WO2001003759A1

WO2001003759A1 - Automatic weaning system for respirator using fuzzy theory control, and recorded medium on which automatic weaning program is recorded

Info

Publication number: WO2001003759A1
Application number: PCT/JP1999/003724
Authority: WO
Inventors: Tadashi Nemoto; Yasuhiko Shimizu; Seiichi Matsunobe
Original assignee: Tapic International Co., Ltd.
Priority date: 1998-01-21
Filing date: 1999-07-09
Publication date: 2001-01-18
Also published as: JPH11206884A

Abstract

Weaning for liberation from a respirator is rationally and efficiently carried out without relying on experience and scent. A fuzzy set and fuzzy rules are made in advance based on the fuzzy theory and stored in a computer (control modules B4, B5, B6). According to the fuzzy set and fuzzy rules, and from at least the heart rate (HR), the respiratory rate, the oxygen saturation (SpO2), and/or the measured values (A1 to A12) of the tidal air exchange, the condition parameters and trend parameters thereof are calculated. The control parameter is derived from the center of gravity of the overall control diagram of the control and trend parameters, thereby controlling the respirator.

Description

Specification

Automatic venting system for ventilators using fuzzy logic control and recording media recording automatic winning programs

The present invention relates to a system for performing automatic winning control of a ventilator, and more specifically, to a technique for converting a ventilator into an automatic winning system using fuzzy logic control. Background art

Most patients managed in hospitals' intensive care units, etc., are often treated with respirators, whose respiration is managed because of respiratory disorders. The ultimate goal of ventilator placement in such cases is to get “pinning” free of ventilator use. Here, "Peaning-Weaning" means that the patient can breathe on his own, no longer needs the assistance and assistance of the machine, and removes the machine; It means to become.

In general, the use of a ventilator improves the patient's condition, and attempts to initiate inning are based on the condition of controlled breathing (where the patient's breathing is completely substituted by the ventilator). It means to shift to the state of assisted breathing. There are two types of assisted breathing: pressure support (PS), which assists the inspiratory effort in the patient's breathing, and Synchronized Intermittent Mandatory Vent i laiion (Synchronized Intermittent Mandatory Vent i laiion). The SI MV method has been implemented. In other words, if the significance of the above-mentioned “pinning” is more specifically shown, it is the above-mentioned PS method, SIMV method, or a kind of trial-and-error from controlled breathing.

〇N-〇FF This means that the patient's breathing is gradually shifted to spontaneous breathing using the F method or the like, and the patient is separated from the machine.

At present, there is no standard method of inning. When performing the coating using the PS method, SI MV method or ON-FF method, Whether the level of assistance should be reduced step by step or not depends on the patient's condition and the personal bias of the physician's skills and experience.

As a result, it is extremely difficult to formulate a general theoretical formula for the steps and processes that need to be taken in innings, and there is currently no theory that can be accepted by all physicians.

However, there is a desire to automate inning, and several proposals have been reported. For example, those examples are 1992, East, TDetal (published Chest, 101: 697-71 0'92), and 1994, Lint on D. Metal (published Chest, 100: 1096-1099 '94). Disclosure of the invention

Nevertheless, the automated techniques disclosed in these papers simply apply conventional respiratory physiology to control algorithms, and are not only complex, but also clinical. The severe situation is that the physician has not performed well (it has not worked) enough to surpass the results of performing innings based on experience or intuition.

Moreover, in reality, the success or failure of inning that relies on manual experience and intuition and passion is dependent on the skill of each physician. That is urgent.

By the way, what I learned from daily examination of patients with respiratory insufficiency was that almost all patients who were able to judge that their condition was able to withstand the inning and that they were able to judge early after the start of ventilator use were able to do the inning. And the fact that breathing mode is not so important, and whether you can actually inning or not depends on how closely you observe the patient.

In other words, even if the physician decides that the patient can do the painting, even if vaguely, the patient is rubbed around the bedside all the time, communicating with the patient, and setting the ventilator little by little. At one point I realized that if I could change it boldly, the patient would be able to win.

And the rationale for the doctor's decision to change the ventilator settings at this time is: It is a sign (life sign) and oxygenation capacity of the lungs, and it is thought that it never uses respiratory physiological theory. If this judgment is correct, ゥ inning is not a theory but a process.

The present invention is not a theory of mining, but it seeks out the secrets of the minimum necessary process to execute this process effectively and effectively, and uses a computer to effectively use the ventilator. Automatic control is the aim of the present invention.

The present invention seeks to achieve this by introducing fuzzy one theory in order to effectively promote the automation of the winning process in a ventilator, which has not been realized despite the strong needs in the past.

The outline of fuzzy theory is a concept intended to roughly and quantitatively quantify the concept that humans subjectively express and express in words or codes and to handle it with a computer according to the amount of information. .

In addition, fuzzy control is a method that uses fuzzy theory to algorithmize control-related activities conventionally performed by humans. By this method, the same control activities as those of a skilled person are to be realized by a computer.

The following is an example of this concept.

