Project

The purpose of the project is to develop "Hippocrates" expert system for the support of physician's decision-making, the system will be used at the stage of diagnosis, prevention and treatment of various diseases.

The introduction of "Hippocrates" system into clinical practice will allow to personalize the approach to each patient, reduce the risk of medical errors and clinical complications.

A feature of the basic technologies that underlie the project and determine its novelty is the combination of technology of deep machine learning while processing large amounts of data with the United Medical Knowledge Base (UMKB). The UMKB is based on the systems of classifiers, medical ontologies, a unique model of knowledge representation and algorithms similar to doctor's thinking.

The main advantage of "Hippocrates" is that the system retains the ability to operate with a lack of baseline data about the patient, and that is quite often encountered in clinical practice. In such cases, knowledge derived only from evidence-based medicine is often not applicable for decision- making. This requires fundamental knowledge in the field of medicine and the use of methods of deduction. This is what the doctor does when taking decisions with insufficient baseline data, he uses the deduction. Therefore, we have built the algorithms of our system similar to doctor's thinking. That's why we train the system step by step, we fill the knowledge base layer by layer, like a student is taught at a medical school. These are medical ontologies, classifiers, systematic knowledge in the field of anatomy, physiology, human pathophysiology, and, of course, clinical knowledge accumulated on the basis of evidence-based medicine.

How does it work? The system is integrated into the medical information system of a clinic (MIS) , it analyzes the electronic health records of patients (EHRs) and tracks the treatment process in the background mode. Based on the EHR the system forms a virtual image of a patient in a semantic representation. Depending on the amount of data received at a certain level, the reactivity and resistance of the organism are simulated. In this model, possible pathological processes are started in correspondence with the clinical picture described in the patient's EHR. After analyzing these data, the system returns a preliminary diagnosis with recommendations for additional research to clarify the diagnosis, as well as recommendations for managing the patient. With each arrival of new data about the patient (diaries, research results, etc.), his virtual image optimizes. Upon reaching a certain confidence, the system issues a clinical diagnosis with recommendations for treatment and management of the patient, taking into account his personal characteristics. On his workstation the doctor sees the recommendations of the system, as well as information about how the system generated this conclusion with a detailed visualization of the intermediate stages of solving the problem. The use of expert system in clinical practice will allow to personalize the approach to each patient and reduce the risk of medical errors and clinical complications. The special technology used to extract information from medical texts will allow the system to understand the EHR in any format and run it on the basis of any MIS. At different stages of project implementation and as the UMKB is being filled, the following individual innovative products have been developed and are being developed:

1. Electronic Clinical Pharmacologist (ECF) is a system for supporting the decision-making of a doctor for prescribing pharmacotherapy. The ECF is integrated into the medical information system of the medical institution, it keeps track of drug prescriptions in the background mode and issues recommendations on the doctor's workstation.With the use of ECF in clinics, the costs of the medical institution for the purchase of medicines are reduced through more rational prescriptions of the doctors, the risk of complications and side effects of medications are reduced, the duration of doctor's reception time is decreased, and the quality of medical care is increased. The ECF system has been developed, successfully clinically tested and is being used in medical institutions. More information about the product can be found at: ecp.umkb.com

2. The system of "smart" electronic prescriptions and control of the distribution of medicines (PHARMTAXI) - a unified network that unites various participants in the pharmaceutical industry (doctors, patients, pharmacies, pharmaceutical companies and medical institutions) into a special logistics and allows you to control information and material flows simultaneously. On the one hand, electronic prescriptions are exchanged between the links of the chain "doctor ---> patient ---> pharmacy", and on the other hand, the distribution of medicines through the supply chain "pharmaceutical producer ---> pharmacy ---> patient" is controlled. This makes it possible to check the efficiency and safety of the drug therapy prescribed to the patient, as well as the authenticity of the medications which the patient is taking.The system has been developed and is being piloted in the regions of the Russian Federation. More details about the product can be found at: pharmtaxi.com

3. Electronic Therapist (ET) isa decision support system for diagnosing diseases. Based on the patient's complaints, anamnesis and the results of laboratory research, the system determines the probable pathology, makes a preliminary diagnosis and forms an epicrisis in order to refer the patient to a relevant specialist. The "Electronic Therapist" system will function in polyclinics and can be integrated into user interface of the Unified Medical Information and Analytical System. This will allow a patient to undergo primary diagnostics and make an appointment with the necessary specialist, thus relieving the district therapists. This product is being developed.

