Patient care with artificial intelligence?
The new IT model can make the lives of asthma patients and the work of asthma doctors easier, but in the medium term, it can be extended to several areas of chronic care, shows the two-year research and development project that has already been closed, but whose results are being published now by CompuTREND Ltd, known primarily for its healthcare IT solutions.
The research and development tender, called KFI_16-1-2017-0113, supported by the Hungarian National Research, Development and Innovation Office, the methodology of asthma patient care under scrutiny and developed data analysis and artificial intelligence solutions that supplement and make the care process much easier.
During the development, a Hungarian IT solution was created that is based on medical text analysis, which predicts therapeutic risks and explores the factors affecting the effectiveness and efficiency of drug therapy. Based on the results, it will be possible to use the advanced data analysis options in other long-term therapeutic areas as well.
Structured data from medical text
The aim of the research and development project was first to create a framework for data extracted from the complex medical documentation of children with asthma or food allergies, as well as real-time information collected in the framework of the continuous spirometer examination and care system carried out in the home environment, related to the disease, recorded by the patient relying on condition descriptions, it provides the attending physician with additional information about the progress of the therapy, as well as predicts the inadequate effectiveness of the drug therapy and the need for a change.
The most important goal of the research program was the construction of a model suitable for medical text analysis in Hungarian, which creates structured data from the textual information of the medical documentation, analyzes it, and compares it with the literature data, as well as with the results of the given therapeutic area. By creating the foundations of medical text analysis in Hungarian, it became possible to give medical and professional research in various fields a new data boost with the big data-based analysis of the electronically available and ever-expanding health data.
Automation of healthcare processes
In a global context, with the continuous growth of the population, the significant increase in the duration of living with chronic diseases, and the rapid development of technology, healthcare expenditures are increasing explosively, and developed economies spend an ever-increasing proportion of Gross National Product on healthcare. The rationalization of healthcare expenditures, thus automating the processes used in decision-making, has become a key issue.
CompuTREND Ltd, which holds a dominant marketing position in the field of health informatics, especially in the field of health economic informatics, and has accumulated significant professional experience, has therefore launched a research and development program for the implementation of big data-based systems.
The healthcare industry – and in this respect Hungary is by no means lagging behind Western Europe – collects, manages and stores a huge amount of data. The common characteristic of these data is that they can be found in different, isolated places, in hospitals, in the databases of family doctors’ offices or private healthcare providers, and the databases of clinical examination and research sites. Structured, unstructured, textual, and image-based, they consist of instrument signals, and this heterogeneous data set makes it extremely difficult to analyze data, extract information, draw appropriate conclusions, and fully map the cause-and-effect relationship system.
However, with the help of big data analysis, a solution to this problem can become available: in addition to structured health data, the information contained in unstructured health data, typically textual descriptions, can also be used for the most accurate decision-making.
The textual element of the medical documentation, the status, epicrisis descriptions, and radiological and histological findings contain a lot of information that cannot be extracted with the help of traditional data mining tools because the majority of medical documentation is a free text field, i.e. an unstructured data set. In the research and development project, CompuTREND, therefore, set out to develop a solution that is able to build a structured database from unstructured data in an automated manner – moreover, it also takes into account the peculiarities of the Hungarian language, so it can be used smoothly in Hungarian healthcare.
The big data model
In the course of the research, the company developed an adapted algorithm and methodological model based on big data processing, which can collect relevant information about the patient’s condition, symptoms affecting the quality of life, appropriate dosage, interactions, and side effects of the drugs taken for the attending physician even between two doctor-patient meetings and indicates as soon as possible if medical intervention is required due to some risk factor.
The algorithm can analyze the symptoms and complaints that determine the patient’s quality of life, the relationships between measurable vital parameters, and how they can be used to predict the worsening of the condition. The research and development project focused primarily on asthma and food allergies, but the big data model can also be used to support the care of other groups of patients, primarily with a standardized database suitable for describing symptoms and complaints and methodological description that is also suitable for use in other fields.
Results of the research
In the two-year research and development program, it was possible to create an operating model that provides a solution for the integration of currently unprocessed data and information from the textual data of health documentation generated during patient care into healing processes. At the same time, an IT-based care model was created, in which the patient itself must actively participate in the treatment process, and by consistently following the therapy, you can expect to achieve an adequate quality of life.
The care model offers the possibility to integrate the measurement data of external devices, in the case of asthma care, the measured data of the spirometer, and also includes a solution suitable for logging the patient’s symptoms and complaints in real-time, which enables effective monitoring of various diseases at home, as well as deteriorations, accurate prediction of therapeutic risks.
Although the care model was developed for patients suffering from respiratory and food allergies and asthma, the methodology also creates an opportunity to support cardiovascular diseases, diabetes, blood coagulation disorders, and other chronic diseases, so that the various participants have access to information on the effectiveness and efficiency of the therapy.
The IT solution created within the framework of the research and development project created the basis for the process that, after reading the complete medical findings, the lay patient automatically receives the interpretation of the medical material in a structured way.