Method and apparatus for detecting vascular needle dislodgment

Abstract

A method for vascular needle disengagement detection is disclosed, which includes performing blood treatment and / or analysis on a patient (104) using a blood analysis loop (102) at a predetermined blood flow and recording a blood line pressure data set (202) at a predetermined time interval using a blood line pressure sensor (108) of the blood analysis loop (102). The method further includes filtering the recorded blood line pressure data set (202) using a processor configured to filter the recorded data set using a predetermined filter function, and validating the filtered blood line pressure data set (202) using a processor further configured to validate the filtered blood line pressure data set (202) based on a validation function. A potential vascular needle disengagement (618) is then detected based on the detection function resulting from the validating. Finally, the potential vascular needle disengagement (618) is confirmed or disproved by a static blood line pressure measurement performed by stopping the blood flow and keeping the blood line of the blood analysis loop (102) isolated from the patient (104).

Description

[Technical field]

[0001] The present invention relates to medical procedures. More particularly, the present invention relates to a method for vascular needle dislodgment detection, particularly for venous needle dislodgment. Furthermore, the present invention relates to an apparatus for vascular needle dislodgment detection, particularly for venous needle dislodgment detection. [Background technology]

[0002] In many medical procedures, blood is extracted from a patient. For example, hemodialysis therapy is used to remove waste, toxins and excess water from the patient via a blood analysis loop. The blood analysis loop pumps the patient's blood and reinfuses the blood into the patient after blood treatment by the hemodialysis machine. Needles are inserted into the patient's body, e.g., into arteries and veins, to transport the patient's blood to and from the hemodialysis machine. Summary of the Invention [Problem to be solved by the invention]

[0003] Needle dislodgement, particularly venous needle dislodgement, can occur accidentally during such blood treatments. If venous needle dislodgement, dislocation or migration occurs, it can rapidly become life threatening for the patient if not addressed promptly. Prior art devices and methods for needle dislodgement detection can lack sensitivity, accuracy and specificity for needle dislodgement. High sensitivity and accuracy of detection is essential to ensure and facilitate rapid response times to minimize blood loss caused by needle dislodgement.

[0004] It may therefore be desirable to provide an improved method for vascular needle dislodgment detection that mitigates at least some of the above-mentioned shortcomings of conventional systems and procedures. [Means for solving the problem]

[0005] The method aims to provide an improved method for detecting accidental blood vessel, e.g., venous or arterial needle detachment through improved blood line pressure data processing.

[0006] This is achieved by the subject matter of the independent claims, further embodiments of which are defined in the dependent claims and in the following description.

[0007] According to a first aspect of the present disclosure, there is provided a method for vascular needle dislodgment detection, the method comprising: performing blood treatments and / or analyses on a patient using a blood analysis loop at a given blood flow; recording a blood line pressure data set at predetermined time intervals using a blood line pressure sensor of the blood analysis loop; filtering the recorded bloodline pressure data set using a processor, the processor being configured to filter the recorded bloodline pressure data set using a predefined filter function, in particular a Kalman filter; validating the filtered bloodline pressure data set using a processor, the processor being configured to validate the filtered bloodline pressure data set based on a validation function, in particular a Wald test; Detecting potential vascular needle dislodgement based on a detection function λ; Confirm or refute potential vascular needle dislodgement by static vascular pressure measurements performed by stopping blood flow and keeping the blood line of the blood analysis loop separate from the patient; Includes.

[0008] Detecting a potential vascular needle dislodgment involves a comparison between a detection function λ and a predefined threshold B.

[0009] In other words, to detect vascular needle dislodgement, a bloodline pressure reference function may track internal slow pressure changes using process statistics.

[0010] In the context of this disclosure, vascular needle disengagement may refer to venous needle disengagement, arterial venous disengagement, or any other needle disengagement of a needle connected to any type of blood vessel of a patient.

[0011] In the context of this disclosure, the term "blood analysis loop" may refer to any device that draws blood from a patient and ultimately reinserts the blood into the patient, with or without blood treatment or blood analysis.

