The Impact of COVID-19 on Periocular NonMelanoma Skin Cancer in the Veteran Population

Purpose : Despite an increasing incidence of skin cancer over the last decade, studies have reported a decline in the diagnosis and treatment of skin cancer during the COVID19 pandemic. We performed a retrospective cohort study using a large population-based cohort from the Veterans Health Administration (VHA) to determine how the pandemic has affected tumor size and morbidity in veterans with periocular non-melanoma skin cancer. Methods : Electronic health records from all VHA sites were accessed through the VA Informatics and Computing Infrastructure (VINCI). Data were stored in the Observational Medical Outcomes Partnership (OMOP) model and queried via SQL Server. ICD-10 and current procedural terminology codes were used to identify patients who received Mohs surgery for periocular basal cell carcinoma (BCC) or squamous cell carcinoma (SCC) between 08/01/2018 and 09/10/2021. A combination of structured algorithms and manual review were used to extract patient demographics, lesion characteristics, and surgical outcome at three time points, ie. pre-COVID, early, and late COVID. Unpaired t-tests were used to assess statistical significance. Results : Patient characteristics were similar between pre- and post-COVID cohorts in terms of gender, age, race, and tumor type. The average number of Mohs periocular surgeries performed per week were 23.1% (7.31 vs 5.62) and 13.1% (7.49 vs 6.51) lower in the early and later pandemic, respectively, compared to similar pre-COVID timeframes by month (Figure 1). Mean lesion size (maximum diameter) was 1.35 cm larger post-COVID compared to pre-COVID (95% CI 0.19 2.51, P=0.022);however, the defect size remained similar (Figure 2). Stratifying by tumor type, the same trends were noted in BCC, particularly early in the pandemic. However, mean SCC lesion and defect sizes did not vary over time. Conclusions : Periocular Mohs surgery rates declined in the COVID pandemic across VHA. Lesions were larger particularly in the earlier phase of the pandemic for BCC. Future analyses using this cohort will attempt to determine if telehealth and travel time were associated with distinct outcomes.

COVID-19 Social Distance Proximity Estimation Using Machine Learning Analyses of Smartphone Sensor Data

Airborne transmittable diseases such as COVID-19 spread from an infected to healthy person when they are in proximity to each other. Epidemiologists suggest that the risk of COVID-19 transmission increases when an infected person is within 6 feet from a healthy person and contact between them lasts longer than 15 minutes (also called Too Close For Too Long (TC4TL). In this paper, we systematically investigate Machine Learning (ML) methods to detect proximity by analyzing publicly available dataset gathered from smartphones' built-in Bluetooth, accelerometer, and gyroscope sensors. We extract 20 statistical features from accelerometer and gyroscope sensors signals and 28 statistical features of Bluetooth signal, which are classified to determine whether subjects are closer than 6 feet as well as the subjects' context. Using machine learning regression, we also estimate the range between the subjects. Among the 19 ML classification and regression methods that we explored, we found that ensemble (boosted and bagged trees) methods perform best with accelerometer and gyroscope data while regression trees ML algorithm performs best with the Bluetooth signal. We further explore sensor fusion methods and demonstrate that the combination of all three sensors achieves a higher accuracy of range estimation than when using each individual sensor. We show that proximity (< 6ft or not) can be classified with 72%-90% accuracy using the accelerometer, 78%-84% accuracy using gyroscope sensor, and with 76%-92% accuracy with the Bluetooth data. Our model outperforms the current state-of-the-art methods using neural networks and achieved a Normalized Decision Cost Function (nDCF) score of 0.34 with Bluetooth radio and 0.36 with sensor fusion.

Passive Smartphone Contact Tracing and Continuous COVID-19 Infection Risk Assessment

COVID-19 has infected millions since November 2019. The virus spreads through close contact with those who are infected. People are often unaware of or lose track of their behaviors, which increase their risk of infection. Passive methods to continuously monitor and track dangerous user behaviors, maintain and update a COVID risk score can enable high risk users to take preventive measures early. At the organization or institution level, such systems can provide insights on organization-wide patterns that exacerbate disease spread. This paper presents our vision of pervasive, continuous infectious disease contact tracing, risky behavior tracking and continuous risk score calculation from contact, place visit information and reported behaviors. As a specific example, we describe the research, design and development of the android app Goatvid Trace that continuously gathers smartphone sensor data, using it to calculcate smartphone users' risk of exposure to COVID-19. Machine Learning methods for proximity detection from smart- phone Bluetooth RSSI signals are also described. GoatVid trace was deployed and evaluated on a small university community. Our evaluation study found that the mean COVID-19 risk score of college students in our study was 25.6%. Our risk score model correlated well with subject questionnaire responses with an R of 0.6166. The machine learning models for proximity detection estimated distances between two phones with a Cross Validation RMSE of 1.58766. © 2021 IEEE.

