If we think of Electronic Health Record (EHR) not only as an infrastructure for sharing documents, but an intelligent platform for exploiting the present information, we need to change our architecture.
The IHE XDS protocol, which is the basis of many EHRs, was defined before the current AI technologies and is focused on sharing HL7 CDA documents. In order to exploit the potential of modern technologies, it is necessary to design an EHR based on the HL7 FHIR standard.
FHIR, an acronym for Fast Healthcare Interoperability Resources, was developed specifically to meet the new challenges of digital healthcare, including automated support for clinical decisions and machine information processing.
The current version of FHIR (R4) includes 145 types of resources, grouped into five categories: foundation, base, clinical, financial, specialized. Some of these are particular important for smart EHR:
- Patient
- Encounter
- Condition (Problem)
- AllergyIntolerance
- Procedures
- DiagnosticReport
- Medication
- MedicationRequest
Other resources that should be considered are:
- EpisodeOfCare
- CarePlan
- CareTeam
- Task
- Appointment
- Observation
An EHR based on FHIR could expose these resources to hospital information systems, territorial systems, EMRs primary care, medical apps, allowing a better and more effective interoperability than IHE XDS today allows.
The presence of these resources and the underlying data model would allow the use of current artificial intelligence technologies to extract and highlight relevant clinical information, organize it and correlate it with each other, and perform predictive analysis.
This is a research field in which important companies such as Google, IBM, Microsoft, Apple and several universities are working on. For this reason, in next articles, we will deepen the topic by illustrating the possible applications of AI technologies in the clinical data processing.
Stay tuned on digital-health.blog!
Digital-health 