iHealth 2016 Presentation Sessions

The clinical informatics solutions are team-based operations. Where a team is listed, the presenter is noted. iHealth 2016 presentations cover the following topics:

Thursday, May 5

S03: Presentations – Usability

A Novel Process for Electronic Health Record (EHR) Adoption which Challenges Traditional Adoption and Support Paradigms. 

Joel Gordon, Mayo Clinic 

Traditional methods of EHR post-go-live support have focused on classroom events, email updates, non-structured elbow-to-elbow, and eDemo tools. These tools do not gauge impact, have limited data on individual effectiveness and do not have proper incentives for participation. We have developed and use a novel user focused evaluation tool to assess for EHR best practices, re-train individuals to these best practices, gather outcomes on effectiveness and incent providers to participate. The outcomes are shared here.

Factors Associated with Clinician Stress and Burnout: Results from the MS-Squared Study (Minimizing Stress, Maximizing Success from Health Information and Communications Technologies)

Philip Kroth, University of New Mexico (Presenter)
Nancy Morioka-Douglas, Stanford University
Sharry Veres, Centura Health
Sara Poplau, Minneapolis Medical Research Foundation
Katherine Pollock, Cabrini Partnership Programs
Mark Linzer, Hennepin County Medical Center 

We report the results of the MS-Squared Study (Minimizing Stress, Maximizing Success from Health Information and Communications Technologies) where we conducted 2 clinician focus groups at each of 3 institutions with three different EHR vendor systems. Our goal was to assess the impact modern health information and communications technologies have had on clinicians' work stress, job satisfaction, and burnout tendency.

The Role of the Patient and Community in Health Information Technology Success 

Theresa Cullen, Regenstrief Institute, Indiana University 

This session will highlight two successful HIT open source projects (OpenMRS and RPMS, the Indian Health Service HIT solution) that have a novel approach to development: the integration of patient/community members in the HIT product lifecycle. From the inception of ideas /requirements for functionality through product delivery and deployment, community members are essential to ensuring that software can be used effectively and efficiently to improve health outcomes for individual patients as well as communities.

S05: Presentations – Care Coordination

The HIMSS Transition of Care Framework

Michael Victorff, University of Colorado School of Medicine, COPIC, Inc.

This program discusses the Transitions of Care Framework (2015), developed by the TOC Work Group of the Healthcare Information and Management Systems Society (HIMSS). To our knowledge, this is the first, general tool to offer a unified analytic perspective on TOC applicable to all healthcare settings and circumstances. It enables planners and evaluators a format for sharing insights and methods and comparing outcomes across disparate settings.

The Longitudinal Care Plan Cycle: Understanding the Interface Between Patients, Caregivers, and Healthcare Teams in Chronic Disease Care

Kim Unertl, Vanderbilt University (Presenter)
Christopher Simpson, Vanderbilt University
Laurie Nova, Vanderbilt University

Managing a chronic disease is a complex, long-term challenge shared by patients, informal caregivers, and healthcare providers. While providers make treatment recommendations, patients and caregivers often have responsibility for implementing those recommendations. Through a multi-year research study focused on chronic disease care in academic and community settings, we identified a six-stage Longitudinal Care Plan Cycle. This model can assist in developing patient-centered tools and processes for chronic disease management, both in healthcare and home/community settings.

S07: Presentations – Quality Improvement

Clinical Documentation Improvement (CDI) Initiatives That Drive Quality Performance and Financial Reimbursement

Angie Curry, CoxHealth

Clinical documentation is at the core of every health care encounter. A Clinical Documentation Improvement (CDI) program helps providers get to the specificity and accuracy needed to support coding activities, while the patient is still inpatient.

The audience will gain a better understanding of a successful CDI program, the use of strategies and automated tools that help improve documentation, and how quality documentation impacts Quality Core Measures reporting and mitigates revenue risk for their organization.

