AMIA 2015 Panels

Didactic Panels

This panel format is on a specific aspect of theory, application, or experience pertaining to any aspect of biomedical or health informatics, or provides interdisciplinary viewpoints that cut across traditional themes.

Sunday, November 15, 2015

3:30 p.m. – 5:00 p.m.

S02: The Sociotechnical Perspective in Biomedical Informatics: What do We Understand and Measure?

J. Aarts, University at Buffalo/Erasmus University; J. Ancker, Weill Cornell Medical College; A. Kushniruk, University of Victoria; D. Sittig, University of Texas Health Science Center; J. Ash, Oregon Health & Science University

Sociotechnical theory has become accepted in the last fifteen years for understanding and evaluating the complexity of design and implementation of health information technology (HIT) in complex health care organizations. The IOM report on information technology and patient safety argues that safety is the product of the sociotechnical system and its constituent parts. The core concept of sociotechnical theory is interdependency. Sociotechnical theory raises a number of methodological questions, since it sits uneasy with accepted research paradigms in biomedicine that focus on measurable clinical outcomes research. The panel seeks to address the value of sociotechnical theory, the underlying research methodologies and to give practical guidance on how studies involving a sociotechnical perspective can be conducted and published.

S03: Wearable Health Data and the Quantified Self: What Role in the Clincial Context?

E. Chiauzzi, PatientsLikeMe; K. Patrick, University of California; C. Bloss, Qualcomm Institute/Calit2, University of California, San Diego; E. Ramirez, QS Labs

The popularity of health self-tracking through the use of consumer health devices has primarily been confined to fitness and wellness contexts, but there has been increasing interest in leveraging this health data in clinical decision-making and care delivery. Because personal tracking devices provide a glimpse into the everyday behavior of patients, there is potentially great value in integrating this data with traditional medical data. Such integration may be particularly useful with patients experiencing chronic diseases, as their conditions require daily self-management that include activity goals, sleep monitoring, an understanding of calorie expenditure, and tracking of medication adherence. However, there are many questions that are unresolved. There is a need for clarity about how such data can be utilized in clinical encounters. There are numerous devices available to consumers and they vary greatly in features and data quality. The patient perspective has not been fully leveraged, as devices generally target a fitness market. Users of tracking devices may not be fully cognizant of issues around privacy, data sharing, and data control. Most importantly, does the use of self-tracking devices affect disease outcomes?

Monday, November 16, 2015

8:30 a.m. – 10:00 a.m.

S12: Public Implementation Resources for Genomic Medicine

J. Peterson, Vanderbilt University Medical Center; M. Williams, Geisenger Health System; C. Overby, University of Maryland School of Medicine/Geisenger Health System; R. Freimuth, I. Kullo, Mayo Clinic

Genomic Medicine programs have proliferated as the evidence for implementing compelling pharmacogenomics use cases or targeted therapy accumulates. However, current electronic health records (EHRs) are under-developed for manipulating genomic data or communicating results to patients and providers. Three NIH-funded networks – Electronic Medical Records and Genomics (eMERGE), Pharmacogenomics Research Network (PGRN), and Implementing GeNomics In PracTicE (IGNITE) -- are creating health information technology (HIT) resources to help fill these gaps. Members of these consortia, which span academic medical centers and integrated health systems, are working to create educational tools, patient engagement technologies, clinical decision support resources, and common data models to exchange structured genomic results and guidelines. During this panel, speakers will review 1-2 resources that are available publically, and discuss how to leverage these resources to implement a new precision medicine or translational program. A panelist will describe methods and resources to encode genomic data and knowledge so that is can be shared between EHR environments. Secondly, a repository for storing and indexing genomic clinical decision support artifacts will be presented as well as efforts to use Infobuttons to provide just-in-time genomic medicine education. Finally, a panelist will present engagement technologies which can assist with communicating genomic results directly to patients.

S13: Harmonization of ICD-11 and SNOMED CT - Not just Mapping! Practical and Theoretical Lessons and Benefits to Users and Implementors

A. Rector, University of Manchester; J. Campbell, University of Nebraska Medical Center; B. Ustun, World Health Organization; C. Chute, Johns Hopkins University; H. Solbrig, Mayo Foundation

ICD and SNOMED CT were designed for different purposes – ICD for statistical reporting and epidemiological studies; SNOMED CT for documenting clinical care. They conform to different semantics – ICD is a closed mono-hierarchy in which sibling entities must be mutually exclusive and jointly exhaustive. SNOMED is an open polyhierarchy based on description logic. ICD has a strong requirement to focus on identifying diagnosis and underlying cause of death; SNOMED has an equally strong requirement to reflect the evolution of clinical understanding. ICD’s usage has clearly outgrown its historical structure. Updates occur at long intervals, and migration between versions is costly. To meet these problems and harmonize with SNOMED, ICD-11 uses a radically new three-component architecture: A “Common Ontology” that will correspond to an agreed subset of SNOMED, a “Foundation Component” structured around this ontology, and “Linearizations” that correspond to the existing classification, plus algorithms that link these component. The result is expected to simplify generating maps between SNOMED and ICD and to provide a stable platform for the evolution of ICD. The development has required clarifying the semantics of both ICD and SNOMED and has important implications for anyone working with either or attempting to harmonize other systems with them.

