[go: up one dir, main page]

WO2021262725A1 - Systèmes et procédés de segmentation d'une population d'utilisateurs sur la base de variations temporelles de niveaux de biomarqueurs - Google Patents

Systèmes et procédés de segmentation d'une population d'utilisateurs sur la base de variations temporelles de niveaux de biomarqueurs Download PDF

Info

Publication number
WO2021262725A1
WO2021262725A1 PCT/US2021/038494 US2021038494W WO2021262725A1 WO 2021262725 A1 WO2021262725 A1 WO 2021262725A1 US 2021038494 W US2021038494 W US 2021038494W WO 2021262725 A1 WO2021262725 A1 WO 2021262725A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
biomarker
variation
symptom
experiencing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US2021/038494
Other languages
English (en)
Inventor
Alec Mian
Stephan DONOGHUE
Gabriel BOUCHER
Amparo CASANOVA
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Curelator Inc
Original Assignee
Curelator Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Curelator Inc filed Critical Curelator Inc
Publication of WO2021262725A1 publication Critical patent/WO2021262725A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems

Definitions

  • Some current techniques for users to manage disease symptoms include keeping a written record of times when disease symptoms occur as well as times when the user engages in potential triggering (and/or perhaps mitigating) activities.
  • Other current techniques can include a user keeping an electronic diary of disease symptoms, disease triggering / mitigating activities, along with perhaps other disease monitoring and management related data.
  • Some platforms for managing disease symptoms involve providing recommendations to users to mitigate the effects of one or more symptoms.
  • Digital health management applications can require and/or rely on inputs for a user population in order to track how individual users in the user population experience symptoms. Tracking in this manner can allow a health management application to analyze each user’s data and provide each user with meaningful information about the user’s health and healthcare. For example, some digital health management applications require users to enter data into the health management application (e.g., via a user interface) and/or wear one or more data gathering devices that report data to the health management application (e.g., a fitness tracker, blood glucose monitor, or other device configured to track health-related data). This manner of obtaining data and any resulting analysis can permit the health management application to predict when a user might experience a symptom. Health management applications can provide one or more recommendations to assist in preventing or mitigating a predicted symptom.
  • data gathering devices e.g., a fitness tracker, blood glucose monitor, or other device configured to track health-related data.
  • IBS Irritable Bowel Syndrome
  • asthma each have episodic recurrences characterized by two or more phases of experiencing the symptom.
  • a migraine attack follows an interictal phase with no migraine symptoms, and the migraine attack can include three phases: a prodromal phase (usually 48 hours directly- preceding a migraine headache) associated with few non-headache symptoms (also referred to as “premonitory ' features”); a migraine headache phase associated with symptoms of a migraine headache; and a postdromal phase defined as a period directly following resolution of a migraine attack which is associated with non- headache symptoms.
  • Some healthcare management applications disadvantageous ⁇ attempt to provide predictions and/or recommendations to a user without leveraging the cyclical nature of experiencing the symptom, and therefore do not use all of the tracked information to arrive at a more reliable prediction and/or recommendation.
  • a method includes receiving, by a computing device, a plurality of inputs indicative of relationships between a symptom and variations in biomarker levels in individual users of a user population, wherein the plurality of inputs indicate, for each individual user, one or more biomarker levels, times of experiencing the variations in biomarker levels, and times of experiencing the symptom.
  • the method includes determining a plurality of biomarker-variation segments for the user population.
  • the method includes grouping users from the user population into the plurality of biomarker-variation segments based on a statistical analysis of the user population, wherein the biomarker-variation segments relate variations in biomarker levels over time to time-cycles of experiencing the symptom for a subset of the user population, and wherein the symptom is experienced by each user during a plurality of phases that sequentially progress during each time-cycle.
  • the method includes determining a biomarker- variation segment for a particular user based on inputs received from the particular user.
  • the method includes providing, by the computing device, a recommendation relating to an intervention for the symptom to the particular user based on the biomarker-variation segment determined for the user.
  • a non-transitory computer readable medium has stored thereon instructions executable by a computer system to cause the computer system to perform functions.
  • the functions include receiving a plurality of inputs indicative of relationships between a symptom and variations in biomarker levels in individual users of a user population, wherein the plurality of inputs indicate, for each individual user, one or more biomarker levels, times of experiencing the variations in biomarker levels, and times of experiencing the symptom.
  • the functions include determining a plurality of biomarker-variation segments for the user population.
  • the functions include grouping users from the user population into the plurality of biomarker-variation segments based on a statistical analysis of the user population, wherein the biomarker- variation segments relate variations in biomarker levels over time to time-cycles of experiencing the symptom for a subset of the user population, and wherein the symptom is experienced by each user during a plurality of phases that sequentially progress during each time-cycle.
  • the fimctions include determining a biomarker-variation segment for a particular user based on inputs received from the particular user.
  • the functions include providing, by the computing device, a recommendation relating to an interv ention for the symptom to the particular user based on the biomarker-variation segment determined for the user.
  • a system includes means for performing functions.
  • the functions include receiving a plurality of inputs indicative of relationships between a symptom and variations in biomarker levels in individual users of a user population, wherein the plurality of inputs indicate, for each individual user, one or more biomarker levels biomarker levels, times of experiencing the variations in biomarker levels, and times of experiencing the symptom.
  • the functions include determining a plurality of biomarker- variation segments for the user population.
  • the functions include grouping users from the user population into the plurality of biomarker-variation segments based on a statistical analysis of the user population, wherein the biomarker-variation segments relate variations in biomarker levels over time to time-cycles of experiencing the symptom for a subset of the user population, and wherein the symptom is experienced by each user during a plurality of phases that sequentially progress during each time-cycle.
  • the functions include determining a biomarker- variation segment for a particular user based on inputs received from the particular user.
  • the functions include providing, by the computing device, a recommendation relating to an interv ention for the symptom to the particular user based on the biomarker-variation segment determined for the user.
  • the systems and methods disclosed herein pror ide the above-described features and functionality and other benefits and capabilities.
  • the disclosed systems and methods provide improvements to shortcomings of existing methods for symptom management within the healthcare field by leveraging relationships between cycles of experiencing a disease symptom and variations in biomarker levels for each given user.
  • segmenting a user population based on characteristic variations in one or more biomarkers over time can pemiit improvements to predicting onset of one or more symptoms.
  • Figure 1 shows a system for tracking biomarker levels in a user population, according to an example embodiment.
  • Figure 2 shows an example client device, according to an example embodiment.
  • Figure 3 shows an example method, according to an example embodiment.
  • Figure 4 shows example variations in biomarker levels over time for different segments of a user population.
