Methods and systems for infection detection

Abstract

A method for determining whether a user is infected includes: receiving, from a wearable device, time series data representing a physiological variable measured by the wearable device; determining a first parametric function characterizing a distribution of a first portion of the time series data, at least one parameter of the first set of parameters characterizing a skew of the first parametric function; determining a second parametric function characterizing a distribution of a second portion of the time series data, at least one parameter of the second set of parameters characterizing a skew of the second parametric function, where the first portion of the time series data represents a baseline window and the second portion represents an observation window; and determining whether the user is currently infected according to at least one measure of a difference between the first set of parameters and the second set of parameters.

Description

Technical Field

[0001] The present disclosure generally relates to methods and systems for early detection of infection. Background Art

[0002] Early detection of infectious diseases such as COVID-19, RSV, and influenza is crucial for curbing their spread. Typically, carriers of these diseases are contagious long before they have symptoms, which means that the disease spreads before the carrier realizes the infection and is able to self-isolate. However, early detection of infection is difficult because symptoms are often short-lived and their duration varies from individual to individual, making it difficult to model in typical machine learning classification algorithms. In some cases, the severity of the symptoms may also be subtle, which may result in low classification accuracy. In addition, determining an individual's health—i.e., non-infected—status may have many confounding factors (such as a user failing to consistently wear a particular wearable or a user being infected continuously), especially when data is collected over a short period of time. Summary of the Invention

[0003] Therefore, there remains a need for early detection of infectious diseases, which can explain the changes, short-lived and subtle symptoms during the early stages of infection. The various embodiments and implementations described in the present disclosure relate to a method or system for early detection of infectious diseases using data about one or more physiological variables of a user collected over time, for example, via a wearable device. The system receives the physiological variable data and organizes it into distributions collected over different time windows, baseline windows, and observation windows. The shapes of the distributions of the two windows, in particular the properties of the "tails" of the distributions, are compared to determine whether the user is currently carrying an infection. The distributions can be compared by characterizing the maximum likelihood fit of each distribution using a parametric function, such as a skewed normal function. The differences between the parameters of the parametric functions can then be input into a predictive model to detect the early presence of an infection.

[0004] According to one aspect, a method for determining whether a user is infected based on data collected by a wearable device worn by the user includes the following steps: receiving time series data representing physiological variables measured by the wearable device from the wearable device; determining a first parameter function defined by a first set of parameters, the first parameter function characterizing the distribution of a first portion of the time series data, and at least one parameter in the first set of parameters characterizing the skew of the first parameter function; determining a second parameter function defined by a second set of parameters, the second parameter function characterizing the distribution of a second portion of the time series data, and at least one parameter in the second set of parameters characterizing the skew of the second parameter function, wherein the first portion of the time series data represents a baseline window and the second portion represents an observation window; and determining whether the user is currently carrying an infection based on at least one measure of the difference between the first set of parameters and the second set of parameters.

[0005] In an example, determining a first parameter function includes: dividing a first portion of the time series data into a plurality of portions; determining a plurality of parameter functions that respectively characterize the distribution of each of the plurality of portions, wherein each of the plurality of parameter functions is defined by a set of parameters that at least characterize a skewness of the parameter function, wherein the first parameter function is determined as a mean or median of a set of parameters of the plurality of parameter functions.

[0006] In an example, the first distribution function and the second distribution function are each skew normal functions.

[0007] In an example, the at least one measure of difference between the first set of parameters and the second set of parameters includes determining at least one of: Kullback-Leibler divergence, effect size, and bulldozer distance.

[0008] In an example, determining whether the user is currently suffering from an infection comprises using at least one measure of difference between the first set of parameters and the second set of parameters as an input to a predictive model.

[0009] In an example, the predictive model is trained using data measured from a plurality of users.

[0010] In the example, the predictive model is a classification model.

[0011] In an example, the first portion of the time series data is collected over at least three weeks.

[0012] In an example, the second portion of the time series data is collected over a period of at most three days.