10:30 a.m. – 12:00 p.m.

S24: Natural Language Processing for Phenotype Extraction: Challenges in Extraction and Representation

H. Hochheiser, R. Jacobson, University of Pittsburgh; G. Savova, Pediatrics, Children's Hospital Boston; J. Denny, Vanderbilt University; N. Washington, Lawrence Berkeley National Labs

Previous work has demonstrated the benefits of using Natural Language Processing (NLP) techniques for extracting diagnoses from clinical texts and assigning patients to categories for cohort identification and GWAS studies using electronic health record (EHR) data. However, there is a growing awareness that understanding cancers and other complex or rare diseases will require richer phenotypic models that describe the development, progression, and location of specific signs and symptoms. Computing over these “deep phenotypes” will require NLP methods that can interpret individual documents, summarization approaches that can combine information from multiple notes and document types (e.g., radiology, pathology, and clinical notes) to build a longitudinal patient, and structured data models that will facilitate the integration between NLP and existing discrete data and comparison of these phenotypic models. This panel will discuss challenges inherent in developing both data models and the NLP tools. Strategies involving the use of NLP pipelines (cTAKES), clinical data models (FHIR), and phenotype data exchange formats will be discussed. Attendees at this panel will gain an understanding of the NLP challenges in extracting deep phenotype information, strategies for representing phenotype data, and emerging systems and tools that are applying these techniques to challenging translational problems.

S25: The Clinical Quality Framework Initiative to Harmonize Decision Support and Quality Measurement Standards: Defined Standards, Pilot Results, and Moving Beyond Quality Improvement

K. Kawamoto, University of Utah; M. Hadley, MITRE Corporation; T. Oniki, Intermountain Healthcare; J. Skapik, Office of the National Coordinator for Health IT

The Clinical Quality Framework (CQF) is a public-private partnership sponsored by the Office of the National Coordinator for Health IT (ONC) and the Centers for Medicare & Medicaid Services (CMS) to create a harmonized set of standards for clinical decision support (CDS) and electronic clinical quality measurement (eCQM). At AMIA 2014, experts presented preliminary results from the first year of this multi-year initiative. In this panel, CQF leaders will describe the standards developed, including the HL7 Clinical Quality Language standard for logical expression, the HL7 FHIR Quality Profiles for data exchange, the HL7 Quality Improvement and Clinical Knowledge model for logical inferencing, and several implementation guides leveraging these foundational standards. The panelists will then present findings from pilot deployments of the standards in various clinical domains ranging from cardiology to radiology, preventive care, infectious diseases, and oncology. The panelists will also describe partnerships with relevant initiatives beyond the quality domain, including the ONC Data Access Framework initiative, the Health Services Platform Consortium (HSPC), and the Clinical Information Modeling Initiative (CIMI).

1:45 p.m. – 3:15 p.m.

S34: Informatics Approaches to Supporting Emerging Accountable Health Care Delivery Methods

G. Kuperman, New York Presbyterian/Columbia University; D. Bates, Brigham and Women's Hospital/Partners HealthCare System/Harvard University; D. Kaelber, The MetroHealth System/Case Western Reserve University; D. Dorr, Oregon Health and Science University

Medicare, Medicaid and commercial payers are creating incentives for efficient care delivery models and the health care market place is responding rapidly. The new incentives require health care delivery organizations – often for the first time – to include efficiency of care delivery as a key strategic goal. These dynamics are changing the way provider organizations are delivering care. New information infrastructures -- including analytics, work flow support for new care roles (e.g., care coordinators), information systems support for team-based work flows, information-enabled interactions with new business partners, and improved interactions directly with patients -- are emerging to manage patients in the new environment. The learning objective for this panel is for attendees to develop a better understanding of how leading organizations are addressing changes in health care payment models and the informatics-enabled approaches they are taking to support workflow changes, team-based care and the need for data-driven organizational and health care management.

S35: Recent Advances in Computational Drug Repositioning

P. Zhang, IBM T.J. Watson Research Center; A. Butte, University of California San Francisco; N. Shah, Stanford University; N. Tatonetti, Columbia University; H. Xu, The University of Texas Health Science Center at Houston

Computational drug repositioning is a promising and efficient tool for discovering new uses from existing drugs and holds the great potential for precision medicine in the age of big data. The explosive growth of large-scale genomic and phenotypic data, as well data of small molecular compounds with granted regulatory approval, is enabling new developments for computational repositioning. To achieve the shortest path towards new drug indications, advanced data processing and analysis strategies are critical for making sense of these heterogeneous molecular measurements. Despite the progress simulated by big data analytics, there is clearly room for technical improvement with regard to computational drug repositioning methods. Furthermore, to materialize the true potential and impact of these methods, much work is needed to show that they can be successfully adopted into practical applications. In this panel, participants will summarize the recent advances in computational drug repositioning and identify challenges and opportunities. Panel participants will synthesize their perspectives on these key issues and likely future developments in this area, explore a diverse set of topics, and engage in thoughtful discussion with the audience.