  • Figure 5 shows an example method, according to an example embodiment.
  • Example methods and systems are described herein. It should be understood that the words “example,” “exemplary,” and “illustrative” are used herein to mean “serving as an example, instance, or illustration.” Any embodiment or feature described herein as being an “example,” being “exemplary,” or being “illustrative” is not necessarily to be construed as preferred or advantageous over other embodiments or features.
  • the example embodiments described herein are not meant to be limiting. It will be readily understood that the aspects of the present disclosure, as generally described herein, can be arranged, substituted, combined, separated, and designed in a wide variety of different configurations, all of which are explicitly contemplated herein.
  • a disease symptom is an observable manifestation of a particular disease or disorder.
  • a disease symptom can be characterized by multiple characterization metrics, including but not limited to one or more of: (i) a time (or range of times) when the user experienced the disease symptom; (ii) a severity ' of the disease symptom; (iii) aspects or characteristics describing the disease symptom; and/or (iv) whether the disease symptom was accompanied by other related disease symptoms (and perhaps disease factors and/or disease triggers, which are described in further detail below).
  • the characterization metrics for the migraine headache symptom can include any one or more of: (i) when the headache occurred; (ii) how long the headache lasted; (iii) the intensity and/or severity of the headache; (iv) the location of the headache along the user’s head; and/or (v) whether the headache was accompanied by other related symptoms such as nausea or dizziness, and if so, the time, duration, intensity/severity of the accompanying symptoms.
  • Disease symptoms for other chronic diseases can include different characterization metrics. More generally, a symptom can be a physical manifestation that is not necessarily associated with a particular disease or disorder. For example, a user of a healthcare management application can experience a symptom without knowing, which, if any, known disease or disorder is causing the symptom.
  • a disease factor is any event, exposure, action, or conduct related to and/or performed or experienced by a user that has the potential to influence, affect, or cause the user to experience a disease symptom, or in some cases, prevent the user from experiencing a disease symptom.
  • Disease factors can include both: (i) voluntary or modifiable conduct and/or experiences by the user over which the user has at least some control, such as emotional states (anger, boredom, stress, anxiety, etc.), consumption of a particular food product, ingestion of a particular therapeutic agent, application of a particular therapeutic agent, ingestion of a particular dietary supplement or drug, performance of a particular physical activity, and/or exposure to a particular chemical agent; and (ii) involuntary or un-modifiable conduct and/or experiences, such as exposure to environmental factors (e.g., smog, sunlight, rain, snow, high or low humidity, or high or low temperatures), ingestion or other exposure to mandatory therapeutic agents or drugs (e.g., drags to treat or maintain other diseases), and effects of other diseases or physical conditions over which the user has little or perhaps effectively no control.
  • environmental factors e.g., smog, sunlight, rain, snow, high or low humidity, or high or low temperatures
  • mandatory therapeutic agents or drugs e.g., drags to
  • a disease factor can also be characterized by multiple characterization metrics, and different disease factors can have different characterization metrics.
  • the characterization metrics can include, for example: (i) when the user consumed the food or drug; and/or (ii) how much of the food or drag the user consumed.
  • Characterization metrics for an exposure-based disease factor can include, for example: (i) when the user was exposed; (ii) the intensity (e.g., bright sunlight) of the exposure; and/or (iii) the duration of the exposure.
  • disease factors can also include premonitory symptoms or warning signs that do not actually cause the user to experience a disease symptom but are closely associated with onset of a disease symptom for a particular user.
  • disease symptoms may be associated with one or more premonitory symptoms.
  • a premonitory symptom might be a craving for sweet foods before the user experiences the migraine headache.
  • the sweet craving does not cause the migraine, but instead is likely related to some physiological change associated with the migraine attack.
  • a particular physical manifestation felt by the user can be a disease symptom or a disease factor depending on its position in the physiological pathway.
  • reduced physical activity can be a disease factor because it tends to cause a disease symptom associated with obesity (e.g., excess fat stores).
  • reduced physical activity can be a disease symptom that is caused by, for example, osteoarthritis that often follows obesity due to increased fat mass.
  • symptoms can be observed in a user that are not associated with a particular disease.
  • a user can wish to hack a particular symptom that she has identified as her “most bothersome symptom” (i.e., a symptom that the user feels is most disruptive or distracting out of a set of symptoms that the user might experience).
  • a user can track the symptom itself, along with potential factors that potentially are associated with the symptom. In this manner, statistical associations can be formed between various symptoms and factors without each symptom necessarily corresponding to a particular disease.
  • the term “symptom factor” can be related to or include any event, exposure, action, or conduct related to and/or performed or experienced by a user that has can influence, affect, or cause the user to experience a symptom, or in some cases, prevent the user from experiencing a symptom, without the symptom necessarily being associated with a particular disease or disorder.
  • the disease factors described herein and defined above can more generally be referred to as symptom factors. Accordingly, any forthcoming description that involves disease factors as they pertain to disease symptoms can more generally be understood as symptom factors and corresponding symptoms.
  • a disease trigger is a disease factor that has been determined, for example through statistical analyses or other methods, to have a sufficiently strong association with a particular disease symptom for an individual user so as to become user-specific information of high interest and clinical use to the user.
  • a disease trigger can be strongly associated with causing the user to experience the particular disease symptom, or at least increasing the risk or likelihood that the user will experience the particular disease symptom.
  • a disease protector can be strongly associated with preventing the user from experiencing the particular disease symptom, or at least decreasing the risk or likelihood that the user will experience the particular disease symptom; such disease triggers can be referred to herein as protectors because they tend to reduce the user’s likelihood of experiencing the disease symptom.
  • a disease trigger for a user is a disease factor having a determined univariate association with a disease symptom for the user, where the determined univariate association demonstrates a statistically significant hazard ratio or odds ratio (greater than 1) or equivalent regression coefficient (greater than 0) at a fixed significant level (e.g. p-value less than 0.05) in univariate regression models or in equivalent multivariable models.
  • NAF no-association factor
  • a phase of a particular symptom can refer to a discrete period in which a particular physical manifestation or manifestations are expected in relation to a given symptom.
  • Some symptoms may manifest in a cyclical manner denoted by a plurality of recurring phases.
  • migraine headaches, Irritable Bowel Syndrome (IBS) and asthma each have episodic recurrences characterized by two or more phases of experiencing the symptom.
  • experiencing a disease symptom can be understood as experiencing a level of severity of the disease symptom, and may include a level of severity relative to prior periods of experiencing the symptom.
  • an interictal phase between migraine attacks can typically be associated with no manifestations of a migraine headache or corresponding disease symptoms.
  • phases might not be tied to biomarker levels in any discrete way.
  • different phases can be sequential and cyclical, and can be predictable in terms of order (e.g., phase 1 always precedes phase 2) even if they are less predictable in terms of duration (e.g., how long phase 3 will last for a particular cycle).
