Background. Mobile Health (mHealth) is a hot research and development topic due to massive spread of smartphones and their high technical performance. Mobile apps for health vary in their functions, from dietary assistance and planning physical activities to “take a drug” reminding and others.

Aim: Assessing the capabilities of existing mobile apps for collecting personal medical data (lab results, radiography scans etc.).

Materials and methods. We performed a systematic search and analysis of mobile apps, designed for storing medical data. We surfed through the App Store and the Google Play, research articles were searched at PubMed, Embase, eLIBRARY.

Results. This review covers mobile apps for storing and analysis of patient’s medical data. Such software products are used to store health information in one place. The most promising are applications that digitize data and analyze them. Dynamic monitoring of lab results makes it possible to control chronic diseases, track changes in the body, diagnose a disease and evaluate prognosis.

Conclusions. The use of medical mobile apps in the healthcare system will reduce economic costs and at the same time increase the availability of medical care in segregated areas.

Introduction. Cardiovascular diseases remain the leading cause of death globally due to the global trend of aging. Mobile medicine — mHealth — is gaining more popularity each year. In this paper, we consider the effectiveness of mHealth in the prevention of cardiovascular events.

Materials and methods. The study was performed in accordance with PRISMA checklist. Original studies published in 2018- 2022 were considered for inclusion in systematic review. Еlibrary, PubMed, Scopus, Google Scholar и ResearchGate were searched for the studies.

Results. Systematic review included 15 original clinical studies. Number of studies per year was as follows: 2021 — 47%, 2020 — 40%, 2019 — 13%. Sample size in the studies varied from 28 to 28 189 people (median — 333). Age of participants ranged from 45 to 68.5 years (М±SD = 59.9±2.1 years), follow-up period was 1.5–36 months (М±SD = 9.4±2.5 years). Reliability of studies results was 0–100% (median — 69.7%). The proportion of studies with reliability of 100% was 46.7%, less than 50% — 26.7%. No effectiveness from the use of mHealth was found in 6.7% of the studies, with 13.3% of included studies didn’t evaluate effectiveness at all. The majority of studies came from USA (20%), United Kingdom (13.3%), China (13.3%). Others came from Russia, Belgium, Germany and Australia (6.7% each). Artificial intelligence was used in 13.3% of the studies.

Conclusion. This systematic review demonstrated significant benefit from using mHealth for prevention of cardiovascular diseases compared to standard approaches.

The paper covers international experience in regulating the use of medical data for the development of artificial intelligence systems (AI) using machine learning methods. High-quality medical data sets are required for successful implementation of AI in medical practice and for higher efficiency of clinical and managerial decision-making. Such data sets are impossible to acquire, store and use without appropriate legal and regulatory framework that takes into account the interests of all participants at each stage of the development and use of AI.

The review of foreign legislations was carried out for the countries — leaders of the macro-regions, which were selected based on the higher metrics of the AI market. Today, there are different approaches to protecting medical data, with the most well-known being industry and cross-industry approaches (USA and EU respectively). In order to keep a proper balance between patient safety and the possibility of collecting medical data for developers, a regulatory framework for both crossbeing industry and cross-industry regulation needs to be formed.

The article presents originally developed intelligent decision support tools designed for precancerous lesions and tumors of the oral mucosa diagnostics.

Background. There is well recognized need for better primary health care in patients with suspected maxillofacial tumors.

Aim: develop intelligent decision support tools for tumors of the oral mucosa diagnostics.

Methods. Developed method was based on generalization of the practicing doctors experience. Diagnostic parameters of the disease were studied during the first stage of the development. All diagnostic parameters were divided into three large groups: patients complaints, patient’s work-up, risk factors and patients lifestyle. At the next stage of analysis each data group was presented as a parameter set. Each parameter set and each singular parameter were assigned a weight coefficient by expert evaluation method. The sum of the weighting coefficients for each group of parameters and for each singular parameter was equal to one for convenience. Use of these coefficients enabled transition made to indicators that allow assessing the expectation of confirmation of the expected prognosis.

Results. Using weight data we developed production knowledge models and implemented the fuzzy set technique, which resulted in models that allow assessing the degree of confidence in the diagnosis. This approach will ensure that doctors get informed decisions summarizing the collective knowledge of medical experts. Such intelligent solutions can only be considered as a kind of hint to a specialist, instead of the only, uncontested option. As the system functions, the models will be refined, which will increase the efficiency of the expert system.