S36: The Public Health Informatics Workforce: Challenges, Initiatives, and the Path Forward

B. Dixon, Indiana University Fairbanks School of Public Health/Regenstrief Institute; S. Grannis, Regenstrief Institute/Indiana University School of Medicine ; M. LaVenture, Minnesota Department of Health; B. Brand, Public Health Informatics Institute; A. Davidson, Denver Public Health

Informatics professionals make significant contributions to detecting, monitoring and improving population health. Although informatics professionals exist in local, state and federal public health agencies, we know little about them or their needs. Yet there exist several disparate education and training programs in public or population health informatics (PHI) that seek to prepare the public health workforce for designing, implementing, and using PHI systems. This panel presents diverse views on the PHI workforce and its needs from local, state, federal and international levels. Current data on the characteristics and information needs of the PHI workforce will be presented. Then panelists will discuss the various initiatives and programs that seek to educate or train the PHI workforce. The panel will raise questions and discuss opportunities about how to better organize information about training opportunities and the role that AMIA should play in supporting the existing and future PHI workforce.

3:30 p.m. – 5:00 p.m.

S46: Informatics Research and Innovation in a Commercial Electronic Health Record: The Experience of Three Organizations using Epic

A. Wright, D. Bates, Brigham and Women's Hospital/Harvard Medical School; E. Kirkendall, Cincinnati Children's Hospital Medical Center/University of Cincinnati; D. Dorr, Oregon Health and Science University; P. DeVault, Epic Systems Corp.

Historically, many of the key innovations in informatics came from organizations with self-developed clinical information systems. Having a self-developed system gives organizations considerable flexibility to tailor and extend their software and explore new and potentially innovative approaches. However, maintaining a self-developed system is expensive, poses risks to generalizability and makes scaling challenging. Recognizing these issues, most academic medical centers have moved, or are in the process of moving, to commercial electronic health records (EHRs).  In this panel, we present the experience of three organizations which are actively implementing innovative approaches to improve care: Partners HealthCare, Cincinnati Children's Hospital Medical Center and Oregon Health and Science University. All three organizations, at one time, had pioneering self-developed systems but switched to Epic, one of the most popular EHRs used worldwide.  The presenters will detail their organizations’ experience innovating on the Epic platform through the lens of several case studies, introduce the range of technical approaches for developing and integrating innovative tools in Epic, discuss Epic’s approaches to working with innovators and cover issues in organization governance of innovation activities.  Though the focus of the panel is the experience with Epic, the lessons learned apply to all commercial EHRs.

S48: ClinicalTrials.gov: Adding Value through Informatics

N. Smalheiser, University of Illinois at Chicago; V. Huser, National Library of Medicine; A. McCray, Harvard Medical School; A. Tasneem, Duke University; C. Weng, Columbia University

ClinicalTrials.gov is a repository of registered clinical trials maintained by NLM containing detailed descriptions of trial sponsorship, design, and results (when available). ClinicalTrials.gov plays an increasingly important, pivotal role in evidence-based medicine, and serves a diverse audience ranging from clinical researchers, who are designing and conducting new trials and recruiting patients; systematic reviewers, who are summarizing the best available evidence regarding safety and efficacy; bio-entrepreneurs, who are looking for drug repurposing or new therapeutic opportunities; and patients, who may be looking for a suitable clinical trial that might accept them. This panel will present an overview of ClinicalTrials.gov and discuss several ongoing lines of informatics research that are adding value -- for example, using text mining to improve the computability of eligibility criteria, design attributes and outcome results and connect these with patient EHR data; linking a registered trial with the publications arising from that trial; and performing aggregate analyses across trials to extract reusable design knowledge, understand design patterns and trends, and uncover systematic biases. The panelists will also discuss challenges and opportunities for further evolution of ClinicalTrials.gov, particularly in light of emerging trends such as patient-centered clinical trials, or the use of unpublished trial data in meta-analyses.

Tuesday, November 17, 2015

8:30 a.m. – 10:00 a.m.

S57: The Best of Imaging Informatics Research 2015

C. Kahn, University of Pennsylvania; B. Erickson, Mayo Clinic

Over the past decade, there have been rapid advances in imaging informatics research. As the storage and display of multidimensional image data has become commonplace and “off the shelf”, imaging informatics has focused increasingly on information not found in pixels and voxels. For example, new ontologies are now available to capture information in the imaging report. New structured data capture technologies and natural language processing capabilities provide discrete data that can be analyzed in concert with other data from the electronic medical record. Likewise, novel quantitative image analysis techniques give us new methods to diagnose clinical conditions and discern their time course. Machine learning methods applied both to images and imaging reports have accelerated our ability to extract meaningful clinical information from images. Several more practical advances, such as standardized work flow and analytic tools have modernized the practice of radiology and other clinical imaging specialties. Advances in imaging informatics research will be reviewed in this session.

S58: Looking Back and Moving Forward: A Review of Public and Global Health Informatics Literature and Events

B. Dixon, Indiana University Fairbanks School of Public Health/Regenstrief Institute; J. Pina, RTI International;  J. Richards, Centers for Disease Control and Prevention; H. Kharrazi, Johns Hopkins School of Public Health; A. Turner, University of Washington School of Public Health

Over the past decade, the disciplines of public health and global health informatics have rapidly expanded and continue to evolve within the field of biomedical informatics. Increased attention and activity by the U.S. Centers for Disease Control and Prevention as well as many health ministries, the World Health Organization, and non-governmental organizations are generating new knowledge and lessons regarding the development, implementation, and use of information systems in health care delivery around the globe. Thus, a growing body of literature now contains important insights and lessons from international informatics activities, stimulating the need to synthesize the knowledge for the field. In this panel, a review of recent literature in the areas of public health and global health informatics will be presented. Key articles revealing trends, methods, and lessons will be summarized to bring attendees up-to-date on the use of informatics to improve population health as well as care in low resource settings.

