Method and apparatus for identifying the risk of post-cardiotomy cardiogenic shock in a patient

Abstract

A method and system for predicting whether a patient is likely to develop post-cardiotomy cardiogenic shock (PCCS) are described. The method includes receiving medical information about 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 configured to output a risk assessment that the patient is likely to develop PCCS, and outputting an indication of the risk assessment.

Description

Technical Field

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

Background Art

[0002] Post - cardiotomy cardiogenic shock (PCCS) is relatively rare but is a significant cause of death in patients undergoing cardiac surgery. Although not consistently defined in the literature, PCCS is associated with circulatory failure following cardiac surgery that necessitates mechanical circulatory support (e.g., support using an extracorporeal membrane oxygenation (ECMO) device, a ventricular assist device (VAD), etc.) and / or administration of high - dose inotropic substances.

Summary of the Invention

Means for Solving the Problems

[0003] In some embodiments, a method for predicting whether a patient is likely to develop post - cardiotomy cardiogenic shock (PCCS) is provided. The method includes receiving medical information for 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 configured to output a risk assessment that the patient is likely to develop PCCS, and outputting an indication of the risk assessment.

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

[0005] In another aspect, the risk assessment is a categorization of a patient into a risk group among a plurality of risk groups, and outputting an indication of the risk assessment includes outputting an indication of the risk group to which the patient belongs. In another aspect, the trained classification model includes a trained neural network. In another aspect, the trained classification model includes a trained random forest model. In another aspect, the trained classification model includes a trained multivariate regression model. In another aspect, the method further includes receiving additional medical information and retraining the trained classification model based on the additional medical information. In another aspect, the additional medical information includes medical information regarding a plurality of patients at a medical facility where a cardiac surgery was performed. In another aspect, the method further includes providing a user interface configured to display values for one or more features, receiving user input via the user interface, changing one or more of the values for the one or more features, at least partially based on the changed one or more values, simulating a risk assessment of a likelihood that a patient will develop PCCS, generating a simulated risk assessment, and displaying the simulated risk assessment on the user interface. In another aspect, outputting an indication of the risk assessment includes outputting a cumulative score associated with the risk assessment.

[0006] In some embodiments, a method of training a risk model for predicting whether a patient is likely to develop post - cardiotomy cardiogenic shock (PCCS) is provided. The method includes 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 regarding at least two risk groups of the patient, using the selected training data to train a risk model, and outputting the trained risk model.

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

[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, the non-transitory computer-readable medium being encoded with a plurality of instructions that, when processed by the at least one hardware computer processor, implement a method. The method includes extracting one or more features from medical information regarding a patient and providing the one or more features as input to a 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, there is provided at least one non-transitory computer-readable medium encoded with a plurality of instructions that, when processed by at least one hardware computer processor, implement a method. The method includes extracting one or more features from medical information regarding a patient and providing the one or more features as input to a trained classification model configured to output a risk assessment that the patient is at high risk of developing PCCS, and outputting an indication of the risk assessment.

Brief Description of the Drawings

[0010]

Figure 1

[0011]

Figure 2

[0012]

Figure 3

Mode for Carrying Out the Invention

[0013] Detailed Description A patient undergoing cardiac surgery may have an increased probability of developing postoperative cardiogenic shock, also known as post-cardiotomy cardiogenic shock (PCCS). Currently, it is unknown which patients are most susceptible, and surgeons and other healthcare providers are left to contemplate whether to anticipate the occurrence of PCCS or wait until it occurs. Unfortunately, patients who develop PCCS are at a significantly increased risk of mortality and morbidity.

[0014] The inventors recognize 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 may enable a healthcare provider (e.g., a physician) to provide appropriate prophylactic cardiac support (e.g., the use of a mechanical circulatory device) to patients having a predicted higher risk of developing PCCS. Such prophylactic support may provide a reduction in the margin of cardiac complications and / or an increase in the likelihood of survival following surgery as compared to the conventional reactive approach where cardiac support is not provided to the patient until evidence of PCCS is shown. A risk model as described herein may identify baseline factors for predicting PCCS and may be trained to set appropriate weights for each factor. As will be appreciated, the risk model may be implemented as a multivariable model. Thus, the embodiments disclosed herein relate to a method for determining the risk of a patient developing PCCS. In some embodiments, a computer may be developed to enable a physician to input patient baseline information regarding certain disclosed factors and predict the risk that a patient may suffer from PCCS.

