Methods for determining a patient's risk score

Abstract

This disclosure generally relates to a computer-implemented method for updating a patient's treatment model. This disclosure also relates to a corresponding computer system and computer program product.

Description

[0001] Case Analysis This disclosure is a divisional application of the invention patent application filed on December 2, 2020, with application number 202080084103.4 and the invention title "Method for Determining a Patient's Risk Score". Technical Field

[0002] This disclosure generally relates to a computer-implemented method for determining a patient's risk score. This disclosure also relates to corresponding computer systems and computer program products. Background Technology

[0003] Over the past few decades, healthcare spending has grown rapidly, and various initiatives have been proposed to at least slow this growth. For example, such initiatives might focus on implementing a relatively high threshold for determining when to provide appropriate treatment to an individual, while still attempting to maintain the quality of healthcare within the system at a desired level.

[0004] As an alternative, doctors, nurses, or any other skilled therapists or medical advisors may attempt to advise individuals on related changes that could have a positive health impact on them, thereby reducing the risk that they will have to seek treatment within the healthcare system.

[0005] In order to determine when to treat a patient and when not to treat a patient, some form of pre-assessment is required.

[0006] When assessing an individual, such as a physician, nurse, or anyone assisting the patient, personal experience, guidelines, and best practices are used to define the individual's current condition as objectively as possible, and may influence recommendations for treatment. While physicians or therapists possess extensive knowledge bases, they are human and may sometimes be unaware of the latest developments in their field, thus failing to grasp an individual's overall condition, such as all relevant medical information. Furthermore, currently available best practices may, in some cases, rigidly impose personalized treatment on an individual.

[0007] Recently, digital solutions have been introduced to assist doctors or therapists, significantly reducing the subjectivity of their decisions while allowing for increased "resolution" of available best practices, enabling doctors or therapists to make decisions based on a larger amount of data. Such digital solutions also allow for the inclusion of all relevant medical information about an individual when defining their current state.

[0008] US20180165418 presents examples of available digital solutions for recommending associated changes. Specifically, US20180165418 discloses a system for collecting data that directly characterizes an individual's health, as well as associated data related to factors that may affect that individual's health. The system uses the collected factor data to construct a vector of features (“health vector”), which indicates and reflects an individual's health status over time. The system can also assess differences in an individual's health vector at different points in time to generate a health vector change. Using an individual's health vector and the health vector change, the system determines the individual's current health score, which characterizes the individual's overall health at that point in time (e.g., within a range from very healthy to very unhealthy). By periodically generating health scores based on updated health vector information, the system also constructs a trend of changes in an individual's health (“health score trend”) as the individual's health score changes over time. The system compares an individual’s health score trend data with the health score trend data of a population (i.e., one or more population groups) that reflects similar situations, and generates suggestions for actions or changes that can be taken by the individual based on the comparison and the behavioral patterns of the compared groups. These actions or changes can both improve the individual’s health and be adopted by the individual.

[0009] However, the solution proposed in US20180165418 has some common drawbacks. First, the solution proposed in US20180165418 rigidly assesses individuals, ultimately leading doctors / therapists to decide to bypass the possible recommendations of the digital solution in order to "keep it safe and secure" and ensure the individual's satisfaction.

[0010] Secondly, the solutions proposed in US20180165418 are only general recommendations for individuals and do not address the actions required when an individual is already hospitalized or requires actual treatment within the healthcare system. Therefore, the solutions proposed in US20180165418 will fail to address the problem of increasing healthcare spending, especially when actual treatment needs to be provided to individuals within the healthcare system.

[0011] Given the issues mentioned above, there appears to be room for physicians to further improve digital solutions in order to balance assessment reliability and healthcare quality, with the overall intention of providing the most appropriate type of treatment to an individual based on their current health / condition. Summary of the Invention

[0012] According to one aspect of this disclosure, a computer-implemented method is used to alleviate the above-mentioned problems. This method, executed by a control unit, is used to determine a patient's risk score. The method includes the following steps: receiving a first set of personal parameters at a control unit, the first set of personal parameters indicating the patient's current or previous state; using the control unit, forming a personal patient model based on the first set of personal parameters; using the control unit, determining a matching level between the personal patient model and each of a plurality of different predetermined general patient models, each of the general patient models having a predetermined patient risk score; using the control unit, selecting at least one general patient model, the at least one general patient model having a matching level above a predetermined threshold; and using the control unit, determining the patient's risk score based on the selected at least one general patient model.

