Monitoring for tremor and sweating attacks

Abstract

A patient monitoring method includes receiving (103, 109) input data indicative of a patient's tremor attacks 52 and sweat attacks 54, recording (105, 111) the respective time points at which the tremor attacks and sweat attacks occur, and generating (115, 117) an output based on the alternating pattern of tremor attacks and sweat attacks from each recorded time point of tremor and / or seizures. Also disclosed are computer program products for implementing such methods and monitoring systems for implementing such methods.

Description

[Technical Field]

[0001] The present invention relates to a computer-implemented method for monitoring tremor and sweating attacks in a patient.

[0002] The invention further relates to a computer program product for implementing such a method on a monitoring system, and to a monitoring system implementing such a method. [Background technology]

[0003] A febrile patient is defined as one whose core body temperature (CBT) exceeds the normal range due to an elevated temperature set point. Fever can be caused by many conditions, ranging from minor to life-threatening. These include viral infections such as the common cold, urinary tract infections, meningitis, malaria, and appendicitis, as well as bacterial and parasitic infections, among others. Non-infectious causes include vasculitis, deep vein thrombosis, medication side effects, and cancer, among others. Fever is thought to provide a host advantage in surviving bacterial infections and is regulated by the body's thermoregulatory strategies, in which sweating and shivering, in particular, play a major role.

[0004] US Patent Application Publication No. 2018 / 0242850 discloses obtaining perfusion parameters from patient temperature monitoring, thereby determining core body temperature, and therefrom characterizing the user's symptoms, such as fever.

[0005] International Patent Publication WO 2017 / 172837 A1 discloses a system, method, and device that detects changes or the absence of changes in a patient's endogenous setpoint body temperature using heat transfer parameters or energy expenditure of a device that provides controlled hypothermia, normothermia, or hyperthermia. When the body's core temperature falls below the interthreshold zone, a series of coordinated responses occur, including, among other things, surface vasoconstriction, shivering, and metabolic thermogenesis. When the body's core temperature rises above the interthreshold zone, a series of coordinated responses occur, including, among other things, surface vasodilation and sweating.

[0006] A patient's CBT exhibits different time-dependent behavior depending on the nature of the condition. For example, fever associated with infection is caused by an increase in the set point due to cytokines released into the bloodstream by circulating white blood cells. However, when the source of infection is eliminated, i.e., activation of the host immune system decreases, cytokine concentrations in the bloodstream decrease and hypothalamic inhibition decreases. This lowers the set point, leading to sweating and a lower set point. This, in turn, results in a CBT that exhibits periodic behavior with fever episodes over time. Long-term monitoring of the CBT, also known as a fever curve, can provide medical professionals, such as clinicians and nurses, with information relevant to diagnosis.

[0007] In critical hospital settings, especially in intensive care units (ICUs), CBT is measured continuously. This is done by invasive measurement. In less critical settings, especially in general wards, invasive methods are discouraged due to the risk of infection. Because reliable, continuous, noninvasive methods for measuring CBT have not been established, CBT is measured by spot checks, usually performed only 1–3 times per day. As a result, fever curves cannot be reconstructed. Summary of the Invention [Problem to be solved by the invention]

[0008] The present invention aims to provide a method for non-invasive patient monitoring.

[0009] The present invention further aims to provide a computer program product for implementing such a method on a monitoring system, and a monitoring system for monitoring a patient, which implements such a method. [Means for solving the problem]

[0010] The invention is defined by the independent claims. The dependent claims define advantageous embodiments. A surrogate fever curve and / or other information is output which a medical professional can use to draw conclusions about the patient. This output based on tremor and sweating attacks is used in an evaluation by the medical professional to obtain diagnostically relevant information about the progression of the patient's fever. The diagnostically relevant information helps the medical professional to recognize the type of the patient's condition, for example, when defining a medication regimen for the patient or when assessing the effectiveness of an administered medication regimen, for example, by monitoring changes in the surrogate fever curve after medication administration.

[0011] According to an embodiment, there is provided a computer-implemented method for monitoring the progression of a patient's fever, the computer-implemented method comprising, using a processor device, receiving input data indicative of the patient's attacks of tremors and sweating as indicators of the patient's fever progression over a time interval, recording each time point at which the attacks of tremors and sweating occur, and generating an output (e.g., an output indicative of the patient's fever progression) from each recorded time point of the attacks of tremors and / or sweating based on an alternating pattern of attacks of tremors and sweating.

[0012] Embodiments of the present invention are based on the insight that the detection of tremor attacks and sweat attacks of a patient, such as a patient or other entity, at different times during a monitoring period (i.e., the time intervals described above) can be interpreted as an indicator of changes in the patient's CBT (i.e., the progression or change in the patient's fever state) without the need to record the patient's actual CBT. In some embodiments, a surrogate fever curve for the patient is constructed by recording the time points at which these tremor attacks and sweat attacks occur. Such a fever curve is considered a surrogate fever curve because it is not based on actual CBT measurements, but on an inference of changes in the patient's CBT from the detection of the tremor attacks and sweat attacks described above.

[0013] In one embodiment, the computer-implemented method further performs, using the processor device, the step of determining a periodicity of the detected alternating pattern of tremor attacks and sweating attacks based on their respective occurrence at different times during a time interval. Based on this, the medical professional identifies a type of fever from the determined periodicity of the alternating pattern. For example, by identifying a type of fever based on the elapsed time between subsequent tremor attacks, subsequent sweating attacks, and / or between tremor attacks and subsequent sweating attacks (or vice versa), additional diagnostically relevant information is extracted from the surrogate fever curve and presented to the medical professional, for example, to assist the medical professional in effectively treating the patient's condition.