[0012] In particular, the term "blood vessel" may refer to a blood vessel in the human body. The term "blood line" may refer in particular to the (partially or completely) extracorporeal part of the (blood) circuit (or, respectively, the blood analysis loop).

[0013] The blood analysis loop may comprise several components and / or elements. The blood analysis loop may generally comprise a blood line and a blood pump. It may further comprise an inlet for, for example, introducing an anticoagulant into the blood flowing through the loop. The blood analysis loop may also comprise a blood purifier, treatment and / or analyzer. Such a blood purifier or machine may further be connected to a dialysate reservoir and a waste receiving container. The blood analysis loop may also comprise a blood line clamp, for example a venous clamp.

[0014] The blood analysis loop may also include one or more interfaces for interacting with a person and / or another device, such as a user interface and / or an interface for wireless communication.

[0015] The blood analysis loop may include two or more blood line pressure sensors all configured to perform recording of blood line pressure data sets at predetermined time intervals, which may function, for example, as a redundant measurement.

[0016] The processor used for the method for vascular needle dislodgment detection may be physically connected to (and / or be an integral part of) the blood analysis loop. Alternatively or additionally, an external processor may be provided and may interact with the blood analysis loop. The processor may be part of, for example, a computing device and / or a mobile device equipped with a wireless and / or wire-based interface. Alternatively and / or additionally, the blood line pressure sensor may be equipped with a wireless and / or wire-based interface.

[0017] The blood line pressure sensor may be a venous pressure sensor or an arterial pressure sensor. If the blood analysis loop includes a blood treatment or analyzer, the blood line pressure sensor may be located downstream of the blood treatment or analyzer. The blood pump may be located upstream of the blood line pressure sensor.

[0018] The filter function may be any mathematical algorithm that uses a series of measurements or a data set observed over time to generate an estimate of an unknown variable. A filter function according to the present invention may be more accurate than a filter function based only on a single measurement. Preferably, the filter function is a filter function based on a recursive algorithm. More preferably, the filter function is a Kalman filter function. Alternatively, the filter function may be any least squares based filter function.

[0019] After the dataset of recorded bloodline pressures has been filtered, the filtered dataset may be tested in a further step via a test function, such as, for example, a Wald test function. Alternatively or additionally, a Lagrange multiplier test and / or a likelihood ratio test may also be used as test function. Preferably, the test performs the following calculation:

[0020]

number

[0021] If λ(k)<0, then λ(k) may be reinitialized to 0. (k) may be a function that models the pressure change due to venous needle withdrawal, and σ 2 (k) may be the variance of the detection function λ(k).

[0022] A vascular needle detachment occurrence may be detected when the detection function (k) exceeds a threshold B, where the threshold B is given by:

[0023]

number

[0024] where α is the false alarm probability and β is the missed alarm probability. Note that the thresholds may be defined alternatively.

[0025] The threshold value B may be user controlled / controllable. Alternatively or additionally, the threshold value may be a blood analysis loop dependent parameter.

[0026] The proposed method may use an algorithm based on a combination of a filter function, e.g., a Kalman filter, and a test function, e.g., a Wald test function. The method for vascular needle dislodgement detection has been found to have a very high accuracy, in particular in confirming true vascular needle dislodgement and / or disproving potential false vascular needle dislodgement. The detection delay time of the method may be small, e.g., including 5 or less, in particular 3 or less (recording) steps, and / or lasting e.g., 50 seconds or less, preferably 30 seconds.

[0027] The sampling time T (=time difference between two samples) may be at least 2 seconds, or at least 5 seconds, or at least 8 seconds, and / or 20 seconds or less, or 12 seconds or less, for example 10 seconds.

[0028] According to one embodiment, the method further comprises initializing the filter function based on at least one statistical parameter, which is preferably blood analysis loop dependent.

[0029] The at least one statistical parameter may be, for example, the variance of the measurement noise and / or the variance of the process noise (i.e., measurement derivative noise, which in particular models the internal dynamics of the blood analysis loop), where both the process noise and the measurement noise may be time-dependent functions. The variance of the process noise may be calculated by dividing the recorded pressure data set into time segments and taking the slope of the linear regression for each time segment.

