Various embodiments described in the present disclosure relate to systems, devices, controllers and methods incorporating statistical classifiers for predicting a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition.
Over the past decade, individual care providers and health care organizations have recognized that an untreated patient unstable/deteriorating that takes place in low acuity wards is a rising problem. These patients typically go unnoticed for a variety of reasons, including low nurse to patient staffing ratios and inexperience among care providers. To address this problem, early warning score (EWS) guides have been developed that are capable of tracking subtle changes in vital signs that might otherwise go unnoticed. EWS guides (e.g., modified EWS guides and national EWS guides) have shown some efficacy in practice and have become the standard of care in some countries (e.g., the United Kingdom).
Unfortunately, adoption of EWS guides has been limited due to the high false rate alarms and overall low sensitivity. These limitations are a result in the heterogeneity among patients and current studies suggest risk assessment should be tailored to specific patient groups (e.g., respiratory illness, cardiovascular illness, septic, etc.) to be more effective.
According to the foregoing, an object of the various embodiments described in the present disclosure is to compute general independent vital sign risk scores from one or more vital signs and/or to compute a general independent vital risk scores based the general independent vital sign risk scores and one or more patient features.
For purposes of describing and claiming the present disclosure:
(1) the terms of the art of the present disclosure including, but not limited to, “vital sign”, “patient feature”, “artificial intelligence”, “statistical classifier” and “risk score” are to be broadly interpreted as known and appreciated in the art of the present disclosure and exemplary described in the present disclosure;
(2) more particularly, the term “vital sign” broadly encompass a sign as understood in the art prior to and subsequent to the present disclosure that either indicates the status of a body's vital life-sustaining functions or has been adopted in medical practice to assess the well-being of a patient. Examples of known vital signs include, but are not limited to, a heart rate, a systolic blood pressure, a respiration rate, a blood oxygen saturation (SPO2), temperature and laboratory/scientific/experimental values/measures/quantifications assessing the well-being of a patient;
(3) more particularly, the term “patient feature” broadly encompass important aspect(s) of a patient medical history and current clinical assessment. Examples of a patient feature include, but are not limited to, clinical diagnosis(ses) of disease(s) or condition(s), result(s) of laboratory test(s) and medication prescription(s);
(4) more particularly, the term “statistical classifier” broadly encompasses a machine learning model, as known in the art of the present disclosure or hereinafter conceived, that is trained in accordance with the present disclosure for predicting which category among a set of categories a new observation belongs. Examples of a statistical classifier include, but are not limited to, a Naive Bayes classifier, a logistic regression classifier, a random forest classifier and a gradient boosting classifier;
(5) more particularly, the term “risk score” broadly encompasses a score rendered by a statistical classifier that is representative of a level of risk that a new observation belongs to a category among a set of categories;
(6) the term “vital sign risk score” broadly encompasses a risk score rendered by a statistical classifier that is representative of a level of risk that a new observation of a particular vital sign belongs to a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition; and
(7) the term “independent vital signal risk score” broadly encompasses a risk score for a particular vital sign rendered by a statistical classifier independent of observation(s) of other vital sign(s).
One embodiment of the present disclosure is a patient risk prediction controller employing a memory storing an artificial intelligence engine including a general statistical classifier and a personal statistical classifier. The general statistical classifier is trained on one or more vital signs to render general independent vital sign score(s), and the personal statistical classifier is trained on one or more patient features to render personal independent vital signa score(s).
The patient risk prediction controller further employs one or more processors. In operation for a singular vital sign, the processor(s) apply a trained general statistical classifier to the singular vital sign to render a singular general independent vital sign risk score. Thereafter, for a singular patient feature, the processor(s) apply a trained personal statistical classifier to the singular general independent vital sign risk score and the singular patient feature to derive a singular personal independent vital sign risk score from an integration of the singular patient feature into the singular general independent vital sign risk score. For plural patient features, the processor(s) apply a trained personal statistical classifier to the singular general independent vital sign risk score and the plural patient features to derive plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into the singular general independent vital sign risk score.
Alternatively in operation for plural vital signs, the processor(s) apply a trained general statistical classifier to the plural vital signs to render plural general independent vital sign risk scores. Thereafter, for a singular patient feature, the processor(s) apply a trained personal statistical classifier to the plural general independent vital sign risk scores and the singular patient feature to derive plural personal independent vital sign risk scores from an individual integration of the singular patient feature into each general independent vital sign risk score of the plural general independent vital sign risk scores. For plural patient features, the processor(s) apply a trained personal statistical classifier to the plural general independent vital sign risk scores and the plural patient features to derive the plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into each general independent vital sign risk score of the plural general independent vital sign risk scores.
A second embodiment of the present disclosure is a non-transitory machine-readable storage medium encoded with instructions for execution by one or more processors of an artificial intelligence engine including a general statistical classifier and a personal statistical classifier. Again, the general statistical classifier is trained on one or more vital signs to render general independent vital sign score(s), and the personal statistical classifier is trained on one or more patient features to render personal independent vital signa score(s).
