Method and apparatus for identifying high-risk patients with respect to percutaneous coronary intervention

Abstract

A method and apparatus are provided for identifying whether a patient is eligible for protected percutaneous coronary intervention (protected PCI). The method includes receiving medical information regarding the patient, extracting one or more features from the received medical information, classifying the patient with respect to eligibility for protected PCI based at least in part on the one or more extracted features, generating a patient classification, and outputting an indication of the patient classification.

Description

Technical Field

[0001] The present disclosure relates to identifying patients with respect to protected percutaneous coronary intervention.

Background Art

[0002] Percutaneous coronary intervention (PCI) is a minimally invasive procedure used to open occluded coronary arteries. Examples of PCI include balloon angioplasty, angioplasty using a stent, and atherectomy. Patients suffering from occluded coronary arteries may be candidates for PCI, some patients are suitable for normal PCI, other patients are suitable for high-risk PCI depending on various risk factors, and still other patients are not suitable for high-risk PCI due to major risk factors.

Summary of the Invention

Means for Solving the Problems

[0003] What is described herein is a system and method for identifying patients who are at high risk for complications and who may benefit from mechanical circulatory support during percutaneous coronary intervention (PCI).

[0004] In some embodiments, a method is provided for identifying whether a patient is eligible for protected percutaneous coronary intervention (protected PCI (登録商標) ). The method includes receiving medical information about the patient, extracting one or more features from the received medical information, and classifying the patient with respect to eligibility for protected PCI (登録商標) based at least in part on the one or more extracted features to generate a patient classification, and outputting an indication of the patient classification.

[0005] On one side, the medical information regarding a patient includes one or more of an electronic medical record, a laboratory report, an electrocardiogram report, and a medical imaging report. On one side, the medical information includes structured data and unstructured data. On one side, extracting one or more features includes extracting at least some of the one or more features from the unstructured data using natural language processing. On one side, extracting one or more features includes extracting at least some of the one or more features using natural language processing. On one side, extracting at least some of the one or more features using natural language processing includes performing natural language processing across multiple types of medical information.

[0006] On one side, one or more features include values associated with a left ventricular ejection fraction and / or one or more co-existing diseases of the patient. On one side, one or more features further include one or more of the diagnostic information. On one side, one or more features further include at least one cardiac value, one or more angiographic values, and one or more of one or more features of a heart disease.

[0007] On one side, the method further includes determining whether the patient meets one or more inclusion criteria and / or one or more exclusion criteria, classifying the patient, and generating a patient classification, which includes generating a patient classification based at least in part on whether the patient meets one or more inclusion criteria and / or one or more exclusion criteria. On one side, classifying the patient includes that the patient undergoes protected PCI when the patient does not meet at least one of the one or more inclusion criteria and / or when the patient meets at least one of the one or more exclusion criteria. (登録商標)including determining that the patient is not eligible with respect to. In one aspect, at least one of the one or more inclusion criteria is at least partially based on one or more of the extracted features. In one aspect, the one or more inclusion criteria includes at least one criterion based on left ventricular ejection fraction.

[0008] In one aspect, classifying a patient includes providing one or more of the extracted features as an input to at least one model, wherein the patient classification is generated at least partially based on an output of the at least one model. In one aspect, the at least one model is a trained statistical model. In one aspect, the method further includes receiving data indicating whether the patient has received or would be suitable for protected PCI (登録商標) and updating the at least one model at least partially based on a comparison of the received data and the generated patient classification. In one aspect, the at least one model is configured to output a score, and the patient classification is generated at least partially based on the score output from the at least one model. In one aspect, the method further includes determining whether the score output from the at least one model exceeds a threshold, and classifying the patient as eligible with respect to protected PCI (登録商標) when it is determined that the score exceeds the threshold. (登録商標)

[0009] In one aspect, extracting one or more features includes applying feature-specific extraction logic to received medical information to extract one or more features. In one aspect, applying feature-specific extraction logic includes selecting a procedure report from the received medical information, searching the selected procedure report for one or more keywords, and extracting a feature value for a particular keyword based on the identification of one or more keywords within the selected procedure report. In one aspect, the method further includes identifying a plurality of keywords of one or more keywords within the selected procedure report and selecting a particular keyword of the plurality of keywords based on priority information associated with the plurality of keywords within the feature-specific extraction logic, and extracting a feature value of one or more features includes extracting a feature value for a particular keyword. In one aspect, the method further includes determining that a particular keyword is associated with an absolute feature value and extracting the absolute feature value as the feature value. In one aspect, the method further includes determining that a particular keyword is associated with a range of values and extracting a feature value as a representative value from the range of values. In one aspect, the method further includes determining that a particular keyword is associated with text, accessing a lookup table that includes a mapping of words to values, and extracting a feature value based at least in part on the particular keyword and words within the text associated with the mapping within the lookup table.

[0010] On one side, the medical information regarding the patient is the first medical information received at the first time, and the method further includes receiving, at a second time after the first time, second medical information regarding the patient, and at least partially reclassifying the patient regarding the eligibility of the patient for protected PCI based on the received second medical information and generating an updated patient classification, and outputting an indication of the updated patient classification. On one side, receiving the second medical information includes receiving the second medical information from the patient's electronic medical record. On one side, receiving the second medical information includes receiving the second medical information via a user interface provided by a computer-implemented system. On one side, the second medical information includes at least one updated value regarding one or more features extracted from the first medical information. On one side, the method further includes extracting one or more updated features from the second medical information, and the reclassification is performed at least partially based on the one or more updated features.

[0011] On one side, the method further includes providing a user interface configured to display values regarding one or more features, receiving user input via the user interface and changing one or more of the values regarding one or more features, and at least partially simulating the classification of the patient regarding the eligibility of the patient for protected PCI based on the changed one or more values and generating a simulated patient classification, and displaying the simulated patient classification on the user interface.