(Fuzzy theory)

In general, when dividing a continuous numerical value into several categories, the boundary is given by a single numerical value, and the numerical value has no flexibility in the system. For example, when defining temperature, if less than 10 degrees is defined as low (cold), 10 degrees or more and less than 25 degrees is just good, if 25 degrees or more is defined as high (hot), 9.9 degrees Although it is cold, if it rises just 0.1 degrees from there it will be just right and it will be inconsistent with human sensation.

This discrepancy occurs because the boundaries are given by a single number and there is no flexibility. Fuzzy theory is a theory aimed at how to treat human senses mathematically by giving flexibility to the boundaries.

Assuming that this problem is treated fuzzy, when it comes to “cold”, let's say that the degree of cold is 1.0, for example, where everyone thinks it's cold without any discussion. Set the condition that you think is not cold to be 0. This condition is represented by temperature, for example, 1.0 for temperatures below 5 degrees and 0 for cold above 15 degrees, and the temperature in between is expressed as a continuously decreasing number between 1.0 and 0. . The line segment defined by this numerical value may be a straight line or, for example, a parabola, but generally it seems that there are many straight lines. Similarly, it is defined as just hot.

In this way, when the temperature rises from 9.9 degrees to 10 degrees, the index of "cold" decreases a little, and the index of "just good" increases slightly. Will be shown.

As a degree of belonging to a subset (group) such as cold or hot in the whole set of temperature, the index such as cold or hot is called a membership value, and the membership value of cold is 0.71. And so on.

The fuzzy theory consists of a fuzzy set, which is a numerical table for determining this membership value (λ ° lame), and fuzzy rules that define the relationship between each parameter. The fuzzy set and the fuzzy rule are the basis of the engineering application of fuzzy theory, and are the most reflective of the skills of the program creator. Next, I will explain the fuzzy used in this program with specific examples of each parameter. BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 is a system block diagram showing an overview of the automatic coining system of the present invention.

FIG. 2 is an interrelation diagram showing the hardware-related parts in the system of FIG. 1 in blocks.

FIG. 3 is a block diagram showing the concept of the main part of the present invention.

FIG. 4 is a schematic diagram of a Control variable determination program.

FIG. 5 is a diagram showing a fuzzy winning algorithm.

FIG. 6 is a diagram showing a fuzzy set for the parameters of HR.

FIG. 7 is a diagram showing a fuzzy set of Control values.

FIG. 8 is a flowchart of the main program. FIG. 9 is a flowchart of the fuzzy variable determination program.

FIG. 10 is a flowchart of the breathing mode control program.

FIG. 11 is a flowchart of the PSV change program.

FIG. 12 is a flowchart of the SIMV change program.

Figure 1 3 is a flowchart of the F i 0 ₂ change program.

FIG. 14 is a flowchart of an interrupt instruction program by an alarm. FIG. 15 is a diagram showing the difference between the setting conditions instructed according to the rules of the present invention in the pressure support and the actual settings made by the doctor. BEST MODE FOR CARRYING OUT THE INVENTION

(Outline of concept of ventilator with fuzzy control ゥ inning system) にあたって In proceeding with inning, consider control based on the concept that "appropriate respiratory condition is reflected in stable physical condition". The reason why such a concept is focused on is that the evaluation of the respiratory condition of a patient has traditionally focused on the goodness and badness of individual respiratory parameters such as blood gas data and airway pressure. In practice, when clinicians make judgments, they take into account the fact that they are focusing on the whole body vital signs.

Here, whole body vital signs are vital signs. In the first embodiment of the present invention, among various vital signs, in particular, four items of heart rate (HR), tidal volume (TV), respiratory rate (f), and oxygen saturation (S p〇 ₂ ) Focusing on the eyes, selecting these data and fuzzifying them, and calculating the fuzzy data according to the fuzzy rules, a parameter called Condition (hereinafter, this parameter is abbreviated as Cond) I do. However, it is not necessary to limit to these four items, and other vital signs can be selected and used. Also in conjunction with this, HR, a TV, f, S p0 ₂ of the changing trend Trend (trend) data, say dHR, dTV, di, an d S P_〇 ₂ fuzzy reduction, as well as the Tend (Tendency) parameters Calculate overnight. Similarly to the above, the change tendency of other items may be obtained.

With these two parameters obtained, a decision to control the ventilator is made. Fuzzy again according to the rules and fuzzy set, and calculate the final parameter called Control. This control indicates the current general condition of the patient and the overall "favorable / bad" relationship of the changes. That is, the whole vital sign is evaluated based on fuzzy theory and quantified.

However, regarding the control method itself at the final stage for changing the mode of the ventilator using this parameter called Control, instead of directly applying Control to the setting change of each mode, the general steps '' Use a bi-step algorithm. No fuzzy logic is used in the final stage of changing the setting of this mode. The present invention further SI MV (Synchronized Intermittent Mandatory Ventilation) , PS (Presoure Support), change of F I_〇 _{_{2 (Fract ion of Inspired 0 2}} ) is carried out, but Uiningu is not performed by necessarily ON- OF F. The switch may be made switchable so that the ON-OF control can be removed by an alarm or at the discretion of the physician.

