The use of historical data from the market of the region under consideration
The use of specialized and proprietary natural language processing libraries
1
Regular internal expert medical assessment of classification correctness
2
3
At MAINS LAB, we have been working with data on provided medical services for years, and over the past 5 years, we have accumulated tens of millions of data records that very accurately describe treatment schemes and the frequency of service assignments in different countries. However, using this data "directly" is not possible due to the various taxonomies used to denote services and diagnoses.
At MAINS LAB, the system for unifying medical services is based on:
2. How Can Big Data Help?
Examples of such processing include:
The service "CT of pelvis with IV contrast medium" may be referred to as:
· ct pelvis without contrast-13074
· c.t. pelvis with cont
· ct pelvis + contrast
· pelvis scan, etc.
The service "Measurement of LDL cholesterol" may be referred to as:
· ldl
· cholesterol
· l.d.l.c
· low-density lipoprotein cholesterol (LDL cholesterol), etc.
We have done extensive work to clean and unify the data, bringing service names and diagnosis codes into a consistent format, which has opened access to previously hidden frequency patterns. For example, based on the data, we can say that prescribing the antibiotic Klavox (Amoxicillin) for colds is necessary in only 5% of cases. Statistically significant deviations from this figure in the higher direction may indicate unethical practices on the part of the clinic or physician.