Student Lecture Series
Autumn 2017 • Tuesdays 12:00pm - 12:50pm • HSB T530
Using Mixed Reality to Improve Medical Practices
Imagine resizing the TV on your wall from 20" to 70" by pinching to zoom, or following a personalized painted line on the road as you drive from point A to point B. These are just two of many applications promised by mixed reality (MR) headsets, a relatively new generation of computing technologies that use holograms and new forms of human computer interactions to facilitate the manipulation and interpretation of digital information. However, little is known about MR's potential to improve patient care. In this research we develop and test two applications, Holographic Pre-procedural Planning Environment (HOPE) and Holographic Surgical Guidance System (HSGS), to explore how MR can facilitate the understanding of patient anatomy and surgical tools.
Trestle: A System for Integrating Historical Spatio-Temporal Data to Support Disease Surveillance
Increasingly, disease surveillance projects lean heavily on geospatial analysis and on having access to accurate information across spatial and temporal intervals, to support in-depth analysis. Unfortunately, modern Geographic Information Systems often struggle to efficiently manage and query complex spatio-temporal data from diverse data sources. This presents a tremendous challenge for public health researchers as they are often forced to either treat the world as a set of static snapshots, or deal with the complexity of manually integrating multiple years of historical spatial data.
Previous attempts at addressing this problem have focused on the development of theoretical models for managing changing temporal states of spatial objects and implementing basic temporal support into existing spatial analysis applications. Our approach builds upon these efforts and leverages advances in graph databases and knowledge representation engines. The result is Trestle, an integrated system that enables combining disparate geospatial datasets (such as administrative boundaries, road networks, and census tracts) into unified spatio-temporal objects. This enables users to efficiently access the correct object attributes for a given spatial and temporal region.
Our initial work has shown the feasibility of developing a robust implementation of a spatio-temporal data management application, which is capable of handling a number of datasets used in disease surveillance and epidemiology. Utilizing the Global Administrative Units Layers dataset from the United Nations, a prototype user portal and application programming interface has been developed that allows domain experts to access the necessary data within their existing research workflows.
By enabling the integration and management of complex spatio-temporal data, we hope to enable faster and more robust access to the types of data necessary to support disease surveillance efforts. Future work will focus on improving ease of use and enabling support for complex data access between related spatial objects.
From Data to Decision: Improving Decision Making with Electronic Health Records Data
In this talk, I will present my recent research efforts in the fields of biomedical and health informatics, human-computer interaction design, and artificial intelligence. During my PhD studies, my research has been focused on the early-stage development of innovative technologies that enable healthcare transformation from reactive and hospital-centered to preventive and person-centered. For the rest of the talk, I will focus on describing my approaches using design methods and mathematical models with Electronic Health Records (EHR) data for improving decision making. I will conclude with my future research directions, including the development of tools that leverage EHR data by reflecting the needs and barriers of multiple stakeholders (e.g., patients, health providers, and clinical researchers), and the development of mathematical models for predicting the spread of multiple types of cancers by integrating clinical data.
Examining the Feasibility of Internet of Things Technologies to Support Aging
The older adult population is one of the fastest growing demographic groups in the United States. Associated with this aging population are changes in health and wellness. Older adults face challenges such as chronic health conditions, reduced mobility, and cognitive decline. Technological solutions are valuable resources to assist older adults in maintaining their quality of life. One approach involves the Internet of Things (IoT) connected sensors which are designed to detect and record individuals’ activities and status within their living spaces. Despite the promise of these technologies to improve health outcomes and quality of life in older adults, there still remains a challenge in understanding older adults’ perceptions and concerns. We propose to conduct a pilot study to demonstrate feasibility and understand older adults’ preferences and needs using the IoT connected sensors within their home. The specific aims of this project are to: 1) Assess the feasibility of an IoT smart sensor system to monitor older adults in their residential setting; 2) Examine older adults’ perceived level of obtrusiveness of an IoT smart sensor system and how this perception may change over time and after exposure to such a system; 3) develop recommendations for the design of a system that will allow older adults to access and potentially control a home-based IoT smart sensor system.
The Wild West of Genomics
The genomic revolution has changed the way we do biology. My goal is to walk you through some common methods and how they look in the context our research. We will explore various sequencing methods such as RNA-seq, DNA-seq, ChiP-seq, and HiC, and see where data analysis was a decade ago, and where it has moved since then. We will also peak under the hood of genome assemblers, and see how reference sequences are generated, problems with them, and the cutting edge of genome improvement. The field of genomics is ever-evolving, with each new technique solving some old questions but creating just as many new ones. Welcome to the wild west of genomics!
