Showing papers by "Mitre Corporation published in 2021"

The Effects of Virtual Reality, Augmented Reality, and Mixed Reality as Training Enhancement Methods: A Meta-Analysis.

Abstract: ObjectiveThe objective of this meta-analysis is to explore the presently available, empirical findings on transfer of training from virtual (VR), augmented (AR), and mixed reality (MR) and determin...

COVID-19 Vaccine Discourse on Twitter: A Content Analysis of Persuasion Techniques, Sentiment and Mis/Disinformation.

Abstract: This research aims to understand the persuasion techniques used in Twitter posts about COVID-19 vaccines by the different vaccine sentiments (i.e., Pro-Vaccine, Anti-Vaccine, and Neutral) using the Elaboration Likelihood Model, Social judgment Theory, and the Extended Parallel Process Model as theoretical frameworks. A content analysis was conducted on a data set of 1,000 Twitter posts. The corpus of Tweets was examined using the persuasion frameworks; tweets that were identified as emanating from bots were further examined. Results found Anti-Vaccine messages predominantly used Anecdotal stories, Humor/Sarcasm, and Celebrity figures as persuasion techniques, while Pro-Vaccine messages primarily used Information, Celebrity figures, and Participation. Results also showed the Anti-Vaccine messages primarily focused on values related to the categories of Safety, Political/Conspiracy Theories, and Choice. Finally, results revealed Anti-Vaccine messages primarily used Perceived Severity and Perceived Susceptibility, which are fear appeal elements. The findings for messages by bots were comparable to the messages in the larger corpus of tweets. Based on the findings, a response framework-Health Information Persuasion Exploration (HIPE)-is proposed to address mis/disinformation and Anti-Vaccine messaging. The results of this study and the HIPE framework can inform a national COVID-19 vaccine health campaign to increase vaccine adoption.

Scalable and explainable legal prediction

Abstract: Legal decision-support systems have the potential to improve access to justice, administrative efficiency, and judicial consistency, but broad adoption of such systems is contingent on development of technologies with low knowledge-engineering, validation, and maintenance costs This paper describes two approaches to an important form of legal decision support—explainable outcome prediction—that obviate both annotation of an entire decision corpus and manual processing of new cases The first approach, which uses an attention network for prediction and attention weights to highlight salient case text, was shown to be capable of predicting decisions, but attention-weight-based text highlighting did not demonstrably improve human decision speed or accuracy in an evaluation with 61 human subjects The second approach, termed semi-supervised case annotation for legal explanations, exploits structural and semantic regularities in case corpora to identify textual patterns that have both predictable relationships to case decisions and explanatory value

Minding Your Own Business? Mindfulness Decreases Prosocial Behavior for People With Independent Self-Construals.

Abstract: Mindfulness appears to promote individual well-being, but its interpersonal effects are less clear. Two studies in adult populations tested whether the effects of mindfulness on prosocial behavior differ according to individuals' self-construals. In Study 1 (N = 366), a brief mindfulness induction, compared with a meditation control condition, led to decreased prosocial behavior among people with relatively independent self-construals but had the opposite effect among those with relatively interdependent self-construals. In Study 2 (N = 325), a mindfulness induction led to decreased prosocial behavior among people primed with independence but had the opposite effect among those primed with interdependence. The effects of mindfulness on prosocial behavior appear to depend on individuals' broader social goals. This may have implications for the increasing popularity of mindfulness training around the world.

Reduced Incidence of Long-COVID Symptoms Related to Administration of COVID-19 Vaccines Both Before COVID-19 Diagnosis and Up to 12 Weeks After