  • a biomarker can refer to a measurable substance in a user, or physiological state of the user, or a perceived stressor experienced by the user, or a specific combination of these, whose presence (or a degree thereof) is indicative of one or more phenomena experienced by the user.
  • a “substance” in this context can be a detectable molecule produced by the human body that can be measured using a sensor device.
  • a “physiological state” can include a condition or state of the body or bodily functions of a user.
  • a “stressor” can be an emotional or physical prompt experienced by a user or a physiological response to such a prompt.
  • a “stressor” in the context of a migraine can include one or more of stress, anxiety, heart rate variability', breathing pattern, hearing sensitivity ' , and sleep quality.
  • Other examples of biomarkers are possible as well.
  • a biomarker can be more broadly understood as including biological signs of a given disease or condition.
  • a biomarker level, and observed patterns in such biomarker levels during phases of a symptom can be leveraged to predict the symptom and to recommend an intervention to prevent or mitigate the predicted symptom.
  • Different biomarkers or combinations of biomarkers may be predictive of a symptom for different users.
  • Embodiments of the systems and methods disclosed herein generally include: (i) receiving a plurality of inputs indicative of relationships between a symptom and variations in biomarker levels in individual users of a user population, wherein the plurality of inputs indicate, for each individual user, detected biomarker levels, times of experiencing the variations in biomarker levels, and times of experiencing the symptom, (ii) determining a plurality of biomarker- variation segments for the user population, (iii) grouping users from the user population into the plurality of biomarker-variation segments based on a statistical analysis of the user population, wherein the biomarker-variation segments relate variations in biomarker levels over time to time-cycles of experiencing the symptom for a subset of the user population, and wherein the symptom is experienced by each user during a plurality of phases that sequentially progress during each time-cycle, (iv) determining a biomarker-variation segment for a particular user based on inputs received from the particular user, and (iv)
  • FIG. 1 shows a system 100 for tracking biomarker levels in a user population, according to an example embodiment.
  • System 100 includes a computing device 102, a plurality of client/sensor devices that are communicatively coupled to the computing device 102, and a database 114.
  • the plurality of client/sensor devices include client/sensor device 110 and client/sensor device 112.
  • Computing device 102 includes one or more processor(s) 104, a memory 106, and instructions 108.
  • the processor(s) 104 can include, for example, central processing unit(s) or CPU(s), one or more general purpose processors, or one or more specialized processors.
  • Memory 106 can include a tangible non-transitory computer readable memory.
  • the instructions 108 can be stored on the memory 106 and be executable by the processor(s) 104 to perform functions associated with computing device 102.
  • computing device 102 is depicted as a single device, it should be understood that computing device 102 can represent one or more servers, a server system, a cloud-based processing system, or a plurality of computing devices operating collectively.
  • database 114 is depicted as being separate from computing device 102, it should be understood that, in some contexts, database 114 can be incorporated into computing device 102, perhaps as part of the memory 106.
  • the plurality' of client/sensor devices can include one or more client devices associated with users of a healthcare management application (e.g., a client device might be a smartphone or tablet controlled by a user), and one or more sensor devices configured for determining biomarker levels for a given user.
  • a healthcare management application e.g., a client device might be a smartphone or tablet controlled by a user
  • sensor devices configured for determining biomarker levels for a given user.
  • computing device 102 can be associated with hosting and/or configuring a health management application deployed on one or more client devices.
  • Computing device 102 can receive inputs (e.g., inputs entered into the healthcare management application via a client device or data from a sensor device) that forms a dataset for use in a statistical analysis of one or more symptoms of a user population.
  • the user population can include a plurality ' of users associated with different accounts in the health management application.
  • Computing device 102 can be configured to perform this statistical analysis using a machine learning model, and can have an architecture designed for such modeling.
  • computing device 102 can be implemented with a deep learning architecture that uses a neural network to analyze the inputs and provide an output based on prior training.
  • the neural network can be trained to segment users in the user population in a supervised, semi-supervised, or unsupervised manner.
  • Other machine learning techniques can be implemented in computing device 102.
  • computing device 102 can segment the user population using regression analyses. Other ways of segmenting the user population are possible.
  • computing device 102 can use the segments of the user population to predict whether individual users might experience a symptom in a given timeframe and to correspondingly recommend one or more actions of a user. Examples of these determinations and functions are described below with respect to Figure 3, Figure 4, and Figure 5.
  • FIG. 2 shows an example client device 200 according to an example embodiment.
  • the client device 200 can be a smartphone, tablet, desktop or laptop computer, or any other type of computing device with the capability of generating, gathering, and/or presenting data disclosed and described herein to a user as well as performing any ancillary functions that can be required for effective implementation of the user population segmentation, symptom prediction, and recommendation methods disclosed and described herein.
  • Client device 200 includes hardware 206 comprising: (i) one or more processors (e.g., a central processing unit(s) or CPU(s) and/or graphics processing unit(s) or GPU(s)); (ii) tangible non-transitory computer readable memory (otherwise referred to as non-transitory computer readable media); (iii) input/output components (e.g., speaker(s), sensor(s), display(s), headphone jack(s) or other interfaces); and (iv) communications interfaces (wireless and/or wired).
  • processors e.g., a central processing unit(s) or CPU(s) and/or graphics processing unit(s) or GPU(s)
  • tangible non-transitory computer readable memory otherwise referred to as non-transitory computer readable media
  • input/output components e.g., speaker(s), sensor(s), display(s), headphone jack(s) or other interfaces
  • communications interfaces wireless and/or wired
  • the hardware 206 components of the computing device 202 are configured to ran software, including an operating system 204 (or similar) and one or more applications 202a, 202b (or similar) as is known in the computing arts.
  • One or more of the applications 202a and 202b can correspond to computer-executable program code that, when executed by the one or more processors, cause the client device 200 to perform one or more of the functions and features described herein, including but not limited to any (or all) of the features and functions of methods 300 and 500, as well as any other ancillary features and functions known to persons of ordinary skill in the computing arts that can be required or at least desired for effective implementation of the features and functions of methods 300 and 500, even if such ancillary features and/or functions are not expressly disclosed herein.
  • Figure 3 shows an example method 300, according to an example embodiment.
  • certain functional blocks of method 300 are illustrated in a sequential order, these blocks can in some instances be performed in parallel, and/or in a different order than those described herein.
  • the various blocks can be combined into fewer blocks, divided into additional blocks, and/or removed based on the desired implementation.
  • v arious blocks are depicted in terms of their function, they can be understood as corresponding to different parts of a system.
  • different blocks can be carried out as functional modules performed by a unitary device or can be carried out by one or more computing devices (e.g. smartphones, tablets, personal computers, or the like), servers, databases, or a combination thereof.
