The invention relates to a method for controlling a bi-level apparatus, and bi-level apparatus. Specifically, this invention relates to the reliable detection of inspiration and expiration despite the pressure fluctuations generated by the bi-level apparatus.
Like the slightly less complex CPAP-apparatus, bi-level apparatus serve the pneumatic splinting of the respiratory tract to avoid obstructive respiratory troubles during sleep.
The CPAP (continuous positive airway pressure) therapy was developed for the treatment of apneas and is described in Chest. Volume No. 110, pages 1077-1088, October 1996 and in Sleep, Volume No. 19, pages 184-188. A CPAP-apparatus generates a positive overpressure up to approximately 30 mbar by means of a compressor or turbine and administers the same, preferably via a humidifier, via a hose and a nose mask, to the respiratory tract of the patient. This overpressure is to make sure that the upper respiratory tract remains fully opened during the whole night, so that no apneas disturbing the patient's sleep will occur (DE 198 49 571 A1). The required overpressure depends, inter alia, on the sleeping phase and the body position of the sleeping person.
Frequently, the overpressure is a nuisance to the patient. To adjust the overpressure as low as possible but as high as necessary, so-called auto-CPAP-apparatus (compare
Another approach to make the pneumatic splinting for the patient as comfortable as possible relates to so-called bi-level apparatus. Bi-level apparatus support the respiration of the patient by administering a slightly higher pressure to the patient during the inspiration as compared to the expiration.
Due to the different pressures during the inspiration and the expiration bi-level apparatus require the determination of the inspiration and expiration phases. In the bi-level apparatus known from WO 98/35715 A1 and EP 0 656 216 A2 the time derivative of the airflow is compared with threshold values so as to make a distinction between inspiration and expiration. According to WO 98/35715 A1 this method used to be the common one in the prior art (bridging paragraph from page 1 to page 2).
In connection with bi-level apparatus it is technically difficult to exactly detect the transitions between inspiration and expiration by means of the airflow, because the pressure is to be changed exactly during these transitions. By changing the pressure, the air in the respiratory hose and the patient's lungs is slightly compressed or expanded, so that the airflow generated by the respiration is superimposed by an airflow generated by the pressure change. The airflow generated by the pressure change is now particularly high exactly at those times that are to be detected with exactness.
It is also problematical that a derivative acts like a high-pass filtering, which results in a roughening of the signal so that noise occurs more strongly. Due to the roughness of the signal the simple comparison with threshold values may provide incorrect results. In WO 02/083221 A2 the derivative is therefore combined with a low-pass filtering, which is called “estimation of the derivative”. On the other hand, a low-pass filtering has the disadvantage that it delays the rise or fall of a signal.
It is desirable to detect the transition moments between inspiration and expiration more exactly and thus determine the inspiration and expiration phases more exactly.
According to an embodiment of the invention a method for controlling a bi-level apparatus is provided. The airflow is repeatedly measured thereby obtaining measured airflow values. An expiration and an inspiration threshold value are chosen. The airflow is repeatedly compared with the expiration threshold value during an expiration mode. Switching to an inspiration mode takes place if the airflow is greater than the expiration threshold value in a comparison with the expiration threshold value. The airflow is repeatedly compared with the inspiration threshold value during the inspiration mode. Switching to the expiration mode takes place, if the airflow is smaller than the inspiration threshold value in a comparison with the inspiration threshold value.
According to another embodiment of the invention a respiratory apparatus comprises a fan, a pressure sensor, a flow sensor and a processor. The pressure sensor determines an overpressure under which air is provided by the fan. The flow sensor measures an airflow. The processor is connected to the pressure sensor and the flow sensor. A pressure signal is fed from the pressure sensor and an airflow signal from the flow sensor to the processor. The processor stores a sequence of instructions and is suitable for executing the instructions. The processor repeatedly obtains measured airflow values from the airflow signal. The processor chooses an expiration threshold value and an inspiration threshold value. The processor repeatedly compares the airflow values with the expiration threshold value during an expiration mode. The processor switches to an inspiration mode if an airflow value is greater than the expiration threshold value in a comparison with the expiration threshold value The processor repeatedly compares the airflow values with the inspiration threshold value during the inspiration mode and switches to the expiration mode if an airflow value is smaller than the inspiration threshold value in a comparison with the inspiration threshold value. An advantage of comparing the airflow with threshold values is that a roughening of the airflow by a time derivative and a delay caused by the low-pass filtering are avoided.
The choosing of the minimum airflow during the preceding expiration phase as threshold value immediately after switching to an inspiration mode, and the choosing of the maximum airflow during the preceding inspiration phase as threshold value immediately after switching to an expiration mode prevent that a further switching to the other respiratory mode takes place too fast as a result of the airflow fluctuations inherent in the pressure change.
