Development of information from health-related functional abstractions based on intra-individual temporal variance heterogeneity

Abstract

A method for automatically abstracting and selecting an optimal set of variance-related functions that are an indicator of an individual outcome in healthcare, wherein the method comprises: generating, by one or more processors, an abstracted set of variance-related candidate-patient functions, wherein the abstracted set of variance-related candidate-patient functions are temporally heteroscedastic functions; optimizing, by one or more processors, each patient function from the abstracted set of variance-related candidate-patient functions by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized, wherein this optimization produces an optimal abstracted set of variance-related patient functions from the time period in which the variances and heteroscedasticity of each patient function are maximized;by one or more processors comparing the optimal abstracted set of variance-related patient functions with a historical dataset for a patient population to create a predictive set of variance-related patient functions, wherein the predictive set of variance-related patient functions predicts a health-related target outcome of the patient population; by one or more processors generating a current optimal patient set of variance-related patient functions for a current patient; by one or more processors comparing the optimal set of variance-related patient functions for the patient population with the current optimal patient set of variance-related patient functions for the current patient;In response to the optimal set of variance-related patient functions for the patient population, which matches the current optimal patient set of variance-related patient functions for the current patient within a predefined limit, one or more processors determine whether the health-related target outcome matches a predefined health-related target outcome for the current patient; and in response to the health-related target outcome matching the predefined health-related outcome for the current patient, one or more processors issue an alert regarding the predefined health-related outcome for the current patient.

Description

TECHNICAL AREA The present disclosure relates to the field of computers and, in particular, the use of computers for analyzing data. Specifically, the present disclosure relates to abstracting and selecting optimal sets of variance-related functions with respect to medically treated patients. BACKGROUND OF THE INVENTION Disease self-management programs and intervention / care plan monitoring programs are limited by their inability to systematically utilize patient-generated data, particularly data requiring informed interpretation of the temporal context of the measurement (examples include, but are not limited to, a patient's weight over time, cholesterol levels, blood glucose levels, etc.). While existing technologies (multiple mobile apps and web-based portals) assist in capturing and storing relevant data, their ability to determine meaningful metrics that are highly specific to that individual is limited or nonexistent. This is because these technologies fail to account for the individual's specific circumstances regarding disease progression, medication profile, and other aspects of care that impact clinical performance indicators (CPIs). Publication WO 2011 / 124 758 A1 concerns a computer-executable method for characterizing the severity of a cancer marker in a tissue sample based on the tissue's gene expression data using a reference database. The method comprises: reading data from the tissue sample into computer memory, selecting a cancer marker, reading information into computer memory that defines a list of genes, including at least one gene indicative of the cancer marker within a selected cancer type, comparing at least one gene expression value from the sample with the corresponding gene expression value in the reference database, and returning a value or description that defines the position of the sample among the reference values. Document US 2013 / 0231953A1 concerns a system, procedure, and software product for matching members of a population, such as patients, based on similarities between the members. Patients are assigned to a bipartite graph, with patient nodes connected by weighted edges to clustered factor nodes and categorically grouped. When a new patient request is received, a similarity measure for each other patient is generated for each cluster by comparing the cluster edges. The cluster similarity measures are aggregated for each patient to obtain a global approximation measure for every other patient. Based on this global approximation measure, a list of the most similar patients is displayed, and feedback on the measurement can be provided. Document US 2007 / 0149952A1 concerns a method for controlling the administration of a pharmacological agent for the treatment of epilepsy. The method includes: processing one or more signals from a patient to characterize the patient's propensity for a future seizure, and, based on the measurement of an increased propensity for a future seizure, enabling the patient's access to the pharmacological agent from a dispensing device. BRIEF SUMMARY OF THE INVENTION The invention is based on the objective of creating a method, system, and computer program product for automatically abstracting and selecting an optimal set of variance-related functions that serve as an indicator for an individual outcome and personalized plan selection in healthcare. This objective has been achieved by the features of the independent claims. Embodiments of the invention are specified in the dependent claims. A procedure, system, and / or computer program product abstracts and automatically selects an optimal set of variance-related functions that are an indicator of individual outcomes and personalized plan selection in healthcare. An abstracted set of variance-related candidate patient functions is generated, exhibiting temporal heteroscedastic features. Each patient function from the abstracted set of variance-related candidate patient functions is optimized by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized. This optimization creates an optimal abstracted set of variance-related patient functions from the time period in which the variances and heteroscedasticity of each patient function are maximized.The optimal abstracted set of patient variance functions is compared to a historical dataset for a patient population to create a predictive set of patient variance functions, where the predictive set of patient variance functions predicts a health-related target outcome for the patient population. A current optimal patient set of patient variance functions is created for a current patient. The optimal set of patient variance functions for the patient population is compared to the current optimal patient set of patient variance functions for the current patient.In response to the optimal set of variance-related patient functions for the patient population, which matches the current optimal patient set of variance-related patient functions for the current patient within a predefined limit, a determination is made as to whether the health-related target outcome matches a predefined health-related target outcome for the current patient. If the health-related target outcome matches the predefined health-related outcome for the current patient, an alert is issued regarding the predefined health-related outcome for the current patient. BRIEF DESCRIPTION OF THE DRAWINGS Next, one or more embodiments of the invention are described by way of example only, with reference to the accompanying drawings, wherein: Fig. 1 represents an exemplary system and network in which the present disclosure can be implemented; Fig. 2 illustrates an exemplary architecture and process for developing information from health-related functional abstractions; Fig. 3 represents a simulated sequence of patient health measurements; Fig. 4 illustrates an estimated trend variance for the patient health measurements shown in Fig. 3; Fig. 5 represents another simulated sequence of patient health measurements; Fig. 6 represents a variance trend over time (VARiance trend Over Time - VAROT) of the patient health measurements shown in Fig. 5.Figure 7 is a table of VAROT measurements according to permutations of different incremental time periods from different observation windows used for the measurements shown in Figure 5; and Figure 8 is a high-level flowchart of one or more operations performed by one or more processors to abstract and select an optimal set of variance-related functions that are an indicator of an individual outcome and personalized plan selection in healthcare. DETAILED DESCRIPTION The present invention may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium (or storage media) containing computer-readable program code to induce a processor to execute aspects of the present invention. The computer-readable storage medium can be a concrete unit capable of retaining and storing instructions for use by an instruction execution unit. For example, a computer-readable storage medium can be, but is not limited to, an electronic storage unit, a magnetic storage unit, an optical storage