Adherence to Sinusitis Treatment Guidelines in the evisit Setting: A Quality Improvement Project 

Kevin Smith, Zipnosis

The purpose of this quality improvement project was to evaluate the effect of selected interventions for improving adherence to a clinical practice guideline for the management of acute bacterial rhinosinusitis (ABRS) in the evisit (electronic visit) setting. A pre and post-intervention evaluation design was used. Review of the structured clinical data extract demonstrated a 3.3% improvement in adherence to the ABRS clinical guideline from 95.25% adherence pre intervention to 98.4% post intervention.

S08: Presentations – Data Sciences 1

IBM Watson Analytics Academic Program

Robert Hoyt, University of West Florida (Presenter)
Dallas Snider, University of West Florida

Data analytics requires a substantial background in statistics and computer science. Multiple studies, however, have shown that most physicians lack data analytical expertise. This is likely because many clinicians did not receive an adequate experience in statistics during training. In 2015 IBM released Watson Analytics that delivers advanced statistics based on the Statistical Package for the Social Sciences (SPSS). The new entry of Watson Analytics into the field of analytical software products provides users with enhanced functions, not available in many existing programs. For example, Watson Analytics automatically analyzes datasets and determines data quality and the optimal statistical approach. Users can request exploratory, predictive and/or visual analytics. Using natural language processing (NLP), users are able to submit additional questions for analyses in a quick response format. This analytical package is available free to academic institutions (faculty and students) that plan to use the tools for non-commercial purposes.

In this presentation we will discuss the features of this analytical package and compare it to other analytical platforms using validated healthcare datasets. A real world application demonstration will also be part of the presentation. In addition, we will discuss how IBM Watson Analytics is used at our university to help teach data analytics in Health Informatics and Data Mining courses.

Applying NLP to Identify Complex Care Patients

Andrew Placona, Evolent Health (Presenter)
Shantanu Phatakwala, Evolent Health

This presentation examines the incremental value add of NLP in population health space for the identification of patients to care management programs. Utilizing repeatable Data Science processes, it is possible to integrate unstructured data using natural language techniques. In doing so it is possible to create a structured dataset incorporating these features that is more accurate Precision Medicine.

S09: Presentations – Workflow

The Relationship of Provider Typing Speed to Adoption and Use of a Structured Documentation Tool

Joel Gordon, Mayo Clinic

Providers are having increased pressure to utilize self-entry documentation (SED) through vended EHR products1. These tools are exciting as they offer assistance and immediacy compared to traditional dictation/transcription documentation systems2. Similarly, there is voice recognition tool evolution occurring3. We have measured typing speeds of providers from a multi-clinic ambulatory primary care practice. Our results compare different typing speed measurements and evaluated correlation with willingness to and level of migration to the novel self-entry documentation tool. 

Analytics for Optimization of in-Hospital Workflows: An Application in Rehabilitation Therapy Service

Krishnaj Gourab, Johns Hopkins Medicine (Presenter)
Christina Kokorelis, Johns Hopkins Medicine
John Adamovich, Johns Hopkins Medicine
Albert Mears, Johns Hopkins Medicine
Barbara Rucizka, Johns Hopkins Medicine

Analytics to frontline clinicians can facilitate in prioritizing resources for patients who would yield the most clinical and operational value for a clinical intervention. We will present improved process/operations outcomes (decreased number of "high priority" patients being missed by rehabilitation therapy services (RTS), time saving for clinicians, improved bed utilization in the hospital's inpatient rehabilitation unit) and demonstrate the current iteration of the analytics tool that was implemented for RTS in an acute care hospital.

S11: Presentations – Data Sciences 2

Applying Data Science to Electronic Health Records

William Feaster, CHOC Children's Hospital, Orange, CA (Presenter)
Louis Ehwerhemuepha, CHOC Children's Hospital, Orange, CA

In this presentation, we will discuss our findings from selected pilot projects undertaken by the Information Systems Department at CHOC Children's (Orange, CA) to understand 30-day readmissions, clinical factors affecting patient satisfaction and hospital acquired venous thromboembolism. Furthermore, we will discuss lessons learned from the interplay between our medical experts, system analysts, and data scientists/analysts.