[0013] The general idea of ​​this disclosure is to determine a patient's risk score, where a pre-assessment of the patient serves as the primary input. This risk score can then be used within a healthcare system to provide the most appropriate treatment to the patient. According to this disclosure, compared to existing technologies, the determination of a patient's risk score involves not only a pre-assessment but also a process of matching patient-related data with multiple different universal patient models. These different universal patient models are pre-formulated, possibly through close collaboration with experts in different fields, and are generally viewed as being associated with different patient behaviors and outcomes, such as in cases of inappropriate treatment. Furthermore, these different universal patient models are typically not based on knowledge relevant to a single patient, but rather on general knowledge (often anonymized) related to a large group of patients and the expected (combined) outcomes for those patients.

[0014] Therefore, according to this disclosure, a patient's individual model (depending on the data collected for the patient) is matched with multiple different general patient models, and at least one general patient model is selected, wherein the at least one general patient model has a matching level above a predetermined threshold. Thus, this approach does not determine a patient's risk score solely based on direct assessment of the patient, but rather ensures that the patient's assessment is placed within a "larger context" by matching the patient's specific behaviors with a "group" of patients who exhibit / perform in a similar manner.

[0015] Therefore, this disclosure allows for the determination of a patient's risk score based on general patient behavior, rather than relying solely on the individual patient. Consequently, the advantage of this approach is its ability to reliably predict a patient's anticipated future behavior and how best to manage such behavior to minimize complications. Matching to different predetermined general patient models can also be seen as a way to filter out possible variations in the patient's individual parameters, as these variations may have previously been determined to have little impact on the patient's future.

[0016] Therefore, the purpose of this disclosure can be to ensure improved quality of care provided to patients while ensuring a reduction in “overtreatment,” thereby alleviating the overall burden on the healthcare system. Furthermore, considering newly identified “best practices,” this disclosure can be implemented in a highly flexible manner, potentially ensuring the continuous introduction of “new” or “updated” universal patient models.

[0017] In the context of this disclosure, the phrase “a set of personal parameters indicating a patient’s current or previous state” should be interpreted broadly and includes any type of patient-related information that has been or is being collected. Such information may include, for example, clinical data of the patient collected over a predetermined period of time (including any data from seconds / hours to the patient’s entire lifespan, such as data collected during different physician appointments and / or hospitalizations), including but not limited to patient vital signs, number of hospitalizations, laboratory results, and prescribed medications. Further information that may be relevant to use includes, for example, heart rate data, electrocardiogram (EKG / ECG) data, respiratory rate data, patient temperature data, pulse oximetry data, and blood pressure data.

[0018] Of course, other patient-related parameters may also be used, such as: BMI (body mass index), urinary control, urinary incontinence, skin type, visual risk area, sex and age, malnutrition screening (MTS), activity level, other physical conditions, mental status, activity level, sensory perception, moisture of patient body parts, nutrition, friction and shear of patient body parts, body temperature, information related to previous pressure sores, perfusion (blood flow), diabetes, tissue perfusion and oxygenation, hygiene, hemodynamics, etc.

[0019] Preferably, in some embodiments, the set of personal parameters may include at least one of images and video sequences of the patient. However, it is suitable to allow caregivers to input other patient-related information, such as information related to the parameters listed above. Preferably, images and / or videos may be collected using, for example, a camera arranged to communicate with a control unit, wherein the control unit may apply an image processing scheme to extract, for example, the parameters listed above. In some embodiments, the image processing scheme may be suitable for standardizing previously collected patient data.

[0020] Furthermore, the term "personal patient model" should also be interpreted broadly, encompassing in one embodiment a set of personal parameters defining a patient. However, in another embodiment, a personal patient model may be defined as a "container" of the patient's personal parameters, for example, as a string of personal parameters organized according to predetermined criteria to improve fit with a generic patient model.

[0021] Furthermore, the term "control unit" should be interpreted broadly and may include any means for providing computing power to perform the scheme according to this disclosure. Thus, the control unit (corresponding to any means for providing processing power) may be implemented within a server, within a client device (e.g., a computer or mobile device), or shared between a server and a client device, as will be further discussed in the detailed description of this disclosure.

