WO2016009240A1 - A clinical waveform algorithm comparing method and device - Google Patents

A method and a device for testing trial clinical waveform algorithms on existing medical devices at run time are disclosed. The method comprises a first step of receiving at least one medical signal data by an event processing engine. The second step involves branching the medical signal data into a first and a second branch data. Next, a step of executing a certified factory algorithm available on the existing medical device to process the first branch data and delivering a factory algorithm result as a first output is carried out. Similarly, in the next step a trial clinical waveform algorithm is executed to process the second branch data and deliver a research algorithm result as a second output. A final step of comparing the factory algorithm result with the research algorithm result by displaying the first output on a primary display and the second output on a research display is executed.

A CLINICAL WAVEFORM ALGORITHM COMPARING METHOD AND DEVICE

The present technology relates to methods and devices for testing trial clinical waveform algorithms on existing medical devices.

In an ordinary hospital setup a physician uses a pre-programmed medical device to measure medical signals generated by a patient. These medical devices are hard programmed with certified factory algorithms at the time these devices are manufactured at the factory and delivered to the hospitals. These certified factory algorithms have passed the necessary tests and have been approved by the regulatory governmental bodies that certify whether these algorithms are fit for use on patients. The physician simply sends in the input through input probes of the medical devices, or any other means for input, thereafter the signals are processed by the factory algorithms available with the device and subsequently output is generated accordingly. This output is hereafter observed and studied by the physician to diagnose the patient's medical condition.

Traditionally, implementing new algorithms in software is time consuming with testing performed only in simulation. Most algorithms are usually written in Matlab and subsequently converted to C code to include in the clinical application. Additionally, code changes are required to be made to the application which further require time to build and test before being shipped to the client and then deployed. This takes a significant amount of time and discourages research customers from using such applications for clinical waveform algorithm testing/trials. Cardiologists export the data acquired during a study to a file and then use Matlab to run their algorithms and test them at a later point of time and not during the patient study or run time.

While using medical devices, for examples devices used by cardiologists working on clinical waveforms, a physician may observe that the factory algorithm being used in the medical device, which is embedded into the device for processing medical signals, can be improved if certain changes could be made to the factory algorithm the result obtained can be more satisfactory. These changes can be based on the observation, study and research performed by the physician. The physician then suggests these changes to the device manufacturer for incorporating them into the factory algorithm and thus, have the devices reprogrammed by the device manufacturer. Once the device manufacturer receives such suggestions for improving or changing the factory algorithms, he reprograms the software to incorporate the changes into the factory algorithms and thus the device is reprogrammed with the new algorithm and then the device is sent back to the physician.

Typically this process may take a long time and may vary from a few weeks to as long as a year or may be more. Once the physician receives the device he performs tests using the trial clinical waveform algorithm to check the efficiency of the new algorithm over the old factory algorithm. Once the physician is satisfied of the result he gives his approval to the device manufacturer who then sends the device for all the necessary regulatory approvals and checks needed for all medical instruments before they are put to use in hospitals or clinics. In this long process crucial time is lost in sending the suggestions to the device manufacturer, getting the device reprogrammed with the trial algorithm and testing the efficiency of the new algorithm and if necessary making further suggestions to the device manufacturer for any further changes necessary and finally sending the device for manufacturing with the trial waveform algorithm and receiving the reprogrammed device from the manufacturer. This not only results in a huge loss of time but the physician is also compelled to work with a less efficient factory algorithm when there is an option to use a better algorithm which would give superior results.

United States Patent Application 20120171650 Al discloses a medical waveform development system comprising a medical waveform recording system for obtaining recorded medical waveforms. The system further comprises a physiological simulator for accessing recorded medical waveforms and for processing modified medical waveforms. The system also comprises an invasive workbench that allows a user to access the recorded medical waveforms and modified medical waveforms via software applications that are accessible on a computer. The idea behind this invention is to support electrophysiological research, training and development.

Provision for on-the-job development and testing of clinical waveform algorithms during run time is not available today. The doctor or the physician has to wait until the manufacturer reprograms the device and sends back, this turn around or waiting time leads to loss of crucial time in medical field and creates a sense of dissatisfaction and a growing discontent amongst physicians for having to work with pre-programmed algorithms which may not be giving the most effective results as desired by the physician. Therefore, there is a need to for a systemic approach that supports testing of trial clinical waveform algorithms on existing medical devices at run time, when the device is being used on patients, without affecting the workflow so that real on-the-job testing of the algorithms can be done at par with the existing factory algorithms.