[0030] According to one embodiment, the method further comprises initializing the filter function when (whenever) a treatment parameter of the blood analysis loop is changed, in particular based on at least one statistical parameter, which may be blood analysis loop dependent.

[0031] The treatment parameters may include blood flow parameters (e.g., blood flow rate and / or blood flow rate) and / or blood pump parameters, in particular blood pump delivery rate (and / or blood pump rotation speed). Each time a treatment parameter is modified or changed, the filter function may be initialized or reinitialized. Thus, the filter function may have a time-dependent dynamic behavior. For example, the filter function may have a wide bandwidth at the beginning of the treatment and gradually narrow its bandwidth (especially as process knowledge increases). Increasing process knowledge may mean, for example, that the accuracy and / or efficiency of the filter function increases over time.

[0032] According to one embodiment, the method comprises: Automatically resume blood flow if potential vascular needle dislodgement is refuted Further includes:

[0033] In particular, if an actual vascular needle disengagement did not occur, the blood analysis loop may be controlled, such as restarting treatment and / or restarting the blood pump. The blood analysis loop may be controlled and / or operated by a processor. For example, the processor may operate a blood pump of the blood analysis loop based on whether a detected vascular needle disengagement was refuted.

[0034] This may have the advantage that medical personnel are not required to resume blood flow and / or restart the blood pump if the detected blood needle dislodgment happens to be a false positive.

[0035] According to one embodiment, the method comprises: If a potential vascular needle dislodgement is identified, the blood analysis loop remains stopped and / or an alarm is triggered Further includes:

[0036] Stopping the blood analysis loop may mean that the blood pump is stopped, preferably automatically. For example, the processor may trigger the blood pump and preferably (all) further elements of the blood analysis loop to stop when a potential vascular disengagement is identified. In particular, stopping the blood analysis loop may mean that the blood line clamp is closed.

[0037] Additionally, an alarm, e.g., as a visual or audio alarm, may also be triggered. The processor may trigger a mobile device or an alarm system to issue an alarm. The alarm may further include information about the patient for whom needle dislodgement is a concern. In particular, the alarm includes information about the geographic location of the patient, e.g., his / her room, and / or the treatment the patient is receiving.

[0038] According to one embodiment, the pre-defined threshold B is a function of the probability of false vascular needle dislodgment detection and of the probability of missed vascular needle dislodgment detection.

[0039] It is further contemplated that the predetermined thresholds may be reinitialized or initialized when treatment parameters of the blood analysis loop are changed.

[0040] According to one embodiment, the detection function λ is a time-dependent function that depends on the time-dependent difference between the recorded bloodline pressure data set and the filtered recorded bloodline pressure data set.

[0041] In other words, the detection function may depend on the function (k), i.e., the difference between the recorded venous pressure and the filtered recorded venous pressure, and may therefore be given by the following equation:

[0042]

number

[0043] where Z(k) may be the recorded blood line pressure, VP f (k) may be the filtered recorded blood line pressure.

[0044] According to one embodiment, the blood flow is constant during each predetermined time interval.

[0045] The blood flow may be controlled using the blood pump of the blood analysis loop.

[0046] It may be preferable for the blood pump of the blood analysis loop to be controlled to pump blood at a constant blood flow, particularly during each predetermined time interval of the blood treatment.

[0047] The blood pump may be controlled or operated by the processor or a further processor.

[0048] According to one embodiment, when the detection function λ exceeds a threshold B with a predetermined false alarm probability α and missed alarm probability β, detecting a potential vascular needle dislodgment is triggered based on at least the detection function λ resulting from the testing.

[0049] In the context of the present disclosure, the predefined confidence interval may include two alarm limits, one upper and one lower. The confidence interval may depend on the blood flow, i.e. on the configuration of the blood pump of the blood analysis loop. Exceeding the above upper and lower limits may imply an additional alarm, warning the user about a vascular access problem.

