WO2009115960A1

WO2009115960A1 - Method and system for processing a patient signal derived from a patient sensor

Info

Publication number: WO2009115960A1
Application number: PCT/IB2009/051041
Authority: WO
Inventors: Rolf Neumann; Guenter Gegner; Harald Greiner; Gerhard Tivig
Original assignee: Philips Intellectual Property & Standards Gmbh; Koninklijke Philips Electronics N. V.
Priority date: 2008-03-17
Filing date: 2009-03-13
Publication date: 2009-09-24

Abstract

A method for processing a patient signal derived from a patient sensor (100) sensing a physiological parameter of a patient, comprising the following steps: processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes, and for each purpose, further processing the respective processed signal according to a method which is unique for the respective purpose, yielding a unique purpose output for each purpose. Accordingly, the invention provides for such a method for processing a patient signal derived from a patient sensor sensing a physiological parameter of a patient which is versatile and efficient for multi purposes.

Description

METHOD AND SYSTEM FOR PROCESSING A PATIENT SIGNAL DERIVED FROM A PATIENT SENSOR

FIELD OF THE INVENTION

The invention relates to the field of processing a patient signal, and especially to the field of patient monitoring and controlling. BACKGROUND OF THE INVENTION In a classic patient monitoring device, e.g. a pulse oximeter, the goal of the signal processing is to provide the user with a patient parameter, e.g. saturation value, for diagnosis purposes. This includes the presentation of the patient parameter as well as corresponding alarming and trending information.

For example, a pulse oximeter delivers a base signal which is first processed in order to yield a pre-processed signal which then undergoes a saturation algorithm for diagnosis yielding a SpO2 value for diagnosis. There might be further averaging of the SpO2 value for a display output, further conditioning for alarming purposes, and further averaging and data reduction for trending. This means that there is further post-processing of the SpO2 value by averaging, conditioning and data reduction in order to deliver outputs (display, alarming, trending) for diagnosis.

In this example the difference in the post-processing of the patient signal (SpO2) might be that the averaging for the display output is a traditional time-based averaging, whereas the data reduction needed for the trend presentation of the data might involves a median filtering, but for alarming the extreme values might be of interest and there is only a delay or hysteresis involved in the conditioning for the alarming.

Based on one base signal or a set of multiple base signals, e.g. based one sensor signal, there is sometimes the possibility to derive more that only one physiological value of the patient signal. For example, a pulse oximeter can derive the saturation value, the pulse rate and a perfusion index from the one set of base signals obtained with the pulse oximeter sensor. There are also devices that use separate signal processing for a common patient signal. Furthermore, there are also methods available which use multi-algorithmic analysis of the signal for the processing for the patient parameter for one purpose, but all with the goal to finally provide a single patient parameter optimized for the one purpose, i.e. diagnosis.

If the patient signal shall be used for more than one purpose, e.g. for diagnosis and as an input for a closed loop controller, the currently available methods offer two options:

First, the patient signal provided for diagnosis purposes is also used as input for the closed loop controller. The disadvantage is that the patient signal for diagnosis is optimized for this specific purpose and does not provide the optimal performance for the closed loop system. Either one accepts the non-optimal performance or one tries to apply algorithms to generate a more optimized control input based on the diagnosis patient signal. It is obvious that all the processing and filtering done in the algorithm for diagnosis cannot be inversed and this solution cannot make use of all the signal properties that might be helpful for a closed loop controller.

Second, to obtain a patient signal in parallel to the diagnosis purpose, a second system with a separate signal acquisition is used that itself is now optimized for the closed loop purpose. This option is somehow cumbersome as typically this involves an additional sensor at the patient and sometimes an additional device.

SUMMARY OF THE INVENTION

It is the object of the invention to provide such a method and system for processing a patient signal derived from a patient sensor sensing a physiological parameter of a patient which are versatile and efficient for multi purposes.