For a singular vital sign, the encoded medium includes instructions for applying a trained general statistical classifier to the singular vital sign to render a singular general independent vital sign risk score. Thereafter, for a singular patient feature, the encoded medium further includes instructions for applying a trained personal statistical classifier to the singular general independent vital sign risk score and the singular patient feature to derive a singular personal independent vital sign risk score from an integration of the singular patient feature into the singular general independent vital sign risk score. For plural patient features, the encoded medium further includes instructions for applying a trained personal statistical classifier to the singular general independent vital sign risk score and the plural patient features to derive plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into the singular general independent vital sign risk score.
Alternatively for plural vital signs, the encoded medium includes instructions for applying a trained general statistical classifier to plural vital signs to render plural general independent vital sign risk scores. Thereafter, for a singular patient feature, the encoded medium further includes instructions for applying a trained personal statistical classifier to the plural general independent vital sign risk scores and the singular patient feature to derive plural personal independent vital sign risk scores from an individual integration of the singular patient feature into each general independent vital sign risk score of the plural general independent vital sign risk scores. For plural patient features, the encoded medium further includes instructions for applying a trained personal statistical classifier to the plural general independent vital sign risk scores and the plural patient features to derive the plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into each general independent vital sign risk score of the plural general independent vital sign risk scores.
A third embodiment the present disclosure is a patient risk prediction method executable by an artificial intelligence engine including a general statistical classifier and a personal statistical classifier. Again, the general statistical classifier is trained on one or more vital signs to render general independent vital sign score(s), and the personal statistical classifier is trained on one or more patient features to render personal independent vital signa score(s).
For a singular vital sign, the patient risk prediction method involves an application of a trained general statistical classifier to the singular vital sign to render a singular general independent vital sign risk score. Thereafter, for a singular patient feature, the patient risk prediction method further involves an application of a trained personal statistical classifier to the singular general independent vital sign risk score and the singular patient feature to derive a singular personal independent vital sign risk score from an integration of the singular patient feature into the singular general independent vital sign risk score. For plural patient features, the patient risk prediction method further involves an application of a trained personal statistical classifier to the singular general independent vital sign risk score and the plural patient features to derive plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into the singular general independent vital sign risk score.
Alternatively for plural vital signs, the patient risk prediction method involves an application of a trained general statistical classifier to plural vital signs to render plural general independent vital sign risk scores. Thereafter, for a singular patient feature, the patient risk prediction method involves an application of a trained personal statistical classifier to the plural general independent vital sign risk scores and the singular patient feature to derive plural personal independent vital sign risk scores from an individual integration of the singular patient feature into each general independent vital sign risk score of the plural general independent vital sign risk scores. For plural patient features, the patient risk prediction method involves an application of a trained personal statistical classifier to the plural general independent vital sign risk scores and the plural patient features to derive the plural personal independent vital sign risk scores from an individual integration of each patient feature of the plural patient features into each general independent vital sign risk score of the plural general independent vital sign risk scores.
Also for purposes of describing and claiming the present disclosure:
(1) the term “controller” broadly encompasses all structural configurations, as understood in the art of the present disclosure and hereinafter conceived, of a main circuit board or an integrated circuit for controlling an application of various principles of the present disclosure as subsequently described in the present disclosure. The structural configuration of the controller may include, but is not limited to, processor(s), non-transitory machine-readable storage medium(s), an operating system, application module(s), peripheral device controller(s), slot(s) and port(s); and
(2) the terms “data” and “signals” may be embodied in all forms of a detectable physical quantity or impulse (e.g., voltage, current, magnetic field strength, impedance, color) as understood in the art of the present disclosure and as exemplary described in the present disclosure for transmitting information and/or instructions in support of applying various principles of the present disclosure as subsequently described in the present disclosure. Data/signal communication encompassed by the present disclosure may involve any communication method as known in the art of the present disclosure including, but not limited to, data/signal transmission/reception over any type of wired or wireless communication link and a reading of data uploaded to a computer-usable/computer readable storage medium.
The foregoing embodiments and other embodiments of the present disclosure as well as various features and advantages of the present disclosure will become further apparent from the following detailed description of various embodiments of the present disclosure read in conjunction with the accompanying drawings. The detailed description and drawings are merely illustrative of the present disclosure rather than limiting, the scope of the inventions of present disclosure being defined by the appended claims and equivalents thereof.
In order to better understand various example embodiments, reference is made to the accompanying drawings, wherein:
To facilitate an understanding of the present disclosure, the following description of
Referring to
For purposes of the present disclosure, vital signs 12 broadly encompass a signs that indicate the status of a body's vital life-sustaining functions. Examples of vital signs 12 include, but are not limited to, a heart rate, a systolic blood pressure, a respiration rate, a blood oxygen saturation (SPO2) and temperature.
