Method and device for identifying risk of cardiac shock after cardiac incision of patient

Abstract

The present disclosure describes methods and systems for predicting whether a patient is likely to have a post-cardiotomy cardiac shock (PCCS). The method comprises the following steps: receiving medical information of a patient; extracting one or more features from the received medical information; providing the one or more features as input to a trained classification model, the trained classification model configured to output a risk assessment of a potential PCCS of the patient; and outputting an indication of the risk assessment.

Description

Technical Field

[0001] The present disclosure relates to identifying patients at risk for developing post-cardiotomy cardiogenic shock. Background Art

[0002] Postcardiotomy cardiogenic shock (PCCS) is a relatively rare but important cause of death in patients who have undergone cardiac surgery. Although inconsistently defined in the literature, PCCS is associated with circulatory failure after cardiac surgery that requires mechanical circulatory support (e.g., support with an extracorporeal membrane oxygenation (ECMO) device, ventricular assist device (VAD), etc.) and / or administration of high-dose inotropes. Summary of the invention

[0003] In some embodiments, a method for predicting whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS) is provided. The method comprises: receiving medical information of a patient; extracting one or more features from the received medical information; providing the one or more features as input to a trained classification model, the trained classification model being configured to output a risk assessment of the patient being likely to develop PCCS; and outputting an indication of the risk assessment.

[0004] In one aspect, the patient's medical information includes one or more of an electronic health record, a laboratory report, a medical procedure report, a doctor's note, and a medical imaging report. In another aspect, the medical information includes structured data and unstructured data. In another aspect, extracting one or more features includes extracting at least some of the one or more features from the unstructured data using natural language processing. In another aspect, the one or more features include left ventricular ejection fraction and / or total bilirubin level. In another aspect, the method further includes receiving data indicating whether the patient has PCCS, and retraining the trained classification model based at least in part on the received data. In another aspect, the risk assessment includes a numerical value, and outputting an indication of the risk assessment includes displaying the numerical value and / or information based on the numerical value on a user interface. In another aspect, the method further includes performing classification of the patient into a risk group in a plurality of risk groups based on the numerical value, and outputting an indication of the risk assessment includes outputting an indication of the patient's risk group. In another aspect, performing classification of the patient into a risk group includes determining whether the numerical value is above a threshold, and classifying a high risk of PCCS when determining that the numerical value is above the threshold. In another aspect, outputting an indication of the patient's risk group includes displaying a color-coded indication of the risk group on a user interface.

[0005] On the other hand, the risk assessment is to classify the patient into a risk group in a plurality of risk groups, and the indication of outputting the risk assessment includes outputting the indication of the risk group of the patient. On the other hand, the trained classification model includes a trained neural network. On the other hand, the trained classification model includes a trained random forest model. On the other hand, the trained classification model includes a trained multivariate regression model. On the other hand, the method further includes receiving additional medical information, and retraining the trained classification model based on the additional medical information. On the other hand, the additional medical information includes medical information of a plurality of patients who have undergone cardiac surgery at a medical facility. On the other hand, the method further includes: providing a user interface configured to display the value of one or more features; receiving user input via the user interface to change one or more of the values ​​of one or more features; simulating the risk assessment of the patient's possible occurrence of PCCS based at least in part on the change of one or more values ​​to generate a simulated risk assessment; and displaying the simulated risk assessment on the user interface. On the other hand, the indication of outputting the risk assessment includes outputting a cumulative score associated with the risk assessment.

[0006] In some embodiments, a method for training a risk model to predict whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS) is provided. The method includes: receiving a data set of patient medical information; selecting training data from the data set of patient medical information based on PCCS criteria and defined data fields, wherein the training data includes patient medical information of at least two risk groups of the patient; training the risk model using the selected training data; and outputting the trained risk model.

[0007] In one aspect, the method further includes defining a plurality of PCCS criteria and generating at least two risk groups for a patient based on the PCCS criteria. In another aspect, the method further includes receiving input regarding data fields to be defined via a user interface, and defining the data fields based at least in part on the received input. In another aspect, the method further includes validating the trained model using at least some patient medical information that was not used to train the model, and outputting the trained risk model includes outputting a validated trained model. In another aspect, the method further includes receiving an indication to update the trained risk model, and retraining the risk model in response to receiving an indication to update the trained risk model. In another aspect, the risk model includes a neural network. In another aspect, the risk model includes a random forest model. In another aspect, the risk model includes a multivariate regression model. In another aspect, the method further includes receiving additional medical information, and retraining the trained risk model based on the additional medical information. In another aspect, the additional medical information includes medical information of a plurality of patients who underwent cardiac surgery at a medical facility.