This object is achieved by a method for processing a patient signal derived from a patient sensor sensing a physiological parameter of a patient, comprising the following steps: processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes, and for each purpose, further processing the respective processed signal according to a method which is unique for the respective purpose, yielding a unique purpose output for each purpose.

This means that the invention overcomes the limitations of the prior art by providing more than one output of the same patient signal, each processed in a specialized way to be optimized for a specific purpose. This allows using only one patient sensor and one signal acquisition part, e.g. the patient connection that is already used for diagnosis purposes. On the other hand, the patient signal can be optimized for several purposes without sacrificing performance for another purpose. Preferred purposes comprise: traditional patient monitoring, patient monitoring for polysomnography or sleep studies, input for a controller, preferably a closed loop controller, and most preferably a controller that requires very robust values with high confidence levels, input for a multi-parameter alarm optimization algorithm that preferably does not allow significant processing delays on the one side, but requires a high confidence on the other side, input for a multi-parameter algorithm that preferably does not allow significant processing delays, input for a data analysis algorithm, which preferably requires a high fidelity of the patient signal.

It is to be noted that in the context of the present invention, the term "patient" does not only apply to human beings but also to animals. Further, the term "patient" does not mean that the respective person/animal is disease-ridden and, thus, also healthy persons who make part of a medical system will be referred to as "patients".

According to a preferred embodiment of the invention, the different processed signals comprise different data based on the same physiological parameter. This means that there can be different types of data, or same types of data calculated in a different way. For example, SpO2 can be calculated in different ways for diagnosis and control purposes, respectively. Further, the patient signal from a pulse oximeter can yield SpO2 and pulse rate, i.e. different types of data.

In general, processing can be done with only one algorithm. However, according to a preferred embodiment of the invention, at least one way of processing the patient signal for a purpose comprises processing the patient signal with multiple sub algorithms and selecting one of the signals which result from the different sub algorithms for further processing. This can be helpful in order to improve signal quality.

Further, according to a preferred embodiment of the invention, at least one unique purpose output is used as an input for an advanced processing of another patient parameter. Furthermore, according to a preferred embodiment of the invention, at least one unique purpose output is used as input for advanced alarming of another patient parameter. Finally, according to a preferred embodiment of the invention, at least one unique purpose output is used together with another parameter to obtain an additional patient parameter. This eliminates the need for a respective sensor. For example, a pulse wave propagation delay measured with optimized pulse oximeter processing and normal or optimized ECG processing can be used to derive an estimation of the mean pressure of the patient.

Above mentioned object is further met by a system for processing a patient signal, with a patient sensor for sensing a physiological parameter of a patient, multiple first processing units for processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes, and for each purpose, a respective second processing unit for processing the respective processed signal according to a method which is unique for the respective purpose, yielding a unique purpose output for each purpose.

Preferred embodiments of the system according to the invention result from the preferred embodiments of the method according to the invention as described above.

Especially, according to a preferred embodiment of the invention, the first processing units are each adapted for yielding different processed signals which comprise different data based on the same physiological parameter. Further, according to a preferred embodiment of the invention, at least one of the first processing units is adapted for processing the patient signal with multiple sub algorithms and selecting one of the signals which result from the different sub algorithms for further processing in the second processing unit.

Moreover, according to a preferred embodiment of the invention, an advanced processing unit is provided which is adapted for inputting at least one unique purpose output for advanced processing of another patient parameter. Further, it is preferred that an advanced alarming processing unit is provided which is adapted for inputting at least one unique purpose output for advanced alarming of another patient parameter. Finally, a patient parameter calculation unit is provided which is adapted for inputting at least one unique purpose output together with another parameter to obtain an additional patient parameter. BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. In the drawings:

Fig. 1 schematically shows the general principle of the invention, Fig. 2 shows a schematic depiction of a first preferred embodiment of the invention, and

Fig. 3 shows a schematic depiction of a second preferred embodiment of the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

From Fig. 1, the general principle of the invention can be seen. There, schematically a system for processing a patient signal is shown which comprises a patient sensor 100 for sensing a physiological parameter of a patient and multiple first processing units 101, 102 for processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes. As set out in more detail in the following, such different purposes can be diagnosis, control, support for other patient signal etc. Further, for each purpose, a respective second processing unit 103, 104 is provided which is suited for processing the respective processed signal according to a method which is unique for the respective purpose and, thus, yields a unique purpose output for each purpose, i.e. adapted for diagnosis, control, support for other patient signal, respectively.