[0014] Such times are recorded by a user manually generating input data, such as a confirmation signal, during the occurrence of a patient's tremor or sweating episode. However, in a particularly advantageous embodiment, receiving input data includes receiving patient monitoring data from a sensor device that monitors the patient's tremor and sweating as indicators of the patient's progression of fever over a time interval, processing the received patient monitoring data to detect tremor and sweating episodes, and recording each time point at which a tremor or sweating episode detected in the patient monitoring data occurs, thereby eliminating the need for manual input for patient monitoring.

[0015] The computer-implemented method may further comprise, using the processor device, determining at least one of the duration and intensity of each of the tremor and sweating attacks, and optionally calculating an indicator of fever intensity from the determined at least one of the duration and intensity of each of the tremor and sweating attacks, which provides valuable insight into the patient's ability to resist fever caused by, for example, an infection, expressed as a trend in the severity and / or duration of such attacks over time.

[0016] In a further refinement, the patient monitoring data includes information regarding the patient's degree of insulation, and the computer-implemented method further performs the step of using the processor device to determine at least one of the duration and severity of each of the tremor attack and the sweat attack as a function of the information regarding the patient's degree of insulation. For example, whether the patient is covered with one or more blankets and / or wearing a certain amount of clothing will affect the severity and duration, and these external influences can be detected to more accurately assess the true nature of the tremor attack or sweat attack.

[0017] In certain embodiments, the computer-implemented method further includes using a processor device to determine the tremor frequency of each tremor attack, and determining whether the tremor attack is an indicator of a change in the patient's body temperature due to illness based on the determined tremor frequency. This is based on the insight that the tremor frequency caused by fever is different from the frequency of other involuntary tremor-like body movements, such as shivering. This further ensures that the tremor attacks included in the proxy fever curve are truly related to the changes in the patient's CBT caused by fever.

[0018] Preferably, sweating episodes are identified based on the patient's sweat rate, as this is a particularly accurate way of determining such sweating episodes. For example, the computer-implemented method further includes using a processor device to determine the patient's sweating base rate from the received patient monitoring data during a tremor episode, and detecting a sweating episode by detecting the patient's actual sweat rate in the received patient monitoring data that exceeds the sweating base rate by a defined amount. This is particularly advantageous because it provides a baseline of sweating rate specific to the patient, based on the insight that a patient's sweat rate during a tremor episode is usually minimal. Therefore, comparing the actual sweat rate with this base rate can detect a sweating rate that is significant to the patient, making the detection method robust to different sweating rates between patients.

[0019] Alternatively or additionally, the computer-implemented method may perform the steps of using a processor device to compare the patient's actual sweat rate with a defined sweat rate threshold, and detecting a sweat bout when the actual sweat rate exceeds the defined sweat rate threshold to ensure that the detected sweat rate is sufficiently high so that the detection is associated with a decline in the patient's CBT, i.e., a genuine sweat bout indicating thermal sedation of the patient.

[0020] In a further embodiment, the computer-implemented method further performs, with the processor device, the steps of determining a signal strength of a signal associated with a tremor attack over the time interval, determining a total amount of sweat produced by the patient during a sweating attack over the time interval, determining a sweat rate of the patient over the time interval from the determined total amount of sweat produced by the patient, calculating a ratio between the determined signal strength and the determined sweat rate, and including the calculated ratio in the output. This ratio can be used to reconstruct a surrogate fever curve that allows, for example, prediction of fever severity and effectiveness of a drug response.

[0021] The patient monitoring data may further include core body temperature data relating to the patient's core body temperature measured over a time interval using a core body temperature sensor in the sensor device, and the computer-implemented method may further include, using a processor device, a step of including the patient's core body temperature data in the output. This may allow, for example, determining the lag between the onset of a tremor attack or sweat attack and the associated change in the patient's CBT, as monitored by the core body temperature sensor, from which the patient's thermoregulatory ability may be derived. Furthermore, CBT measurements may help distinguish between a tremor attack caused by a fever (which increases the patient's CBT) and a tremor attack caused by the patient being cold (which does not increase the patient's temperature set point). This may be used, for example, to verify whether a tremor attack is reliably associated with the onset of a fever attack, thereby further improving the accuracy of patient monitoring. Furthermore, including CBT information over a finite period of time, such as a period encompassing several fever cycles, may be used to calibrate or train a surrogate fever curve using the processor device. For example, CBT data may be used as ground truth for training artificial intelligence or machine learning algorithms. Such calibration or training therefore leads to the generation of a surrogate fever curve that can accurately estimate the actual CBT. Such CBT information is collected when a patient is continuously monitored in an intensive care setting, for example, during spot checks or invasive procedures.

[0022] In another embodiment, the computer-implemented method further includes the steps of: receiving, using the processor device, auxiliary information through the communication interface device at time intervals, the auxiliary information including at least one of a measurement of an ambient temperature to which the patient is exposed and an activity level of the patient; determining whether a tremor or sweating attack is a reliable indicator of a change in the patient's body temperature due to illness based on the actual auxiliary information at the time of the tremor or sweating attack; and including the tremor or sweating attack in the step of generating an output only if the tremor or sweating attack is determined to be a reliable indicator of the patient's fever progression. This can eliminate tremors or sweating attacks caused by factors other than fever from the proxy fever curve, improving the accuracy of the proxy fever curve.

[0023] According to another aspect, there is provided a computer program product including a computer-readable storage medium having computer-readable program instructions embodied therein that, when executed on a processor unit of a patient monitoring system, cause the processor unit to perform a method according to any of the embodiments described herein. Such a computer program product can be used to configure the monitoring system to perform a method according to an embodiment of the invention.