[0050] A further aspect of the disclosure relates to an apparatus for vascular needle disengagement detection. The apparatus comprises a blood analysis loop and a processor configured to control the blood analysis loop and detect potential vascular needle disengagement. The blood analysis loop is configured to perform blood therapy and / or analysis on the patient. The blood analysis loop further comprises a blood pump, a blood line pressure sensor, and a blood line clamp. The pressure sensor is configured to record a blood line pressure data set of the patient's blood and transmit the recorded blood line pressure data set to the processor. The processor is further configured to confirm or disprove the potential vascular needle disengagement and operate the blood line clamp and / or the blood pump based on confirming or disproving the potential vascular needle disengagement detection.

[0051] In general, the apparatus may be configured to carry out the methods as described above and below.

[0052] It should be noted that the processor may be further configured to operate, control and / or actuate some or all of the elements of the blood analysis loop.

[0053] According to one embodiment, the processor is further configured to filter the recorded bloodline pressure data set using a predefined filter function, in particular a Kalman filter.

[0054] According to one embodiment, the processor is further configured to test the difference between the recorded bloodline pressure data set and the filtered bloodline pressure data set based on a test function, in particular a Wald test.

[0055] According to one embodiment, the processor is further configured to instruct the pressure sensor to stop the blood pump of the blood analysis loop, keep the blood line clamp open, and perform static vascular pressure measurements, such as to confirm or refute potential vascular needle dislodgement.

[0056] According to one embodiment, a processor is communicatively connected to the blood analysis loop.

[0057] The communication connection between the processor and the blood analysis loop may be accomplished via a wireless connection and / or a cable connection. It is noted that the communication connection between the processor and the blood analysis loop is preferably a bidirectional connection such that data and / or information may be communicated (transmitted) in both directions.

[0058] The processor may be communicatively connected to several elements of the blood analysis loop, such as, for example, a blood pump, a blood line clamp, and / or a blood line pressure sensor of the blood analysis loop.

[0059] According to one embodiment, the processor is further configured to trigger an alarm when a potential vascular needle dislodgment is identified.

[0060] In general, any feature, function, and / or element described above and below with reference to one embodiment of the disclosure applies equally to any other embodiment of the disclosure, as described above and below.

[0061] It should be noted that different aspects of the present invention are disclosed with reference to different subject matter. In particular, some aspects are disclosed with reference to apparatus claims, while other aspects are disclosed with reference to method claims. However, a person skilled in the art may derive from the above and following descriptions that, unless otherwise specifically disclosed, any combination between features related to different categories of subject matter, in addition to any combination of features belonging to one category of subject matter, is considered to be disclosed by the present disclosure. In particular, combinations between features of apparatus claims and features of method claims may be disclosed.

[0062] These and other aspects of the invention will be apparent from and elucidated with reference to the exemplary embodiments described hereinafter.

[0063] The present invention will now be described with reference to the accompanying drawings, which contain a background description of the invention and specific embodiments thereof, and the scope of the invention is not limited to the specific features disclosed in the context of the drawings. [Brief description of the drawings]

[0064] [Figure 1] 1 illustrates an apparatus for vascular needle dislodgment detection according to an exemplary embodiment. [Diagram 2] FIG. 1 illustrates a blood line pressure data set recorded using a method according to an exemplary embodiment. [Diagram 3] FIG. 1 illustrates a state-of-the-art statistical blood line pressure mathematical model. [Figure 4] FIG. 1 illustrates a blood line pressure mathematical model according to an exemplary embodiment of the present invention. [Diagram 5] FIG. 13 illustrates a bloodline pressure data set according to a further exemplary embodiment of the present invention. [Figure 6a] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 6b] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 7a]1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 7b] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 8a] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 8b] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 9a] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 9b] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 10] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 11] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 12a] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. [Figure 12b] 1 illustrates a simulation of an apparatus for vascular needle dislodgment detection in accordance with an exemplary embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0065] The drawings are only schematic and are not to scale. As a rule, identical or similar parts, elements and / or steps are provided with identical or similar reference numbers in the figures.

[0066] FIG. 1 illustrates an apparatus for vascular needle dislodgment detection according to an exemplary embodiment.

