Biotechnology Software As A Medical Device Law In Russia.
1. Introduction
Software as a Medical Device (SaMD) is an important area of modern biotechnology, healthcare regulation and digital medicine. It refers to software designed for medical purposes that performs functions such as diagnosing diseases, monitoring patients, predicting medical conditions, analysing biological information or supporting clinical treatment without necessarily being incorporated into a physical medical device.
In Russia, SaMD regulation operates within the broader legal framework governing medical devices, public health, personal data, information technology, clinical research and medical product registration.
Examples include:
AI software that detects cancer through medical imaging.
Genomic analysis software used to identify genetic disorders.
Bioinformatics applications that analyse DNA and RNA sequences for clinical purposes.
Software that predicts cardiovascular disease.
Digital pathology systems that identify abnormal cells.
Software supporting personalised medicine and biotechnology-based treatment.
Applications that monitor patients and provide clinically significant recommendations.
The central legal question is whether software performs a medical function and therefore falls within the regulatory definition of a medical device, rather than merely providing general information or administrative assistance.
Russia's regulatory framework must also be considered alongside the Eurasian Economic Union (EAEU) medical-device registration system.
2. Legal Framework Governing SaMD in Russia
The principal legislation and regulatory instruments are as follows.
| Legal instrument | Regulatory significance |
|---|---|
| Federal Law No. 323-FZ of 21 November 2011 | Establishes the legal definition and regulation of medical devices under Article 38. |
| Government Resolution No. 1684 of 30 November 2024 | Establishes the Russian state-registration procedure for medical devices. |
| EAEU Commission Decision No. 173 of 22 December 2015 | Establishes risk-based classification of medical devices. |
| EAEU Commission Decision No. 65 of 23 May 2023 | Introduces specific risk-classification rules for software medical devices. |
| EAEU Commission Recommendation No. 25 | Provides criteria for determining whether software qualifies as a medical device. |
| Federal Law No. 152-FZ of 27 July 2006 | Governs personal data processing, including health-related personal information. |
| Federal Law No. 323-FZ | Also governs patient rights, medical confidentiality and healthcare delivery. |
| GOST R 59766-2021 | Provides a framework for categorising risks associated with SaMD. |
Important regulatory authority: The Federal Service for Surveillance in Healthcare (Roszdravnadzor) is responsible for medical-device registration, market surveillance and monitoring of medical-device safety.
11reg.roszdravnadzor.ru+2
2.1 Article 38 of Federal Law No. 323-FZ
Article 38 is the foundation of Russian medical-device regulation.
It includes special software intended by its manufacturer for:
Disease prevention.
Diagnosis of medical conditions.
Disease treatment.
Medical rehabilitation.
Monitoring human health.
Medical research.
Supporting changes or replacement of anatomical structures or physiological functions.
The legislation does not restrict medical-device regulation to physical equipment.
Consequently, software can independently qualify as a medical device even if it is installed on a smartphone, computer, hospital server or cloud platform.
For example, software that analyses MRI scans and identifies potentially malignant tumours may fall within Article 38 because it performs a diagnostic function.
roszdravnadzor.gov.ru+1
2.2 Government Resolution No. 1684 of 2024
Government Resolution No. 1684 establishes the current national registration framework.
It expressly includes special software among products subject to medical-device registration when the manufacturer's intended purpose falls within the statutory medical purposes.
Roszdravnadzor administers the registration process.
The registration documentation for software medical devices includes relevant information about:
Intended medical purpose.
Software functionality.
Technical characteristics.
Software versions.
Instructions for use.
Quality, safety and effectiveness.
Clinical and technical evidence.
Risk classification.
The rules also recognise software using artificial intelligence as a distinct regulatory category in the registration documentation.
ПП РФ 1684 NEW.doc+1
3. When Does Software Become a Medical Device?
Not every healthcare-related software application is legally a medical device.
The EAEU criteria are particularly important in determining whether independent software qualifies as SaMD.
Under the relevant criteria, four conditions must be satisfied.
First: Independent software product
The software must be a computer program or software module, irrespective of its hardware platform or method of delivery.
It may operate through:
Mobile applications.
Desktop applications.
Cloud computing.
Web-based systems.
Artificial intelligence platforms.
Software-as-a-Service (SaaS).
Second: Not merely an integral part of another medical device
The software must not simply be an integral component of another medical device.
Software controlling the internal functioning of a medical instrument may instead be regulated as part of that instrument.
Third: Medical purpose
The manufacturer must intend the software to be used in providing medical care.