It is an object of the present invention to present a framework that provides the ability to write trial clinical waveform algorithms, deploy them in an existing medical device or system and test these algorithms during real ongoing study or case.

The object of the invention is achieved by a method and a device for testing trial clinical waveform algorithms on existing medical devices at run time as disclosed according to this invention. The advantage offered by the disclosed method is that without any impact on the ongoing study it ensures maximum patient safety and minimal turnaround time, which includes the time taken for requiring the application code to be changed, built, tested, delivered and deployed onto the medical device. The disclosed method comprises of a first step of receiving at least one medical signal data by an event processing engine. The second step involves branching of the medical signal data into a first and a second branch data. Next, a step of executing a certified factory algorithm available on the existing medical device to process the first branch data and delivering a factory algorithm result as a first output is carried out. Similarly, in the next step a trial clinical waveform algorithm is executed to process the second branch data and deliver a research algorithm result as a second output. A final step of comparing the factory algorithm result with the research algorithm result by displaying the first output on a primary display and the second output on a research display is executed.

An event processing engine, or a complex event processing engine (CEP), is an event processing unit that combines data from multiple sources to infer events or patterns that suggest more complicated circumstances. CEP identifies and analyzes cause-and-effect relationships among events in real time. The goal of CEP engine is to identify meaningful events and respond to them as quickly as possible. It provides for filtering, correlating and aggregating of real-time event data in a low latency environment.

The proposed event processing engine or the complex event processing (CEP) engine is being used for dynamically adding and removing trial clinical waveform algorithms into existing medical devices and applications thereby reducing the time a clinical researcher needs to implement and test an algorithm during an ongoing case, i.e. during run time. This would make the device disclosed according to this invention a preferred buy for hospitals which perform cardiac research and similar high end clinics. The physicians would not only be able to use the device on patients for medical procedures but also be able to carry out research and testing of trial algorithms on the same device which would promote research, development and growth in the field of clinical waveform technology. Use of CEP engine provides many advantages. It is highly suitable for ID or one dimensional waveform processing. A ID waveform can be represented as a continuous stream of data. For instance, an ECG signal or an internal blood pressure signal is a sequence of data being generated by the ECG or pressure leads.

Using CEP we can achieve near real time computation of data streams and generate results. One can also add queries at runtime using CEP engine. It provides support for multiple queries. It has the ability to chain the output of one query to the input of another query thereby building complex algorithms. All these features of CEP together make it an ideal candidate for computing clinical ID waveform algorithms.

Another advantage of using CEP engine is the time advantage offered by Using a CEP from a clinical researcher's perspective. It processes multiple algorithms simultaneously without any delay. A CEP engine allows dynamic queries to be written and inserted into the processing engine at runtime. This allows a clinical researcher to write the algorithm and insert into the processing stage at any time. If we provide a framework or a software tool, where a researcher can write the algorithm as a CEP engine query, for instance in LINQ if using Streaminsight, we can insert the query for processing without having the need to alter code, build, test or deliver and deploy the software.

Moreover, the CEP engine provides the ability to see the algorithm performance in comparison to existing ones in real time. By providing a section for clinical trials in the application, the clinical researchers/cardiologists will get a novel way to try out new trial algorithms using existing medical devices and system, even in a live environment.

And above all, use of CEP engine would also ensure that patient safety is taken care of during such trial waveform algorithm testing. CEP allows different queries to run independently of each other. By making the queries which are certified correct, and shipped from the factory, to run as standing queries, i.e. queries which run independently and always, with predefined and fixed sources (input) and sinks (output), we can ensure that the new or trial clinical waveform algorithm queries do not impact these in any way. This would remove any hazard issues for the patient.

In one embodiment of the method the trial clinical waveform algorithm is entered via a user interface configured to receive the trial clinical waveform algorithm as an input. The user interface provides an easy platform to the user to make entry of the trial algorithms into a factory programmed existing medical device. It acts as a direct entry point for a physician into the device.

In a further embodiment the user interface is a console for writing the trial clinical waveform algorithm. The user, i.e. the doctor or the physician using the medical device, can make an entry of the trial clinical waveform algorithm that requires testing through the console. This console acts as an input device for convenient entering of the test or trial algorithm into the device. The console can have buttons, a touch screen and/or a display screen. The user can view the trial algorithm on the console and accordingly modify the trial algorithm so as to achieve the desired output results.