  • method 300 can be facilitated using inputs provided via a user interface on a client device associated with a given user (e.g., client device 200). These inputs, for example, can be made in response to input prompts provided by the client device.
  • a user’s computing device receives such a message or notification (e.g., an input prompt)
  • the user’s computing device displays the message or notification within a graphical user interface (GUI) on the computer device.
  • GUI graphical user interface
  • some messages can further include software-based links that launch user interface screens for the user to input his or her experience of a biomarker identified in the message, a time of experiencing the biomarker, and perhaps a level of experiencing the biomarker.
  • some messages can include software links that, when activated by the user (e.g., by touching the link within the message or by touching a link on a notification), cause the client device to generate and display one or more GUI screens that enable the user to input data characterizing his or her experience of a biomarker.
  • a user can provide information that indicates time- variations of biomarker levels, which can facilitate operation of method 300.
  • a sensor device can be used to provide real time, or near real time, data to facilitate operations of method 300.
  • block 302 can be performed to acquire population data.
  • a plurality of client devices and/or sensor devices can be used to acquire data from a population of users of a healthcare management application.
  • the healthcare management application can be hosted by a computing device (e.g., computing device 102) and deployed on a plurality of client devices.
  • the population data can indicate changes in biomarker levels over time for each user, and can also indicate which phases of a time-cycle of a symptom each user is experiencing.
  • the population data may include information that denotes whether a user is experiencing an interictal phase, prodromal phase, migraine headache phase, or postdromal phase of a migraine at a given time.
  • An example scenario can involve tracking of stress levels in a patient population versus a time cycle of a migraine.
  • a migraine can c orrespond to a time-cycle that includes an interictal phase associated with little to no symptoms of a migraine, a prodromal phase associated with a period directly preceding a migraine headache, a migraine headache phase associated with symptoms of a migraine headache, and a postdromal phase associated with a period directly following a migraine headache.
  • Each user can track stress levels or other biomarker levels throughout the time-cycle using the healthcare management application.
  • Block 304 can be performed to perform a statistical analysis on the acquired population data.
  • the statistical analysis can involve training a machine learning model with a portion of the population data (“training data”) to determine relationships between variations in biomarker levels and the time-cycle.
  • training data a portion of the population data
  • biomarker levels can vary as a time-cycle progresses, and these variations may indicate when the user may enter a phase of the time- cycle typically associated with negative aspects of the symptom.
  • the migraine headache phase may be associated with disabling head pain.
  • the statistical analysis can reveal a time phase (e.g., the prodromal phase) when intervention can be effective for the user to mitigate or prevent the headache.
  • Block 306 can be performed to generate biomarker-variation segments based on the statistical analysis.
  • the biomarker-variation segments relate biomarker levels 308 to a time- cycle 310 for a disease symptom.
  • some users can experience stress levels differently at different phases of a migraine cycle.
  • a timeline of stress levels can be established relative to a migraine cycle for a particular user, and the particular user can have a similar timeline to some other users in the user population.
  • segmenting the user population in this manner can be performed using a cluster analysis of the population. Grouping the user population in this manner can have significant practical implications because some subsets of the population can predictably experience a migraine attack while other subsets are less predictable.
  • Determining each phase of the time-cycle 310 can include tracking inputs from a client device and/or inputs from a sensor device.
  • the inputs can provide indications of whether a user is experiencing a symptom or an event, such as a migraine headache, abdominal pain, restricted breathing, or another recurring symptom.
  • the inputs may also indicate one or more symptom factors that may be premonitory symptoms for the symptom or event. These are symptom factors that may have been previously determined as statistically predictive of the onset of a symptom for a given user.
  • Detecting a premonitory symptom with the inputs may indicate that the user is in the prodromal phase of a migraine cycle.
  • Predictive symptoms can be observed in other recurring symptoms, such as those presenting in IBS or asthma.
  • variations in biomarker levels can be used for determining the phase of a symptom. For example, for a given user, prior patterns of biomarker levels may consistently align with each phase of experiencing the symptom such that a statistical association exists between the biomarker levels and the different phases of the symptom. The biomarker levels can then be used for determining which phase of the symptom a user is experiencing.
  • detecting what phase of a time-cycle a user is experiencing may include dividing the time-cycle into a pre-attack/ relapse phase (e.g., within X days of attack/relapse occurrence, where X is disease or symptom specific), an attack/relapse phase (e.g., days with disease specific symptoms present), and a post-attack/relapse phase (e.g., Y days following the last day with symptoms, where Y is disease specific) and interictal days (e.g., non- symptomatic days that are not part of the pre-attack/relapse phase or the post- attack/relapse phase).
  • a pre-attack/ relapse phase e.g., within X days of attack/relapse occurrence, where X is disease or symptom specific
  • an attack/relapse phase e.g., days with disease specific symptoms present
  • a post-attack/relapse phase e.g., Y days following the last day with symptoms, where Y is disease specific
  • interictal days e
  • determining the time-cycle for a given symptom may include tracking each time a symptom is experienced, and assigning phases to timeframes before (interictal phase and pre-attack/relapse phase), during (attack/relapse phase), and after (post-attack/relapse phase). Establishing these timeframes may be based on previously established benchmarks for a given disease or disorder.
  • a pre-headache (prodromal) phase can correspond to two days prior to the first day with migraine pain
  • a migraine headache phase may include migraine days during which the user experiences a headache and/or aura
  • a post-headache (postdromal) phase can correspond to two days following the last day with migraine pain
  • an interictal phase may correspond to other, non-migraine, days.
  • Block 312 can be performed to determine a biomarker-variation segment for a user. This can include receiving inputs from the user to a trained machine learning model and receiving a classification of the user. The classification can categorize the user in one of the biomarker-variation segments based on the inputs received from the user, for example, based on a pattern of biomarker levels associated with the user. In another example, categorizing a user can include determining that the user has similar biomarker levels (e.g., within 1 standard deviation) as the average biomarker levels in the biomarker-variation segment during multiple phases of the time-cycle.
  • biomarker levels e.g., within 1 standard deviation
  • Classifying users in this manner can permit improved accuracy in recommendations provided to users, or in deciding how effective an intervention is likely to be.
  • the user can be categorized in a segment of the population that responds well to specific migraine therapeutic intervention recommendations administered during the prodromal phase of a migraine cycle, but other population segments might not respond positively to such recommendations, and can rely instead on alternative acute interventions (e.g., using medication within prescribed dosages from a clinician).
  • User-specific measures can be recommended as well.
  • a recommendation may relate to a stress reduction program for the user.
  • Block 314 can be performed to provide a recommendation to the user.
  • feedback can be received from users in the segment of a particular individual user.
  • the feedback can rate how effective a therapeutic or acute intervention has been at various stages of the time-cycle for the symptom (e.g., at an interictal, prodromal, or migraine headache phase of the migraine cycle).