As described above, Fig. 4 shows an overview of the main parts of the control method (algorithm) applied by the Control parameter for the PS method and the SIMV method.

(Overview of the entire automatic coining system of the present invention)

FIG. 1 is a block diagram showing an overview of the automatic coining system of the present invention, and FIG. 2 is a block diagram showing an outline of hardware-related parts in the system.

In FIG. 1, A shows a drive part for a respirator to be subjected to the fuzzy control of the present invention, and B shows a control part according to the present invention.

Gas mixer A 1, in order to supply oxygen planned concentration to the patient, is used to mix the oxygen 〇 ₂ and air air.

The oxygen 〇 ₂ sensor A 2 is a device for checking whether or not the intake gas mixed by the gas mixer A 1 has an actually expected oxygen concentration. It also has a feedback function to the oxygen concentration controller B1 to keep the oxygen concentration at a set constant value.

The gas accumulator A3 stores the gas actually inhaled by the patient once and acts as a kind of tank to stabilize the gas supply. The gas supplied to the patient now gets the driving pressure. Bacteria Fill A4 is used to remove bacteria from the gas supplied to the patient.

The flow sensor A5 works in conjunction with the flow control valve A6 to ensure that the gas delivered to the patient is actually the expected amount. This device has a feedback function for the flow valve controller B2.If the supply is excessive, the flow control valve A6 is closed according to the current flow rate through the flow valve controller B2. Open the valve.

Humidifier A7 moderately humidifies the gas supplied to the patient.

The airway pressure gauge A8 senses the spontaneous breathing of the patient and determines whether the exhalation valve controller B3 controls the exhalation valve A9 normally according to the respiration mode set by the respiration control module B4. Provide information to Moreover, at best, mean airway pressure is used to monitor patient airway injuries, as it is important in preventing them, and is also used to sense air leaks from ventilator circuits. Exhalation valve A 9 closes when the patient is in inspiration and opens when the patient is in exhalation. This ensures that the ventilator provides gas to the patient and does not interfere with the patient's exhalation.

The flow sensor A10 measures the patient's tidal volume TV. After this, the patient's exhalation is released to the atmosphere. The information here interlocks with the exhalation valve controller B3 to ensure that the patient's exhalation frequency and tidal volume satisfy the set values. The information here is also used as vital signs used in the program. Expiratory valve A 9 also improves positive respiratory status by applying positive pressure to the patient's expiratory phase during PEEP (positive end-expiratory pressure); Technique or CPAP (Cont in continuous positive airway pressure; continuous positive airway pressure; for patients with spontaneous breathing who are intubated in the airway, improving the respiratory condition by always maintaining positive air pressure in the airway Is also used when applying The end-expiratory pressure set by the breathing mode control module B4 acts on the exhalation valve A9 via the exhalation-valve controller B3 to apply pressure to the patient's end-expiration. This pressure is constantly monitored by an airway pressure gauge A8, and a feedback function is provided to an expiration controller B3. The oxygen saturation S p 0 ₂ sensor A 11 and the heart rate HR sensor A 12 are not provided in the conventional ordinary ventilator, but are specifically provided to improve the implementation accuracy of the present invention. It is a thing.

The oxygen controller B1, the flow valve controller B2, and the intake valve controller B3 are also provided as standard equipment in conventional ventilators.

The breathing mode control module B4 is used to select and adjust the breathing mode specified by the input control panel B7 or the breathing mode specified by the fuzzy coining control module B6. The respirator setting conditions given from the fuzzy winning control module B6 of the present invention are input to the information processing means (CPU) for respirator setting in B4, and are actually Control. Therefore, the fuzzy inning control module B6 receives all information from A1 to A12. If an abnormality occurs based on the information, an alarm is issued through the alarm mechanism B9. The information of the airway pressure gauge A8 and the flow sensor A10 indicating the control status required for the ventilator performed by the respiratory control module B4 and the status of the actual operation of the ventilator are fuzzy. It is visualized through the monitor B8 of the inning control module B6 and facilitates the doctor's monitoring. Block denoted by reference numeral 2 1 in FIG. 2, HR, TV, f, S p 0 2 the measuring Josuru sensor portion, for example, the sensor probe and the digital converter (A 2 in FIG. 1, A 5, A 8, A10 and A11A12).

The block indicated by the same reference numeral 22 is an information processing part (control computer part) and corresponds to B 4, B 5 and B 6 in FIG. 1, and operates a fuzzy algorithm program and a ventilator setting program. It consists of a CPU, memory, etc. Block indicated by reference numeral 2 3 is a state display monitor portion was equivalent to B 8 of FIG. 1, direct HR, TV, f, S P_〇 _2, P aw (airway pressure), in addition to equal This is a part that displays the state determined by the fuzzy algorithm and the ventilator setting information according to it.