[0015] Turning now to the figures, FIG. 1 shows a flowchart of a process 100 for classifying the risk of a patient undergoing PCCS based on medical information, according to some embodiments of the present technology. At act 110, electronic medical information regarding a patient may be received. The electronic medical information may include, but is not limited to, an electronic health record (EHR), laboratory reports, medical procedure (e.g., electrocardiogram) reports, physician notes, and medical imaging reports. In some embodiments, this information may be automatically read from an electronic patient record. In other embodiments, act 110 may include receiving information entered into a patient computer by a physician. In yet other embodiments, the received electronic medical information may include first information automatically read from an electronic patient record and second information manually entered into a user interface (e.g., a user interface configured to display a patient computer) by a physician or other healthcare provider.

[0016] The process 100 may then proceed to act 120, where one or more features are extracted from the received medical information. The one or more features extracted may include, but are not limited to, cardiac function values such as left ventricular ejection fraction (LVEF), total bilirubin level, presence of chronic pulmonary disease, New York Heart Association (NYHA) classification, occurrence of cardiogenic shock, Society of Thoracic Surgeons (STS) predicted mortality risk score, whether the patient has had or will have mitral valve surgery, and / or whether the patient has had or will have reoperation. FIG. 2 shows a chart of exemplary features for predicting PCCS that may be used in some embodiments. Other features for predicting PCCS may be used in other embodiments in addition to, or as an alternative to, those shown.

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

[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 the text in the received medical information and infer the information determined as the feature. In some cases, 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] Process 100 then proceeds to act 130, where the patient is classified based on one or more features that have been at least partially extracted (e.g., as being at high risk of developing PCCS). The extracted features can be provided as inputs to a model (e.g., a trained machine learning model) trained to output a patient classification (also referred to herein as a "risk assessment"), or as inputs to an algorithm that weights the features to determine a score upon which the classification of the patient is based. In some embodiments, values for two or more of the extracted features can be combined to generate a derived feature that does not exist within the received medical information. For example, the values for two or more of the extracted features can be added, multiplied, subtracted, divided, or otherwise combined in any other manner for generating a value for the derived feature. The value for such a derived feature can be used to train a machine learning model and / or provided as an input to a trained machine learning model according to some embodiments. 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. Any single type of model or combination of multiple types of models can be used as a risk model according to the techniques described herein. In some embodiments, the trained machine learning model(s) can use an ensemble method, a regression method, or any other suitable method to identify the feature(s) and / or combination of features that are most predictive of the patient having a high likelihood of developing PCCS (e.g., by associating the feature with a greater weight).

[0020] In some embodiments, the classification assigned to a patient may depend at least in part on the values of the extracted data. For example, patients with an LVEF <= 35%, such as LVEF <= 25%, may have a higher risk of PCCS. Similarly, patients with a NYHA class III or class IV assessment may also have a higher risk of PCCS. Deteriorating (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 mortality risk (STS PROM) of >= 5% or >= 10% may also indicate a higher risk of PCCS. Additionally, patients who have had or will undergo mitral valve surgery may exhibit a higher risk of PCCS for that patient.

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

[0022] In some embodiments, the data associated with each of the features (e.g., categories) may be weighted uniformly. In other embodiments, the data associated with each of the features may be weighted differently. For example, features with a higher likelihood of causing PCCS (e.g., patients who have undergone re - surgery 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 should be understood, other thresholds and / or features may be used in some embodiments that employ an algorithm-based risk model, and the extracted and / or derived features are considered in a manner that is discretized (e.g., by comparing the values of the features to a threshold). In embodiments that employ 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 a comparison to a threshold). Additionally, the trained model itself may learn the predicted values of combinations of features rather than combinations of features that are explicitly defined as derived features provided as inputs to the model when assessing whether a patient is likely to develop PCCS. Thus, the model may learn (e.g., by setting appropriate weights) that a combination of features that may not individually indicate that a patient is likely to develop PCCS can nevertheless be a good indicator that a patient is a high-risk (or low-risk, medium-risk, or some other risk category) PCCS patient.

