Анализ исследований по нормативным документам для сбора и хранения больших мультимодальных данных
Artificial Intelligence is the one fast evolving driving field of computer science, transforming medical practice and healthcare as a whole. The development of Artificial Intelligence in Kazakhstan’s healthcare sector depends on the availability of a robust regulatory framework for managing large-scale multimodal medical data, including electronic medical records, medical imaging, and genomic data. This article analyzes the challenges of implementing Artificial Intelligence technologies in Kazakhstan’s healthcare system and provides practical policy recommendations. The policy brief compares the current legislation and standards of Kazakhstan with international benchmarks for this emerging field. The analysis reveals that while Kazakhstan has an adequate regulatory foundation for data protection, there are critical gaps in Artificial Intelligence regulation, data interoperability, and workforce preparedness. The authors propose a multi-level governance model and a detailed algorithm for the implementation, testing, and monitoring of Artificial Intelligence technologies. The results include practical recommendations for developing state policies and improving the regulatory base to position Kazakhstan as a regional leader in the ethical and effective use of Artificial Intelligence in healthcare. This brief is intended for policymakers involved in the implementation of innovative technologies in healthcare to support evidence-based and effective decision-making, as well as for healthcare professionals, digitalization experts, and the general public interested in health innovation policy. What is the problem? When considering the introduction of new technologies based on multimodal data and Artificial Intelligence applications in healthcare, IT is essential to understand the underlying legal and regulatory framework. The key issue is the discrepancy between Kazakhstan’s evolving legislative base and the stringent international requirements for the processing of multimodal medical data used for training and validating Artificial Intelligence systems. The main challenges include: The following approaches could be considered to improve the situation: - including ICD-related issues in educational programmes for healthcare professionals; - introducing mandatory continuing education and certification for healthcare professionals and experts; - conducting regular audits of the quality of ICD data entry; - developing an interdepartmental programme to improve the quality of mortality data. Regulatory fragmentation and lack of specialized Artificial Intelligence regulation Laws and standards are dispersed across multiple documents, making compliance difficult. There are no targeted regulations covering the full data lifecycle for AI, including management, storage, and versioning of data libraries and terminologies (reference sets, classifiers). Missing standards for data quality, model validation, clinical trials, explainability, and algorithmic monitoring. Systemic gaps in data governance and absence of key roles Key roles for managing medical data—Data Steward, Data Annotator, Clinical Data Analyst—are not legally defined or institutionalized. This results in a lack of accountability for data quality, annotation, and security, rendering data unsuitable for building reliable AI models. Technical and semantic fragmentation of data. Incomplete integration of key international standards (Health Level Seven Fast Healthcare Interoperability Resources (HL7 FHIR), Systematized Nomenclature of Medicine – Clinical Terms (SNOMED CT), Logical Observation Identifiers Names and Codes (LOINC)) leads to fragmented and non-standardized data that cannot be scaled or harmonized. Critical shortage of interdisciplinary specialists There is a severe deficit of professionals combining expertise in data science/AI with clinical medicine and medical data standards. In Kazakhstan, decisions on the introduction of technology are made by the Ministry of Health after approval by developers and approval of the introduction of technology from such structures as the National Center for Drug Expertise, the National Scientific Center for Health Development and the Joint Commission on the Quality of Medical Services. Policy options establishment of a national regulatory framework for Artificial Intelligence in healthcare; development of a national centralized infrastructure – a repository/platform for the collection, storage, and governance of multimodal medical data; development of human resource capacity in the field of digital health and Artificial Intelligence. Vision for implementing policy options Assessing the feasibility of the proposed policy options in light of existing barriers and opportunities, we conclude that a phased and integrated approach to implementing all three strategic directions is necessary. These policy options are not mutually exclusive but form a logical sequence in which each stage builds upon the previous one, creating a comprehensive and sustainable Artificial Intelligence AI ecosystem in Kazakhstan’s healthcare system.
Сведения
- Опубликовано
- 2025 · С. jhd030
- Язык
- Русский
- ISSN
- 2663-1768 · 2663-1776 · 2225-9929
- Издатель
- Salidat Kairbekova National Research Center for Health Development