10:30 a.m. – 12:00 p.m.

S61: User-centered Methods to Optimize Clinical Decision Support: Examples from Pediatrics with Applicability to All Care Settings

D. Karavite, The Children's Hospital of Philadelphia ; E. Shelov, L. Utidjian, The Children's Hospital of Philadelphia/University of Pennsylvania; J. Michel, E. Lourie, The Children's Hospital of Philadelphia

Unintended consequences of health information technology (HIT) and clinical decision support (CDS) are frequently a result of poor design. Usability problems in electronic health records (EHR) have gained national attention. The need to increase effectiveness and reduce harm is vital and urgent. User-centered methodologies provide an array of methods to design systems around the requirements of the users and are critical to developing HIT that is safe and effective. In a series of vignettes, the panelists will describe applications of user-centered methods during the design of CDS projects in a multispecialty pediatric healthcare system. The panelists will review common user-centered design methods, suggest how to choose methodologies, and discuss both the benefits and challenges of following a user-centered approach in busy clinical settings. Panelists will describe examples of unintended consequences of CDS interventions and user-centered methods to prevent them. The panel moderator is a formally-trained in user-centered methods and All panelists are pediatric informaticists with experience in the design and implementation of HIT/CDS interventions.

S62: Developing Natural Language Processing Systems for Healthcare

G. Gobbel, R. Reeves, Tennessee Valley Health System VA/Vanderbilt University; D. Finch, James A Haley Veterans' Hospital; W. Chapman, University of Utah; J. Garvin, University of Utah/VA Salt Lake City Health Care System

Natural language processing (NLP) systems can help uncover medical knowledge buried within clinical free text. Despite multiple demonstrations of the clinical utility of NLP, its use for biomedical research and patient care remains limited. A detailed overview of the steps involved and potential challenges associated with creating an NLP system could help potential users of such systems; they could use this knowledge to assess feasibility and identify the types of expertise required for system development and implementation. This panel will delineate four key tasks commonly required when developing a clinical NLP system: 1) document selection; 2) annotation; 3) NLP software development; and 4) clinical implementation. Presenters will explain the role of each task in NLP system development, define the subtasks required, describe approaches for estimating the effort and time requirements, and delineate approaches and tools for assisting with task completion. Attendees will acquire an understanding of the common requirements for NLP system development, an appreciation of potential tools to help with the process, and be better able to assess whether and how to approach using NLP in their own biomedical centers.

1:45 p.m. – 3:15 p.m.

S70: Needs of the Digital Native: Adolescents and Access to PHRs

C. Smith, University of Wisconsin-Madison; F. Bourgeois, Boston Children's Hospital; P. Charney, Bellevue College; P. Brennan, University of Wisconsin-Madison

PHRs present particular challenges in pediatrics and the pediatric PHR is a particularly timely, while controversial, issue in medical informatics. This is true because minor patients present “unique health privacy and confidentiality standards,” different users within a family unit have different rights to access different information, and there are “extensive variations” in PHR implementation and access across the United States. This panel considers the challenges, benefits and risks of access to PHRs by adolescent patients. In order to address the needs of this population, we have to support access policies that handle age-specific changes in PHR ownership, as well as encourage EHR functionality that allows appropriate identification and classification of confidential information. We also offer two examples of how these challenges play out in context of adolescents’ lived experiences. How is care coordinated for adolescents with chronic health conditions when PHRs are not available? Can PHRs play a role in assisting adolescents and young adults living with chronic illness to develop self-advocacy and self-awareness?

S71: Rapid Development and Implementation of Critical Information Systems for Ebola Treatment Centres in West Africa: Lessons for Future Events

J. Teich, Harvard Medical School/Elsevier; H. Fraser, University of Leeds/Harvard University; E. Perakslis, Harvard University; S. Oza, London School of Hygiene and Tropical Medicine; D. Jazayeri, ThoughtWorks Uganda, Ltd.

In response to the large West African Ebola outbreak beginning in 2014, Ebola Treatment Centres (ETCs) were established to provide efficient care for suspected and confirmed Ebola patients while minimizing risk of cross infection to staff and other patients. Critical clinical information entry and review tasks such as assessments, ordering, and administration of drugs and fluids must be performed rapidly in highly challenging conditions because workers wear personal protective equipment (PPE) which reduces dexterity, vision and comfort. Additionally, it is vital to have information tools to conduct important functions, such as contact tracing to limit the spread of new cases and trials of new diagnostic/therapeutic approaches. In this panel, members of two teams providing rapid development and implementation at different centres describe the unique situational needs and challenges, software and hardware considerations and solutions, user engagement throughout testing and rollout, and approaches to advanced functionality such as analytics and clinical decision support. The audience will be better able to plan future implementations in challenging environments; to include information management in advance planning for future crises; and should be able to leverage available technologies and strategies for effective design and deployment of crisis information systems.

3:30 p.m. – 5:00 p.m.