For purposes of the present disclosure, patient features 23 broadly encompass important aspects of a patient medical history and current clinical assessment. Examples of patient features 23 include, but are not limited to, clinical diagnosis(ses) of disease(s) or condition(s), result(s) of laboratory test(s) and medication prescription(s). In practice, a singular patient feature 23 may consist of singular important aspect of the patient medical history and current clinical assessment (e.g., a singular clinical diagnosis of a disease, or a result of a singular laboratory test or a singular medication prescription), or may consist of an accumulation of plural important aspects of the patient medical history and current clinical assessment (e.g., plural clinical diagnoses of a disease or results of plural laboratory tests or plural medication prescriptions, or any combination of clinical diagnosis(ses), lab result(s) and medication prescription(s)).
Still referring to
In a first set of embodiments, general statistical classifier 40 is constructed in accordance with the principles of the present disclosure to compute a general independent vital sign risk score (GVRS) 43 for a singular vital sign 12 (e.g., heart rate), and is trained in accordance with the principles of the present disclosure on a general patient population of the singular vital sign 12 whereby the general independent vital sign risk score 43 quantifies a probability of classifying the singular vital sign 12 as a stable/non-deteriorating patient condition (i.e., a patient condition deemed in practice harmless to a patient's health) or an unstable/deteriorating patient condition (i.e., a patient condition deemed in practice as potentially hazardous/dangerous to a patient's health). The training associated with the stable/non-deteriorating patient condition may be directed to patients recovering or recovered from a health emergency (e.g., a heart attack or a stroke) and/or a surgery (e.g., heart transplant or a coronary bypass), and the training associated with the unstable/deteriorating patient condition may be directed to deceased patients, patients transferred to a higher acuity and/or patients which required a call for a rapid response team.
For the first set of embodiments, general statistical classifier 40 may be further constructed in accordance with the principles of the present disclosure to derive the general patient risk score 44 from the singular general independent vital sign risk score 43 in any manner suitable for an informative reporting of the general patient risk score 44 quantifying a general stable/non-deteriorating patient condition or a general unstable/deteriorating patient condition. For example, the general patient risk score 44 may be equal to the singular general independent vital sign risk score 43 or a normalization of the singular general independent vital sign risk score 43.
In a second set of embodiments, general statistical classifier 40 is constructed in accordance with the principles of the present disclosure to separately compute a general independent vital sign risk score 43 for each vital sign 12 among plural vital signs 12 (e.g., heart rate, systolic blood pressure, respiration rate, blood oxygen saturation (SPO2) and temperature), and is trained in accordance with the principles of the present disclosure on a general patient population of the plural vital signs 12 whereby each general independent vital sign risk score 43 independently quantifies a probability of classifying a corresponding vital sign 12 as a stable/non-deteriorating patient condition (i.e., a patient condition deemed in practice as harmless to a patient's health) or an unstable/deteriorating patient condition (i.e., a patient condition deemed in practice as potentially hazardous/dangerous to a patient's health). Again, the training associated with the stable/non-deteriorating patient condition may be directed to patients recovering or recovered from a health emergency (e.g., a heart attack or a stroke) and/or a surgery (e.g., heart transplant or a coronary bypass), and the training associated with the unstable/deteriorating patient condition may be directed to deceased patients, patients transferred to a higher acuity and/or patients which required a call for a rapid response team.
For the second set of embodiments, general statistical classifier 40 may be further constructed in accordance with the principles of the present disclosure to derive the general patient risk score 44 from the plural general independent vital sign risk scores 43 in any manner suitable for an informative reporting of the general patient risk score 44 quantifying a general stable/non-deteriorating patient condition or a general unstable/deteriorating patient condition. For example, the general patient risk score 44 may be an aggregation of the plural general independent vital sign risk scores 43 in the form of a summation of the plural general independent vital sign risk scores 43, or a normalization of a summation of the plural general independent vital sign risk scores 43.
Still referring to
In a first set of embodiments, personal statistical classifier 40 is constructed in accordance with the principles of the present disclosure to derive a singular personal independent vital sign risk score (not shown in
For purposes of the present disclosure, a weighted function of the singular patient feature 23 broadly encompasses a quantification of the singular patient feature 23 that further personally refines the singular general independent vital sign risk score 43 as a probability of classifying the singular vital sign 12 as a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition. In practice, the weighted function may be simple (e.g., a binary number indicating an absence or a presence of a particular clinical diagnosis, a particular lab result or a particular medication prescription) or complex (e.g., a multivariate expression of various categories of a clinical diagnosis, numerous test ranges of lab results and various dosages of a medication prescription). For example, the weighted function of the singular patient feature 23 may be a product of simple or complex coefficient(s) and the singular patient feature 23.
