Poster
# 101
Poster |
|
6th Internet World Congress for Biomedical Sciences |
Jose Luis Ruiz Gonzalez(1), Francisco Taboada(2), Antonio M. López(3), Alberto Diez(4)
(1)Universidad de Oviedo - Oviedo. Spain
(2)Hospital Central de Asturias - Oviedo. Spain
(3)(4)University of Oviedo - Gijón. Spain
|
 |
[Medical Electronics & Engineering] |
| |
[New Technology] |
Worldwide economical crisis and progressive reduction of available beds in hospitals has forced Intensive Care Units to improve the functions and efficiency of their activities. These changes need the improvement of two interrelated areas: technological development with high cost-effectiveness relation and information systems that allow a more efficient use of technology.
The demands to which Intensive Medicine is faced up are the result of a particularly difficult dichotomy in modern medicine. While there are more means available to treat critical patients each time, financial resources which support the advanced care of these patients are being progressively limited due to economical a budget reasons.
It is in this context that we have posed the simultaneous integration of several monitoring systems on a critical patient that can bring maximum effectivity with minimal risks, and which can contribute to future advances in our capability to treat patients that are referred to the Intensive Medicine Unit. To sum up, we try to combine emergent technologies with tools for database usage.
Critical patients who are referred to a Critical Care Unit often have a deficit in tisular perfusion and an initial compensating response characterized by vasoconstriction, independently that the initial cause was a consequence of hipovolemia, trauma, low output syndrome, or something else. In any of these situations the inadequate tisular perfusion can lead to tisular hipoxia, bad micro-circulation distribution, reperfusion damage, multiorganic failure and death.
Even though the clinical picture of shock is easily diagnosed when it is established, it is much more difficult to recognize in its initial stages when the symptoms are imprecise. This means important delays in its treatment and, as a consequence, worse results. If early monitoring and diagnose can be achieved by a simultaneous combination of different systems, then tisular hipoperfusion situations can be early solved and thus survival can increase.
The main goal of this study is the establishment of a is the development of a method of multiple monitoring, both bloody and bloodless, which allows an early diagnosis and an aggressive treatment from the beginning in hipoperfusion situations. Previous studies have shown that optimizing oxygen delivery preoperatively and keeping it in supra-normal values during postoperative reduced mortality by 75%. In contrast with these results, other studies showed that there wasn´t any improvement in patients referred to an ICU after developing a multiorganic failure, in spite of the therapeutic efforts made.
Thus, our work has been pointed towards the implementation of major components of circulation inside a program that allowed capture and storage of physiological signals by a computer. These major components include variables that inform of haemodynamics, respiratory situation and oxygen carriage and consumption.
Cardiac output is calculated using TFI (thoracic fluid index) and the outline of a pulse wave. Systolic, diastolic and mean arterial pressures through an arterial catheter and a pressure monitor.
TFI method uses a tetra-polar electrode system, which are located in the neck and thorax of the patient, and through which alternative current of 4 mA at 100 KHz goes. This allows a signal of ECG and the first derivative of the impedance, which is afterwards processed to obtain heart rate, cardiac output and stroke volume.
Through the pulse wave method, which requires inserting a catheter into the femoral artery, cardiac output can also be obtained, in addition to intra-arterial pressures. That method requires a pre-calibration through by getting the cardiac output through transpulmonary thermodilution. The monitor also informs of the state of myocardial contractility through the derivative pressure/time.
To get the transcutaneous oxygen saturation, a conventional pulsioximeter is used.
Oxygen consumption is studied through indirect calorimetry. The same device informs of CO2 production. El O2 is carriage is deducted once cardiac output, hemoglobin and O2 saturation are known.
Pressure, flux and volume signals are obtained through respirators, if the patient is being subjected to mechanical ventilation
The goal of the system is to be able to read the data from different monitoring systems and to store all the data in a centralized way, an thus be able to work together with all the information.
The system is made up of a central computer, in which the database with all the information is stored, and one (or more) computers connected to it by a LAN (Local Area Network), which read the data from the monitors and sends it to the database.
The communication between the monitors and the computers that capture the data is done through the serial protocol RS-232-C, for which the monitors are already prepared. The input to the computer is done though an RS-232 multiplex, which allows all the data to be able to enter to the computer using a single communications port.
Graphically, the global configuration of the system can be seen in fig. 1.
This capture system has been designed in a different layer structure, in which every layer is specialized in a different task. This structure makes it very easy the incorporation of new monitors, as well as making the system as insensible as possible to communication failures.
Three levels have been designed:
1) The lowest level is formed by a group of modules, each one specialized in the dialog which a certain monitor type, and which will be in charge of the acquisition of the data obtained by the measuring device. The use of this level makes the addition of a new monitor quite easy, because all that must be done is the development of an specific module for communicating with it, and which will automatically integrated with the rest of levels.