In one embodiment of the method the trial clinical waveform algorithm is entered via a user interface configured to receive the trial clinical waveform algorithm as an input. The user interface provides an easy platform to the user to make entry of the trial algorithms into a factory programmed existing medical device. It acts as a direct entry point for a physician into the device.

In a further embodiment the user interface is a console for writing the trial clinical waveform algorithm. The user, i.e. the doctor or the physician using the medical device, can make an entry of the trial clinical waveform algorithm that requires testing through the console. This console acts as an input device for convenient entering of the test or trial algorithm into the device. The console can have buttons, a touch screen and/or a display screen. The user can view the trial algorithm on the console and accordingly modify the trial algorithm so as to achieve the desired output results.

In an embodiment of the method, the trial clinical waveform algorithm executes independently from the certified factory algorithm to deliver the research algorithm result independent of the factory algorithm result. This ensures that the existing medical device keeps working normally and displaying results using algorithms that are pre-approved and medically certified, referred here as certified factory algorithms. Thus, by using the medical signal data from the patients as an input to test the trial clinical waveform algorithm, the medical device is playing a dual role wherein the first role of working with the certified factory algorithm is unaffected by the second role which involves the use of the CEP to test the trial clinical waveform algorithm.

In yet another embodiment, the medical signal data is selected from a group consisting of ECG signal data, IBP signal data, IECG signal data. These are only a few examples, however any other appropriate medical signal data can also be used.

According to an embodiment, the certified factory algorithm and the trial clinical waveform algorithm execute at runtime. This ensures that a real time result is obtained and no time is lost in data gathering and executing. In a further embodiment both the algorithms execute simultaneously. This will help the user to analyze and compare the time taken by both the algorithms to arrive at the output.

According to an embodiment, a device for testing trial clinical waveform algorithms on existing medical devices is disclosed. The device comprises of an event processing engine, a first algorithm unit, a second algorithm unit, a primary display and a research display. The event processing engine receives at least one medical signal data and branches the medical signal data into a first branch data and a second branch data. The first algorithm unit contains a certified factory algorithm and the second algorithm unit contains the trial clinical waveform algorithm, wherein the first and the second algorithm units supply the corresponding algorithms to the event processing engine to process the first and the second branch data respectively. The primary display displays a factory algorithm result obtained from processing of the first branch data by the certified factory algorithm. The research display displays a research algorithm result obtained from processing of the second branch data by the trial clinical waveform algorithm.

In an embodiment of the device the second algorithm unit further comprises a console for writing the trial clinical waveform algorithm. This console provides an interface to the user to enter the trial algorithms that he wants to test and also make appropriate changes to the trial algorithms accordingly.

In another embodiment, the primary display and the research display are a part of a single display unit. This feature enables the user to view the output generated by both the algorithms on the same display screen and make observations and comparisons to understand which algorithm gives better results.

In yet another embodiment of the device, the device further comprises a comparison unit for comparing the factory algorithm result with the research algorithm result. This further helps the user to evaluate the results obtained from using the trial clinical waveform algorithm in view of the results obtained from the certified factory algorithm to arrive at a conclusion on the effectiveness of using the trial clinical waveform algorithm over the certified factory algorithm. In one embodiment the comparison unit is a processor. In another embodiment the comparison unit is a comparator. The above-mentioned and other features of the invention will now be addressed with reference to the accompanying drawings of the present invention. The illustrated embodiments are intended to illustrate, but not limit the invention. The drawings contain the following figures, in which like numbers refer to like parts, throughout the description and drawing.

FIG.l is a schematic diagram showing a device for testing trial clinical waveform algorithms on existing medical devices.

FIG.2 shows a flowchart containing the steps according to the disclosed method.

As seen in FIG.l, the device 1 comprises of an event processing engine 2, a first algorithm unit 6, a second algorithm unit 8, a primary display 10 and a research display 12. The event processing engine 2 receives at least one medical signal data 3 and branches the medical signal data into a first branch data 4 and a second branch data 5. Both the first 4 and the second 5 branch data contain the same medical signal data 3. The first algorithm unit 6 contains a certified factory algorithm 7 and the second algorithm unit 8 contains the trial clinical waveform algorithm 9, wherein the first 6 and the second 8 algorithm units supply the corresponding algorithms 7, 9 to the event processing engine 2 to process the first 4 and the second 5 branch data respectively. The primary display 10 displays a factory algorithm result 11 obtained from processing of the first branch data 4 by the certified factory algorithm 7. The research display 12 displays a research algorithm result 13 obtained from processing of the second branch data 5 by the trial clinical waveform algorithm 9.