For the first set of embodiments, the personal statistical classifier 50 is further constructed in accordance with the principles of the present disclosure to derive the personal patient risk score 54 from the singular personal independent vital sign risk score in any manner suitable for an informative reporting of the personal patient risk score 54 quantifying a personal stable/non-deteriorating patient condition or a personal unstable/deteriorating patient condition. For example, the personal patient risk score 54 may be equal to the singular personal independent vital sign risk score or a normalization of the singular personal independent vital sign risk score.
In a second set of embodiments, personal statistical classifier 40 is constructed in accordance with the principles of the present disclosure to derive plural personal independent vital sign risk scores (not shown in
For purposes of the present disclosure, a weighted function of each patient feature 23 broadly encompasses a quantification of each patient feature 23 that further personally refines the singular general independent vital sign risk score 43 as a probability of classifying the singular vital sign 12 as a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition. Again, in practice, the weighted function may be simple (e.g., a binary number indicating an absence or a presence of a particular clinical diagnosis, a particular lab result or a particular medication prescription) or complex (e.g., a multivariate expression of various categories of a clinical diagnosis, numerous test ranges of lab results and various dosages of a medication prescription). For example, the weighted function of each of the plural patient features 23 may be a product of simple or complex coefficient(s) and a corresponding patient feature 23.
For the second set of embodiments, the personal statistical classifier 50 is further constructed in accordance with the principles of the present disclosure to derive the personal patient risk score 54 from the plural personal independent vital sign risk scores in any manner suitable for an informative reporting of the personal patient risk score 54 quantifying a personal stable/non-deteriorating patient condition or a personal unstable/deteriorating patient condition. For example, the personal patient risk score 54 may be an aggregation of the plural personal independent vital sign risk scores in the form of a summation/a product of the plural personal independent vital sign risk scores or a normalization of a summation/a product of the plural personal independent vital sign risk scores.
In a third set of embodiments, personal statistical classifier 40 is constructed in accordance with the principles of the present disclosure to derive plural personal independent vital sign risk scores (not shown in
For purposes of the present disclosure, a weighted function of the singular patient feature 23 broadly encompasses a quantification of the singular patient feature 23 that further personally refines each general independent vital sign risk score 43 as a probability of classifying each vital sign 12 as a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition. Again, in practice, the weighted function may be simple (e.g., a binary number indicating an absence or a presence of a particular diagnosis, a particular lab result or a particular medication) or complex (e.g., a multivariate expression of various categories of a diagnosis, numerous test ranges of lab results and a number of a particular type of medication). For example, the weighted function of the singular patient feature 23 may be a product of simple or complex coefficient(s) and the singular patient feature 23.
For the third set of embodiments, the personal statistical classifier 50 is further constructed in accordance with the principles of the present disclosure to derive the personal patient risk score 54 from the plural personal independent vital sign risk score in any manner suitable for an informative reporting of the personal patient risk score 54 quantifying a personal stable/non-deteriorating patient condition or a personal unstable/deteriorating patient condition. For example, the personal patient risk score 54 may be an aggregation of the plural personal independent vital sign risk scores in the form of a summation/a product of the plural personal independent vital sign risk scores or a normalization of a summation/a product of the plural personal independent vital sign risk scores.
In a fourth set of embodiments, personal statistical classifier 40 is constructed in accordance with the principles of the present disclosure to derive plural personal independent vital sign risk scores (not shown in
For purposes of the present disclosure, a weighted function of each patient feature 23 broadly encompasses a quantification of each patient feature 23 that further personally refines each general independent vital sign risk score 43 as a probability of classifying each vital sign 12 as a stable/non-deteriorating patient condition or an unstable/deteriorating patient condition. Again, in practice, the weighted function may be simple (e.g., a binary number indicating an absence or a presence of a particular clinical diagnosis, a particular lab result or a particular medication prescription) or complex (e.g., a multivariate expression of various categories of a clinical diagnosis, numerous test ranges of lab results and various dosages of a medication prescription). For example, the weighted function of each of the plural patient features 23 may be a product of simple or complex coefficient(s) and a corresponding patient feature 23.
For the fourth set of embodiments, the personal statistical classifier 50 is further constructed in accordance with the principles of the present disclosure to derive the personal patient risk score 54 from the personal independent vital sign risk scores in any manner suitable for an informative reporting of the personal patient risk score 54 quantifying a personal stable/non-deteriorating patient condition or a personal unstable/deteriorating patient condition. For example, the personal patient risk score 54 may be an aggregation of the plural personal independent vital sign risk scores in the form of a summation/product of the plural personal independent vital sign risk scores or a normalization of a summation/a product of the plural personal independent vital sign risk scores.
Referring to
During a stage S72 of flowchart 70, artificial intelligence engine 30 receives a singular vital sign 12 from a vital sign source 10, or plural vital signs 12 from one or more vital sign sources 10. In practice, vital sign sources 10 may be any type of source capable of sensing, detecting or otherwise monitoring a vital sign of a patient 11. Examples of vital sign sources 10 include, but are not limited to, heart rate sensors, electrocardiograms, blood pressure sensors, respiratory rate sensors, pulse oximeters and thermometers. In practice, the vital sign(s) 12 may be communicated by techniques known in the art prior to and subsequent the present disclosure at any time suitable for ascertaining a condition of patient 11 (e.g., in real-time or post-study).