[0013] According to another aspect, a non-transitory storage medium storing program code, which, when executed by a processor, determines whether a user is infected based on data collected by a wearable device worn by the user, the program code comprising the following steps: receiving time series data representing physiological variables measured by the wearable device from the wearable device; determining a first parameter function defined by a first set of parameters, the first parameter function characterizing the distribution of a first portion of the time series data, at least one parameter in the first set of parameters characterizing a skew of the first parameter function; determining a second parameter function defined by a second set of parameters, the second parameter function characterizing the distribution of a second portion of the time series data, at least one parameter in the second set of parameters characterizing a skew of the second parameter function, wherein the first portion of the time series data represents a baseline window and the second portion represents an observation window; and determining whether the user is currently carrying an infection based on at least one measure of the difference between the first set of parameters and the second set of parameters.

[0014] In an example, determining a first parameter function includes: dividing a first portion of the time series data into a plurality of portions; determining a plurality of parameter functions that respectively characterize the distribution of each of the plurality of portions, wherein each of the plurality of parameter functions is defined by a set of parameters that at least characterize a skewness of the parameter function, wherein the first parameter function is determined as a mean or median of a set of parameters of the plurality of parameter functions.

[0015] In an example, the first distribution function and the second distribution function are each skew normal functions.

[0016] In an example, the at least one measure of difference between the first set of parameters and the second set of parameters includes determining at least one of: Kullback-Leibler divergence, effect size, and bulldozer distance.

[0017] In an example, determining whether the user is currently suffering from an infection comprises using at least one measure of difference between the first set of parameters and the second set of parameters as an input to a predictive model.

[0018] In an example, a predictive model is trained using data measured from multiple users. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In the drawings, like reference numerals generally refer to the same parts throughout the different views. The drawings illustrating features and modes of implementing various embodiments should not be construed as limiting other possible embodiments within the scope of the appended claims. Furthermore, the drawings are not necessarily drawn to scale, with emphasis generally being placed upon illustrating the principles of the various embodiments.

[0020] Figure 1 Depicted is a system for early detection of infection according to an example.

[0021] Figure 2 Depicted is a flow diagram of a method for early detection of infection according to an example.

[0022] Figure 3A Depicted are example histograms and maximum likelihood fits of baseline and observation window temperatures measured for COVID-19 positive users.

[0023] Figure 3B Depicted are example histograms and maximum likelihood fits of baseline and observation window temperatures measured for COVID-19 negative users.

[0024] Figure 4A Depicted are example Kullback-Leibler divergences between the baseline and observation window fitted functions obtained from COVID-19 positive and negative cases.

[0025] Figure 4B Depicted are example effect sizes between baseline and observation window fitted functions obtained from COVID-19 positive and negative cases.

[0026] Figure 4C Depicted are example earth mover's distances between baseline and observation window fitted functions obtained from COVID-19 positive and negative cases.

[0027] Figure 5 Depicted is a portion of a flow chart of a method for early detection of infection according to an example.

[0028] Figure 6 Depicted is a set of example maximum likelihood fits to temperature values ​​measured over a number of three-day periods computed between one week and six months before a single participant's COVID-19 test, along with the mean and median of the maximum likelihood fits. DETAILED DESCRIPTION

[0029] refer to Figure 1 , illustrates a system 100 for early detection of infectious diseases according to aspects of the present disclosure. Figure 1, an example system 100 is shown in FIG. As shown, system 100 includes a wearable device 102, a mobile device 104, and a remote server 106. In various embodiments and implementations, the wearable device 102, worn by a user, can detect various physiological variables (such as body temperature, heart rate, oxygen saturation, respiratory rate, etc.) and send data representing the physiological variables to the mobile device 104, which can relay the data to the remote server 106. In an alternative example, the wearable 102 can send the data directly to the remote server 106. The remote server 106 can analyze the data to determine whether the user wearing the wearable 102 is carrying an infection.

[0030] The analysis may include: (1) calculating the distribution function of the baseline window of data and the distribution function of the observation window of data; (2) finding a parametric function that defines the maximum likelihood fit of the two distribution functions, the parametric function being represented by a parameter that includes at least the skewness of the fit; and (3) finding at least one measure of the difference between the parameters of the fits of the two distribution functions. A determination is then made as to whether the user carries an infection based on the measure (or measures) of difference between the parameters of the parametric functions. This step may include inputting the measure (or measures) of difference into a predictive model (such as a classification model). Applicants have recognized and appreciated that differences in the shape of the distribution functions, and in particular differences in the tails of the distribution functions, are useful for early detection of infection. In combination Figure 2 and Figure 5 The method described in more detail can account for subtle and transient symptoms present in the early stages of infection with high accuracy. Once an infection has been detected, the user can be notified, such as through a display 108 on the wearable device 102 or a message that can be displayed on the display 110 of the mobile device 104 (e.g., through a notification from a dedicated application).