[0008] In some embodiments, a computer-implemented system for predicting whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS) is provided. The system includes: at least one hardware computer processor; and at least one non-transitory computer-readable medium encoded with a plurality of instructions that, when processed by the at least one hardware computer processor, perform a method. The method includes: extracting one or more features from the patient's medical information; providing the one or more features as input to a trained classification model, the trained classification model configured to output a risk assessment that the patient is likely to develop PCCS; and outputting an indication of the risk assessment.

[0009] In some embodiments, at least one non-transitory computer-readable medium is provided, encoded with a plurality of instructions that, when processed by at least one hardware computer processor, perform a method. The method includes: extracting one or more features from medical information of a patient; providing the one or more features as input to a trained classification model, the trained classification model being configured to output a risk assessment that the patient is likely to develop PCCS; and outputting an indication of the risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flow chart illustrating a process for triaging patients regarding their risk of developing post-cardiotomy cardiogenic shock (PCCS) in accordance with some embodiments of the present technology.

[0011] Figure 2 Examples of predictive factors that may be used in some embodiments of the present technology to predict a patient's risk of developing PCCS are described.

[0012] Figure 3 A flow chart illustrating a process for training a model for assessing a patient's risk of developing PCCS in accordance with some embodiments of the present technology. DETAILED DESCRIPTION

[0013] Certain patients who undergo cardiac surgery may have an increased probability of developing postoperative cardiogenic shock, also known as postcardiotomy cardiogenic shock (PCCS). Currently, it is unclear which patients are most susceptible, and surgeons and other healthcare providers can only rely on gestalt to predict the occurrence of PCCS or wait until it occurs. Unfortunately, patients who develop PCCS have a significantly increased risk of mortality and morbidity.

[0014] The inventors have recognized the benefits of using a risk model (e.g., using a large sample size of cardiac surgery patients) to predict which patients are most likely to develop PCCS. The output of the risk model can enable a healthcare provider (e.g., a physician) to provide appropriate preemptive cardiac support (e.g., using a mechanical circulatory device) to patients with a predicted higher risk of developing PCCS. Such preemptive support can provide patients with an opportunity to reduce cardiac complications and / or increase the chance of postoperative survival compared to the conventional reactive approach of not providing cardiac support until the patient shows evidence of PCCS. A risk model as described herein can identify baseline factors for predicting PCCS and can be trained to set appropriate weights for each factor. As will be appreciated, the risk model can be implemented as a multivariate model. Thus, embodiments disclosed herein relate to a method for determining a patient's risk of developing PCCS. In some embodiments, a calculator can be developed to allow a physician to input patient baseline information for certain disclosed factors and predict the patient's risk of experiencing PCCS.

[0015] Now turn to the diagram, Figure 1 A flow chart showing a process 100 for classifying a patient's risk of experiencing PCCS based on medical information in accordance with some embodiments of the present technology. In act 110, electronic medical information of a patient may be received. The electronic medical information may include, but is not limited to, electronic health records (EHRs), laboratory reports, medical procedure (e.g., electrocardiogram) reports, physician notes, and medical imaging reports. In some embodiments, this information may be automatically retrieved from an electronic patient record. In other embodiments, act 110 may include receiving information typed into a patient calculator by a physician. In still other embodiments, the received electronic medical information may include first information automatically retrieved from an electronic patient record and second information manually typed into a user interface (e.g., a user interface configured to display a patient calculator) by a physician or other healthcare provider.

[0016] Next, process 100 may proceed to act 120, where one or more features are extracted from the received medical information. The one or more extracted features may include, but are not limited to, cardiac function values, such as left ventricular ejection fraction (LVEF), total bilirubin level, presence of chronic lung disease, New York Heart Association (NYHA) classification, occurrence of cardiogenic shock, Society of Thoracic Surgeons (STS) predicted mortality risk score, whether the patient has undergone or will undergo mitral valve surgery, and / or whether the patient has undergone or will undergo reoperation. Figure 2 A graph showing exemplary features for predicting PCCS that may be used in some embodiments. Other features for predicting PCCS may additionally or alternatively be used in other embodiments.