[0012] In some embodiments, a computer-implemented system is provided for identifying whether a patient is eligible for protected percutaneous coronary intervention (PCI). The system comprises 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, implement a method. The method comprises extracting one or more features from medical information about a patient and, at least in part, classifying the patient with respect to eligibility for protected PCI based on the one or more extracted features and generating a patient classification and outputting an indication of the patient classification. (登録商標) including classifying the patient with respect to eligibility for protected PCI, generating a patient classification, and outputting an indication of the patient classification.

[0013] In some embodiments, 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 is disclosed. The method comprises extracting one or more features from medical information about a patient and, at least in part, classifying the patient with respect to eligibility for protected PCI based on the one or more extracted features and generating a patient classification and outputting an indication of the patient classification. (登録商標) including classifying the patient with respect to eligibility for protected PCI, generating a patient classification, and outputting an indication of the patient classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0014]

Figure 1

[0015]

Figure 2

[0016]

Figure 3

[0017]

Figure 4

[0018]

Figure 5

[0019]

Figure 6

[0020]

Figure 7

[0021]

Figure 8

[0022]

Figure 9

[0023]

Figure 10

[0024]

Figure 11

Embodiments for Carrying Out the Invention

[0025] Detailed Description As is known, percutaneous coronary intervention (PCI) can be used to open a patient's occluded coronary artery, such as a patient suffering from coronary artery disease. Unfortunately, using conventional risk factor analysis, a significant portion of patients who would benefit from PCI are not considered suitable for high-risk PCI due to high risk factors. The inventors recognize and understand that such patients may be suitable for PCI if, during the procedure, the patient receives mechanical circulatory support (e.g., via a percutaneous mechanical heart pump) to temporarily assist the heart during the procedure and ensure that blood flow is maintained to critical organs. Such a procedure can be referred to as "protected PCI" in that the heart pump plays a role in protecting the patient's heart during PCI. (登録商標) To achieve this goal, some embodiments are directed to techniques based on data for identifying patients who may be candidates for protected PCI, even though they have higher risk factors that would normally exclude them from high-risk PCI. Predicting patients who would benefit from the use of a mechanical heart pump during PCI would enable those patients to benefit from having PCI, thereby improving patient treatment and surgical outcomes. (登録商標)

[0026] Figure 1 shows protected PCI according to some embodiments of the present technology. (登録商標) (登録商標)FIG. 100 is a flowchart of a process for identifying a patient with respect to protected PCI. In act 110, it is determined whether the patient meets one or more inclusion and / or exclusion criteria for being considered a candidate for protected PCI (登録商標) Any suitable inclusion / exclusion criteria may be used, examples of which are provided herein. In some embodiments of the present technology, only inclusion criteria are considered in act 110, and exclusion criteria are not considered. If it is determined in act 110 that the patient does not meet the inclusion / exclusion criteria, process 100 proceeds to act 120, where it is determined that the patient is not a suitable candidate for protected PCI (登録商標) If it is determined in act 110 that the patient meets the inclusion / exclusion criteria, process 100 proceeds to act 130, where a hemodynamic functional decline score is determined. In some embodiments, the hemodynamic functional decline score is determined using one or more trained models (e.g., one or more machine learning models) configured to capture one or more values or features extracted from medical information (e.g., electronic medical records, laboratory reports, medical imaging reports, etc.) as input and output a hemodynamic functional decline score. Non-limiting examples of medical information that may be used to determine the hemodynamic functional decline score are schematically shown in FIG. 2. Additional types of medical information that may be used to determine the hemodynamic functional decline score are described in more detail herein in connection with FIGS. 4-6. Any suitable model may be used to determine the hemodynamic functional decline score, and embodiments of the present technology are not limited in this regard.

[0027] Process 100 then proceeds to act 140, where it is determined whether the determined hemodynamic functional decline score exceeds a threshold. If it is determined in act 140 that the score does not exceed the threshold, process 100 proceeds to act 120, where it is determined that the patient is not a candidate for protected PCI (登録商標) If it is determined in act 140 that the score exceeds the threshold, process 100 proceeds to act 110, where it is determined that the patient is a candidate for protected PCI (登録商標)Proceed to act 150, which is identified as a candidate regarding [the matter]. As described in connection with process 300 shown in FIG. 3, in addition to determining whether the score exceeds a threshold value, other criteria may be considered when determining whether a patient is identified as a candidate regarding protected PCI (登録商標) when determining whether the patient is identified as a candidate regarding [the matter].

[0028] In some embodiments, the protected PCI (registered trademark) recommendation determined in act 150 is provided to healthcare providers (e.g., physicians such as cardiologists or interventional cardiologists, medical assistants, and / or care managers). The protected PCI (registered trademark) recommendation may be provided in any suitable manner. For example, an indicator representing the protected PCI (registered trademark) recommendation may be automatically included in the patient's electronic medical record. In some embodiments, the indicator may include a text alert (e.g., "Protected PCI is recommended"). In other embodiments, the indicator may include a visual indicator such as a green - red indicator system used to provide the protected PCI (registered trademark) recommendation to healthcare providers, where "green" indicates that the patient is a good candidate regarding protected PCI (registered trademark) and "red" indicates that the patient is not a good candidate regarding protected PCI (registered trademark). In some embodiments, the techniques described herein may be used to analyze the medical information of multiple patients in a healthcare facility (e.g., a hospital) to identify and / or prioritize patients regarding protected PCI (registered trademark), and a summary report of those patients most suitable regarding protected PCI (registered trademark) may be provided to healthcare providers (e.g., physicians) to facilitate clinical decisions regarding the treatment of the patients.