[0022] Preferably, in one embodiment of this disclosure, the selection step includes selecting a generic patient model with the highest matching level. Therefore, in some embodiments, a particular generic patient model can be precisely identified as the most relevant model, and the risk score is based on this match. In some embodiments, such implementation may be preferred, for example, where it is desirable to quickly determine a patient's risk score.

[0023] However, as an alternative, the risk score can be determined based on more than one selected generic patient model, for example, based on a combination of at least two selected generic patient models. In such an embodiment, it may be desirable to apply a weight to each of the selected generic patient models, where the weight may, for example, depend on the matching level. Clearly, this implementation may provide further improvements in the reliability of the determined risk score, but on the other hand, it may result in slightly more processing and therefore slightly slower processing compared to selecting only a single generic patient model.

[0024] In some embodiments, a predetermined threshold can be used to ensure that the match remains at at least at a specific baseline level. In other words, if the match is not low enough—that is, when no true match is produced when the individual patient model is compared with multiple different predetermined universal patient models—this information can be used as an indication that the physician should manually assess the patient's risk score without relying on the current protocol. In other words, a low match level can also be seen as an indicator that the individual parameters indicating the patient's current or previous state are incorrect or otherwise unreliable, and that it is appropriate to collect further / new patient-related information before proceeding to determine the risk score.

[0025] In one embodiment of this disclosure, the method further includes the steps of: defining low-risk, medium-risk, and high-risk categories using a control unit; and assigning a risk category to a patient by comparing the determined patient's risk score with a predefined range of risk scores for the different categories using the control unit. Other categories may be included and fall within the scope of this disclosure. Such other categories may, for example, include an intermediate "higher-risk category" between the medium-risk and high-risk categories. In some cases, the use of risk categories can help allow, for example, caregivers to quickly obtain information related to how to act on the patient, where, for example, different actions may have been previously (e.g., during training) assigned to different categories without having to interpret the "risk score number" (e.g., between 0 and 100, or otherwise defined). Therefore, if a patient is determined to be in a high-risk category, a caregiver can quickly take action to manage the patient.

[0026] Thus, in one embodiment of this disclosure, the scheme may further include the step of: using a control unit to formulate a recommended treatment for the patient based on the selected patient risk category, wherein the recommended treatment differs for different risk categories. In other words, instead of recommending treatment for all risk categories, this scheme excludes cases where the recommended treatment is only provided if "truly" needed, thereby reducing the overall burden on the healthcare system because providing treatment only to patients who truly need it will significantly reduce the total cost of treatment. It should be understood that the risk score ranges for different categories can be dynamic, meaning that the risk score ranges can change over time or be used to provide the most appropriate treatment to the patient with the aim of averaging total cost.

[0027] Preferably, the proposed treatment may include at least one of the patient's "pre-treatment" or a treatment product / treatment regimen for the patient. In the context of this disclosure, the term "pre-treatment" can refer to any form of treatment provided to the patient before, for example, knowledge of the disease. Such pre-treatment may include anything from nutritional recommendations to hygiene instructions. Similarly, the term "treatment product / treatment regimen" should be interpreted broadly to include any form or means suitable for use in connection with the active treatment of the patient, such as its use in treating the patient's wounds. Taking wound products as an example, wound products may include, for example, wound dressings, bandages, topical medications, treatment methods in combination with specific types of wound dressings, etc. Furthermore, current or future treatment products / treatment regimens are possible and fall within the scope of this disclosure.

[0028] In some embodiments, the method further includes the steps of: receiving a second set of personal parameters at a control unit, the second set of personal parameters indicating the patient’s status after receiving the recommended treatment; using the control unit to determine the patient’s health progress based on the first set of personal parameters and the second set of personal parameters; and using the control unit to compare the determined patient’s health progress with a predetermined health progress defined for at least one selected general patient model.

[0029] According to this disclosure, a time difference, for example, between 1 hour and 90 days, may be allowed between the collection of the first set of personal parameters and the collection of the second set of personal parameters. However, the mentioned time difference is merely an example, and it can certainly be shorter and longer. In one embodiment, the allowable time difference depends on the recommended treatment. Furthermore, it should be understood that the system may use more than the first and second sets of personal parameters, such as a third set of personal parameters, so that the time difference between data collection times can be fixed or variable.