Abstract: Both clinical trials and studies leveraging real-world data have repeatedly confirmed the three COVID-19 vaccines authorized for use by the Food and Drug Administration are safe and effective at preventing infection, hospitalization, and death due to COVID-19 and a recent observational study of self-reported symptoms provides support that vaccination may also reduce the probability of developing long-COVID. As part of a federated research study with the COVID-19 Patient Recovery Alliance, Arcadia.io performed a retrospective analysis of the medical history of 240,648 COVID-19-infected persons to identity factors influencing the development and progression of long-COVID. This analysis revealed that patients who received at least one dose of any of the three COVID vaccines prior to their diagnosis with COVID-19 were 7-10 times less likely to report two or more long-COVID symptoms compared to unvaccinated patients. Furthermore, unvaccinated patients who received their first COVID-19 vaccination within four weeks of SARS-CoV-2 infection were 4-6 times less likely to report multiple long-COVID symptoms, and those who received their first dose 4-8 weeks after diagnosis were 3 times less likely to report multiple long-COVID symptoms compared to those who remained unvaccinated. This relationship supports the hypothesis that COVID-19 vaccination is protective against long-COVID and that effect persists even if vaccination occurs up to 12 weeks after COVID-19 diagnosis. A critical objective of this study was hypothesis generation, and the authors intend to perform further studies to substantiate the findings and encourage other researchers to as well.

Reducing COVID-19 Airborne Transmission Risks on Public Transportation Buses: An Empirical Study on Aerosol Dispersion and Control

Abstract: This study is one of the first COVID-19 related bus studies to fully characterize cough aerosol dispersion and control in the highly turbulent real-world environment of driving regular bus routes o...

Rydberg Atom Electric Field Sensors for Communications and Sensing

Abstract: Rydberg atom electric field sensors are projected to enable novel capabilities for resilient communications and sensing. This quantum sensor is small-size, highly sensitive, and broadly tunable, and it has the potential for performing precision vector electric field and angle-of-arrival measurements. While these atomic electric field sensors will not replace traditional receivers in commodity applications for RF signal reception, these sensors could be an enabling technology in niche application spaces. This review outlines the principles of operation of atomic electric field sensors and compares their performance capabilities to traditional RF receivers. It also highlights recent research and development efforts in atomic electric field sensing and identifies applications for which these sensors are projected to impact communications and remote sensing.

A Review of Commercial and Emerging Atomic Frequency Standards

Abstract: Atomic frequency standards are used to generate accurate and precise time and frequency, enabling many communications, synchronization, and navigation systems in modern life. GPS and other satellite navigation systems, voice and data telecommunications, and timestamping of financial transactions all rely on precise time and frequency enabled by atomic frequency standards. This review provides a snapshot and an outlook of commercial and emerging atomic frequency standards. We provide a concise summary of the performance and physics of operation of current atomic frequency standards. In addition, we discuss examples of emerging frequency standard technologies and prototype demonstrations with a focus on technologies expected to provide commercial or military utility within the next decade. We include a comparison of performance versus size and power for current atomic frequency standards. We develop and discuss an empirical relationship between frequency standard performance and product size.

The strategy and change interface: understanding “enabling” processes and cognitions

Abstract: The aim of this special issue is to better understand the strategy and change interface, in particular, the (sub)processes and cognitions that enable strategies to be successfully implemented and organizations effectively changed. The ten papers selected for this special issue reflect a range of scholarly traditions and, thus, as our review and integration of the relevant literatures, and our introductions to the ten papers demonstrate, they shed light on the strategy and change interface in starkly different ways. Collectively, the papers give us more insight into the recursive activities, and structural, organizational learning and cognitive mechanisms that are encouraged or deliberately established at organizations to allow their people to successfully implement a strategy and effect change, including achieve greater levels of horizontal alignment. Moreover, they demonstrate the benefits associated with establishing platforms and/or routines designed to overcome decision-makers’ cognitive shortcomings while implementing a strategy or making timely adjustments to it. We conclude our editorial by identifying some yet unanswered questions.

Flexible Mechanisms for Remote Attestation

Abstract: Remote attestation consists of generating evidence of a system’s integrity via measurements and reporting the evidence to a remote party for appraisal in a form that can be trusted. The parties that exchange information must agree on formats and protocols. We assert there is a large variety of patterns of interactions among appraisers and attesters of interest. Therefore, it is important to standardize on flexible mechanisms for remote attestation. We make our case by describing scenarios that require the exchange of evidence among multiple parties using a variety of message passing patterns. We show cases in which changes in the order of evidence collection result in important differences to what can be inferred by an appraiser. We argue that adding the ability to negotiate the appropriate kind of attestation allows for remote attestations that better adapt to a dynamically changing environment. Finally, we suggest a language-based solution to taming the complexity of specifying and negotiating attestation procedures.