Upon receipt of a singular vital sign 12, general statistical classifier 40 executes a general vital sign risk scoring 41 of the singular vital sign 12 to render a singular general independent vital sign risk score 43 as previously described in the present disclosure. Subsequently, general statistical classifier 40 executes a general patient risk scoring 42 to compute general patient risk score 44 as previously described in the present disclosure. For example, during scoring 41, general statistical classifier 40 may implement a Naive Bayes classification, a logistic regression classification, a random forest classification or a gradient boosting classification of the singular vital sign 12 to render the singular general independent vital sign risk score 43. Subsequently, during scoring 42, general statistical classifier 40 may compute general patient risk score 44 as the singular general independent vital sign risk score 43.
Upon receipt of plural vital signs 12, general statistical classifier 40 executes general vital sign risk scoring 41 of the plural vital signs 12 to render plural general independent vital sign risk scores 43 as previously described in the present disclosure. Subsequently, general statistical classifier 40 executes a general patient risk scoring 42 to compute general patient risk score 44 as previously described in the present disclosure. For example, during scoring 41, general statistical classifier 40 may individually implement a Naive Bayes classification, a logistic regression classification, a random forest classification or a gradient boosting classification of each of the plural vital signs 12 to render the plural general independent vital sign risk scores 43. Subsequently, during scoring 42, general statistical classifier 40 may implement a summation/a product of the plural general independent vital sign risk scores 43 to compute general patient risk score 44.
Still referring to
Upon receipt of a singular patient feature 23, personal statistical classifier 50 executes a personal vital sign risk scoring 51 of the singular patient feature 23 and a general independent vital sign risk score 43 or plural general independent vital sign risk scores 43, whichever is applicable, to render a singular personal independent vital sign risk score 53 or plural personal independent vital sign risk scores 53, respectively, as previously described in the present disclosure. Subsequently personal statistical classifier 50 executes a personal patient risk scoring 52 to compute personal patient risk score 54 as previously described in the present disclosure. For example, during scoring 51, personal statistical classifier 50 may implement a weighted function of singular patient feature 23 and a compute a product of the weighted function of the single patient feature 23 and the singular general independent vital sign risk score 43 or of the plural general independent vital sign risk score(s) 43, whichever is applicable, to render the singular personal independent vital sign risk score 53 or the plural personal independent vital sign risk scores 53. Subsequently, during scoring 52, personal statistical classifier 50 may equate the singular personal independent vital sign risk score 53 as personal patient risk score 54 or may implement a summation of the plural personal independent vital sign risk scores 53, whichever is applicable.
Upon receipt of plural patient features 23, personal statistical classifier 50 executes a personal vital sign risk scoring 51 of the plural patient features 23 and a general independent vital sign risk score 43 or plural general independent vital sign risk scores 43, whichever is applicable, to render the plural personal independent vital sign risk scores 53 as previously described in the present disclosure. Subsequently personal statistical classifier 50 executes a personal patient risk scoring 52 to compute the personal patient risk score 54 as previously described in the present disclosure. For example, during scoring 51, personal statistical classifier 50 may generate a weighted function of each of the plural patient features 23 and a compute an individual product of the weighted function of each of the plural patient feature 23 and the singular general independent vital sign risk score 43 or of the plural general independent vital sign risk score(s) 43, whichever is applicable, to render the plural personal independent vital sign risk scores 53. Subsequently, during scoring 52, personal statistical classifier 50 may implement a summation of the plural personal independent vital sign risk scores 53 to compute the personal patient risk score 54.
Still referring to
In a second embodiment, artificial intelligence engine 30 executes a patient risk reporting 78 as shown in
To facilitate a further understanding of the present disclosure, the following description of
For clarity purposes, the following description of
Referring to
In operation, statistical classifier 141a is constructed and trained to input heart rate (HR) signal 112a to thereby render a general heart rate risk score (GHRRS) 143a.
Statistical classifier 141b is constructed and trained to input a blood pressure (BP) signal 112b to thereby render a general blood pressure risk score (GBPRS) 143b.
Statistical classifier 141c is constructed and trained to input a respiratory rate (RR) signal 112c to thereby render a general respiratory rate risk score (GRRRS) 143c.
Statistical classifier 141d is constructed and trained to input a blood oxygen saturation (SPO2) signal 112d to thereby render a general blood oxygen saturation risk score (GSPRS) 143d.
Statistical classifier 141e is constructed and trained to input a temperature (TEMP) signal 112e to thereby render a general temperature risk score (GTPRS) 143e.