[0067] The device 100 of Fig. 1 may be used to perform the methods described above and below. In particular, the device 100 of Fig. 1 may be used to perform blood treatment and / or analysis on a patient 104 using a blood analysis loop 102 at a predetermined blood flow. The blood analysis loop 102 may pump blood toward a blood treatment and / or analysis machine 114, etc., using a blood pump 118. During blood treatment and / or analysis, blood line pressure data sets are recorded at predetermined time intervals using a blood line pressure sensor 108 of the blood analysis loop 102. Reference numeral 109 denotes an optional blood line arterial pressure sensor.

[0068] The predetermined time interval may be referred to as a sampling step. The predetermined time interval may be, for example, 10 seconds, 8 seconds, or 6 seconds. The blood line pressure data set may include 1000 sampling steps. Thus, the recording of the vascular pressure data set may last, for example, 160 minutes.

[0069] The recorded bloodline pressure data set may be transmitted to a processor such that the processor may filter the data set. The data set may be transmitted wirelessly or using wires. The processor is configured to filter the recorded data set using a predefined filter function. A preferred filter function may be a Kalman filter (see FIG. 4).

[0070] The filtered data set is then tested by the processor based on a test function, which may be the so-called Wald test function.

[0071] Based on the detection function λ resulting from the calibration step, a potential venous or arterial disengagement may be detected. However, at this stage of the method, it may happen that the recorded pressure changes that lead to the detection of needle disengagement do not relate to an actual needle disengagement.

[0072] Therefore, a verification step is performed. After a potential needle dislodgement is detected, a static vascular pressure measurement is performed to confirm or disprove the potential needle dislodgement. To do so, blood flow is stopped while the blood line of the blood analysis loop 102 is not isolated from the patient 104. In other words, the blood pump 118 is stopped, but the blood clamp 106, e.g., the venous clamp 106, remains open during the static vascular pressure measurement.

[0073] If potential vascular needle disengagement is confirmed, the blood line clamp 106 may be closed and an alarm may be triggered, but if potential vascular needle disengagement is disproved, the blood pump 118 may be restarted and blood therapy continued.

[0074] The arrows in Figure 1 indicate the direction of blood flow into the blood analysis loop 102. Additionally, the blood analysis loop 102 of Figure 1 includes an inlet for introducing an anticoagulant 116 into the blood flowing through the loop. A blood treatment and / or analyzer 114 may be disposed on the vascular line of the blood analysis loop 102. The blood treatment and / or analyzer 114 is connected to the dialysate preparation device 110 and the waste receiving device 112.

[0075] 1 may be controlled, operated and / or actuated in whole or in part by a processor that may be located external or internal to the blood analysis loop. Accordingly, it is contemplated that the blood analysis loop may be equipped with one or more communication interfaces, such as a wireless interface (not shown).

[0076] 2 illustrates a bloodline pressure data set 202 recorded using a method according to an exemplary embodiment. In particular, the recorded bloodline pressure data set 202 is shown relative to the time of recording on the horizontal axis. Time may be shown in units of seconds. Pressure on the vertical axis may be shown in units of mmHg. The recorded pressure data set shown may be one of many data sets measured during a blood treatment or blood analysis.

[0077] A statistical parameter of the method according to the exemplary embodiment may be the variance of the slope of a number of data set segments. The internal fluctuations of the blood pressure measured or recorded in the patient's blood vessels may be represented by the so-called process noise. The process noise may be estimated by the variance of the slope of each data set segment lasting a given time, for example 1000 seconds or for example 100 sampling steps. The respective slopes may be calculated using a linear regression model 204.

[0078] Thus, for the pressure data set of Figure 2, ten slope coefficients may be calculated. Based on that, the variance of the slope may be calculated, which gives the process noise.

[0079] Calculating the slopes based on the linear regression model at each dataset segment may be performed for each recorded pressure dataset, and the global process noise may be calculated as the average of all process variances. Note that the variance of the slopes may be calculated based on the linear slope coefficients of each slope. The linear slope coefficients may be given in units of mmHg / sec.

[0080] In the description of Figures 3-12, it is assumed that the blood lines are venous lines (i.e., the blood vessels may be veins), but this should not be understood as a limitation: other blood vessels / blood lines may still be considered.

