National E-Health system

Seeing as there was no legacy system to upgrade or build upon, Edenlab engineers had to build the entire system infrastructure from scratch. It was challenging but also beneficial as we could use a modern technology stack.

The base infrastructure is built on a microservices architecture and has two layers:

• a central component (open API) which includes national registries, classifiers and terminology services, central data storage, etc.;
• connected, private EHR systems, such as medical and laboratory information systems, information systems for patients, and NHS.

The architecture was deployed to Kubernetes on Docker. We used Elixir as our primary back-end language for better scalability and speed.

In order to eliminate the inefficient paper-based workflow, there was a need to digitize all medical data (any health-related information associated with regular patient care or as part of a clinical trial program).

We built a national clinical data repository (NCDR), which at its core is a high-loaded system for storing and exchanging medical information at the national level in the HL7 FHIR format. This national EHR system operates at two levels:

• EDS-signed medical records (which are legally enforceable and record physician responsibility)
• normalized medical data in HL7 FHIR format

A patient summary solution was introduced to facilitate the availability of critical data to the doctors for treatment. It contains the most crucial information about the patient from the NCDR system and is to be used, for example, during emergencies or first-time visits.

As part of the national E-Health project, aimed to digitize the healthcare system of Ukraine, there was a need to create a fast, secure, and transparent way for patients to receive their prescriptions in accordance with the the National Affordable Medicine Reimbursement Program (a government-run prescription reimbursement program for citizens).

The system allows physicians to create a digital prescription that can then be dispensed by pharmacies. Patients receive a code on their phones, via SMS, that is then used to get their medications at the pharmacy. The prescription record is also saved to the eHealth database, and the pharmacy can be reimbursed by the government. With the use of the ePrescription, 27% of all electronic prescriptions are repaid within one hour.

The medications registry is a part of the solution and can easily be integrated with any other private medication registry. We built a program-management module that controls the reimbursement of medication. Any new medical prescription program can be smoothly added. A rule engine was built to allow users to create any configuration depending on different population segments.

In addition to being able to reimburse medication, we added the capability for the reimbursement of medical devices of any kind, such as glucose meters for example. 

The reimbursement algorithm ensures the secure creation and approval of reimbursement contracts through a series of validations and verifications. Once validated, the reimbursement contract must be approved by both the NHS and MSP before being finalized and stored securely, ensuring compliance and accuracy throughout the process.

eReferrals work in a similar way; the referral is created in the cloud and afterwards a patient can choose a service provider who then confirms the eReferral.

This way, interoperability between service providers is achieved, and the results are accessible for the requester. The service provider can be reimbursed by the government according to the pre-configured programs.

Based on the medical data provided in the eReferral, such as test results, a diagnostic report can be created in the system. This report can then be accessed and interpreted by the physician requesting the referral. The reports use the LOINC code.

In order to improve patient quality of life and avoid possible complications from diabetic diseases in the EHR, a relevant patient disease history was implemented which teaches patients to hold recommended consultations with specialists and simplifies access to free insulin.

Data security and integrity is crucial in healthcare. We implemented an algorithm based on blockchain that protects the data in registries from unauthorized changes. The entered data is authenticated by an electronic digital signature.

For example, when a patient-doctor agreement enters the database, hash keys are created based on this data. Each day, a block of hash keys is created which are used further in the newer agreements entered into the system. Thus, the system ensures that even the super administrator of the system cannot change the agreements.

The EHR system operates as a platform for medical information systems connected via REST APIs. This nature of interacting with multiple clients puts a special strain on the already sensitive issue of patient medical data security. The system implements the pseudonymization mechanism recommended by the European GPDR regulation – the separation of personal and medical data. Thus, even a hypothetical intruder who gains access to medical data will not be able to link it to a specific patient.

The Attribute-Based Access Control (ABAC) approach to access control has been implemented, allowing for flexible configuration of access policies to patient medical data and allowing the patient to determine with whom to share his or her medical information.

In accordance with the legislation, the National Health Service of Ukraine (NHSU) had to improve measures to ensure targeted and effective use of funds under the medical guarantee programs. That’s why a tool was developed for reviewing and analyzing medical data in the electronic data repository to meet the needs of the clinical and practical monitoring department of the NHSU.

 Defining forbidden groups: in order to extend the rules on restrictions on access to patient medical data in EHR, namely to restrict access to information on human immunodeficiency virus (HIV) and mental health, our team implemented functionality for coding diagnoses and medical records that contain such diagnoses. This also takes into account the likelihood of new regulations that can establish a separate regulation of access to information about other diseases that are not currently subject to such regulation.

To ensure the accuracy and reliability of information in the central database of the electronic healthcare system, we implemented a process of verifying a person’s data. All information of all registries is subject to verification in order to ensure the management of data relevance, in particular, detection of duplication of records and other inaccurate or incomplete information.

First, we built a duplicate-identification model which defines the probability of duplicate records. The model employs supervised machine learning for deduplication of data and merging duplicate records.

There are two processes for identifying duplicates: online (checks for duplicates before patient profile creation) and offline (performs a thorough duplicate search).

Third-party applications and patient portals are authorized to interact with the NHS system by verifying access tokens and user permissions. The system ensures the validity of access tokens and checks whether they have the correct scope (such as viewing or editing patient details). If any issues are found (e.g., expired tokens or missing permissions), appropriate error messages are returned. The system also ensures that patient data access is secure by validating that the token contains a person ID and that the patient’s status is active.

Users can interact directly with patient data through third-party applications, enabling them to view, edit, and manage patient declarations, such as signing or terminating agreements. When updates are made to patient data, the system ensures request validity by verifying active status, checking for potential conflicts (like unauthorized confidant updates), and applying schema validations for fields such as tax IDs, addresses, and documents.

Additionally, the system generates secure upload links for document updates and renders a printout form if required for patient or confidant verification. It prevents duplicate requests and handles pending or canceled person requests efficiently.