Fourth: Interpretation of clinical data
The software must interpret data automatically, including through artificial intelligence, or according to parameters established by a medical professional, in a way that affects clinical decision-making.
eec.eaeunion.org+1
Examples of software classification
| Software functionality | Likely regulatory treatment |
|---|---|
| Hospital appointment scheduling | Generally not SaMD |
| Electronic storage of patient records | Generally not SaMD by itself |
| Displaying laboratory results without interpretation | Generally not SaMD |
| Converting medical measurements into different units | Generally not SaMD |
| AI-based cancer detection | SaMD |
| Software interpreting genetic mutations for clinical decisions | SaMD |
| Software predicting stroke risk | SaMD |
| Software recommending treatment based on patient data | Potentially SaMD |
| Software performing clinical image interpretation | Potentially SaMD |
The distinction is important because the regulatory classification depends on the intended medical function, not simply the fact that software is used in a hospital.
4. Risk Classification of Software Medical Devices
Russia follows a risk-based regulatory approach, taking account of the potential harm that may result from incorrect software operation.
The EAEU framework establishes four principal risk classes.
SaMD risk classification
Class 1
Low risk
Software with comparatively low potential consequences for patients.
Class 2a
Medium risk
Software providing information that can influence clinical decisions in less critical circumstances.
Class 2b
Elevated risk
Software supporting decisions where incorrect results may have serious clinical consequences.
Class 3
High risk
Software involving the highest potential risk, including AI-based software under the applicable EAEU classification rules.
The classification of non-AI software considers:
The medical purpose of the software.
The clinical situation in which it is used.
The significance of the information generated.
The consequences of an incorrect clinical decision.
The potential harm to an individual or public health.
An important feature of the EAEU rules introduced through Decision No. 65 of 2023 is that software medical devices using artificial intelligence are assigned Class 3 under the relevant classification framework.
This has substantial consequences for manufacturers developing AI-based biotechnology products in Russia.
Webiomed+2
5. Registration Procedure for SaMD in Russia
Registration is an essential legal requirement before a medical device can ordinarily be placed into circulation in Russia.
The registration authority is Roszdravnadzor.
Main registration stages
Identification of the product
Determine whether the software qualifies as an independent medical device under Article 38 and the applicable EAEU criteria.
Determination of risk class
Identify the applicable classification and justify the risk class based on intended use and clinical consequences.
Preparation of technical documentation
Prepare the software description, architecture, intended users, functionality, version information and operating instructions.
Technical testing and validation
Establish that the software performs its intended functions reliably and consistently.
Clinical evaluation
Provide appropriate evidence supporting clinical effectiveness and safety.
Submission to Roszdravnadzor
Submit the registration dossier and required supporting materials.
Regulatory examination and decision
The regulator examines the required evidence and determines whether registration can be granted.
Post-market monitoring
Monitor software performance, adverse events, defects and safety-related updates.
Special requirements for software documentation
Software manufacturers must pay particular attention to:
Software identification and version numbering.
Intended purpose and clinical functionality.
Software architecture.
Input and output data.
Verification and validation.
Risk-management documentation.
Cybersecurity.
Clinical performance.
Software modification procedures.
User instructions and limitations.
EAEU Decision No. 29 of 12 February 2016 contains specific requirements concerning the technical documentation of software medical devices.
promeddev.com+2
6. Artificial Intelligence and Biotechnology Software Regulation
Artificial intelligence is particularly important in Russian biotechnology because AI-based software can analyse biological and clinical datasets and produce recommendations that influence medical treatment.
Examples include:
AI-based genomic sequencing interpretation.
Machine-learning cancer detection.
Automated pathology analysis.
Predictive genetic disease modelling.
AI-supported drug-response prediction.
Personalised treatment recommendation systems.
AI-based embryo and reproductive-health analysis.
6.1 AI-based medical software
The regulatory framework treats AI-enabled SaMD as a high-risk category under the applicable EAEU classification rules.
Manufacturers must therefore address:
Accuracy of AI predictions.
Reliability of training datasets.
Algorithm validation.
Data quality.
Bias and performance limitations.
Explainability and clinical interpretability.
Human oversight.
Software modification and retraining.
For example, if an AI system analyses genomic information and predicts that a patient has a hereditary cancer risk, the manufacturer must establish that the system's intended medical function is supported by appropriate evidence.
A disclaimer stating that the software is only informational will not necessarily remove medical-device obligations if its actual intended purpose and functionality are medical.
6.2 AI updates and algorithm modifications
One significant regulatory issue concerns continuous learning.
Suppose a Russian biotechnology company develops software for identifying cancer cells and subsequently changes its machine-learning model.