According to FIG.l, the second algorithm unit 8 further comprises a console 14 for writing the trial clinical waveform algorithm 9. This console 14 provides an interface to the user to enter the different types and variations of the trial algorithms 9 that the user wants to test and also make appropriate changes to the trial algorithms 9 accordingly.

As seen in FIG.l, the primary display 10 and the re-search display 12 are a part of a single display unit 15. The device as shown in FIG.l further comprises a comparison unit 16 for comparing the factory algorithm result 11 with the research algorithm result 13.

The method 100, as seen in FIG.2, comprises of a first step 101 of receiving at least one medical signal data 3 by an event processing engine 2. The second step 102 involves branching of the medical signal data 3 into a first 4 and a second 5 branch data. Next, a step 103 of executing a certified factory algorithm 7 available on the existing medical device to process the first branch data 4 and delivering a factory algorithm result 11 as a first output is carried out. Similarly, in the next step 104 a trial clinical waveform algorithm 9 is executed to process the second branch data 5 and deliver a research algorithm result 13 as a second output. A final step 105 of comparing the factory algorithm result 11 with the research algorithm result 13 by displaying the first output on a primary display 10 and the second output on a research display 12 is executed.

Although the invention has been described with reference to specific embodiments, this description is not meant to be construed in a limiting sense. Various modifications of the disclosed embodiments, as well as alternate embodiments of the invention, will become apparent to persons skilled in the art upon reference to the description of the invention. It is therefore contemplated that such modifications can be made without departing from the embodiments of the present invention as defined.

Claims :

1. A method (100) for testing trial clinical waveform algorithms on existing medical devices, the method (100) comprising:

a step (101) of receiving at least one medical signal data (3) by an event processing engine (2);

a step (102) of branching the medical signal data (3) into a first (4) and a second (5) branch data;

a step (103) of executing a certified factory algorithm (7) available on the existing medical device to process the first branch data (4) and deliver a factory algorithm result (11) as a first output;

a step (104) of executing a trial clinical waveform algorithm (9) to process the second branch data (5) and deliver a research algorithm result (13) as a second output; a step (105) of comparing the factory algorithm result (11) with the research algorithm result (13) by displaying the first output on a primary display (10) and the second output on a research display (12).

2. The method (100) according to the above claim, wherein the trial clinical waveform algorithm (9) is entered via a user interface configured to receive the trial clinical waveform algorithm (9) as an input.

3. The method (100) according to the above claim, wherein the user interface is a console (14) for writing the trial clinical waveform algorithm (9).

4. The method (100) according to any one of the above claims, wherein the trial clinical waveform algorithm (9) executes independently from the certified factory algorithm (7) to deliver the research algorithm result (13) independent of the factory algorithm result (11).

5. The method (100 according to any one of the above claims, wherein the medical signal data (3) is selected from a group consisting of ECG signal data, IBP signal data and IECG signal data.

6. The method (100) according to any one of the above claims, wherein the certified factory algorithm (7) and the trial clinical waveform algorithm (9) execute at runtime.

7. A device (1) for testing trial clinical waveform algorithms on existing medical devices, the device (1) comprises of:

an event processing engine (2) to receive at least one medical signal data (3) and to branch the medical signal data (3) into a first (4) and a second (5) branch data;

a first algorithm unit (6) containing a certified factory algorithm (7) and a second algorithm unit (8) containing the trial clinical waveform algorithm (9), wherein the first (6) and the second (8) algorithm units supply the corresponding algorithms to the event processing engine (2) to process the first (4) and the second (5) branch data respectively; a primary display (10) for displaying a factory algorithm result (11) obtained from processing of the first branch data (4) by the certified factory algorithm (7);

a research display (12) for displaying a research algorithm result (13) obtained from processing of the second branch data (5) by the trial clinical waveform algorithm (9).

8. The device (1) according to the above claim, wherein the second algorithm unit (8) further comprises a console (14) for writing the trial clinical waveform algorithm (9).

9. The device (1) according to any one of claims 7 or 8, wherein the primary display (10) and the research display (12) are a part of a single display unit (15).

10. The device (1) according to any one of claims 7 to 9, wherein the device (1) further comprises a comparison unit (16) for comparing the factory algorithm result (11) with the research algorithm result (13).