[0081] Figure 3 illustrates a venous pressure mathematical model. In particular, Figure 3 illustrates a filter function by which a recorded venous pressure data set can be filtered.

[0082] The variable X is the venous blood line pressure state, e.g., the venous pressure model state: X=(x1,x2). The variable x1 represents the combination of the venous pressure of the patient 104 and the pressure drop due to the connection of the blood line to the patient's blood vessel, and is dependent on time k. The variable x2 is the (time) derivative of x1, and therefore represents the rate of venous pressure change over time. The vector function F is represented by the box in FIG. 3 and is given by:

[0083]

number

[0084] where T is the sampling time and w(k) is the process noise, which models the internal fluctuations of blood pressure. Then, the venous pressure VP(K) or Z(k) affected by the measurement noise r(k) is given by

[0085]

number

[0086] Note that the process noise is only used when calculating x2.

[0087] 4 illustrates a blood line pressure mathematical model according to an exemplary embodiment of the present invention. In particular, FIG. 4 illustrates a Kalman filter function according to an exemplary embodiment of the present invention, by which a recorded venous pressure data set may be filtered.

[0088] The vector function K(k) is represented by a box in Figure 4 and is calculated based on the variance Q of the process noise w(k), i.e. an estimate of the variance of the gradient as described above, and the variance R of the measurement noise r(k). Note that the vector K is also a function of time k. The characteristics of the Kalman filter shown in Figure 4 change over time. Thus, the Kalman filter can closely track the output of any initialization and reduce / smooth out (at least partially) the measurement noise.

[0089] Function (k) is the difference between the recorded venous pressure and the filtered recorded venous pressure, therefore γ(k)=Z(k)-VP f (k) and may be used by the test function.

[0090] 5 illustrates a bloodline pressure dataset, e.g., a venous pressure dataset, according to a further exemplary embodiment. In particular, FIG. 5 illustrates a recorded venous pressure dataset 202 obtained by a method according to an exemplary embodiment of the present invention. The grey line in FIG. 5 represents a filtered venous pressure dataset 504, obtained by using a filter function, e.g., the filter function according to FIG. 4.

[0091] 6a and 6b show a simulation of a device for vascular needle dislodgment detection according to an exemplary embodiment. In FIG. 6a, a venous pressure data set 202 is shown. Time may be shown in units of seconds. Pressure on the vertical axis may be shown in units of mmHg. A corresponding filtered venous pressure data set 504 is also shown in the graph of FIG. 6a. The dotted lines 604 represent what are typically set as alarm limits for the venous pressure. It is noted that these limits are automatically set by default and may be manually adjusted in some cases. The two limits are sometimes referred to as confidence intervals. The thicker black line 602 represents the blood pump flow rate set 602, which is constant throughout the treatment period in the example of FIG. 6a and FIG. 6b.

[0092] Figure 6b corresponds to the result of a test function applied to the treatment shown in Figure 6a, in particular the Wald test function used in the method according to the exemplary embodiment. The horizontal axis of Figure 6b is time, for example given in seconds, and the vertical axis is the detection function λ.

[0093] The test function used, for example the Wald test function, is given by:

[0094]

number

[0095] If λ(k)<0, then λ(k) is reinitialized to 0. (k) is a function that models the pressure change due to venous needle withdrawal and is a constant, e.g., μ(k)=-40mmHg. σ 2 (k) is the variance of the detection function λ(k), i.e., a function that depends on Q, R and time k. The function γ(k) is given by equation (3). Finally, an occurrence 608 of venous needle detachment is detected when (if) the detection function (k) exceeds a threshold B 606, where the threshold B is given by:

[0096]

number

[0097] where α is the false alarm probability and β is the missed alarm probability. In Figure 6b, an occurrence 608 of venous needle dislodgement is detected when the detection function (k) exceeds a threshold B. As is evident from Figure 6a, this occurrence would not have been detected if the detection was based solely on the commonly set alarm limits 604, i.e., confidence intervals.