The block indicated by reference numeral 24 corresponds to the alarm mechanism, which is marked with the symbol B9 in Fig. 1, and stops the alarm and coining when each vital sign reaches the set alarm condition. , For starting emergency settings Information to the control computer.

The block indicated by reference numeral 25 is a part of the ventilator itself, and has a configuration based on a long-term ventilator set for both a computer-controlled metered volume and a Z-compressed pressure having respective pinning modes such as PS and SIMV. .入出力 Equipped with an input / output mechanism (terminal) for the purpose of automating inning.

(Outline of the program)

The main program is roughly divided into two parts, the first half and the second half. 3 and 8, the first half of the program portion 4 Tsunoba Ital sign (heart rate HR, tidal volume TV, respiratory rate f, the oxygen saturation S P_〇 ₂₎ was continuously measured, using fuzzy logic, the measured values and the respective change value (d HR, d TV, df , d S P_〇 ₂₎ from the finally to create a single Cont ro l parameter Isseki. This value is a comprehensive assessment of a patient's general condition based on their vital signs and trends. In short, is the patient now in the power of "good" or "bad"? Is a numerical representation of

When a physician considers a patient's condition, the physician does not place much importance on the values of individual vital signs, and considers this patient as “good feeling” or “stiffness” from the overall feeling. The first half of this program can be said to embody the physician's thinking, and the resulting figures can be said to be quantifications of the “good and bad” relationship. In that sense, it processes vital signs in a way very similar to how doctors think.

The second half of the program uses the parameters given in the first half of this program and actually changes the settings of the ventilator. Settings can be changed independently for oxygen concentration, SIMV count, and pressure support level. In this embodiment, no fuzzy logic is used. From the value given by Control (can take a value from 150 to +50) and the maximum and minimum values of each artificial respirator preset by the physician, the normal steps are taken. Change the settings using the step algorithm.

The reason why fuzzy theory is not used here is that changes to fuzzy rules and fuzzy sets require a certain amount of specialized knowledge, and manuals are used for such changes. You need to work for a relatively long time. However, changing the step-by-step algorithm is easy. In addition, if the physician needs to change the method of changing settings using Control because of the need to accommodate a particular patient, for example, most of the program changes will be sufficient for this final change. However, no fuzzy theory is used here so that any doctor can rewrite the program relatively easily.

One of the features of this program is that setting change rules can be easily created for any doctor. Other more experienced physicians may want to change the original settings as exemplified in the present invention, but in such a case, if the setting change rules do not have the flexibility, it would be troublesome. Become. Anyone can easily rewrite the setting change rules with one hand, and the machine will be useless at all.

(Algorithm)

As shown in Fig. 4, the main part of the present invention is to apply a fuzzy set and a fuzzy rule to the heart rate HR, the respiratory rate f, the oxygen saturation S p 0 ₂ and the tidal volume TV, (Condition; state) is to determine a parameter called overnight. This parameter overnight is a parameter representative of only the current state of the patient. Heart rate HR in the same way, respiration number of times f, oxygen saturation S p0 _2, tidal volume TV Torre Ndode Isseki, dHR, df, Trend (trend) from d S p0 _2, a parameter called Tendency (Tend) Is calculated. This quantifies the change in the vital signs of the patient from past to present.

Table 1 shows an example of an assignment table for calculating membership values. Table 2 shows examples of membership values after fuzzification used to determine Cond parameters and Trend parameters, and Table 3 shows examples of fuzzy rules for evaluating Cond (state). ing. Next, apply the fuzzy set and rules (for example, Table 4 for the combination) to the Cond parameter and the Trend parameter to calculate the final parameter called Control. This is a numerical value that indicates the overall condition of the patient by judging the current situation of the patient and its changes in a comprehensive manner. This is a paradigm that this Control is given in the first half of the above program (see the flowchart in Fig. 8). See).

Next, in the second half of the program, this Control and the four vital signs (heart rate HR, tidal volume TV, respiratory rate f, oxygen saturation S p〇 ₂ ), current ventilator settings, Based on the time elapsed since the setting change, the program decides whether or not to change one of the three setting conditions: the inspired oxygen concentration, the pressure support (PS) level, or the SIMV level (number of times). Judge whether to change the degree. Then, based on the judgment, the ventilator is actually changed. Regarding this actual ventilator setting change, check with the physician before changing the setting and change it after obtaining consent, or change it automatically at the discretion of the computer without the doctor's consent. Can be done either with a single switch. In addition, a program has been created so that doctors can always freely change ventilator settings (see the flowchart in Fig. 8).

(Overview of fuzzy control program)

The relationship of the program for controlling the automatic coining system of the present invention with the hardware portion will be outlined with reference to flowcharts.

This program is roughly divided into the flow of the main program in Fig. 8 showing the overall flow, and Figs. 9 to 14 showing the flow of the subroutine program in the flow. The subroutine program consists of a "fuzzy variable determination program", a "breathing mode control program", a "PSV change program", a "SIMV change program", a "Fi change program", and an "alarm interrupt instruction program". The entire program resides in the fuzzy coining control module (B6) shown in Fig. 1, and controls the data control module (B5) that controls data A1 to A12 and the input control panel (B7) for the ventilator. ) And outputs it to the breathing mode control module (B4), and automatically controls the breathing mode via the subroutine program shown in Figs.