[0030] Furthermore, at least some generic patient models can be provided with associated health progression. In other words, generic patient models with (or without) relevant treatment recommendations / suggestions may have defined expectations regarding how they are tailored to a patient. According to this embodiment, how a patient actually responds to the recommended treatment can be compared to the expected response (depending on the selected generic patient model). This comparison can then be used to further develop the scheme according to this disclosure. In other words, in some embodiments, such as when the patient's health progression substantially corresponds to a predetermined health progression of the selected generic patient model, the selected generic patient model can be "validated." Validation may include updating the selected generic patient model using further data or minor adjustments.

[0031] However, collecting and storing patient progress can also be equally useful when a patient's health progress deviates (significantly) from the predetermined health progress of a chosen universal patient model. In such cases, for example, a starting point can be formed where the deviated health progress can be considered a new situation compared to the previously expected outcome.

[0032] The patient-related information collected later can be used not only to update / adjust / validate the general patient model, but also, in contrast, to allow different institutions and / or organizations to benchmark against other institutions and / or organizations. Therefore, in some embodiments, it may be desirable to ensure that the collected patient-related information remains strictly anonymous.

[0033] In one embodiment, updating / adjusting a general patient model may include applying a machine learning process. In other words, instead of having a physician (or technician) create a new general patient model, the system itself can generate such a model or an iteration of an existing general patient model. For example, in some cases, a general patient model may be subdivided into two (or even more) sub-models (if further data is provided), showing that different evaluations can be made in different situations. The machine learning process may be an unsupervised machine learning process, a supervised machine learning process, and / or based on a convolutional neural network (CNN) or a recurrent neural network (RNN). Further implementations are possible and fall within the scope of this disclosure.

[0034] According to another aspect of this disclosure, a computer-implemented method is also provided, executed by a control unit, for determining a patient's risk score, wherein the method includes the steps of: receiving at the control unit a first set of personal parameters, the first set of personal parameters indicating the patient's current or previous state; using the control unit, matching the first set of personal parameters against a plurality of different predetermined universal patient models, each of the universal patient models having a predetermined patient risk score; using the control unit to select at least the universal patient model that best matches the personal parameters; and determining the patient's risk score based at least on the selected universal patient model. This aspect of the disclosure provides advantages similar to those discussed above with respect to the preceding aspects of the disclosure. Even so, according to this aspect of the disclosure, a slightly different method is provided in which personal parameters are directly matched against a plurality of different predetermined universal patient models, without including individual patient models. In some cases, for example, when the type of personal parameters is expected to be the same / similar across all collected examples, this implementation may be preferred.

[0035] According to another aspect of this disclosure, a computer-implemented method is provided, executed by a control unit, for reducing patient-related healthcare costs. The method includes: receiving at the control unit a first set of personal parameters, the first set of personal parameters indicating a patient's current or previous state; using the control unit, forming a personal patient model based on the first set of personal parameters; using the control unit, determining a matching level between the personal patient model and each of a plurality of different predetermined general patient models, each of the general patient models having a predetermined patient risk score; using the control unit, selecting at least one general patient model, the at least one general patient model having a matching level above a predetermined threshold; using the control unit, determining a patient's risk score based on the selected at least one general patient model; using the control unit, defining low-risk, medium-risk, and high-risk categories; using the control unit, assigning a risk category to the patient by comparing the determined patient's risk score to predefined risk score ranges for the different categories; and using the control unit, recommending treatment for the patient only when the patient is assigned a high-risk category. This aspect of the disclosure also provides advantages similar to those discussed above with respect to previous aspects of the disclosure.

[0036] Furthermore, according to another aspect of this disclosure, a computer system suitable for determining a patient's risk score is provided. The computer system includes a control unit adapted to: receive a first set of personal parameters indicating the patient's current or previous state; form a personal patient model based on the first set of personal parameters; determine a matching level between the personal patient model and each of a plurality of different predetermined general patient models, each of the general patient models having a predetermined patient risk score; select at least one general patient model having a matching level above a predetermined threshold; and determine the patient's risk score based on the selected at least one general patient model. This aspect of the disclosure provides advantages similar to those discussed above with respect to the preceding aspects of the disclosure.