unit, an electromagnetic storage unit, a semiconductor storage unit, or any suitable combination thereof.A non-exhaustive list of more specific examples of computer-readable storage media includes the following: a portable computer floppy disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static working memory (SRAM), a portable CD-ROM, a DVD drive (DVD), a memory stick, a floppy disk, a mechanically encrypted unit such as punched cards or raised structures in a groove with instructions recorded thereon, and any suitable combination of the foregoing. A computer-readable storage medium, as used herein, need not be designed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g.,Light pulses passing through an optical fiber cable) or electrical signals transmitted through a wire. The computer-readable program instructions described herein can be downloaded to respective data processing units from a computer-readable storage medium or to an external computer or storage device via a network, such as the internet, a local area network, a wide area network, and / or a wireless network. The network may include copper transmission lines, fiber optic cables, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each data processing unit receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage on a computer-readable storage medium within the respective data processing unit. Computer-readable program instructions for performing operations of the present invention can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or either source code or object code written in any combination of one or more programming languages, including Java, Smalltalk, C++, or the like, and conventional procedural programming languages ​​such as the programming language "C" or similar programming languages. The computer-readable program instructions can be executed entirely on the user's computer, partially on the user's computer as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server.In the latter scenario, the remotely located computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be established to an external computer (for example, via the internet using an internet service provider). In some embodiments, electronic circuits, including, for example, programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), can execute the computer-readable program instructions using state information from the computer-readable program instructions to personalize the electronic circuit in order to implement aspects of the present invention. Aspects of the present invention are described herein with reference to illustrations of flowcharts and / or block diagrams of processes, devices (systems), and computer program products according to embodiments of the invention. It is understood that each block in the illustrations of flowcharts and / or block diagrams, and combinations of blocks in the illustrations of flowcharts and / or block diagrams, can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other devices that process programmable data to create a machine such that the instructions executed via the processor of the computer or other devices that process programmable data create means for carrying out the functions / actions specified in the flowchart and / or block or blocks of the block diagram.These computer-readable program instructions can also be stored in a computer-readable storage medium capable of controlling a computer, programmable data processing device and / or other units to function in a particular manner, such that the computer-readable storage medium with the instructions stored therein constitutes a manufactured item, including instructions that implement the function / action specified in the flow chart and / or the block or blocks of the block diagram. The computer-readable program instructions can also be loaded onto a computer, other programmable data processing device, or other unit to cause the execution of a series of operations on the computer, other programmable device, or other unit to produce a computer-implemented process, such that the instructions executed on the computer, other programmable device, or other unit implement the functions / actions specified in the flowchart and / or block(s) of the block diagram. With reference to the figures, and in particular to Fig. 1, a block diagram of an exemplary system and network is presented that can be used by and / or in the implementation of the present invention. It should be noted that some or all of the exemplary architecture, including the hardware and software shown for and in a computer 102, can be used by a software-providing server 150 and / or a data storage system 152. The example computer 102 contains a processor 104 connected to a system bus 106. The processor 104 can use one or more processors, each with one or more processor cores. A video adapter 108, which controls / supports a display 110, is also connected to the system bus 106. The system bus 106 is connected via a bus bridge 112 to an input / output (I / O) bus 114. An I / O interface 116 is connected to the I / O bus 114. The I / O interface 116 enables data exchange with various I / O units, including a keyboard 118, a mouse 120, a mounting frame 122 (which can contain storage units such as CD-ROM drives, multimedia interfaces, etc.), a printer 124, and one or more external USB ports 126.While the format of the ports connected to the I / O interface 116 can be any known to experts in the field of computer architecture, in one embodiment some or all of these ports are Universal Serial Bus (USB) ports. As shown, the computer 102 is capable of exchanging data with a software-providing server 150 using a network interface 130. The network interface 130 is a hardware network interface, such as a network interface card (NIC), etc. A network 128 can be an external network such as the Internet or an internal network such as an Ethernet or virtual private network (VPN). A hard disk interface 132 is also connected to the I / O bus 106. The hard disk interface 132 forms an interface with a hard disk 134. In one embodiment, the hard disk 134 fills a system memory 136, which is also connected to the system bus 106. The system memory is defined as a lowest level of volatile memory in the computer 102. This volatile memory contains, but is not limited to, other higher levels of volatile memory (not shown), including cache memory, registers, and buffer memory. Data filling the system memory 136 includes an operating system (OS) 138 of the computer 102 and application programs 144. The operating system 138 includes a shell 140 for providing transparent user access to resources such as application programs 144. In general, the shell 140 is a program that provides an interpreter and an interface between the user and the operating system. Specifically, the shell 140 executes commands entered into a command-line user interface or commands from a file. Therefore, the shell 140, also known as the command processor, is generally the highest level in the hierarchy of operating system software and acts as a command interpreter. The shell provides a system prompt, interprets commands entered via the keyboard, mouse, or other user input devices, and sends the interpreted command(s) to the appropriate lower levels of the operating system (e.g., a kernel 142) for processing.It should be noted that although Shell 140 is a line-oriented, text-based user interface, the present invention also supports other user interface modes, such as graphics, speech, gestures, etc. As shown, the operating system 138 also contains the kernel 142, which contains lower levels of functionality for the operating system 138, including the provision of essential services required by other parts of the operating system 138 and application programs 144, including memory management, process and task management, disk management, and mouse and keyboard management. The application programs 144 contain a renderer, which is shown in an exemplary manner as a browser 146. The browser 146 contains program modules and instructions that enable a World Wide Web (WWW) client (i.e., the computer 102) to send and receive network messages to and from the Internet using Hypertext Transfer Protocol (HTTP) messaging, thereby enabling data exchange with the software-providing server 150 and other computer systems. The application programs 144 in the system memory of computer 102 (as well as the system memory of the software-providing server 150) also contain logic for intra-individual temporal heteroscedasticity analysis (IITVHAL) 148. The IITVHAL 148 contains code for implementing the processes described below, including those shown in Figures 2, 3, 4, 5, 6, 7 to 8. In one embodiment, computer 102 can download the IITVHAL 148 from the software-providing server 150, including on an on-demand basis, whereby the code in the IITVHAL 148 is only downloaded when it is needed for execution.Furthermore, it should be noted that in one embodiment of the present invention, the software-providing server 150 performs all functions pertaining to the present invention (including the execution of IITVHAL 148), thereby relieving the computer 102 of having to use its own internal data processing resources for the execution of IITVHAL 148. It should be noted that the hardware elements shown in Computer 102 are not