[0024] Process 100 may then proceed to act 140, and the patient classification determined at 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 the patient classification, the output of the model may include a numerical value, and outputting the patient classification may be implemented by outputting an indication of the numerical value and / or a range of numerical values associated with the numerical value output from the model. For example, the numerical value and / or range of numerical values may be displayed on a user interface of a computing device so that a healthcare provider can consider the risk assessment when making a determination regarding whether a patient would benefit from additional cardiac support to reduce the risk of developing PCCS. In some embodiments, the output of the model may be provided as a percent 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) that indicate to some extent the risk to the patient with respect to PCCS. The risk category associated with the patient may be output in any suitable manner. For example, visual indicators (e.g., red, yellow, green) may be displayed (e.g., on a user interface, in the patient's electronic medical record, etc.) to indicate the likelihood that the patient exhibits PCCS. In some embodiments, both a 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 exhibits PCCS. In some embodiments, process 100 may be performed for each of a plurality of patients in a healthcare 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, such as while the patient is in the hospital. In such cases, the patient may have the same output in act 140. In other cases, the output in act 140 may vary over time depending on the received medical information (in act 110) and the classification in act 130. Accordingly, a physician may subsequently treat the patient differently depending on the output in act 140 over time.

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

[0028] In some embodiments, patients who were initially classified as having "borderline" or a medium risk with respect to the PCCS can be reclassified after additional information becomes available. In such cases, the classification is updated after such additional information becomes available, and the new classification can be relayed to the hospital staff.

[0029] Regarding embodiments of the present technology that employ a model or algorithm in which the extracted features can be weighted to determine patient classification, the model / algorithm can be updated (e.g., retrained or adjusted). For example, the model / algorithm can be updated based on feedback regarding whether the classification provided by the model / algorithm matched or did not match the assessment of a healthcare professional who reviewed the risk of a patient with respect to the PCCS.

[0030] Figure 3 illustrates a process 300 for training a model (e.g., a machine learning model) that can be used to assess the risk that a patient will develop PCCS, according to some embodiments of the present disclosure. Process 300 can begin with an act 310 where PCCS criteria can be defined. As described herein, PCCS may not have a consistent definition within the scientific literature. Thus, a patient's medical record may not explicitly identify whether the patient developed PCCS following cardiac surgery. Accordingly, as described herein, the inventors recognize the benefit of generating criteria for arriving at such a definition. In some embodiments, the PCCS criteria can include two classes of definitions, namely, PCCS (positive class) and non-PCCS (negative class). In other embodiments, more than two classes can be defined. For example, in addition to the PCCS and non-PCCS classes defined above, the PCCS criteria can include a "Maybe PCCS" class and a "Maybe No PCCS" class. Including such classes can provide insight into how the model can perform with respect to patients who can be characterized by a broader definition of PCCS, for example, than that used with respect to the "PCCS" and "non-PCCS" classes.

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

[0032] In the example shown in FIG. 3, two groups of patients, namely, the PCCS patient group and the non-PCCS patient group, are generated based on whether a patient within a dataset of patient medical data meets the PCCS criteria. For example, a patient may be included in the PCCS patient group when the patient has any one of the following criteria: an intra-aortic balloon pump (IABP) inserted after surgery, a catheter-based heart assist device inserted after surgery, a veno-arterial extracorporeal membrane (VA-ECMO) device inserted during or after surgery for heart failure or a veno-venous extracorporeal membrane (VV-ECMO) device converted to VA-ECMO during or after surgery, an unplanned assistive artificial heart (VAD) implanted in conjunction with cardiovascular surgery, or any one of cardiac death or death during initial cardiac surgery. For example, a patient may be included in the non-PCCS patient group when all of the following are true: the patient was in the intensive care unit for less than 10 days, no prolonged ventilation was required, no readmission to the hospital was required, no end-organ dysfunction was detected, no open-chest procedure was required, no heart assist device was required, the patient survived and was discharged from the hospital, and no readmission to the intensive care unit was required. It should be understood that the foregoing PCCS criteria are merely illustrative, and any suitable PCCS criteria may be used to define patient groups for use in training a risk model 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, the PCCS criteria may be used to generate three patient groups corresponding to risk categories, namely, "low risk", "medium risk", and "high-risk", and the model may be trained to identify patients within those three patient groups using individual patient data as described herein.