[0098] Figures 7a and 7b show an example similar to the one of Figures 6a and 6b. It is noteworthy that, again, the occurrence of venous needle disengagement 610 is detected while the venous pressure is still consistently higher than the lower of the two alarm limits 604. In Figure 7b, it is clear that there is a detection delay. The detection delay may be, for example, three sampling steps, e.g., 30 seconds.

[0099] 8a and 8b show further examples of simulations of an apparatus for vascular needle dislodgment detection according to an exemplary embodiment. In FIG. 8a, an initial pressure rise 612 is evident, which may be due to a vascular access problem, but should not be associated with venous needle dislodgment. If the venous pressure had exceeded the alarm limit 604, it would have triggered the detection of a vascular access problem. However, as is evident from FIG. 8b, the detection function λ did not exceed the threshold B, and therefore a potential venous needle dislodgment was not detected.

[0100] As is evident from Figures 9a and 9b, the filter function, for example a Kalman filter, can advantageously track relatively large changes in venous pressure without triggering false needle disengagement detection.

[0101] Further characteristics of the filter function used are illustrated by the example shown in Fig. 10. The filter function, in particular the Kalman filter, has a time-dependent dynamic behavior, initially with a wide bandwidth and (progressively) narrower bandwidth (as the process knowledge increases). This characteristic is illustrated by the example in Fig. 10, where the filter function is reinitialized every time the user applies a change in the blood pump speed 602. Even with this procedure, no false positives occur and therefore no false alarms are triggered.

[0102] FIG. 11 shows a detail of the previous example of FIG. 10 related to the first blood flow change. In particular, the high robustness of the method according to an exemplary embodiment of the invention is highlighted by FIG. 11. FIG. 11 shows a detail of the re-initialization of the filter after the blood flow change 602, where the filter function follows the measurement of the start of the blood pump change very closely and gradually recovers its standard filtering ability in the next few steps, for example the next 6-7 steps. During the re-initialization of the filter function, i.e. during the transition, the filter function may be in a fast tracking mode that leaves the detection of the withdrawal practically disabled. This is a temporary situation and may last for several steps, for example 6-7 steps (about 1 minute).

[0103] 12a and 12b show an example of a situation that results in a false vascular needle disengagement detection. A detected potential vascular needle disengagement 618 occurs when λ(k)>B. In the example shown by FIG. 12a and 12b, the detected potential vascular needle disengagement can be refuted by a static vascular pressure measurement performed by stopping the blood flow and keeping the blood line of the blood analysis loop 102 not isolated from the patient 104, i.e., by keeping the blood line clamp open. As shown in FIG. 12a, eliminating the potential vascular needle disengagement 618 allows the blood treatment to continue.

[0104] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do 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 measures cannot be used to advantage. Any reference signs in the claims should not be interpreted as limiting the scope.

[0105] Parentheses and the term "particularly" indicate optional features (this does not imply that the absence of such parentheses or expressions means that other features are mandatory in the respective context). [Explanation of symbols]

[0106] I. List of References 100 Device for detecting vascular needle dislodgement 102 Blood Analysis Loop 104 patients 106 Blood Line Clamp 108 Blood line (venous) pressure sensor 109 Blood line (arterial) pressure sensor 110 Dialysis fluid preparation device 112 Waste receiving device 114 Blood treatment and / or analysis machines 116 Anticoagulants 118 Blood Pump 202 Bloodline Pressure Dataset 204 Linear Regression Model 504 Filtered Bloodline Pressure Dataset 602 Pump Flow Set 604 Alarm Limit 606 Threshold 608 Detected vascular needle detachment B. Threshold w(k) process noise r(k) measurement noise Q Process noise variance R Measurement noise variance λ(k) test function The function used by the γ(k) test function Z(k) Recorded blood line pressure VP f (k) Filtered recorded blood line pressure T Sampling time μ(k) is a function that models the pressure change due to blood vessel detachment. σ 2 (k) Variance of the test function x1(k) Blood line pressure x2(k) is the rate of change of blood line pressure over time