[0037] In possible embodiments of this disclosure, the computer system is a mobile electronic device, such as at least one of a "dedicated electronic device," a mobile phone, a tablet computer, etc. Alternatively, the computer system may be a computer (e.g., a laptop computer) equipped with a camera, as discussed above, for acquiring images or video sequences of a patient's wound. For example, the computer system may be configured to be operated by a caregiver.

[0038] In a preferred embodiment of this disclosure, the computer system includes a graphical user interface (GUI) adapted to provide a caregiver with instructions for obtaining a first set of parameters for the patient. Then, following the processing steps discussed above, the GUI may be adapted to present information indicating risk scores and / or risk categories.

[0039] According to another aspect of this disclosure, a computer program product is provided, comprising a non-transitory computer-readable medium on which computer program means are stored, the computer program means being configured to operate a computer system suitable for determining a patient's risk score, the computer system including a control unit, wherein the computer program product includes: code for receiving at the control unit a first set of personal parameters, the first set of personal parameters indicating a patient's current or previous state; code for forming a personal patient model based on the first set of personal parameters using the control unit; code for determining, using the control unit, a matching level between the personal patient model and each of a plurality of different predetermined universal patient models, each of the universal patient models having a predetermined patient risk score; code for selecting at least one universal patient model using the control unit, the at least one universal patient model having a matching level above a predetermined threshold; and code for determining a patient's risk score based on the selected at least one universal patient model using the control unit. This aspect of the disclosure also provides advantages similar to those discussed above with respect to the preceding aspects of the disclosure.

[0040] Preferably, the control unit is a microprocessor. Similarly, the computer-readable medium can be any type of storage device, including removable non-volatile random access memory, hard disk drive, floppy disk, CD-ROM, DVD-ROM, USB storage, SD memory card, or one of the similar computer-readable media known in the art.

[0041] Further features and advantages of this disclosure will become apparent when examined in light of the appended claims and the following description. Those skilled in the art will recognize that different features of this disclosure may be combined to create embodiments different from those described below without departing from the scope of this disclosure. Attached Figure Description

[0042] Various aspects of this disclosure, including its particular features and advantages, will be readily understood from the following detailed description and accompanying drawings, wherein: Figure 1 A computer system according to a presently preferred embodiment of the present disclosure is conceptually illustrated; Figure 2 A possible client device is disclosed, comprising a graphical user interface applying the concepts of the present invention; and Figure 3 This is a flowchart illustrating the steps of performing a method according to a currently preferred embodiment of the present disclosure. Detailed Implementation

[0043] The present disclosure will now be described more fully below with reference to the accompanying drawings, in which presently preferred embodiments of the disclosure are illustrated. However, the disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided to be thorough and complete and to fully convey the scope of the disclosure to those skilled in the art. Similar reference numerals always refer to similar elements.

[0044] Now turn to the attached diagram, especially Figure 1 A computer system 100 suitable for determining a patient's risk score is conceptually shown.

[0045] Computer system 100 includes server 106 and client devices 112. Server 106 includes some form of control unit 108, which provides computing power, and server 106 is configured to communicate with database 110. Client devices 112 are configured to network with server 106, for example, using the Internet. Figure 1 In this embodiment, client device 112 is operated by a caregiver (not shown); however, any user, such as any form of caregiver, may be permitted to operate client device 112.

[0046] Network communication can be wired or wireless, including wired connections such as LANs, WANs, Ethernet, IP networks, and wireless connections such as WLAN, CDMA, GSM, GPRS, 3G mobile communication, 4G mobile communication, 5G mobile communication, Bluetooth, infrared, or similar.

[0047] Client device 112, in Figure 2 The device is described in further detail and shown as a mobile phone, including a graphical user interface (GUI) and a camera 204. The client device 112 also includes some form of control unit 206, which provides computing capabilities. The GUI is preferably adapted to present instructions and information to, for example, a caregiver, such as for acquiring images of the patient 106 using the camera 204, for receiving other patient data input by the caregiver, and for displaying information related to the patient's risk score and / or risk category.