[0029] In some embodiments, protected PCI (登録商標) Feedback provided by one or more healthcare providers based on the recommendation may be used to improve the accuracy of the model used to determine the hemodynamic function degradation score. For example, if a healthcare provider determines that a patient is a good candidate regarding protected PCI (登録商標) (or alternatively, protected PCI (登録商標)If one agrees with the recommendation (that it is not a good candidate), that information may be used to update (e.g., retrain) the model used to determine the hemodynamic functional decline score to further improve the accuracy of the model based on the feedback. In some cases, this feedback may be provided to the patient's EHR, which is configured to transmit the feedback to the system. In such cases, the EHR may also be configured to monitor whether the doctor is screening the patient for follow-up, whether follow-up has been performed, whether the patient desires to wait, and / or whether a procedure to be performed has been performed. Again, one or more of these parameters may be used to retrain the model. They may also be incorporated into the scoring (e.g., see discussion regarding FIG. 11 where reclassification of patients may be enabled through the system) and / or patient support. As will be appreciated, in other cases, the feedback may be provided to a computer-implemented (or mobile device implemented) system that performs the classification. In such cases, the system may be configured to transfer such data to the patient's EHR.

[0030] FIG. 3 illustrates a flowchart of a process 300 for identifying patients related to protected PCI according to some embodiments of the present technology. In the exemplary process shown in FIG. 3, certain features and values are such that the patient is a candidate for protected PCI (登録商標) is illustrated. In the exemplary process shown in FIG. 3, certain features and values are such that the patient is a candidate for protected PCI (登録商標)It is considered for process 100 to determine whether it should be identified as a candidate regarding . In act 310, the inclusion criteria may include having a left ventricular ejection fraction (LVEF) of less than or equal to 45% and having a diagnostic cardiac catheterization report in their electronic medical records. No exclusion criteria are specified. If it is determined that the patient meets these inclusion criteria, process 300 proceeds to act 320 where a hemodynamic dysfunction score is determined for the patient. As discussed regarding process 100, the hemodynamic dysfunction score may be determined as the output of one or more models that capture features or values as inputs based on the medical information regarding the patient. The input features or values provided as inputs to the model may or may not be related to the inclusion criteria considered in act 310.

[0031] Process 300 then proceeds to act 330 to determine whether the patient's LVEF is less than or equal to 35%. In act 330, if it is determined that the LVEF is less than or equal to 35%, process 300 proceeds to act 360 where the patient is identified as a candidate regarding protected PCI (registered trademark). In act 330, if it is determined that the patient's LVEF is not less than or equal to 35% (i.e., the LVEF is between 35% and 45%), process 300 proceeds to act 340 to determine whether the hemodynamic dysfunction score exceeds threshold X. In act 340, if it is determined that the score is less than threshold X, process 300 proceeds to act 350 where the patient is identified as not being a suitable candidate regarding protected PCI (登録商標) In act 340, if it is determined that the score exceeds threshold X, process 300 proceeds to act 360 where the patient is identified as a candidate regarding protected PCI (登録商標) In act 360, the patient is identified as a candidate regarding protected PCI.

[0032] Figure 4 shows protected PCI based on medical information according to some embodiments of the present technology (登録商標)FIG. 400 is a flowchart of a process for classifying patients as good / bad candidates for a procedure. At act 410, electronic medical information about the patient is received. The electronic medical information may include, but is not limited to, an electronic health record (EHR), laboratory reports, medical procedure reports (e.g., electrocardiogram reports), physician notes, and medical imaging reports. Process 400 then proceeds to act 420 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, diagnostic information, cardiac function values (e.g., LVEF, cardiac power), co-morbidities (e.g., anemia, chronic obstructive pulmonary disease (COPD), hypertension), angiographic information, and cardiac disease descriptors (e.g., long-term calcified lesions and mitral regurgitation). 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., compiled based on ontology using a label, field, or other metadata associated with the data that can be used for feature extraction).

[0033] In some embodiments of the present technology, at least some of the one or more features are extracted using natural language processing (NLP) techniques. Such NLP techniques may be used to analyze the text within the received medical information and infer the information determined as a feature. NLP may be used to analyze individual types of medical information (e.g., medical reports) and / or may be used across multiple types of medical information to extract one or more features. In some cases, a first set of features may be extracted using one or more manual techniques and a second set of features may be extracted using automated (e.g., NLP) techniques. Automated techniques for extracting one or more features from a patient's electronic medical information (e.g., an electronic health record) are described in more detail herein with respect to FIG. 10.

[0034] Process 400 then proceeds to act 430, where the patient is classified (e.g., as a good candidate for protected PCI (登録商標) based on one or more of the extracted features (e.g., as being a good candidate for protected PCI). The extracted features may be provided as input to a model (e.g., a trained machine learning model) trained to output a patient classification, or as input to an algorithm that weights the features to determine a score on which the classification of the patient is based. Process 400 then proceeds to act 440, where the patient classification determined at act 430 is output. The patient classification may be output in any suitable manner, examples of which are described above in connection with act 150 of process 100 shown in FIG. 1. For example, visual indicators (e.g., red, yellow, green) may be associated with the patient to indicate the classification. Alternatively, process 400 may be performed for each of a plurality of patients within a medical facility, and a list of patients classified as being most likely to benefit from protected PCI (登録商標) may be output at act 440.