exhaustive but are intended to be representative, highlighting essential components required for the present invention. For example, Computer 102 may include alternative memory units such as magnetic cartridges, DVDs, Bernoulli boxes, and the like. These and other variations are intended to fall within the scope of the present invention. With reference to Fig. 2, an exemplary architecture and process for developing information on health-related functional abstractions is presented. A system 200, which in one embodiment is the computer 102 shown in Fig. 1, contains a general population component 202 and an individual patient component 204. Within the general population component 202 and the individual patient component 204 are one or more processors (such as the processor 104 shown in Fig. 1, but not shown in Fig. 2) that perform one or more of the described steps 1 to 5. In step 1, an abstraction of a candidate function is generated. Abstracted / generated candidate functions change over time. That is, the candidate function abstraction creates a model of how one or more biological functions change for a patient over time to form an abstracted set of variance-related candidate patient functions. As described here, these variances of the patient functions are temporally heteroscedastic (i.e., they change differently over different time periods and according to how the time periods are subdivided for analysis). The variances can be univariate or multivariate. As an example, consider a univariate model in which a single type of biological event is measured. An example of a univariate model is a measured low blood cell count (the single type of biological event). A low blood cell count often leads to a significant proliferation of hematopoietic stem cells, which often results in leukemia (the endpoint). Thus, when a patient has a low blood cell count (i.e., a reduced number of red blood cells and / or white blood cells), the body generates more hematopoietic stem cells. These hematopoietic stem cells are precursor cells from which red blood cells (erythrocytes) and white blood cells (e.g., lymphocytes) are formed. In the case of white blood cells, the hematopoietic stem cells form immature white blood cells called blasts. These blasts then differentiate into mature white blood cells.When a patient is exposed to radiation or other environmentally induced mutagenic substances while their blood-forming stem cells are transforming into immature white blood cells (blasts), these blasts are at risk of mutation and an abnormal increase in number (i.e., leukemia). Therefore, repeated negative spikes (i.e., a decrease) in a patient's blood cell count are an indicator of a higher risk of leukemia. A multivariate model, as the name suggests, uses several biological events that exhibit variances. Consider, for example, a patient who has undergone general anesthesia for surgery. General anesthesia can affect several patient functions, including problem-solving ability, memory (short- and long-term), mood, and so on. By quantitatively measuring such functions (e.g., through functional magnetic resonance imaging (fMRI), writing / speech tests, etc.), fluctuations in these multiple abilities can be measured. As described here, such fluctuations (variances) can be used to predict an ultimate endpoint (e.g., level of cognitive health) for a population of patients and / or a specific patient. This variance in biological functions, which is used in one or more embodiments of the present invention for predicting endpoints, may be present according to the magnitude of the change (on an amplitude basis) or the frequency of the change (on a frequency basis). In one embodiment, the measured variances are therefore amplitude-based. This means that an event can fluctuate across different ranges. For example, a red blood cell count might fluctuate between 3.0 (million cells per microliter) and 6.0 during a first, longer period, and between 4.0 and 5.0 during a second, longer period. Thus, the amplitude-based variance during the first longer period (6.0 - 3.0 = 3.0) is greater than during the second longer period (5.0 - 4.0 = 1.0). This variance is therefore referred to as "amplitude-based variance." In one embodiment, the measured variances are frequency-based. That is, an event (e.g., a decrease in blood cell count, measured cognitive ability, etc.) can fluctuate with varying frequencies, so that the variance of the measured event is more widespread (i.e., more frequent) at certain times than at others. For example, blood cell counts may cyclically decrease to a level X every 7 days during a first extended period and every 3 days during a second extended period. Thus, the frequency of variance during the second extended period (every 3 days) is greater than during the first extended period (every 7 days). This variance is therefore referred to as "frequency-based variance." Referring again to Fig. 2, the complete function set is optimized (step 2) once a complete function set has been generated (i.e., according to the change of one or more patient attributes over time). This optimization is performed by analyzing selected variance functions from the complete function set (i.e., the abstracted / generated patient functions). This optimization includes identifying when certain variances are maximized. In one or more embodiments of the present invention, this optimization uses a Variance Trend Over Time (VAROT) algorithm, which is discussed in detail below. The VAROT analyzes variances according to the length of observation windows and the incremental time periods contained therein. That is, it is assumed that there are three time periods (observation windows) in which patient functions are monitored.The variances in these patient functions not only fluctuate between the three time domains, but the variances also depend on which intermediate time periods (incremental time periods) are used in each of the time domains. Once the optimized functional subset is created (i.e., a model showing time points at which variances are maximized), as described in step 3 of Figure 2, input data sources are filtered from a general population to match this data with the optimized functional subset. This identifies real-world data that match the optimized functional subset, including the predicted endpoint. In other words, step 3 finds databases containing the optimized functional subset (including maximized variances) as well as data describing the predicted endpoint (e.g., the onset of a disease in the populations described by the input databases) that occurs in patients whose functional subsets match those of the optimized functional subset. As described in step 4 of Fig. 2, the populated optimized functional subset (i.e., the “functional population”) is then compared with data from a database 206 and / or a database 208 for an individual patient. In one embodiment, the database 206 and / or the database 208 are provided by the data storage system 252 shown in Fig. 1. The database 206 contains data from the Electronic Health Records / Personal Health Records (EHR / PHR) for a particular patient. Data from the database 206 include historical data about that particular patient, including laboratory results, X-rays, physician reports, etc. The database 208 contains real-time data about a patient, obtained from wearable cardiac monitors, blood glucose meters, and other sensors that measure real-time conditions for a patient.The data from database 206 and / or database 208 are used to generate an optimized functional subset that has a similar format to the one created in step 2 for a broad patient population. If a match is found between the optimized functional subset created for the current patient and the optimized subset created for the general population (from step 2), an alert is triggered. In one embodiment, this alert indicates that such a match exists only if the optimized functional subset exceeds a certain measurement baseline for that patient. For example, a particular patient might have a heart rate that regularly fluctuates into the abnormally low range.However, database 206 confirms that this patient has a “sports heart”, in which bradycardia is simply caused by the patient’s high level of physical fitness and is not pathological. As described in step 5 of Fig. 2, it is determined whether the optimized functional subset actually matches key performance indicators (KPIs) for a specific patient. For example, suppose the user wants to know if a patient is at risk of stroke. The data from databases 206 and 208 can generate several different optimized functional subsets for the current patient. However, only the optimized functional subset "stroke" has an endpoint that is useful for predicting the risk of the patient suffering a stroke. Similarly, step 5 corrects the optimized functional subset for the general population (step 3) with data for the current patient, since the current patient is also part of the general population. If a match is found between a specific optimized functional subset from the general population (containing the desired KPIs) and the optimized functional subset for the current patient, an individually tailored plan (alert, intervention, therapy, treatment) is created for the current patient (Block 210). Further details regarding steps 1 to 5 shown in Fig. 2 are