Search

Objective In response to COVID-19, the informatics community united to aggregate as much clinical data as possible to characterize this new disease and reduce its impact through collaborative analytics. The National COVID Cohort Collaborative (N3C) is now the largest publicly available HIPAA limited dataset in US history with over 6.4 million patients and is a testament to a partnership of over 100 organizations. Materials and Methods We developed a pipeline for ingesting, harmonizing, and centralizing data from 56 contributing data partners using 4 federated Common Data Models. N3C data quality (DQ) review involves both automated and manual procedures. In the process, several DQ heuristics were discovered in our centralized context, both within the pipeline and during downstream project-based analysis. Feedback to the sites led to many local and centralized DQ improvements. Results Beyond well-recognized DQ findings, we discovered 15 heuristics relating to source Common Data Model conformance, demographics, COVID tests, conditions, encounters, measurements, observations, coding completeness, and fitness for use. Of 56 sites, 37 sites (66%) demonstrated issues through these heuristics. These 37 sites demonstrated improvement after receiving feedback. Discussion We encountered site-to-site differences in DQ which would have been challenging to discover using federated checks alone. We have demonstrated that centralized DQ benchmarking reveals unique opportunities for DQ improvement that will support improved research analytics locally and in aggregate. Conclusion By combining rapid, continual assessment of DQ with a large volume of multisite data, it is possible to support more nuanced scientific questions with the scale and rigor that they require.

Recently, the number of observed malware samples has rapidly increased, expanding the workload for malware analysts. Most of these samples are not truly unique, but are related through shared attributes. Identifying these attributes can enable analysts to reuse analysis and reduce their workload. Visualizing malware attributes as sets could enable analysts to better understand the similarities and differences between malware. However, existing set visualizations have difficulty displaying hundreds of sets with thousands of elements, and are not designed to compare different types of elements between sets, such as the imported DLLs and callback domains across malware samples. Such analysis might help analysts, for example, to understand if a group of malware samples are behaviorally different or merely changing where they send data.To support comparisons between malware samples' attributes we developed the Similarity Evidence Explorer for Malware (SEEM), a scalable visualization tool for simultaneously comparing a large corpus of malware across multiple sets of attributes (such as the sets of printable strings and function calls). SEEM's novel design breaks down malware attributes into sets of meaningful categories to compare across malware samples, and further incorporates set comparison overviews and dynamic filtering to allow SEEM to scale to hundreds of malware samples while still allowing analysts to compare thousands of attributes between samples. We demonstrate how to use SEEM by analyzing a malware sample from the Mandiant APT1 New York Times intrusion dataset. Furthermore, we describe a user study with five cyber security researchers who used SEEM to rapidly and successfully gain insight into malware after only 15 minutes of training.

The Food and Drug Administration and Department of Defense were interested in detecting sub-optimal use of long-acting beta-agonists (LABAs) in asthmatics within the Military Health System (MHS). Visualizing the patterns of asthma medication use surrounding a LABA prescription is a quick way to detect possible sub-optimal use for further evaluation. Analysis of 100 asthma patients was conducted in EventFlow to display and summarize timepoint and interval data. Epidemiologists reported that EventFlow was a powerful tool for rapidly visualizing possible patterns of sub-optimal LABA use that can be targeted for intervention.

Objective: To explore approaches for integrating and visualizing time-oriented medication data in narrative and structured formats and to address related issues on handling temporal abstraction, granularity, and uncertainty. The ultimate goal is to improve medication reconciliation by providing clinicians with more accurate medication information in patient care. Methods: An event taxonomy was generated to capture different combinations of clinical and temporal uncertainties. A prototype of a temporal visualization system was implemented using an open source software package called Timeline. Medications were parsed and mapped to the event taxonomy, and then represented in Timelines. Seventy-five medications from narrative discharge summary reports and seventy-nine medications from structured orders were used as data input for temporal visualization. Five physicians served as domain experts and answered ten proof-of-concept survey questions. Results: Overall positive feedback from experts suggested the potential value of the proposed timeline visualization method. Challenges were also identified, and future work will include reconciliation of medications from various sources based on temporal attributes and medication classification.