[0035] In some embodiments, the patient may be flagged for additional follow-up (e.g., if the patient is a "borderline" case that is close to the state recommended for protected PCI (登録商標) ). As described herein in connection with FIG. 11, in some embodiments, patients initially classified as not recommended for protected PCI (登録商標) may be reclassified as the patient's medical status changes over time. For example, as additional medical information associated with the patient becomes available (e.g., by being entered into the patient's electronic medical record and / or by being directly entered into a computer-implemented (or mobile device-implemented) system configured to output a patient classification), patients initially classified as not recommended for protected PCI (登録商標) may be considered for reclassification based on intervening medical data (e.g., additional cardiac catheterization laboratory reports, additional blood test data, etc.) during the reclassification.(登録商標) may be reclassified as recommended with respect to. In this way, the classification status of patients in a medical facility is tracked over time, and patients who may benefit from the protected PCI (登録商標) procedure may be dynamically identified.

[0036] As will be understood, patients may also exist as cases of interest to them (e.g., patients classified as recommended with respect to protected PCI (登録商標) ), and if the physician requests additional information before recommending the procedure to that patient, a flag for additional follow-up may be set. Similarly to the above, once the additional information is received, the patient may be reclassified again (e.g., using similar or different classifications), and the physician may make a recommendation to that patient based on the classification.

[0037] In some embodiments of the present technology, different classification techniques may be used in different healthcare systems and / or multiple classification techniques may be used in a single healthcare system. FIG. 5 shows two such techniques, namely, protected PCI (登録商標)A "screening" classification technique configured to primarily (or only) analyze an electronic health record (EHR) to identify patients at a health center for further manual screening to assess eligibility with respect to [[ID=]], and a "model" or "algorithm" classification technique having a higher level of complexity compared to the screening classification technique in that a patient is classified using a model or algorithm based on extracted features, as described, for example, in connection with the processes of FIGS. 1, 3, and 4. The screening technique may be less complex than the algorithmic approach in the sense that, among other things, the screening technique may operate on limited patient data (e.g., EHR data only), the screening technique may provide a simpler feature extraction process than the algorithmic technique, and the output of the screening technique may not be as complex and / or predictive as the output of the algorithmic technique, and thus further follow-up (either by a person or another classification process) may be required for the screening technique but not necessarily for the algorithmic technique.

[0038] FIG. 6 shows that different healthcare systems (e.g., hospitals, medical clinics, etc.) protect PCI according to the method described herein. (登録商標)Illustrate schematically that one or more classification techniques can be used to identify patients who will benefit therefrom. For example, following two exemplary classification techniques described in FIG. 5, some healthcare systems may adopt a "screening" classification technique, while other healthcare systems may adopt an "algorithm" classification technique. Still other healthcare systems may adopt both techniques or a combination thereof. As shown in FIG. 6, the screening classification technique may have a simpler workflow and have limited patient data (although its use is not so limited), and may be advantageous to use within a healthcare system. For example, a local clinic or other medical facility without a laboratory and / or medical imaging equipment may be well-suited to using the screening classification technique. However, it should be understood that the use of the less complex screening technique is not so limited. For example, a large-scale healthcare system may benefit from protected PCI (登録商標) The screening classification technique may be adopted to identify a subset of patients who may benefit from the protected PCI procedure. The medical information of the subset of patients identified by the screening classification technique is then further analyzed using an algorithm classification technique (or some other classification technique) to refine the subset of patients and identify a smaller subset of patients who are eligible for protected PCI (登録商標) It should be understood that some healthcare systems may adopt only one type of classification technique (e.g., only the screening technique, the algorithm technique, or some other technique not shown).

[0039] FIGS. 7 and 8 show protected PCI using the techniques described herein (登録商標)Illustrates an exemplary workflow for classifying patients as eligible or ineligible with respect to []. In the exemplary workflow shown in FIG. 7, a treatment coordinator may collaborate with an interventional cardiologist (IC). As an overview of the workflow, the results of the screener classification technique may be reviewed by the treatment coordinator from the IC clinic. Based on that review, a subset of the cases "left for candidate consideration" may be queried to the IC for review, and then they are reviewed by the IC. In some embodiments, the cases left for candidate consideration may be determined based at least in part on an automated analysis of data extracted from the patient's electronic chart and / or other electronic medical information, examples of which are described herein. In some embodiments, automatically analyzing the extracted data includes associating a score or scores greater than one with the set of extracted data. For example, the extracted data may be associated with one or more categories, and the score may be applied to each category of the extracted data. Non-limiting examples of categories associated with the extracted data include, but are not limited to, gender, age, family history of heart disease, LVEF value, LVEF time series record, co-existing diseases (e.g., cardiac co-existing diseases (e.g., successful resuscitation after a cardiac event), non-cardiac major co-existing diseases (e.g., end-stage renal disease (ESRD), peripheral vascular disease (PVD), other co-existing diseases), combinations of co-existing diseases, presence of one or more diseases (e.g., coronary artery disease, left main disease, and / or chronic total occlusion), findings from ECHO (e.g., TTE, TEE, stress), whether a cardiac surgeon or interventional cardiologist has been requested for a diagnosis, number of diagnostic catheterizations, readmission to the observation hospital, heart failure drug treatment, number of other cardiac drug treatments, or whether the patient has received a previous coronary artery bypass graft (CABG), PCI procedure, open heart procedure, and / or a procedure for implanting a cardiac device. In some embodiments, the score assigned to a category of the extracted data depends at least in part on the value of the extracted data.For example, regarding the extracted LVEF value, the first score may be used when LVEF <= 40%, and the second score may be used when LVEF is 41% - 50%. In some embodiments, the score associated with the extracted data may be selected to represent the level of risk associated with the specific data. For example, a higher score may be assigned to each category of data having a value corresponding to a higher risk factor for the patient. In some embodiments, a cumulative score (e.g., across multiple categories) may be provided for review. In some embodiments, the cumulative score may be determined by adding the scores associated with each category. In some embodiments, the data associated with each category may be weighted evenly such that the cumulative score is determined by the summation of the individual category scores. In other embodiments, the data associated with one or more categories may be weighted higher than other categories such that the cumulative score represents the weighted sum of the individual category scores. In this regard, the first category may be weighted more heavily than the second category if the first category can correlate to a higher risk factor for the patient than the second category.