Though the first processing units 101, 102, and second processing units 103, 104 are shown as different units, this is for only for reasons of comprehensibility. Actually, the first processing unit 101 and the second processing units 103 can be integrated into a common unit. The same applies for other first processing units and second processing units which are connected in series.

According to the principle of the invention, an advanced processing unit 105 can be provided which is adapted for inputting at least one unique purpose output for advanced processing of another patient parameter. Further, an advanced alarming processing unit 106 can be provided which is adapted for inputting at least one unique purpose output for advanced alarming of another patient parameter. Furthermore, a patient parameter calculation unit 107 can be provided which is adapted for inputting at least one unique purpose output together with another parameter to obtain an additional patient parameter. This will be described in more detail in the following in connection with the preferred embodiments of the invention.

From Fig. 2 a schematic depiction of a first preferred embodiment of the invention can be seen which shows processing of base signal 1 derived from a pulse oximeter. This base signal is pre-processed into a pre-processed signal 2 and then further processed in a saturation algorithm for diagnosis 3 and a saturation algorithm for control 4, i.e. for diagnosis and control purpose, respectively. This yields a SpO2 value for diagnosis 5 and a SpO2 value for control 6. In other words, the patient signal 1 derived from the pulse oximeter is processed in two different ways for different purposes, i.e. diagnosis and control, respectively, using different processing algorithms, i.e. an algorithm for diagnosis 3 and an algorithm for control 4, respectively, yielding different processed signals which are suited for different purposes, i.e. SpO2 value for diagnosis 5 and SpO2 for control 6, respectively.

The SpO2 for diagnosis value 5 is further processed according to methods which are typical and unique for diagnosis purposes, i.e. according to a postprocessing 7, steps of averaging 8, conditioning 9, and averaging and data reduction 10 are performed, respectively. Then, averaged data is used for display output 11, conditioned data is used for alarming 12, and averaged and reduced data is used for trending 13, generally referred to as the outputs for diagnosis 14. The SpO2 for control value is further processed in a conditioning step 15 and then output to a controller 16. This means that for each purpose, i.e. diagnosis and control, respectively, further processing of the respective processed signal, i.e. SpO2 for diagnosis 5 and SpO2 for control 6, respectively, is done according to a method which is unique for the respective purpose, i.e. averaging 8, conditioning 9 and averaging and data reduction 10 for the diagnosis purpose, and conditioning 15 for the control purpose, yielding a unique purpose output for each purpose, i.e. display out 11, alarming 12 and trending 13 as outputs for diagnosis 14, and output to controller 16, respectively.

From Fig. 3, a second preferred embodiment of the invention can be seen which is more complex. Similar to the first preferred embodiment described above, a base signal 1 derived from a pulse oximeter is pre-processed into a pre-processed signal 2 and then further processed in a saturation algorithm for diagnosis 3 and a saturation algorithm for control 4, i.e. for diagnosis and control purpose, respectively. In contrast to the first preferred embodiment described above, the saturation algorithm for diagnosis 3 and the saturation algorithm for control 4 both comprise sub algorithms for diagnosis 21, 22, 23, and sub algorithms for control 25, 26, 27, respectively.