Driver trust in and training for advanced driver assistance systems in Real-World driving

Abstract: Two-hundred and twenty-three participants completed an online survey regarding their experiences with advanced driver assistance systems (ADAS) on their personal vehicles, with focus on 1) drivers’ trust in 13 ADAS technologies, and 2) perceived effectiveness of currently used methods of training. Eighteen drivers participated in focus groups designed to probe more deeply into survey responses. Results of the survey showed that participant ratings of trust increased significantly with longer vehicle ownership, but participants who experienced unexpected ADAS technology behavior rated their trust over time significantly lower on ADAS technologies with the exception of rear collision avoidance. The majority (75.8%) of participants reported receiving some ADAS instruction at their vehicle dealership, but only 16.6% indicated it was formal. Participants who received formalized training reported it to be significantly more effective than those who received informal overviews of their systems. Use of trial and error and the owner’s manual were the most frequently reported methods of learning outside of dealership training. Responses indicated that the lack of content tailored to trim-specific vehicle features in owner’s manuals was a barrier to effective use.

ADE Eval: An Evaluation of Text Processing Systems for Adverse Event Extraction from Drug Labels for Pharmacovigilance

Abstract: The US FDA is interested in a tool that would enable pharmacovigilance safety evaluators to automate the identification of adverse drug events (ADEs) mentioned in FDA prescribing information. The MITRE Corporation (MITRE) and the FDA organized a shared task—Adverse Drug Event Evaluation (ADE Eval)—to determine whether the performance of algorithms currently used for natural language processing (NLP) might be good enough for real-world use. ADE Eval was conducted to evaluate a range of NLP techniques for identifying ADEs mentioned in publicly available FDA-approved drug labels (package inserts). It was designed specifically to reflect pharmacovigilance practices within the FDA and model possible pharmacovigilance use cases. Pharmacovigilance-specific annotation guidelines and annotated corpora were created. Two metrics modeled the experiences of FDA safety evaluators: one measured the ability of an algorithm to identify correct Medical Dictionary for Regulatory Activities (MedDRA®) terms for the text from the annotated corpora, and the other assessed the quality of evidence extracted from the corpora to support the selected MedDRA® term by measuring the portion of annotated text an algorithm correctly identified. A third metric assessed the cost of correcting system output for subsequent training (averaged, weighted F1-measure for mention finding). In total, 13 teams submitted 23 runs: the top MedDRA® coding F1-measure was 0.79, the top quality score was 0.96, and the top mention-finding F1-measure was 0.89. While NLP techniques do not perform at levels that would allow them to be used without intervention, it is now worthwhile exploring making NLP outputs available in human pharmacovigilance workflows.

Problems with Evidence Assessment in COVID-19 Health Policy Impact Evaluation (PEACHPIE): A systematic strength of methods review