In practice, statistical classifiers 141a-141e may implement a statistical classifier as known in the art prior to and subsequent to the present disclosure that is constructed and trained in accordance with the principles of the present disclosure to render the plural general independent vital sign risk scores 143a-143e respectively for plural vitals sign 112a-112e. Examples of various embodiments of statistical classifiers 141a-141e include, but are not limited to, a parallel network of Naive Bayes classifier, a parallel network of logistic regression classifiers, a parallel network of random forest classifier and a parallel network of gradient boosting classifiers.
For clarity purposes, the following description of the parallel network of statistical classifiers 141a-141e will be as a parallel network of Naive Bayes classifiers. Nonetheless, from the description of the parallel network of Naive Bayes classifiers, those having ordinary skill in the art of the present disclosure will appreciate how to apply description the present disclosure for making and using numerous and various additional embodiments of the parallel network of statistical classifiers 141a-141e including, but not limited to, a parallel network of logistic regression classifiers trained in accordance with the present disclosure via logistic/sigmoid function(s) as known in the art prior to and subsequent to the present disclosure, a parallel network of random forest classifiers trained in accordance with the present disclosure via decision trees as known in the art prior to and subsequent to the present disclosure and a parallel network of gradient boosting classifiers trained in accordance with the present disclosure on prediction models as known in the art prior to and subsequent to the present disclosure.
In one embodiment of general statistical classifier 140 as shown in FIG. 3B, each statistical classifier 141a-141e generates a training histogram 145 of continuous values having a density axis 145a, a vital sign axis 145b and a risk curve axis 145c. A mixture model (e.g., a gaussian model, a log-normal mixture, an exponential mixture, an alpha mixture and a beta mixture) is used to fit a stable/non-deteriorating class C0, distribution 146 of stable/non-deteriorating training values of an assigned vital sign and an unstable/deteriorating class C1 distribution 147 of unstable/deteriorating training values of the assigned vital sign plotted within the training histogram 145. A vital sign risk curve 148 is calculated in accordance with the following log odds ratio equation [1], the following normalized probability equation [2] of the following normalized probability equation [3]:
For equations [1]-[3], P(Xi|C0) is the probability of observing the assigned vital sign for the stable/non-deteriorating class C0, P(Xi|C1) is the probability of observing the assigned vital sign for the unstable/deteriorating class C1, and P(Xi) is the probability of observing the assigned vital sign.
In a second embodiment of general statistical classifier 140 as shown in
In practice, risk score adder 142 is any type of adder as known in the art prior to and subsequent to the present disclosure that is constructed accordance with the principles of the present disclosure to compute a general patient risk score 144 as a summation of the plural general independent vital sign risk scores 143a-143e.
Referring to
During a stage S172 of flowchart 170, statistical classifiers 141a-141e independently renders general independent vital sign risk scores 143a-143e respectively for vital sign 112a-112e.
For a log odds ratio embodiment 173a, statistical classifiers 141a-141e independently renders the plural general independent vital sign risk scores 143a-143e respectively for vital sign 112a-112e in accordance with the following equations [4]-[8]:
For a normalized probability embodiment 173b, statistical classifiers 141a-141e independently renders general independent vital sign risk scores 143a-143e respectively for vital sign 112a-112e in accordance with the following equations [9]-[13] for either the stable/non-deteriorating class C0 or the unstable/deteriorating class C1:
During a stage S174 of flowchart 170, risk score adder 142 compute general patient risk score 144 as a summation of general independent vital sign risk scores 143a-143e.
For a log odds ratio embodiment 175a, risk score adder 142 computes general patient risk score 144 as summation of the plural general independent vital sign risk scores 143a-143e in accordance with the following equation [14a] or the following equation [14b]:
For equation 14[b], log (P(C1)/P(C0)) represents a term for biasing the GRS by the overall prevalence of unstable/deteriorating class C1.
For a normalized probability embodiment 175b, risk score adder 142 computes general patient risk score 144 as a logarithmic summation of general independent vital sign risk scores 143a-143e in accordance with the following equation [15] for either the stable/non-deteriorating class C0 or the unstable/deteriorating class C1:
For equation [15b], log (P(C1)/P(C0)) again represents a term for biasing the GRS by the overall prevalence of unstable/deteriorating class C1.
Referring to
In practice, weighted function multiplier 151a is constructed and trained to input general heart rate risk score (GHRRS) 143a and plural weighted functions 156 to compute a personal heart rate risk score (PHRRS) 153a for each weighted function 156.
Weighted function multiplier 151b is constructed and trained to input general blood pressure risk score (GBPRS) 143b and the plural weighted functions 156 to thereby render a personal blood pressure risk score (PBPRS) 153b for each weighted function 156.
Weighted function multiplier 151c is constructed and trained to input general respiratory rate risk score (GRRRS) 143c and the plural weighted functions 156 to thereby render a personal respiratory rate risk score (PRRRS) 153c for each weighted function 156.