These sub algorithms 21, 22, 23, 25, 26, 27 process the pre-processed data in different ways, respectively. Both, in the diagnosis case and in the control case, a respective decider 24, 28 is provided to decide for one signal achieved by the sub algorithms 21, 22, 23, 25, 26, 27. This way an SpO2 value for diagnosis 5 and a SpO2 value for control 6 are achieved, respectively, which are further processed in a way similar to the first preferred embodiment described above.

Further to the SpO2 value for diagnosis 5 and the SpO2 value for control 6, according to the second preferred embodiment of the invention, a pulse rate value for diagnosis 17 and a pulse rate value for ECG alarm support 18 are compiled. For that, the pre-processed signal 2 is input into a pulse rate algorithm for diagnosis 35 which comprises two sub algorithms 29, 30 and a decider 31, and into a pulse rate algorithm for ECG alarm support 36 which comprises two sub algorithms 32, 33 and a decider 34. This way, above mentioned pulse rate value for diagnosis 17 and pulse rate value for ECG alarm support 18 are achieved.

In dependence on their respective purposes, i.e. diagnosis and ECG alarm support, respectively, the pulse rate value for diagnosis 17 and the pulse rate value for ECG alarm support 18 are further processed. Further processing of the pulse rate value for diagnosis 17 is done in the "diagnosis way" comprising steps 7 to 14 as described above with respect to the further processing of the SpO2 value for diagnosis 5.

In other words, steps 7 to 14 represent the method which is unique for the respective purpose, i.e. diagnosis. In contrast to that, the pulse rate value for ECG alarm support 18 is further processed according to a method which is typical an unique for the purpose "ECG alarm support". This means that the pulse rate value for ECG alarm support 18 undergoes a conditioning step 19 followed by output to ECG alarm system 20. According to the second preferred embodiment of the invention, the

SpO2 for diagnosis purpose and the pulse rate for diagnosis purpose are used in a traditional patient monitoring way for presentation of the current patient state, alarming and patient data trending. The SpO2 for control purpose is used as input for a closed loop system that controls the FiO2 concentration that is administered to the patient. The SpO2 value for this purpose is processed in such a way that provides for a very robust output, i.e. when a value is reported it has very high confidence and if the confidence is to low the controller is informed that no new SpO2 data is currently available. To obtain this robust SpO2 value the response time is rather slow and a relative high amount of drop-out times, i.e. times where no SpO2 value is provided, is accepted. This behavior would typically not be accepted for the patient monitoring purpose, i.e. for diagnosis.

Further, according to the second preferred embodiment of the invention, the pulse rate for ECG alarm support purpose is used as input for the ECG alarm processing in such a way that as long as valid pulses can be measured, a potential ECG asystoly alarm is suppressed for a predefined time. With this pulse rate input the ECG alarm system can better differentiate between true asystoly conditions and signal disturbance conditions that in the past have caused false alarms. To be used as input for such a critical processing, the pulse rate value for this purpose has to be very robust, i.e. only at high confidence levels the pulse rate is provided/used by the alarm system, and the processing delay has to be minimal. This behavior would typically not be accepted for the patient monitoring purpose, i.e. for diagnosis, because of the expected high amount of drop-out times and the very dynamic behavior of the numerical values.

Another embodiment of the invention for which patient parameter data used as input for advanced diagnostic algorithms is an early warning score using "resting respiration" as set out in the following:

The respiration rate of a patient (RESP) is routinely used for monitoring purposes and this measurement provides alarming capabilities to inform the user when the patient condition becomes critical and might need immediate intervention. For the immediate threshold alarming capability and also the presentation of the patient's current status, the user expects the RESP value to represent the actual respiration rate of the patient.

According to this embodiment of the invention, there is an enhanced additional alarming functionality, i.e. an early warning score, that uses several parameters in combination to provide an early indication that the patient condition is degrading and that the patient might be on risk to collapse in the future. As the patient is typically not yet in a critical condition when the early warning score triggers an alarm, the expected response time of the clinicians is more in the order of an hour and not a few minutes. Therefore the required updated rate of the input parameters for this purpose is very relaxed compared to the standard continuous real-time monitoring.