Abstract: Introduction Assessing the impact of COVID-19 policy is critical for informing future policies. However, there are concerns about the overall strength of COVID-19 impact evaluation studies given the circumstances for evaluation and concerns about the publication environment. This study systematically reviewed the strength of evidence in the published COVID-19 policy impact evaluation literature. Methods We included studies that were primarily designed to estimate the quantitative impact of one or more implemented COVID-19 policies on direct SARS-CoV-2 and COVID-19 outcomes. After searching PubMed for peer-reviewed articles published on November 26 or earlier and screening, all studies were reviewed by three reviewers first independently and then to consensus. The review tool was based on previously developed and release review guidance for COVID-19 policy impact evaluation, assessing what impact evaluation method was used, graphical display of outcomes data, functional form for the outcomes, timing between policy and impact, concurrent changes to the outcomes, and an overall rating. Results After 102 articles were identified as potentially meeting inclusion criteria, we identified 36 published articles that evaluated the quantitative impact of COVID-19 policies on direct COVID-19 outcomes. The majority (n=23/36) of studies in our sample examined the impact of stay-at-home requirements. Nine studies were set aside because the study design was considered inappropriate for COVID-19 policy impact evaluation (n=8 pre/post; n=1 cross-section), and 27 articles were given a full consensus assessment. 20/27 met criteria for graphical display of data, 5/27 for functional form, 19/27 for timing between policy implementation and impact, and only 3/27 for concurrent changes to the outcomes. Only 1/27 studies passed all of the above checks, and 4/27 were rated as overall appropriate. Including the 9 studies set aside, reviewers found that only four of the 36 identified published and peer-reviewed health policy impact evaluation studies passed a set of key design checks for identifying the causal impact of policies on COVID-19 outcomes. Discussion The reviewed literature directly evaluating the impact of COVID-19 policies largely failed to meet key design criteria for useful inference. This was largely driven by the circumstances under which policies were passed making it difficult to attribute changes in COVID-19 outcomes to particular policies. More reliable evidence review is needed to both identify and produce policy-actionable evidence, alongside the recognition that actionable evidence is often unlikely to be feasible.

Estimating the effect of protected bike lanes on bike-share ridership in Boston: A case study on Commonwealth Avenue

Abstract: While many studies have studied the connection between cyclist ridership and the built environment, few findings provide relevant quantitative guidance to decision-makers. This study examines the effect of a single intervention (installation of a protected bike lane) in Boston, Massachusetts, on the nearby ridership of 'BlueBikes', a local bicycle sharing system (bikeshare). Bikeshare activity along the new protected bike lane almost tripled in the year following installation; however, ridership on routes unaffected by the new bike lane also saw dramatic increases in ridership. Using a differences-in-differences comparison, which assumes the bike lane had no influence on adjacent routes, suggests that the causal impact of the new bike lane increased bikeshare ridership +80% on affected routes. These quantitative estimates represent credible upper and lower bounds on the effect of replacing a conventional bike lane with a protected bike lane. Additional analysis also suggests that the influence of the bike lane is strongest when trip origins and destinations are a minimal distance (under 1.6 km) away from the bike lane, which may be useful information in planning bicycle networks.

Likert or Not? How Using Likert Rather Than Biposlar Ratings Reveal Individual Difference Scores Using the Godspeed Scales

Abstract: We examined how a revised presentation method of The Godspeed Scales affected results accruing from such tests which concern user judgments of anthropomorphism, animacy, likability, perceived intelligence, and perceived safety of robots. Through the use of Likert-type scales, rather than the original semantic- or bipolar-scale structure, we correlated results in order to determine which word pairs in the original scales were truly opposite in their meanings, where true opposites were anticipated to possess a strongly negative correlation. Results showed that individual differences in each participant’s baseline tendency to choose a rating exerted the strongest relationship with their overall scores. When those differences were accounted for the majority of the word-pairs used in the Godspeed Scale had negative correlations. These findings indicate that individual differences, rather than features of the robot per se, played the largest part in predicting how people will perceive any particular robot.

COVID-19 Telehealth Impact Study: Exploring One Year of Telehealth Experimentation

