InterSystems Virtual Summit 2020: Day Three Focus Session Recap

Day One Summary	Day Two Summary	Day Three Summary

2020 marks the first year ever of InterSystems’ Global Summit going virtual! Join us from October 20th to November 5th as we recap various sessions of this year’s event.

---

Healthy Data in a Crisis: Manifest MedEx Uses HealthShare for COVID-19 Response

Summary by: Jen Carter, Executive of Marketing, J2 Interactive

When COVID-19 reached the United States, multiple communities in California were among the earliest hotspots.

In those early days, care teams were unsure how to identify COVID-19 patients and distinguish them from other coronavirus cases. Moreover, the systems they rely on didn’t yet have unambiguous diagnosis codes and lab codes for COVID-19. In the absence of defined standards, various provider organizations, labs, and EHR vendors scrambled to address this gap on their own. The result was messy, duplicative, and incomplete clinical data that made it extremely difficult to form an accurate public health assessment. The Golden State urgently needed data it could trust.

Manifest MedEx (MX), California’s largest health information network, was ready to answer the call. The MX exchange, powered by InterSystems HealthShare, combines clinical and payer data for 23 million patients across the State of California. With help from J2, MX was able to normalize COVID-19 diagnosis and testing data across its population and deliver timely reports to authorities at the state and county level.

The flexibility of the HealthShare platform was critical to MX’s ability to meet the challenges presented by the pandemic. First, HealthShare made it possible to map source-specific COVID-19 diagnosis and test codes to the standard ICD-10 and LOINC codes that eventually emerged. HealthShare also facilitated the translation of free-text test results into clear positives and negatives, as well as the disambiguation of antibody tests from antigen tests. MX was able to make this data available to internal and external stakeholders not only using the data visualization tools built into HealthShare Health Insight, but also with other reporting applications, such as Tableau.

In addition to enhanced reporting, the joint MX / J2 team implemented notifications to instantly alert providers whenever one of their patients tests positive for COVID-19. The team also extended the HealthShare Clinical Viewer to display a bright, visually distinctive banner for any patient with a positive result. This visual cue ensures that even in an acute care setting such as an ED or an ambulance, care teams will immediately know the patient’s testing status.

The role of HIEs in public health has been more critical than ever during COVID-19. Manifest MedEx was able to adapt quickly to the data needs of its community because of their investment in a robust, versatile technical infrastructure anchored by InterSystems HealthShare. MX’s commitment to healthy data made it possible to deliver life-saving insights to providers and public health officials when they were needed most.

---

Working with FHIR Profiles in IRIS for Health

Summary by: Erin Nielsen, Senior Consultant, J2 Interactive

In most cases, when you add an extension to a FHIR resource, you will also want to add a search parameter for that field. Without it, you won’t be able to search. To do this process, Dawn explained:

1. Start with a package in npm format. It can either be a package you get from the FHIR Package Registry or a custom one you create the JSON files for.
2. Import the package through ‘Package Configuration’ in the management portal.
3. Add additional packages to the FHIR endpoint in the management portal.
4. Manually re-index the endpoint from the terminal.

---

Using Alteryx with InterSystems IRIS to Craft Your Analytics

Summary by: Aaron Milberger, Principal Consultant, J2 Interactive

Alteryx develops software used to blend, prep, analyze, and deploy business intelligence (BI) solutions. It is used by analysts and data scientists, providing self-service tools used in analysis, discovery, visualization, modeling, and integration. InterSystems’ demonstrated the integration between Alteryx and InterSystems IRIS on a sample application used to assign the patient’s discharge disposition.

Alteryx collected encounter data from IRIS and merged it with a set of readmissions tracked in a different application. From the merge of these two data sets, Alteryx built a predictive model calculating the likelihood of a readmission.

The predictive model was then sent to InterSystems, where it was incorporated into the application, providing an estimate of the likelihood that a patient would need to be readmitted in the next 30 days. This estimation helped the hospital user select the appropriate location to discharge the patient.

All the integration and analytic work in Alteryx was performed without writing any code.

---

Extending Interoperability with Productions Extensions & Containers

Summary by: Alexis Tate, Senior Consultant, J2 Interactive

In IRIS 2020.3, InterSystems introduced the Production EXtension (PEX) framework. The PEX framework allows developers to use multiple languages to develop interoperability productions. Previously, InterSystems ObjectScript was the only language that could be used. Currently, in IRIS 2020.3, the PEX-supported languages are Java, .Net, and InterSystems ObjectScript, with Python availability next year. Developers can build productions entirely in one language or use a mix of the available languages to create the components. The Java and .Net production components are called at runtime using the PEX framework to communicate with other such components.

The presenter touched on the topic of deploying using Docker containers. Its benefit is that it allows for multi-stage builds, which enable the deployment of incremental updates.

---

Customer Innovations in a COVID Landscape

Summary by: Nathan Elliott, Senior Consultant, J2 Interactive

HBI and Healthix have been collaborating to put Healthix’s 253 million monthly inbound messages, 7 million monthly patient record updates, and over 9 million monthly clinical alerts to use to aid early detection of COVID-19 contributing risk factors. The partnership led to capitalizing on risk models that found connections between Kawasaki Disease and Multisymptom Inflammatory Syndrome (MIS-C) among pediatric patients.