These data information is constantly output to the monitor (B8) to display the patient's condition—the patient's condition determined by the fuzzy control program and the contents of control to the ventilator. The input data is output from B5 to the alarm module (B9). If it is determined that the state is dangerous, an alarm is displayed and the setting is changed by interrupting the breathing mode module (B4).

The fuzzy variable determination program in FIG. 9 and the breathing mode control program in FIG.

① Change of control interval setting,

② Change of membership function of fuzzy set,

③Change of fuzzy rule book,

変更 Change of breathing mode control algorithm

Subroutine programs have been compiled so that they can be easily executed depending on the condition.

The alarm interrupt command program monitors changes in the patient's condition during inning, and when the condition is judged to be critical, stops inning as soon as possible to prevent the patient from becoming dangerous. This is also organized as a subroutine.

The numerical values and conditions specifically shown in the above program are only examples and can be easily changed, and always keep in mind the change and improvement of the program. The flow of this program is shown in the flowcharts of Figs. 8 to 13 and corresponded to the block diagram of the related ventilator (Fig. 1).

The program uses a fuzzy rule table and a fuzzy set as described above. The parameters used are heart rate HR, oxygen saturation S p〇 ₂ , respiration rate f, tidal volume TV, and dHR, dS p〇 ₂ , df, d TV. First, one parameter set called Cond is made from the first four parameters, and a parameter set called Tend is determined from the latter four parameters (Trend). Then, the final parameter Control is determined from Cond and Tend.

(Cond decision)

First, the heart rate HR, the oxygen saturation S p0 ₂ of arterial blood, respiratory rate f, tidal volume TV, a fuzzy set for, Te従Tsu in Table 1 prepared in advance membership value (information processing means) Suppose that the decision is made as shown in Fig. 6 (Fig. 6 This is an example).

In FIG. 6, assuming that HR-115 is now, the membership value for "normal" given by this is 0.25, and the membership value for "high" is 0.75. Similarly, assume that other membership values are determined for each value as shown in Table 2 and stored in the information processing means.

Next HR, S P_〇 _2, f, the fuzzy rules as shown in Table 3 relates TV, advance and stored in the information processing means to prepare Me pre.

Table 3 is an excerpt from the fuzzy rule book for finding the Cond used in this program. There are five Cond that can be this excerpt for the given four parameter values over time: “best”, “good”, “normal”, “small” and “bad”. There are a total of eight combinations of parameters, and the membership value of Cond obtained from one combination is given as the minimum value of the four parameters that determined that Cond. For example,

(HR, S P_〇 _2, f, TV) = (high, low, normal, normal) = (0.75, 0.3, 0.8, 1.0)

The Cond obtained from is "bad" and its membership value is 0.3, which is the minimum of 0.75, 0.3, 0.8, and 1.0. Similarly, Table 4 shows the membership values for all combinations.

The membership value finally given to the possible Cond is the maximum value given in Table 4. For example, "bad" is in Table 4, 0.2,

0.3 is given, of which 0.3 is the maximum value. The membership value given here is

(Best, good, normal, bad, bad, bad) = (0.25, 0.7, 0.25, 0.3, 0.2). Cond was determined in this way.

(Trend's decision)

Trends are determined in exactly the same way, according to membership values and rule books that are also prepared in advance. Now suppose that the Trend is determined as follows:

(Improvement tendency, universal, worsening tendency, collapse) = (0.2, 0, 5, 0.4, 0 · 1) (Control decision) The above process means that in FIG. 4, trend data (shown at 42) indicating the current vital signs and their changes input at 41 are fuzzy as shown at 43 in FIG. In this case, a fuzzy set for converting the data of each vital sign as shown by 44 is required. Next, as in 46, the Cond and Trend are actually determined as in 45 using the rule table for determining Cond and Trend.

In this case, the determined Cond and Trend are represented by the category one (improvement tendency etc.) and their membership value (0.2et) as described above, and these two factors are used to determine Trend and Cond. Used to determine the Control, as with vital signs.

As in HR = 1 15, control is determined by applying a fuzzy set and fuzzy rules for control determination to “improvement tendency” = 0.2 (47, 48). For example, it is determined as follows.

(Large increase, increase, flat, decrease) = (0.4, 0.6, 0.3, 0.2) The control value and each vital sign determined in this way are output to the monitor at 49, and It is passed to the control program of the pinning mode. Then, it is assumed that the fuzzy set of Control is defined as shown in FIG. (Non-fuzzy)

Cut each parameter in this graph by its membership value and enclose it in bold lines as shown in Figure 7. The center of gravity of this figure is the value obtained by defuzzifying the final parameter Overnight Control. Finding this center of gravity is the ultimate goal of the series. In this case, ♦ indicates the center of gravity, and the value of Control is + 17% of the center of gravity.