given below. Step 1: Functional abstraction A functional abstraction defines a specific candidate patient function for predicting a particular condition or event. With reference to Fig. 3, a diagram 300 represents a simulated sequence of a patient's health measurements. These patient health measurements can be derived from the patient's medical history (e.g., from database 206 shown in Fig. 2) and / or from raw sensor data (e.g., passed through database 208 shown in Fig. 2). The measurements can be values ​​from a blood count, vital signs (temperature, pulse, and respiratory rate), insulin levels, etc. In one embodiment, the patient functions are univariate (i.e., they involve only a single type of patient measurement). In another embodiment, the patient functions are multivariate (i.e., they consider several types of patient measurements). It is therefore assumed that diagram 300 is a simulated sequence of measurements x (i.e., a single patient function) with a length of 150 days from start to finish, generated using a normal distribution with a constant mean (mu = 100) and a non-constant variance over time. In this example, the observation window begins on day 30 (the first vertical dashed line) and ends on day 120 (the last vertical dashed line). The observation window is divided into three periods (dt), where dt = 30 days. Thus, the first period lasts from day 30 to day 60; the second period lasts from day 60 to day 90; and the third period lasts from day 90 to day 120. In this example, the period type is set to "discrete" (i.e.,It has a fixed period starting from a point "0" instead of a "rolling" period that resets to a new day each time to examine the next 30 days from the current new day. Finally, it is assumed that a condition is defined to stipulate that each period must have at least 10 measurements (s = 10) for the measurements to be valid. Figure 4 presents a Chart 400 illustrating an estimated trend variance for the patient health measurements shown in Figure 3. Chart 400 illustrates the estimated variance and its trend over time. The three triangles shown are example variances in each of the three time periods described above for Chart 300. The slope of line 402 through the triangles is positive, indicating an increasing amount of variance measured / captured in Chart 300. Line 402, fitted using the ordinary least squares (OLS) method, is the estimated trend variance over time (VAROT) for the data shown in Chart 300. It should be noted that the VAROT shown in diagram 400 is only an estimate, as it does not take into account the subdivisions of the three time periods represented by the triangles in diagram 400. An optimized version of the VAROT takes such subdivisions into account, as described below. The VAROT is abstracted from a sequence of measurements indexed by time for a predefined observation window. Generally, the VAROT is written as a function where: x is a sequence of time-indexed measurements; ts is a starting point of an observation window; wl is the length of the observation window; dt is an incremental period in one or more of the observation windows; pt describes a condition for the period type (either discrete or rolling period); and unds describes a sparsity condition (minimum requirement for data availability in each period). However, the VAROT shown in Fig. 4 is only a statistical approximation. To create a more useful VAROT, the VAROT is optimized, resulting in an optimized subset of functions (see step 2 in Fig. 2). Step 2: Functional optimization Obtaining a full sequence of measurements does not reveal the subordinate time period with the steepest variance trend in the patient's history. The VAROT abstracted from a subordinate time period with a larger variance gradient (in absolute values) is likely to be more closely related to the patient's future outcome. Therefore, an optimization framework seeks the optimal parameter set that returns the strongest VAROT signals in the patient's time-indexed measurements. With reference to Fig. 5, Diagram 500 represents another simulated sequence of a patient's health measurements. An incidental observation notes that as time progresses, a greater amount of amplitude variance appears to be present. However, within the entire 300-day period depicted in Diagram 500, there may be certain intervals where the amplitude changes more significantly. It is therefore assumed that there is a peak in the range between 70 and 130, and the peaks immediately before and after this 70 / 130 variance are between 80 and 120. Thus, the 70 / 130 interval (with a change of 60 points) and its 80 / 120 neighbors (with a change of 40 points) have a variance range difference of 20 (60 - 40) points.Further assuming there is also a peak value in the range between 80 and 120 (with a change of only 40 points), but the peak values ​​before and after this 80 / 120 peak are only 90 / 100. Thus, the 80 / 120 range (with a change of 40 points) and its 90 / 100 neighbors (with a change of 10 points) have a variance range difference of 30 (40 - 10) points. This means that although the absolute range of variation for the 70 / 130 peak (60 points) is higher than for the 80 / 120 peak (40 points), the change in the range between previous and subsequent peaks is greater for the 80 / 120 peak (with a variance range difference between it and neighboring peaks of 30) than for the 70 / 130 peak (with a variance range difference between it and neighboring peaks of 20). The VAROT formula described herein is used to identify where such maximum variance range differences occur. It is assumed that the following VAROT formula was used for the data points shown in diagram 500 in Fig. 5: Using these values, a diagram 600 in Fig. 6 represents the VAROT values ​​for patient health measurements shown in Fig. 5. It should be noted that the plotted points in diagram 600 may be color-coded according to a legend 602 and indicate the time points at which the VAROT is at a maximum value (indicating maximum variances in the recorded data), such as between time points 100 and 125. It should also be noted that the VAROT result is at a minimum value around time point 150 (indicating minimum variances in the recorded data). Thus, a table 700 in Fig. 7 shows VAROT measurements according to permutations of different incremental time periods from different observation windows used for the measurements shown in Fig. 5.As shown in Table 700, the maximum variance (as indicated by the VAROT value 68.66) occurs between time 90 (ts) and time 180 (wl = 90) when this period is divided into blocks of 25 days (dt = 25). Step 3: Functional Population As described herein, after creating an optimized functional subset using the VAROT formula, the optimized functional subset is configured to receive input data sources for identification across the general population. Thus, databases matching the abstracted / candidate trends created in steps 1 and 2 populate a database identified as such, thereby making the data available at the individual level for specific patients. This data-driven approach is used when data for an individual needs to be derived to make reliable assessments of an intervention. It should be noted that data from Electronic Health Records (EHRs), Personal Health Records (PHRs), and device data can be obtained for both the general population and specific patients. As also described above, univariate and multivariate data can be used for VAROT functional abstraction. Certain key design factors considered during function creation can be used as a starting point for analyzing a variance matrix over time (e.g., Table 700 shown in Fig. 7) generated by the VAROT algorithm. This means that when setting up the parameters for the VAROT algorithm, the following are taken into account: the number of available readings; the frequency of available readings; the observation time (i.e., total observation period – from ts to ts + wl); incremental time (i.e., daily, weekly, monthly, quarterly – dt); data sensitivity (i.e., how much data are affected by environmental conditions, seasonal changes, individual patient actions, etc.); the design of the time interval (wl); and allowable fluctuation levels (i.e.,Neglect of anomalous peaks that exceed a predefined threshold and are therefore likely artifacts; type of device used to obtain real-time readings; permissible levels of sparsity in the data (s); length of the observation window (wl): moving window or discrete window (pt); viewing before / after meals (i.e., patient activities that affect readings, such as diet, drinking habits, exercise, etc.); and knowledge of response variables (i.e., additional information that explains why variance may occur). Step 4: Setting the alert As described herein, baseline measurement data can be used to understand normal variance and to establish upper and lower control limits. This means that an alert is generated when the optimized functional subset of a current patient matches the optimized functional subset of the general population for patients reaching a specific endpoint (e.g., developing a disease). Once a trend in variance is detected, appropriate quality control charts and alerts are set up. Based on individual calibrations using variance techniques / alerts, triggers are created so that the healthcare provider can identify areas of concern during case management. In