[0031] Wearable device 102 may include any device worn by a user or attached to the user's body that is suitable for measuring one or more physiological variables useful for achieving early detection of infection, as described in the present disclosure. Examples of such wearables include mobile devices that can be worn on the wrist to track physiological variables, such as smart watches or fitness trackers. However, other wearables such as smart clothing, smart patches, glasses, jewelry, earbuds, etc. are also within the scope of the present disclosure. Wearable device 102 may include one or more sensors for measuring physiological variables and a module for wirelessly transmitting data measured by the sensors to an external device (e.g., via Bluetooth). (However, it is contemplated that wearable device 102 may store such data and transmit the data when connected to another device by a wire.)

[0032] like Figure 1As shown, wearable device 102 can be connected to a mobile device 104, such as a smartphone or tablet. For example, such a device is typically paired with a smartphone for operation. Mobile device 104 can then be connected to one or more remote servers 106 via an Internet connection (e.g., via Wi-Fi or cellular connection), where early detection analysis can be performed. Alternatively, instead of a mobile device, wearable device 102 can be connected to any other suitable intermediate computing device, such as a laptop or desktop computer. In some examples, wearable device 102 can be connected directly to remote server 106 (e.g., similarly via Wi-Fi or cellular connection) rather than through an intermediate device.

[0033] In some alternative examples, a connection to a local server can be used instead of the remote server 106 (e.g., cloud computing). For example, in a hospital environment, a local server that is part of a local patient management system can be used to communicate with wearables worn by health workers and / or patients for early detection of infection. In other examples, the remote or local server can be omitted, and the calculations can be performed on the wearable itself or on a mobile or other computing device. In addition, the early detection analysis can be distributed among multiple devices, such as between the wearable device 102 and the remote server 106 or between the mobile device 104 and the remote server 106; other combinations are also conceivable.

[0034] Figure 2 A method 200 for early detection of infection using data collected by a wearable device is described. Figure 1 As described, the steps of method 200 can be performed by a remote server that receives physiological data collected by the wearable device and sent from a device paired with the wearable or directly from the wearable; however, in various examples, the paired device, the wearable, or another computing device (such as a laptop or desktop computer or a local server) can perform the analysis. In addition, the steps of method 200 need not be performed by one device, but can be distributed among multiple devices, including wearables, mobile devices or other computing devices, and local or remote servers. To this end, the steps of method 200 can be stored in one or more non-transitory storage media (e.g., located within a remote server or another device that performs the method) and executed by one or more processors in communication with the non-transitory storage media.

[0035] At step 202, time series data (e.g., vector data) representing physiological variables measured by a wearable device is received. The physiological variables can be any such variables useful for detecting the presence of an infection according to the following steps, such as body temperature, heart rate, and oxygen saturation. However, in practice, body temperature has been found to be most useful for early detection of infection. Although this step assumes that some device other than the wearable device is performing the calculations, this step can be performed later in the method, or can be omitted entirely, insofar as one or more steps are performed on the wearable device.

[0036] At step 204, a parametric function defined by a set of parameters that includes a distribution of a baseline window that skews and characterizes the time series data can be determined. The baseline window can be a portion of the physiological data collected over a period of time sufficient to form a data set representing values ​​that are normal for the user (i.e., when the user is not carrying an infection). In an example, the baseline data set can be collected over a period of ten days. However, confounding factors such as short infections (e.g., a user catches a minor cold during the baseline window) may interfere with collecting the baseline data. To account for this, a much larger window can be used, such as three months. In practice, a baseline window of twenty-one days has been found to be sufficient to accurately determine a user's baseline.

