Modeling is the common approach for predicting not only the population health, but also the social and economic burden of disease, which is an important argument while making decisions in health care and prevention.
Aim: To develop the model for predicting cardiovascular risk, applicable for the assessment of clinical and socio-economic effects of preventive and therapeutic actions at the level of the whole population or part (region, city, group of patients).
Material and methods: An analytical model for making decision was performed by using a Markov model consisting of Markov states and probabilities of transition from one state to another within a certain time interval. The model included risk factors and cardiovascular diseases (blood pressure, cholesterol, smoking) and probabilities of transition between them. Data was standardized by age for both males and females. Multivariate sensitivity analysis was performed. The literature search conducted using eLIBRARY.RU (http://elibrary.ru) and CyberLeninka (http://cyberleninka.ru). Consultations with experts in the field of coronary heart disease, stroke, heart failure were carried out.
Results: The model, allowing to compare the outcomes of two scenarios (absence/presence of intervention). The model included risk factors: arterial hypertension, smoking, hypercholesterolemia, and important CVD: coronary artery disease, myocardial infarction, unstable angina, heart failure, chronic heart failure after myocardial infarction, transient ischemic attack, stroke, atrial fibrillation. There was absorbent state - death. At the output from the model the patient state was defined as the sum of the Markov states characteristics during the model time horizon. Each result had the cost and outcome, which values could be calculated by simulation modeling ("cohort simulation"). The data analysis from prospective study had shown that mortality increases with age, as expected, but in different age groups impact of cardiovascular causes was different and declined with age. In the case of the blood pressure there was the expected increase of the death risk with the growth of pressure levels, both for males and females, except for males 60-64 years old who had a minimal risk of death at the blood pressure 140-149/90-99 mmHg, and among males with normal blood pressure the risk was higher. Smoking was associated with an expected increase of the death risk among all age groups in both sexes. In males, aged 40-64 years, the death risk was higher at the normal levels of cholesterol (2-5 mmol/l), than at the cholesterol levels equal 5-7 mmol/l. There were no data sources to assess probability of occurrence of the risk factors (hypertension, smoking, hypercholesterolemia) in patients who did not have these factors previously in our studies, and available literature. This requires the prospective studies on at least two slices of surveys (not just with the endpoint analysis). Analysis of the literature on search of prospective Russian studies that would evaluate the probability of transition from one state to another, and consultations with experts have identified that currently conducted studies do not provide all the necessary probability of transition on the basis of national data. In the absence of local data for the model is acceptable to use the results of meta-analyzes of international studies.
Conclusion: Markov model will allow for prediction the effectiveness of different interventions, including their socio-economic consequences. The created model will allow in the future to make changes with the appearance of the results of new studies or new data in order to improve modeling accuracy.
Keywords: Markov model; cardiovascular risk; death; modeling; prognosis; social and economic damage.