Weighted function multiplier 151d is constructed and trained to input general blood oxygen saturation risk score (GSPRS) 143d and the plural weighted functions 156 to thereby render a personal blood oxygen saturation risk score (PSPRS) 153d for each weighted function 156.
Weighted function multiplier 151e is constructed and trained to input general temperature risk score (GTPRS) 143e and the plural weighted functions 156 to thereby render a personal temperature risk score (PTPRS) 153e for each weighted function 156.
In practice, each weighted function multiplier 151 is any type of multiplier as known in the art prior to and subsequent to the present disclosure that is constructed in accordance with the principles of the present disclosure to compute personal independent vital sign risk scores 153a-153e as a product of a corresponding general independent vital sign risk scores 143a-143a and each of the plural weighted functions 156.
In practice, risk score adder 152 is any type of adder as known in the art prior to and subsequent to the present disclosure that is constructed accordance with the principles of the present disclosure to compute a personal patient risk score 154 as a logarithmic summation of personal heart rate risk scores 153a-153e.
In practice, weight matrix generator 155 is any type of arithmetic logic unit as known in the art prior to and subsequent to the present disclosure that is constructed in accordance with the principles of the present disclosure to generate a weighting function of diagnosis patient feature 123a informative of a cardiac clinical diagnosis, a lab results patient feature 123b informative of results of a cardiac laboratory test, and a medication patient feature 123c information of a prescribed cardiac medication.
In practice, weight matrix generator 155 encodes a patient feature and applies the encoded patient feature to a weighted coefficient that is priori determined through logistic regression with regularization during the training of personal statistical classifier 150 (
Further in practice, weighted coefficient(s) for a particular patient feature 123 may be determined for all of the vital signs 141a-141e (
In one embodiment, weight matrix generator 155 implements a binary encoding or one-hot encoding of categorical variable(s) or continuous variable(s) for each patient feature 123. For example, for diagnosis patient feature 123a, a binary encoding may be “0” for an absence of a categorical variable of a diagnosed cardiac disease and may be “1” for a presence of a categorical variable of a diagnosed cardiac disease. By further example, for lab results patient feature 123b, a one-hot encoding may be used for multiple continuous variables of results of a cardiac laboratory tests. By even further example, for medication patient feature 123c, a binary encoding may be “0” for a no-use categorical variable of a prescribed cardiac medication and may be “1” for a use categorical variable of a prescribed cardiac medication.
Referring to
Referring to
In a universal weighting embodiment 273a, weighted matrix generator 155 generates a weighting coefficient ViDiagnosis*f(yDiagnosis) from diagnosis patient features 123a for all vital signs, a weighting coefficient ViLab Results*f(yLab Results) from lab results patient features 123b for all vital signs, and a weighting coefficient ViMeds*f(yMeds) from meds patient features 123c for all vital signs.
In a vital sign embodiment 273b, for heart rate 112a (
For blood pressure 112b (
For respiratory rate 112c (
For blood oxygen saturation 112d (
For temperature 112e (
During a stage S274 of flowchart 270, weighted matrix generator 155 communicates the weighted functions Vij*f(yi) to each weighted function multipliers 143a-143e. The communication follows a matrix of weighted functions Vij*f(yj) arranged by columns of vital signs 112a-112e and rows of patient features 123a-123c as shown, or vice-versa.
Referring to
For a log odds ratio embodiment 375a, weighted function multipliers 151a-151e independently computes personal independent vital sign risk scores 153a-153e respectively for general independent vital sign risk scores 143a-143e in accordance with the following equations [16]-[20]:
For a normalized probability embodiment (
During a stage S374 of flowchart 370, risk score adder 152 compute personal patient risk score 154 as a summation of the plural personal independent vital sign risk scores 153a-153e.
For log odds ratio embodiment 375a, risk score adder 152 computes personal patient risk score 154 as a summation of the plural personal independent vital sign risk scores 153a-153e in accordance with the following equation [26]:
PRS=ΣVi,jf(yi)*log(P(Xi|C1)/P(Xi|C0)) [26]
For a normalized probability embodiment 375b, risk score adder 152 computes personal patient risk score 154 as a logarithmic summation of personal independent vital sign risk scores 153a-153e in accordance with the following equation [27] for either the stable/non-deteriorating class C0 or the unstable/deteriorating class C1:
PRS=ΣVi,jf(yj)*log(P(Xi|C1)/P(Xi) [27]
To further facilitate an understanding of the present disclosure, the following description of
In practice, a patient risk prediction controller of the present disclosure may be embodied as hardware/circuitry/software/firmware for implementation of a patient risk prediction method of the present disclosure as previously described herein. Further in practice, a patient risk prediction controller may be customized and installed in a server, workstation, etc. or programmed on a general purpose computer.
In one embodiment as shown in
The processor 81 may be any hardware device capable of executing instructions stored in memory or storage or otherwise processing data. As such, the processor 81 may include a microprocessor, field programmable gate array (FPGA), application-specific integrated circuit (ASIC), or other similar devices.