[0040] Information related to the score may be provided to the treatment coordinator in any suitable manner. For example, the treatment coordinator may be provided with information that provides an indication of the score and / or explains the method by which the score was calculated based on the extracted data within a user interface. Regarding cases left for candidate consideration, the IC may interact with the patient for patient outreach or consult with a physician or cardiologist regarding the identified cases. Based on a review of the results of the screener classification technique, the user may interact with the user interface and change one or more aspects of the scoring, and such changes may be reflected in the candidate consideration of the cases identified during the workflow. In some embodiments, the status assigned to a particular patient may also or alternatively be used to update the candidate consideration of the cases identified during the workflow.

[0041] In the exemplary workflow shown in FIG. 8, a medical assistant may collaborate with a cardiologist. As an overview of the workflow, the results of the screener classification technique may be reviewed by the medical assistant from the cardiologist's clinic. Based on that review, a subset of the cases "left for candidate consideration" may be referred to the cardiologist for review, and thereafter they are reviewed by the cardiologist. Similar to the workflow in FIG. 7, the cases left for candidate consideration may be determined, at least in part, based on an automated analysis of data extracted from the patient's electronic chart and / or other electronic medical information, examples of which are described herein. For cases left for candidate consideration, the cardiologist may interact with the patient or query the patient's IC for further consideration and / or opinions. Based on the review of the results of the screener classification technique, a user (e.g., a cardiologist) may interact with the user interface and change one or more aspects of the scoring, and such changes may be reflected in the candidate consideration of the cases identified during the workflow. In some embodiments, the status assigned to a particular patient may also or alternatively be used to update the candidate consideration of the cases identified during the workflow.

[0042] As will be appreciated, with respect to the workflows in both FIGS. 7 and 8, patients may in some instances be enumerated as "borderline" as described with respect to FIG. 11 and reclassified after additional information becomes available. In such instances, the classification is updated after such additional information becomes available and the new classification may be relayed to the hospital staff.

[0043] With respect to 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 is such that the classification provided by the model / algorithm is protected PCI (登録商標)It may be updated (e.g., retrained or adjusted) based on feedback as to whether it matched or did not match the assessment of the healthcare provider who reviewed the patient eligibility regarding. Figure 9 shows the comparison between the predictions or recommendations output by the model / algorithm and the actual values for patients who did / did not receive protected PCI (登録商標) and who would / would not be appropriate based on published criteria or physician labeling. Figure 9 further illustrates an exemplary method by which the model / algorithm can be adjusted to reduce false positives and false negatives by applying feedback to adjust the model / algorithm.

[0044] As shown, the model / algorithm performs well when there are true positives or true negatives, i.e., when the output of the model / algorithm and the actual values match. False negatives (patients who received or were appropriate for protected PCI, but the model / algorithm did not identify the patient as being eligible for protected PCI (登録商標) can occur, for example, when the model / algorithm is not sensitive enough (e.g., not weighted enough) to the features most strongly correlated with the protected PCI (登録商標) decision. In such cases, the weighting for those features can be adjusted as appropriate to improve the predictions output from the model / algorithm. Other aspects of the model / algorithm may also be updated as described in Figure 9. False positives (patients who did not receive or were not eligible for protected PCI, but the model / algorithm identified the patient as being eligible for protected PCI (登録商標) can occur, for example, when the model / algorithm is too sensitive (e.g., weights too heavily) to a certain feature. In such cases, the weighting for those features can be adjusted as appropriate to improve the predictions output from the model / algorithm. (登録商標) (登録商標) (登録商標)

[0045] ​(For example, in connection with act 420 of process 400 shown in FIG. 4) As described herein, some embodiments may relate to extracting one or more features from electronic medical information to facilitate patient classification. The inventors have recognized and understood that, for example, when simple keyword searches or medical task code-based extractions are used, the information is represented within the electronic medical information and the variability of the ways in which it is described can complicate the extraction of one or more features. For example, the medical diagnosis of coronary artery disease (CAD) may not be consistently shown within a patient's electronic health record (EHR). If an extraction technique for determining whether a patient has CAD is configured to search the patient's EHR only with respect to the presence / absence of specific medical task codes corresponding to CAD, for example, some patients who have the characteristics of CAD but do not have a clear diagnosis using those specific task codes in their EHRs may be missed. Some embodiments herein may extract a feature (e.g., whether a patient has CAD) from electronic medical information by examining additional data elements such as drug therapies and / or procedures related to the feature (e.g., procedures related to CAD).

[0046] FIG. 10 illustrates a process 1000 for multi-grained extraction of one or more features from electronic medical information, according to some embodiments. In act 1010, electronic health information regarding a plurality of patients may be analyzed to separate the patients into a plurality of cohorts, an example of which is shown in Table 1. [Table 1]