one embodiment, alerts are used to prompt the development of a personalized care plan for the patient based on the most accurately predicted VAROT function for that patient. This, in turn, can help develop the intervention scope and potentially serve as a basis for evidence generation to optimize interventions. In one embodiment, alerts serve as a basis for developing compliance programs that, using self-efficacy intervention or any coordinated care, form a basis for patient self-management. Step 5: Functional learning for adaptation When the optimized functional subset of the current patient is aligned with an optimized functional subset for the general population (of medical patients), the system checks and reconfirms that the selected abstraction is the correct one for the individual. This means that confirmation is made that the optimized functional subset for the general patient population leads to a desired endpoint (key performance indicator - KPI) (e.g., predicting a specific disease state). It should also be noted that different data readings are triggered by different events. For example, patient data may be read when a patient undergoes surgery, starts a specific medication, begins physiotherapy, etc. This results in a ts (described above) that influences which data is considered, thereby creating time periods, which in turn triggers a check to determine whether the selected function is the optimal one. It is important to note that the current VAROT process allows the system to differentiate patients according to their medical needs. This means that by predicting the likelihood of a specific class of patients reaching a particular endpoint (e.g., developing a disease condition) based on the strength of their VAROT scores, medical resources can then be allocated accordingly. Thus, in one embodiment, the process described herein uses statistical modeling techniques (e.g., mixed modeling) to segment patients based on the optimized set derived from the VAROT algorithm, data availability, and data completeness to predict the same outcome. As described herein, it should be noted that despite an analysis being carried out at the population level, intervention techniques can be applied at the individual level. With reference to Fig. 8, a high-level flowchart of one or more operations performed by one or more processors to abstract and select an optimal set of variance-related functions that are an indicator of an individual outcome and personalized plan selection in healthcare is presented. Following the initial block 802, an abstracted set of variance-related candidate-patient functions is generated by one or more processors (block 804). This abstracted set of variance-related candidate-patient functions are temporally heteroscedastic functions. The term "temporally heteroscedastic functions" is defined as functions that change 1) according to the time of a particular event at which they occur (according to the variables ts and wl in the VAROT algorithm described here), and 2) according to the time intervals at which the functions are measured (according to the variable dt in the VAROT algorithm). As described in Block 806, one or more processors optimize each patient function from the abstracted set of variance-related candidate patient functions by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized. The optimization creates an optimal abstracted set of variance-related patient functions from the time period in which the variances and heteroscedasticity of each patient function are maximized. For example, the VAROT formula in Diagram 600 in Fig. 6 identifies the variance of a given patient function as heteroscedastically maximized (i.e., 68, 66 are reached) in the time between a time marker 90 and a time marker 180 when this time period is partitioned into time segments of 25 units (see Table 700). As described in block 808 of Fig. 8, one or more processors then compare the optimal abstracted set of variance-related patient functions with a historical dataset for a patient population to create a predictive set of variance-related patient functions. As described herein, a predictive set of variance-related patient functions predicts a health-related target outcome for the patient population. As described in block 810 of Fig. 8, one or more processors then generate a current optimal patient set of variance-related patient functions for a current patient. As described in block 812, one or more processors compare the optimal set of variance-related patient functions for the patient population with the current optimal patient set of variance-related patient functions for the current patient. If a match is found (query block 814) (i.e.,If the optimal set of variance-related patient functions for the patient population matches the current optimal patient set of variance-related patient functions for the current patient within a predefined limit, one or more processors then determine whether the health-related target outcome matches a predefined health-related target outcome for the current patient (Block 816). This means that an investigation is performed to confirm that the candidate patient with variance-related functions actually leads to a desired KPI (e.g., predicting a diagnosis of a specific disease (Query Block 818)). If there is a match between the health-related target outcome and the predefined health-related outcome for the current patient, as described in block 820, one or more processors then issue an alert regarding the predefined health-related outcome for the current patient. This alert may be a warning regarding an increased risk of a disease, a recommended course of action to prevent / treat the disease, etc. The process ends at the closing block 822. In one embodiment of the present invention, the period in which the variances and heteroscedasticity of each patient function are maximized is identified by: generating a plurality of time-segment sizes by one or more processors; generating a plurality of subordinate time-segment sizes by one or more processors; creating various permutations of the plurality of time-segment sizes with the plurality of subordinate time-segment sizes by one or more processors; and identifying an optimal combination of a given time-segment size with a given subordinate time-segment size in which the variances and heteroscedasticity of each patient function are maximized. In one embodiment of the present invention, one or more processors, based on historical data for the current patient, create a normal variance in the current optimal patient set of variance-related patient functions for the current patient, wherein the normal variance is not predefined as a prediction of a disease state for the current patient. For example, the current patient may have a slow heart rate, which is "normal" (i.e., not harmful) for the current patient. One or more processors determine whether the current optimal patient set of variance-related patient functions exceeds the normal variance for the current patient.In response to the determination that the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance, one or more processors issue the alert regarding the predefined health-related outcome for the current patient. In one embodiment of the present invention, the predefined health-related outcome for the current patient is the implementation of a medical treatment plan to cure a disease condition from which the current patient is suffering. In this embodiment, the method further comprises: determining, by one or more processors, whether the implementation of the medical treatment plan has cured the current patient's disease condition within a predefined time period; and, in response to the determination that the implementation of the medical treatment plan has not cured the current patient's disease condition within a predefined time period, selecting, by one or more processors, a new set of variance-related patient functions for the current patient to generate a new current optimal patient set of variance-related patient functions for the current patient. In one embodiment of the present invention, one or more processors identify a trend in the time-heteroscedastic functions, wherein a positive trend indicates a temporal increase in variances for the time-heteroscedastic functions, whereas a negative trend indicates a temporal decrease in variances for the time-heteroscedastic functions, and wherein the positive trend and the negative trend describe changes in the amplitude of the variances for the time-heteroscedastic functions over time. In response to the detection of a positive trend in the time-heteroscedastic functions, one or more processors issue an alert regarding the predefined health-related outcome for the current patient. In one embodiment of the present invention, the abstracted set of variance-related candidate-patient functions for the general population, as well as variance-related patient functions for the current patient, is generated by one or more processors by maximizing a variance trend over time, wherein: where x = measured values ​​of a predefined measured patient characteristic, ts = a starting point of an observation window for observing the predefined measured patient characteristic, wl = a length of the observation window; dt = an incremental period length for a subunit of the