[0033] Process 300 can then proceed to act 314, where a plurality of data fields can be defined for use as training data. For example, a dataset of patient medical data can include a large number (e.g., 100, 200, 300, 500) of data fields that can potentially be used to train a risk model. Since not all data fields may be equally relevant to predicting the risk of developing PCCS, the inventors recognize and understand that it may be advantageous to use a smaller number of data fields (i.e., less than all possible data fields). The data fields can be defined in any suitable manner. For example, perturbation techniques can be used to determine which data fields, when included as inputs to the model, contribute to producing an accurate risk assessment as reflected in the output of the risk model. Other techniques can also be used as an alternative. For example, one or more of the fields with mostly empty data, fields for unrelated features, fields related to the features used in defining the PCCS criteria, fields related to post-intervention features, or highly correlated fields may not be used to train the risk model. In some embodiments, missing values in the dataset can be imputed, the class ratios in the training data can be balanced, and / or the feature values for the defined data fields can be normalized prior to being used as training data.

[0034] Process 300 may then proceed to act 316, where 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) with respect to the patient population defined in act 312. At least some patients (e.g., 10%, 20%, 30%, etc.) in the first patient population (e.g., the PCCS patient population) and the second patient population (e.g., the non-PCCS patient population) may be held out by not being used to train the model. Data regarding the held-out patients may be used to obtain performance scores (e.g., sensitivity, specificity) of the trained model. Process 300 may then 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 risk of PCCS for new patients not included in the training data set or the held-out data set. An example of using a model trained to determine a patient classification associated with developing PCCS is described herein in connection with process 100 of FIG. 1.

[0035] As described herein, the inventors recognize that it may be beneficial to update (e.g., retrain) a model used to assess the risk of developing PCCS. For example, if some of the criteria initially used to define the patient population are not available in a new patient data set, the PCCS criteria may be refined to define the patient population used to retrain the model. Similarly, different data fields may need to be defined in act 314 for a particular hospital system or different electronic medical record types or formats. In addition, it may be beneficial to adjust the model for a particular hospital system and improve the predictive ability 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 over the course of a year, use patient data from a large number of cardiac surgery patients, and adjust the risk model to be specific to the patient population of the hospital system, which may be advantageous.

[0036] Turning back to process 300 shown in FIG. 3, at act 322, it can be determined whether the risk model should be updated. If at act 322 it is determined that the model need not be updated, process 300 can end. Otherwise, process 300 can proceed to act 324 where it can be determined whether the PCCS criteria should be updated. If at act 324 it is determined that the PCCS criteria should be updated (for example, because some of the PCCS criteria are not available within the patient dataset used for retraining), process 300 can proceed to act 310 and process 300 can be updated based on the current patient dataset in which the PCCS criteria are used to update the model. If at act 324 it is determined that the PCCS criteria need not be updated, process 300 can proceed to act 326 where it can be determined whether the data fields used to retrain the model should be updated. If at act 326 it is determined to update the data fields used for training, process 300 can proceed to act 314 and the data fields can be defined based on the current patient dataset used to update the risk model. If at act 326 it is determined that the data fields do not need updating, process 300 can proceed to act 316 and the model is trained using the current patient dataset selected to retrain the model. As should be understood, the risk model can be retrained at any suitable interval, using any suitable training dataset, and with respect to any suitable patient population such that the predictive ability of the risk model is aligned 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 over time for a particular patient, the risk model can be trained using time series data that describes the values for defined data fields at multiple time points.

[0037] Although some aspects and embodiments of the technology described in this disclosure have been described as above, it should be understood that various modifications, corrections, and improvements will readily occur to those skilled in the art. Such modifications, corrections, and improvements are intended to be within the spirit and scope of the technology described herein. For example, those skilled in the art can readily envision various other means and / or structures for implementing the functions described herein and / or obtaining one or more of the results and / or advantages, and each such variation and / or modification is considered to be within the scope of the embodiments described herein. Those skilled in the art will be able to recognize or confirm many equivalents of the specific embodiments described herein using no more than routine experimentation. Therefore, it should be understood that the foregoing embodiments are presented by way of example only, and that embodiments of the present invention may be practiced otherwise than as specifically described within the scope of the appended claims and their equivalents. In addition, any combination of two or more of the features, systems, articles, materials, kits, and / or methods described herein is included within the scope of this disclosure if such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent.