[0017] As will be further appreciated, the received medical information may include structured data (e.g., based on an ontology, for example, organized using tags, fields, or other metadata associated with the data that can be used for feature extraction) and / or unstructured data (e.g., physician's notes typed electronically in a free text field).

[0018] As will be further appreciated, one or more features may be extracted using natural language processing (NLP) techniques. Such NLP techniques may be used to analyze text in received medical information to infer information determined to be features. In some examples, a first set of features is extracted using one or more manual techniques and a second set of features is extracted using automated (e.g., NLP) techniques.

[0019] Next, process 100 proceeds to act 130, where the patient is classified (e.g., classified as having a high risk of developing PCCS) based at least in part on the extracted one or more features. The extracted features may be provided, for example, as input to a model (e.g., a trained machine learning model) that is trained to output a patient classification (also referred to herein as a "risk assessment"), or may be provided as input to an algorithm that weights the features to determine a score based on which the patient's classification is based. In some embodiments, the values ​​of two or more extracted features may be combined to create a derived feature that is not present in the received medical information. For example, the values ​​of two or more extracted features may be added, multiplied, subtracted, divided, or combined in any other manner to produce a value for the derived feature. According to some embodiments, the value of such a derived feature may be used to train a machine learning model and / or may be provided as input to a trained machine learning model. Examples of trained machine learning models include, but are not limited to) neural networks (e.g., deep neural networks), logistic regression models, random forest models, naive Bayes models, and decision tree models. According to the techniques described herein, any single type of model or combination of multiple types of models may be used as a risk model. In some embodiments, the trained machine learning model may use ensemble methods, regression methods, or any other suitable method to identify features and / or combinations of features that are most predictive (e.g., by associating them with larger weights) that a patient is likely to develop PCCS.

[0020] In some embodiments, the classification assigned to the patient may depend at least in part on the value of the extracted data. For example, a patient with an LVEF of <= 35% (e.g., LVEF <= 25%) may have a higher risk of PCCS. Similarly, a patient with a grade III or IV NHYA rating may have a higher risk of PCCS. Worsening (e.g., higher) total bilirubin levels may also indicate a higher risk of PCCS. For example, a total bilirubin level>2.5 may indicate a higher risk of PCCS. An STS predicted risk of mortality (STS PROM) of>=5% or>=10% may also indicate a higher risk of PCCS. Additionally, a patient who has undergone or is undergoing mitral valve surgery may indicate a higher risk of PCCS for the patient.

[0021] In some embodiments, a cumulative score may be determined based on the value of the extracted data. Figure 1 As shown in , in some embodiments, instead of or in addition to patient classification (see action 140 ), a cumulative score may be provided to a physician (see action 150 ).

[0022] In some embodiments, the data associated with each of the features (e.g., categories) may be weighted evenly. In other embodiments, the data associated with each of the features may be weighted differently. For example, a feature with a greater likelihood of causing PCCS (e.g., patients undergoing reoperation or patients with LVEF <= 35% or <= 25%) may be weighted higher than another feature (e.g., a feature with a lower likelihood of causing PCCS).

[0023] As will be appreciated, other thresholds and / or features may be used in some embodiments employing an algorithm-based risk model, wherein the extracted and / or derived features are considered in a discretized manner (e.g., by comparing the value of the feature to a threshold value). In embodiments employing a trained machine learning model (e.g., a multivariate linear regression model, a neural network-based model, etc.), the values ​​of the extracted and / or derived features may be considered in a continuous manner (e.g., by not requiring comparison to a threshold value). Additionally, when assessing whether a patient is likely to develop PCCS, the trained model itself may learn the predictive value of a combination of features, rather than a combination of features that are explicitly defined as derived features provided as input to the model. In this way, the model may learn (e.g., by setting appropriate weights) that a combination of features that individually may not indicate that a patient is likely to develop PCCS may still be a good indicator that the patient is a high risk (or low risk, moderate risk, or some other risk category) PCCS patient.