  • the recommendation can reflect this by suggesting that the user copy the most highly rated interv ention method. For example, this can involve providing a notification to the user via a user interface of a client device.
  • method 300 provides an adaptive way of identifying which segment of a user population in which to categorize a given user. Further, by first establishing whether the segment is receptive to intervention techniques, method 300 permits an increased likelihood of selecting patients that will respond to specific interv entions, while potentially redirecting other users to different options. While method 300 focuses on a single biomarker, it should be understood that many biomarkers can be tracked, and users can be segmented into different groups for each biomarker. Similarly, users can be segmented based on different groups of two or more tracked biomarkers. This can permit a more robust recommendation system.
  • method 300 is described using symptoms that might correspond to a particular disease or disorder (e.g., a migraine), it should be understood that method 300 can be more generally applied to symptoms experienced by a user of the healthcare management system.
  • method 300 can include or focus on tracking a most bothersome symptom identified by a user.
  • symptom analysis can further encourage user engagement by focusing on suggestions related to a symptom that the user would like to prioritize.
  • This can further permit a healthcare management platform to provide feedback to a user even where a symptom is not particularly tied to a disease or disorder, and the healthcare management application can thus address symptoms, such as chronic pain, that are not always decisively tied to a specific disease or disorder.
  • method 300 allows for robust and adaptive analysis of symptoms while encouraging user engagement.
  • method 300 is described in terms of population data and usage thereof to segment the population and to provide recommendations to individual users, it should be understood that, after receiving data from a population of users and performing a statistical analysis on the population data, individual data received from a single user can be used to categorize that user and lead to a recommendation. For example, during a first period of time, a machine learning model can be trained using the population data, then, during a second period of time, the individual data can be provided as an input to the trained machine learning, which outputs the a category for the user, or a recommendation for the user. Thus, after an initial training phase for the machine learning model, an actionable recommendation can be provided using only inputs from the individual user.
  • Figure 4 shows a set 400 of example variations in biomarker levels over time for different segments of a user population.
  • set 400 includes six groups: first group 402, second group 404, third group 406, fourth group 408, fifth group 410, and sixth group 412.
  • the groupings can correspond to biomarker variations during a migraine cycle, and in this example the biomarker that is tracked is stress level.
  • the stress levels can be received from a client device based on user inputs that track stress level on a scale of 1 to 10, where 10 is the highest level of stress and 1 is the lowest level of stress.
  • Other types of biomarkers can be tracked using sensor devices.
  • sleep quality can be tracked using an accelerometer in a client device or other sensor device placed on or near a user’s bed.
  • the heart rate of a user can be tracked using a heart rate monitor, such as a heart rate monitor on a smartwatch.
  • the client device can act as one or more sensor devices and also receive inputs from a user.
  • the same device can be used to determine a phase in a time-cycle of the symptom and to determine biomarker levels for the user. Further details regarding tracking the phases of a time-cycle of a symptom are provided below with respect to Figure 5.
  • each segment of the user population shows a different pattern of stress level across different phases of a migraine.
  • Phase P0 corresponds to an interictal phase associated with little to no symptoms of a migraine
  • phase P I corresponds to a prodromal phase associated with a period directly preceding a migraine headache
  • phase P2 corresponds to a headache phase associated with symptoms of a migraine headache
  • phase P3 corresponds to a postdromal phase associated with a period directly following resolution of a migraine headache.
  • Phase P0, P1, P2, and P3 are characterized as cyclical because P3 periodically transitions to P0 after a migraine headache of the user. Different symptoms can have different phases, or can lack discrete phases.
  • first group 1 is characterized by users having higher stress levels during phases P0, P1, and P2, but having reduced stress levels during P4.
  • first group 402 is responsive to interventions, such as therapeutic or acute interventions.
  • a user that is grouped in first group 402 can be a candidate for a recommendation to intervene.
  • a timing of the intervention can correspond to maximizing or minimizing the benefit for an intervention.
  • the users in first group 402 can receive the maximum benefit from a therapeutic interv ention during P I and a maximum benefit from an acute intervention during P2.
  • the healthcare management application can thus provide different recommendations at different times depending on which segment a user falls into.
  • a biomarker might act as a disease trigger for a symptom.
  • high stress levels can cause a migraine headache.
  • recommendations can be tailored to this type of association.
  • second group 404 includes users that experience migraine headaches after reaching a threshold stress level during PI .
  • a recommendation might be provided during P0 that encourages a user to engage in meditation or another stress-reducing activity. This can help to prevent or delay a migraine attack.
  • a given group might not receive a benefit from recommended interventions.
  • third group 406 might not have statistically significant relief from migraine symptoms in response to a therapeutic intervention.
  • a recommendation for third group 406 can include alternative steps, such as to be ready for a potentially upcoming migraine headache. This can allow users to have an acute treatment available for upcoming symptoms, or to consult a clinician for alternative treatment prior to experiencing a migraine headache.
  • selecting each user for a given segment of the user population can be performed in accordance with a statistical analysis of a user population. This can manifest as grouping users together that share similar patterns of biomarker levels over the course of a time-cycle of a symptom. For example, this can include training a machine learning model with a portion of population data in order to form discrete categories of users or provide one or more outputs used for a clustering analysis of the population data. Providing inputs from an individual user to the trained machine learning model can then result in the user being grouped with a particular segment associated with a given biomarker. This same process can be performed for several different biomarkers to increase the likelihood that a beneficial recommendation can be provided to the user.
  • Figure 5 shows an example method 500, according to an example embodiment.
  • the blocks are illustrated in a sequential order, these blocks can in some instances be performed in parallel, and/or in a different order than those described herein. Also, the various blocks can be combined into fewer blocks, divided into additional blocks, and/or removed based on the desired implementation. Additionally, the example method 500 shows a client device performing some steps and a server performing other steps, but in alternative embodiments, some of the steps performed by a client in example method 500 can be performed by a server and vice versa.
  • each block can represent a module, a segment, or a portion of program code, which includes one or more instructions executable by a processor or computing device for implementing specific logical functions or steps in the method.
  • the program code can be stored on any type of computer readable medium, media, or memory, for example, such as a storage device including a disk or hard drive or other type of memory, such as flash memory or the like.
  • the computer readable medium can include non-transitory computer readable medium, for example, such as computer-readable media that stores data for short periods of time like register memory, processor cache and Random Access Memory (RAM).
  • the computer readable medium can also include non-transitory media, such as secondary or persistent long term storage, like read only memory (ROM), optical or magnetic disks, compact-disc read only memory (CD-ROM), and/or flash memory for example.
