06 November 2020

Digital Medicine in Russia

Is there a "figure" in medicine?

Olga Popova, it-world.ru

The Association of Developers and Users of AI in Medicine "National Database of Medical Knowledge" (NBMZ) conducts many projects, including the development of data sets, the formation of "sandboxes" within the experimental legal regime, cooperation with Russian manufacturers of medical equipment and the search for solutions and teams for them, pilot projects in clinics for SPD. Talking with the Director of Project work of the Association Andrey Almazov, we hoped to hear optimistic figures of the victories of Russian medicine in the digital world. But they heard something completely different.

At what stage is the digital transformation of the healthcare system in Russia now? Is it possible to compare with world indicators?

On basic... there is no digital transformation, there is a protracted attempt at informatization. In the world, the difference was visible even during our trip to HIMSS in 2012, and now what was then shown there as a prospect has already entered, if not into life, then into wide approbation. And we are all discussing how we can arrange electronic medical records and get statistics from primary documentation... And the world is developing bioinformatics, genome research and various devices and services that help to obtain information about the state of the body and analyze medical data. We are also trying, but we have a couple dozen such solutions per country, and until recently there was exactly one thing with registration certificates, if about the so–called AI systems, and if more broadly, a few more. But they are not used properly yet, it is extremely difficult to integrate them into the public health system.

And most importantly – we don't have a market! The state allocates money for informatization/digitalization/transformation, etc. of healthcare and develops them itself in the person of state-affiliated structures, the policy of closed admission of players prevails, no one is going to create a market, there is one state everywhere, and now look what is happening with the CHI and the admission of private clinics there. And in informatization, everything is even worse: there are continuous attempts of "single suppliers" on all issues, even that pathetic market of MIS (medical information systems – VM), which somehow appeared in 25 years, is now being finished off in the regions. But it's better to ask the manufacturers themselves about this, they will tell you everything in detail and justify everything.

Strongly. But is there a Single Digital healthcare Circuit? Who is responsible for the integration of systems within this digital circuit?

To be honest, it's most like trying to get money for the development of something that didn't work out the first time. To admit that the Unified State Health Information System (Unified State Information System in the Field of Healthcare – VM) has not yet been created in the planned form – it is impossible, it is impossible to allocate money for continuation, so it was necessary to replace the concepts, and this digital circuit appeared. Who is responsible for integrating the systems of other departments is a very interesting question, I don't know. But this is a serious task, because for effective management of the industry, it is required that not only general reference books are used in the system of the Ministry of Health, but in general there are a lot of intersections in social and public administration. And now ask any developer of so-called medical information systems how many reference books on one topic they have to "map" inside a system or an integration bus and what is the weight of the Ministry of Health's standards for other, related departments? Or why not link the Unified State Register of Legal Entities and the register of medical organizations, if each medical organization is included there and at least the issue of updating is removed? Actually, such a task, as I understand it, has been set within the framework of the ECDC, but why couldn't it be done right away, and not produce stuffing boxes in which it is difficult to ensure the availability of complete and up-to-date information.

What other problems are there with the figure in the regions?

The lack of funds and a single technological and organizational policy that is specified in detail in the regulations and regulations. We are, in fact, on the very edge of the abyss, into which regions and developers are already beginning to fall, and the Ministry of Health is no longer in control of the situation. Here and there local (regional!) riots, accusations and threats of criminal cases, refusals of developers to accompany the mess that eventually divorced. An attempt to impose a single supplier on everyone, and not brought to its logical conclusion and its full responsibility, but at the same time rhetoric about the integration of best practices, but without the ability for regions to independently determine technological policy, and against the background of their still weak competence in this. Only certain regions somehow cope with the task and deal with the truly objective problems of informatization (the complexity of implementation on the ground, chaos in processes, sabotage of those who do not want transparency and control of medical organizations, etc., etc.), and if at the same time they also lag behind in creating general and working NSIS and regulations in general, as well as to consider informatization as a separate entity from healthcare management, the complexity of the problem grows by two orders of magnitude.

And what about electronic medical cards, which have been talked about a lot lately?

It is unclear to whom and why they are needed, except to use text stuffing, subsequent printing and pasting into a paper card. The transition to a full-fledged EDI requires significant efforts on the ground, but there is no real need for it and sufficient resources. Therefore, all this is being sabotaged. And then, are you sure that we really want transparent statistics in the healthcare system from the EMC? No, we, as citizens, of course, want to... Well, probably the main thing is that there is no patient orientation here yet. It is declared that the patient will have access to his medical information, but so far it has not been done anywhere really and conveniently, secondly, the patient does not really need it now. This is necessary when the patient has switched to the ideology of controlling his health status from the ideology of "repair by appeals" and when it will be easy for him to change a doctor, hospital, insurance company, it is easy to get remote consultations. For all the other key players in the system, in fact, it is more or less unprofitable, and therefore the locomotive of such an introduction should be either a serious reform in the healthcare organization, (by the way, however, it is not a fact that this is a blessing – the system needs to be developed quietly and evolutionarily, the price of all reforms and shocks in this industry is significant and it is measured by lives, the responsibility is colossal, and here the principle of "do no harm" is also needed), or the growth of self-awareness and the demands of patients.

But there are systems to support medical and managerial decision-making. Are they developing somehow?