[0038] The embodiments described above can be implemented in any of a number of ways. One or more aspects and embodiments of the present disclosure involving the implementation of a process or method may utilize program instructions executable by a device (e.g., a computer, a processor, or other device) to perform or control the performance of the process or method. In this regard, various inventive concepts, when executed on one or more computers or other processors, may be embodied as one or more programs encoded on a computer-readable storage medium (or multiple computer-readable storage media) (e.g., computer memory, one or more floppy (registered trademark) disks, compact disks, optical disks, magnetic tapes, flash memory, circuit configurations within a field programmable gate array or other semiconductor device, or other tangible computer storage media) that implement a method of implementing one or more of the various embodiments described above. The computer-readable medium or media may be transportable such that the program or programs stored thereon can be loaded onto one or more different computers or other processors to implement various ones of the aspects described above. In some embodiments, the computer-readable medium may be a non-transitory medium.

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

[0040] Furthermore, it should be understood that the computer can be embodied in any of several forms, such as, by way of non-limiting example, a rack-mounted computer, a desktop computer, a laptop computer, or a tablet computer. Additionally, the computer can be embedded within a device that is generally not regarded as a computer, such as a personal digital assistant (PDA), a smartphone, or any other suitable portable or fixed electronic device, but has suitable processing capabilities.

[0041] In addition, a computer may have one or more input and output devices. These devices can be used, inter alia, to present a user interface. Examples of output devices that can 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 can be used for a user interface include a keyboard and pointing devices such as a mouse, touchpad, and digitizing tablet. As another example, a computer may receive input information through speech recognition or in other audible formats.

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

[0043] Also, as described, some aspects can be embodied in one or more ways. The acts performed as part of a method can be ordered in any suitable way. Accordingly, even if shown as sequential acts in an illustrative embodiment, embodiments can be constructed in which some acts are performed simultaneously, including in a different order than that shown.

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

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

[0046] As used herein, the phrase "and / or" as used within the specification and claims should be understood to mean "either or both" of the elements so joined, i.e., elements that in some cases coexist conjunctively and in other cases exist disjunctively. A plurality of elements listed using "and / or", i.e., "one or more" of the elements so joined, should be construed in the same manner. Other elements may optionally exist, whether or not they are related to the specifically identified elements, other than those specifically identified by the clause "and / or". Thus, by way of non-limiting example, reference to "A and / or B", when used in conjunction with non-limiting language such as "comprising", can refer in one embodiment to "A only" (optionally including elements other than B), in another embodiment to "B only" (optionally including elements other than A), and in yet another embodiment to "both A and B" (optionally including other elements), etc.

[0047] As used herein, as in the specification and claims, the phrase "at least one" referring to a list of one or more elements should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but does not necessarily include at least one of every element specifically recited in the list of elements, nor does it exclude any combination of elements in the list of elements. This definition also allows for the possibility that elements other than those specifically identified in the list of elements referred to by the phrase "at least one" may optionally be present, whether related or unrelated to those specifically identified elements. Thus, by way of 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") can, in one embodiment, mean "at least one, optionally including more than one, A, with no B present" (optionally including elements other than B), in another embodiment, mean "at least one, optionally including more than one, B, with no A present" (optionally including elements other than A), in yet another embodiment, mean "at least one, optionally including more than one, A", and "at least one, optionally including more than one, B" (optionally including other elements), etc.

[0048] Also, the grammar and terminology used in this specification are for the purpose of explanation and should not be construed as limiting. The use of "including", "comprising", "having", "containing", "involving", and variations thereof in this specification means including the items listed hereinafter, their equivalents, and additional items.

[0049] In the claims and the above specification, all transitional phrases such as "comprising", "including", "carrying", "having", "containing", "involving", "holding", "composed of", and equivalents should be understood to be non-restrictive, i.e., to mean "including but not limited to". Only the transitional phrases "consisting of" and "consisting essentially of" shall be considered restrictive or semi-restrictive 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, or the temporal order in which acts of a method are performed, and is used only as a label to distinguish one claim element having a certain name from another element having the same name (in the absence of the use of ordinal terms) for the purpose of distinguishing claim elements.