[0024] According to yet another aspect, a monitoring system for monitoring a patient is provided. The monitoring system includes a processor device that receives input data indicative of the patient's tremor and sweating attacks as indicators of the patient's fever progression over a time interval, and the monitoring system further includes a computer program product according to any of the embodiments described herein, wherein the processor device executes computer-readable program instructions. Thus, such a monitoring system can generate a proxy fever curve for the patient from the patient's tremor and sweating attacks captured in the patient monitoring data provided by the sensor device without collecting CBT information. The monitoring system provides an output (e.g., an output including information related to the proxy fever curve and / or diagnosis derived from the tremor and sweating attacks) to a user interface based on the proxy fever curve, such that the output can be evaluated at the user interface by, for example, a medical professional, or the output can be stored in a data storage architecture, such as an electronic patient record, for timely retrieval by a medical professional.

[0025] The monitoring system further includes a communication interface device that receives input data in the form of patient monitoring data from a sensor device that monitors tremor and sweating as an indication of the patient's fever progression and transfers the patient monitoring data to the processor device, and the monitoring system optionally includes the above-mentioned sensor device, whereby the patient is monitored without the need for manual intervention, thus providing a complete solution for monitoring the patient's fever progression based on monitoring the patient's tremor and sweating attacks, as described above. The sensor device includes at least one of a camera, a sensor pad integrated into a bed, a wearable sensor including a sweat sensor, and a motion sensor that detects tremor, and preferably the sensor device includes a wearable sensor. [Brief explanation of the drawings]

[0026] Embodiments of the invention will now be described in more detail and by way of non-limiting example with reference to the accompanying drawings, in which:

[0027] [Figure 1] FIG. 1 shows a schematic diagram of a monitoring system according to an embodiment. [Figure 2] FIG. 2 shows a schematic example of a fever curve (y-axis is core body temperature, x-axis is time). [Figure 3] FIG. 3 illustrates, in schematic form, certain aspects of the monitoring system of FIG. 1 in more detail. [Figure 4] FIG. 4 shows a flowchart of a fever progression monitoring method according to an embodiment. [Figure 5] FIG. 5 shows a graph illustrating the construction of a surrogate heating curve according to an example embodiment. [Figure 6] FIG. 6 shows a graph illustrating the construction of a surrogate heating curve according to another example embodiment. [Figure 7] FIG. 7 shows a graph illustrating the construction of a surrogate heating curve according to yet another example embodiment. [Figure 8] FIG. 8 shows a flowchart of an aspect of a fever progression monitoring method according to an embodiment. [Figure 9] FIG. 9 is a graph showing several fever index curves (y-axis is reconstructed core body temperature index, x-axis is time). DETAILED DESCRIPTION OF THE INVENTION

[0028] It should be understood that the figures are schematic only and are not drawn to scale, and that the same reference numerals have been used throughout the figures to indicate the same or similar parts.

[0029] FIG. 1 schematically illustrates an arrangement for monitoring the progression of a fever in a patient 1 using a monitoring system 10. The monitoring system 10 typically includes a data processing architecture 30 (e.g., a computer system, a server system, etc.) that includes a communication interface device 32 in communication with a processor device 34. The communication interface device 32 may include any suitable number of data communication interfaces, such as one or more P2P communication interfaces such as a Bluetooth® interface, one or more wireless communication interfaces such as a Wi-Fi® interface, or one or more wired communication interfaces such as an Ethernet® interface, for connecting the data processing architecture 30 to a local area network, the Internet, or the like. The processor device 34 may include one or more processing elements, such as one or more processors, processor cores, etc., that process patient monitoring data provided by the sensor device 20.

[0030] The sensor device 20 (which may form part of the monitoring system 10) typically provides patient monitoring data such that the processor device 34 can detect tremor attacks 52 and sweat attacks 54 in the patient 1 being monitored, to enable the processor device 34 to construct a surrogate fever curve 50 for the patient 1, as shown schematically in FIG. 2, as will be described in more detail below. The patient monitoring data is typically provided to the data processing architecture 30 via one or more data communication links between the sensor device 20 and the communication interface device 32. The communication interface 32 forwards the received patient monitoring data to the processor device 34. In order to provide patient monitoring data from which the processor device 34 can derive the tremor and sweat attacks in the patient 1, the sensor device 20 includes one or more sensors for this purpose.

[0031] For example, sensor device 20 includes a camera 21 pointed at patient 1. Camera 21 detects the patient's tremor, for example, by detecting patient movement in images provided by camera 21, and also detects sweating, for example, by changes in light reflection caused by the formation of sweat droplets on exposed portions of the patient's skin, such as the patient's forehead, detected in images provided by camera 21. Sensor device 20 includes a wearable sensor 22 attached to the patient's skin that provides the necessary patient monitoring data. An example embodiment of such a wearable sensor 22 is shown in FIG. 3. FIG. 3 illustrates wearable sensor 22 including a motion sensor 221, such as an EMG or ECG sensor or an accelerometer, capable of detecting patient movement (e.g., tremor), a sweat sensor 222 capable of detecting sweating in the area of the patient's skin to which wearable sensor 22 is attached, and optionally at least one additional sensor 223 that provides additional sensor data, as described in more detail below.

[0032] The sweat sensor 222 measures sweat flow rate or sweat volume after a defined collection time, or indirectly measures sweat parameters, for example, by measuring skin conductance. When including both a motion sensor 221 and a sweat sensor 222, the wearable sensor 22 should be placed on a part of the patient's body where both tremor and sweating can be reliably measured, such as the patient's chest or neck. The sensor device 20 includes a sensor 23 that is integrated into bedding, for example, a mattress 3 on which the patient 1 rests, to detect patient movement while the patient 1 is in bed. From this patient movement, tremor can be detected. Any suitable combination of such sensors 21, 22, and 23 can be deployed. In a preferred embodiment, the sensor device 20 includes at least the wearable sensor 22.