(Change of actual ventilator settings)

The value of Comrol obtained here now, further S p 0 _2, and the change value d S P_〇 _2, again again calculated by the intake oxygen concentration. Let the recalculated value be called X.

For example, if Control> 0, S p〇 ₂ <92, and d S pO ₂ <0, then X = Control X I.5. Now, suppose that X = +25. At this stage of calculating X, no fuzzy theory is used. The program is F i〇 ₂ , SI MV and PS level settings can be changed, but the actual setting to be changed depends on the current vital sign, the elapsed time since the last change, and the current setting conditions.

Fig. 3 summarizes the points so far. First, the vital signs of the patient are sensed every 5 seconds, and the average value of the vital signs every 5 seconds for 5 minutes is

It is fuzzified to calculate Cond and Trend, and Control is finally given.

In the previous example, the value of + 17% indicated by the center of gravity is the result of this Control. This is passed to the ventilator setting program to calculate X. And the ventilator is actually changed, which in turn changes the patient's condition, which is sensed and fed back every 5 seconds.

Now, let us say that SIMV has been changed from these conditions, resulting in an increase in the number of times. Suppose the current SIMV is 10 times, and the upper and lower limits are 25 and 3 times, respectively. The S I MV after the change is given by the following equation. n ew SI MV = (upper limit value-current value) XX% + current value = (25-10) X 0.25 + 10 = 13.75 times / min. Instead, the change is usually rounded to a number, for example, taking a multiple of four.

Here, a multiple of 4 is assumed. Then, 1.3.75 should be rounded to 12, and the changed SIMV = 12. The above is the route from the actual parameter setting to the setting condition change.

Right now, it is rounded to a multiple of four, but it can be set to take a multiple of three, in which case it will be 15 times. In this way, the SI MV count is a multiple of a certain integer, but the doctor can freely select what multiple to use. For patients who do not find it completely safe to completely automate the configuration changes, the physician can easily configure the computer to ask the physician for permission to change the settings. In addition, the physician is always free to change the ventilator settings. And, of course, changing this setting will take precedence over computer decisions. Likewise configuration changes of PS and F I_〇 ₂ is also performed.

(Alarm and safety function) 自動 When automating innings, the most important thing is not to be able to actually perform innings, but to be patient, safe and secure. Therefore, this program also has a number of safety mechanisms (see Fig. 14). First, regarding the alarm function, during program running, the patient is monitored for vital signs every 5 seconds and fuzzy every 5 minutes. Is done. The value for fuzzification every 5 minutes uses the 5-minute average of vital signs monitored every 5 seconds.

Here the patient

① Alarm on vital sign value every 5 seconds,

② Vital sign value alarm every 5 minutes,

③ Alarm on trends in vital sign values every 5 minutes,

アラーム Alarm on the parameter Control that came out fuzzified,

It is monitored by four alarm mechanisms. In particular, the fourth type of alarm II, which has never existed before, is considered to be used independently and can be used as a monitor alarm in various fields as a multiple intelligent alarm mechanism.

The alarm in ① is activated when the value of the patient vital sign monitored every 5 seconds exceeds a preset upper and lower limit. The alarm in ② works when the patient vital sign value averaged every 5 minutes exceeds a preset upper and lower limit. The alarm in ② works independently of the alarm in ①, and unlike the alarm in ①, which monitors the instantaneous patient condition, it monitors the condition for a relatively long time. The alarm in (3) is an alarm related to changes in patient vital signs averaged every 5 minutes. Even if the patient's vital signs fall within the upper and lower limits set in (2) and (3), the changes fluctuate greatly. Work when you do. For example, if a patient's heart rate changes from 70 to 120 in 5 minutes, even if 70 and 120 were within the alarm, 50 The increase is judged to be abnormal, and an alarm is issued.

アラーム is a completely new type of alarm that has never existed before. As a result of fuzzifying the patient's vital signs, the final parameter obtained, Control, indicates the overall condition of the patient. Evil can be judged. Therefore, this value itself is It can be used as an indicator for the overall patient state, regardless of the condition, and it can be added with an alarm function. There is no alarm of this type, and it is considered that it is possible to take out only this alarm mechanism and apply it to monitoring mechanisms in other medical and engineering fields.

As another alarm,

アラーム Alarm on the first vital sign at the time of the program evening,

アラーム Alarms on status changes after increasing oxygen concentration,

アラーム There is an alarm for vital sign change after all setting changes.

Now, suppose that the patient's condition became worse for ⑥ and ⑦, and the oxygen concentration was increased. Naturally, an increase in oxygen saturation is expected, but it is not possible to determine whether the increase in oxygen saturation of the patient was caused by an increase in oxygen concentration or due to improvement in general condition. In order to avoid this difficulty, the program cancels all trend data once after the oxygen concentration is increased, and restarts the trend data from the beginning.

However, if the oxygen saturation decreases at this time, it is absolutely abnormal. Therefore, the sixth alarm works especially when oxygen saturation drops. In addition, alarms about vital sign changes after setting changes are indispensable, since setting changes always cause the patient's condition to deteriorate.

