Description
Project abstract (1000 Characters): The Pediatric Patient Summary (PPS) SMART on FHIR® app aims to help clinicians and families improve care and care coordination for children and youth with special health care needs. Coordinating care for the 11.2 million U.S, children with complex health care problems is a challenge. Optimal information sharing is critical for improving management and quality of life for these patients and for reducing costs. The PPS app provides an integrated display of patient clinical information obtained from multiple health care organizations and EHR systems through a state health information exchange (HIE) using HL7 FHIR and Clinical Document Architecture (CDA) standards. In its current version, the PPS provides summarized and reconciled information about patient’s problems, medications, encounters, and allergies. The PPS was developed by the University of Utah Departments of Pediatrics and Biomedical Informatics and deployed in a real care setting at the University of Utah Health in August 2019.
Project rationale, impact and innovation (3500 Characters): An estimated 15% of children and youth in the US (11.2 million) have special health care needs and require health related services beyond that required by children generally. These patients often receive care from multiple specialties, disciplines, and health care systems, leading to fragmented and voluminous health records. Poor care coordination can compromise the quality of that care due to delays in diagnosis and treatment, duplicative or unnecessary evaluations, and inappropriate, or potentially dangerous interventions. This can lead to reduced patient and physician satisfaction, and increased costs. Knowing the range of care a patient is receiving and coordinating with other providers presents challenges. Current EHRs do not provide sufficient capabilities to enable a patient’s care team to effectively gather and share information across multiple institutions on complex patient care.
EHRs and HIEs have dramatically increased the volume of information available but often do not provide clinicians with sufficient tools that help them to efficiently find, sort, and make sense of relevant information. Currently providers, particularly those in large health care markets with multiple delivery organizations, must search several screens in multiple clinical systems to locate needed information. Often, key information may resides in PDF-formatted files that are difficult to find and that require opening and visually scanning the record to know what they contain. Generic or insufficiently specific labeling of these PDF documents makes the data even more difficult to locate. Finally, the requirement for additional logins and long load times create delays that are unmanageable in busy care settings where visit times are scheduled every 15 minutes.
Sharing information among all clinical providers (e.g. primary care, subspecialists, mental health, therapists, emergency care personnel), parents or guardians, and other relevant parties (e.g., school nurses, daycare providers) is important to coordinating care and optimizing outcomes and may reduce the effort and time required of providers to prepare for visits, develop plans of care, refer patients to other health care providers, and provide consultations.
The EHR-integrated SMART on FHIR® PPS platform addresses information access challenges related to complex pediatric patients by providing a succinct, accurate, and up-to-date compilation of patient information from multiple providers and care settings. The PPS plots a historical timeline of the patient's encounters, problems, and medications to enable the patient’s care-team to gain a quick understanding of the patient’s situation at a glance and also includes structured tables with important details. This way of displaying both high-level overarching information along with highly focused details aims to support rapid sensemaking while at the same time allowing users to drill down to specifics – a design attribute consistent with Schneiderman’s “mantra”. In this way, the PPS app supports the dual cognitive processes for visit planning and medical decision making.
Project design and implementation (7000 characters): Overall, we followed an iterative, user-centered design approach that included (i) interviews and observations with pediatricians, care coordinators, and parents of children with special health care needs; (ii) iterative design of low-fidelity prototypes that evolved into high-fidelity prototypes based on target user feedback; and (iii) usability heuristic evaluation. Finally, we integrated the design work with the PPS architecture using a SMART on FHIR® client and backend services as described below.
User needs analysis
To inform the design of the PPS we started with a needs analysis by eliciting the information needs and goals of primary care providers, care coordinators, and parents in the course of caring children and youth with special health care needs. A key information goal was the desire to understand a patient’s situation, including their diagnoses, treatment plan, recent encounters, social context, and care team.
User requirements study
Next, a requirements study was conducted to identify the specific elements to include in the PPS display and to inform the design. Participants were asked to demonstrate in their EHR how they prepared for a patient’s visit and insights related to improving the EHR to support their needs were elicited. The study revealed patterns of information retrieval and a reliance on the problem list, medication list, and scanning of encounter notes to get a sense of the patient’s health care needs, care team, and recent health encounters. Difficulty with accessing information from outside sources was observed due to the need for multiple logins, lack of meaningful encounter-note labels, almost meaningless organization of documents, as well as difficulty with reconciliation of long medication lists that were not organized by ingredient and lacked adequate filtering methods.
Mockups
We mocked up the initial draft of the PPS with a health event timeline as the first feature. This strategy was based on our observations and the finding by Weir et al. that clinical system designers should support pattern-matching by integrating information into single views that easily capture the gist of a situation. We also included the patient’s problem list, medications, allergies and encounter notes all in one page. In addition, we included a link to the patient’s growth chart and infobutton links to external resources providing care guidelines, practice tools, and patient education resources specific to the patient’s problems and medications, as well as relevant local services.
Heuristic usability evaluation
A heuristic evaluation was conducted by members of the University of Utah’s Socio-Technical Service Line. Each team member individually evaluated the system based on Nielsen’s 10 design heuristics. The results were compiled and the design team met to review the results and identify feasible solutions to identified problems, including misalignment of page components, difficulty with scroll features, hidden nesting, small font size, linking to inactive websites, and the need for improved labeling. Problems identified were either addressed or verified by the user evaluation described below.