  • non-transitory media such as secondary or persistent long term storage, like read only memory (ROM), optical or magnetic disks, compact-disc read only memory (CD-ROM), and/or flash memory for example.
  • the computer readable media can also be any other volatile or non-volatile storage systems.
  • the computer readable medium can be considered a computer readable storage medium, for example, or a tangible storage device.
  • the blocks of example method 500 can be performed by one or more processors associated with one or more computing devices.
  • a server system can carry out the functions performed by each block.
  • the serv er system can include one or more computing devices, such as one or more servers, one or more storage devices, such as one or more databases, and one or more networking devices, such as one or more modems and/or routers.
  • a single one of these devices can carry out the functions described in relation to the blocks of method 300.
  • the various devices in the server system can collectively carry out these functions.
  • method 500 includes receiving a plurality of inputs indicative of relationships between a symptom and variations in biomarker levels in individual users of a user population.
  • a computing device such as computing device 102, can receive the inputs.
  • each input can be received from a client device controlled by a user or a sensor device that tracks biometrics of the user.
  • the plurality of inputs indicate, for each individual user, detected biomarker levels, times of experiencing the variations in biomarker levels, and times of experiencing the symptom.
  • These inputs can be representable in a manner similar to that presented in Figure 4. For example, timestamps can be matched between the biomarker data and time-cycle data of the symptom.
  • time-cycle data of the symptom can include symptom factors that indic ate which phase of a time-cycle of the symptom the user is experiencing.
  • premonitory symptoms can be used for determining or confirming a phase of the time-cycle.
  • the phases can be predetermined relative to each user experiencing the symptom.
  • the phases can be determined based on the biomarker levels determined from the inputs received from the user. [0059]
  • method 500 includes determining a plurality of biomarker-variation segments for the user population.
  • the biomarker-variation segments can be categories associated with one or more biomarkers, and a subset of the user population can be grouped into each category.
  • Each category can correspond to a pattern of biomarker levels such as those shown in Figure 4, and can be derived from the inputs received from the user population.
  • Individual users may experience different patterns of biomarker levels during different occurrences of the symptom. For example, during a first occurrence of a migraine, a user may experience a pattern similar to that of first group 402 shown in Figure 4, and during a second occurrence, the user may experience a pattern similar to that of second group 404.
  • a machine learning model can be applied to identify similar patterns of biomarker levels during each time-cycle of a symptom experienced by the user population. In this manner, a set of biomarker- variation segments can be established for each biomarker based on the total number of times the user population experiences a symptom.
  • method 500 includes grouping users from the user population into a plurality of biomarker-variation segments based on a statistical analysis of the user population. For example, grouping the users into the biomarker- variation segments can be performed using a machine learning model.
  • the biomarker-variation segments relate variations in biomarker levels over time to time-cycles of experiencing the symptom for a subset of the user population. For example, the segments can be determined based on providing time-aligned biomarker and time-cycle data (as shown in block 306 of Figure 3) to a neural network, and receiving an output that classifies each user in accordance with the aligned data.
  • the symptom is experienced by each user during a plurality of phases that sequentially progress during each time-cycle.
  • a given symptom can be experienced in a plurality of phases. These phases can vary in terms of symptom factors experienced at different times and the intensity of these symptom factors. Accordingly, a healthcare management platform can determine which phase the user is experiencing based on inputs from a client device, and by establishing patterns of variation in sequential biomarker measurements. Within examples, patterns in these sequential measurements can be established over a period of hours or days.
  • the user may experience different patterns of biomarker levels for different occurrences of the symptom. For example, during a first occurrence of a migraine, the user may experience a pattern similar to that of first group 402, and during a second occurrence, the user may experience a pattern similar to that of second group 404. A dominant pattern for each user may emerge. Accordingly, each pattern of biomarker levels can be categorized for each user, and each user may be grouped according to the dominant pattern associated with that user (e.g., the category most often associated with the user). In this manner, recommendations provided to the user have a greater chance of accuracy.
  • method 500 includes determining a biomarker-variation segment for a particular user based on inputs received from the particular user. For example, the particular user can be categorized into a biomarker-variation segment based on an output from a trained machine learning model.
  • the machine learning model can be trained based on a portion of population data received from a population of users.
  • the biomarker-variation segment can be associated with a dominant pattern of biomarker levels for the user.
  • method 500 includes providing a recommendation to the particular user based on the biomarker-variation segment determined for the user.
  • the recommendation can relate to an intervention for the symptom.
  • a recommendation can correspond to advising a user to take an action, such as to use a therapeutic interv ention, an acute interv ention, or to consult a clinician.
  • the recommendation can be provided as part of a notification sent to a client device and presented to the particular user on a user interface of the client device. These types of recommendations may be reserved for segments that are predictive of the user experiencing a symptom.
  • Each user may be grouped into a plurality of biomarker-variation segments corresponding to a plurality of tracked biomarkers, but only some of these may be predictive of experiencing the symptom for the user. Further, in some examples, a combination of biomarker patterns (e.g., stress levels and sleep duration) may be predictive for the user, but individually the biomarker patterns (e.g., only stress levels or only sleep duration) might not be predictive. Accordingly, providing the recommendations to a user may be based on a combination of biomarker- variations segments for the particular user. [0064] Within examples, grouping users from the user population into the plurality of biomarker-variation segments includes using the statistical analysis to determine the plurality of biomarker-variation segments.
  • determining each biomarker-variation segment includes determining a time-based signature of biomarker levels that is common to a subset of the user population.
  • the signature can relate to an average biomarker level at each phase of a time-cycle of the symptom for a given subset of the user population.
  • grouping users into the plurality of biomarker-variation segments can further include allocating each user in the user population to a particular biomarker-variation segment based on comparing inputs received from the user to the time-based signature of biomarker levels for the particular biomarker-variation segment.
  • this can involve determining a time- based signature for the biomarker-variation segment based on an average biomarker level during each phase of the symptom.
  • the time-based signature can be representable as shown in groups 402-412 of Figure 4.
  • the inputs from the particular user can reflect biomarker levels experienced by the particular user during each phase of the symptom.
  • the particular user can be placed in the biomarker-variation segment if the inputs align within a threshold level of the time-based signature, or in the biomarker-variation segment that best aligns with the user inputs.
  • Each user that falls within a threshold value (e.g., one standard deviation) of one or more average values of the signature might be placed in a corresponding biomarker-variation segment.