Conclusions. The results of the study propose a new approach to classification, as well as highlight the structure of the parameters allowing suspicion of a malignancy in patients. We also developed new formal diagnostic method of maxillofacial cancer which suggest further patient management. Production knowledge base for automated diagnosis of the oral mucosa pathologies was created.

Practical application. This research results could be used as an additional tool that allows the doctor to verify diagnosis or treatment strategy of patients.

Background. The development and implementation of medical information systems make it possible to simplify and automate many processes in medical organizations. At the same time, the amount of data on patients’ health is constantly accumulating which allows solving many problems related to the prediction and diagnosis of diseases.

Aim. To study approaches to processing of Russian unstructured medical texts and to predicting certain groups of diseases based on machine learning methods.

Initial data consisted of an array of depersonalized data from medical organizations in the Orenburg region containing 119,780 records. Three approaches to probabilistic forecasting of groups of diseases based on unstructured medical texts of patient complaints in Russian were studied: rule-based approach, logistic regression-based approach and approach using BERT transformer models.

Results. Comparative analysis showed that показывает, logistic regression-based approach combined with TfidfVectorizer method had the best results in Precision (0,8296), F1-score (0,8269) and Matthews’s correlation coefficient (0,7695).

Conclusion. Traditional rule-based approach was the least effective (Precision = 0,7182) among the studied methods, but at the same time it allowed to interpret the results of the classifier as visualization of the decision tree. Logistic regression-based approach (Precision = 0,8296) and approach using BERT transformer models (Precision = 0,8164) showed the best classification results and can be further used as a basis for building and developing medical decision support systems and find application in medical practice.

Background. Strategic management of healthcare which, among all, includes modern digital technologies, has increasing importance in the light of non-communicable diseases burden.

Aim. To develop an tool which evaluates and predicts the digital maturity of healthcare, accounting for the strategic importance of combating chronic non-communicable diseases.

Materials and methods. We performed an analytical study, making a systematic review of international experience, analysis and adaptation of the principles of monitoring digital maturity. The validity and reliability of the developed index was assessed, with international experts being on the team.

Results and discussion. A patient-oriented index of the healthcare system digital maturity has been developed and validated. For the first time this index was used to measure the system state and dynamics, and to predict the development of the healthcare digital transformation in Turkmenistan. Systematic analysis of maturity of the healthcare digitalization was also a first in Turkmenistan. Authors found positive dynamics of the system going from “low” status in 2018 to “developing” status in 2021. An optimal development scenario aiming at achieving “mature” and “innovative” statuses by 2026 was established using forecasting.

Conclusion. The validity and reliability of the patient-centered index of the healthcare digital maturity was 0.92 [95% CI 0.88; 0.94] and 0.91 [95% CI 0.87; 0.94], respectively.

Aim: to develop and test a methodology for assessing the maturity of healthcare software based on artificial intelligence (AI).

Materials and methods. The methodology for developing a maturity matrix for AI-based healthcare software is based on published data and on an analysis of our own practical experience obtained during the «Experiment on the use of innovative technologies in the field of computer vision for the analysis of medical images and further application in the Moscow healthcare system « in 2021–2022. We studied study results from 35 separate software products based on AI, covering key areas of radiology.

Results. We developed a maturity matrix that takes into account the indicators of technical stability — the proportion of technological defects, and the diagnostic component — the area under the characteristic curve. This model has been tested in 35 software products based on AI, with 40% of the products having achieved maturity. The dynamics of development was assessed for 24 software products based on AI: 15 of them (62%) were in the zone of diagnostic stagnation; 8 (33%) — in the zone of high diagnostic and technical potential, 1 (4%) — in the zone of low diagnostic and technical potential, and 1 (4%) worsened the technical component with the increase in diagnostic potential.

Conclusion. A methodology for assessing the maturity of AI for healthcare has been developed based on the performance and quality assessment of 35 software products. This methodology includes a maturity matrix and a method for assessing the clinical and technical transformation of maturity, which makes it possible to evaluate an AI-based software product both discretely (simultaneously) and in dynamics.