Assessment of the accuracy of low-cost store-and-forward teledermatology consultation

Background: Telemedicine has the potential to revolutionize the delivery of dermatologic care to underserved areas. Objective: Our purpose was to compare diagnoses from two types of dermatology consultations: telemedicine using store-and-forward (SAF) technology, and traditional face-to-face (FTF) office visits. Methods: Skin conditions were imaged with a consumer-gt-ade digital camera. A standardized template was used to collect historical data. Information was stored in a secured database for access by 2 or 3 board-certified dermatologists. Results from the FTF visit were used to assess the accuracy of the SAF diagnoses. Results: A total of 106 dermatologic conditions in 92 patients were included. Concordance between FTF and SAF diagnoses was high, ranging from 81% to 89% for all 3 dermatologists. Clinically relevant disagreement occurred in only 4% to 8% of cases. Remaining disagreements did not affect patient care. Diagnostic confidence and image quality affected agreement. When cases of high confidence were analyzed separately, agreement increased to 88% to 100%. This increase was substantiated by means of a chi-square test between the high confidence and low confidence groups, which demonstrated statistical significance (P < .005) for all dermatologists. Similarly, when cases of above average image quality were considered, agreement increased to 84% to 98%. Again this difference was substantiated by means of a chi-square test between adequate and poor images, with statistical significance for two dermatologists (P < .001). Accuracy was comparable between disease types with the exception of benign neoplasms, which demonstrated agreement of 22% to 46%. Conclusion: These data support the use of existing digital technology to construct an accurate SAF teledermatology system. The inexpensive camera and widely available computer equipment make this an extremely affordable system. Furthermore, participating dermatologists appear well aware of system limitations, as reflected in the increased agreement for high confidence cases. Additional investigation of the accuracy of teledermatology for benign neoplasms is warranted.