Deep Learning Anomaly Detection methods to passively detect COVID-19 from Audio

The world has been severely affected by COVID-19, an infectious disease caused by the SARS-Cov-2 coronavirus. COVID-19 incubates in a patient for 7 days before symptoms manifest. The identification of the presence of COVID-19 is challenging as its symptoms are similar to influenza symptoms such as cough, cold, runny nose, and chills. COVID-19 affects human speech sub-systems involved in respiration, phonation, and articulation. We propose a deep anomaly detection framework for passive, speech-based detection of COVID-related anomalies in voice samples of COVID-19 affected individuals. The low percentage of positive cases and extreme imbalance in available COVID audio datasets present a challenge to machine learning classifiers but create an opportunity to utilize anomaly detection techniques. We investigate COVID detection from audio using various types of deep anomaly detectors and convolutional autoencoders. Contrastive loss methods are also explored to force our models to learn discrepancies between COVID and non-COVID cough data representations. In contrast with prior work that controlled data collection, our work focuses on crowdsourced datasets that are true representatives of the general population. In rigorous evaluation, the variational autoencoder with the elliptic envelope as the anomaly detector analyzing Mel Filterbanks audio representations performed best with an AUC of 0.65, outperforming the state of the art. © 2021 IEEE.

Passive COVID-19 Assessment using Machine Learning on Physiological and Activity Data from Low End Wearables

COVID-19 has now infected over 165 million people and killed over 3.5 million people. While public health interventions have reduced its spread and vaccines are being deployed, passive detection methods are needed to detect infections and early track its resurgence. Wearables that are widely owned can gather various physiological and activity data, presenting an opportunity to detect COVID-19 unobtrusively. COVID-19 infection causes deviations in the vital physiological signs and activity patterns of infected users. However, similar deviations of these same variables can also be affected by non-COVID factors, confounding the signals. In this paper, we investigate the feasibility of predicting COVID-19 infection to detect abnormalities in heart rate, activity (steps), and sleep data available on low-end wearables by using machine learning. Prior work utilized data such as oxygen saturation that is only available on clinical-grade equipment or expensive wearables. We extracted 43 statistical features (standard deviation, mean, slope) and behavioral (min/max/avg length of sedentary and active bouts, sleep duration, no. of awake/asleep/restless samples) from wearable sensor data. We classified these features using machine learning classification and anomaly detection algorithms. Physical activity features were the most predictive (min length of the sedentary and active bout), yielding an AUC-ROC of 78% [specificity=74%, sensitivity=69%] when classified using Gradient Boosting Machines (GBMs). We also found that sleep irregularities had low discriminative performance. COVID-19 detection using inexpensive wearables can facilitate population-level interventions. © 2021 IEEE.

DeepSEAS: Smartphone-based Early Ailment Sensing Using Coupled LSTM AutoEncoders

Infectious diseases epidemics such as the current COVID-19 pandemic have an immense impact on all facets of life. Consequently, the current dearth of effective and timely public health surveillance methods, especially at the individual level, have been accentuated, prompting research into supplementary methods. Sensor-rich, ubiquitously owned smartphones can now gather large volumes of data that has been utilized for passive and continuous physical and mental health assessment. In this paper, we propose a Deep learning based Smartphone Early Ailment Sensing (DeepSEAS) framework that predicts a smart-phone user's future manifestation of influenza-like biological symptoms (e.g. coughing and sneezing) a day early while they are still asymptomatic. DeepSEAS works by analyzing a subject's historical one-day smartphone sensor and mobility data. First, we utilize the mean shift clustering algorithm to create clusters of users with similar social and behavioral traits such as their socialization levels, social media presence, eating and working out habits. Then, DeepSEAS employs an end-to-end trainable LSTM Autoencoder (LSTM AE) coupled with a Feed Forward Neural network classifier, a chieving a sensitivity of 7 8% i n correctly identifying users who will manifest biological symptoms a day later. DeepSEAS facilitates up-to-date influenza s urveillance at the individual level, which could transform the current healthcare system. Early detection can enable asymptomatic users to be alerted, notified and isolated, which could reduce disease transmission. © 2020 IEEE.