Predicting Change in Glycemic Control among Previously Controlled Type 2 Diabetes Patients: A Machine Learning Approach

Che Ngufor, Mayo Clinic

Accurate and timely early warning of future changes in Glycated haemoglobin (HbA1c) over time and in response to treatment is important for early diagnosis of diabetes and for monitoring the disease progression. In this study, we propose new machine learning techniques that incorporate random effects to predict changes in HbA1c using large longitudinal data of patients with previously controlled type 2 diabetes.

Friday, May 6

S12: AMIA Public Policy Talk - The Obama Administration and the 114th Congress: How will Washington End a Busy Chapter of Health Informatics Policymaking?

Doug Fridsma, AMIA
Paul Fu, Harbor-UCLA
Jeffery Smith, AMIA

As the Obama administration and the 114th Congress come to a close, regulators and legislators have limited time to ensure their policy ideas are solidified in the Federal Register or as Public Law. Come learn from AMIA experts about which policies and political issues will drive the health informatics policy agenda through the close of the year.

S14: Presentations – Mobile Technology

Evaluation of Mobile Continuous Vital Sign Surveillance on Clinical Workflow

Rosemary Kennedy, Sotera Wireless (Presenter)
Ida Androwich, Loyola University

Studies confirm that patient deterioration is preceded by periods of vital sign (VS) instability. This has increased the use of mobile continuous VS monitoring. There is a paucity of literature demonstrating impact of such technology on workflow. This study compared manual VS collection with continuous VS monitoring in non-critical care. A savings of 9.2 nursing hours per day was achieved using continuous VS monitoring. The time savings were analyzed within context of the nursing process.

The Need for a Mobile Medical Photography Standard

Oliver Aalami, Stanford University

Photos are powerful and tell a story. Experts estimate that there will be more digital photos taken in 2015 by mobile devices than ever taken on film. More and more clinical photos are also being taken of physical findings such as wounds and rashes and are often shared amongst the care team. The efficiency in data capture, and interpretation is unparalleled. The problem is that these photos are not medical grade and are rarely integrated.

Introducing iPET: Interactive Patient Engagement Technologies

Perry Gee, Dignity Health
Frances Patmon, Dignity Health (Presenter)
Tina Rylee, Dignity Health

Hospitals and health systems are implementing technologies for patients and families who are interacting with acute care environments. Our team is calling these systems Interactive Patient Engagement Technologies or "iPET." The iPET systems have a wide range of functionality but overall promote patient engagement and education. We will define types of iPET and the initial ethnographic studies our team has conducted to better understand how these systems are being used by patients and nursing staff.

S15: Presentation – Data and Network Security

Healthcare – the New Target in Cybercrime

JoEllen Frain, Mayo Clinic

From information breaches to phishing scams, cyber-attacks targeting patient data are getting more sophisticated by the day. Understand the who behind the threat, what they are after and what role we as individuals play in protecting our data and ourselves from this criminal activity.

S16: Presentations - Population Health

Information Visualization for Population Health Management via Cognitively-guided Disease Risk Assessment

Rema Padman, Carnegie Melon University

Increased availability of patient data in electronic databases presents an opportunity to develop substantial cognitively-guided predictive analytics to improve chronic disease risk assessment for population and individual health management. This study explores the evaluation of a methodology and tool that offers an interactive visual interface for clinicians to access, explore and compare risk predictions at the population, individual and intervention levels for patients with diabetes and its complications in the context of many risk factors.

Use of a Local Learning Healthcare System to Improve the Health of Patient and Community

Amy Sitapati, University of California San Diego (Presenter)
Barbara Berkovich, University of California San Diego

Learning Healthcare System and Patient Centered Medical Home (PCMH) principles were drivers of the University of California San Diego Health (UCSDH) improvement of the clinic level HIV viral load and nine national quality measures. Here, we describe a case study of HIV care delivery, and the population health tools that made this possible. 