The additional comparison of the derivative of the airflow with different threshold values during the inspiration and expiration phases, too, prevents an undesired switching to the other respiratory mode.
To achieve a noise suppression, the measured airflow is initially subjected to a median filtering and then averaged prior to calculating the derivative of the airflow. The median filtering advantageously suppresses outliers. The combination of a median filtering over a few measured values and a subsequent averaging over double the number of measured values is an optimum compromise between calculation expenditure, filtering time and required averaging.
The additional inclusion of the actual pressure increases the reliability of the switching between inspiration and expiration mode. One aspect is here that no switching to the inspiration mode is to take place in the event of coughing or sneezing. The actual pressure hereby increases above the set pressure because the pressure control loop is too sluggish to compensate such fast pressure fluctuations. The other aspect is that a switching to the inspiration mode is to take place if the actual pressure is below the set pressure and the airflow has strongly increased in a third predetermined time interval.
The lowering of the threshold value for the airflow during the expiration shortly after switching to the expiration mode renders the method increasingly sensitive to a switching back to inspiration mode. This advantageously corresponds to the average duration of an expiration phase.
In a correspondingly advantageous manner the threshold value for the airflow is raised approximately proportionally to the current airflow during the inspiration until the current airflow has reached a maximum. Upon reaching the maximum, the threshold value is approximately kept constant.
Preferred embodiments of the invention will be explained in more detail by means of the attached drawings below. In the drawings:
The architecture of apparatus 31 follows the trend in the field of electronics to digitalize sensor signals as fast as possible and then to perform the signal processing digitally. The program executed by the microprocessor 39 includes an internal control loop according to which the speed of the fan is controlled such that the actual pressure measured by the pressure sensor 36 corresponds to a set pressure as well as possible. The set pressure is predetermined by other program parts. If the set pressure is predetermined without taking into account the inspiration and the expiration the apparatus 31 is a CPAP-apparatus. If, in addition, the airflow is evaluated to optimize the set pressure an auto-CPAP-apparatus is concerned. If different set pressures for inspiration and expiration are predetermined the apparatus 31 is a bi-level apparatus. Thus, it is merely a question of the other program parts whether the apparatus 31 works as an auto-CPAP- or as a bi-level apparatus.
A preprocessing takes place in step 55, which will be explained in more detail by means of
The preprocessing 55 is shown in more detail in
In step 63 a first moving average value Ave5i (Ave=average) over the last 500 measured airflow values Fi is calculated according to formula (1). This corresponds to an averaging of the air flow over five seconds.
Similarly, in step 64, another moving average value Ave1i is calculated from the 100 airflow values measured in the last second according to formula (2):
In step 65 a median filtering over the last ten measured airflow values is performed in accordance with formula (3). In a median filtering the average value or the arithmetic average of both average values is returned or processed further. A median filtering is more complicated than an average value calculation. However, the result is practically not affected by outliers, while outliers occur in the calculation of the average value.
Fmed,i=Median(Fi-9 . . . , Fi) (3)
Then, a moving average value
Finally, in steps 67 and 68, another two moving average values HighAvei and LowAvei respectively are calculated over 1500 measured airflow values Fi each in accordance with formula (5) or (6) respectively, which corresponds to a time span of 15 seconds. The special thing about these averagings is that only airflow values Fi measured during the inspiration are included in HighAvei and that only airflow values Fi measured during the expiration are included in LowAvei. If the variable bInspiration is true, the method according to the invention is in the inspiration mode, so that the measured airflow values Fi are used to calculate HighAvei. Otherwise, the method is in the expiration mode, so that the measured airflow values Fi are used to calculate LowAvei.
The expiration processing in step 57 is explained by means of
In step 72, a so-called offset Offi is calculated for each index i according to formula (7):
Offi=(HighAvei−Ave5[00]i)/6 (7)
In another embodiment, the Offset Offi may be calculated according to formula (7′) in step 72.
Offi=(HighAvei−LowAvei)/10 (7′)
Next, in step 78, the memory AveHold=Ave1i is set for three seconds, if tex>1 s and Fi>Ave5i and Fi=Ave1i. The latter three conditions are verified in steps 73 to 75. From this follows that also Ave1i>Ave5i. tex here stands for the time since the last switching to the expiration mode. The time at which AveHold=Ave1i is set is stored in memory tAH0 (AH=AveHold) in step 79. If one of the conditions verified in steps 73 to 75 is not fulfilled, it is verified in step 76 whether the three seconds have already expired. If so, memory AveHold=0 is set in step 77.