[0048] Control unit 108 and control unit 206 may include a general-purpose processor, a special-purpose processor, circuitry including processing components, a set of distributed processing components, a set of distributed computers configured for processing, etc. The processor may be or include any number of hardware components for performing data or signal processing or for running computer code stored in memory. Memory may be one or more devices for storing data and / or computer code to perform or facilitate the various methods described herein. Memory may include volatile or non-volatile memory. Memory may include database components, object code components, script components, or any other type of information structure for supporting the various activities described herein. According to exemplary embodiments, any distributed or local storage device may be used with the systems and methods of this specification. According to exemplary embodiments, memory may be communicatively connected to the processor (e.g., via circuitry or any other wired, wireless, or network connection) and includes computer code for performing one or more processes described herein.

[0049] Furthermore, in one embodiment, preferably, the computer system 100 is implemented as a cloud-based computing system, wherein the server 106 is a cloud server. Therefore, computing power can be allocated among multiple different servers (not shown), and the location of the servers does not need to be explicitly defined. As described above, computing power can also be allocated between servers and client devices.

[0050] Furthermore, the advantage of using a cloud-based solution is the inherent redundancy. In other words, by applying a distributed approach to both the server and the user / operator, security is enhanced because it is typically impossible to attach to a specific operating site (physical or computer attack), as existing solutions typically have both a server and a user / operator at the designated operating site.

[0051] During the operation of computer system 100, further reference Figure 3 The process can be initiated by a caregiver, for example, by providing a first set of personal parameters through the GUI of the client device 112, the first set of personal parameters indicating the patient’s current or previous state, and then received by the control unit 108 of the server and / or the control unit 206 of the client device 112 (S1).

[0052] Subsequently, the control unit 108 / 206 may form a personal patient model based on the first set of personal parameters (S2). As described above, in some embodiments, the personal patient model may be a pre-assessment for the patient, or in another embodiment, the personal patient model may simply be a data string or vector holding the first set of personal parameters.

[0053] Then, the individual patient model is matched with multiple different predetermined universal patient models (S3), each of which has a predetermined patient risk score. In some embodiments, the multiple different predetermined universal patient models may be stored together with database 110 and / or with a storage module included in client device 112.

[0054] Matching an individual patient model with multiple different predetermined general patient models allows for the determination of a matching level. In some embodiments, the matching can be multidimensional, where a first set of individual parameters is matched with multiple different parameters associated with multiple different predetermined general patient models. It is possible that the first set of individual parameters does not necessarily correspond to parameters of multiple different predetermined general patient models, and that multiple different predetermined general patient models do not necessarily have the same type of parameters. Therefore, to find matches, it may be necessary to match parameters across multiple dimensions. Preferably, the matching level should take this into account, and in some embodiments may include determining the Euclidean distance between the different parameters.

[0055] Once the matching level is determined, at least one generic patient model is selected (S4). Even so, a prerequisite exists: only one or more generic patient models are selected, which have a matching level higher than a predetermined threshold. As mentioned above, this threshold can be dynamic and depends on the current implementation. Therefore, the threshold can range from 0 to 100 (where matching levels have a similar range).

[0056] After selecting at least one generic patient model, a patient's risk score can be determined (S5). The risk score determination will be based on the selected at least one generic patient model, but a combination of more than one generic patient model may also be allowed. In this implementation, different generic patient models may have different weights, for example, based on the matching level of each generic patient model.

[0057] Risk scores can be standardized between 0 and 100. Other ranges are, of course, possible and fall within the scope of this disclosure. Risk scores can also be used to determine a patient's risk category, possibly by allowing different ranges of the total range to be used for the risk score to correspond to different risk categories. In some embodiments, a risk score between 0 and 50 may correspond to a low-risk category, a risk score between 51 and 75 to a medium-risk category, and a risk score between 76 and 100 to a high-risk category. The ranges provided are for illustrative purposes only. It may be desirable to provide some form of treatment, at least to patients in the high-risk category.

[0058] This disclosure enables the rapid and efficient determination of a patient's risk score in an efficient manner, relying not only on patient-related data but also including matching schemes with multiple predetermined universal patient models. The advantages of this scheme include the ability to reliably predict a patient's expected future behavior and how best to manage such behavior to minimize complications. Matching with different predetermined universal patient models can also be seen as a way to filter out possible variations in the patient's individual parameters, as these variations may have previously been determined to have little impact on the patient's future.