Development of a Health Registry Ontology

Barbara Berkovich, University of California San Diego (Presenter)
Amy Sitapati, University of California San Diego

Health registries are important tools in the organization and aggregation of health data, and are a foundational element of the Learning Health System. In 2015, the terms, "registry" or "registries" were found in 47 articles in the Journal of the American Medical Informatics Association (JAMIA). However the content, curation, and use cases of the referenced registries vary widely. The Health Registry Ontology will contribute to the literature by defining key classes and properties.

S18: Presentations – Clinical Decision Support

Incorporating Medication Indication into CPOE: What Do We Need to Build?

Gordon Schiff, Brigham and Women's Hospital (Presenter)
Lynn Volk, Partners HealthCare
Neri Pamela, Partners HealthCare
Aaron Nathan, Brigham and Women's Hospital
Kevin Kron, Partners HealthCare
Mary Amato, Massachusetts College of Pharmacy and Health Sciences University
Alejandra Salazar, Brigham and Women's Hospital
Enrique Seoane-Vasquez, Brigham and Women's Hospital
Tewodros Eguale, Brigham and Women's Hospital
Sarah McCord, Brigham and Women's Hospital
Rosa Rodriquez-Monguio, UMass Amherst
Adam Wright, Brigham and Women's Hospital

Currently, medication orders lack information about the drug indication. There are compelling reasons and broad support for incorporating indications into CPOE, however, there are numerous complexities in developing such systems. System design considerations will be presented that have been identified through working with six stakeholder webinar panel meetings that included representation from physicians, pharmacists, patients, EHR and knowledge compendium vendors, HIT and human factors researchers, and patient safety experts.

Dynamic Predictive Analytics to Prevent Chemotherapy-Induced Nausea and Vomiting

Abu Mosa, University of Missouri (Presenter)
Illhoi Yoo, University of Missouri
A. Mosharraf Hossain, University of Missouri

Chemotherapy-induced nausea and vomiting (CINV) is one of the most dreadful and unpleasant side-effect of chemotherapy. The consequences of CINV include impaired life-quality, poor social life, increased healthcare cost, and denial of chemotherapy due to unendurable CINV. In this project, we have developed a novel and dynamic CINV prediction engine using patient-specific factors. The prediction performance of the system outperformed many popular prediction algorithms and all the CINV prediction studies published in the literature.

Efficacy of Clinical Alerts Designed to Reduce Inappropriate Urinary Cultures in Catheterized Patients in the ICU

Mark Parkulo, Mayo Clinic (Presenter)
Launia White, Mayo Clinic
Colleen Thomas, Mayo Clinic
March Rucci, Mayo Clinic
Walter Hellinger, Mayo Clinic
Tara Creech, Mayo Clinic
Daniel Daughtry, Mayo Clinic
Darlene Care, Mayo Clinic

Catheter Associated Urinary Tract Infection (CAUTI) is a common hospital acquired complication in Intensive Care Unit (ICU) patients. It is also a publicly reported quality measure. Many patients in the ICU will develop a fever. As an evaluation for that fever providers often order multiple cultures including urinary cultures in patients who currently have an indwelling urinary catheter. The Critical Care and Infectious Disease Societies of America recommend against culturing these patients except in specific circumstances. Often these cultures will grow organisms that reflect colonization, not infection, resulting in misdiagnosis, inappropriate antibiotic treatment, and a mislabeling of patients as CAUTI patients for quality reporting. In July 2014 we implemented a provider alert designed to educate providers about the evidence based indications for urinary cultures in catheterized patients as well as providing them with the opportunity within the alert to continue or withdraw the order. A pre- and post-alert comparison revealed a statistically significant decrease in urinary cultures in catheterized ICU patients (19.3% vs. 14.3%) (p-value 0.0005). We also noted a substantial improvement in our CAUTI quality reporting following implementation of the alert.