[0037] The baseline data set can be organized into a distribution, and a parametric function that accounts for the skewness of the distribution can be fit to it. The distribution can be determined by assigning a value to each interval of physiological values, which represents the total instances that the interval occurred within the time series. So, taking temperature as an example, for each measured temperature interval - for example 34°, 34.2°, 34.4°, 34.6°C - a number representing the number of times the temperature was measured can be assigned to that interval. So, if over the course of three months a temperature of 34.8 is measured 1500 times, the number 1500 would be assigned to 34.8. This data can then be normalized by the total number of measurements, yielding the frequency of each measurement occurring at each interval. Plotting this data will produce a histogram, such as the one discussed below Figure 3A and Figure 3B shown.

[0038] A parametric function that characterizes the distribution of the measured temperature values ​​can then be determined. As will be described in more detail below, the skewness of the maximum likelihood fit curve (i.e., its measure of the asymmetry of the distribution) is particularly useful for determining the presence of infection, and therefore the parametric function selected can include a parameter that characterizes the skewness of the function.

[0039] One such parametric function is the skew normal function, which in this example may be defined by the following equation (1):

[0040]

[0041] where Φ is the cumulative density function, φ is the Gaussian density function, and μ and τ are the mean and reciprocal of the standard deviation of the normal distribution, respectively. The additional parameter α is the skewness parameter; in the limit α = 0, the Gaussian distribution is restored. Other equations for the skewed normal function are known and are contemplated as alternatives. In addition, other parametric functions that take into account the skewness of the distribution, such as the exponentially modified normal distribution, can be used.

[0042] At step 206 , the method of step 204 may be repeated for an observation window of the data, and thus, a parametric function defined by a set of parameters that includes the skew and characterizes the distribution of the observation window of the time series data may be determined.

[0043] The observation window can be the portion of physiological data collected during a period of time sufficient to detect that the user is carrying an infection when compared to the baseline window, as described in steps 208 and 210. In practice, a length of three days has been found to be effective, however, shorter (e.g., one day) or longer (e.g., four days) lengths are contemplated. However, the observation window typically cannot be longer than the infection period because data unrelated to the infection status may corrupt the data within the observation window.

[0044] It is generally preferred that the baseline window terminate some time period (e.g., three days) before the start of the observation window to avoid including data that may indicate infection status in the baseline window. However, it is conceivable that there may be a certain amount of overlap between the data forming the baseline window and the data forming the observation window, particularly where the baseline window is acquired over a longer period of time (e.g., several months).

[0045] Similar to step 204, the distribution can be normalized by the number of measurements completed in the observation window. Normalization of both the baseline measurement data and the observation measurement data allows comparison of the two windows. Because the baseline window is typically much longer than the observation window, the total number of measurements within the baseline will be much higher, preventing a meaningful comparison without normalization. In addition, as in step 204, a parametric function that characterizes the distribution and accounts for skewness (such as a skewed normal distribution (as defined by equation (1) or other definitions accepted in the art) or an exponentially modified normal distribution) can be used to characterize the distribution.

[0046] exist Figure 3A and Figure 3B An example of a distribution and maximum likelihood fit of a time series of a physiological variable (as defined by a skew normal function)—as described in conjunction with steps 204 and 206—is shown in . Figure 3A Specifically, the temperature distribution for the temperature of the COVID-19 positive user example is shown. Specifically, the bars represent the distribution of normalized baseline and observation window temperature values. The fit to these distributions is a line representing a skewed normal function obtained from a maximum likelihood fit to the data. Figure 3B The skewed normal function characterizing the distribution of COVID-19 negative users’ baseline and observation window is shown. Figure 3B In the COVID-19 negative samples, there is close to total overlap between the skew normal functions. However, in Figure 3A In the COVID-19 positive example, the skewed normal function representing the observation window distribution is skewed to the right relative to the skewed normal function representing the baseline window distribution. Differences can also be noticed in the peaks, where the baseline window has a higher peak than the observation window. It is these differences, and particularly the differences in skewness, that can be exploited by characterizing the distribution using a parametric function that includes a skewness value.

