The innovation of the model proposed by graph md is based on the integration between the representation of knowledge through semantic chains with the connectivity provided by neural networks. This integration establishes the increase in the abstraction of the connective process proposed by the artificial neural network. The new abstraction presents connective nodes based on different levels defined by an ontology and implemented by a graph that stores knowledge in the respective conceptual maps.
In the clinical care form, it can be seen that knowledge representation is carried out through conceptual maps. The automated analysis of information banks (images, examination records, and clinical data), through this model, allows establishing the degree of relationship of a specific context of a patient, the respective most relevant possibilities, proposing a standardization and simplification of the diagnosis. The idea of using the proposed technology is to help relate this basic information, providing specialists with the relevant information at all times. The objective is to contribute to the development of technology to help the specialist doctor to make an accurate diagnosis. Below are some aspects related to the complexity of creating these databases: overcoming ethical issues, information ownership, system usability, as well as data privacy and security. A strategy for solving these problems was initially provided by Alex Pentland from MIT, who contextualized the use of the large volume of information related to people's behavior associated with mobile computing and/or the Web. The author's contribution to this proposal is to shift the focus of the problem. Numerous initiatives to implement an integrated electronic medical record solution come up against the problems of a large centralized system, which should allow the integration of all health institutions through a network.
The graph.md platform has an initiative from another point of view. The focus is on the person having ownership of their health data. This person can allow access to their data to the doctor who is performing a consultation and/or diagnosis. On the other hand, the doctor who performed the diagnosis also has the right to the data he/she generated for the patient. This property would not give the patient the right to change the information that would be digitally signed by doctors through a digital certification process (BlockChain) associated with their CRM. This information sharing is the solution's usage strategy. Thus, it is possible to make the graph.md super app is available not only to the patient but also to the doctor and/or institutions such as Emergency Care Units, Basic Health Units, clinics, and hospitals. Each of the elements (patients, doctors, and hospitals) has an isolated virtual instance of graph.md, thus each of them is guaranteed to keep their data in the timeline. Knowledge synthesis creates new instances to perform diagnostic analyses by intelligent agents. Only the data necessary for the diagnosis enters the synthesis base, and only the sufficient data necessary for the clinical records of the events (consultations, exams, diagnoses, etc.) is returned to the elements; the nodes with their properties and relationships, therefore, exchange information with different points of view.
In the patient element, which has all its clinical events, there is an event of a specific doctor who provided a certain diagnosis at a certain time. On the other hand, this same doctor is also an element, and thus, has all the diagnoses assigned to all his patients. And thinking about that specific patient, mentioned initially in this example, here within the isolation of the virtual instance of the doctor is just one event. The institution follows the same idea, encompassing all its doctors and patients. Thus, no boundaries are crossed, a fact that guarantees the isolation of relationships and minimizes the computational complexity of the system. Other positive impacts of the model from this idea are presented in other sections of the website. But it is worth noting that in this way the complex system is limited to the universe of 10k diseases and 4k symptoms, always working with only a single patient and a single doctor at any given time. This does not mean that the system loses performance, since 35k concepts of medical diagnostic knowledge are complementary to the 15k direct terms, which currently totals 50k terms within the graph.md platform. to support medical diagnosis, with over a thousand scientific articles supporting the terms defined on the platform.
We work continuously to aggregate new knowledge from medical specialists, which is why the platform will always be in continuous evolution, storing new concepts at every moment that represent the synthesis of medical knowledge applied to patients, overcoming the physical barriers of the institution and adding value to people's preventive health in their own homes through electronic devices that begin to monitor, in a personalized and constant way, the subjective and objective health perception of each patient. [analysis e2 audiobook]
