Clinical thinking of a cardiologist in the digital age

Full Text

RELEVANCE

Cardiology is undergoing a technological revolution. Algorithms analyze ECG in suspected acute coronary syndrome with 90% sensitivity and 92% specificity [1]. This matches the performance of experienced cardiologists. Automated systems calculate left ventricular ejection fraction with 92% accuracy, enabling classification of heart failure according to ESC 2023 into three categories: HFrEF (LVEF ≤40%), HFmrEF (LVEF 41-49%), and HFpEF (LVEF ≥50%) [2]. Clinical decision support systems analyze thousands of parameters and propose treatment protocols based on similar cases [3].

Algorithms identify correlations that are difficult to detect through visual interpretation. Monitoring systems are shifting medicine from a reactive to a preventive approach [2]. However, the risks are substantial. Physicians may blindly trust algorithms without critical analysis [3, 4].

Constant reliance on digital systems threatens atrophy of diagnostic skills that have been developed over decades [4].

A clinical example: A 58-year-old patient presented with retrosternal chest pain. An algorithm analyzed the ECG in 10 seconds and generated the output: "ST-elevation myocardial infarction, probability 94%." The cardiologist confirmed the diagnosis, assessed hemodynamic status, and directed the patient to the catheterization laboratory. Door-to-balloon time was 62 minutes (within the ESC 2023 guideline of <90 minutes). AI accelerated diagnosis, but the physician made the decision.

This raises the question: How will clinical reasoning change with AI implementation? Can cardiologists preserve independent judgment?

The answer determines the quality of care and the future role of physicians in healthcare [6]. The European Society of Cardiology and the Russian Society of Cardiology have established a position: the physician remains the ultimate decision-maker, with AI serving as an assistive tool (ESC 2023, RSC 2023) [8, 9].

AIM

To analyze the transformation of clinical reasoning in cardiologists during the implementation of digital technologies.

To determine an AI system integration model that expands physician competencies while preserving independent clinical judgment.

MATERIALS AND METHODS

We conducted a literature review covering 2020-2025. Sources included international peer-reviewed journals (PubMed) and Russian resources (eLIBRARY.RU, CyberLeninka). Searches were performed in Russian and English. Terms used included: clinical reasoning, cardiology, artificial intelligence in cardiology, clinical decision support systems, healthcare digitalization, hybrid intelligence in medicine, AI and ethics in cardiology.

RESULTS

Our literature analysis begins with the question: What constitutes clinical reasoning in physicians?

Physicians synthesize observations, knowledge, experience, and intuition. Research demonstrates a dual-process cognitive model distinguishing between fast (intuitive) and slow (analytical) thinking [4, 5].

A junior cardiologist employs an analytical approach, systematically analyzing data through differential diagnostic schemes [5]. This requires substantial cognitive resources.

An experienced specialist makes diagnoses instantaneously, relying on mental patterns formed through thousands of hours of practice [4].

Artificial intelligence systems in cardiology are based on analysis of massive clinical datasets. Machine learning algorithms process thousands of ECGs and identify patterns associated with specific diagnoses. AI trains on large samples and achieves high accuracy in classification tasks.

However, physicians make decisions under conditions of uncertainty, incomplete information, and ethical dilemmas. AI processes data faster than humans but lacks critical thinking [5, 7].

Despite high accuracy (90-95%), algorithms do not account for clinical context and ethical considerations. Example: AI recommends aggressive chemotherapy for an 85-year-old cancer patient because statistically this prolongs life. A physician considers quality of life, comorbidities, patient preferences, and may choose a palliative approach. AI does not comprehend human suffering.

We believe the optimal AI integration model is hybrid intelligence. Physician and algorithm work collaboratively. AI performs routine data analysis, identifies patterns, and generates preliminary conclusions [3].

The physician integrates this information with the patient's clinical context, evaluates ethical aspects, and makes decisions according to ESC 2023 and RSC 2023 recommendations [8, 9]. Physicians must understand the algorithm's principles, limitations, and circumstances when it can be trusted [5, 7].

A concrete example of hybrid intelligence: AI analyzes an ECG and generates the conclusion "myocardial infarction with 92% probability." The physician verifies the patient's clinical presentation (pain, history), confirms the diagnosis, and prescribes treatment. AI proposes a reperfusion protocol based on 10,000 similar cases. The physician adjusts the protocol accounting for the specific patient's renal insufficiency.

The physician makes the decision, but AI expands their capabilities.

Russia is actively implementing AI technologies in healthcare. By the end of 2024, medical organizations in 84 Russian regions employ 412 AI-based solutions [9]. These systems analyze medical images, interpret ECGs, and predict cardiovascular event risk according to algorithms based on RSC recommendations [9, 10].

The regulatory framework in the Russian Federation prohibits autonomous functioning of AI in clinical practice [9, 10]. Systems are designed as assistive tools.

CONCLUSION

Our analysis demonstrates that clinical reasoning in cardiologists is not disappearing in the era of digitalization. It is undergoing profound transformation.

Reasoning remains the foundation of practice, but its methodology is changing. Physicians are no longer the sole source of knowledge but are becoming more critical. They serve as integrators of knowledge with the lived reality of individual patients.

The perspective for 21st-century cardiology is physician-technology collaboration. The future cardiologist will utilize AI to expand diagnostic and prognostic capabilities while preserving independent judgment and responsibility according to RSC 2023 recommendations. The physician must remain the decision-making subject.

About the authors

Kristina Andreevna Sadovski

Belgorod State National Research University

Email: mafin.alen@yandex.ru
ORCID iD: 0009-0002-5409-4981
Russian Federation, 85, Pobeda str., 308015, Belgorod, Russian Federation

Vera Aleksandrovna Markovskaya

Belgorod State National Research University

Author for correspondence.
Email: markovskaya@bsuedu.ru
ORCID iD: 0000-0003-4410-9318

PhD in Biology, Assoc. Prof., Head of the Department of Pathology

Russian Federation, 85, Pobeda str., 308015, Belgorod, Russian Federation

References

Supplementary files

There are no supplementary files to display.