EHR Integration 101: What to Look for When Choosing Speech Recognition Software
Medical speech recognition should be sensitive and advanced enough to capture what clinicians need to document, and versatile enough to work across the multiple software applications they use. For many healthcare organizations, finding speech recognition software that meets those demands is difficult. Organizations often use multiple electronic health record (EHR) and other clinical applications. A clinician may work in an EHR, a specialty application, a browser-based tool, a virtual desktop environment, or several systems during the same day.
Because clinicians work across multiple applications, compatibility is a critical consideration when evaluating speech recognition software. It is not enough for a vendor to simply claim integration with Epic or another major EHR. Healthcare organizations should evaluate how the software interacts with the applications clinicians use, which features and workflows are supported in each environment, and whether additional interfaces, configuration, or development are required.
Fusion Narrate is Dolbey’s cloud-based real-time speech recognition solution, designed to work across the applications clinicians use throughout the documentation workflow. It has been tested with more than 100 applications, including EHRs, LIS, RIS/PACS, and browser and desktop applications.
What Does EHR Integration Mean for Speech Recognition?
Integration can vary significantly by speech recognition product, application, and clinical workflow. Products may support different integration approaches, and more than one approach may be used within the same environment depending on the application and functionality required. Some common forms of integration that clinicians should consider when comparing speech recognition software include:
Cursor-based text entry. Recognized text is entered at the cursor in supported text fields. This can allow speech recognition to work across multiple applications without a dedicated interface for each one.
Application-specific integration. The speech product connects through an API or other application-specific interface. This can support deeper workflow integration or exchange of information with the EHR, but the connection is generally tied to that specific application and must be configured and maintained for the applicable environment. Clinicians who also work in other applications may need separate integrations or may not be able to carry the same speech-enabled workflow across those systems.
EHR-embedded tools. Some speech capabilities are delivered within a particular EHR environment. Organizations should determine whether the same functionality is available when clinicians work outside that EHR.
For real-time, or front-end, speech recognition, the basic workflow is straightforward: the clinician dictates and sees the recognized text appear immediately in the application. The clinician can review and edit the documentation while working rather than waiting for a transcription workflow to be completed.
But text entry is only one part of the workflow. Depending on the product, clinicians may also use voice commands, templates, shortcuts, or automation to move through applications and complete repetitive tasks.
Look Beyond the Primary EHR
Healthcare organizations often begin by asking whether a speech recognition solution is compatible with their primary EHR. While that is an important consideration, it does not capture the full picture of application compatibility. Clinicians frequently move between multiple applications within a single documentation workflow, including:
- Laboratory information systems
- Radiology information systems and PACS
- Specialty clinical applications
- Web-based portals
- Microsoft Office applications
- Remote or virtual desktop environments
- Legacy applications that remain part of the clinical workflow
For that reason, an organization should evaluate the applications its clinicians actually use rather than relying only on a generic list of supported EHR systems.
Fusion Narrate takes an application-agnostic approach. Recognized text can be placed at the cursor in supported text fields, allowing the same speech profile to be used across supported applications rather than limiting speech recognition to a single EHR. Fusion Narrate has been tested with more than 100 applications across healthcare and general business workflows.
Questions to Ask About EHR and Application Compatibility
A product demonstration should include the applications and workflows your clinicians use every day. Asking specific questions can help distinguish basic speech-to-text functionality from broader workflow support.
- Where can recognized text be inserted? Is speech recognition limited to particular EHR fields, or can clinicians dictate wherever text entry is supported?
- Does the product require an API or custom interface for each application? If so, determine what is involved in adding or updating those interfaces.
- What happens when clinicians move between applications? Can they use the same speech profile, vocabulary, commands, and preferences?
- How does the software work in Citrix, Remote Desktop, or other virtual environments?
- What voice commands or workflow automation are available? Can users create shortcuts for repetitive tasks or multi-step workflows?
- Can administrators configure and distribute workflows centrally?
- What happens when an EHR or other application is upgraded? Ask what testing or configuration may be required.
- Are there additional charges for interfaces, configuration, or support for additional applications?
- Which applications outside the EHR has the vendor tested? Ask to see the applications and workflows that matter to your organization during a live demonstration.
The answers can have a significant effect on both usability and total deployment effort.
Speech Recognition Should Fit the Workflow
Traditional speech recognition is sometimes evaluated primarily on recognition accuracy. While accuracy is important, the surrounding workflow of the speech recognition software is also critical.