The modification may affect:
Diagnostic accuracy.
Intended purpose.
Clinical performance.
Risk classification.
Safety characteristics.
The manufacturer must assess whether the modification requires amendments to its registration documentation or further regulatory assessment.
An update cannot automatically be treated as an ordinary technical improvement.
7. Personal Data and Genetic Information Protection
Biotechnology SaMD frequently processes highly sensitive personal information.
This includes:
DNA sequencing results.
Genetic predisposition information.
Patient medical histories.
Laboratory results.
Biomarker information.
Reproductive-health information.
Medical imaging.
Clinical trial information.
7.1 Federal Law No. 152-FZ
The Russian Personal Data Law, Federal Law No. 152-FZ, establishes requirements governing the processing of personal data.
Health-related information is generally classified as special-category personal data.
Genetic information also requires careful treatment under the applicable rules governing personal data and medical confidentiality.
A biotechnology software manufacturer processing such information must consider:
Lawful grounds for processing.
Consent where required.
Purpose limitation.
Data minimisation.
Appropriate security safeguards.
Access restrictions.
Retention requirements.
Rules governing international transfers.
7.2 Medical confidentiality
Article 13 of Federal Law No. 323-FZ protects medical confidentiality.
A software provider working with hospitals must ensure that its access to confidential patient information is legally justified and appropriately controlled.
For example, a cloud-based genomic analysis platform cannot assume that a hospital's possession of patient data automatically authorises every subsequent commercial use of that data.
7.3 Cross-border data transfers
A Russian SaMD company transferring patient information to foreign cloud servers must examine the requirements of Federal Law No. 152-FZ, including applicable localisation and cross-border transfer rules.
This is particularly relevant to international biotechnology companies collaborating with research institutions, pharmaceutical manufacturers and genetic laboratories.
roszdravnadzor.gov.ru+1
8. Clinical Evidence and Scientific Validation
Clinical validation is an important part of SaMD registration.
A manufacturer must establish that the software is sufficiently reliable and safe for its intended medical purpose.
For biotechnology software, evidence may include:
| Evidence | Purpose |
|---|---|
| Analytical validation | Establishes accuracy of data processing. |
| Clinical validation | Establishes relevance to actual medical practice. |
| Software verification | Checks whether the software meets its specifications. |
| Software validation | Establishes suitability for its intended use. |
| Performance testing | Measures accuracy, sensitivity and specificity where applicable. |
| Cybersecurity testing | Evaluates software security and vulnerabilities. |
| Risk assessment | Identifies potential harm from software failure. |
Consider a genomic diagnostic application that identifies mutations associated with inherited diseases.
The manufacturer should establish that:
The underlying genetic reference data are reliable.
The algorithm correctly identifies relevant variants.
The results are clinically meaningful.
The software communicates uncertainty appropriately.
The output is suitable for its intended medical purpose.
A software product that produces technically impressive results is not automatically clinically effective.
9. Manufacturer Liability and Legal Accountability
SaMD manufacturers may face different forms of legal responsibility.
9.1 Civil liability
Under the Russian Civil Code, manufacturers may face liability where defective products cause legally recognised harm.
Relevant provisions include:
Article 1064 – General grounds of liability for causing harm.
Article 1095 – Liability for harm caused by defects in goods, work or services.
Article 1096 – Persons responsible for harm caused by defects.
Article 1098 – Grounds for exemption from liability.
For example, suppose diagnostic software incorrectly identifies a malignant tumour as benign and the patient suffers injury because of delayed treatment.
Depending on the facts, liability may arise from defective software, inadequate validation, misleading instructions or other legally relevant failures.
9.2 Administrative liability
Regulatory violations may result in administrative consequences under the Code of Administrative Offences of the Russian Federation.
Potentially relevant provisions include:
Article 6.28 – Violation of established rules concerning the circulation of medical devices.
Article 6.33 – Illegal circulation of falsified, substandard or unregistered medical devices, where the statutory elements are satisfied.
Article 19.5 – Failure to comply with lawful regulatory orders.
The precise provision depends on the nature of the violation.
9.3 Criminal liability
Where defective medical software contributes to serious injury or death, criminal liability may arise if the relevant elements of an offence are established.
Potentially relevant provisions include:
Article 109 – Causing death by negligence.
Article 118 – Causing serious harm to health by negligence.
Article 238 – Providing services or selling products that do not meet safety requirements, where applicable.
Criminal liability is not automatic merely because an algorithm makes an incorrect prediction. The prosecution must establish the relevant elements, including the legally required causal connection and fault.