[0033] The data processing architecture 30 may further be communicatively coupled to one or more user interfaces 41, 42. The user interfaces display the results of the processing of the patient monitoring data by the processor device 34 (e.g., a surrogate fever curve 50). To this end, the processor device 34 typically generates an output based on the processing results, e.g., including the processing results and information derived therefrom. This output is transferred to the particular user interface 41, 42 via the communication interface device 32. For example, the user interface 41 may be a mobile communication device including an app for visualizing the processing results of the processor device 34, which a medical professional, such as a nurse or doctor, can use to evaluate the fever of the patient 1. The user interface 42 may be a user terminal, such as a personal computer, through which the medical professional can receive the processing results from the processor device 34. Alternatively or additionally, the processor device may store the processing results in a data storage arrangement 43, e.g., a network-connected database that stores the patient monitoring data, e.g., an electronic patient record, so that the processing results can be accessed at any appropriate time, e.g., using the user interface 41 or 42. Other suitable architectures will be apparent to those skilled in the art.

[0034] In another embodiment (not shown), the sensor device 20 is omitted and the processor device 34 can instead receive input data manually generated, for example, by the patient 1 or a medical professional such as a nurse or doctor, using an input device such as a remote control, an electronic communication device such as a smartphone, etc., so that the person operating such input device can signal the occurrence of a tremor attack 52 or a sweat attack 54 in the patient 1 on the input device, and the processor device need only record the time points at which such input is received in order to construct the surrogate fever curve 50.

[0035] The monitoring system 10 implements one or more embodiments of a method 100 of the present invention. A flowchart of the method 100 is shown in FIG. 4. The method 100 begins at step 101. In step 101, for example, the sensor device 20, if present, is enabled and provides the aforementioned patient monitoring data to the processor device 34 via the communication interface device 32. In step 103, the processor device 34 processes the patient monitoring data received from the sensor device 20 to determine whether a tremor attack 52 of the patient 1 can be detected in the patient monitoring data. This processing can be performed continuously or periodically. That is, the patient monitoring data can be processed continuously or periodically sampled, such as once every 1 to 5 minutes, for processing by the processor device 34. If a tremor is not detected, the processor device 34 continues to search for a tremor attack 52 of the patient 1 in the received patient monitoring data. However, if the processor device 34 detects the occurrence of a tremor attack 52, the method proceeds to step 105. In step 105, the processor device records the time when a tremor attack 52 is detected, for example, as derived from patient monitoring data provided by the sensor device 20 or from manually generated input data indicative of the tremor attack 52. This time may be expressed in any suitable format, for example, as elapsed time since patient monitoring began, actual time of day, etc.

[0036] The method 100 proceeds to step 107, in which the processor device 34 determines the patient 1's sweating baseline rate during a tremor attack 52. This is based on the insight that during such a tremor attack 52, the patient 1's CBT, i.e., temperature set point, is elevated, which typically corresponds to minimal associated sweating. Therefore, determining the sweating rate during such a tremor attack 52 provides a reliable baseline for determining the patient 1's subsequent sweating attacks 54, as the patient's actual sweating attacks 54 will result in sweating rates that are significantly higher than the baseline sweating rate. Determining such a baseline sweating rate has the advantage that a patient-specific sweating baseline is defined, thereby eliminating the need to consider different sweating rates for each patient in order to reliably detect the patient's sweating attacks. However, in alternative embodiments, the detection of such a sweating baseline rate may be omitted. In this case, the patient 1's sweating attacks 54 can be detected using a defined sweating rate threshold. The sweating rate threshold may be general, gender-specific, specific to the location of the wearable sensor 22 on the patient's body, or the like.

[0037] In step 109, the processor device 34 processes the patient monitoring data received from the sensor device 20 to determine whether a sweating episode 54 of the patient 1 can be detected in the patient monitoring data. This processing may be performed continuously or periodically. That is, the patient monitoring data may be processed continuously or may be sampled periodically, such as once every 1-5 minutes, for processing by the processor device 34. For example, the processor device 34 may determine an actual sweating rate of the patient 1, compare this actual sweating rate with the sweating base rate determined in step 107, and determine that a sweating episode 54 is occurring if the actual sweating rate exceeds the sweating base rate by a defined amount (e.g., a defined multiple such as 4-6 times).

[0038] Alternatively or additionally, the processor device 34 may compare the actual sweat rate with a defined sweat rate threshold and determine that a sweat bout 54 is occurring if the actual sweat rate exceeds the defined threshold, thereby confirming that the sweat rate is sufficiently high (e.g., greater than 0.7 nL / min for sweat glands) to determine with sufficient confidence that a sweat bout 54 is occurring. Such a sweat rate may be determined using the wearable sensor 22, for example, by determining the actual sweat rate or a parameter related to changes in sweat rate, such as skin conductance, or using the camera 21, for example, by determining changes in reflected light from camera images and deriving the sweat rate from this change.

[0039] If a sweating attack 54 is not detected, the processor device 34 continues to search for a sweating attack 54 in the patient 1 in the received patient monitoring data at step 109. However, if the processor device 34 detects the occurrence of a sweating attack, for example, in the patient monitoring data provided by the sensor device 20 or from manually generated input data flagging a sweating attack 54, the method proceeds to step 111 where the processor device 34 records the time when the sweating attack 54 was detected. This time may be expressed in any suitable format, such as the time elapsed since monitoring of the patient 1 began, the time elapsed since a previously detected tremor attack 52, the actual time of day, etc.

[0040] As shown in step 113, the recording of the times at which tremor attacks 52 and sweat attacks 54 are detected continues, in which case method 100 returns to step 103 and processor device 34 processes the patient monitoring data to detect the next tremor attack 52 or awaits manual flagging of such an attack as described above. Method 100 may alternatively proceed to step 115, in which processor device 34 constructs a surrogate fever curve 50 for patient 1 from the recorded times at which patient 1's tremor attacks 52 and sweat attacks 54 occurred. Note that, to avoid concerns, construction of the surrogate fever curve 50 is performed after monitoring patient 1 for a fixed time interval or in parallel with monitoring patient 1 over that time interval. Processor device 34 constructs the surrogate fever curve 50 based on the alternating pattern of tremor attacks 52 and sweat attacks 54 and the periodicity of this alternating pattern.

