In May-July 2020, the Webiomed team within the framework of a joint pilot project of the National Medical Knowledge Base association with the Murmansk region conducted a study. The topic was as follows: “Assessment of medical data to determine risk factors and perform cardiovascular risk stratification in the Murmansk Regional Clinical Hospital named after P. A. Bayandin".
An automated cardiovascular diseases (CVD) risk assessment was performed on a sample of 1805 people. The distribution in this sample was distributed as follows: 61% men, 39% women. The age varied from 18 to 96 years old, the average age was 62 years. 57.6% of patients were aged from 40 to 65 years.
Webiomed CDDS was able to successfully estimate the risk of CVD in 78.8% of the records in the dataset. It was impossible to estimate the remaining 21.2% due to the lack of features for calculation.
The assessment of the CVD risks by Webiomed CDDS has an integrated approach, which lies in risk analysis using machine learning models and the most common risk scores.
The study led to the following conclusions:
- The overall risk assessment using a variety of scales and models evaluates CVD risk more fully in comparison with the SCORE.
- 550 patients with a full set of data required for CDV risk assessment (30.47%) represent a very high-risk group. The likelihood of developing serious cardiovascular diseases and their complications in such a group is very high.
- Among patients with high and very high CVD risk, 1 patient has a low absolute risk on the SCORE, and 62 - a moderate risk. These are direct reserves for the fight against CVD since physicians in their practice are most often focused on the assessment of CVD risk according to SCORE, and if such an assessment has a low or moderate risk, such patients may not be taken into account from the point of CVD prevention.
- The reason for the impossibility of assessing risks in some cases was the lack of data in medical records, which may be a reserve for strengthening preventive measures among the population, whose age characteristics (average age 62 years) indicate the maximum attention to primary prevention.