[0047] At step 208, a determination is made as to whether the user currently carries an infection based on at least one measure of the difference between the first set of parameters and the second set of parameters. Any suitable measure of parameter difference may be used, such as Kullback-Leibler divergence, effect size, and bulldozer distance (also known as the Wasserstein metric), although other suitable metrics may be used. For example, effect size may be defined by equation (2):

[0048]

[0049] where μ and σ are the mean and standard deviation of the baseline and observation window functions, respectively. Notably, the effect size only quantifies the difference between the Gaussian components of the skew-normal function and ignores differences in the corresponding tails. In contrast, the Kullback-Leibler divergence and the Earthmovers distance account for differences in the skewness of the skew-normal function.

[0050] Figure 4A 、 4B Figures 4C and 4C depict the distributions of the calculated values ​​of Kullback-Leibler divergence, effect size, and bulldozer distance for a large number of COVID-19 negative users (denoted by n) and a large number of COVID-19 positive users (denoted by p), respectively. For each measure of variation, these distributions illustrate consistent differences in values ​​between COVID-19 negative and positive users.

[0051] The difference measure(s) of the parameters can be used to determine whether a given user carries an infection (such as COVID-19) by, for example, inputting the difference measure(s) into a predictive model, such as a classification model (although other suitable predictive models may be used). The predictive model will determine whether a given user carries an infection based on the difference measure(s). Such predictive models are well known, and a suitable predictive model may be used.

[0052] At step 210, the user is alerted that the user is carrying an infection. For example, the user may be alerted by a notification sent via the wearable itself (e.g., displayed on its screen or through or in combination with some haptic feedback) or by a notification sent via a mobile device or computer (e.g., a notification from a dedicated app, or a text message or email).

[0053] Figure 5 Depicted are steps of an example method 500 for determining a parameter function 204 (or 206) based on a large data set. More specifically, when calculating a parameter function for baseline data, a large number of data points (typically hundreds of megabytes) need to be stored in memory to form the distribution to which the parameter function is fitted. This is memory and processing intensive. Therefore, instead of storing all of the data, the data can be chunked (i.e., divided into smaller portions) as it is received. The parameter function can then be determined for the smaller portions of the data as it is received, where the parameters of the function are stored rather than the data representing the physiological variables. Once the parameter function for the chunked baseline window is determined, the median or mean of the stored parameters can be used to approximate the parameter function for the larger data set.

[0054] Thus, at step 502, the portion of data forming the baseline is divided into smaller portions as it is measured (e.g., every three days, although other time periods can be used). At step 504, the same parametric function fitting process described in step 204 is applied to the smaller portion of the data, resulting in a parametric function defined by a set of parameters, at least one of which describes the skewness of the function fitted to the smaller portion of the data. The parameters can then be stored instead of retaining the portion of data in memory. This operation is repeated for each portion of the data, such that multiple parametric functions are fitted to multiple portions of the data as the data is measured, with the parameters stored for each portion instead of the data to which the function was fitted. At step 506, the mean or median of the stored parameters is found. For example, if a skew normal distribution is found for each portion, the following parameters are stored for each portion: μ, μ, α. Thus, the mean or median μ, the mean or median μ, and the mean or median α can be found. The resulting parameters for the mean or median represent an approximation of the parameters for the complete data set for the baseline window.

[0055] This can be Figure 6 , which shows a set of exemplary maximum likelihood fits to temperature values ​​measured over a number of three-day periods calculated between one week and six months before a single participant's COVID-19 test, along with the mean and median of the maximum likelihood fits. Dashed lines 602 and 604 represent the maximum likelihood fits resulting from the mean and median values ​​of these parameters, respectively.

[0056] The above methods and systems can be deployed within a larger machine learning system that prepares the user's physiological variables to provide advanced tracking of the user's health by utilizing a large data set of thousands of similar users. For example, Philips North America (Cambridge, MA) has collaborated with the Department of Defense to develop a system for rapid threat exposure analysis - called RATE - to monitor the happiness and readiness of soldiers, as reported in Nature magazine in the journal Scientific Reports on March 8, 2022 (Conroy, B., Silva, I., Mehraei, G. et al. "Real-time infection prediction with wearable physiological monitoring and AI to aid military workforce readiness during COVID-19" (SciRep 12, 3797 (2022))). In addition to the large-scale analysis of artificial intelligence systems, the above methods and systems also provide the additional benefit of allowing the development of personalized baselines for each user.