10. Case Laws and Judicial Interpretation
An important research limitation must be recognised: Russian reported judicial decisions specifically addressing the independent classification and registration of AI-based biotechnology SaMD remain limited. The following cases are relevant to the wider legal principles governing medical-device registration and circulation. They should not be represented as direct judicial decisions on AI diagnostic algorithms.
Case 1: Arbitration Court of Perm Krai – Case No. A50-30273/2022
Relevant medical-device registration precedent
Decision: 6 October 2023.
Appellate proceedings: Seventeenth Arbitration Court of Appeal, Case No. 17АП-425/23, 19 December 2023.
Legal issue:
The dispute concerned the supply of medical equipment and alleged defects preventing its proper use. The proceedings involved the application of Article 38 of Federal Law No. 323-FZ and the legal requirements governing medical-device circulation.
Legal principle:
The court considered the statutory definition of medical devices and the regulatory requirements applicable to their circulation.
The judgment demonstrates the significance of lawful medical-device status and compliance with registration requirements.
Relevance to SaMD:
Although this was not a software-specific case, its statutory reasoning is relevant when examining whether software constitutes a medical device and whether the registration requirements have been fulfilled.
ГАРАНТ+1
Case 2: Sixteenth Arbitration Court of Appeal – Case No. A63-7306/2023
Decision: 4 July 2024.
Legal issue:
The dispute involved medical equipment and the application of the Russian medical-device registration rules.
The court considered the relevance of Government Resolution No. 1416, which governed registration under the earlier national framework.
Legal principle:
The registration certificate and the identification of the medical device are important elements of lawful circulation.
Relevance to SaMD:
This reasoning is relevant to software manufacturers because registration must correspond to the product and its approved intended purpose.
For example, a manufacturer cannot simply assume that registration of a medical imaging system automatically authorises the independent marketing of a separate diagnostic AI product.
The scope of the registration documentation is therefore important.
ГАРАНТ
Case 3: Regulatory interpretation under EAEU Commission Recommendation No. 25
This is not a judicial judgment, but it is an important interpretative authority.
Legal issue: Whether software should be classified as a medical device.
Principle: Software is classified as a medical device when the prescribed criteria are satisfied, including independent software status, medical intended purpose and interpretation of data affecting clinical decisions.
The criteria distinguish medical software from administrative applications and basic data-display functions.
Relevance:
This provides a practical legal test for companies developing:
AI genomic platforms.
Digital pathology applications.
Genetic diagnostic software.
Clinical decision-support systems.
eec.eaeunion.org+1
Case 4: EAEU Commission Decision No. 65 of 23 May 2023
This is also a regulatory instrument rather than a court judgment.
It amended the software medical-device risk-classification framework under Decision No. 173.
Legal significance:
It introduced specific rules for software medical devices and established the four risk classes.
The treatment of AI-based SaMD as Class 3 is especially important for manufacturers developing diagnostic biotechnology algorithms.
Webiomed+1
Judicial research conclusion: The first two cases are actual arbitration decisions concerning medical-device regulatory compliance, while the latter two are regulatory authorities rather than court cases. A legal researcher should not cite them as judgments specifically deciding SaMD classification.
11. Intellectual Property Rights and Software Ownership
Biotechnology SaMD development also raises intellectual-property questions.
Relevant Russian legislation includes Part IV of the Civil Code of the Russian Federation.
Important provisions include:
Article 1259 – Copyright-protected works.
Article 1261 – Computer programs as copyright-protected works.
Article 1295 – Service works.
Article 1350 – Patentability of inventions.
Article 1358 – Exclusive rights to inventions.
Software source code, algorithms expressed in protectable forms and associated documentation may receive copyright protection.
However, an abstract mathematical algorithm or medical method does not automatically qualify for patent protection.
Important legal disputes
Potential disputes include:
Ownership of AI diagnostic software.
Employee-created biotechnology software.
Licensing of genomic databases.
Ownership of machine-learning training datasets.
Copyright infringement involving software code.
Patent disputes involving computer-implemented inventions.
Commercial exploitation of clinical research software.
For example, if a Russian biotechnology company employs researchers to develop an AI-based cancer detection platform, ownership questions may arise concerning employment agreements, software development contracts, database rights and intellectual-property assignments.
12. Regulatory Challenges Facing Biotechnology SaMD Companies
Several significant legal challenges arise in the Russian market.
| Challenge | Legal concern |
|---|---|
| AI diagnostic accuracy | Reliability and patient safety |
| Genetic data interpretation | Personal data protection |
| Cloud-based medical software | Data security and localisation |
| Continuous algorithm updates | Registration amendments |
| International software licensing | Intellectual-property and regulatory compliance |
| Clinical validation | Evidence of effectiveness |
| Cyberattacks | Patient safety and information security |
| Software failure | Civil, administrative or criminal liability |
| Cross-border research | Data-transfer and regulatory requirements |
Particular challenge: AI accountability
Consider a biotechnology company that develops software to predict hereditary cancer.
The software incorrectly interprets a genetic variant and provides a misleading clinical recommendation.
Three separate legal questions arise:
Did the manufacturer comply with medical-device registration and validation requirements?
Did the healthcare professional appropriately evaluate the software's recommendation?
Did the software defect or professional conduct cause legally recognised harm?
These questions must be assessed separately. Russian law does not automatically transfer all responsibility to either the software manufacturer or the medical professional.
13. Future Development of SaMD Law in Russia
Several emerging developments are likely to influence Russian biotechnology regulation.
1. AI-based genomic diagnostics: Increasing use of machine learning for genetic disease identification will increase the importance of clinical validation and algorithm accountability.
2. Personalised medicine: Software interpreting patient-specific biological information will require stronger evidence of clinical performance.
3. Digital pathology: Automated identification of cancer cells and tissue abnormalities will raise questions about clinical reliability and professional responsibility.
4. Cloud-based medical software: International data hosting and remote access will require careful compliance with Russian personal-data requirements.
5. Continuous-learning algorithms: Regulatory frameworks will need to address changes to AI models after initial registration.
6. Biotechnology cybersecurity: Software vulnerabilities may become patient-safety issues rather than merely technical problems.
7. International regulatory cooperation: EAEU harmonisation will remain important for manufacturers seeking access to multiple Eurasian markets.
14. Critical Legal Analysis
Russian SaMD regulation demonstrates the transition from traditional medical-device regulation towards software-based medical technologies.
Article 38 of Federal Law No. 323-FZ provides the foundation by recognising special software within the medical-device framework.
Government Resolution No. 1684 of 2024 establishes the national registration requirements, while EAEU Decisions No. 173 and No. 65 provide an important risk-classification framework.
Three legal tensions deserve particular attention.
First, innovation versus patient safety. Strict registration and evidence requirements protect patients, but may increase development costs and delay market entry for smaller biotechnology companies.
Second, algorithmic innovation versus regulatory certainty. AI models can change after deployment. Manufacturers therefore need clear procedures for assessing whether software modifications require regulatory approval.
Third, technological development versus privacy rights. Genomic and clinical datasets are essential for medical AI, but their use must comply with personal-data protection and medical-confidentiality requirements.
The central regulatory challenge is to ensure that software innovation does not compromise patient safety, scientific reliability or individual rights.
15. Conclusion
Biotechnology Software as a Medical Device Law in Russia is governed principally by Federal Law No. 323-FZ, Government Resolution No. 1684 of 2024 and the relevant Eurasian Economic Union regulatory instruments.
The framework recognises that medical software can constitute an independent medical device even when it operates through cloud platforms, smartphones or artificial intelligence systems.
Its legal regulation involves several interconnected areas:
Medical-device registration.
Risk classification.
AI accountability.
Clinical validation.
Genetic-data protection.
Medical confidentiality.
Cybersecurity.
Intellectual-property rights.
Manufacturer liability.
Post-market surveillance.
For biotechnology companies, the most important legal consideration is determining whether their software performs an independent medical function affecting clinical decision-making. If it does, the applicable medical-device registration and safety obligations must be carefully examined before commercial distribution.
16. Important Legal References and Research Sources
The following sources are particularly useful for further academic research.
Federal Law No. 323-FZ – Fundamentals of Healthcare Protection in Russia.
Government Resolution No. 1684 of 30 November 2024 – Medical Device Registration Rules.
EAEU Commission – Criteria for Classifying Products as Medical Devices.
EAEU Decision No. 29 of 12 February 2016.
GOST R 59766-2021 – Software as a Medical Device: Risk Categorisation.
Roszdravnadzor – Medical Device Registration.
Research: State Registration of Software as a Medical Device under EAEU Rules.
Final academic observation: For a comparative biotechnology-law dissertation, Russian SaMD regulation is particularly significant because it combines national medical-device registration, EAEU harmonisation, AI risk classification and genetic-data protection within one regulatory framework.
Legal research qualification: This is an academic overview, not a formal Russian legal opinion. For an actual product launch, the current consolidated legislation, applicable EAEU registration route and product-specific classification should be checked. The cited arbitration cases concern general medical-device compliance, not direct judicial rulings on independent SaMD classification.

comments