[0047] As shown in the exemplary cohort of Table 1, for cohort 3 in which patients have a low LEVF value (e.g., < 35%) and did not undergo PCI procedures, all of the patients within cohort 3 are protected PCI due to that low LEVF value (登録商標)Despite being considered potential candidates for the procedure, only one-third of those patients (patients within cohort 3B) had CAD diagnoses described in their electronic medical information. Some embodiments use additional information (e.g., medication treatment information, procedure information) within a patient's electronic medical information to determine whether a patient is associated with a particular characteristic (e.g., whether the patient has coronary artery disease) when such an association may not be explicit within the patient's electronic medical information. For example, in the exemplary cohort shown in Table 1, additional electronic medical information (e.g., procedure reports) regarding 2,000 patients within cohort 3A was analyzed to determine whether some or all of the patients had CAD (and thus may be eligible for protected PCI) despite not having an explicit CAD diagnosis described in their medical records. The additional electronic medical information may include, but is not limited to, electrocardiograms, transthoracic echocardiograms (TTEs), cardiovascular stress echocardiograms, nuclear stress tests, cardiovascular stress tests, Doppler ultrasound tests, coronary diagnostic angiograms, and cardiac catheterization reports. The additional electronic medical information may include structured data and / or unstructured data. (登録商標) Returning to process 1000, after separating patients into multiple cohorts, process 1000 may proceed to act 1020 where feature-specific extraction logic is used to extract one or more features of interest from the additional electronic medical information. For example, the feature-specific extraction logic for the feature "LVEF" may search for the presence of a TTE within the patient's electronic medical information, and if found, information associated with the LVEF may be extracted from the TTE.

[0048]

[0049] ​Information associated with a particular feature may be extracted from a patient's electronic medical information in any suitable manner. In some embodiments, the extraction is performed using natural language processing. For example, one or more reports or other documents as defined within feature-specific logic may search for one or more keywords and identify information associated with the feature to be extracted. When the feature-specific logic is configured to search for multiple keywords, the feature-specific logic may assign priorities to the multiple keywords such that the keyword with the higher assigned priority will be selected for extraction of its corresponding value when the multiple keywords are present in the electronic medical information. Thus, process 1000 may proceed to act 1030, in which it is determined whether the electronic medical information (e.g., TTE) contains multiple keywords. In act 1030, if it is determined that the electronic medical information contains multiple keywords, process 1000 may proceed to act 1040, in which a particular keyword is selected based on the priority information associated with the multiple keywords. For example, as described above, a keyword with a higher priority may be selected. In act 1030, if it is determined that the electronic medical information does not contain multiple keywords, or after selection of a particular keyword in act 1040, process 1000 may proceed to act 1050, in which it is determined whether the selected keyword is associated with an absolute value for the feature. If it is determined that the keyword is associated with an absolute value for the feature, process 1000 may proceed to act 1060, in which the value associated with the feature for the keyword is extracted as the feature value. Otherwise, process 1000 may proceed to act 1070, in which it is determined whether the keyword is associated with a range of values. In act 1070, if it is determined that the keyword is associated with a range of values, process 1000 may proceed to act 1080, in which the feature value is determined as a representative value from the range of values associated with the keyword. Otherwise, process 1000 may proceed to act 1090, in which a lookup table is used to determine the value for the feature.For example, electronic medical information may include one or more words (e.g., normal, excellent, lower normal limit, poor, severe, abnormal, etc.) that qualitatively describe a value regarding a feature. The value extracted regarding the feature may be based on a look-up table that maps one or more of these words (e.g., based on a range defined within a medical document) to the value.

[0050] Acts 1020 - 1090 may be repeated for a plurality of features of interest to extract feature values from a patient's electronic medical information. The extracted feature values may then be used, at least in part, to classify the patient (e.g., as a patient regarding protected PCI (登録商標) as described herein with respect to act 430 within process 400 of FIG. 4).

[0051] The inventors have recognized and understood that patients initially classified as not recommended for protected PCI (登録商標) may later be classified as patients recommended for protected PCI due to intervening events between the time the patient was initially classified and a later time. Thus, the inventors have recognized that patient classification (e.g., whether a patient is recommended for protected PCI (登録商標) should be updated over time to account for such intervening events. (登録商標)understands the benefits of the system such that the classification (whether a candidate regarding (登録商標) or not) can be updated as additional information (e.g., additional electronic medical information) associated with the patient becomes available. For example, a physician or other healthcare provider may enter additional medical information into a user interface associated with the patient's electronic chart, and the patient classification may be updated based on the additional medical information. In other instances, a physician may order one or more tests to be performed and later added to the patient's electronic chart. Still in other instances, a physician or other healthcare provider may directly enter additional information into a computer-implemented (or mobile device-implemented) system that performs patient classification. In such instances, updating the patient classification over time as new information associated with the patient becomes available may enable long-term tracking of the patient's classification status. For example, initially classified as not a candidate regarding (登録商標) protected PCI, "borderline" patients may have their status re-evaluated based on additional information so that they can be reclassified as candidates regarding

[0052] FIG. 11 illustrates a process 1100 for updating patient classification based on additional information, according to some embodiments of the present technology. In act 1110, a patient may be classified (e.g., as a good / bad candidate regarding (登録商標) protected PCI) based at least in part on medical information associated with the patient. For example, a patient may be classified as a good / bad candidate regarding (登録商標)It may be classified as a bad candidate regarding it, and its embodiments are described in this specification. Process 1100 may then proceed to act 1112 where additional information associated with the patient is received. For example, a physician may interact with the user interface of a computer system configured to implement one or more than one of the classification techniques described in this specification (e.g., screener classification technique, algorithm classification technique) and input at least a portion of the additional information. Alternatively, at least a portion of the additional information may be entered into the patient's electronic health record (EHR), and the additional information may be provided from the EHR as input to the classification technique. In some embodiments, one or more than one of the classification techniques described in this specification may be integrated with the EHR (e.g., as a module or plugin associated with the EHR) such that when the additional information is entered into the EHR, an updated classification is automatically and / or generated in response to a user request, and a patient classification may be generated. In such embodiments, receiving the additional information in act 1112 of process 1100 may occur simultaneously with receiving the additional information as it is entered into the EHR. Act 1112 may also occur simultaneously with other acts of process 1100 (e.g., acts 1114 and 1116 as described below).