Claims

1. 1. A method for vascular needle dislodgment detection, comprising: performing blood therapy and / or analysis on a patient (104) using a blood analysis loop (102) at a predetermined blood flow; - recording a blood line pressure data set (202) at predetermined time intervals using a blood line pressure sensor (108) of said blood analysis loop (102); - filtering the recorded bloodline pressure data set (202) using a processor, the processor being configured to filter the recorded bloodline pressure data set using a predefined filter function, in particular a Kalman filter; - validating the filtered bloodline pressure data set (202) using the processor, the processor being configured to validate the filtered bloodline pressure data set based on a validation function, in particular a Wald test; Detecting potential vascular needle dislodgement (618) based on a detection function λ; confirming or disproving the potential vascular needle dislodgment (618) by static blood line pressure measurements performed by stopping the blood flow and keeping the blood line of the blood analysis loop (102) not isolated from the blood vessel of the patient (104); and detecting a potential vascular needle dislodgment (618) comprises a comparison between the detection function λ and a predefined threshold B.

2. Initializing the filter function based on at least one statistical parameter, the at least one statistical parameter being blood analysis loop dependent. The method of claim 1 further comprising:

3. - initializing the filter function based on at least one statistical parameter when a treatment parameter of the blood analysis loop (102) is changed, the at least one statistical parameter being blood analysis loop dependent.

2. The method of any one of the preceding claims, further comprising:

4. Automatically re-starting the blood flow if the potential vascular needle dislodgment (618) is refuted.

2. The method of any one of the preceding claims, further comprising:

5. If the potential vascular needle dislodgment is identified, the blood analysis loop (102) remains stopped and / or an alarm is triggered.

2. The method of any one of the preceding claims, further comprising:

6. 2. A method according to any one of the preceding claims, wherein the predetermined threshold B is a function of the probability of detecting a false vascular needle dislodgment and of the probability of detecting a missed vascular needle dislodgment.

7. 10. The method of claim 1, wherein the detection function λ is a time-dependent function and depends on a time-dependent difference between the recorded bloodline pressure data set (202) and the filtered recorded bloodline pressure data set (504).

8. the blood flow is constant during each predetermined time interval; 2. The method of any one of the preceding claims, wherein the blood flow is controlled using a blood pump (118) of the blood analysis loop (102).

9. The method of any one of the preceding claims, wherein the detection of a potential vascular needle dislodgment is triggered based at least on the detection function λ resulting from the testing when the detection function λ leaves a threshold B with a predetermined false alarm probability α and missed alarm probability β.

10. An apparatus (100) for vascular needle dislodgment detection, comprising: A blood analysis loop (102); a processor configured to control the blood analysis loop (102) and detect potential vascular needle dislodgment (618); Equipped with the blood analysis loop (102) is configured to perform blood therapy and / or analysis on a patient (104); The blood analysis loop (102) includes a blood pump (118), a blood line pressure sensor (108), and a blood line clamp (106); the bloodline pressure sensor (108) is configured to record a bloodline pressure data set (202) of blood of the patient (104) and transmit the recorded bloodline pressure data set (202) to the processor; The processor is further configured to confirm or disprove the potential vascular needle disengagement (618) and operate the blood line clamp (106) and / or the blood pump (118) based on confirming or disproving potential vascular needle disengagement detection.

11. The apparatus (100) of claim 10, wherein the processor is further configured to filter the recorded bloodline pressure data set (504) using a predefined filter function, in particular a Kalman filter.

12. The apparatus (100) according to any of claims 10 to 11, wherein the processor is further configured to test the difference between the recorded bloodline pressure data set and the filtered bloodline pressure data set based on a test function, in particular a Wald test.

13. The apparatus (100) of any one of claims 10 to 12, wherein the processor is further configured to instruct the blood line pressure sensor (108) to stop the blood pump (118) of the blood analysis loop (102), to leave the blood line clamp (106) open, and to perform a static vascular pressure measurement, such as to confirm or disprove the potential vascular needle disengagement (618).

14. The apparatus (100) of any one of claims 10 to 13, wherein the processor is communicatively connected to the blood analysis loop (102).

15. The apparatus (100) of any one of claims 10 to 14, wherein the processor is further configured to trigger an alarm when the potential vascular needle dislodgment (618) is identified.

Images