Finland
September 1, 2023
CAFALY DUAL
September 1, 2023Interoperability on Artificial intelligence
The digitalization of medicine promises significant advances for global health. In combination with artificial intelligence, cloud computing, and big data analytics, this data holds great potential for personal health, healthcare, and health-related environments worldwide. (1) Medical data has value only if it can be converted into meaningful information. So, high-quality datasets, seamless communication across IT systems, and standard data formats are required for the information processing of both humans and machines. Nevertheless, large parts of today’s medical data could be more useful: Isolated data pools with incompatible systems make it difficult to exchange, process, and interpret. To reach the full potential of health-related information and AI, an interconnected data infrastructure with rapid, reliable, and safe interfaces, there is a need for international standards for data exchange and medical terminologies to define unambiguous vocabularies for the communication of medical information. Digital medicine only sometimes needs sophisticated analytics or complex AI algorithms. Making the correct information available to the right person at the right time can significantly improve patient care. (1)
Interoperability in Digital Medicine
Interoperability of health data is essential; it can be defined as the ability of various systems to exchange information and use appropriately the information that has been exchanged. The most significant barrier to applying AI and big data to medicine is not a lack of algorithms but relevant data. (1)
Technical interoperability
Technical interoperability implies basic data exchange capabilities between systems (for example, moving data from a USB stick to a computer). These capabilities require communication channels and protocols for data transmission. (1)
Syntactic interoperability
Syntactic interoperability relates to the format and structure of the data. The structured exchange of healthcare-related data is supported by international standards development organizations (SDOs), like Health Level Seven International (HL7) or Integrating the Healthcare Enterprise (IHE). Two examples of these types of systems are HL7’s Fast Healthcare Interoperability Resources (FHIR), which defines around 140 common healthcare concepts, and OpenEHR, which allows medical professionals and health IT experts to define clinical content using archetypes specifications of clinical concepts, including a portable language for querying, the Archetype Querying Language (AQL)). (1)
Semantic interoperability
Semantic interoperability implicates the use of medical terminologies, nomenclatures, and ontologies. They ensure that the meaning of medical concepts can be shared across systems, providing a common digital language for medical terms that is, ideally, understandable to humans and machines worldwide. SNOMED CT works as a general-purpose language for advancing semantic interoperability in medicine and healthcare; also, it can be complemented by domain-specific terminologies such as, for example, Logical Observation Identifiers Names and Codes (LOINC) for laboratory observations, the Identification of Medicinal Products (IDMP) for medicines, the HUGO Gene Nomenclature Committee (HGNC) for genes or the Human Phenotype Ontology (HPO) for phenotypic abnormalities. (1)
Organizational interoperability
At the highest level, interoperability involves organizations, legislations, and policies. Exchanging and managing data across health IT systems is not an end but should help healthcare professionals work more efficiently and improve patients’ health. The appropriate exchange of standard business processes and workflows enables seamless healthcare provision across institutions. At this point, interoperability could be mandatory if not enforced worldwide. (1)
Future directions
Interoperability will be essential concerning the multiple components of a typical clinical workflow. Allowing integration will require a set of standards between the different algorithms and simultaneously enable algorithms to be run on different equipment worldwide. The lack of early efforts to optimize interoperability will result in the practical ineffectiveness of AI technologies. (2) Looking to the future, the FHIR framework is critical for the implementation of AI-based health technologies and healthcare-related areas that utilize electronic data frequently, the same way as DICOM and PACS became essential for the exchange of digital medical images. (2)