[0053] In cases where the classification system is implemented as part of the EHR, it should be understood that the plugin may also be configured to perform calculations after a defined period (e.g., every 5, 10, 15, 20 minutes) after new patient information is added to the patient record.

[0054] After receiving additional information in act 1112, process 1100 may proceed to act 1114 where an updated patient classification may be determined, at least in part, based on the additional information. For example, the additional information may be provided as input to one or more than one of the classification techniques described herein and an updated patient classification may be determined. The additional information may be considered in any suitable manner by the classification technique. For example, the additional information may be used to replace a portion of the information used to generate an initial patient classification (e.g., the patient classification determined in act 1110). As a specific example, an updated LVEF value for the patient may be extracted from a new TTE that was not available when the initial patient classification was determined, and the updated LVEF value for the patient may be used in place of the LVEF value used during the initial patient classification determination. In some embodiments, the additional information may augment the data and / or features used within the initial patient classification data without replacing any specific values. Process 1100 may then proceed to act 1116 where the updated patient classification may be output. As discussed in connection with process 400 in FIG. 4, the updated patient classification may be output in any suitable manner, examples of which are described above in connection with act 150 of process 100 shown in FIG. 1. For example, visual indicators (e.g., red, yellow, green) may be associated with the patient to indicate the classification. As described herein in connection with other processes, process 1100 may be implemented for each of a plurality of patients within a medical facility, and a list of patients classified as most likely to benefit from protected PCI (登録商標) It should be understood that a list of patients classified as most likely to benefit from protected PCI may be output in act 1116. Thus, all or a subset of the patients in a medical facility are tracked over time and as new information about the patients in the medical facility becomes available for consideration by the classification technique, a current list of patients who may benefit from protected PCI (登録商標) may be identified.

[0055] In some embodiments, a computer-implemented system configured to perform one or more of the patient classification techniques described herein presents a user interface to enable a physician or other healthcare provider to test a "hypothetical" scenario by adjusting the values of one or more features used to generate a patient classification. For example, the user interface may display values for a plurality of features extracted from a patient's EHR that are used to generate the current patient classification for the patient. The healthcare provider may then interact with the user interface and change one or more of the values displayed on the user interface (e.g., by typing or otherwise entering a new value into a field of the user interface), and the patient may be reclassified (e.g., as a simulation) based at least in part on the changed values. In some embodiments, the user interface may present a slider element for representing a value for one or more of the features, and the healthcare provider may change the value by interacting with the slider element. By enabling the healthcare provider to simulate how changes in various feature values affect the patient classification, the healthcare provider may gain a deeper understanding of how different features contribute to the patient classification decision and provide insights regarding specific features for monitoring within a patient that may be recommended for protective PCI (登録商標) with respect to, and may provide insights regarding specific features for monitoring within a different patient.

[0056] 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 described herein, and such variations and / or corrections are each 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 nothing more than routine experimentation. Therefore, the foregoing embodiments are presented by way of example only, and it should be understood 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.

[0057] 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, when various concepts of the present invention are executed on one or more computers or other processors, a method for implementing one or more of the various embodiments described above is implemented, and is encoded using one or more programs, a computer-readable storage medium (or plural computer-readable storage media) (e.g., a computer memory, one or more floppy (registered trademark) disks, a compact disk, an optical disk, a magnetic tape, a flash memory, a circuit configuration within a field programmable gate array or other semiconductor device, or other tangible computer storage medium) may be embodied. The computer-readable medium or plural media can be transportable such that the program or plural programs stored thereon can be loaded onto one or more different computers or other processors for implementing various ones of the aspects described above. In some embodiments, the computer-readable medium may be a non-transitory medium.

[0058] The embodiments described above of the present technology can be implemented in any of a number of ways. For example, the embodiments may 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 with microcode or software to implement the functions enumerated above, and when the controller corresponds to multiple components of the system, it may be implemented in a combination of ways.

[0059] 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 may 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 that has suitable processing capabilities.

[0060] In addition, the 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, the computer may receive input information through speech recognition or in other audible formats.

[0061] Such computers may be interconnected by one or more networks in any suitable form, including local area networks such as corporate networks or wide area networks, and intelligent networks (IN) or the Internet. Such networks may be based on any suitable technology, may operate according to any suitable protocol, and may include wireless networks, wired networks, or fiber optic networks.

[0062] Also, as described, some aspects may be embodied as one or more methods. The acts performed as part of the method may be ordered in any suitable manner. Thus, even if shown as sequential acts in an illustrative embodiment, embodiments may be constructed in which acts are performed in a different order than shown, including performing some acts simultaneously.

[0063] All definitions 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.

[0064] As used herein, the indefinite articles "a" and "an" as used in the specification and claims should be understood to mean "at least one" unless clearly indicated otherwise.

[0065] As used herein, the phrase "and / or" as used in 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 disjunctively. A plurality of elements recited 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 related or unrelated to those specifically identified by the "and / or" clause. Thus, by way of non-limiting example, a 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.

[0066] As used herein, as in the specification and claims, the phrase "at least one", referring to a list of one or more elements, is to 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 listed 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.

[0067] 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 is meant to encompass the items listed hereinafter and their equivalents as well as additional items.

[0068] 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.

[0069] The use of ordinal terms such as "first", "second", "third", etc. in the claims to modify claim elements does not by 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 identifying whether a patient is eligible for protected percutaneous coronary intervention (PCI), the method comprising: Receiving medical information about the patient; Extracting one or more features from the received medical information; Classifying the patient with respect to the patient's eligibility for protected PCI and generating a patient classification, at least in part based on the one or more extracted features; Outputting an indication of the patient classification A method comprising.

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, an electrocardiogram report, 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 extracting at least some of the one or more features from the unstructured data using natural language processing.