4. The system for predicting the risks of diseases and complications (RPS). The system will determine the risks of the development of diseases and complications in the user, taking into account the individual characteristics of his organism.In case of increased risk, the system notifies the user about the likelihood of a dangerous disease and issues certain recommendations. For example, recommendations for the research to exclude the likelihood of a dangerous disease and recommendations on the methods of prevention to reduce the risk of developing the disease. The integration of the system into social networks will allow to identify users with increased risk of cardiovascular diseases and cancer in the background mode. The users leave a huge amount of information in social networks, some of which is helpful for this screening, for example, gender, age, habits, region of residence, lifestyle, etc. After asking a few more leading questions, the system will perform a full screening to identify a risk group for the particular user. Regular mass screening of the population will increase the likelihood of identifying patients at the early stages of the disease.Thus, timely prevention and early treatment will reduce mortality from cancer and cardiovascular diseases.

5. The expert system for personal medical tracking of a user- "Personal doctor." The system will provide services of a personal physician to any user, and if necessary, accompany the person from birth. Remembering the individual characteristics of its user, the system evaluates the risks of developing the most common diseases (cancer, heart attacks and strokes, sudden cardiac death, allergic reactions, infectious diseases, etc.). At elevated risk, the system alerts the user and/or the attending physician of the likelihood of dangerous diseases and gives specific recommendations for prevention. On the basis of complaints of the patient, the system determines the likely pathology and makes an appointment for the patient with a relevant specialist or the attending doctor. Tracking the performance of prescriptions of the attending doctor (taking medications, regimen, examinations) the system notifies both sides about violations.

Market

The volume and dynamics of the market: the domestic market for Decision Support Systems (DSS) is gradually beginning to develop and, according to targets of the HealthNet road map, the revenue in this market is expected to reach $ 23.6 billion.

Today, the research and development of Medical Information Systems (MIS) and expert systems are actively being conducted. Most MISs, installed in clinics, solve issues of documents circulation and reporting.

The vast majority of expert systems used in health care today are of a strictly specialized nature. Their use, as a rule, is limited to a narrow circle of doctors, specialists in the relevant fields, and is aimed at highly specialized diagnostic tasks, systematic monitoring and collection of targeted information and training of doctors.

There are online services aimed at determining the disease by its symptoms (http://simptomus.ru, bots in the Telegram). The last stage of development of expert systems in health care is directly linked to the development of big data as applied to the accumulated medical knowledge. In this niche the closest analogue of the proposed Hippocrates system is IBM Watson Health.

The global CDSS market is expected to reach $ 21B by 2020, an average growth of 25% during the forecast period (2014 to 2020). Drivers of growth will be the increasing demand for analytical IT solutions that will become an effective tool to address the issue of medical errors and make the right decisions when providing medical care which in turn will help to reduce the costs and increase the quality of medical care (by: marketsandmarkets.com). The key players in this market are as follows: Agfa Healthcare (Belgium), Athenahealth, Inc. (USA), Allscripts Healthcare Solutions, Inc. (USA), Carestream Health, Inc. (USA), Cerner Corporation (USA), Epic (USA), GE Healthcare (UK), McKesson Corporation (USA), MEDITECH (USA), NextGen Healthcare Information System LLC (USA), Novarad Corporation (USA), Philips Healthcare (Netherlands), Siemens Healthcare (Germany), Wolters Kluwer (USA), and Zynx Health (USA).

Regarding the pharmaceutical market, it is expected that by 2021 global spending on prescription drugs will reach about $ 1.5 trillion per year, or $ 370 billion more than in 2016 (IMS data). In view of these forecasts, we believe that in the coming years without the DSS no competitive solution is possible in the field of MIS.

The main consumers: b2b - general and highly specialized medical facilities (polyclinics and hospitals), pharmacies, MIS manufacturers, pharmaceutical companies, medical insurance companies; b2c - doctors, patients, pharmacists; b2g - government agencies- MIS integrators.

Countries and regions: Today, the system of classifiers of UMKB is mainly implemented in the Russian language and the knowledge base is filled from Russian sources. However the platform is multilingual and can support complementarity with foreign classifiers. Today, the complementarity between the objects of the UMKB classifiers and the terms of the international reference book SNOMED-CT has reached 60%. After completing this process, the system will be able to analyze the medical literature in English, which will allow you to download information from external sources, such as Elsevier, FDA (FAERS), NICE, Clinical Trials.