Abstract: Importance: This three-part study characterizes the widespread implementation of telehealth during the first year of the COVID-19 pandemic, giving us insight into the role of telehealth as we enter a stage of “new normal” healthcare delivery in the U.S. Objective: The COVID-19 Telehealth Impact Study was designed to describe the natural experiment of telehealth adoption during the pandemic. Using a large claims data stream and surveys of providers and patients, we studied telehealth in all 50 states to inform healthcare leaders. Design, Setting, Participants: In March 2020, the MITRE Corporation and Mayo Clinic founded the COVID-19 Healthcare Coalition (C19HCC), to respond to the pandemic. We report trends using a dataset of over 2 billion healthcare claims covering over 50% of private insurance activity in the U.S. (January 2019-December 2020), along with key elements from our provider survey (July-August 2020) and patient survey (November 2020 - February 2021). Main Outcomes and Measures: There was rapid and widespread adoption of telehealth in Spring 2020 with over 12 million telehealth claims in April 2020, accounting for 49.4% of total health care claims. Providers and patients expressed high levels of satisfaction with telehealth. 75% of providers indicated that telehealth enabled them to provide quality care. 84% of patients agreed that quality of their telehealth visit was good. Results: Peak levels of telehealth use varied widely among states ranging from 74.9% in Massachusetts to 25.4% in Mississippi. Every clinical discipline saw a steep rise with the largest claims volume in behavioral health. Provision of care by out-of-state provider was common at 6.5% (October-December 2020). Providers reported multiple modalities of telehealth care delivery. 74% of patients indicated they will use telehealth services in the future. Conclusions and Relevance: Innovation shown by providers and patients during this period of rapid telehealth expansion constitutes a great natural experiment in care delivery with evidence supporting widespread clinical adoption and satisfaction on the part of both patients and clinicians. The authors encourage continued broad access to telehealth over the next 12 months to allow telehealth best practices to emerge, creating a more effective and resilient system of care delivery.

Reducing COVID-19 Airborne Transmission Risks on Public Transportation Buses: An Empirical Study on Aerosol Dispersion and Control

Abstract: This study is one of the first COVID-19 related bus studies to fully characterize cough aerosol dispersion and control in the highly turbulent real-world environment of driving regular bus routes on both a school bus and a transit bus. While several other bus studies have been conducted, they were limited to clinical contact tracing, simulation, or partial characterization of aerosol transmission in the passenger areas with constraint conditions. When considering the risk of transmission of SARS-CoV-2 (COVID-19) and other highly infectious airborne diseases, ground based public transportation systems are high-risk environments for airborne transmission particularly since social distancing of six feet is not practical on most buses. This study demonstrates that wearing of masks reduced the overall particle count released into the bus by an average of 50% or more depending on mask quality and reduced the dispersion distance by several feet. The study also demonstrates an 84.36% reduction in aerosol particles and an 80.28% reduction in the mean aerosol residence time for some test cases. We conducted 84 experimental runs using nebulized 10% sodium chloride and a mechanical exhalation simulator that resulted in 78.3 million data points and 124 miles of on-the-road testing. Our study not only captures the dispersion patterns using 28 networked particle counters, as well as quantifies the effectiveness of using on-board fans, opening of various windows, use of face coverings or masks, and the use of the transit bus HVAC system. This work also provides empirical observations of aerosol dispersion in a real-world turbulent air environment, which are remarkably different than many existing fluid dynamics simulations, and also offers substantial discussion on the implications for inclement weather conditions, driver safety, retrofit applications to improve bus air quality, and operational considerations for public transportation organizations.

A Complex Network Framework for Validated Assessments of Systems of Systems Robustness

Abstract: The complexity of modern systems of systems (SoS) requires the ability to quickly and effectively evaluate robustness in architecture alternatives This article presents a framework for rapid, quantitative comparisons of robustness in SoS architectures that leverages complex network methods for assessing robustness and design of experiments (DoE) techniques for validating their use in the scenario of interest We consider both single-layer and multilayer network representations of SoS and focus on algebraic connectivity, inverse average path length, and largest connected component size as measures of robustness Two case studies are used to illustrate our framework and assess its utility: a command, control, communications, computer, intelligence, surveillance, and reconnaissance (C4ISR) simulation and a multilayer message-passing network simulation We find that most of the considered network metrics capture expected robustness trends, though their ability to capture these trends is often affected by the scenario of interest These results demonstrate the potential value of complex network methods for lightweight analysis of robustness in SoS architecture alternatives, when appropriately supported by DoE methods for understanding their limitations

Lex Rosetta: transfer of predictive models across languages, jurisdictions, and legal domains

