AI-enabled predictive analytics and risk assessment in healthcare

D. Gavrilov; L. Serova, I. Korsakov, A. Gusev, R. Novitsky, T. Kuznetsova CARDIOVASCULAR DISEASES PREDICTION BY INTEGRATED RISK FACTORS ASSESSMENT BY MEANS OF MACHINE LEARNING // Vrach.2020.-№5-P.41-45 Copy

Aim. To develop a model by machine learning to predict the risk of cardiovascular diseases (CVD) and validate the model using Russian medical data. Materials and methods. The data set was obtained from the Framingham study, consisting of 4,363 patients without CVD, 852 (19.5%) of which died of myocardial infarction and stroke within 10 years of observation. Incoming model features: gender, age, systolic blood pressure, cholesterol, smoking, body mass index, heart rate. The original data set was divided into 2 parts: the training data set (80% of the records) and the validate data set (the remaining 20%). Additionally, the model was evaluated by an external data set included 411 depersonalized patient data from the Russian citizens. Results. The WML.CVD.Score model was created by the serial neural network with one input, two hidden and one output layer. Accuracy results on a training dataset: Accuracy 81.15%, AUC 0.80. The same indicators on the validate data set were: Accuracy 81.1%, AUC 0.76. Test results for the test data set: Accuracy 79.07, AUC 0.86. On the Russian test data, the AUC for the SCORE scale was 0.81 versus 0.86 for the developed model, which showed the validity of the use of machine learning in order to increase the predictive model. Conclusion. The developed model has demonstrated high accuracy to CVD predicting in both internal and external validation. The model can be used in medical practice for patients in Russia.

Gavrilov D., Gusev A., Korsakov I., Novitsky R., and Serova L., "Feature Extraction Method from Electronic Health Records in Russia", in Proceedings of the FRUCT’26, ISSN 2305-7254, ISBN 978-952-69244-2-7, pp. 497-500, April 2020 Copy

The medical language is the basis of the electronic medical records (EHR), and up to 70 percent of the information in these records were writing in natural language, in the free text part. The last few years have seen a surge in the number of accurate, fast, publicly available name entity recognition (NER) parsers. At the same time, the use of NER parsing in natural language processing (NLP) applications has increased. It can be difficult for a non-expert to select a good “off-the-shelf” parser. We present a method of using statistical NER parsers on a medical corpus of Russian. We developed a new tool that gives a convenient way to extract NER from unstructured medical documents.

Korsakov I., Gusev A., Kuznetsova T., Gavrilov D., Novitskiy R. Deep and machine learning models to improve risk prediction of cardiovascular disease using data extraction from electronic health records // European Heart Journal, Volume 40, Issue Supplement_1, October 2019, ehz748.0670 Copy

Advances in precision medicine will require an increasingly individualized prognostic evaluation of patients in order to provide the patient with appropriate therapy. The traditional statistical methods of predictive modeling, such as SCORE, PROCAM, and Framingham, according to the European guidelines for the prevention of cardiovascular disease, not adapted for all patients and require significant human involvement in the selection of predictive variables, transformation and imputation of variables. In ROC-analysis for prediction of significant cardiovascular disease (CVD), the areas under the curve for Framingham: 0.62–0.72, for SCORE: 0.66–0.73 and for PROCAM: 0.60–0.69. To improve it, we apply for approaches to predict a CVD event rely on conventional risk factors by machine learning and deep learning models to 10-year CVD event prediction by using longitudinal electronic health record (EHR). Acces: https://doi.org/10.1093/eurheartj/ehz748.0670

Gusev A. V., Gavrilov D. V., Korsakov I. N., Serova L. M., Novitsky R. E., Kuznetsova T.Yu. Prospects for the use of machine learning methods for predicting cardiovascular disease // Information technologies for the Physician. 2019 №3. p. 41-47 Copy

Morbidity and mortality from cardiovascular diseases (CVD) has remained the leading rate in recent decades worldwide. Primary prevention methods based on the management of cardiovascular risk factors are most effective in reducing the burden of CVD. In preventive medicine for risk management of CVD use the riskometers – scales that was obtained as a result of long prospective studies. But the practical application of the developing scales has showed the limitations in the forecast accuracy. Machine learning makes it possible to improve the accuracy of cardiovascular risk prediction due to nonlinear relationships of their deeper adjustment between risk factors and disease outcomes. 2236 patients’ data were used. We trained the model on the features used in the Framingham scale construction. We compared the resulting model and the Framingham scale for the accuracy of the cardiovascular event prediction. Thus, according to the ROC analysis for the Framingham scale, the indicators are as follows: precision Accuracy: 70,0%, the AUC: 0.59. At the same time for the model obtained using machine learning similar indicators were: Accuracy: 78,8%, AUC: 0.84. Thus, the use of machine learning algorithms including deep learning algorithms can significantly improve the accuracy of cardiovascular risk prediction of trained models

Gusev A.V., Kuznetsova T.Yu., Korsakov I.N. Artificial Intelligence in Risk Assessment of Cardiovascular Diseases // Journal of Telemedicine and E-Health. 2018. No. 3 (8). S. 85-90. Copy

Cardiovascular disease (CVD) is one of the leading causes of death and disability in most countries of the world, including Russia. According to Russian Federal State Statstics Service (Rosstat) in 2016, 904 thousand people died from diseases of the circulatory system in Russia, which amounted to 47.8% in the structure of the causes of mortality. A significant proportion of this morbidity and mortality could be prevented through early diagnosed and prevention strategies, both for people with established disease and for those at high risk of developing disease.

Gusev A.V., Pliss M.A. The main recommendations for the creation and development of information systems in healthcare based on artificial intelligence // Doctor and information technology. 2018. No3. - Page 45-60 Copy

Artificial intelligence is becoming one of the main drivers in solving serious problems of medicine and health, such as inadequate resources, further improving efficiency, quality and speed of work. All over the world, more and more solutions are being developed in this area. However, the more new products appear, the more questions and problems arise. The work analyzes some foreign publications and research results, which studied the main problems associated with the creation and implementation of artificial intelligence in health care. As a result of the analysis, a number of practical recommendations were formulated that will help increase the likelihood of successful creation and introduction of such products in the practical link of health.

Gusev A.V., Dobridnyuk S.L. Artificial Intelligence in Medicine and Health Care // Information Society, No. 4-5, 2017. pp. 78-93 Copy

The article discusses the prospects for using artificial intelligence technologies in medicine and healthcare. The history of artificial intelligence development is described. The technologies of machine learning and neural networks are analyzed. The review of already implemented artificial intelligence projects is provided. The forecast of the most promising directions of the artificial intelligence technologies development, according to the authors, for the near future is given.

Gusev A.V. Prospects for neural networks and deep machine learning in creating solutions for healthcare // Physician and information technology. 2017. No3. Page 92-105 Copy

The paper gives an overview of the prospects of using neural networks and deep machine learning in the creation of artificial intelligence systems for healthcare. The definition and explanations on the technologies of machine learning and neural networks are given. The review of already implemented artificial intelligence projects is presented, as well as the forecast of the most promising directions of development in the near future

Gusev A.V., Zarubina T.V. Support for the adoption of medical decisions in the medical information systems of a medical organization // Doctor and information technology. 2017. No2. Page 60-72 Copy

In the article the review of various possibilities of support of acceptance of medical decisions in medical information systems of the medical organizations is presented. The description of functional requirements and prospects in terms of increasing the effectiveness of medical information systems in the informatization of clinical work of doctors