5. The method according to claim 1, wherein extracting one or more features includes extracting at least some of the one or more features using natural language processing.

6. The method according to claim 4 or 5, wherein extracting at least some of the one or more features using natural language processing includes performing natural language processing across multiple types of the medical information.

7. The method according to claim 1, wherein the one or more features include a left ventricular ejection fraction and / or a value associated with one or more co-existing diseases of the patient.

8. The method according to claim 7, wherein the one or more features further include one or more of the diagnostic information.

9. The method according to claim 8, wherein the one or more features further include one or more of at least one cardiac value, one or more angiographic values, and one or more features of one or more heart diseases.

10. Determining whether the patient meets one or more inclusion criteria and / or one or more exclusion criteria Further comprising, Classifying the patient and generating a patient classification includes, at least in part, generating the patient classification based on whether the patient meets the one or more inclusion criteria and / or the one or more exclusion criteria, the method of claim 1.

11. Classifying the patient includes determining that the patient is not eligible for protective PCI when the patient does not meet at least one of the one or more inclusion criteria and / or when the patient meets at least one of the one or more exclusion criteria, the method of claim 10.

12. At least one of the one or more inclusion criteria is based, at least in part, on the one or more extracted features, the method of claim 10.

13. The one or more inclusion criteria include at least one criterion based on left ventricular ejection fraction, the method of claim 12.

14. Classifying the patient is providing the one or more extracted features as an input to at least one model, wherein the patient classification is generated, at least in part, based on the output of the at least one model, the method of claim 1.

15. The at least one model is a trained statistical model, the method of claim 14.

16. receiving data indicating whether the patient has received or would be suitable for protective PCI, and updating the at least one model, at least in part, based on a comparison of the received data and the generated patient classification further comprising the method of claim 14.

17. The at least one model is configured to output a score, and the patient classification is generated, at least in part, based on the score output from the at least one model, the method of claim 14.

18. determining whether the score output from the at least one model exceeds a threshold, and classifying the patient as eligible for protective PCI when it is determined that the score exceeds the threshold further comprising the method of claim 17.

19. The medical information regarding the patient is first medical information received at a first time, and the method further comprises receiving, at a second time after the first time, second medical information regarding the patient; at least partially reclassifying the patient with respect to the patient's eligibility for protected PCI based on the received second medical information and generating an updated patient classification; outputting an indication of the updated patient classification The method according to claim 1, comprising. **Claim 20** The method according to claim 19, wherein receiving the second medical information comprises receiving the second medical information from the patient's electronic medical record. **Claim 21** The method according to claim 19, wherein receiving the second medical information comprises receiving the second medical information via a user interface provided by a computer-implemented system. **Claim 22** The method according to claim 19, wherein the second medical information comprises at least one updated value regarding one or more features extracted from the first medical information. **Claim 23** further comprising extracting one or more updated features from the second medical information and wherein reclassifying is performed at least partially based on the one or more updated features. The method according to claim 19. **Claim 24** Extracting one or more features comprises applying feature-specific extraction logic to the received medical information to extract the one or more features The method according to claim 23, comprising. **Claim 25** Applying feature-specific extraction logic comprises selecting a procedure report from the received medical information; searching the selected procedure report for one or more keywords; extracting a feature value for a particular keyword based on the identification of the one or more keywords within the selected procedure report The method according to claim 24, comprising. **Claim 26** identifying a plurality of keywords of the one or more keywords within the selected procedure report; selecting a particular keyword of the plurality of keywords based on priority information associated with the plurality of keywords within the feature-specific extraction logic and Extracting the feature value of the one or more than one feature includes extracting the feature value regarding the specific keyword, according to the method described in claim 25. [

27. ] Determining that the specific keyword is associated with an absolute feature value; Extracting the absolute feature value as the feature value; The method according to claim 25 or claim 26, further comprising the above steps. [

28. ] Determining that the specific keyword is associated with a range of values; Extracting the feature value as a representative value from the range of values; The method according to claim 25 or claim 26, further comprising the above steps. [

29. ] Determining that the specific keyword is associated with text; Accessing a look-up table including a mapping of words to values; Extracting the feature value at least partially based on the specific keyword and the words in the text associated with the mapping in the look-up table; The method according to claim 25 or claim 26, further comprising the above steps. [

30. ] Providing a user interface configured to display values regarding the one or more than one feature; Receiving user input via the user interface and changing one or more than one of the values regarding the one or more than one feature; At least partially simulating the classification of the patient regarding the eligibility of the patient for protected PCI based on the one or more than one changed value, and generating a simulated patient classification; Displaying the simulated patient classification on the user interface; The method according to claim 1, further comprising the above steps. [

31. ] A computer-implemented system for identifying whether a patient is eligible for protected percutaneous coronary intervention (PCI), the system comprising: At least one hardware computer processor; At least one non-transitory computer-readable medium, the at least one non-transitory computer-readable medium being encoded with a plurality of instructions, which, when processed by the at least one hardware computer processor, implement a method, the method comprising: extracting one or more features from the medical information regarding the patient; classifying the patient with respect to the eligibility of the patient for protective PCI and generating a patient classification, at least in part, based on the one or more extracted features; outputting an indication of the patient classification; and at least one non-transitory computer-readable medium comprising; a system comprising. **Claim 32** At least one non-transitory computer-readable medium, the at least one non-transitory computer-readable medium being encoded with a plurality of instructions that, when processed by at least one hardware computer processor, implement a method, the method comprising: extracting one or more features from the medical information regarding the patient; classifying the patient with respect to the eligibility of the patient for protective PCI and generating a patient classification, at least in part, based on the one or more extracted features; outputting an indication of the patient classification; and at least one non-transitory computer-readable medium comprising.

Images