Informatics for Health Policy
Health policy has a direct and meaningful impact on population health. Capturing, analyzing, and understanding data effectively may help strengthen health policy processes. Informatics solutions, including health information management systems and dashboards, can lead to improvements in policy and, therefore, the health of a country. In my seminar talk, I will present the use case of designing a solution for a global health informatics project, including some high level functionality and challenges that have surfaced. I will also provide some background about the project setting and context about the health policy life cycle. Finally, I would like to illicit feedback from the audience on initial dissertation ideas, interesting research questions, and next steps.
NO CLASS (WEEK OF AMIA)
Revealing Hidden White Blood Cell Count Phenotypes for Gene Discovery: Deep Phenotyping with Latent Class Mixed Modeling
Deep phenotyping for gene discovery aims to identify more precise phenotypes to increase the power to detect and the effect size of associated genetic variants. One strategy to deepen phenotype is to harness longitudinal data to interrogate phenotypes that present as disease or developmental trajectories. However, trajectory heterogeneity may be difficult to discern in large, observational datasets using standard statistical methods. Latent class mixed modeling (LCMM) is a method that can identify unobserved heterogeneity in longitudinal data and attempts to classify individuals into groups based on a linear model of repeated measurements over time. We applied LCMM to repeated white blood cell (WBC) count measures derived from the electronic medical record among participants of the National Human Genetics Research Institute (NHRGI) electronic MEdical Record and GEnomics (eMERGE) network study, revealing two WBC count trajectory phenotypes. Advancing these phenotypes to a genome-wide association study, we found novel genetic associations between WBC count trajectory class membership and regions on chromosome 1p34.3 and chromosome 11q13.4. The chromosome 1 region contains CSF3R, which encodes the granulocyte colony stimulating factor receptor. This protein is a major factor in neutrophil stimulation and proliferation. The association on chromosome 11 is in an intron of RNF169; its gene product is involved in the regulation of double strand break DNA repair.
The Feasibility of Using Personal Health Records with Homebound Older Adults
Patient activation, or an individual’s willingness and ability to take actions to maintain their health and wellness, is a primary component of the patient-centered health system. Activated patients are more likely to report positive experiences with their medical providers, have better health outcomes, and spend less on healthcare services. Although all patients can benefit from services that support increased activation, recent literature has found that some patient populations are more likely to include activated patients than other populations.
Homebound older adults face more barriers to patient activation than their non-homebound peers. Because people who are homebound are unable to leave their homes without significant assistance, regularly accessing clinic-based medical services is difficult. In addition, as a population, homebound older adults have more chronic diseases, physical and cognitive impairments, and challenges with activities of daily living than non-homebound older adults.
The number of older adults who are homebound is on the rise, and they are a growing proportion of the older adult patient population. Therefore, more research is needed to understand how consumer health information tools can used with this population to support activation and improve health outcomes.
This dissertation explores whether personal health records can be used by homebound older adults following a hospitalization. In the first paper, I discuss what features would be necessary for a PHR to meet the needs of homebound older adults, and I assess commercial PHRs for these features. Paper two describes a usability study conducted on two commercial PHR systems from the perspective of a homebound older adult and their care team. Paper three evaluates the feasibility of partnering with home health agencies to recruit homebound older adults following a hospitalization. Finally, paper four describes the feasibility of conducting a PHR pilot study with homebound older adults and their caregivers. In addition, this paper also describes an analysis of the preliminary effectiveness of PHRs in a homebound older adult care environment. Together, these papers further our understanding of the unique challenges and opportunities for personal health records in the homebound older adult population.
Telemedicine: Virtual Care Clinics
Healthcare organizations are increasingly adopting telemedicine services to increase access and reduce costs. Virtual care clinics (VCCs) are a direct to consumer telemedicine service that is gaining traction. In this talk, I will provide a brief overview of virtual care clinics along with a few examples. Using insights from a recent qualitative study conducted with some early adopters of virtual care clinics I will discuss the target market for these clinics and marketing strategies adopted by healthcare organizations. Finally, I will conclude with a discussion on some challenges that exist for successful patient engagement with telemedicine.
Failure-to-Rescue: Detecting Early Signs of Patient Deterioration
Failure-to-Rescue (FTR) is the research area focusing on preventable patient deterioration. Along with the sheer amount of clinical data available in electronic form and recent breakthroughs in machine learning research, data-driven forecast on patient’s progress has been getting more accurate. In this talk, I’ll present on how machine learning approaches have been applied to predict patient outcome and disease-onset. Moreover, I’ll go over potential ways to improve the current clinical practice based on what we learned from FTR cases using machine learning approach.
An Ontological Approach to Personalizing Health Dialog
A combination of recent advances in speech recognition and machine learning is ushering in a new era in human-computer interaction, conversational UI. While conversational agents have been a long-standing goal of general artificial intelligence, generating health-specific dialog presents many unique challenges that require a more structured approach than current general domain methods. However, past attempts within the medical domain have often relied on entirely scripted dialog thus limiting their maintainability and application to multiple sub-domains. In this talk, I will describe my current approach to developing a dialog system for patient health education through unscripted document explanation.