Patient Factors Associated With Teledermatology Visit Type and Submission of Photographs During the COVID-19 Pandemic: Cross-sectional Analysis

Background: The COVID-19 pandemic necessitated the widespread adoption of teledermatology, and this continues to account for a significant proportion of dermatology visits after clinics have reopened for in-person care. Delivery of high-quality teledermatology care requires adequate visualization of the patient's skin, with photographs being preferred over live video for remote skin examination. It remains unknown which patients face the greatest barriers to participating in a teledermatology visit with photographs. Objective: The aim of this study was to identify patient characteristics associated with type of telemedicine visit and the factors associated with participating in teledermatology visits with digital photographs versus those without photographs. Methods: We performed a cross-sectional analysis of the University of Pennsylvania Health System electronic health record data for adult patients who participated in at least 1 teledermatology appointment between March 1, 2020, and June 30, 2020. The primary outcomes were participation in a live-interactive video visit versus a telephone visit and participation in any teledermatology visit with photographs versus one without photographs. Multivariable logistic regression was performed to evaluate the associations between patient characteristics and the primary outcomes. Results: In total, 5717 unique patients completed at least 1 teledermatology visit during the study period;68.25% (n=3902) of patients participated in a video visit, and 31.75% (n=1815) participated in a telephone visit. A minority of patients (n=1815, 31.75%) submitted photographs for their video or telephone appointment. Patients who submitted photographs for their teledermatology visit were more likely to be White, have commercial insurance, and live in areas with higher income, better education, and greater access to a computer and high-speed internet (P<.001 for all). In adjusted analysis, older age (age group >75 years: odds ratio [OR] 0.60, 95% CI 0.44-0.82), male sex (OR 0.85, 95% CI 0.75-0.97), Black race (OR 0.79, 95% CI 0.65-0.96), and Medicaid insurance (OR 0.81, 95% CI 0.66-0.99) were each associated with lower odds of a patient submitting photographs for their video or telephone visit. Older age (age group >75 years: OR 0.37, 95% CI 0.27-0.50) and Black race (OR 0.82, 95% CI 0.68-0.98) were also associated with lower odds of a patient participating in a video visit versus telephone visit. Conclusions: Patients who were older, male, or Black, or who had Medicaid insurance were less likely to participate in teledermatology visits with photographs and may be particularly vulnerable to disparities in teledermatology care. Further research is necessary to identify the barriers to patients providing photographs for remote dermatology visits and to develop targeted interventions to facilitate equitable participation in teledermatology care. ©Jordan E Lamb, Robert Fitzsimmons, Anjana Sevagamoorthy, Carrie L Kovarik, Daniel B Shin, Junko Takeshita. Originally.

Provably Secure Contact Tracing with Conditional Private Set Intersection

The novel coronavirus COVID-19 spreads easily through personal contact, requiring the use of contact tracing to track the spread of the disease. Many existing approaches either trust a public health authority with private data, or publish patients’ data, leading to privacy breaches. Private Set Intersection based on Homomorphic Encryption is a promising solution, but it is limited because the management of keys is challenging and further filtering of contacts is not included. We present a protocol for secure and private conditional contact tracing, allowing the tracking of users’ contacts subject to extra conditions. We construct and apply our new primitive of Conditional Private Set Intersection and combine it with a Trusted Execution Environment (TEE) to construct a protocol with provable security and a high degree of functionality. Our approach moves the memory- and computation-intensive portions of contact tracing out of the TEE to a cloud server. We also present how multi-hop contact tracing can be done with minimal user communication. Our proof-of-concept implementation with Microsoft SEAL allows users to perform their computation in less than 9 min, and the cloud’s per-user computation can be as little as 11 min for a population of 50,000 users with 500 infected (assuming 40 contacts/user) in a day. With other HE libraries/schemes that allows customized parameter sets, our protocol will show much higher scalability. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.