2) Above these, there is a coordinating module. It is in charge of the control of the capture modules and the management of the information provided by them. This module gives the starting and ending signals of the capture, as well as filter the information obtained by each module, extracting the data and storing it in an intermediate buffer for its insertion in the database.
3) Finally, in the highest level, and working independently of the previous layers, there is a module which extracts the information provided by the previous layer and stores it in the database. This module makes the capture system insensitive to network failures because, if we didn´t use an intermediate storage, a problem in the communication with the database would lead to data loss. This upper module also allows capturing data from computers not connected to the central one, which could be inserted a posteriori.
This modular structure can be seen in detail in fig. 2.
All the monitoring devices we era working with have an RS-232 serial connection port. In addition, we have observed that their communications protocol is based on the transmission of the data coded in ASCII. Some of the devices need to be configured, but this is done also through the transmission of ASCII characters.
Even though some of the monitors provide measures of analog data, those values are also given in ASCII format, which simplifies the process, making it unnecessary to process analog data. That way, A/D cards and digitalization of the signals is avoided.
The design of the database has been done following tow main guides. In one hand, we pretend to store every available data, that is, all the values that the monitors measure, so in future we have enough available data for queries o correlation. On the other hand, the database must be open to the future inclusion of new monitors, without having to modify it.
The most basic index over which the database is ordered is the patient´s history number, which is the number correspondent to each one´s clinical history. This number is actually used to refer to each patient in the existent data archives. In addition to the patient´s information, the applied treatments and the diagnosis are registered, so all the information necessary to carry out future studies is available.
We define session ("sesión") as the series of data, continuous in time, coming from a group of different monitors, containing all the data read from them from the beginning of the data capture until it is stopped. The basic idea of the data capture is to be able to make sessions of data read from the patients in different moments, to be able to study those data and its evolution taking into account the applied treatments.
The same way, we define taking ("toma") as the series of data inside a session belonging to each of the monitoring devices, in such a way that a session is, thus, the group of a series of takings, belonging each of them to the same patient and a different monitor. Each of the takings is made up of a series of measures ("medidas"), grouped in different tables depending on the monitor they come from. Each of the measures is the group of data sent by a monitor in a certain moment, and which represents the values of different variables of the patient in that instant.
Finally, we pretend to include in the database, in auxiliary tables, all the information concerning the system´s behavior, including that relative to the configuration of the monitors. That way, it´s unnecessary to have configuration files in each of the capturing computers involved in the system. This solution makes the system easier to update, or a new monitoring device to be included, because all that has to be done is to modify a series of values in the database, and not in some files in every capturing equipment.
Implementation of medical algorithms
Instead of implementing medical algorithms inside the system, we have decided to create an interface to export the data in the database to an external statistics program. This way, advance can be taken from all the capabilities of these programs, and getting an open system, not restricted to a group of certain fixed algorithms coded into the application.
To access the database, instead of using query languages like SQL, unknown to the personal of the ICU, we have decided to create a screen that allows, in a graphical way, to specify which information is wanted.
On the other side we have studied and worked some alternative options based in techniques of artificial intelligence (AI), as would be the application of evolutionary algorithms or artificial neural nets, to be able to make studies that would be impossible to achieve using just classical techniques. In this point much information about mentioned techniques has been compiled and some test cases have been evaluated.
The user interface has been designed in a continuous work of the technical team and the ICU personal. We have taken into account that the interface has to be friendly as well as easy and intuitive, creating an easy-to-use system and promoting that way its usage.
The interface has been created using the capabilities of the modern GUIs (Graphic User Interfaces) and windowing systems, as well as integrating the user interaction inside an easy-to-modify web-browser.
Some screenshots of the user interface can be seen in fig. 3 and fig. 4.
Even though the system is not absolutely finished yet, it is in an advanced development stage, requiring only some final integration, debugging and integral checking before being used in the ICU.
The proposed system means a great advance in monitoring patients, allowing the integrated show of different variables, what will imply more efficiency in the treatment of patients in state of shock.
In addition, massive storage of information will allow making studies of the evolution of patients and the impact of different treatments, directed all of it towards increasing the life expectancy of the patients referred to the ICU. It also opens the door to the usage of non-classical techniques for processing information and modeling systems. These techniques, although not accepted in many forums, represent clearly an open door in many no-way-out streets, so its application deserves being studied in the medical environment.
We deeply thank Universidad de Oviedo for financing this project under the regional interest program.
We also need to thank...
Área de Ingeniería de Sistemas y Automática.
Departamento de Ingeniería Eléctrica, Electrónica, de Computadores y Sistemas.
Campus Universitario de Viesques.
Universidad de Oviedo.
Unidad de Cuidados Intensivos.
Hospital Central de Asturias.
...without whose collaboration this project wouldn´t have been possible.
|
|
|
[Medical Electronics & Engineering] |
| |
|
[New Technology] |