Claims

1. A method for predicting whether a patient is likely to develop post - cardiotomy cardiogenic shock (PCCS), the method comprising: receiving medical information about the patient; extracting one or more features from the received medical information; providing the one or more features as an input to a trained classification model configured to output a risk assessment that the patient is likely to develop PCCS; outputting an indication of the risk assessment A method comprising the above.

2. The method according to claim 1, wherein the medical information about the patient includes one or more of an electronic medical record, a laboratory report, a medical procedure report, a physician's note, and a medical imaging report.

3. The method according to claim 1, wherein the medical information includes structured data and unstructured data.

4. The method according to claim 3, wherein extracting one or more features includes using natural language processing to extract at least some of the one or more features from the unstructured data.

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

6. receiving data indicating whether the patient has developed PCCS; retraining the trained classification model at least partially based on the received data The method according to claim 1, further comprising the above.

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

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

9. Categorizing the patient into a risk group includes determining whether the numerical value exceeds a threshold, The method according to claim 8, further comprising classifying the high risk related to PCCS when it is determined that the numerical value exceeds the threshold.

10. Outputting the indication of the risk group for the patient includes displaying the color-coded indication of the risk group on a user interface, according to the method of claim 8.

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

12. The trained classification model includes a trained neural network, according to the method of claim 1.

13. The trained classification model includes a trained random forest model, according to the method of claim 1.

14. The trained classification model includes a trained multivariate regression model, according to the method of claim 1.

15. Receiving additional medical information, Retraining the trained classification model based on the additional medical information, The method of claim 1 further includes.

16. The additional medical information includes medical information about a plurality of patients in a medical facility where cardiac surgery has been performed, according to the method of claim 15.

17. Providing a user interface configured to display values for the one or more features, Receiving user input via the user interface and changing one or more of the values for the one or more features, At least partially simulating a risk assessment that the patient is at high risk of developing PCCS based on the changed one or more values and generating a simulated risk assessment, Displaying the simulated risk assessment on the user interface, The method of claim 1 further includes.

18. Outputting the indication of the risk assessment includes outputting a cumulative score associated with the risk assessment, according to the method of claim 1.

19. A method for training a risk model for predicting whether a patient is at high risk of developing post-cardiotomy cardiogenic shock (PCCS), the method comprising: Receiving a dataset of patient medical information, Selecting training data from the dataset of the patient medical information based on the PCCS criteria and defined data fields, wherein the training data includes patient medical information regarding at least two risk groups of a patient; Training the risk model using the selected training data; Outputting the trained risk model; A method comprising the above steps.

20. Defining a plurality of PCCS criteria; Generating at least two risk groups of the patient based on the PCCS criteria; The method according to claim 19, further comprising the above steps.

21. Receiving an input regarding the data fields for definition via a user interface; Defining the data fields at least partially based on the received input; The method according to claim 19, further comprising the above steps.

22. Further comprising validating the trained model using at least some patient medical information not used for training the model, wherein outputting the trained risk model includes outputting the validated and trained model. The method according to claim 19.

23. Receiving an indication and updating the trained risk model; Retraining the risk model in response to receiving the indication and updating the trained risk model; The method according to claim 19, further comprising the above steps.

24. The method according to claim 19, wherein the risk model includes a neural network.

25. The method according to claim 19, wherein the risk model includes a random forest model.

26. The method according to claim 19, wherein the risk model includes a multivariate regression model.

27. Receiving additional medical information; Retraining the trained risk model based on the additional medical information; The method according to claim 19, further comprising the above steps.

28. The method according to claim 20, wherein the additional medical information includes medical information regarding a plurality of patients in a medical facility where cardiac surgery has been performed.

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 for implementing a method when processed by the at least one hardware computer processor, the method comprising: Extracting one or more features from medical information regarding the patient; Providing the one or more features as an input to a trained classification model configured to output a risk assessment that the patient is at high risk of developing PCCS; Outputting an indication of the risk assessment At least one non-transitory computer-readable medium including A system comprising.

30. At least one non-transitory computer-readable medium encoded with a plurality of instructions for implementing a method when processed by at least one hardware computer processor, the method comprising: Extracting one or more features from medical information regarding a patient; Providing the one or more features as an input to a trained classification model configured to output a risk assessment that the patient is at high risk of developing PCCS; Outputting an indication of the risk assessment At least one non-transitory computer-readable medium including.

Images