Next, the safety mechanism is roughly divided into three.

A. ゥ Stop inning and start emergency setting.

B. 中止 Stop the inning and follow the patient with the final ventilation setting.

C. In case of worsening vital sign after setting change, ゥ continue inning, automatically back to the setting before setting change.

Except that alarm ⑦ is linked to safety mechanism C, ① to ⑥ are tied to A and B variously depending on their values. Of course, ⑦ and A and B can be connected. The connection relationship and the alarm setting conditions can be changed relatively easily. In addition, each doctor can freely set the emergency setting.

(Retro-Scientific · Clinical · Trial)

A clinical trial was performed on three patients. The oxygen concentration and the pressure support were compared 324 times, respectively, and the SI MV was compared 70 times. Compare books The actual setting conditions are subtracted from the setting conditions given by the program along the rules created by the inventor and the rules created by other doctors, and the results are graphed and illustrated (Fig. 15). See).

A very high peak is observed where the difference is zero (coincidence). The setting conditions specified in accordance with the rules of the present inventor were as follows: oxygen concentration: 305 times (94%), pressure support: 248 times (76%), and SIMV: 52 times (74%). And consistent with the actual settings made by the doctor. FIG. 15 shows the difference (cm H ₂₀ ) between the rule of the present invention and the actual setting in the pressure support on the horizontal axis, and the number of times on the vertical axis. this is

1. Cannot continuously measure vital signs,

2. The physician is able to achieve a match rate in spite of a rather unfavorable condition for the program, such as making a whirlwind call to promptly change settings, for example.

(Flow chart showing the flow of processing)

FIG. 8 is a flow chart showing the flow of the main program of the automatic ventilation system for a ventilator using the fuzzy logic control of the present invention, and FIGS. 9 to 13 are flow charts showing the flow of the subroutine program. Although it is an expression that can track the flow of the program in chronological order, the explanation may be complicated.Therefore, in the explanations up to the preceding paragraph, the technical idea of the invention was conceptualized in an abstract manner and expressed by simple examples. However, please be aware that the entities are the same.

(Evaluation)

Next, evaluation of the basic concept will be described. As to the question, "Can I really change the settings with just four such simple vital signs?" First, the machine changes the settings of the ventilator and guides the patient to the inning. I said. In other words, the present invention relates to a ventilator, and does not involve other elements of the inning, such as nutrition and mental care. So, if you are going to ニング inning ’with these elements, this question is correct. However, if the settings are changed, the physician will, in fact, see almost nothing other than these four, at most as much as arterial blood gas analysis or the patient's complexion. The reason we chose these four is that

1. Non-invasive to the intubated patient;

2. Continuous monitoring is possible,

3. There is no need to explain the significance, rationality or interpretation,

4. Can be measured at any facility,

5. Economically cheap,

6. And it is important to note that it is actually possible to change the ventilator settings with only these four,

Above the heart rate HR is a program of the first embodiment, tidal volume TV, respiratory rate f, it went all set to change the measurement of four vital signs Oxygen saturation S p 0 _2.

Table 5 shows an example of the membership value of the second example when the measurement of tidal volume (TV) is omitted from the four vital signs in the first example, Table 6 shows an example of the rule book, and the combination Table 7 shows examples of membership values for.

Table 8 respectively of the third embodiment was omitted likewise measurement of oxygen saturation S p O ₂

Shown at 10.

In the flowcharts shown in FIG. 8 to FIG.

Next, I will explain the advantages of this program over doctors. Doctors can't work 24 hours due to their high salary, but the program is not tired, complains and doesn't play. You can make more frequent decisions than doctors and make fine-grained setting changes. So this program may be close to a nurse. List the application areas expected from this program.

This is probably the first automated decision-making program in the medical field,

1. Renormalize large variables into one parameter using fuzzy logic.

2. Based on this parameter, make one or more decisions independently from the current situation,

There are two features. If a large number of variables are the vital signs of the patient, and one parameter is the parameter that indicates the general condition of the patient, Controll, then many vital signs can be monitored at once, and This makes it possible to create an unprecedented monitor system in which the parameter “l” is used to perform various specific operations using the parameter, and an alarm is also performed using the parameter.

1. Monitor in intensive care unit

2. CCU monitor

3. Prognosis of patients who need to monitor many parameters overnight (such as pregnant women and newborns at birth)

(Features of the present invention)

1) Inning, which was thought to be impossible to automate until now, has been algorithmized by judging good or bad respiratory condition based on fuzzy theory from the patient's whole body condition and automated.

2) This will reduce the clinician's frequent setting change effort, and also make the biased setting method of the physician's personal ability, knowledge, etc., objective, and optimize artificial respiration for inning. Capability to execute device setting conditions.

3) When new knowledge is accumulated in the future, the ability of the machine can be easily upgraded by changing the algorithm.

4) Using a fuzzy logic for vital signs, which are a large number of variables, and adopting a comprehensive monitoring system that is independent of individual vital signs.