To be sensitive as an early indication that the patient is degrading the early warning score uses sensitive thresholds for the different input parameters. To avoid unnecessary false positive alarms the RESP value that is to be used for this early warning score method is the so called "resting respiration". The resting respiration means the respiration value of a patient that is observed when the patient is resting. The typical way when this value is obtained manually is that the doctor/nurse ensures that the patient is resting while the measurement is taken.

For an automated calculation of the early warning score an algorithm is used that can detect when the patient is in a resting condition. An enhanced algorithm to calculate the resting respiration is not just a simple post processing of the real-time RESP value that is used for monitoring purposes. For example, a simple average of the real-time RESP value could potentially be false high if the averaging time is short and it would be too damped if the average time is long. An enhanced algorithm for the resting respiration has access to the "raw data" to allow applying rules to determine the points in time when the patient is in a resting condition and calculate the "resting respiration" accordingly. On the other hand, the RESP value for monitoring is not generated from the resting respiration value, as there are "non-resting" conditions for a monitored patient that can be life threatening and which need immediate intervention of the doctor/nurse. Further, there is another area of application for the invention, i.e. patient parameter data used as input for advanced algorithm of another patient parameter. According to an embodiment of the inventin, maternal pulse rate from SpO2 measurement is used for coincident check of fetal heart rate obtained via ultrasound as set out in the following: There is a potential problem with fetal heart rate (fHR) monitoring obtained via an ultrasound measurement: It is possible that not the heart of the fetus is focused by the ultrasound beam, but a large vessel of the mother. If this occurs the midwives might falsely be reassured that the unborn is doing fine and they are not aware that they don't monitor the fetus, which might actually be in trouble. To overcome this problem there is the possibility to warn the midwives if the fetal monitor device encounters that the fetal heart rate is the same as the maternal heart or pulse rate. To make such a coincident check it would be best if the maternal pulse rate would have similar dynamic properties as the fetal heart rate measurement. Such a measurement of the maternal pulse rate could be derived from a pulse oximeter sensor applied to the mother. As the fetal heart rate needs to have a dynamic behavior in a more or less in the order of the beat-to-beat resolution this would be the requirement also for the maternal pulse rate for an optimal coincident check.

To fulfill this requirement such an algorithm is optimized to be as fast as an beat-to-beat pulse rate value to show the maximum dynamic behavior. A possible approach is to use the raw data and apply a threshold based beat finder and once a beat could be identified calculate the current pulse rate value for that beat. If the signal would be heavily disturbed by artefacts, this algorithm cannot provide a valid output, but in general the duration when no output is possible will be limited to several seconds or in the worst case to a few minutes. This is a sufficient coverage for the coincident check described above.

On the other hand, the mother also often needs monitoring of her condition, especially if sedation is used during labor. To monitor the mothers condition the properties of the pulse rate values are significant different. Those pulse rate values do not need to be as dynamic, but should be much more robust and the times of no value available should be minimized even under artefact conditions. Such a "monitoring algorithm" uses more raw data to analyze the signal over a longer period of time, e.g. collect several heart beats to do the calculations or a time window of 5 to 15 seconds to do a FFT transformation of the signal. The pulse rate values resulting from such an algorithm are not a dynamic as the beat-to-beat type algorithms described above, but somehow more robust against artefacts and still responsive enough to ensure an alarming with a delay only in the order of 10-30 seconds. Again, a good solution for this application is to use two different algorithms that both use the raw data from the same sensor, but which are optimized differently in order to provide the same physiological parameter with differently optimized properties for each purpose.

The applications of the invention are the combination of a standard patient monitoring device that typically is used for diagnosis purposes (as described above) with an additional application that is intended to make use of the same patient parameter, but the properties (e.g. the dynamic properties) needed by this additional application are different to the properties of the standard patient monitoring. Therefore, a different processing of the base signal(s) can optimize the performance of the additional application, i.e. the additional purpose.