  • a correlation between biomarker levels for the particular user and the signature of biomarker levels can be used to determine which segment best fits the user. Users can be allocated to the segment with which the user’s biomarker levels have a greatest correlation.
  • the statistical analysis includes determining statistical associations between times of experiencing the variations in biomarker levels and times of experiencing the symptom based one or more of univariate or multivariable regression analyses applied to inputs received from the user population.
  • the statistical analysis includes training a machine learning model using training data associated with the user population, and applying the trained machine learning model to the plurality of received inputs to determine the plurality of biomarker- variation segments.
  • the statistical analysis can incorporate a supervised, semi- supervised, or unsupervised machine learning model.
  • method 500 further includes determining a relationship between a particular biomarker-variation segment and a therapeutic intervention for the symptom, determining that the particular user is in the particular biomarker-variation segment.
  • a first biomarker-variation segment can include users that are responsive to interventions, while a second biomarker segment includes users that are not responsive to recommended interventions.
  • providing the recommendation to the particular user includes providing a therapeutic intervention recommendation to the particular user responsive to determining that the particular user is in the particular biomarker-variation segment.
  • a healthcare management application can direct the user to attempt an interv ention if the user is in the first segment, but does not recommend an intervention for the user if the user is in the second segment.
  • method 500 can also include determining a recommendation set of users based on the users being in one or more of the biomarker-variation segments.
  • the recommendation set can include a plurality of users that are part of biomarker-variation segments that are responsive to recommended interventions.
  • method 500 further includes adding the particular user to the recommendation set of users based on the particular user being in the particular biomarker-variation segment, receiving an input indicative of the particular user experiencing less than a threshold level of effect of applying the therapeutic intervention recommendation relative to experiencing the symptom; and removing the particular user from the recommendation set based on the particular user experiencing less than the threshold level of effect of applying the therapeutic intervention recommendation.
  • a client device may provide a first prompt for the user to input whether the user used the recommendation (e.g., took the recommended therapeutic intervention) and a second prompt, perhaps at a later time, for the user input an extent to which the symptom was experienced (e.g., an intensity or duration of a migraine headache).
  • the extent to which the symptom was experienced can be compared to one or more past instances of experiencing the symptom. If less than threshold improvement (e.g., 10% less pain or duration of the migraine headache) is experienced, then the user can be removed from the recommendation set.
  • threshold improvement e.g. 10% less pain or duration of the migraine headache
  • the recommendation can initially be provided based on the user being in the particular biomarker-variation segment, but if the recommendation does not provide a benefit to that particular user, the user can be removed from the recommendation set of users to avoid the recommendation being applied unnecessarily to the particular user.
  • method 500 further includes determining that there is not a relationship between a particular biomarker-variation segment and a therapeutic intervention, and determining that the particular user is in the particular biomarker-variation segment.
  • providing the recommendation to the particular user comprises providing a recommendation to consult a resource (e.g., to contact a clinician or a healthcare provider) responsive to determining that the particular user is in the particular biomarker- variation segment.
  • the recommendation can instead recommend an acute intervention or to consult with a clinician for additional options.
  • the biomarkers tracked by the health management application can include one or more of stress, anxiety, heart rate variability, breathing pattern, and sleep quality for a given user.
  • the plurality of biomarker-variation segments can include a set of biomarker-variation segments for each biomarker.
  • grouping users from the user population into the plurality of biomarker-variation segments includes grouping each user of the user population into a single biomarker- variation segment within each set of biomarker-variation segments. In this manner, a healthcare management application can group each user into a plurality of biomarker-variation segments, thereby increasing the chances of making a beneficial recommendation to the user.
  • grouping users from the user population into the plurality of biomarker-variation segments based on the statistical analysis of the user population includes performing a cluster analysis of the user population using (i) detected biomarker levels, (ii) times of experiencing the variations in biomarker levels, (iii) times of experiencing the symptom, and (iv) types of biomarkers as parameters for clustering users.
  • a cluster analysis can be performed on an output of the statistical analysis, or this clustering can be performed as part of the statistical analysis.
  • the symptom relates to a migraine
  • the time-cycles of experiencing the symptom comprise a plurality of phases of a migraine.
  • method 500 can further include determining variation profdes for each user in the user population based on inputs received during each of the plurality of phases of a migraine for the user population.
  • the variation profdes track variations in biomarker levels relative to the time-cycles of experiencing the symptom.
  • the variation profdes can be representable as depicted in Figure 4.
  • method 500 can further include determining, for each variation profile, an association between the variation profile and a therapeutic intervention, w'herein providing the recommendation to the particular user comprises providing the recommendation based on determining associations between the variation profiles and the therapeutic intervention.
  • the systems and methods described herein generally relate to software application technology, and more particularly to healthcare management software.
  • Software applications are typically most effective where users remain engaged and when churn rates are low.
  • healthcare management applications rely on consistent data from individuals with a patient population to arrive at meaningful conclusions. Accordingly, it is desirable within the software application and healthcare industry to promote user/patient engagement.
  • One way of promoting user engagement is by providing beneficial recommendations based on data provided by the user. The above-described examples achieve this by first establishing which biomarker-variation segment a user belongs to, then providing recommendations to the user based on which segment the user is a part of.
  • a healthcare management application “rewards” users for engagement, and thus encourages continued use of the software.
  • providing recommendations to the user before the user begins to experience painful aspects of a symptom e.g., a migraine headache
  • the described embodiments further drive user engagement by providing feedback that is personalized for each user.
  • the systems and methods described herein provide specific implementations directed towards improving the technical fields of (i) software applications and (ii) healthcare management platforms.
  • example embodiments can involve (i) tracking a first number of biomarkers for each user, (ii) determining a lack of association between biomarker level variations and a symptom, and (iii) tracking a second number of biomarkers that is less than the first number of biomarkers.
  • This may further involve optimizing a machine learning model by removing one or more parameters corresponding to biomarkers that do not have an effect on a symptom. In this manner, computing requirements of the system can be reduced by determining which biomarkers are predictive of users experiencing a symptom.
  • Reducing tracked biomarkers for one or more users also involves reducing input prompts sent to a client device, which reduces data storage needs of the system and reduces edge computing strains on client devices resulting from tracking the biomarkers.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Investigating Or Analysing Biological Materials (AREA)