S80: Health Information Technology and Large-scale Adverse Events

D. Sittig, University of Texas Health Sciences Center; P. Kilbridge, The Advisory Board Company; J. Scott, Veterans Health Administration; F. Magrabi, Macquarie University

There is an urgent need to address large-scale adverse events associated with health information technology (HIT) because the opportunity for harm to numerous patients is intensified with rapid implementation of systems worldwide. This panel will review current evidence about large-scale events including their impact on care delivery and consequences for patients. It will then use a case study of a high profile failure to discuss lessons for safe implementation and operation of present day HIT systems. The panel will then turn its attention to processes for detection and management of HIT events. And finally it will examine strategies for prevention and mitigation of large-scale events. The panel will aim to improve understanding about large-scale events and to transfer learnings about evidence-based best practice methods to improve preparedness, detection, response and recovery from HIT events, and to identify gaps and areas for further work.

S81: Collaboration and Health Information Technologies: Towards Defining and Operationalizing the Collaboration Space

C. Kuziemsky, University of Ottawa; M. Reddy, Northwestern University; K. Siek, Indiana University; S. Collins, Partners Healthcare Systems

Despite calls for increased collaborative care delivery it is still a challenge to operationalize collaboration, and more specifically, to design and evaluate HIT to support collaboration. Reasons for that include a lack of integrated studies on different aspects of collaboration and a lack of research with an explicit focus on collaboration. However, studying collaboration can be challenging given the range of processes, providers and settings where collaboration takes place. To advance research on collaboration and HIT we need ways to bound studies of collaboration according to the different type of collaborations and the overall collaboration space where they exist. This panel will provide insight on how to define and operationalize the collaboration space. We will discuss the structure of the collaboration space and how it provides bounding for studies of collaboration as well as discuss clinical collaboration and patient centered participatory medicine as two specific contexts of collaboration spaces.

S82: Data Quality in Clinical Data Research Networks (CDRNs)

A.B. McCoy, Tulane University School of Public Health and Tropical Medicine; R. Jain, Louisiana Public Health Institute; J. Doctor, University of Southern California; M. Kahn, University of Colorado; L. Waitman, University of Kansas Medical Center

National initiatives are increasingly leveraging the vast amounts of data availed by health information technology. PCORnet, the National Patient-Centered Clinical Research Network is one such example, and it aims to improve the ability to perform comparative effectiveness research by integrating data from clinical data research networks (CDRNs) and patient-powered research networks (PPRNs). One key component to the success of these networks will be ensuring the quality of the electronic health record and patient reported outcome data over time and across organizations. Clinical research informaticians have indicated challenges to data quality stemming from data integration, data capture, data exchange, and data standardization. Discussion items include the importance of data quality, a framework and plan for assessing data quality, the effects of health information system infrastructure on data quality, and data quality considerations for patient reported outcomes. Shared learnings from PCORnet stakeholders committed to the network’s excellence should help inform the national dialogue.

S83: State of the Art of Clinical Narrative Report De-Identification and Its Future

O. Uzuner, State University of New York, Albany; J. Aberdeen, MITRE Corporation; S. Meystre, University of Utah/VA Health System; M. Kayaalp, NIH

While automatic de-identification systems exist, release of de-identified data usually requires significant multi-round expensive effort for validation. To overcome this barrier, we need a consensus on the parameters of successful automatic de-identification. Although we can establish such parameters relative to error rates of human annotators, it is ultimately a policy question whose answer needs to be vetted by the public. When personal identifiers are substituted with surrogates or pseudonyms, it could be very difficult to spot the residual identifiers missed by the de-identifier, but in absolute terms, it is difficult to ensure that de-identified clinical text contains no references that might indirectly identify the patient; hence, de-identified clinical text is usually shared through a data use agreement. When such an agreement is in place, dates and some address parts can be left identified in a limited data set. If we reframe the de-identification problem by focusing on the pertinent identifiers, we may smooth the path to data sharing. Installing and running a clinical de-identification system may require substantial expertise, which small institutions and clinical scientists may lack. Although we would like to develop capable systems with sophisticated functionalities, we also should strive for simplicity for routine de-identification tasks.

Wednesday, November 18, 2015

8:30 a.m. – 10:00 a.m.

S93: Clinical Decision Support: How to Apply Standards to Deliver Knowledge-driven Interventions

R. Jenders, Charles Drew University/UCLA, G. Del Fiol, K. Kawamoto, University of Utah; H. Strasberg, Wolters Kluwer Health

Clinical decision support (CDS) is receiving increasing attention as a way to help improve clinical practice and health behaviors. The use of health information technology (HIT) standards for encoding data, representing knowledge and delivering knowledge-based interventions can help facilitate implementation of CDS. However, many standards from numerous standards development organizations (SDOs) exist that are variously incorporated into vendor software, and consensus on the use of these standards is lacking. Moreover, newly proposed standards elaborated in the past year in CDS and related areas such as clinical quality measurement have increased the complexity of this domain. The recent emergence of the Health Level Seven Fast Healthcare Interoperability Resources (FHIR) standard for patient data representation is particularly pertinent.  Accordingly, the panel members, who are clinicians and co-chairs of the Health Level Seven International (HL7) CDS Work Group, will address two learning objectives. First, attendees will learn key details of extant and proposed HIT standards that are applicable to CDS which they may need to use in their work. Second, the attendees will recognize how these standards have been applied in actual implementations that are used to provide CDS that can be used to improve the quality of clinical care and outcomes.