In another embodiment, it is additionally examined as to whether memory AveHold=0, before AveHold is set. This causes memory AveHold to remain unchanged after being set.
In another embodiment AveHold will be reset, i.e. set to 0, when switching to the inspiration mode.
In
TLow=InMaxwenn0≦tex<0.25 s (8)
If 0.25 s<tex<1 s the expiration threshold value TLow is lowered during the time span 15 approximately linearly with time according to formula (9). Ave5AtSC is equal to Ave5i with tex=0, i.e. at the moment of switching from the inspiration mode to the expiration mode. The function max (x,y) returns x, if x>y and otherwise y.
TLow=TLow,Initial*(1−tex)+max(Ave5i+Offi,Ave5AtSC)*tex (9)
Next, in steps 86, 88 and 90, the expiration threshold value TLow is calculated for the time segments 16, 17 and 18 in accordance with the following formulae 10 to 12.
TLow=max(Ave5i+Offi,Ave5AtSC)if1s≦tex<2.5 s (10)
TLow=max(Ave5i+Offi,AveHold+Off)if2.5s≦tex<7 s (11)
TLow=Ave5i+Offiif 7 s≦tex (12)
Apart from the expiration threshold value TLow it will be determined according to the following C or JavaScript Code (compare JavaScript Das umfassende Referenzwerk, David Flanagan, Translator Ralf Kuhnert et al., O'Reilly, Köln, ISBN 3-930673-56-8) whether the set pressure SetPressure predefined by parts of the software has already been reached and the actual pressure measured by the pressure sensor 36 remains stable on the set pressure. The code line numbers on the right-hand edge of the page are not part of the code, but merely serve reference purposes. In step 80, the following code is run through every 10 ms.
Even though the set pressure SetPressure is reduced stepwise when switching from the inspiration mode to the expiration mode, the actual pressure ActualPressure remains initially unchanged and is then gradually reduced to the new set pressure. During this phase the variable PressureReached=0. After the checking !PressureReached in code line 1, where PressureReached is inverted by “!”, it is checked in this phase in code line 2 whether ActualPressure<=SetPressure, which is not the case as yet. ActualPressure has dropped to or below SetPressure only in a later code run. Now, PressureReached is incremented in code line 2, i.e. =1 is set. In the next code run, after the checking in code line 1, PressureReached is incremented in code line 4. Then, the inverted variable PressureStable is verfied in code line 5. PressureStable is initially 0. Therefore, it is subsequently verified in code line 6 whether ActualPressure>=SetPressure. If the actual pressure ActualPressure overshoots below SetPressure, this condition is not fulfilled till the end of the overshoot. In the next code run, the condition !PressureStable is no longer fulfilled, so that code line 6 is skipped. PressureReached and PressureStable are further incremented in the following code runs until the condition ActualPressure>SetPressure+LowerPressureThreshold Stable in code line 7 is fulfilled. LowerPressureThresholdStable here is a predetermined value in the range from 0.5 to 1 mbar. This may happen either with a second overshoot where the actual pressure again rises by LowerPressureThreshold Stable above the set pressure. A second possibility is coughing or sneezing. The pressure fluctuations then occurring are too fast for the pressure controlling of the bi-level apparatus, so that they are not adjusted. The variable PressureStable is evaluated in step 95 as to prevent a switching to the inspiration mode in the event of coughing or sneezing.
One of the following three conditions must be fulfilled to allow the switching from the expiration mode to the inspiration mode:
These conditions are also shown in the flowchart of
The conditions 2 and 3 are structured similarly and are verified in steps 93 and 95 to 97. According to these conditions a switching to the inspiration mode takes place if the airflow is increased (Fi>Fi-9+NF) when the patient takes a deep breath and the actual pressure simultaneously drops below the set pressure. In dependence on the variable PressureStable a differently high drop of the actual pressure below the set pressure is required in steps 96 and 97. LowerPressureThresholdUnstable and LowerPressureThresholdStable are predetermined constants.
The inspiration processing shown in
THigh=ExMinif0≦tin≦0.25 s (13)
Upon the expiration of the first 0.25 seconds it is verified in step 104 whether the previous maximum airflow is greater during the current inspiration phase InMax than Ave5AtSC. Ave5AtSC is the moving average value over 500 measured airflow values at the time of the preceding switching from the expiration mode to the inspiration mode. If this is the case, the inspiration threshold value THigh is calculated in step 106 according to formula (14). Otherwise, the inspiration threshold value THigh is calculated in step 105 according to formula (15).