observation window; pt = period type for the observation window, wherein the period type is selected from a population consisting of a discrete period and a rolling period, and s = a sparsity condition defining a required minimum number of data points for x in the incremental period in the observation window. In one embodiment of the present invention, the starting point of the observation window described in the VAROT formula is triggered by a predefined event related to the current patient. In one embodiment of the present invention, this predefined event related to the current patient is the start of a pharmacological protocol applied to the current patient. In one embodiment of the present invention, this predefined event related to the current patient is a surgical procedure performed on the current patient. In one embodiment of the present invention, this predefined event related to the current patient is a nutritional event occurring for the current patient. As described herein, the present invention describes a method and system that support the abstraction, construction, and populations of new functions by highlighting the variability of key performance indicators over time (heteroscedasticity), thereby enabling the use of insights gained from this function for the development, monitoring, and adaptation of services in care management, such as compliance. The system also includes a learning component that uses individual historical data to assess the sensitivity of the selected functional abstractions. The data-driven approach described herein enables the recording of the key figures and their associated temporal context without having to define theoretical models, and also offers the possibility of continuous monitoring and modification of the chosen abstractions. Use cases Clinical diagnoses and prognoses A concept underlying the present invention is that parameters of a biological model describing a previous development of a system (or an organism) serve as predictors of endpoints. This prediction can be univariate or multivariate. Univariate example A low red blood cell count leads to a significant increase in blood-forming stem cells. Since the probability of mutations (which ultimately lead to leukemia) is high under radiation exposure, certain measurable features of the dynamics of red blood cell counting could be considered risk factors for leukemia, e.g., the rate and maximum decline in the number of red blood cells in peripheral blood. Multivariate example Multivariate data collected on various human cognitive functions and their variances, measured over time, can be used to determine the long-term effects of anesthesia on perception. Some measurements obtained from general analyses of cognitive tests serve as predictors of the patient's future cognitive health and / or quality of life. The present invention uses two fundamental arguments in the analysis of variances (or other general variables) in functional abstractions and their applications: statistical and biological. Statistical analysis Statistical analysis creates predictors based on statistics to determine their predictive power at the endpoint. A logistically or linearly fitted line (e.g., using the difference between the last and penultimate values ​​of covariates, i.e., variance of previous measurements) is initially used as a trend line for variance trends as a predictor for the endpoint. These variances can be based on increased frequency variance or increased variance between data points (reduced interval between two consecutive data points). That is, within a given time period, many variances can occur ("increased frequency variance"), or there can simply be a "reduced interval between two consecutive data points" (i.e., within a predefined time subset in a period), without considering how many variances occur over the entire period. It should be noted that in one or more embodiments, mixed models for patient segmentation based on a significant abstraction of variance factors are used to predict the same outcome. This means that the VAROT formula described herein can identify certain populations / patients for whom a specific predefined outcome is likely. Biological analysis Although the present invention has been described as being based on statistical tools, it must be clear that the underlying data are based on biological / medical evidence, such that a correlation exists between the data attribute variability and the endpoint. That is, parameters of a biological model describe previous developments of a system (or an organism) that serve as predictors of endpoints in one or more embodiments. Examples of such biological analyses include, but are not limited to, the following exemplary applications: Radiation exposure: Data collected from counting reduced red blood cells under radiation exposure, as well as from accelerated stem cell regeneration to compensate for the loss of red blood cells, can be an indicator of an increased risk of leukemia.The low number of blood cells leads to a significant increase in blood-forming stem cells. Since the probability of mutations (which ultimately lead to leukemia) is high under radiation exposure, certain measurable characteristics of the dynamics of blood cell counting are considered risk factors for leukemia, e.g., the rate and maximum decline in the number of blood cells in peripheral blood. Kidney failure: Blood pressure data collected during surgery can indicate a higher risk of kidney failure. It is clinically known that prolonged periods of low blood pressure lead to kidney failure. Therefore, surgical protocols involving sub-normal blood pressure are used as a predictor of kidney failure. Heart disease: Continuously high blood pressure is less problematic than fluctuating blood pressure. A calculated variance is a better predictor of heart disease than the actual measured values. Cognitive functions (multivariate data): Data collected on various human cognitive functions (feeling, thinking, etc.) and their variances, measured over time, are used to determine the long-term effects of anesthesia on perception. Some measurements obtained from general analyses of cognitive tests (e.g., using factor analysis or latent class analysis) serve as predictors of the patient's future cognitive health and / or quality of life. All of these use cases can utilize the VAROT formula described herein to predict one or more specific outcomes / results. Personalized treatment Based on the prediction of outcome / consequence / result / endpoint, identified through the VAROT-based process described herein (i.e., capturing variations over time for individual predictions), personalized care plans and adherence programs can then be developed. Creating a tailored treatment plan or specific intervention leads to a favorable clinically feasible outcome for the provider or the patient. For example, depending on the variations over time for functions where weight management is a response variable, a personalized treatment plan leading to lifestyle and dietary changes can be implemented. One or more embodiments of the present invention are thus beneficial in the field of personalized medicine / predictive medicine. The goal of predictive medicine is to forecast the probability of a future illness so that healthcare professionals and patients themselves can take the initiative by modifying their lifestyles and utilizing more frequent medical monitoring. For example, a full-body skin examination every six months by a dermatologist or internist can be ordered if an increased risk of melanoma is detected in the patient. Similarly, an ECG and a cardiac examination by a cardiologist can be ordered if an increased risk of cardiac arrhythmia is detected in a patient.Similarly, MRIs or mammograms can be ordered every six months if a patient is found to have an increased risk of breast cancer. The data analysis can therefore be used in the field of personalized medication / predictive medicine using the VAROT-based process described herein. The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, processes, and computer program products according to various embodiments of the present invention. In this respect, each block in the flowchart or block diagrams can represent a module, segment, or code section containing one or more executable instructions for implementing the specified logical function(s). It should also be noted that in some alternative implementations, the functions specified in the block may occur in a different order than shown in the figures. For example, two blocks shown consecutively may actually be executed essentially in parallel, or the blocks may sometimes be executed in reverse order, depending on the functionality involved.It is also noted that each block in the block diagrams and / or in the illustration of the flowchart and combinations of blocks in the block diagrams and / or the illustration of the flowchart can be implemented by special systems based on hardware that perform the specified functions or actions, or combinations of special hardware and computer instructions. The terminology used herein serves only to describe certain embodiments and is in no way intended to limit the present invention. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It is further understood that the terms "indicates" and / or "indicating" as used in this patent specification indicate the presence of identified features, integers, steps, processes, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, processes, elements, components, and / or groups thereof. The corresponding structures, materials, actions, and correspondences of all means or step-plus-function elements in the following claims shall include all structures, materials, or actions for performing the function in combination with other claimed elements, as specifically claimed. The description of the various embodiments of the present invention has been prepared for illustrative and descriptive purposes; however, it is by no means intended to be exhaustive or limited to the present invention as disclosed. Many modifications and variations are obvious to those skilled in the art without deviating from the scope of protection and the inventive concept of the present invention.The embodiment was selected and described to best explain the principles of the present invention and its practical application, and to enable other persons skilled in the art to understand the present invention in various embodiments with different modifications suitable for the intended specific use. Furthermore, it should be noted that all methods described in this disclosure can be implemented using a VHDL (HVSIC Hardware Description Language) program and a VHDL chip. VHDL is an exemplary development input language for field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), and other similar electronic devices. Therefore, each software-implemented method described herein can be emulated by a hardware-based VHDL program, which is then applied to a VHDL chip, such as an FPGA. Having thus described in detail the embodiments of the present invention of the present application with reference to their illustrative embodiments, it is obvious that modifications and variations are possible without deviating from the scope of protection of the present invention as defined in the claims in the Annex.

Claims

A method for automatically abstracting and selecting an optimal set of variance-related functions that are an indicator of an individual outcome in healthcare, wherein the method comprises: generating, by one or more processors, an abstracted set of variance-related candidate-patient functions, wherein the abstracted set of variance-related candidate-patient functions are temporally heteroscedastic functions; optimizing, by one or more processors, each patient function from the abstracted set of variance-related candidate-patient functions by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized, wherein this optimization produces an optimal abstracted set of variance-related patient functions from the time period in which the variances and heteroscedasticity of each patient function are maximized;by one or more processors comparing the optimal abstracted set of variance-related patient functions with a historical dataset for a patient population to create a predictive set of variance-related patient functions, wherein the predictive set of variance-related patient functions predicts a health-related target outcome of the patient population; by one or more processors generating a current optimal patient set of variance-related patient functions for a current patient; by one or more processors comparing the optimal set of variance-related patient functions for the patient population with the current optimal patient set of variance-related patient functions for the current patient;In response to the optimal set of variance-related patient functions for the patient population, which matches the current optimal patient set of variance-related patient functions for the current patient within a predefined limit, one or more processors determine whether the health-related target outcome matches a predefined health-related target outcome for the current patient; and in response to the health-related target outcome matching the predefined health-related outcome for the current patient, one or more processors issue an alert regarding the predefined health-related outcome for the current patient. The method of claim 1, wherein the time period in which the variances and heteroscedasticity of each patient function are maximized is identified by: generating a plurality of time-segment sizes by one or more processors; generating a plurality of subordinate time-segment sizes by one or more processors; creating multiple permutations of combinations of the plurality of time-segment sizes with the plurality of subordinate time-segment sizes by one or more processors; and by one or more processors identifying an optimal combination of a given time-segment size with a given subordinate time-segment size in which the variances and heteroscedasticity of each patient function are maximized. The method of claim 1, further comprising: by one or more processors, based on historical data for the current patient, creating a normal variance in the current optimal patient set of variance-related patient functions for the current patient, wherein the normal variance has not been predefined as a prediction of a disease state for the current patient; by one or more processors, determining whether the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance; and, in response to determining that the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance, by one or more processors, issuing the alert with respect to the predefined health-related outcome for the current patient. The method of claim 1, wherein the predefined health-related outcome for the current patient is an implementation of a medical treatment plan to cure a disease condition from which the current patient is suffering, and wherein the method further comprises: determining by one or more processors whether the implementation of the medical treatment plan has cured the disease condition of the current patient within a predefined time period; and in response to determining that the implementation of the medical treatment plan has not cured the disease condition of the current patient within a predefined time period, by one or more processors selecting a new set of variance-related patient functions for the current patient to generate a new current optimal patient set of variance-related patient functions for the current patient. The method of claim 1, further comprising: identifying a trend in the temporal heteroscedastic functions by one or more processors, wherein a positive trend indicates a temporal increase in variances for the temporal heteroscedastic functions, whereas a negative trend indicates a temporal decrease in variances for the temporal heteroscedastic functions, and wherein the positive trend and the negative trend describe changes in the amplitude of the variances for the temporal heteroscedastic functions over time; and, in response to the detection of a positive trend in the temporal heteroscedastic functions, issuing an alert by one or more processors with respect to the predefined health-related outcome for the current patient. The method of claim 1, wherein the abstracted set of variance-related candidate-patient functions is generated by one or more processors by maximizing a variance trend over time (VAROT), wherein: VAROT = f ( x , ts , wl , dt , pt , s ) where x = measured values ​​of a predefined measured patient characteristic, ts = a starting point of an observation window for observing the predefined measured patient characteristic, wl = length of the observation window; dt = an incremental time period for a subunit of the observation window; pt = time period type for the observation window, where the time period type is derived from a population is selected that consists of a discrete period and a rolling period, and s = a sparsity condition that defines a required minimum number of data points for x in the incremental period within the observation window. Method according to claim 6, wherein the starting point of the observation window is triggered by a predefined event that is related to the current patient. Method according to claim 7, wherein the predefined event relating to the current patient is the start of a pharmacological protocol applied to the current patient. Method according to claim 7, wherein the predefined event relating to the current patient is an operation performed on the current patient. Method according to claim 7, wherein the predefined event related to the current patient is a nutritional event occurring for the current patient. A computer program product for automatically abstracting and selecting an optimal set of variance-related functions that are an indicator of an individual outcome and personalized plan selection in healthcare, wherein the computer program product comprises a computer-readable storage medium containing computer-readable program code, the program code being readable and executable by a processor to perform a procedure comprising: generating an abstracted set of variance-related candidate-patient functions, wherein the abstracted set of variance-related candidate-patient functions are temporally heteroscedastic functions;Optimizing each patient function from the abstracted set of candidate patient functions based on variance by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized, wherein this optimization creates an optimal abstracted