[0041] For example, as shown generally in FIG. 5 , processor device 34 receives an indication, e.g., from sensor device 20 or by manual input as described above, of the onset of a tremor attack 52 at T=t1. The tremor attack 52 ends at T=t2. The top graph shows the intensity (arbitrary units) (Y-axis) of the tremor attack 52 and the sweat attack 54 as a function of time (X-axis). Processor device 34 further receives an indication, e.g., from sensor device 20 or by manual input as described above, of the onset of a sweat attack 54 at T=t3. The sweat attack 54 ends at T=t4. Processor device 34 further receives an indication, e.g., from sensor device 20 or by manual input as described above, of the onset of a further tremor attack 52 at T=t5. The further tremor attack 52 ends at T=t4. The processor device 34 constructs a surrogate heating curve 50 from the time points at which the tremor attacks 52 and sweat attacks 54 began (i.e., t1, t3, and t3), without taking into account the duration of each attack. This generates a resulting surrogate heating curve 50 having CBT, or more precisely, a projected change in CBT (y-axis), as a function of time (x-axis), as shown in FIG. 5 . As can be seen from this curve, each tremor attack 52 episode is interpreted as a momentary increase in CBT, and each sweat attack episode is interpreted as a momentary decrease in CBT. Of course, alternative approaches are possible, such as fitting a sinusoidal curve based on the recorded time points of the tremor attack 52 and sweat attack 54 episodes. The processor device 34 then generates an output based on the constructed surrogate heating curve 50 in step 117, for example, for transmission to the user interface 41 or 42 and / or storage in the data storage architecture 43. The method 100 then ends in step 119.

[0042] In an improved version, when constructing the surrogate fever curve 50 in step 115, the processor device 34 takes into account the duration of each tremor attack 52 and sweat attack 54. This is shown in FIG. 6, where at the start of each tremor attack 52, the processor device 34 assumes a gradual increase, such as a linear increase, in the patient's CBT until the end of the tremor attack 52. The CBT is then assumed to be constant until the start of the subsequent sweat attack 54. At that point, the processor device 34 assumes a gradual decrease, such as a linear increase, in the patient's CBT until the end of the sweat attack 54 in its construction of the surrogate fever curve 50. Again, as an alternative, a sinusoidal curve could be fitted based on the start and end values (in time) of the tremor attack 52 and the sweat attack 54, respectively.

[0043] In a further refinement, as shown in FIG. 7 , the processor device 34 constructs a surrogate heating curve 50 by taking into account both the duration and intensity of tremor attacks 52 and sweat attacks 54. In this approach, if the patient 1 is sweating heavily, i.e., has a high sweat rate, the recursive approximation of the patient's CBT decreases more quickly than if the patient 1 is only sweating lightly, i.e., has a low sweat rate. As described above, a sweat attack 52 is detected when the actual sweat rate exceeds a sweat reference rate or threshold 55. Similarly, if the patient 1 has a severe tremor, the recursive approximation of the patient's CBT increases more quickly than if the patient 1 has only a slight tremor. The intensity of the tremor attack is determined using power spectral density analysis, as described in more detail below. Because the intensity level of the tremor attack 52 or sweat attack 54 changes over time, the binary on / off approach to constructing the surrogate heating curve 50 of FIG. 5 becomes integral, as represented by Equation (1).

number

[0044] In this equation, t is time, T0 is the body temperature at t=0 (which may be measured, estimated, or a default value), and c1 and c2 are parameters that depend on the patient and the environment (e.g., taking into account the degree of insulation of patient 1, see below), and may also depend on the intensity of tremor and sweating. If an independent CBT measurement is not available, default values are used for T0, c1, and c2. However, if such an independent CBT measurement becomes available, this measurement can be used to determine T0. If multiple independent CBT measurements are available over a period of time (e.g., through continuous or spot-check CBT measurements over a period of time during which oscillation seizures 52 and sweating seizures 54 occur), c1 and c2 can also be estimated for a particular patient, environment, and tremor / sweat seizure intensity level during the CBT measurement. This serves as a calibration for the construction of the surrogate fever curve 50 by the processor device 34 after the CBT measurement has stopped. Thus, in this embodiment, a surrogate fever curve 50 can be provided that includes an estimated absolute value of the patient's CBT on the Y-axis of a graph that shows the surrogate fever curve 50 as a function of time on the X-axis of the graph.

[0045] In a preferred embodiment, the processor device 34 generates information to assist a medical professional in treating the patient, such as determining the type of fever from the recorded time points at which the patient's 1 detected tremor attacks 52 and sweat attacks 54 occurred, in addition to or instead of generating the surrogate fever curve 50, and includes this information in the output generated in step 117, for example, in addition to the surrogate fever curve 50 or as information derived from the surrogate fever curve 50 without including the actual surrogate fever curve 50, to provide the medical professional with additional insight into the patient's underlying condition. This is described in more detail with reference to FIG. 8, which schematically illustrates a decision tree deployed by the processor device 34 to determine the type of fever from the recorded time points. The decision tree begins at step 151, after which the processor device 34 determines the periodicity of the surrogate fever curve 50 in step 153, for example, by determining the elapsed time between successive tremor attacks 52 separated by sweat attacks 54, or the elapsed time between successive sweat attacks 54 separated by tremor attacks 52 from the recorded time points. If the periodicity is less than about 24 hours, as determined in step 155, processor unit 34 determines that the type of fever is attenuation fever, as represented in conclusion 171.