[0024] Next, process 100 may proceed to act 140, where the patient classification determined in act 130 may be output. The patient classification may be output in any suitable manner. For example, when a trained machine learning model is used to determine a patient classification, the output of the model may include a numerical value, and outputting the patient classification may be performed by outputting an indication of the numerical value and / or a numerical range associated with the numerical value output from the model. For example, the numerical value and / or numerical range may be displayed on a user interface of a computing device so that a healthcare provider reviewing the output may consider the risk assessment when making a decision about whether a patient would benefit from additional cardiac support to reduce the patient's risk of developing PCCS. In some embodiments, the output of the model may be provided as a percentage probability that a patient will develop PCCS.

[0025] In some embodiments, the patient classification determined in act 130 may represent one of a plurality of risk categories (e.g., high risk, medium risk, low risk) indicating the degree to which the patient is at risk for PCCS. The risk category associated with the patient may be output in any suitable manner. For example, a visual indicator (e.g., red, yellow, green) may be displayed (e.g., on a user interface, in the patient's electronic health record, etc.) to indicate the likelihood that the patient will exhibit PCCS. In some embodiments, the visual indicator (e.g., red, yellow, green) and a numerical value, numerical range, or percentage may be output in act 140 to indicate the likelihood that the patient will exhibit PCCS. In some embodiments, process 100 may be performed for each of a plurality of patients in a medical facility, and a list of patients classified as most likely to develop PCCS may be output in act 140.

[0026] As will be appreciated, process 100 may be repeated for the same patient over a desired period of time (e.g., while the patient is in the hospital). In such an example, the patient may have the same output at act 140. In other examples, the output of act 140 may differ over time depending on the medical information received (at act 110) and the classification of act 130. Thus, the physician may continue to treat the patient in different ways depending on the output in act 140 over time.

[0027] Information related to the output 140 may be provided to a physician and / or other healthcare provider in any suitable manner. For example, information may be provided to a healthcare provider in a user interface that provides an indication of the output of the risk model and / or describes how the output is determined based on the extracted features. In some embodiments, based on inspection of the output 140 provided on the user interface, a user (e.g., a healthcare provider) may interact with the user interface to change one or more aspects of the risk model, and such changes may be reflected in the output 140. For example, a physician may change one or more criteria to see if it changes the risk assessment score. In some embodiments, a physician may use information about the risk assessment score to order one or more additional tests, and the results of such tests may be considered when determining a patient classification and / or calculating a risk assessment score.

[0028] In some embodiments, patients initially classified as "borderline" or having a moderate risk of PCCS may be reclassified after additional information becomes available. In such an example, the classification may be updated after this additional information becomes available, wherein the new classification is communicated to hospital staff.

[0029] As described herein, for embodiments of the present technology that employ a model or algorithm in which extracted features may be weighted to determine a patient classification, the model / algorithm may be updated (e.g., retrained or tuned.) For example, the model / algorithm may be updated based on feedback as to whether the classification provided by the model / algorithm is consistent or inconsistent with the assessment of a healthcare professional examining the patient's PCCS risk.

[0030] Figure 3 A process 300 for training a model (e.g., a machine learning model) that can be used to assess the risk of a patient developing PCCS is described in accordance with some embodiments of the present disclosure. The process 300 may begin in action 310, where a PCCS criterion may be defined. As described herein, PCCS may not have a consistent definition in the scientific literature. Thus, a patient's medical record may not clearly identify whether the patient has developed PCCS after cardiac surgery. Therefore, as described herein, the inventors have recognized the benefits of creating a criterion for reaching this definition. In some embodiments, the PCCS criterion may include the definition of two categories, namely PCCS (positive class) and no PCCS (negative class). In other embodiments, more than two categories may be defined. For example, in addition to the PCCS and no PCCS categories defined above, the PCCS criterion may include a "possible PCCS" category and a "possible no PCCS" category. The inclusion of such categories may, for example, provide a deep understanding of how the model may perform for patients whose characteristics may be broader than the PCCS and no PCCS categories.

[0031] The PCCS criteria defined in act 310 may be used to determine whether the patient should be classified as one of a plurality of risk categories (e.g., low risk, medium risk, high risk) for developing PCCS. Next, process 300 may proceed to act 312, where a plurality of patient groups are generated based on the PCCS criteria defined in act 310.