Abstract

L'invention concerne un procédé consistant à recevoir une pluralité d'entrées indiquant des relations entre un symptôme et des variations de niveaux de biomarqueurs chez des utilisateurs individuels d'une population d'utilisateurs. Le procédé consiste à déterminer une pluralité de segments de variation de biomarqueurs relatifs à la population d'utilisateurs. Le procédé consiste à regrouper des utilisateurs de la population d'utilisateurs dans la pluralité de segments de variation de biomarqueurs en fonction d'une analyse statistique de la population d'utilisateurs. Le procédé consiste à déterminer un segment de variation de biomarqueur pour un utilisateur particulier en fonction d'entrées reçues de l'utilisateur particulier. Le procédé consiste à fournir une recommandation relative à une intervention concernant le symptôme sur l'utilisateur particulier en fonction du segment de variation de biomarqueur déterminé pour l'utilisateur.
PCT/US2021/038494 2020-06-22 2021-06-22 Systèmes et procédés de segmentation d'une population d'utilisateurs sur la base de variations temporelles de niveaux de biomarqueurs Ceased WO2021262725A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US202063042030P 2020-06-22 2020-06-22
US63/042,030 2020-06-22

Publications (1)

Publication Number Publication Date
WO2021262725A1 true WO2021262725A1 (fr) 2021-12-30

Family

ID=79281736

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2021/038494 Ceased WO2021262725A1 (fr) 2020-06-22 2021-06-22 Systèmes et procédés de segmentation d'une population d'utilisateurs sur la base de variations temporelles de niveaux de biomarqueurs

Country Status (1)

Country Link
WO (1) WO2021262725A1 (fr)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060218010A1 (en) * 2004-10-18 2006-09-28 Bioveris Corporation Systems and methods for obtaining, storing, processing and utilizing immunologic information of individuals and populations
US20080091471A1 (en) * 2005-10-18 2008-04-17 Bioveris Corporation Systems and methods for obtaining, storing, processing and utilizing immunologic and other information of individuals and populations
US20130178386A1 (en) * 2004-04-26 2013-07-11 The Newman-Lakka Cancer Foundation Platelet biomarkers for the detection of disease
US20160324796A1 (en) * 2007-06-11 2016-11-10 Edge Therapeutics, Inc. Compositions and their use to treat complications of aneurysmal subarachnoid hemmorrhage

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130178386A1 (en) * 2004-04-26 2013-07-11 The Newman-Lakka Cancer Foundation Platelet biomarkers for the detection of disease
US20060218010A1 (en) * 2004-10-18 2006-09-28 Bioveris Corporation Systems and methods for obtaining, storing, processing and utilizing immunologic information of individuals and populations
US20080091471A1 (en) * 2005-10-18 2008-04-17 Bioveris Corporation Systems and methods for obtaining, storing, processing and utilizing immunologic and other information of individuals and populations
US20160324796A1 (en) * 2007-06-11 2016-11-10 Edge Therapeutics, Inc. Compositions and their use to treat complications of aneurysmal subarachnoid hemmorrhage

Similar Documents

Publication Publication Date Title
US11587672B1 (en) Tiered time-in-range guiding interfaces for diabetes management
US20150148621A1 (en) Methods and systems for creating a preventative care plan in mental illness treatment
CN113905663B (zh) 监测注意力缺陷伴多动障碍的诊断和有效性
US20160357173A1 (en) Evidence Generation and Data Interpretation Platform
US20230245773A1 (en) System and method for detecting or predicting return in major depressive disorder
US20210151194A1 (en) Dynamic Behavioral Phenotyping for Predicting Health Outcomes
JP2018524137A (ja) 心理状態を評価するための方法およびシステム
EP3005191A1 (fr) Système et procédé de soutien des soins de santé
US12183464B2 (en) Predication of a headache
US20160117469A1 (en) Healthcare support system and method
US10262759B2 (en) Personal health operating system
CN113679348A (zh) 血糖预测方法、血糖预测装置、设备及存储介质
WO2024209468A1 (fr) Prédiction de l'apparition et de la progression de maladies neurodégénératives à l'aide de données de test sanguin et de modèles d'apprentissage automatique
JP2023507730A (ja) 平均ユーザ対話データに基づいてアプリケーション・ユーザの心理状態をリモートでモニタするための方法及びシステム
JP7776421B2 (ja) 予測データと対応する受信データの差異を評価する統合状態推定予測
CN114830245A (zh) 遗漏丸剂剂量检测和相关系统、方法和设备
JP2022524869A (ja) 因果モデルを使用した個別医療
EP3605553B1 (fr) Systèmes et procédés d'évaluation discrète de la santé numérique
Squirrell et al. Blood pressure predicted from artificial intelligence analysis of retinal images correlates with future cardiovascular events
WO2021146527A2 (fr) Systèmes et procédés de réglage de données d'apprentissage sur la base des données de capteur
WO2021262725A1 (fr) Systèmes et procédés de segmentation d'une population d'utilisateurs sur la base de variations temporelles de niveaux de biomarqueurs
CN117119957A (zh) 焦虑和抑郁障碍的诊断和药物治疗有效性监测
Goldstein et al. Classifying individuals based on a densely captured sequence of vital signs: An example using repeated blood pressure measurements during hemodialysis treatment
CN112805785A (zh) 提高患者生活质量、参与度和持续性的疾病因子的早期反馈
US10517514B2 (en) Glycemic response insight detection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21828905

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21828905

Country of ref document: EP

Kind code of ref document: A1