[0046] On the other hand, if the periodicity is determined to be approximately 24 hours, the processor device 34 determines that the type of fever is a daily fever. Next, in step 157, the processor device 34 further evaluates the elapsed time between the tremor attack 52 and the subsequent sweat attack 54 to determine the patient's body's ability to fight the daily fever. Typically, if the patient's immune system is functioning normally, the duration of the elevated CBT or fever is within a few hours, during which the patient's immune system attacks the parasite outbreak. Therefore, if the processor device 34 concludes in step 157 that the duration of the elevated CBT period corresponds to proper functioning of the immune system, the processor device 34 concludes that the fever is a "normal" daily fever, as represented by conclusion 172. On the other hand, if the processor device 34 concludes in step 157 that the duration of the elevated CBT period exceeds a threshold indicating a weakened immune system, the processor device 34 concludes that the fever is a daily fever in weakened patient 1, as represented by conclusion 173.

[0047] Of course, assessment of the state of the patient's immune system is optional, in which case determining the elapsed time between a tremor attack 52 and a subsequent sweat attack 54 is omitted from this assessment. Accordingly, recording of the time of occurrence of a sweat attack 54 is omitted, and assessment of the surrogate fever curve 50 is based solely on the elapsed time between successive tremor attacks 52 (as long as these tremor attacks are separated by a sweat attack 54). Similarly, recording of the time of occurrence of a tremor attack 52 is omitted, and assessment of the surrogate fever curve 50 is based solely on the elapsed time between successive sweat attacks 54 (as long as these sweat attacks are separated by a tremor attack 52).

[0048] If, in step 155, the processor device 34 determines that the periodicity of the patient 1's fever is significantly longer than 24 hours, the processor device 34 concludes that the fever is another type of fever, such as tertian fever, quartan fever, relapsing fever, undulant fever, etc., as represented in conclusion 174. To distinguish between these additional types of fever, a further decision tree (not shown) may be implemented in the processor device 34.

[0049] As described above, in at least some embodiments using sensor device 20 to collect patient monitoring data from which surrogate fever curve 50 is constructed, additional sensors (e.g., additional sensor 223) are used for a number of reasons. For example, sweat sensor 222 of wearable sensor 22 further determines the concentration of analytes of interest (e.g., cytokines or lactate) in secreted sweat. This is used by processor device 34 to further determine the cause of the fever, the severity of the disease, and / or the effectiveness of medication to combat the disease. This information provides useful insight to medical personnel on how to treat patient 1, e.g., by administering a different medication dosage, a different drug dose, or a different medication. Further additional sensors, preferably, but not necessarily, integrated into one or more wearable sensors 22, provide physiological monitoring data such as heart rate, respiratory rate, blood pressure, peripheral oxygen saturation, etc. Including this sensor data in the output generated by processor device 34 further enhances the diagnostic potential of the output.

[0050] The sensor device 20 may further include one or more sensors that provide auxiliary information to the processor device 34, such as one or more motion sensors, a heart rate monitor, or a blood pressure sensor that detects an increase in the patient's activity (e.g., sports activity, strenuous activity such as climbing stairs, etc.), which may trigger a sweating attack rather than a decrease in the patient's body temperature set point due to fever. This auxiliary information allows the processor device 34 to distinguish between fever-related tremor attacks 52 and sweating attacks 54, or attacks that occur for different reasons. An environmental sensor (e.g., an ambient temperature sensor) can be used to determine whether the onset of a tremor or sweating attack is caused by environmental factors, such as low or high ambient temperatures, rather than by illness. The processor device 34 uses this auxiliary information to determine whether the tremor attacks 52 or sweating attacks 54 are reliable indicators of a change in the patient's body temperature due to illness. As a result, if the processor device 34 determines that the tremor attack 52 or sweat attack 54 is a reliable indicator of a change in the patient's body temperature due to illness, rather than one that may be caused by other factors as indicated by the auxiliary information, then the processor device 34 will only include the time points of such tremor attack 52 or sweat attack 54 in the proxy fever curve 50.

[0051] In yet another embodiment, the sensor device 20 further includes a CBT sensor, which may be an additional sensor 223 within the wearable sensor 22 or a separate (wearable) sensor. Including CBT information in the patient monitoring data can provide more information from the time of occurrence of a tremor attack 52 and / or a sweat attack 54. Typically, CBT increases with tremor and decreases with sweating, but there is a lag time between the occurrence of such a tremor attack or sweat attack and the associated change in CBT. This lag time can be used to gain insight into the patient's ability to regulate their body temperature. Furthermore, the CBT information can be used to calibrate the construction of the surrogate fever curve 50 by the processor device 34. For example, it can be used as ground truth for training an artificial intelligence or machine learning algorithm used by the processor device 34 to construct the surrogate fever curve 50.

[0052] Another reason for adding a CBT measurement is to distinguish between tremors due to an elevated set point (resulting in a fever) and tremors due to the patient getting too cold (who has a normal set point). This is particularly useful when establishing the onset of a fever period. One simple way to distinguish between these two possibilities is to perform a single CBT measurement when tremors are detected for an appropriate period (e.g., 10 minutes). If an elevated set point is the cause, the temperature will already be rising, but if the body is cold, it will fall.