[0032] exist Figure 3 In the example shown in , two patient groups are generated based on whether the patient in the data set of patient medical data meets the PCCS criteria, i.e., the PCCS patient group and the non-PCCS patient group. For example, the patient may be included in the PCCS patient group when the patient has any of the following criteria: postoperative insertion of an intra-aortic balloon pump (IABP); postoperative insertion of a catheter-based cardiac assist device; intraoperative or postoperative insertion of a veno-arterial extracorporeal membrane membrane (VA-ECMO) device for heart failure, or a veno-venous extracorporeal membrane membrane (VV-ECMO) device converted to a VA-ECMO device during or after surgery; an unplanned ventricular assist device (VAD) implanted simultaneously with cardiovascular surgery; or cardiac death or death during the initial cardiac surgery. For example, a patient may be included in the non-PCCS patient group if all of the following conditions are met: the patient's time in the intensive care unit is less than 10 days; no prolonged ventilation is required; no re-admission to the hospital is required; no end-organ dysfunction is detected; no open-heart surgery is required; no cardiac assist device is required; the patient is discharged alive; and no re-admission to the intensive care unit is required. It should be understood that the above PCCS criteria are merely exemplary, and any suitable PCCS criteria may be used to define patent groups for training risk models according to the techniques described herein. It should also be understood that any suitable number of patient groups may be generated based on the PCCS criteria, and the use of two patient groups is merely an example. For example, in some embodiments, three patient groups corresponding to the risk categories "low risk", "medium risk", and "high risk" may be generated using the PCCS criteria, and the model may be trained to identify patients in the three patient groups using the corresponding patient data as described herein.

[0033] Next, process 300 may proceed to act 314, where a plurality of data fields may be defined for use as training data. For example, a data set of patient medical data may include a large number (e.g., 100, 200, 300, 500) of data fields that may potentially be used to train a risk model. The inventors have recognized and appreciated that using a smaller number of data fields (i.e., less than all possible data fields) may be advantageous because not all data fields may be relevant to predicting the risk of developing PCCS. The data fields may be defined in any suitable manner. For example, perturbation techniques may be used to determine which data fields, when included as input to the model, contribute to producing an accurate risk assessment as reflected in the output of the risk model. Other techniques may be used alternatively. For example, one or more of a field with mostly empty data, a field for an irrelevant feature, a field for a feature that defines a PCCS criterion, a field for a post-intervention feature, or a highly correlated field may not be used to train a risk model. In some embodiments, missing values ​​in a data set may be imputed, class ratios in training data may be balanced, and / or feature values ​​of defined data fields may be normalized before being used as training data.

[0034] Next, process 300 may proceed to act 316, where the patient medical information associated with the data fields defined in act 314 may be used to train a risk model (e.g., a machine learning model) for the patient group defined in act 312. At least some patients (e.g., 10%, 20%, 30%, etc.) in the first patient group (e.g., the PCCS patient group) and the second patient group (e.g., the non-PCCS patient group) may be excluded for not being used to train the model. The data of the excluded patients may be used to obtain a performance score (e.g., sensitivity, specificity) for the trained model. Next, process 300 may proceed to act 320, where the trained model may be output. For example, the trained model may be stored for later use in assessing the PCCS risk of new patients not included in the training data set or the exclusion data set. This document is combined with Figure 1 Process 100 describes an example of using a trained model to determine a patient classification associated with the occurrence of PCCS.

[0035] As described herein, the inventors have recognized that it may be beneficial to update (e.g., retrain) a model for assessing the risk of developing PCCS. For example, if some criteria used to initially define a patient group are not available in a new patient data set, the PCCS criteria may be refined to define a patient group for retraining the model. Similarly, it may be desirable to define different data fields in action 314 for a particular hospital system or different electronic medical record types or formats. Additionally, it may be beneficial to tune a model for a particular hospital system to improve the predictive power of the model for that particular hospital system. For example, a hospital system may perform a large number (e.g., hundreds, thousands, etc.) of cardiac surgeries in a year, and it may be beneficial to use patient data from a large number of cardiac surgery patients to tune a risk model to make it specific to the patient population of the hospital system.