10:30 a.m. – 12:00 p.m.

S102: Perioperative Clinical Decision Support: Improving Care of the Surgical Patient through Informatics

R. Epstein, Thomas Jefferson University; K. Poterack, Mayo Clinic; P. Guffey, Children's Hospital Colorado; B. Nair, University of Washington; B. Pickering, Mayo Clinic

The perioperative period is a complex environment that can benefit significantly from the implementation of clinical decision support (CDS), given the large volume and velocity of data, the presence of many distractions, and increasing documentation requirements. Panel participants will describe projects they have successfully implemented that relate to four separate domains of perioperative CDS. Implementation of post hoc provider-specific feedback through the use of dashboards will be discussed in the context of improving the quality of patient care. Guidance of provider activities at the point of patient care in the operating room through integration of data from physiologic monitors, anesthesia machines, and the electronic health record and delivery of timely advice will be presented. A CDS system to facilitate compliance with the new SCIP 4 protocol related to control of blood glucose following cardiac surgery will be described. Translation of published research regarding Bayesian statistical predictions for time remaining during surgical cases to a real time whiteboard to facilitate operating room management CDS will be shown. Panelists will discuss issues related to implementing CDS in the context of modern electronic health systems in which access to necessary data in near real-time has proven to be challenging.

S103: The Value of an Open-Source Observational Research Collaboratory: Results from the OHDSI Initiative

J. Duke, Regenstrief Institute; G. Hripcsak, Columbia University; N. Shah, Stanford University; P.  Ryan, Janssen Research and Development

The Observational Health Data Sciences and Informatics collaborative (OHDSI, pronounced ‘Odyssey’) was formed in 2013 with the goal of creating reliable scientific evidence through large-scale analysis of observational health data from around the world. To advance this goal, OHDSI has had to grow rapidly its scientific, technical, and community infrastructure. OHDSI has now grown to over 120 participants from 10 countries. Across the collaborative, there are over 50 databases covering hundreds of millions of patient lives. OHDSI has conducted multiple international network-based observational research studies using this infrastructure. In this panel, we present initial results and lessons learned from the OHDSI initiative.

S104: The Implementation of Online Patient Portals in Safety Net Settings: The Realities of Meaningful Use Certification with Vulnerable Patient Populations

C. Lyles, U. Sarkar, N. Ratanawongsa, University of California, San Francisco; D. Oryn, Petaluma Health Center

This panel will provide insights and early data about the experiences of safety net healthcare settings implementing online patient portals in order to meet Meaningful Use Stage 2 certification deadline during 2015. While the vast majority of US healthcare systems are in the midst of implementing patient access to online patient portal websites to allow patients to view their medical information online and communicate electronically with providers in between office visits, healthcare systems serving predominantly low-income and uninsured/publicly insured face unique challenges with patient as well as provider/staff engagement. In particular, these systems have developed strategies for engaging patients with limited health literacy and/or basic computer/Internet skills, such as dedicated portal training programs. They also have recognized the need for parallel processes for provider/staff engagement.


Interactive Panels

This panel format creates an interactive event between the organizers and an active audience. Panels introduce topics that are either controversial or deemed to elicit varying opinions and questions, as well as active discussions.

Sunday, November 15, 2015

3:30 p.m. – 5:00 p.m.

S01: Open Architecture for Pathways and Care Coordination

M. Burton, K. Toussaint, Mayo Clinic; R. Greenes, Arizona State University/Mayo Clinic; O. Diaz, Healthcare Services Platform Consortium

Current health IT has evolved over many decades as a patchwork of proprietary legacy systems for managing electronic health records (EHRs), providing decision support, facilitating workflow, finances, and quality management, among other functionality. The closed, proprietary nature of the architecture and underlying technologies in these systems has resulted in silos of data, fragmentation of the patient records, difficulty coordinating care, and difficulty evolving to address the needs of the future. In order to better support the needs of clinical pathway development and care coordination, leading health care organizations should work together with leading open technology providers to create open architecture, open interface and open source community cloud computing environment. The presence of such an environment would enable a new generation of health information technology innovation resulting in information technology tools that would be better suited to meet the needs of a health care future where delivering best practice care to patients in every setting is the norm.

Monday, November 16, 2015

10:30 a.m. – 12:00 p.m.

S23: Opportunities for Social Media within Consumer Health Informatics

S. Myneni, University of Texas at Houston; A. Hartzler, GroupHealth Research Institute; L. Mamykina, Columbia University; N. Cobb, Georgetown University; L. Barnes, R. Valdez, University of Virginia, Charlottesville

Social media, including blogs, wikis, discussion forums, and social networking tools, has become an increasingly significant space for health-related interactions among patients and others engaged in their care (e.g., family, friends, peers, and health professionals). This panel will address critical questions surrounding current and future opportunities for social media within consumer health informatics. Panelists will discuss their experiences analyzing health-related interactions on social media and the resulting implications for consumer health informatics design and research: Dr. Hartzler will discuss ethical considerations; Dr. Myneni will discuss use of existing theories and methods for data analysis; Dr. Valdez will discuss novel recruitment and data collection methods; Dr. Mamykina will discuss design guidance drawn from observational learning; Dr. Cobb will discuss methods of outcome measurement and inherent challenges. The session will then serve as a forum for panel and audience engagement with opportunities at the social media and consumer health informatics intersection, including: innovative research designs, strategies for safeguarding user privacy, novel methods, and the role for health professionals and other stakeholder groups in social media interactions. This panel is timely given that the rapid growth of social media for personal health management has enabled new and evolving forms of patient engagement and stakeholder collaboration.