[0053] As is well known, fever is an increase in body temperature during infection. However, the underlying mechanism of fever is more complex and involves the host response to infection. The immune system is activated by the infectious agent, releasing chemokines into the body, leading to the activation of cyclooxygenase (COX), which results in the stimulation of "heat gain thermoeffectors" (i.e., shivering) and the inhibition of "heat loss thermoeffectors" (i.e., sweating). Typically, only core body temperature is measured, which, apart from the aforementioned impractical clinical integration, only provides information regarding the elevation of the patient's temperature set point. Therefore, in a preferred embodiment, the construction of the surrogate fever curve 50 further includes determining the patient's shivering rate and sweating rate to provide an additional layer of diagnostically relevant information to the surrogate fever curve 50. This may, for example, provide information regarding the host response, i.e., the immune response, to the infectious agent. In addition to or as an alternative to measuring patient 1's CBT to determine the lag between tremor and sweating attacks and induced changes in the patient's CBT, as described above, such a host response can be calculated from several parameters, such as the change in time between successive tremor attacks 52, the change in time between successive sweating attacks 54, the change in time between a tremor attack 52 and a subsequent sweating attack 54, and the intensity and duration of such tremor attacks 52 and sweating attacks 54. The intensity of the sweating attacks 54 is based on the sweat rate determined by the sensor device 20. The intensity of the tremor attacks 52 is determined based on Fourier analysis and / or spectral power analysis of the tremor data provided by the sensor device 20. This serves several purposes. First, analysis of such tremor data provides a measure of the frequency of patient 1's tremor. This allows the processor device to distinguish between unintentional muscle movements, such as age-related tremors, and tremors due to an elevated set point compared to patient 1's actual CBT. For example, hand tremors caused by Parkinson's disease can be as high as 9 Hz, with tremor amplitudes and acceleration amplitudes of 19 cm and 20 m / s , respectively. 2During fever or hyperthyroidism, the frequency may increase up to 14 Hz, thereby distinguishing between fever-related shivering and other types of shivering, and the processor device 34 may discard tremor attacks 52 caused by such other types of shivering and consider only fever-induced tremor attacks 52 in constructing the surrogate fever curve 50, for example, by using a band-pass or high-pass filter with an appropriate cutoff frequency (e.g., 10 Hz).

[0054] The ratio of the signal strength of the tremor signal generated by the sensor device 20 to the total sweat volume or sweat rate over time monitored over the patient monitoring time interval can then be used to construct a surrogate fever curve to predict fever severity and treatment response effectiveness.

[0055] In the first step, a fast Fourier analysis (e.g., amplitude or power spectral signal analysis) of the oscillating signal is performed to determine the tremor strength. The power spectral density (PSD) of the tremor signal (PSD_Shivering) represents the power present in the signal as a function of frequency per unit frequency. Power spectral density is usually expressed in watts per hertz (W / Hz). Next, the total power present in the signal over the relevant frequency range (e.g., 10-20 Hz or 5-15 Hz) is calculated.

[0056] In the second step, the ratio of tremor signal strength to sweat rate (nL / s) is determined, and a fever index is calculated according to the following Equations 2 and 3. Such a fever index can be either an index at a given time (Equation 2) or a time-integrated index (Equation 3), i.e., the area under the fever index curve shown in FIG. 9. When the fever index is integrated over time to calculate a cumulative fever index (Equation 2), this time-integrated index provides clinical information related to the intensity of fever attacks or fever intervals in a given time frame of clinical interest. Another option is to calculate the first derivative of the fever index curve, which can provide information about fever dynamics, such as the rate of fever onset or fever recovery.

number

number

[0057] This ratio is calculated because the onset and development of fever is usually proportional to and directly related to tremor, while fever can be assumed to be inversely proportional to sweat rate, decreasing with sweating. This is shown diagrammatically in Figure 9. Three fever index curves are shown, progressing from a CBT with a normal or baseline value of 5 below the upper threshold of normal 7, to a strong fever value of 6 above the critical threshold 8 indicated by a tremor attack 52, and then returning to the normal or baseline value of 5 (indicated by a sweat attack 54).

[0058] From the proxy heat index curves A to C, the time constant τ up or τ down For example, τ up is defined as the time it takes for the surrogate heat curve 50 to pass through the critical threshold 8 starting from the reference value 5, and τ down is defined as the time it takes for the proxy heating curve 50 to pass from the heating value 6 to the upper reference threshold 7. Alternatively, τ down , may be defined as the time it takes for fever to decay to 67% of its maximum value, similar to calculating the decay time of an electrical component such as a capacitor. These time constants therefore carry clinical information regarding the rate at which fever develops and dissipates, such as after drug administration. The surrogate fever curve shown in FIG. 9 can be visualized on user interface 41 or user interface 42, for example, to provide direct feedback to healthcare professionals about treatment response, for example, to enable clinical decision support for the effectiveness of an intervention or drug.

[0059] In one embodiment, the determinations of patient 1's tremor intensity and sweat rate are scaled by processor device 34 based on available information regarding patient 1's insulation. For example, during a tremor attack 52, patient 1 can keep warm by wearing more clothing or adding blankets, which can reduce the intensity of the tremor attack 52. Similarly, during a sweat attack 54, patient 1 can cool down by removing clothing or blankets, which can reduce the sweat rate. To this end, sensor device 20 provides information regarding the patient's degree of insulation, enabling processor device 20 to determine at least one of the duration and severity of each of tremor attack 52 and sweat attack 54 as a function of the information regarding the patient's degree of insulation, as described above. Such information may be provided in any suitable manner. For example, the presence of any blankets or clothing and / or the thickness of the blankets or clothing may be determined from image data provided by camera 21. Alternatively, such information may be generated using wearable sensor 22, for example, by providing measurements of one or more of skin temperature, pressure, and ambient light. In particular, the pressure of the blanket against the skin and ambient light, i.e., light passing through blankets and clothing, are independent of core body temperature and are particularly useful parameters to monitor for the above purposes.

[0060] Embodiments of the present invention provide clinical insight into the thermoregulatory capabilities of a patient's body, for example, to provide insight into the effectiveness of administered medications, the severity of a disease, or to aid in the diagnosis of fever causes. Such clinical insight is useful in various care settings, such as in a hospital, in a hospital-to-home setting, or in a home setting, to help identify appropriate medications, dosages, and timing. For example, clinical information provided by the processor device 34 can help determine whether antibiotics are necessary to treat a patient's condition, thereby reducing the risk of unnecessary administration of such antibiotics (which can undesirably promote antibiotic resistance). However, it should be understood that the use of the present invention is not limited to medical applications and may be utilized to gain insight into the thermoregulatory characteristics of the body of any patient (e.g., human or non-human) in general. In this context, the term "patient" as used herein is intended to apply to all patients whose CBT is monitored in accordance with the teachings of the present application and should not be construed as being limited to those suffering from a disease.