HBI’s risk engine technology has been able to sift through Healthix’s inbound data to flag patients with possible early signs of Kawasaki Disease and MIS-C, which are now associated with higher mortality risk for patients who contract COVID-19. Risk models can be applied to other geographies but are typically customized at a local or regional level based on the type and quality of HIE-level data that is available for the algorithms to use. The development time for new models and subsequent near real-time notifications to care team members about high-risk patients can be completed in a matter of a few weeks.

Questions can be directed to Todd Rogow at trogow@healthix.org, Eric Widen at ewiden@hbisolutions.com, or Jonathan Teich at jteich@intersystems.com.

---

FHIR Implementation Patterns in IRIS for Health

Summary by: Erin Nielsen, Senior Consultant, J2 Interactive

The session outlined different ways that data could be brought into a FHIR repository, focusing on data coming from:

1. A C-CDA document
2. A FHIR resource submitted through a REST API
3. A JSON payload received from a device

In addition, data can be brought into the FHIR repository from HL7v2 or XML format.

An example production setup was discussed, including the routing rule and custom BPL being used. The FHIR resource data is populated in the repository and available via SQL. JDBC can be used to access the FHIR repository data and write reports off of it.

---

ONC & CMS Interoperability Rules

Summary by: Nathan Elliott, Senior Consultant, J2 Interactive

InterSystems provides support for customers with the following tools and releases:

• FHIR 4.0 released with InterSystems HealthShare 2020.1
• US Core FHIR resources are supported and will be conformant with 2020.2
• USCDI updates available as early as 2020.2
• CPCDS support for CARIN Implementation Guide as part of 2020.2/2021.1

Specifically, InterSystems is implementing the following for CMS:

• Patient Access APIs for Payers
• Provider Directory APIs (FHIR support in 2021.1)
• Payer-to-Payer Data Exchange
• Event Notification
• And for ONC, for certified modules, customers must agree to the Conditions and Maintenance of Certification.

Questions can be directed to Lynda Rowe at Lynda.Rowe@InterSystems.com, @lynda_rowe, or https://www.linkedin.com/in/lynda-rowe-509346/.

---

Healthcare Data Standard Designed for the Future

Summary by: Ayesha Lefebvre, Senior Consultant, J2 Interactive

In this presentation, Dr. Leftwich (@DocOnFHIR!) broke down how FHIR first ignited and its fundamentals. These include:

• Resources are discrete data concepts. All FHIR resources are defined in the specification. They exist in a known location on a server specified in a URL.
• Profiles are built for specific use cases. They consist of all resources needed for a use case. In addition, all the value sets that are needed are defined in that use case.
• Extensions can map data unique to a specific use case to an existing FHIR resource as an extension of that resource.

Did you know FHIR can be used in four different interoperability paradigms?
Dr. Leftwich explained:

• FHIR can represent patient information expressed as its own message – for example, when sending data to another system
• FHIR can be used in a document similar to CDA, where FHIR resources represent the data in the document
• FHIR can be a REST API
• FHIR can be used in services like Decision Support Service

Here’s an example of FHIR interoperability: Consider a lab result that is received from a REST API call as FHIR data, which can be placed in a message or document (without transforming) or sent to a DSS that would consume it and use that data in computing a decision. FHIR is machine-readable and can be consumed by those services. No other existing healthcare data standard has all these qualities working without transformations across interoperability paradigms.

Dr. Leftwich first gave a 20,000ft view, then drilled down to explain each concept. He stated:

“Consider a FHIR profile for a procedure, which contains multiple FHIR resources. This profile might be a lab test, an imaging procedure, or a surgery. All procedures have something in common: a person (patient resource) on whom the procedure is performed, a person (practitioner resource) performing the procedure, a clinical context (encounter resource) in which said procedure took place, a reason (condition resource) for the procedure, and for most procedures, there is a result (diagnostic resource).

So this is a 20,000 ft view—these details of that profile (procedure profile) have to exist. To specify these details, some concepts in FHIR, such as terminology bindings, are needed. [Terminology bindings] are when we create a value set for a specific element in a profile that are the values that element must have in an implementation of that profile. Those values come from code systems, referenced as a URL that is stored on a terminology server (which are part of the architecture of FHIR). [This storage] allows implementations to share those FHIR values/codes across regions and systems by placing them on a terminology server for access from each implementation system.”

Concluding remarks referenced the FHIR Implementation Guide, which contains all the rules specific to a certain use case that has been implemented and can be used for validation. While Dr. Leftwich mentioned no specific tools, HL7 has made some available, which you can read about in this post.

---

Day One Summary	Day Two Summary	Day Three Summary

Thank you for joining us for today’s recap. J2 has been a trusted InterSystems Implementation Partner since 2004. Learn more about us and our involvement in InterSystems Virtual Summit 2020.