3:30 p.m. – 5:00 p.m.

S45: Building the Computational Workforce for Precision Medicine

J. Tenenbaum, Duke University; J. Denny, Vanderbilt University; D. Flannery, ACMG; D. Fridsma, AMIA; M. Williams, Geisinger Health System

Biomedical research and practice have never been more interdisciplinary than they are today, and yet interdisciplinary training in biomedicine remains the exception and not the rule. AMIA has a number of initiatives, from working groups to the annual Summit on Translational Bioinformatics, which emphasize the importance of the “bio” end of the biomedical informatics spectrum. However, there is significant opportunity for collaboration with other organizations with complementary expertise. AMIA has significant capabilities in informatics to support precision medicine. The American College of Medical Genetics and Genomics has significant expertise in medical genetics and genomics to improve human health. This panel brings together thought leaders from these two organizations to discuss the challenges in training tomorrow’s workforce for precision medicine, and the possibilities for interprofessional educational opportunities.

S47: Challenges, Successes, and Future Directions of Consumer Health IT Evaluation

U. Backonja, University of Washington; R. Valdez, University of Virginia; W. Riley, National Cancer Institute; K. Goodman, University of Miami; K. Siek, Indiana University; T. Zayas-Cabán, Agency for Healthcare Research and Quality

The accelerated growth and diversity of consumer health IT interventions creates an urgent need to develop evaluation criteria and methods to assess their effectiveness and efficacy. This panel will discuss challenges, successes, and future directions of evaluating consumer health IT interventions. Panelists will discuss their evaluation experiences: Dr. Riley will discuss accounting for the rapid pace of technological change; Dr. Siek will discuss accounting for patients’ workflow and daily living contexts; Dr. Zayas-Cabán will discuss accounting for intervention efficacy and safety; Dr. Goodman will discuss accounting for ethical considerations. The session will then serve as a forum for the panelists and audience to discuss topics related to consumer health IT evaluation, such as how to: modify existing knowledge and evaluation frameworks for consumer health IT evaluation; assess the safety and cost/benefit of the interventions; and balance evaluation rigor and richness with the need to efficiently evaluate rapidly changing technologies. This panel is timely given the rapid uptake of consumer health IT technologies, which necessitates the development of new and innovative methods to evaluate these technologies and interventions.

Tuesday, November 17, 2015

10:30 a.m. – 12:00 p.m.

S59: What Could Go Wrong?: Migrating from One EHR to Another

R. Schreiber, Holy Spirit Hospital/Geisinger Health System; R. Koppel, University of Pennsylvania, J. McGreevey, University of Pennsylvania; R. Schreiber, Holy Spirit Hospital; C. Craven, University of Missouri

Many healthcare institutions face a second crisis in electronic health record system (EHR) implementation: migrating from one mature commercial or home-grown EHR to another. A considerable portion of the healthcare market is now converting to a second or third EHR system. AMIA members, ostensibly, have examined in great detail the challenges of initial EHR selection, implementation, adoption, and the shift away from paper records. Research is missing or sparse to guide informaticists and information technology specialists on the challenges and support for subsequent implementations.  This panel seeks to accomplish the following: Recognize that one important implementation issue is one of migrating from one mature EHR to another; Explore the advantages of such shifts; Discuss extant evidence-based literature on system migration, or “re-implementation”; Identify the conversion risks and costs; Distinguish often unforeseen consequences; Guide others in this common endeavor.  Panelists will explore what is known via anecdotal sources, listserv discussions, personal experience, and the relatively few published studies related to this topic. We shall solicit information from fellow AMIA members and present new information. Panelists will declare their positions on these topics, followed by questions proffered to each other in an attempt to stimulate audience participation in the open discussion period.

S60: Patient Privacy and "De-identified Health Records in the Genomic Era

J. Tenenbaum, Duke University; G. Biggers, Genomera, B. Malin, Vanderbilt University; L. Ohno-Machado, University of California, San Diego; G. Biggers, Genetic Alliance, L. Wolf, Georgia State University College of Law

Current policy allows researchers to use de-identified data from electronic health records for research purposes. Data is considered de-identified if eighteen identifying fields, including name, address, phone number, social security number, etc. are stripped off of the derived dataset. However, in the era of precision medicine, detailed genomic data will increasingly be stored along with patients’ clinical data. Genomic data, particularly whole genome/exome sequencing is, by definition, identifying. The availability of genomic data together with clinical information represents a rich resource for science, but also introduces ethical challenges around patient privacy. How can we manage the research demand for data while respecting patients' concerns, interests, and rights? On what “truths” does everyone agree today, and what are the potential sticking points down the road? What can we, as informatics researchers, do to help? These are just some of the timely and compelling issues that will be addressed by this diverse panel, featuring technical, ethical/legal, community, and patient’s rights perspectives.

1:45 p.m. – 3:15 p.m.