Abstract: In this paper, we examine the use of multi-lingual sentence embeddings to transfer predictive models for functional segmentation of adjudicatory decisions across jurisdictions, legal systems (common and civil law), languages, and domains (i.e. contexts). Mechanisms for utilizing linguistic resources outside of their original context have significant potential benefits in AI & Law because differences between legal systems, languages, or traditions often block wider adoption of research outcomes. We analyze the use of Language-Agnostic Sentence Representations in sequence labeling models using Gated Recurrent Units (GRUs) that are transferable across languages. To investigate transfer between different contexts we developed an annotation scheme for functional segmentation of adjudicatory decisions. We found that models generalize beyond the contexts on which they were trained (e.g., a model trained on administrative decisions from the US can be applied to criminal law decisions from Italy). Further, we found that training the models on multiple contexts increases robustness and improves overall performance when evaluating on previously unseen contexts. Finally, we found that pooling the training data from all the contexts enhances the models' in-context performance.

Overbounding the effect of uncertain Gauss-Markov noise in Kalman filtering

Abstract: Prior work established a model for uncertain Gauss-Markov (GM) noise that is guaranteed to produce a Kalman filter (KF) covariance matrix that overbounds the estimate error distribution. The derivation was conducted for the continuous-time KF when the GM time constants are only known to reside within specified intervals. This paper first provides a more accessible derivation of the continuous-time result and determines the minimum initial variance of the model. This leads to a new, non-stationary model for uncertain GM noise that we prove yields an overbounding estimate error covariance matrix for both sampled-data and discrete-time systems. The new model is evaluated using covariance analysis for a one-dimensional estimation problem and for an example application in Advanced Receiver Autonomous Integrity Monitoring (ARAIM).

Optimizing network microsegmentation policy for cyber resilience

Abstract: This paper describes an approach for improving cyber resilience through the synthesis of optimal microsegmentation policy for a network. By leveraging microsegmentation security architecture, we ca...

Emerging approaches to support dynamic mission planning: survey and recommendations for future research:

Abstract: In the Department of Defense, unmanned aerial vehicle (UAV) mission planning is typically in the form of a set of pre-defined waypoints and tasks, and results in optimized plans being implemented p...

Proof of X: Experimental Insights on Blockchain Consensus Algorithms in Energy Markets

Abstract: The current centralized model of the electricity market is not efficient in performing distributed energy transactions required for the transactive smart grid. One of the prominent solutions to this issue is to integrate blockchain technologies, which promise transparent, tamper-proof, and secure transaction systems specifically suitable for the decentralized and distributed energy markets. Blockchain has already been shown to successfully operate in a microgrid peer-to-peer (P2P) energy market. The prime determinant of different blockchain implementations is the consensus algorithm they use to reach consensus on which blocks/transactions to accept as valid in a distributed environment. Although different blockchain implementations have been proposed independently for P2P energy market in the microgrid, quantitative experimental analyses and comparison of the consensus algorithms that the different blockchains may use for energy markets, has not been studied. Identifying the right consensus algorithm to use is essential for scalability and operation of the energy market. To this end, we evaluate three popular consensus algorithms: (i) proof of work (PoW), (ii) proof of authority (PoA), and (iii) Istanbul Byzantine fault tolerance (IBFT), running them on a network of nodes set up using a network of docker nodes to form a microgrid energy market. Using a series of double auctions, we assess each algorithm's viability using different metrics, such as time to reach consensus and scalability. The results indicate that PoA is the most efficient and scalable consensus algorithm to hold double auctions in the smart grid. We also identified the minimum hardware specification necessary for devices such as smart meters, which may run these consensus algorithms.

Modeling covid-19 for lifting non-pharmaceutical interventions

Abstract: As a result of the COVID-19 worldwide pandemic, the United States instituted various non-pharmaceutical interventions (NPIs) in an effort to slow the spread of the disease Although necessary for public safety, these NPIs can also have deleterious effects on the economy of a nation State and federal leaders need tools that provide insight into which combination of NPIs will have the greatest impact on slowing the disease and at what point in time it is reasonably safe to start lifting these restrictions to everyday life In the present work, we outline a modeling process that incorporates the parameters of the disease, the effects of NPIs, and the characteristics of individual communities to offer insight into when and to what degree certain NPIs should be instituted or lifted based on the progression of a given outbreak of COVID-19 We apply the model to the 24 county-equivalents of Maryland and illustrate that different NPI strategies can be employed in different parts of the state Our objective is to outline a modeling process that combines the critical disease factors and factors relevant to decision-makers who must balance the health of the population with the health of the economy © 2021, University of Surrey All rights reserved