[0036] Return to Figure 3 , it may be determined in act 322 whether the risk model is to be updated. If it is determined in act 322 that the model does not need to be updated, then process 300 may end. Otherwise, process 300 may proceed to act 324, where it may be determined whether the PCCS criteria should be updated. If it is determined in act 324 that the PCCS criteria should be updated (e.g., because some PCCS criteria are not available in the patient data set used for retraining), then process 300 may proceed to act 310, where the PCCS criteria may be updated based on the current patient data set used to update the model. If it is determined in act 324 that the PCCS criteria do not need to be updated, then process 300 may proceed to act 326, where it is determined whether the data fields used to retrain the model should be updated. If it is determined in act 236 that the data fields used for training are updated, then process 300 may proceed to act 314, where the data fields may be defined based on the current patient data set used to update the risk model. If it is determined in act 326 that the data field does not need to be updated, then process 300 may proceed to act 316, where the model is trained using the current patient data set selected for retraining the model. As will be appreciated, the risk model may be retrained with any suitable training data set and for any suitable patient population at any suitable time interval so that the predictive power of the risk model is consistent with the patient population for which a risk assessment of developing PCCS is desired. For example, if it is desired to assess the risk of developing PCCS for a particular patient over time, the risk model may be trained using longitudinal data that describes the values ​​of the defined data fields at multiple points in time.

[0037] Having thus described several aspects and embodiments of the technology described in this disclosure, it will be appreciated that various changes, modifications and improvements will be readily apparent to those skilled in the art. Such changes, modifications and improvements are intended to be within the spirit and scope of the technology described herein. For example, it will be readily conceivable for those skilled in the art to perform functions and / or obtain results and / or one or more of the advantages described herein, and each of such changes and / or modifications is considered to be within the scope of the embodiments described herein. Those skilled in the art will recognize, or be able to determine many equivalents of the specific embodiments described herein using only routine experiments. Therefore, it will be appreciated that the aforementioned embodiments are presented only by way of example, and within the scope of the appended claims and their equivalents, other ways different from the specific descriptions may be practiced to implement the invention embodiments. In addition, if such features, systems, articles, materials, equipment and / or methods are not mutually inconsistent, any combination of two or more features, systems, articles, materials, equipment and / or methods described herein is included within the scope of this disclosure.

[0038] The above embodiments can be implemented in any of numerous ways. One or more aspects and embodiments of the execution of the process or method disclosed herein can utilize program instructions that can be executed by a device (e.g., a computer, a processor or other device) to execute the process or method, or control the execution of the process or method. In this regard, various inventive concepts can be embodied as a computer-readable storage medium (or multiple computer-readable storage media) (e.g., a computer memory, one or more floppy disks, compressed optical disks, optical disks, tapes, flash memories, circuit configurations in field programmable gate arrays or other semiconductor devices, or other tangible computer storage media) encoded with one or more programs, and the one or more programs are executed when executed on one or more computers or other processors to implement one or more of the various embodiments described above. One or more computer-readable media can be transportable so that one or more programs stored thereon can be loaded onto one or more different computers or other processors to implement the various aspects described above. In some embodiments, the computer-readable medium can be a non-transitory medium.

[0039] The above-described embodiments of the present technology may be implemented in any of numerous ways. For example, the embodiments may be implemented using hardware, software, or a combination thereof. When implemented in software, the software code may be executed on any suitable processor or set of processors, whether provided in a single computer or distributed among multiple computers. It should be appreciated that any component or set of components that perform the functions described above may generally be considered to be a controller that controls the functions described above. The controller may be implemented in numerous ways, such as using dedicated hardware, or using general-purpose hardware (e.g., one or more processors) that is programmed using microcode or software to perform the functions described above, and may be implemented in a combination of several ways when the controller corresponds to multiple components of the system.

[0040] Furthermore, it should be appreciated that a computer may be embodied in any of a number of forms, such as a rack-mounted computer, a desktop computer, a laptop computer, or a tablet computer, as non-limiting examples. Furthermore, a computer may be embedded in a device not generally regarded as a computer but having suitable processing capabilities, including a personal digital assistant (PDA), a smart phone, or any other suitable portable or fixed electronic device.