set of patient functions based on variance from the time period in which the variances and heteroscedasticity of each patient function are maximized; Comparing the optimal abstracted set of patient functions based on variance with a historical dataset for a patient population to create a predictive set of patient functions based on variance, wherein the predictive set of patient functions based on variance predicts a health-related target outcome of the patient population; Generating a current optimal patient set of patient functions based on variance for a current patient;Comparing the optimal set of variance-related patient functions for the patient population with the current optimal patient set of variance-related patient functions for the current patient; in response to the optimal set of variance-related patient functions for the patient population matching the current optimal patient set of variance-related patient functions for the current patient within a predefined limit, determining whether the health-related target outcome matches a predefined health-related target outcome for the current patient; and in response to the health-related target outcome matching the predefined health-related outcome for the current patient, issuing an alert regarding the predefined health-related outcome for the current patient. Computer program product according to claim 11, wherein the time period in which the variances and heteroscedasticity of each patient function are maximized is identified by: generating a plurality of time-segment quantities; generating a plurality of subordinate time-segment quantities; creating multiple permutations of combinations of the plurality of time-segment quantities with the plurality of subordinate time-segment quantities; and identifying an optimal combination of a given time-segment quantity with a given subordinate time-segment quantity in which the variances and heteroscedasticity of each patient function are maximized. Computer program product according to claim 11, wherein the method further comprises: creating, based on historical data for the current patient, a normal variance in the current optimal patient set of variance-related patient functions for the current patient, wherein the normal variance has not been predefined as a prediction of a disease state for the current patient; determining whether the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance; and, in response to determining that the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance, issuing the alert with respect to the predefined health-related outcome for the current patient. Computer program product according to claim 11, wherein the predefined health-related outcome for the current patient is an implementation of a medical treatment plan to cure a disease condition from which the current patient is suffering, and wherein the method further comprises: determining whether the implementation of the medical treatment plan has cured the disease condition of the current patient within a predefined time period; and in response to the determination that the implementation of the medical treatment plan has not cured the disease condition of the current patient within a predefined time period, selecting a new set of variance-related patient functions for the current patient to generate a new current optimal patient set of variance-related patient functions for the current patient. Computer program product according to claim 11, wherein the abstracted set of variance-related candidate-patient functions is generated by one or more processors by maximizing a variance trend over time (VAROT), wherein: VAROT = f ( x , ts , wl , dt , pt , s ) where x = measured values ​​of a predefined measured patient characteristic, ts = a starting point of an observation window for observing the predefined measured patient characteristic, wl = length of the observation window; dt = an incremental time period for a subunit of the observation window; pt = period type for the observation window, where the period type is selected from a population consisting of a discrete period and a rolling period, and s = a sparsity condition that defines a required minimum number of data points for x in the incremental period within the observation window. Computer system comprising: a processor, a computer-readable memory, and a computer-readable storage medium on which executable program instructions are stored for: generating an abstracted set of variance-related candidate-patient functions, wherein the abstracted set of variance-related candidate-patient functions are temporally heteroscedastic functions; optimizing each patient function from the abstracted set of variance-related candidate-patient functions by identifying a time period in which the variances and heteroscedasticity of each patient function are maximized, wherein this optimization produces an optimal abstracted set of variance-related patient functions from the time period in which the variances and heteroscedasticity of each patient function are maximized;Comparing the optimal abstracted set of variance-related patient functions with a historical dataset for a patient population to create a predictive set of variance-related patient functions, where the predictive set of variance-related patient functions predicts a health-related target outcome of the patient population; Generating a current optimal patient set of variance-related patient functions for a current patient; Comparing the optimal set of variance-related patient functions for the patient population with the current optimal patient set of variance-related patient functions for the current patient;In response to the optimal set of variance-related patient functions for the patient population, which matches the optimal patient set of variance-related patient functions for the current patient within a predefined limit, determine whether the health-related target outcome matches a predefined health-related target outcome for the current patient; and in response to the health-related target outcome matching the predefined health-related outcome for the current patient, issue an alert regarding the predefined health-related outcome for the current patient. Computer system according to claim 16, further comprising: program instructions for identifying the time period in which the variances and heteroscedasticity of each patient function are maximized by: generating a plurality of time-segment quantities; generating a plurality of subordinate time-segment quantities; creating multiple permutations of combinations of the plurality of time-segment quantities with the plurality of subordinate time-segment quantities; and identifying an optimal combination of a given time-segment quantity with a given subordinate time-segment quantity in which the variances and heteroscedasticity of each patient function are maximized. Computer system according to claim 16, further comprising: program instructions for generating a normal variance in the current optimal patient set of variance-related patient functions for the current patient based on historical data for the current patient, wherein the normal variance has not been predefined as a prediction of a disease state for the current patient; program instructions for determining whether the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance; and program instructions for issuing the alert with respect to the predefined health-related outcome for the current patient in response to determining that the current optimal patient set of variance-related patient functions for the current patient exceeds the normal variance. Computer system according to claim 16, wherein the predefined health-related outcome for the current patient is an implementation of a medical treatment plan to cure a disease condition from which the current patient is suffering, and wherein the computer system further comprises: program instructions to determine whether the implementation of the medical treatment plan has cured the disease condition of the current patient within a predefined time period; and program instructions to select a new set of variance-related patient functions for the current patient in response to the determination that the implementation of the medical treatment plan has not cured the disease condition of the current patient within a predefined time period, in order to generate a new current optimal patient set of variance-related patient functions for the current patient. Computer system according to claim 16, further comprising: program instructions for generating the abstracted set of variance-related candidate-patient functions by maximizing a variance trend over time (VAROT), wherein: VAROT = f ( x , ts , wl , dt , pt , s ) where x = measured values ​​of a predefined measured patient characteristic, ts = a starting point of an observation window for observing the predefined measured patient characteristic, wl = length of the observation window; dt = an incremental time period for a subunit of the observation window; pt = period type for the observation window, where the period type is selected from a population consisting of a discrete period and a rolling period, and s = a sparsity condition that defines a required minimum number of data points for x in the incremental period within the observation window.

Images