The memory 82 may include various memories such as, for example L1, L2, or L3 cache or system memory. As such, the memory 82 may include static random access memory (SRAM), dynamic RAM (DRAM), flash memory, read only memory (ROM), or other similar memory devices.
The user interface 83 may include one or more devices for enabling communication with a user such as an administrator. For example, the user interface 83 may include a display, a mouse, and a keyboard for receiving user commands. In some embodiments, the user interface 83 may include a command line interface or graphical user interface that may be presented to a remote terminal via the network interface 84.
The network interface 84 may include one or more devices for enabling communication with other hardware devices. For example, the network interface 84 may include a network interface card (NIC) configured to communicate according to the Ethernet protocol. Additionally, the network interface 84 may implement a TCP/IP stack for communication according to the TCP/IP protocols. Various alternative or additional hardware or configurations for the network interface will be apparent.
The storage 85 may include one or more machine-readable storage media such as read-only memory (ROM), random-access memory (RAM), magnetic disk storage media, optical storage media, flash-memory devices, or similar storage media. In various embodiments, the storage 85 may store instructions for execution by the processor 81 or data upon with the processor 81 may operate. For example, the storage 85 store a base operating system (not shown) for controlling various basic operations of the hardware.
More particular to the present disclosure, storage 85 may store control modules 87 in the form of general statistical classifier 40 (
Referring to
In operation, patient risk prediction controller 80 inputs medical imaging data 30, planar or volumetric, from medical imaging data sources 80 during a training phase and a phase. Medical imaging data sources 90 may include any number and types of medical imaging machines (e.g., a MRI machine 91, a CT machine 93, an X-ray machine 95 and an ultrasound machine 97 as shown) and may further includes database management/file servers (e.g., MRI database management server 92, CT server 94, X-ray database management server 96 and ultrasound database manager server 97 as shown). In practice, application server 90 or workstation 93, whichever is applicable, may be directly or networked connected to a medical imaging data source 90 to thereby input medical imaging data 30 for patient risk prediction controller 80. Alternatively, a medical imaging data source 90 and application server 90 or workstation 93, whichever is applicable, may be directly integrated whereby the patient risk prediction controller 80 has direct access to medical imaging data 30.
Referring to
In practice, display/display interface 103 displays patient monitoring data as customized by a user via display interface 103 (e.g., keys) and patient risk score(s) as generated by patient risk prediction controller 80 as previously described in the present disclosure. Controller interface 15 (e.g., knobs and buttons) allows the user to apply various therapies (e.g., a shock) to a patient. Port interface 17 allows for the connection by the user to vital sign source(s) 10 for receiving vital signs and to patient feature sources (20) for receiving patient features.
Referring to
Furthermore, it will be apparent that various information described as stored in the storage may be additionally or alternatively stored in the memory. In this respect, the memory may also be considered to constitute a “storage device” and the storage may be considered a “memory.” Various other arrangements will be apparent. Further, the memory and storage may both be considered to be “non-transitory machine-readable media.” As used herein, the term “non-transitory” will be understood to exclude transitory signals but to include all forms of storage, including both volatile and non-volatile memories.
While the device is shown as including one of each described component, the various components may be duplicated in various embodiments. For example, the processor may include multiple microprocessors that are configured to independently execute the methods described in the present disclosure or are configured to perform steps or subroutines of the methods described in the present disclosure such that the multiple processors cooperate to achieve the functionality described in the present disclosure. Further, where the device is implemented in a cloud computing system, the various hardware components may belong to separate physical systems. For example, the processor may include a first processor in a first server and a second processor in a second server.
It should be apparent from the foregoing description that various example embodiments of the invention may be implemented in hardware or firmware. Furthermore, various exemplary embodiments may be implemented as instructions stored on a machine-readable storage medium, which may be read and executed by at least one processor to perform the operations described in detail herein. A machine-readable storage medium may include any mechanism for storing information in a form readable by a machine, such as a personal or laptop computer, a server, or other computing device. Thus, a machine-readable storage medium may include read-only memory (ROM), random-access memory (RAM), magnetic disk storage media, optical storage media, flash-memory devices, and similar storage media.
It should be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative circuitry embodying the principles of the invention. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in machine readable media and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.
Although the various exemplary embodiments have been described in detail with particular reference to certain exemplary aspects thereof, it should be understood that the present disclosure is capable of other embodiments and its details are capable of modifications in various obvious respects. As is readily apparent to those skilled in the art, variations and modifications can be affected while remaining within the spirit and scope of the present disclosure. Accordingly, the foregoing disclosure, description, and figures are for illustrative purposes only and do not in any way limit the present disclosure, which is defined only by the claims.
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2019/074600 | 9/16/2019 | WO | 00 |
Number | Date | Country | |
---|---|---|---|
62732623 | Sep 2018 | US |