[0041] In addition, the computer may have one or more input and output devices. These devices may be used, among other things, to present a user interface. Examples of output devices that may be used to provide a user interface include a printer or display screen for visual presentation of output and a speaker or other sound generating device for audible presentation of output. Examples of input devices that may be used for a user interface include a keyboard and a pointing device, such as a mouse, a touch pad, and a digital tablet. As another example, a computer may receive input information through speech recognition or in other audible formats.

[0042] Such computers may be interconnected in any suitable form via one or more networks, including local or wide area networks, such as enterprise networks, and intelligent networks (IN) or the Internet. Such networks may be based on any suitable technology and may operate according to any suitable protocol and may include wireless networks, wired networks or fiber optic networks.

[0043] In addition, as described, some aspects may be embodied as one or more methods. The actions performed as part of a method may be ordered in any suitable manner. Thus, embodiments may be constructed in which the actions are performed in a different order than described, which may include performing some actions simultaneously, even though shown as sequential actions in an illustrative embodiment.

[0044] All definitions, as defined and used herein, should be understood to control over dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0045] As used herein in the specification and claims, the indefinite articles "a" and "an" should be understood to mean "at least one" unless clearly indicated to the contrary.

[0046] The phrase "and / or" as used herein in the specification and claims should be understood to mean "either or both" of the elements so connected, i.e., elements that are present in conjunction in some cases and separately in other cases. Multiple elements listed with "and / or" should be interpreted in the same manner, i.e., "one or more" of the elements so connected. Other elements may optionally be present in addition to the elements explicitly identified by the "and / or" clause, whether related or unrelated to the elements explicitly identified. Thus, as a non-limiting example, when used in conjunction with open language such as "comprising", a reference to "A and / or B" may refer to only A (optionally including elements other than B) in one embodiment; to only B (optionally including elements other than A) in another embodiment; to both A and B (optionally including other elements) in yet another embodiment; etc.

[0047] As used herein in the specification and claims, the phrase "at least one" in reference to a list of one or more elements should be understood to mean any one or more selected from the elements in the list of elements, but not necessarily including at least one of each element explicitly listed in the list of elements and not excluding at least one element of any combination of elements in the list of elements. This definition also allows for the optional presence of elements other than the elements explicitly identified in the list of elements to which the phrase "at least one" refers, whether related or unrelated to the elements explicitly identified. Thus, as a non-limiting example, "at least one of A and B" (or equivalently, "at least one of A or B" or equivalently "at least one of A and / or B") may refer in one embodiment to at least one (optionally including more than one) A, wherein B is not present (and optionally including elements other than B); in another embodiment, to at least one (optionally including more than one) B, wherein A is not present (and optionally including elements other than A); in yet another embodiment, to at least one (optionally including more than one) A and at least one (optionally including more than one) B (and optionally including other elements); etc.

[0048] In addition, the words and terms used herein are for descriptive purposes and should not be considered limiting. The use of "including," "comprising," or "having," "containing," "involving," and variations thereof herein is intended to encompass the items listed thereafter and their equivalents as well as additional items.

[0049] In the claims, as well as in the foregoing description, all transitional phrases such as "comprises," "comprising," "carrying," "having," "containing," "involving," "having," "consisting of," and the like shall be understood as open-ended, i.e., meaning including but not limited to. Only the transitional phrases "consisting of" and "consisting essentially of" shall be closed or semi-closed transitional phrases, respectively.

[0050] The use of ordinal terms such as "first", "second", "third", etc. in the claims to modify claim elements does not in itself imply any priority, precedence or order of one claim element over another claim element or the temporal order of the acts of performing the method, but merely serves as a label to distinguish one claim element with a particular name from another element with the same name (but used in ordinal terms) to distinguish the claim elements.

Claims

1. A method for predicting whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS), the method comprising: Receive medical information from patients; extracting one or more features from the received medical information; providing the one or more features as input to a trained classification model, the trained classification model being configured to output a risk assessment of the patient's likelihood of developing PCCS; and An indication of the risk assessment is output.

2. The method of claim 1, wherein the medical information of the patient comprises one or more of electronic health records, laboratory reports, medical procedure reports, physician notes, and medical imaging reports.

3. The method of claim 1, wherein the medical information comprises structured data and unstructured data. 4 . The method of claim 3 , wherein extracting one or more features comprises extracting at least some of the one or more features from the unstructured data using natural language processing.