[0061] The above-described embodiments of method 100 executed by processor unit 34 are realized by computer-readable program instructions embodied in a computer-readable storage medium. When executed by processor unit 34 of a computing device 30, such as a patient monitoring terminal, the computer-readable program instructions cause processor unit 34 to perform any embodiment of method 100. Any suitable computer-readable storage medium may be used for this purpose, such as an optically readable medium such as a CD, DVD, or Blu-Ray® disc, a magnetically readable medium such as a hard disk, or an electronic data storage device such as a memory stick. The computer-readable storage medium may be a medium accessible via a network, such as the Internet, such that the computer-readable program instructions are accessible via the network. For example, the computer-readable storage medium may be a network-attached storage device, a storage area network, cloud storage, or the like. The computer-readable storage medium may also be an Internet-accessible service from which the computer-readable program instructions can be obtained. In some embodiments, at least a portion of the computer-readable program instructions are incorporated into processor unit 34 in the form of hardware.

[0062] It should be noted that the above embodiments are illustrative rather than limiting of the present invention, and that those skilled in the art can design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the scope of the claim. The word "comprises" does not exclude the presence of elements or steps other than those listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by hardware comprising several different elements. In a device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. Measures recited in mutually different dependent claims may be advantageously combined.

Claims

1. 1. A patient monitoring system for monitoring the progression of fever in a subject, comprising: receiving patient monitoring data from a sensor device that monitors tremors and sweating of the patient as indicators of the progression of the patient's fever over a time interval; processing the received patient monitoring data to detect tremor attacks and sweating attacks and record the time points at which the tremor attacks and the sweating attacks occurred, respectively; 1. A patient monitoring system including a processor device, the processor device further comprising: and constructing a surrogate fever curve from the recorded time points of the tremor attacks and the sweating attacks based on an alternating pattern of the tremor attacks and the sweating attacks for use in monitoring the progression of the fever of the subject.

2. The processor device determining a periodicity of the alternating pattern of the detected tremor attacks and sweating attacks based on their respective occurrences at different times during the time interval; The patient monitoring system of claim 1 , wherein the determined periodicity of the alternating pattern identifies a type of fever.

3. The processor device 3. The patient monitoring system of claim 1, wherein at least one of the duration and intensity of each of the tremor attacks and sweat attacks is determined.

4. the patient monitoring data includes information regarding the degree of insulation of the patient; The processor device 4. The patient monitoring system of claim 3, wherein at least one of the duration and severity of each of the tremor attack and the sweat attack is determined as a function of the information regarding the degree of insulation of the patient.

5. The processor device 5. The patient monitoring system of claim 1, wherein the tremor frequency of each tremor episode is determined.

6. The processor device 6. The patient monitoring system of claim 1, wherein a sweating episode is identified based on the patient's sweat rate.

7. The processor device determining a sweating base rate for the patient from the patient monitoring data received during a tremor attack; 7. The patient monitoring system of claim 6, wherein a sweat episode is detected by detecting, within the received patient monitoring data, an actual sweat rate of the patient that exceeds the sweat baseline rate by a defined amount.

8. The processor device comparing the patient's actual sweat rate with a defined sweat rate threshold; 8. The patient monitoring system of claim 6 or 7, wherein a sweating episode is detected when the actual sweat rate exceeds the defined sweat rate threshold.

9. The processor device determining a signal strength of a signal associated with a tremor attack over the time interval; determining a total amount of sweat produced by the patient during the sweating bout over the time interval; determining a sweat rate of the patient over the time interval from the determined total amount of sweat produced by the patient; calculating a ratio between the determined signal strength and the determined sweat rate; 9. A patient monitoring system according to claim 6, which outputs the calculated ratio.

10. 10. The patient monitoring system of claim 1, wherein the patient monitoring data further includes core body temperature data relating to the patient's core body temperature measured over the time interval using a core body temperature sensor of the sensor device, and the processor device outputs the core body temperature data.

11. The processor device receiving, by a communication interface device, auxiliary information during the time interval, the auxiliary information including at least one of a measurement of an ambient temperature to which the patient is exposed and an activity level of the patient; determining whether a tremor attack or a sweat attack is a reliable indicator of a change in the patient's body temperature due to illness based on actual auxiliary information at the time of the tremor attack or the sweat attack; 11. The patient monitoring system of claim 1, wherein the patient monitoring system outputs a tremor or sweating attack only if the tremor or sweating attack is determined to be a reliable indicator of the patient's developing fever.

12. 12. The patient monitoring system of claim 11, further comprising a communications interface device that receives input data in the form of patient monitoring data from a sensor device that monitors the patient's tremor and sweating as indicators of a progression of the patient's fever, and that transfers the patient monitoring data to the processor device, the patient monitoring system optionally including the sensor device.

13. 13. A computer readable storage medium having computer readable program instructions embodied thereon, the computer readable program instructions, when executed by a processor unit of a patient monitoring system according to any one of claims 1 to 12, causing the processor unit to: receiving the patient monitoring data from the sensor device monitoring the patient's tremors and sweating as indicators of the patient's fever progression over the time interval; processing the received patient monitoring data to detect tremor attacks and sweating attacks and record the time points at which the tremor attacks and sweating attacks occurred, respectively; constructing a surrogate fever curve from the recorded time points of the tremor attacks and the sweating attacks based on an alternating pattern of the tremor attacks and the sweating attacks for use in monitoring the progression of the fever of the subject; 12. A computer-readable storage medium for implementing a patient monitoring method, comprising:

Images