[0057] The present invention represents a contemplated improvement in the ability of computers or computer systems to detect early-stage infections in users. Previous computers or computer systems have failed due to the transient and subtle nature of the symptoms detected. Thus, the present invention provides a robust and repeatable method for early detection of infectious diseases, which is crucial for curbing the spread of infectious diseases or providing early treatment to infected individuals.

[0058] Although this disclosure focuses on building features based on temperature collected from wearable devices, the methods described herein can be used for any other type of physiological measurement captured over an extended period of time. Furthermore, although a self-organizing parameter function has been used to model temperature, a maximum likelihood fit can be performed on a wide range of parameter functions, and a metric can then be defined to automatically assign the best-performing parameter function to a given physiological input.

[0059] The present invention may be a system, method and / or computer program product. The computer program product may include (one or more) non-transient computer-readable storage media having computer-readable program instructions for causing a system or processor to perform various aspects of the present invention. The computer-readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. The computer-readable program instructions described herein may be downloaded from the computer-readable storage medium to a corresponding computing / processing device, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, and / or a wireless network). Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, devices, systems, and computer program products according to embodiments of the present invention. It should be understood that each block of the flowchart illustrations and / or block diagrams and the combination of blocks in the flowchart illustrations and / or block diagrams may be implemented by computer-readable program instructions.

[0060] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0061] Unless explicitly indicated to the contrary, the terms "a" and "an" as used in the specification and claims herein should be understood to mean "at least one."

[0062] As used in the specification and claims herein, the phrase "and / or" should be understood to mean "either or both" of the elements so coupled, i.e., the elements being present in some cases in combination and in other cases separate. Multiple elements listed with "and / or" should be interpreted in the same manner, i.e., "one or more" of the elements so coupled. Other elements may optionally be present in addition to the elements specifically identified by the "and / or" clause, whether related or unrelated to those elements specifically identified.

[0063] As used in the specification and claims herein, “or” should be understood to have the same meaning as “and / or” as defined above. For example, when separating items in a list, “or” or “and / or” should be interpreted as inclusive, i.e., including at least one, but also including more than one element of a plurality of elements or a list of elements and optionally additional unlisted items. Only terms that clearly indicate the opposite, such as “only one of…” or “exactly one of…”, or, when used in a claim, “consisting of…” will refer to including exactly one element of a plurality of elements or a list of elements. Generally, the term “or” as used herein should be interpreted as indicating an exclusive alternative (i.e., “one or the other but not both”) only when preceded by an exclusive term (e.g., “any of…”, “one of…”, “only one of…”, or “exactly one of…”).

[0064] As used herein in the specification and claims, the phrase "at least one" referring to a list of one or more elements should be understood to mean at least one element selected from any one or more elements in the list of elements, but does not necessarily include at least one element of each and every element specifically listed in the list of elements, and does not exclude any combination of elements in the list of elements. This definition also allows that elements may optionally be present other than the elements specifically identified in the list of elements to which the phrase "at least one" refers, whether related or unrelated to those elements specifically identified.

[0065] It should also be understood that in any method claimed herein that includes more than one step or action, the order of the steps or actions of the method is not necessarily limited to the order in which the steps or actions of the method are described unless explicitly stated to the contrary.

[0066] In the claims and the foregoing description, all transitional phrases such as "comprises," "comprising," "carrying," "having," "containing," "involving," "holding," "including," etc. should be understood as open-ended, i.e., meaning including but not limited to. Only the transitional phrases "consisting of" and "consisting essentially of" should be closed or semi-closed transitional phrases, respectively.

[0067] Although several inventive embodiments have been described and illustrated herein, a person of ordinary skill in the art will readily conceive of various other modules and / or structures for performing the functions and / or achieving the results and / or one or more of the advantages described herein, and each of such variations and / or modifications is considered to be within the scope of the inventive embodiments described herein. More generally, a person skilled in the art will readily understand that all parameters, dimensions, materials, and configurations described herein are intended to be exemplary, and that actual parameters, dimensions, materials, and / or configurations will depend on one or more specific applications in which the teachings of the invention are used. A person skilled in the art will recognize or be able to determine, using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. Therefore, it should be understood that the foregoing embodiments are presented by way of example only, and within the scope of the claims and their equivalents, the inventive embodiments may be practiced in a manner other than that specifically described and claimed. The inventive embodiments of the present disclosure relate to each individual feature, system, article, material, kit, and / or method described herein. In addition, any combination of two or more such features, systems, articles, materials, kits, and / or methods is included within the scope of the invention of the present disclosure, provided that such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent.