5. The method of claim 1, wherein the one or more characteristics include left ventricular ejection fraction and / or total bilirubin level.

6. The method according to claim 1, further comprising: receiving data indicating whether the patient has experienced PCCS; and The trained classification model is retrained based at least in part on the received data.

7. The method according to claim 1, wherein The risk assessment comprises a numerical value, and Outputting an indication of the risk assessment includes displaying the value and / or information based on the value on a user interface.

8. The method according to claim 7, further comprising: Classifying the patient into a risk group of a plurality of risk groups is performed based on the numerical value, wherein Outputting the indication of the risk assessment includes outputting an indication of the risk group for the patient.

9. The method according to claim 8, wherein Performing the classification of the patient into a risk group includes determining whether the value is above a threshold; and The high risk of PCCS is classified when the value is determined to be above the threshold.

10. The method of claim 8, wherein outputting an indication of the risk group for the patient comprises displaying a color-coded indication of the risk group on a user interface.

11. The method according to claim 1, wherein The risk assessment is to classify the patient into a risk group among a plurality of risk groups, and Outputting the indication of the risk assessment includes outputting an indication of the risk group for the patient.

12. The method of claim 1, wherein the trained classification model comprises a trained neural network.

13. The method of claim 1, wherein the trained classification model comprises a trained random forest model.

14. The method of claim 1, wherein the trained classification model comprises a trained multivariate regression model.

15. The method of claim 1, further comprising: receive additional medical information; and The trained classification model is retrained based on the additional medical information.

16. The method of claim 15, wherein the additional medical information includes medical information of a plurality of patients who underwent cardiac surgery at a medical facility.

17. The method of claim 1, further comprising: providing a user interface configured to display values ​​of the one or more characteristics; receiving user input via the user interface to change one or more of the values ​​of the one or more features; simulating a risk assessment of the patient for developing PCCS based at least in part on the altered one or more values ​​to produce a simulated risk assessment; and The simulated risk assessment is displayed on the user interface.

18. The method of claim 1, wherein outputting an indication of the risk assessment comprises outputting a cumulative score associated with the risk assessment.

19. A method for training a risk model to predict whether a patient is likely to experience post-cardiotomy cardiogenic shock (PCCS), the method comprising: receiving a dataset of patient medical information; selecting training data from the dataset of patient medical information based on PCCS criteria and defined data fields, wherein the training data includes patient medical information for at least two risk groups of patients; training the risk model using the selected training data; and The trained risk model is output.

20. The method of claim 19, further comprising: Define multiple PCCS criteria; and The at least two risk groups of patients are generated based on the PCCS criteria.

21. The method of claim 19, further comprising: receiving input regarding the data fields to be defined via a user interface; and The data fields are defined based at least in part on the received input.

22. The method of claim 19, further comprising: The trained model is validated using at least some patient medical information not used to train the model, wherein outputting the trained risk model includes outputting the validated training model.

23. The method of claim 19, further comprising: receiving an instruction to update the trained risk model; and The risk model is retrained in response to receiving the indication to update the trained risk model.

24. The method of claim 19, wherein the risk model comprises a neural network.

25. The method of claim 19, wherein the risk model comprises a random forest model.

26. The method of claim 19, wherein the risk model comprises a multivariate regression model.

27. The method of claim 19, further comprising: receive additional medical information; and The trained risk model is retrained based on the additional medical information.

28. The method of claim 20, wherein the additional medical information includes medical information of a plurality of patients who underwent cardiac surgery at a medical facility.

29. A computer-implemented system for predicting whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS), the system comprising: at least one hardware computer processor; and At least one non-transitory computer-readable medium encoded with a plurality of instructions that, when processed by the at least one hardware computer processor, perform a method comprising: extracting one or more features from the patient's medical information; providing the one or more features as input to a trained classification model, the trained classification model being configured to output a risk assessment of the patient's likelihood of developing PCCS; and An indication of the risk assessment is output.

30. At least one non-transitory computer-readable medium encoded with a plurality of instructions that, when processed by at least one hardware computer processor, perform a method comprising: extracting one or more features from the patient's medical information; providing the one or more features as input to a trained classification model, the trained classification model being configured to output a risk assessment of the patient's likelihood of developing PCCS; and An indication of the risk assessment is output.

Images