A Machine Learning Application for Predicting and Alerting Missed Approaches for Airport Management

Abstract: This study applied machine learning technique to predict the occurrences of Missed Approach (MA) at an airport. In general, MA events are rare, and they typically occur at an aircraft’s descending phase a few minutes before landing. The occurrence of MA is usually indicative of safety-related concerns; however, it may cause excessive delays due to disrupting the arrival sequence during peak hours. This study focused on Denver International Airport (DEN) since MA events occurred more frequently as compared with other airports. Instead of using on-board data of individual aircraft, which is difficult to obtain systematically in real time, this study relied on airport operational data and developed a machine learning model to predict the likelihood of MA occurrences for DEN. The results of this study can inform the design and development of a real-time alerting tool for terminal or tower air traffic managers to take pre-emptive actions for mitigating the impact on terminal efficiency and increasing airport throughput. Lessons learned through this study can be applied to other airports with MA concerns.

Patient and Clinician Characteristics Associated With Secure Message Content: Retrospective Cohort Study.

Abstract: Background: Good communication has been shown to affect patient outcomes; however, the effect varies according to patient and clinician characteristics. To date, no research has explored the differences in the content of secure messages based on these characteristics. Objective: This study aims to explore characteristics of patients and clinic staff associated with the content exchanged in secure messages. Methods: We coded 18,309 messages that were part of threads initiated by 1031 patients with hypertension, diabetes, or both conditions, in communication with 711 staff members. We conducted four sets of analyses to identify associations between patient characteristics and the types of messages they sent, staff characteristics and the types of messages they sent, staff characteristics and the types of messages patients sent to them, and patient characteristics and the types of messages they received from staff. Logistic regression was used to estimate the strength of the associations. Results: We found that younger patients had reduced odds of sharing clinical updates (odds ratio [OR] 0.77, 95% CI 0.65-0.91) and requesting prescription refills (OR 0.77, 95% CI 0.65-0.90). Women had reduced odds of self-reporting biometrics (OR 0.78, 95% CI 0.62-0.98) but greater odds of responding to a clinician (OR 1.20, 95% CI 1.02-1.42) and seeking medical guidance (OR 1.19, 95% CI 1.01-1.40). Compared with White patients, Black patients had greater odds of requesting preventive care (OR 2.68, 95% CI 1.30-5.51) but reduced odds of requesting a new or changed prescription (OR 0.72, 95% CI 0.53-0.98) or laboratory or other diagnostic procedures (OR 0.66, 95% CI 0.46-0.95). Staff had lower odds of sharing medical guidance with younger patients (OR 0.83, 95% CI 0.69-1.00) and uninsured patients (OR 0.21, 95% CI 0.06-0.73) but had greater odds of sharing medical guidance with patients with public payers (OR 2.03, 95% CI 1.26-3.25) compared with patients with private payers. Staff had reduced odds of confirming to women that their requests were fulfilled (OR 0.82, 95% CI 0.69-0.98). Compared with physicians, nurse practitioners had greater odds of sharing medical guidance with patients (OR 2.74, 95% CI 1.12-6.68) and receiving prescription refill requests (OR 3.39, 95% CI 1.49-7.71). Registered nurses had greater odds of deferred information sharing (OR 1.61, 95% CI 1.04-2.49) and receiving responses to messages (OR 3.93, 95% CI 2.18-7.11) than physicians. Conclusions: The differences we found in content use based on patient characteristics could lead to the exacerbation of health disparities when content is associated with health outcomes. Disparities in the content of secure messages could exacerbate disparities in patient outcomes, such as satisfaction, trust in the system, self-care, and health outcomes. Staff and administrators should evaluate how secure messaging is used to ensure that disparities in care are not perpetuated via this communication modality. Trial Registration: