INDUSTRY AND TECHNOLOGY PERSPECTIVE

Abstract

Blood testing generates clinically relevant measurements, but measurement alone does not ensure understanding. Conventional laboratory reports remain organized primarily around individual biomarkers, numerical values, and population reference intervals. They often provide limited explanation of relationships between biomarkers, changes over time, or the questions that patients should discuss with clinicians.

Aima Diagnostics has developed an AI-assisted interpretation layer intended to address this gap. The system integrates biomarker analysis, longitudinal comparison, and constrained plain-language reporting within a framework of clinician oversight. It is designed to support patients, clinicians, laboratories, clinics, and manufacturers of diagnostic equipment without replacing laboratory measurement or clinical judgment.

This article presents the Aima Diagnostics technology, its intended value across the diagnostic ecosystem, its commercial and research collaborations, and its proposed approach to evaluating AI-assisted interpretation. The company aims to contribute to the development of a shared, independently verifiable evaluation standard and, ultimately, a gold standard for the interpretation of blood test results.

Keywords: Aima Diagnostics; laboratory diagnostics; AI-assisted interpretation; blood test results; patient comprehension; longitudinal biomarkers; clinical decision support; point-of-care testing; laboratory medicine

1. The Missing Layer Between Measurement and Understanding

Blood testing generates clinically relevant measurements, but measurement alone does not ensure understanding.

A conventional laboratory report generally presents a list of analytes, numerical values, units, and reference intervals. For laboratory specialists and experienced clinicians, this is a familiar working format. For many patients, it remains difficult to interpret. Even when every value is technically correct, the report may not explain how different results relate to one another, whether a change is new, or which findings deserve further discussion.

Limited health and numerical literacy can make this problem more significant (Nielsen-Bohlman et al., 2004). Patients may misunderstand reference intervals, overreact to isolated abnormalities, overlook patterns requiring attention, or seek explanations from unreliable sources. The result is an interpretive gap between the information contained in a laboratory dataset and the understanding that patients and clinicians can practically derive from it.

Aima Diagnostics was established to address this interpretive gap. The company does not perform laboratory assays or position its technology as a replacement for clinical judgment. Instead, its platform provides an additional interpretation layer between laboratory measurement and subsequent patient or clinical action.

The purpose of this layer is to help transform laboratory results from a collection of measurements into a structured, contextualized, and longitudinal account of what may deserve attention.

2. The Aima Diagnostics Interpretation Technology

The Aima Diagnostics interpretation framework is organized around five connected functions.

2.1 Data ingestion and normalization

Laboratory results can be received through patient upload or through integration with laboratory and clinical information systems. The interpretation process requires consistent handling of biomarker names, units, reference intervals, measurement dates, and patient history.

Normalization is an essential part of this process. Results produced by different laboratories or analytical systems may use different units, reference intervals, naming conventions, and reporting structures. These differences must be addressed before results can be compared reliably.

2.2 Biomarker and relationship analysis

Each biomarker is evaluated against the applicable reference information. The system also examines relationships between biomarkers rather than treating every result as an entirely independent observation.

This approach is relevant because physiological and pathological processes often affect multiple parameters simultaneously. Iron-related indices may require interpretation in the context of inflammatory markers and other relevant clinical information. Glucose-related results may acquire additional meaning when reviewed with other metabolic or hepatic parameters, while lipid fractions are commonly interpreted in relation to one another.

The purpose of relationship analysis is not to generate speculative diagnoses. It is to organize potentially relevant patterns so that they can be communicated cautiously and, where appropriate, reviewed by a clinician.

2.3 Longitudinal interpretation

A single laboratory report shows a moment in time. Clinical interpretation often depends on the direction, rate, persistence, and reversal of change.

Aima Diagnostics preserves longitudinal results so that new measurements can be compared with previous findings. This allows the interpretation layer to distinguish an isolated result from a continuing pattern and to highlight changes that may not be apparent when reports are reviewed separately.

Longitudinal continuity also gives patients a more coherent record of their diagnostic history. Its value should come from better continuity and understanding, not from restricting patients’ ability to export or transfer their data.

2.4 Constrained plain-language reporting

The platform generates reports intended to be understandable without sacrificing clinical caution. Reports are organized around findings within and outside reference intervals, relevant biomarker relationships, longitudinal changes, and appropriate next steps.

The reporting layer operates under explicit constraints. These include avoidance of unsupported diagnostic statements, controlled communication of uncertainty, and mandatory escalation language when a finding may require prompt clinical attention.

The objective is not to encourage self-diagnosis. It is to help patients understand what their results may mean and to support more informed conversations with qualified healthcare professionals.

2.5 Clinician oversight and governance

Medical specialists participate in the development and review of interpretation logic, clinical language, and safety rules. AI specialists contribute to system architecture, quality control, technical evaluation, and the monitoring of model behavior.

Aima Diagnostics treats governance as part of the technology itself. Human oversight, traceability, conservative language, incident review, and control of system changes are preconditions for responsible deployment rather than optional additions. This approach is consistent with broader work on the clinical evaluation and governance of AI in healthcare (Esteva et al., 2019); (Rajpurkar et al., 2022); (Topol, 2019).

3. Improving Patient Understanding

For patients, the intended value of Aima Diagnostics is a clearer explanation of laboratory information.

An interpreted report can help a patient distinguish an isolated deviation from a persistent trend, understand that biomarkers may be clinically related, and identify questions that should be discussed with a clinician. It can also explain when a result does not, by itself, establish a diagnosis.

This approach is designed to improve comprehension without creating false certainty. A responsible report should communicate both what may be important and what remains unknown. It should neither amplify anxiety around minor deviations nor provide inappropriate reassurance when clinical assessment is needed.

Patient benefit should ultimately be evaluated through direct measurement. Relevant outcomes include correct recall of important findings, understanding of recommended next steps, recognition of findings requiring urgent attention, and comprehension across different levels of health literacy. Engagement or repeat testing alone is not sufficient evidence of improved understanding.

4. Supporting Clinicians and Clinical Workflows

Clinicians frequently need to review results produced at different times, by different laboratories, and in different formats. Part of this work involves routine comparison, reconstruction of trends, and repeated explanation of common findings.

Aima Diagnostics is designed to reduce this burden by organizing biomarker relationships and longitudinal changes into a structured, reviewable summary. The intended clinical value is twofold: helping clinicians review relevant information more efficiently and supporting better-informed decisions without removing clinical judgment from the process.

Any claim that the system improves clinical decisions or saves time must be evaluated directly. Time-and-workflow studies can measure the duration of result review and patient explanation before and after implementation. Clinical studies can assess whether important findings are identified more consistently and whether the system introduces omissions, misleading correlations, or inappropriate reassurance.

Clinical decision support is therefore framed as an area for validation, not as a benefit that can simply be inferred from the existence of the technology.

5. An Additional AI Layer for Laboratories and Clinics

For laboratories and clinics, Aima Diagnostics can extend the value of an existing diagnostic service without requiring replacement of analytical equipment.

The conventional patient journey often ends when a laboratory report is delivered. An interpretation layer creates an additional opportunity to provide understandable results, longitudinal comparison, and a clearer rationale for appropriate follow-up. This may strengthen patient relationships while giving laboratories and clinics a service capability that complements their core diagnostic activities.

A commercial case study published by Aima Diagnostics describes a five-week deployment. After six months, the reported repeat-visit rate increased from 12% to 16%, average revenue per patient increased by 20%, and complaints about unclear results fell by half (Aima Diagnostics, n.d.). These observations relate to a single commercial implementation and should not be treated as controlled evidence of clinical effectiveness.

Commercial outcomes also require careful interpretation. More repeat visits or additional testing are not inherently beneficial. Follow-up should be clinically appropriate, and commercial evaluation should be considered alongside patient comprehension, safety, and avoidance of unnecessary investigations.

6. Multidisciplinary Expertise and Research Collaborations

Aima Diagnostics has brought together professionals in medicine and artificial intelligence to develop the technology at the intersection of laboratory medicine, clinical practice, and health data science.

Medical specialists contribute knowledge of biomarker interpretation, clinical context, uncertainty, and patient safety. AI specialists contribute expertise in system development, data processing, quality assurance, and the technical evaluation of interpretation systems. This multidisciplinary structure is important because clinically responsible AI cannot be developed through language generation or software engineering alone.

The company is also developing academic, research, and commercial collaborations in the United States, Scandinavia, the United Kingdom, Ukraine, and other European markets. These relationships include laboratory networks, clinics, universities, and research organizations.

The purpose of these collaborations differs by partner. Commercial implementations provide experience with integration and operational workflows. Academic and research collaborations provide a basis for studying clinical concordance, patient comprehension, workflow effects, and safety. Research activities must be distinguished clearly from commercial deployment: a partnership or pilot is not, by itself, independent clinical validation.

7. Interpretation in Addiction Medicine: The Narcology Use Case

The collaborations of Aima Diagnostics include work with a private addiction treatment center in Ukraine. This setting presents a particularly demanding interpretation context.

Some patients receiving addiction-related care may have histories involving alcohol misuse, smoking, or psychoactive substance use. These factors can coexist with complex combinations of laboratory findings involving liver function, inflammation, metabolism, nutritional status, and other physiological systems. Patients may also have comorbid conditions or receive treatments that affect laboratory results.

In such circumstances, interpretation should not be limited to determining whether each value falls inside or outside its reference interval. Relevant information may include relationships between biomarkers, consistency with previous findings, the persistence of abnormalities, and changes during treatment or recovery.

At the same time, laboratory results must not be used in isolation to infer a patient’s behavior, assign stigmatizing labels, or make unsupported statements about substance use. The role of Aima Diagnostics is to structure and communicate relevant laboratory information cautiously, while leaving diagnosis and clinical assessment to qualified professionals.

This use case illustrates why safety, context, longitudinal analysis, and clinician oversight are central to the development of the platform.

8. A Value Layer for Diagnostic Equipment and Point-of-Care Analyzers

Aima Diagnostics is designed to provide an additional software-based interpretation layer for manufacturers of blood analyzers, including point-of-care systems.

An analyzer produces measurements, but the value delivered to the end user can extend beyond measurement. Manufacturers may be able to combine analytical hardware with structured interpretation, longitudinal reporting, and patient- or clinician-facing explanations.

The integration model includes two principal routes:

1. Integration with new equipment. The interpretation capability can be incorporated into a broader product offering and supplied alongside a new analyzer or its associated digital environment.
2. Integration with installed equipment. The interpretation layer can be connected to existing analytical systems, laboratory infrastructure, or associated software, subject to technical compatibility and configuration-specific validation.

For point-of-care testing, where results may be generated outside traditional laboratory settings, clear reporting and appropriate escalation pathways are especially important. Each integrated configuration must nevertheless be evaluated independently: analytical compatibility, data mapping, intended use, and local regulatory requirements cannot be assumed from one deployment to another.

9. Toward a Shared Evaluation Standard

Aima Diagnostics aims to contribute to the development of a shared, independently verifiable evaluation standard for AI-assisted interpretation of blood test results.

This standard should not be based on marketing claims, visual report quality, or user engagement alone. Four minimum evaluation dimensions are proposed.

9.1 Clinical concordance

System-identified abnormalities and biomarker relationships should be compared with assessments made by qualified clinicians using a prespecified protocol. Evaluation should include appropriate performance measures, confidence intervals, and assessment of agreement between clinical reviewers.

Where a study evaluates diagnostic performance against an independent reference standard, sensitivity, specificity, and other diagnostic-accuracy measures may be appropriate. Agreement with a clinician panel should not, by itself, be described as diagnostic accuracy.

9.2 Readability and comprehension

Reports should be tested with patients, not only assessed with automated readability formulas. Comprehension studies should measure whether patients correctly understand significant findings, uncertainty, and recommended next steps.

9.3 Clinical, behavioral, and workflow outcomes

Evaluation should examine consultation follow-up, appropriateness of repeat testing, clinician review time, patient understanding, and potentially unintended consequences such as anxiety or unnecessary investigation.

9.4 Safety and failure analysis

Developers should maintain a structured taxonomy of errors, including omissions, false relationships, misleading phrasing, inappropriate reassurance, and failures of escalation. Error rates should be reported with clear denominators and consideration of clinical severity.

Studies involving large language models should be reported with reference to TRIPOD-LLM, which addresses transparency in the description of LLM-based research, reproducibility, evaluation, and human oversight (Gallifant et al., 2025). TRIPOD-LLM is a reporting guideline, not a substitute for an appropriate study design or independent validation.

Where Aima Diagnostics is evaluated as a diagnostic system against an independent reference standard, STARD 2015 provides relevant guidance for reporting diagnostic-accuracy studies (Cohen et al., 2016). Its applicability should be determined by the actual study design and intended use of the system.

A shared evaluation standard must be reproducible and open to independent assessment. It should allow different systems to be compared using transparent methods and should require publication of unfavorable as well as favorable findings.

10. Evidence Status, Ethics, and Regulatory Considerations

This article describes the technology, intended value, implementation models, and evaluation program of Aima Diagnostics. It does not report the results of a completed controlled clinical validation.

The degree of evidence varies across applications. Technical deployment demonstrates that integration is feasible; it does not by itself demonstrate clinical benefit. Commercial observations can identify hypotheses for further study, but they cannot establish improvements in comprehension, decision quality, or patient outcomes.

Responsible implementation requires explicit consent where applicable, data minimization, secure data handling, patient access to retained information, and appropriate data-export mechanisms. Deployment must comply with the regulatory and data-protection requirements of each jurisdiction.

The general principles of human oversight, transparency, accountability, and responsibility described by WHO remain relevant to the governance of Aima Diagnostics (World Health Organization, 2021). For systems using generative or large multimodal models, WHO’s 2024 guidance provides additional recommendations concerning their development and deployment in healthcare (World Health Organization, 2024).

In the United States, the FDA’s January 2026 guidance on Clinical Decision Support Software is relevant to the analysis of clinician-facing functions. However, its non-device CDS framework should not be generalized to patient-facing reports without a separate assessment. FDA guidance distinguishes software intended for healthcare professionals from software intended for patients or caregivers (US Food and Drug Administration, 2026). The regulatory status of Aima Diagnostics therefore depends on the specific functionality, intended use, claims, users, and deployment configuration.

AI-assisted interpretation systems also require continuing review after deployment. Changes in laboratory methods, data formats, user populations, or language-generation components may alter performance. Monitoring, version control, incident reporting, and periodic reevaluation are therefore necessary.

11. Conclusion

The clinical value of diagnostic testing depends not only on measurement quality, but also on whether relevant information can be understood, placed in context, compared over time, and reviewed safely.

Aima Diagnostics seeks to provide this interpretation layer for patients, clinicians, laboratories, clinics, and diagnostic-equipment manufacturers. Its multidisciplinary expertise, commercial implementations, research collaborations, and work in complex clinical environments provide a foundation for further development and evaluation.

The long-term objective is not simply to produce more accessible reports. Aima Diagnostics aims to help establish an industry-wide approach to AI-assisted interpretation that is understandable, clinically governed, measurable, and independently verifiable.

A gold standard will not be established by declaration. It must be built through transparent methods, multidisciplinary collaboration, published evaluation, and continuing evidence of safety and usefulness.

About Aima Diagnostics

Aima Diagnostics was founded by Sergey Smirnov. The company is registered in the United States and Norway and conducts commercial and research activities in the United Kingdom, the United States, Scandinavia, Ukraine, and other European markets.

Aima Diagnostics brings together professionals in medicine and artificial intelligence and collaborates with laboratory networks, clinics, universities, research centers, and diagnostic-technology partners. Its mission is to make blood test results more understandable and useful while contributing to the development of a shared, independently verifiable evaluation standard for AI-assisted laboratory interpretation.

Declarations

Nature of the article: Industry and technology perspective.

Evidence status: No new patient-level data or completed controlled clinical-validation results are presented.

Commercial interest: The article describes technology developed and commercialized by Aima Diagnostics.

Data availability: The commercial implementation outcomes discussed in this article are derived from a publicly available Aima Diagnostics case study. No additional patient-level dataset is analyzed or made available.

References

1. Aima Diagnostics. (n.d.). Clinics & labs case study: How AI interpretation turned a one-time test into lifetime loyalty. Commercial case study.
2. Cohen, J. F., Korevaar, D. A., Altman, D. G., et al. (2016). STARD 2015 guidelines for reporting diagnostic accuracy studies: Explanation and elaboration. BMJ Open, 6(11), e012799. doi:10.1136/bmjopen-2016-012799.
3. Esteva, A., Robicquet, A., Ramsundar, B., et al. (2019). A guide to deep learning in healthcare. Nature Medicine, 25(1), 24–29. doi:10.1038/s41591-018-0316-z.
4. US Food and Drug Administration. (2026, January 29). Clinical decision support software: Guidance for industry and Food and Drug Administration staff.
5. Gallifant, J., Afshar, M., Ameen, S., et al. (2025). The TRIPOD-LLM reporting guideline for studies using large language models. Nature Medicine, 31, 60–69.
6. Nielsen-Bohlman, L., Panzer, A. M., & Kindig, D. A. (Eds.). (2004). Health literacy: A prescription to end confusion. National Academies Press.
7. Rajpurkar, P., Chen, E., Banerjee, O., & Topol, E. J. (2022). AI in health and medicine. Nature Medicine, 28(1), 31–38. doi:10.1038/s41591-021-01614-0.
8. Topol, E. J. (2019). High-performance medicine: The convergence of human and artificial intelligence. Nature Medicine, 25(1), 44–56. doi:10.1038/s41591-018-0300-7.
9. World Health Organization. (2021). Ethics and governance of artificial intelligence for health.
10. World Health Organization. (2024). Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models.