The reference numerals 20, 21, 22 and 23 in
Next, prior to the decision in step 107, the following C Code is executed. Like the rest of the inspiration processing, also the code is run through every 10 ms. The code lines 11 to 20 correspond to the above code lines 1 to 10. It must be considered, however, that the set pressure SetPressure is increased stepwise during the transition into the inspiration phase, so that the actual pressure approximates the set pressure from below at the beginning of the inspiration phase. This has the consequence that the relational operators “>” and “<” are currently interchanged.
In code lines 21 to 23 it is measured how long the actual pressure was above the set pressure. In the variable PressureAboveSet the time is counted up. The value of PressureAboveSet must be multiplied by 10 ms to actually obtain the time.
One of the following four conditions must be fulfilled to allow the switching from the inspiration mode to the expiration mode:
3) if ActualPressure>SetPressure+UpperPressureThreshold or
4) if PressureAboveSet≧25 & & Fi<THigh
The usual case for the switching to the expiration mode is that the conditions are fulfilled in steps 107 and 108, i.e. that the airflow Fi is smaller than the inspiration threshold value THigh and the derivative SlopeAvei is smaller than the inspiration derivative threshold value TSlopeDown.
The second condition, which is verified in step 109, merely concentrates on whether the airflow Fi drops below a predetermined threshold.
Moreover, switching to the expiration mode takes place in accordance with the third condition if the actual pressure increases by UpperPressureThreshold above the set pressure. This condition is verified in step 110.
Finally, in accordance with the fourth condition, the switching to the expiration mode takes place if the actual pressure is above the set pressure for more than 0.25 s and the airflow Fi simultaneously drops below the inspiration threshold value THigh, which is verified in step 111.
During the switching to expiration the Boolean variable bInspiration is set to false in step 112.
The invention was explained in more detail by means of preferred embodiments above. A person skilled in the art will appreciate, however, that various alterations and modifications may be made without departing from the gist of the invention. Therefore, the scope of protection will be defined by the hereinafter following claims and their equivalents.
Number | Date | Country | Kind |
---|---|---|---|
10 2004 014 538 | Mar 2004 | DE | national |
This application is a continuation of international application number PCT/DE2005/000534 (publication number: WO 2005/092414 A1) filed on Mar. 23, 2005 and entitled METHOD FOR CONTROL OF A BI-LEVEL DEVICE AND BI-LEVEL DEVICE and claims the benefit of the above-mentioned international application and the corresponding German national patent application number 10 2004 014 538.5 filed on Mar. 23, 2004 and entitled VERFAHREN ZUR STEUERUNG EINES BILEVEL-GERÄTS SOWIE BILEVEL-GERÄT the contents of which are expressly incorporated herein by reference.
Number | Name | Date | Kind |
---|---|---|---|
4448192 | Stawitcke et al. | May 1984 | A |
5433193 | Sanders et al. | Jul 1995 | A |
5494028 | DeVries et al. | Feb 1996 | A |
5647351 | Weismann et al. | Jul 1997 | A |
5865173 | Froehlich | Feb 1999 | A |
6345619 | Finn | Feb 2002 | B1 |
6626174 | Genger et al. | Sep 2003 | B1 |
6626176 | Madaus et al. | Sep 2003 | B1 |
6990980 | Richey, II | Jan 2006 | B2 |
7000612 | Jafari et al. | Feb 2006 | B2 |
7089937 | Berthon-Jones et al. | Aug 2006 | B2 |
7267122 | Hill | Sep 2007 | B2 |
20020185131 | Madaus et al. | Dec 2002 | A1 |
20030221689 | Berthon-Jones | Dec 2003 | A1 |
20050211249 | Wagner et al. | Sep 2005 | A1 |
20050217671 | Fisher et al. | Oct 2005 | A1 |
20060027234 | Gradon et al. | Feb 2006 | A1 |
20060102179 | Rapoport et al. | May 2006 | A1 |
20060237015 | Berthon-Jones et al. | Oct 2006 | A1 |
20080283060 | Bassin | Nov 2008 | A1 |
20090038617 | Berthon-Jones et al. | Feb 2009 | A1 |
Number | Date | Country |
---|---|---|
19849571 | May 2000 | DE |
102 53 946 | Jan 2004 | DE |
0 656 216 | Jun 1995 | EP |
1 005 829 | Jun 2000 | EP |
WO 9835715 | Aug 1998 | WO |
WO 0010633 | Mar 2000 | WO |
0024446 | May 2000 | WO |
0174430 | Oct 2001 | WO |
0200283 | Jan 2002 | WO |
02083221 | Oct 2002 | WO |
Number | Date | Country | |
---|---|---|---|
20070089745 A1 | Apr 2007 | US |
Number | Date | Country | |
---|---|---|---|
Parent | PCT/DE2005/000534 | Mar 2005 | US |
Child | 11526487 | US |