S72: Health Informatics Graduate Program Accreditation: CAHIIM Process and Standards Update

J. Warren, Warren Associates, LLC; S. Johnson, Cornell University; S. Boren, University of Missouri School of Medicine; S. Speedie, University of Minnesota; G. Tusch, Grand Valley State University

Academic program accreditation in higher education is both a process and a status involving a review of a program against formal Standards to determine educational quality, how well students and society are served, and if successful the award of accreditation. The Commission on Accreditation for Health Informatics and Information Management Education (CAHIIM) has offered accreditation at the masters’ degree level in Health Informatics since 2010. AMIA officially joined CAHIIM as a Member organization in January 1, 2015. A revision in the governance structure created a Health Informatics Accreditation Council with the responsibility of establishing and maintaining accreditation Standards, policies and procedures. This Panel will present the Council’s work to date and address questions about how CAHIIM accreditation is conducted, the value of accreditation to the health informatics field, the status of goals for 2015-2016, how the process is conducted, and the types of CAHIIM volunteer opportunities. Participants will have the opportunity to ask questions and provide input on the Council’s initiatives from their varied perspectives.

S73: The Informatics Sculptor & the Clinical Annotator: Effective Annotation Strategies

R. Reeves, N. Gentry, E. Hanchrow, G. Gobbel, Tennessee Valley Healthcare System/Depatment of Veterans Affairs/Vanderbilt University; B. South, University of Utah/Salt Lake City VA Medical Center; S. Bradley, Eastern Colorado Health Care System VA/University of Colorado

Annotation of free text is an operation requiring human labor, complex cognition and topic-specific knowledge, requiring multiple skills sets in order to be effectively performed. This panel organized around the shared perspective that cross-education of research team members in the disciplines involved is a necessary and desirable function of informatics research. Curating clinical data thru analytic processes including annotation depends on multiple levels of data transform, each requiring expertise from the participating roles involved in the project. This panel discussion includes the following roles: 1) annotator, 2) data quality organizer, 3) NLP developer, 4) clinical expert study designer. Panel members representing one of the roles in integrating annotation data into the research project will discuss tasks and responsibilities of their own role within the process, converging on how their own professional perspective impacts and is impacted by the other roles in the project. The barriers to operationalizing annotation goals will be significantly lowered for attendees of this panel session. Those with an interest in initiating an NLP system for their research or in obtaining quality annotation for other purposes will gain an understanding of ways to best make use of each team member’s particular expertise to obtain quality annotations.

Wednesday, November 18, 2015

8:30 a.m. – 10:00 a.m.

S91: Patient Portals: Best Practices and New Directions for Development and Investigation

P. Dykes, S. Collins, A. Dalal, Brigham & Womens Hospital; R. Greysen, University of California, San Francisco; C. Dwyer, Johns Hopkins Hospital

The meaningful use legislation requirements for patient engagement have led to a greater use of personal health records or patient portals in outpatient settings. However, there are limited examples in the literature that describe strategies and e-health tools to provide patients with access to their health information in hospital or inpatient settings. This interactive panel will explore the findings from a recent study that identified inpatient portal best practices across four “early adopter” healthcare organizations in the United States. Session participants will be involved in refining and identifying new directions for the development and investigation of patient portals.

S92: (Authoring) Rules, (Distributed Query) Tools, and Drools: The Challenging New World of High Throughput Phenotyping

J. Pacheco, A. Kho, Northwestern University; J. Pathak, Mayo Clinic; J. Denny, Vanderbilt University; S. Murphy, Partners Healthcare

Deriving phenotypes from electronic health records (EHRs) is in progress within several national networks, including the electronic Medical Records and Genomics (eMERGE) networkthe Pharmacogenomics Research Network (PGRN), the SHARPn (Strategic Health IT Advanced Research Projects) and PCORnet, to name only a few, and those phenotype definitions are also now being shared across those networks. Two main obstacles make it more difficult to share those EHR derived phenotypes across sites, and especially networks: 1) the absence of clear, unambiguous phenotype algorithm definitions represented in a standardized, if not executable, format that can easily be shared, and 2) the use of different standard data representations, and varying data models. This panel will discuss the varying array of methodologies, informatics tools, and data standards and data models for sharing and executing EHR derived phenotypes. The panel will also discuss efforts to harmonize data models and phenotyping informatics tools across the multiple parallel national efforts mentioned here.

S94: Career Opportunities for the Many Paths to Informatics

L. Wiley, Vanderbilt University Medical Center; T. Kelley, Nightingale Apps; V. Lorenzi, Columbia University; V. Mohan, J. Doberne, Oregon Heath and Science University; J. Tenenbaum, Duke University

Informatics is a diverse field with an even more diverse workforce. A demand for informatics professionals continues to increase, the number and types of informatics training opportunities also increase. From traditional informatics training programs to cross-training in informatics and informatics certification programs, there are a number of ways into an informatics career. Following our successful tradition, the AMIA Student Working Group proposes a “career panel” to offer perspectives and advice for students on career opportunities and professional development. This year, the panelists include a board certified physician informatician, an informatics PhD working in academia, a cross-trained PhD-level informatician entrepreneur and a certified informatics professional with industry and academic experience. They will share their career and educational experiences and discuss upcoming trends in informatics careers. This panel will help future and current informatics students and early-career professionals to better prepare for and develop their careers.