Technology Integration
The PPS architecture consists of a SMART on FHIR client and backend services that retrieve FHIR resources from EHR FHIR servers and CDA documents from a state-wide HIE. For the implementation at the University of Utah, we used Epic’s FHIR API to retrieve FHIR resources for data locally stored in Epic; and Epic’s Care Everywhere connection to Utah’s HIE (Utah Health Information Network [UHIN]) to retrieve CDA documents from other health care organizations throughout the state. Once retrieved, CDA data are converted into FHIR resources using the open source cda2fhir package ( https://github.com/srdc/cda2fhir). Next, local and external data are normalized (e.g., mapped to common terminologies, such as RxNorm for medications and SNOMED for problems), ordered [RCC2] and grouped as a cohesive view of the patient data across the differing data providers. Last, the reconciled data are rendered for display in the PPS app. The PPS can be launched on demand from the EHR activity menu. The SMART protocol is used to provide transparent, single sign-on authentication and authentication token. The current implementation uses the Epic® EHR FHIR API.
Project evaluation and sustainability (3500 characters): User evaluation
An early iteration of the PPS was user tested with 3 pediatricians for further refinements. Users were asked to open the PPS inside a patient’s chart and think aloud as they addressed the following items: 1) What do you see? 2) What can you do? 3) Anything that is helpful. 4) Anything that is confusing or that you would change. Participants indicated the timeline supported discovery of important information and patterns that they would not have noticed. Inclusion of the problem specific overview with the problem list was also described as valuable. Sorting of medication by ingredient was thought to be an improvement over the name sorting that the users regularly used.
A mixed-methods post implementation evaluation is underway with pilot user participants. The evaluation consists of 5 components, 1) User rating of PPS unique features, 2) Observation with retrospective think aloud in which clinicians prepare for an upcoming patient visit using the PPS, 3) Structured interview 4) Usability questionnaire evaluating effectiveness, efficiency & satisfaction, and (5) Usage data analysis.
Dissemination and financial sustainability
Sustainability of the PPS app is dependent on factors such as overall HIE use, interoperability with multiple EHRs, perceived value among clinicians and patients, dissemination, and an approach to obtaining continuous financial support. Several channels will be leveraged to help disseminate the PPS to other health care organizations in Utah and nationwide. We will collaborate closely with the Health Services Platform Consortium (HSPC), which endorses the SMART on FHIR® approach proposed in the current project. One of their short-term goals is to establish a vendor-neutral Health care App Store. Intermountain Healthcare is one of the founders of HSPC and Intermountain employees who are members of the current project have leadership roles in HSPC. The PPS is available in the HSPC sandbox and will be made available within the HSPC Health care App Store. Additional dissemination activities include engaging with the Epic® and Cerner® user groups and demonstrating the PPS/Portal at national conferences and webinars. Financial sustainability models will be explored with assistance from the University of Utah Technology Venture and Commercialization (TVC) office. The TVC ranks 1st in the US in technology commercialization among academic institutions and have experience with the commercialization of health IT solutions, including services over open-source platforms.
Twitter project summary (140 characters): The Pediatric Patient Summary SMART on FHIR® App integrates disparate data to support patient-centered care for children with complex needs
How is FHIR used in the App being demonstrated (500 characters)? : The app uses FHIR® resources available in Epic® 2018 from two sources, CareEverywhere and the main Epic FHIR database. The architecture consists of a SMART on FHIR client and backend services that retrieve FHIR resources from EHR FHIR servers and CDA documents from a state-wide HIE. Once retrieved, CDA data are converted into FHIR resources using the open source cda2fhir package (https://github.com/srdc/cda2fhir). The SMART protocol is used to provide tokenized single sign-on authentication.
1. What FHIR release does your application use? (500 characters)?: FHIR® STU 2
What is the data source for the FHIR resources and how are the FHIR resources accessed? (500 characters): FHIR® bundles are retrieved from Epic’s® FHIR API for local data. For external data, Epic’s Care Everywhere connection to Utah’s HIE (Utah Health Information Network [UHIN]) is used to retrive CDA documents from the Utah state HIE, which are then converted in to FHIR bundles using the open source cda2fhir package (https://github.com/srdc/cda2fhir). We retrieve FHIR resource bundles with the following resources: AllergyIntolerance, Condition, Encounter, Patient, Observation, and MedicationOrder.
Any other information about the project we should know about (1500 characters)?: Meaningful Use
The PPS aligns with selected Meaningful Use (MU) Stages 2 and 3 goals as follows:
MU Stage 2: Advance Clinical Processes
Electronic transmission of patient care summaries across multiple settings
Support care coordination among clinicians and other providers of care throughout patients’ communities
Provide encounter alerting to medical home providers
MU Stage 3: Improved Outcomes
Improve quality, safety and efficiency leading to improved health outcomes
Leverage query-based information exchange to support development and updating of focused patient summaries
Increase knowledge and use of local health-related service providers
Provide a platform for collaborative, shared care planning and coordination
Decision support for national high priority conditions
Support clinical decision-making with the summaries and with information to guide clinical management and referral, focused on patients with chronic conditions, particularly those with limited access to subspecialists.
Future work. Planned iterations will include:
Care team viewer
Care plan management
Parent-facing version of the PPS, including the ability to view and annotate patient’s data, and contribute to the patient’s care plan.
Authors:
Teresa Taft (Presenter)
University of Utah
Guilherme Del Fiol, University of Utah
Damian Borbolla, University of Utah
Ryan Cornia, University of Utah
Kensaku Kawamoto, University of Utah
Charlene Weir, University of Utah
Scott Narus, Intermountain Health Care
Mindy Tueller, University of Utah
David Shields, University of Utah
Andrew Iskander, University of Utah
Philip Warner, University of Utah
Chuck Norlin, University of Utah