Adequacy of Existing Surveillance Systems to Monitor Racism, Social Stigma and COVID Inequities: A Detailed Assessment and Recommendations.

The populations impacted most by COVID are also impacted by racism and related social stigma; however, traditional surveillance tools may not capture the intersectionality of these relationships. We conducted a detailed assessment of diverse surveillance systems and databases to identify characteristics, constraints and best practices that might inform the development of a novel COVID surveillance system that achieves these aims. We used subject area expertise, an expert panel and CDC guidance to generate an initial list of N > 50 existing surveillance systems as of 29 October 2020, and systematically excluded those not advancing the project aims. This yielded a final reduced group (n = 10) of COVID surveillance systems (n = 3), other public health systems (4) and systems tracking racism and/or social stigma (n = 3, which we evaluated by using CDC evaluation criteria and Critical Race Theory. Overall, the most important contribution of COVID-19 surveillance systems is their real-time (e.g., daily) or near-real-time (e.g., weekly) reporting; however, they are severely constrained by the lack of complete data on race/ethnicity, making it difficult to monitor racial/ethnic inequities. Other public health systems have validated measures of psychosocial and behavioral factors and some racism or stigma-related factors but lack the timeliness needed in a pandemic. Systems that monitor racism report historical data on, for instance, hate crimes, but do not capture current patterns, and it is unclear how representativeness the findings are. Though existing surveillance systems offer important strengths for monitoring health conditions or racism and related stigma, new surveillance strategies are needed to monitor their intersecting relationships more rigorously.

TAPER: Time-Aware Patient EHR Representation.

Effective representation learning of electronic health records is a challenging task and is becoming more important as the availability of such data is becoming pervasive. The data contained in these records are irregular and contain multiple modalities such as notes, and medical codes. They are preempted by medical conditions the patient may have, and are typically recorded by medical staff. Accompanying codes are notes containing valuable information about patients beyond the structured information contained in electronic health records. We use transformer networks and the recently proposed BERT language model to embed these data streams into a unified vector representation. The presented approach effectively encodes a patient's visit data into a single a distributed representation, which can be used for downstream tasks. Our model demonstrates superior performance and generalization on mortality, readmission and length of stay tasks using the publicly available MIMIC-III ICU dataset.

Children Activity Recognition: Challenges and Strategies.

In this paper, we study the problem of children activity recognition using smartwatch devices. We introduce the need for a robust children activity model and challenges involved. To address the problem, we employ two deep neural network models, specifically, Bi-Directional LSTM model and a fully connected deep network and compare the results to commonly used models in the area. We demonstrate that our proposed deep models can significantly improve results compared to baseline models. We further show benefits of activity intensity level detection in health monitoring and verify high performance of our proposed models in this task.