To scale the ECP and PHARMTAXI projects to the international market, it is sufficient to add the trade names of the drugs registered in the required country, since the work of the main algorithms and semantic links are built around active substances and international non-proprietary names (INN).


Company

JSC "Socmedika" is an IT company, a resident of Skolkovo innovation centre, specializing in creating expert systems in the field of medicine.

In 2011, a group of researchers from Bakulev Center for Cardiovascular Surgery of the Russian Academy of Medical Sciences proceeded to the creation of the United Medical Knowledge Base (UMKB) for using it in medical expert systems.To structure medical knowledge, it became necessary to develop an original model of knowledge representation and medical ontologies. And also to develop a flexible system of medical classifiers.

In early 2012, a group of enthusiasts was formed, who became shareholders of the company, a team of programmers was recruited, a brand and a logo of Socmedika company were created.

The key task was to create an instrument for distant modeling of medical knowledge and attract doctors and experts for working with the knowledge base. This enabled the doctors to fill the base at any time, no matter where they were.

Later on, came the idea to create an Internet portal through which doctors of different specialties could participate in the project on creation of the UMKB.

In the course of filling the knowledge base, the idea of creating UMKB was supported in the scientific community. Thus, in August 2013, the UMKB project was launched and research centres were engaged to contribute to the creation of the United Medical Knowledge Base, each in its own field.

Today, our technologies allow us to create expert systems in the medical field in the shortest possible time.

To implement the Hippocrates system, a number of basic technologies are needed, most of which have already been implemented by JSC Soсmedika  and are described in detail on our website. Of these, the most important technologies are the collection, formalization, integration and updating of the medical knowledge base that underlies the system being developed.

In order to fill the knowledge base and constantly update it, the system uses deep machine learning technologies (artificial intelligence methods) in combination with crowdsourcing technologies, connecting the expert community. The key objective of this project is to create a way to extract meaning from medical texts.

Today, there are a lot of technologies for linguistic analysis of texts, however, analysis of the text at the level of only linguistic rules has not been sufficient to correctly extract the facts from the medical literature. To that end, the system should be based on the necessary basic knowledge in the field of medicine.

That is why our team "teaches" the cognitive expert system step by step, filling the knowledge base layer by layer, similar to how a student is taught in a medical school. The system has accumulated basic information - medical ontologies, classifiers, systematized knowledge in the field of anatomy, physiology and human pathophysiology. This approach will increase the accuracy of determining the meaning when analyzing the text up to 95%.

Investment opportunities

In the process of developing the main product "Hippocrates", the following stages have been completed up to now:

1-st stage. Basic technologies were developed (2012-2015)

  • Basic technologies of modeling, accumulation and integration of medical knowledge were developed.
  • United Medical Knowledge Base - UMKB was created.

2-nd stage. The first expert systems based on UMKB were developed (2015-2016)

  • The expert system of doctor decision support on pharmacotherapy "Electronic Clinical Pharmacologist" (ECP) was developed.
  • A system of "smart" electronic prescriptions and control of the distribution of medicines was developed - PHARMTAXI.
  • Integration of ECP into medical information systems (qMs, Intersystem, etc.)
  • Clinical approbation of the ECP and the first sales of the ECP.
  • To complete the development of the final product, an investment of $ 3.5M is required, which will be spent on the following stages:

3-rd stage. Development of an expert system for diagnosis of diseases "Electronic Therapist". The required investment is $ 1.2M.

4-th stage. Development of an expert system for forecasting the risks of diseases and complications. The required investment is $ 1M.

5-th stage. Development of an expert system for personal medical tracking of a user - "Personal Doctor". The required investment is $ 1.3M.

The company enters self-financing in 8 months after receiving the required investment. In two years from the beginning of financing the net profit will be about $ 0,7M a month.


Team [27]

Gevorg Blejyants
Gevorg Blejyants

General director. A cardiovascular surgeon, the candidate...

Julia Isakova
Julia Isakova

Head of the expert group on modelling UMKB in pharmacology...

Mikael Abgaryan
Mikael Abgaryan

Head of department of collaboration with medical institutes...

Tumanov Nikolay
Tumanov Nikolay

Executive Director. Psychiatrist, candidate of medical sciences...

Artem Panosyan
Artem Panosyan

IT Director. A programmer. In 2001 he graduated from the Moscow...

Muslim Guseynov
Muslim Guseynov

Chief programmer. Career started with making small programs...


Presentation