Consider a clinician who dictates a note but still has to repeatedly use the mouse and keyboard to navigate screens, insert standard text, open templates, or move information between applications. Speech recognition has reduced typing, but it has not necessarily addressed the rest of the documentation workflow.
Fusion Narrate combines real-time speech recognition with configurable workflow tools. Shortcut Builder, Quick Keys, and the patented Vision Click feature can automate single- or multi-step actions within or across applications. These tools can support workflows that extend beyond entering text into the EHR.
That distinction can matter in specialty workflows where clinicians move between several applications during the same reporting process.
Consider the Full Range of Documentation Workflows
Healthcare organizations do not all document the same way. Even within the same organization, different departments may need different things from the software used by their organization.
Real-time speech recognition converts dictation to text as the clinician speaks, allowing the clinician to review and edit the documentation immediately.
Back-end dictation/transcription allows the clinician to record dictation for subsequent transcription or processing. Fusion Narrate can be paired with an optional back-end transcription module that supports traditional transcription workflows, including workflows with speech recognition and correctionist review as well as transcription-only workflows. This allows organizations that need both front-end and back-end workflows to support them within the same platform, so different users and departments can work within a common documentation environment and administrative controls.
Ambient documentation provides another option. Instead of dictating a note, the clinician conducts the patient encounter naturally while the ambient solution captures the conversation and generates a draft report for clinician review and approval. When evaluating a speech recognition vendor, organizations should consider whether ambient documentation is available as part of the same product family or requires a separate vendor and workflow.
The ability to support real-time speech recognition, back-end transcription, and ambient documentation can be particularly valuable for larger organizations. Different clinicians, specialties, and care settings can choose the documentation method that best fits their workflow while the organization maintains a more consistent technology platform and common administrative controls.
Fusion Narrate supports real-time speech recognition and can be paired with Fusion Narrate Text for back-end transcription workflows, while Fusion Narrate ACI provides ambient clinical documentation. This allows organizations to support traditional dictation, real-time speech recognition, ambient documentation, or a combination of these approaches based on the needs of individual users and departments.
Don’t Confuse ‘Compatible’ With ‘Fully Integrated’
Vendor terminology can make product comparisons difficult.
A product described as compatible with an EHR may allow clinicians to dictate into supported text fields. Another product may use an API or application-specific interface to exchange information or enable deeper workflow integration. A third may provide voice-driven navigation or workflow automation. A single product may use more than one of these approaches.
When comparing products, ask the vendor to demonstrate exactly what happens in the applications your clinicians use.
- Where the clinician can dictate.
- How the user moves between fields and applications.
- Whether commands and shortcuts work outside the primary EHR.
- How the system behaves in virtual environments.
- What configuration is required from IT.
- What happens when applications change or are upgraded.
Seeing the actual workflow is more useful than relying on the word integration alone.
How Fusion Narrate Approaches Application Compatibility
Fusion Narrate is cloud-based front-end medical speech recognition with a lightweight client and a speech profile that follows the clinician across supported devices and applications. Its cursor-based approach is designed to allow clinicians to use speech recognition in a broad range of applications rather than limiting documentation to one named EHR.
Fusion Narrate has been tested with more than 100 applications, including EHRs, LIS, RIS/PACS, browser-based tools, and desktop applications. In supported text fields, real-time dictation generally does not require a dedicated application-specific interface. More advanced workflows, including multi-step voice shortcuts and automation, can be configured for particular applications.
Organizations can review Dolbey’s Application Compatibility List and test their own workflow during a free evaluation. Compatibility should still be validated against the organization’s specific environment. EHR configuration, virtualization, security controls, specialty applications, and workflow requirements can differ among healthcare organizations.
Evaluate the Workflow, Not Just the EHR Name
Choosing medical speech recognition is not simply a matter of finding your EHR on a compatibility list.
Look at where clinicians actually work, how many applications they use, what tasks surround the dictation process, and how much configuration will be required to support those workflows. Then ask each vendor to demonstrate those workflows using your real-world applications.
For organizations using a diverse mix of EHRs and specialty systems, broader application flexibility can be just as important as speech recognition itself.
See whether Fusion Narrate works with the applications your clinicians use. Review the Application Compatibility List, try Fusion Narrate free for 14 days, or request a demonstration.
Image Source: Shutterstock /Andrey_Popov