[0059] The control functions of this disclosure can be implemented using existing computer processors, or by a dedicated computer processor of a suitable system, or by a hardwired system, wherein a dedicated computer processor of a suitable system is combined for this purpose or another. Embodiments within the scope of this disclosure include a program product comprising a machine-readable medium for carrying or storing machine-executable instructions or data structures. Such a machine-readable medium can be any available medium accessible by a general-purpose computer, a special-purpose computer, or other machine having a processor. For example, such a machine-readable medium may include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage, or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of machine-executable instructions or data structures and is accessible by a general-purpose computer, a special-purpose computer, or other machine having a processor.

[0060] Although the accompanying drawings may show a sequence, the order of the steps may differ from the depicted order. Furthermore, two or more steps may be performed simultaneously, or partially simultaneously. This variation will depend on the chosen software and hardware system and the designer's choices. All these variations fall within the scope of this disclosure. Similarly, software implementation can be accomplished using standard programming techniques with rule-based logic and other logic to implement various connection steps, processing steps, comparison steps, and decision steps. Moreover, even though this disclosure has been described with reference to specific exemplary embodiments, many different changes, modifications, etc., will become apparent to those skilled in the art.

[0061] Furthermore, by studying the accompanying drawings, the disclosure, and the appended claims, those skilled in the art can understand and implement variations of the disclosed embodiments in practicing this disclosure. Additionally, in the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

Claims

1. A computer system suitable for determining a patient's risk score, the computer system comprising a control unit adapted to: Receive a first set of personal parameters, which indicate the patient’s current or previous status; A personal patient model is formed, wherein the personal patient model is a data string or vector that maintains the first set of personal parameters; Determine the matching level between the individual patient model and each of several different pre-defined general patient models, wherein, Each generic patient model is based on anonymized knowledge associated with a patient population and the expected combination of outcomes for those patients, thereby representing a group of patients that appear or behave in a similar manner, and each of the generic patient models is set to have a predetermined patient risk score; Select at least one general patient model, wherein the at least one general patient model has a matching level higher than a predetermined threshold; The patient's risk score is determined by assigning a predetermined risk score from at least one selected generic patient model to the patient; Define low-risk, medium-risk, and high-risk categories; The patient is assigned a risk category by comparing the identified patient's risk score with a predefined range of risk scores for different categories; as well as A recommended treatment plan is formulated for the patient, wherein the recommended treatment is different for different risk categories.

2. The computer system according to claim 1, wherein, The selection of at least one general patient model includes: selecting a general patient model with the highest matching level.

3. The computer system according to claim 1, wherein, The risk score is determined based on a combination of at least two selected universal patient models.

4. The computer system according to claim 3, wherein, Each of the selected at least two general patient models has a weight to be applied when determining the risk score.

5. The computer system according to any one of claims 1 to 4, wherein, The personal parameters include multiple clinical data points collected from the patient within a predetermined time period.

6. The computer system according to claim 5, wherein, The clinical data includes at least the patient's vital signs, number of hospitalizations, laboratory results, and prescribed medications.

7. The computer system according to claim 6, wherein, The patient's vital signs include at least one of the following: heart rate data, electrocardiogram (EKG / ECG) data, respiratory rate data, patient body temperature data, pulse oximetry data, and blood pressure data.

8. The computer system according to claim 1, wherein, The treatment is administered to the patient only if the patient is assigned to the high-risk category.

9. The computer system according to claim 1, wherein, The control unit is also adapted to: Receive a second set of personal parameters, which indicate the patient's status after receiving the recommended treatment; The patient's health progress is determined based on the first set of personal parameters and the second set of personal parameters; as well as The determined health progress of the patient is compared with the predetermined health progress defined for the selected at least one general patient model.

10. The computer system according to claim 9, wherein, The control unit is also adapted to: Based on the combination of the determined individual patient model and the health progress comparison results, at least one of the general patient models is updated.

11. The computer system according to claim 10, wherein, Updating at least one of the general patient models includes applying a machine learning process.

12. The computer system according to claim 11, wherein, The machine learning process described is an unsupervised machine learning process.

13. The computer system according to claim 11, wherein, The machine learning process described is a supervised machine learning process.

14. The computer system according to claim 11, wherein, The machine learning process is based on convolutional neural networks (CNNs) or recurrent neural networks (RNNs).

Images