And the like. Industrial applicability

In the past, it was a breakthrough from the common sense that automation of inning was considered impossible, embodying the technical idea of determining whether the respiratory condition is good or bad by fuzzy theory from the patient's whole body condition and making it algorithmic. Has realized automation.

As a result, the clinician's frequent setting change effort is reduced, and the biased setting method of the physician's individual competence and knowledge is made objective, and the optimal ventilator for inning Setting conditions can be applied to patients. When technology advances or new knowledge accumulates in the future, the ability of the ventilator hardware to be easily upgraded, for example, by simply changing software such as algorithms. Example assignment table for membership value calculation

Membership value after fuzzification

Example rulebook for condition evaluation

HRS p0 ₂ f TV state HRS p0 ₂ f TV state

(Cond) (Cond) Normal Normal Normal Normal Normal Best High Normal Normal Normal Good Good Normal Normal 冋 Normal Good High High Normal High Normal Normal;

Say S normal low normal normal normal;

Say high low normal normal, normal low high normal / | ヽ high low high normal Member value for combination

Example rulebook for condition evaluation

Membership value for combination

[Best, Good, Normal, Bad, Bad, Bad] = [0.25 0.7 0.2 0.25, 0.3] After fuzzification of each vital sign-Ship value

Example of rulebook for evaluation of application

[Good, Normal, Bad, Worst] [0.25 0.4 0.2 0.6 0.2]

Claims

The scope of the claims

1. An automatic coining system that controls the operation of the ventilator using fuzzy logic,

At least a heart rate (HR), respiratory rate (f), oxygen saturation (S p 0 ₂ ) and / or tidal volume ( TV)

Means for converting the output of each of the sensors into digital information and inputting the information, and controlling fuzzification means stored in advance to fuzzify the digital information, respectively, based on the fuzzified information. An information processing means for deriving at least a first parameter and a second parameter;

Means for obtaining a third parameter value in accordance with the first parameter and the second parameter value obtained from the information processing means, and supplying the third parameter value as control information for a ventilator;

Feedback means capable of providing correction information to the previously entered information in response to a change over time, a change in condition or a judgment of a physician;

An automatic spinning system for a ventilator using fuzzy logic control.

2. Each of the above information and desired parameters

2. The automatic ventilation system for a ventilator using fuzzy logic control according to claim 1, further comprising display means capable of visually displaying.

3. The information processing means includes at least the first parameter and

The fuzzy processing device according to claim 1 or 2, wherein a re-fuzzification process is performed according to a fuzzy set and a fuzzy rule for determining a control of the ventilator based on a subsequent change tendency of the second parameter. Automatic ventilation system for ventilators using theoretical control.

4. Based on a fuzzy set and a fuzzy rule stored in advance in the information processing means, at least a fuzzified heart rate obtained from the sensor.

(HR), respiratory rate (f), further oxygen saturation (S P_〇 ₂₎ and Z or once conversion The information processing means calculates the Cond (condition) parameter from various data such as volume (TV), and calculates the trend parameter of the various data. Claims 1 and 2 including a step of deriving a control parameter, which is an overall parameter parameter of the various parameters, and supplying the parameter parameter to a ventilator. 4. An automatic ventilation system for a ventilator using fuzzy logic control according to paragraph 3 or 3.

5. In the information processing means, the various parameters are evaluated by using fuzzy logic theory as a whole, regardless of their individual changes, and are quantified so as to catch early signs of deterioration of the patient's condition. 5. The automatic ventilation system for a ventilator using fuzzy logic control according to claim 1, wherein the alarm system includes an alarm system according to claim 1.

6. A recording medium recording an automatic coining program for controlling at least expiration and inspiration by a ventilator, wherein the program is configured to obtain respiratory parameters to be instructed to the ventilator, At least the heart rate (HR), respiratory rate (f), and sent to the information processing section measured by further oxygen saturation (S P_〇 ₂₎ and Z or tidal volume (TV) and, respectively it sensors, each Fuzzy output of each of the sensors, causing the information processing means to derive at least a first parameter and a second parameter based on the fuzzified information, and deriving from the information processing means. A third parameter is derived in accordance with the first parameter and the second parameter, and is supplied as control information for the ventilator. Recorded an automatic mining program for controlling at least expiration and inspiration by the ventilator, comprising the step of providing feedback as a correction to the previously entered information in response to a change or a physician's judgment. recoding media.

7. Based on a fuzzy set and a fuzzy rule stored in advance in the information processing means, at least a heart rate (HR), a respiratory rate (f), and an oxygen saturation (S) obtained from the sensor and fuzzified. From the various data such as p 0 ₂ ) and / or single air exchange (TV), Cond (condition) parameter is calculated by the information processing means, and the trend parameter of the various data is calculated. Let me calculate A step of deriving a control parameter (control) parameter, which is an overall parameter of the various data by fuzzification, and supplying the parameter to a ventilator. A recording medium for recording an automatic coining program according to item 5.