If we are talking about Russia, then "managerial" ones are practically not developing in any way, there is no clean data for them, and in general they are not accepted on the basis of real data, you can look at the example of the fight against covid. It has nothing to do with epidemiology and data, it's politics and social technologies. However, it is worth noting that, fortunately, we are not leaders… The problem with data has arisen all over the world due to the total incorrectness of primary data due to the politicized, social and economically conditioned interpretation of them at the time of their occurrence in statistics.

In this case, no support for management decisions by the systems that you have in mind here is required. Support systems for medical decision-making are based on a fairly large enthusiasm of researchers and scientists of doctors, of which we still have enough. The Russian land has not become impoverished. However, all their developments remain developments and do not pass into the category of "products". There is no entrepreneurial culture, there are few so–called technological entrepreneurs, and most importantly, once again, THERE is NO MARKET, and there are no Russian strategic players on it, so the lot of those who have grown up to the product is the sale of technologies in medicine by an international major and return to us as part of their ecosystems and platforms, not even solutions. At the appropriate price. Our Association, with the support of the existing development structures in the Russian Federation and with some support from the Ministry of Economic Development, is trying to offer a solution - a startup factory, a center for testing and acceleration within the experimental legal regime, where we could, as in an incubator, complete and nurture those most promising developments, bringing them to the product. We are also actively negotiating with the structures of Rostec Group on the topic of who, if not them, and being a leading strategic player in Russian and world medicine, and then all our developments will have an option to develop here, and not go to the West and to China. This is a serious comprehensive program, we have its achievements, but support is not yet sufficient, here we need to consolidate the efforts of the state and its actors in the person of not one, but several even state corporations. We are working on it, I can't say more yet. There are prospects, the pace is three times less than we would like and ten times more than necessary .... Two years ago, we were almost not far behind in the DSS market and were closer to the starting point together with the whole world, now we have moved one step, the world is ten, and the gap is rapidly growing. Two or three more years, and that's it... until the next technological break in 30-50 years.

Let's talk about what you do – artificial intelligence. How can he help the doctor?

Bad in any way, he is too self-confident and does not need advice. Therefore, only normative, rigidly from above, as a means of control. A good person can really help, as a hint, as a tool for processing and analyzing the huge amount of data and knowledge that he really needs now to remain a good doctor. And such doctors understand this well, but they need proven and complete solutions, but for now we are still dealing with an experimental phase. Full–fledged clinical studies on how AI can be integrated into practice and what effect, in fact, it will give - not only is there no, but it is not clear how to conduct them, we need a design of such studies, and this is not a simple story. But everything will be, moreover, AI is already being used in individual tasks. It's only worth mentioning that under AI, we have neural networks training here for image recognition and analysis of medical data, the so-called deep AI is still far away. However, what is already available can and should help doctors very much, at least – not to make a mistake, at most – to deal with a complex case and reduce the time and amount of information processing needed for this "manually".

What is data mining in medical programs?

Processing of large arrays of medical records and an attempt to draw conclusions from this about the risks or the presence of conditions and diseases that require intervention (for example, to further examine and take on dispensary registration) and analysis of medical images with an attempt to mark up suspected pathology. All this is further presented to the court of doctors and with good sensitivity (detection of pathology in principle) and specificity (its categorization and concretization), the number of false positive cases allowed by the doctor's workload can be a good help.

How is the problem of trust in artificial intelligence solved in healthcare?

Abstract trust in abstract AI cannot and should not exist. Trust should arise in specific products, this will, as is customary in medicine, be achieved through independent, full-fledged clinical trials. Trust will grow with the development of the practice of application and feedback in medical publications and at medical conferences, as is done in medicine. There is only one way – doctors should tell each other about the use of AI in their clinical practice and increase the number of supporters and users. This is the way of any method in medicine, if we are talking about trust.

What other problems do you have to solve when promoting AI?

There is no task of "promoting AI", there is a task to compensate for failures in the quality or availability of medical care or its improvement. The problem is formulated in a banal way – it is necessary that the technology really works, it is necessary to prove and show it, it is necessary that clinicians use and promote it further, so that it, the technology, is in demand by them.

How are machine learning algorithms used in practice for early disease prediction?

In no way, the healthcare system is in principle not configured for early disease prediction and risk management. In a small part, this is preventive work in the mode of dispensary observation of those who have been identified and registered, but this is no longer a prediction of the disease, it is already a disease, it does not matter whether acquired or congenital. Therefore, for now we can say that technologies can help identify these risks, but the healthcare system does not know what to do with them next and is not ready to increase the detectability by an order of magnitude.

We wanted to talk about "medicine 2.0", but it turned out...

Yes, it's all "medicine 1.1", well, "1.2". Although all these gradations, of course, from the evil one, I just want to note that new medicine requires not only and not so much technology as a paradigm shift. Money should go to the industry when everyone is healthy, and not when the more sick and, most importantly, treated, the better the pharma and medicine. By making healthcare a branch of the market economy, humanity has driven itself into a very tricky trap. The effects of health are not directly extracted within a separate industry, healthcare is an initially subsidized story, the benefits of which should lie outside this system, so as not to provoke a race of technologies and pharma there for the sake of income in healing people. The transition from healthcare to health care will require significant changes in many paradigms in the economic and socio-political structure. Genetics-eugenics, lifestyle and environment – food, water, air, stress and only then medicine. This is how it should be in the modern world.

Portal "Eternal youth" http://vechnayamolodost.ru

Found a typo? Select it and press ctrl + enter Print version