No matter where you sit in the healthcare ecosystem, data integrations are complex and critical to success. Your product or service may speed up revenue cycles, help doctors and patients connect virtually or hold the cure for the plague of the 21st century, but if it can't easily share and consume data with other systems, you won't get very far.
Three trends that complicate health data sharing:
- Health standards are plentiful and evolving. Staying (or getting) up to speed so that you know the best format for each data acquisition and can remain flexible in building interfaces to adapt to variations in deployment can be a full-time job.
- There are a myriad of EHRs and practice management systems. Even a single health system may require you to connect with more than one EHR, each with different messaging formats guarding the data you need.
- The volume of data flowing through these systems is growing, even as digital transformation drives workflows outside of the EHR. In fact, a recent Gartner1 report predicts, "By 2023, 35% of healthcare delivery organizations will have shifted workflows outside the EHR to deliver better efficiency, experience, and outcomes."
To grow, creators of digital health solutions need a thoughtful data integration, or interoperability, strategy that empowers you to overcome these complications. Data exchange with trading partners such as hospitals, payers, public health agencies, and other systems is a required capability for any successful digital health product. While reliable integration can speed up your time to market, it's likely not a skillset that's part of the core competency of your development team.
As you grow, you'll want to consider the complexity of information sharing and if you are best positioned to buy, build, or partner to keep pace. To inform this decision, think about your business requirements and the complexity of each integration project.
Your business plan and product workflow dictate needs
There are many ways to make an integration happen. An important first step is to understand what data you need to surface, and where it should surface, so that it meets your business requirements.
Let's say you've created a telemedicine solution, and you need to integrate with hospital EHRs. You'll need to understand HL7 to interpret where in the EHR's interface message the data that you need is available. Similarly, you'll want to understand how the workflows you are part of interact with the EHR. These workflows can dictate the types of interfaces and standards you'll read from and write to.
There is no one-size-fits-all to digital health information sharing. Each product, each solution, and each idea mixes and matches different requirements. Following are eight critical steps that our teams have found must be fulfilled during every project. The timeline varies by organization.
8 critical steps to every integration
Successful and rapid patient data sharing is a dilemma that requires a plan. Whether you execute that plan with tools built for the job or rely on integration as a managed service, getting to go-live requires you to check every box. As you consider the right path for you, arm yourself with this 8-step project plan.
- Planning & building your integration layer. Assess your product or application to ensure it has features that allow for data exchange. If you don't already have an integration layer, you will need to build it or find a partner who can build it for you. Either way, this layer should only need to be built once, allowing the flexibility to handle integration with different systems you want to connect with now and in the future.
- Planning & paperwork. Initial steps include preparing contracts, signing of business associate agreements, and kick-off meetings.
- Gathering requirements. Data necessities — such as sample HL7 messages, sample JSON, associated data dictionaries, and the vendor's API with associated documentation — must be received.
- Defining infrastructure and secure connectivity, even when that means VPN access. You will need to make decisions about infrastructure deployment and secure connectivity so that those pieces are in place as a foundation for interface configuration.
- Configuring and coding each interface. Each organization's and vendor's endpoints are set up, as well as the interface and its mappings.
- Testing. Sample messages are sent, reviewed, and acknowledged by both inbound and outbound users. Any issues are identified and resolved prior to completed testing.
- Going live. All channels are turned on through migration to production. The organization data feeds are opened, and the vendor can begin to populate the system. The hospital can either insert this data feed into the EHR or another inbound interface.
- Ongoing support. Determine and implement support and alert plans.
Once you've finished your first hospital integration, you've finished your first integration. Each one is different.
For the reasons above, integration will continue to be a time-consuming job for your team. It is unrealistic to assume integrating once to an API will plug you into every hospital because many healthcare organizations don't have the technical expertise to connect to individual vendors' native APIs. Scaling integrations as you grow will be accelerated by giving your talent access to proven tools and industry experts that speak HL7, FHIR, and healthcare in general.
The solution to healthcare integration
Think of interoperability or integration as a journey. It's a process with multiple, interconnected steps. You will likely need help. Believing your team can gain clinical data integration domain expertise is about much more than just learning the technicalities of HL7 or FHIR. You must become domain experts throughout the full process or rely on a trusted interoperability partner.
How can you get started on your interoperability journey? Download 9 Questions to Ask Before Choosing an Interoperability Strategy and discover key questions that can help you land on a robust, secure, and scalable solution.
- Predicts 2021: Healthcare Providers Must Accelerate Digital Transformation to Address Disruption, Sharon Hakkennes, Barry Runyon, Mike Jones, Mark Gilbert, 25 November 2020