Claims

1. A method for determining whether a user is infected based on data collected by a wearable device worn by the user, comprising the following steps: receiving, from the wearable device, time series data representing a physiological variable measured by the wearable device; determining a first parametric function defined by a first set of parameters, the first parametric function characterizing a distribution of a first portion of the time series data, at least one parameter of the first set of parameters characterizing a skewness of the first parametric function; determining a second parametric function defined by a second set of parameters, the second parametric function characterizing a distribution of a second portion of the time series data, at least one parameter in the second set of parameters characterizing a skew of the second parametric function, wherein the first portion of the time series data represents a baseline window and the second portion represents an observation window; as well as Whether the user currently carries an infection is determined based on at least one measure of difference between the first set of parameters and the second set of parameters.

2. The method according to claim 1, wherein Determining the first parameter function includes: dividing the first portion of the time series data into a plurality of portions; Determine a plurality of parameter functions that respectively characterize the distribution of each of the plurality of portions, wherein each of the plurality of parameter functions is defined by a set of parameters that characterize at least a skewness of the parameter function, and wherein the first parameter function is determined as a mean or median of the set of parameters of the plurality of parameter functions.

3. The method according to claim 1, wherein The first distribution function and the second distribution function are each a skew normal function.

4. The method according to claim 1, wherein The at least one measure of difference between the first set of parameters and the second set of parameters includes determining at least one of: Kullback-Leibler divergence, effect size, and bulldozer distance.

5. The method according to claim 1, wherein Determining whether the user is currently suffering from an infection includes using the at least one measure of difference between the first set of parameters and the second set of parameters as an input to a predictive model.

6. The method according to claim 5, wherein: The prediction model is trained using data measured from multiple users.

7. The method according to claim 5, wherein: The prediction model is a classification model.

8. The method according to claim 1, wherein The first portion of the time series data is collected over at least three weeks.

9. The method according to claim 1, wherein The second portion of the time series data is collected over a period of at most three days.

10. A non-transitory storage medium storing program code, wherein when executed by a processor, the program code determines whether a user is infected based on data collected by a wearable device worn by the user, the program code comprising the following steps: receiving, from the wearable device, time series data representing a physiological variable measured by the wearable device; determining a first parametric function defined by a first set of parameters, the first parametric function characterizing a distribution of a first portion of the time series data, at least one parameter of the first set of parameters characterizing a skewness of the first parametric function; determining a second parametric function defined by a second set of parameters, the second parametric function characterizing a distribution of a second portion of the time series data, at least one parameter in the second set of parameters characterizing a skew of the second parametric function, wherein the first portion of the time series data represents a baseline window and the second portion represents an observation window; as well as Whether the user currently carries an infection is determined based on at least one measure of difference between the first set of parameters and the second set of parameters. The non-transitory storage medium according to claim 10 , wherein: Determining the first parameter function includes: dividing the first portion of the time series data into a plurality of portions; Determine a plurality of parameter functions that respectively characterize the distribution of each of the plurality of portions, wherein each of the plurality of parameter functions is defined by a set of parameters that characterize at least a skewness of the parameter function, and wherein the first parameter function is determined as a mean or median of the set of parameters of the plurality of parameter functions.

12. The non-transitory storage medium according to claim 10, wherein: The first distribution function and the second distribution function are each a skew normal function.

13. The non-transitory storage medium according to claim 10, wherein: The at least one measure of difference between the first set of parameters and the second set of parameters includes determining at least one of: Kullback-Leibler divergence, effect size, and bulldozer distance. The non-transitory storage medium according to claim 10 , wherein: Determining whether the user is currently suffering from an infection includes using the at least one measure of difference between the first set of parameters and the second set of parameters as an input to a predictive model.

15. The non-transitory storage medium according to claim 14, wherein: The prediction model is trained using data measured from multiple users.

Images