Examples for such an additional application are the following:

Patient parameter data used for pattern recognition algorithms that require a maximum of fidelity of the provided numerical values.

Especially pulse oximeter data (SpO2) used as input for pattern algorithm as described by Lawrence A. Lynn (Lyntek Technologies) to evaluate desaturation patterns in order to detect upper airway instability in patients. Patient parameter data used for polysomnography and sleep studies.

Patient parameter data used as input for control systems, e.g. closed loop controls. The purpose of controlling is not limited to a single type of control loop and includes several variations and different implementations of control loops, e.g. closed loop controls, supervised controls and open loop controls.

One of ordinary skill in the art will appreciate many variations and modifications within the scope of this invention. This method and system will be used mainly for hospitalized patients, but there are also applications possible for mobile patients in the hospital environment, during transport or at home. Also devices could make use of this invention that are intended for healthy persons or even animals. Further, while the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments.

Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. A method for processing a patient signal derived from a patient sensor

(100) sensing a physiological parameter of the patient, comprising the following steps: processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes, and for each purpose, further processing the respective processed signal according to a method which is unique for the respective purpose, yielding a unique purpose output for each purpose.

2. The method according to claim 1, wherein the different processed signals comprise different data based on the same physiological parameter.

3. The method according to claim 1 or 2, wherein at least one way of processing the patient signal for a purpose comprises processing the patient signal with multiple sub algorithms and selecting one of the signals which result from the different sub algorithms for further processing.

4. The method according to any of claims 1 to 3, wherein the different purposes comprise one or more of: patient monitoring, patient monitoring for polysomnography or sleep studies, input for a controller, especially a closed loop controller, input for a alarm algorithm, especially a multi-parameter alarm algorithm, input for a multiparameter algorithm, input for a data analysis algorithm.

5. The method according to any of claims 1 to 4, wherein at least one unique purpose output is used as an input for an advanced processing of another patient parameter.

6. The method according to any of claims 1 to 5, wherein at least one unique purpose output is used as input for advanced alarming of another patient parameter.

7. The method according to any of claims 1 to 6, wherein at least one unique purpose output is used together with another parameter to obtain an additional patient parameter.

8. A system for processing a patient signal, with a patient sensor (100) for sensing a physiological parameter of a patient, multiple first processing units (101, 102) for processing the patient signal in at least two different ways for different purposes using different processing algorithms and, thus, yielding different processed signals which are suited for different purposes, and for each purpose, a respective second processing unit (103, 104) for processing the respective processed signal according to a method which is unique for the respective purpose, yielding a unique purpose output for each purpose.

9. The system according to claim 8, wherein the first processing units (101, 102) are each adapted for yielding different processed signals which comprise different data based on the same physiological parameter.

10. The system according to claim 8 or 9, wherein at least one of the first processing units (101, 102) is adapted for processing the patient signal with multiple sub algorithms and selecting one of the signals which result from the different sub algorithms for further processing in the second processing unit.

11. The system according to any of claims 8 to 10, wherein the different purposes comprise: patient monitoring, patient monitoring for polysomnography or sleep studies, input for a controller, especially a closed loop controller, input for a alarm algorithm, especially a multi-parameter alarm algorithm, input for a multi-parameter algorithm, input for a data analysis algorithm.

12. The system according to any of claims 8 to 11, wherein an advanced processing unit (105) is provided which is adapted for inputting at least one unique purpose output for advanced processing of another patient parameter.

13. The system according to any of claims 8 to 12, wherein an advanced alarming processing unit (106) is provided which is adapted for inputting at least one unique purpose output for advanced alarming of another patient parameter.

14. The system according to any of claims 8 to 13, wherein a patient parameter calculation unit (107) is provided which is adapted for inputting at least one unique purpose output together with another parameter to obtain an additional patient parameter.