an assessment of how urgent an appointment would have been made given the information provided by the system. A third consultant graded the level of agreement between the telemedicine diagnoses and the face-to-face consultations.

Results: High levels of agreement were found between the diagnoses of the dermatologists using the system to inspect images and those of the dermatologist who saw the patients (77%). Consultants using the system recommended fewer urgent appointments (32% compared with 64%) and felt that in 31% of cases the patient did not need to be seen. In 15% of these cases (5% of the total), however, their diagnosis differed significantly from that of the consultant who saw the patient. Had the system been in use, 14% of patients conventionally assigned a non-urgent appointment would have been seen urgently.

Conclusions: The images allowed a reasonably accurate diagnosis. The software was not reliable (six cases could not be viewed), or easy to use (it took approximately an hour to view 20 cases) but an improved version could be used in triaging outpatient appointments.