Abstract: A computer system selects features of a dataset for predictive modeling. A first set of features that are relevant to outcome are selected from a dataset comprising a plurality of cases and controls. A subset of cases and controls having similar values for the first set of features is identified. The subset is analyzed to select a set of additional features relevant to outcome. A first and second predictive model are evaluated to determine that the second predictive model more accurately predicts outcome, wherein the first predictive model is based on the first set of features and the second predictive model is based on the first set of features and the additional features. The second predictive model is utilized to predict outcomes. Embodiments of the present invention further include a method and program product for selecting features of a dataset for predictive modeling in substantially the same manner described above.

Abstract: A plurality of events are extracted from a plurality of electronic health records associated with a first patient. The extracted plurality of events are analyzed to identify a plurality of stimulus events and a plurality of response events. An association between a first stimulus event and a first response event is determined. A stimulus-response (SR) variable is generated for the first patient based at least in part on the determined association, and the generated SR variable is integrated into one or more predictive cognitive models.

Abstract: Obtain covariates and an outcome data for a population. Partition the population into a plurality of subgroups. Produce outcomes predictions by applying a machine learning model to the covariate data for the population. Establish performance measures based on the outcomes predictions. Compare the performance measures for at least one subgroup to the performance measures for at least one other subgroup. Identify an outlying subgroup for which the machine learning model produces performance measures that are different than the performance measures for one or more other subgroups. Optionally, retrain the machine learning model on additional covariate and outcomes data for the outlying subgroup.

Abstract: A plurality of attributes are extracted from a plurality of electronic health records, where each electronic health record is associated with a patient in a plurality of patients. Additionally, a training data set and a scoring data set are generated based on the plurality of attributes, and a patient similarity model is trained based on the training data set. A precision cohort is identified, where the precision cohort includes patients in the plurality of patients from the scoring data set that are similar to a first patient based on an electronic health record of the first patient and the similarity model. At least one result statistic for each of a plurality of treatments given to patients in the precision cohort is determined, and a first treatment of the plurality of treatments is selected for the first patient based at least in part on the determined result statistics.

Abstract: Techniques for generating an ontology based on biomarker information associated with persons to facilitate improving clinical predictions relating to medical conditions are presented. An ontology generator component (OGC) can extract clinical features associated with patients and their associated times from medical records or databases to develop clinical profiles associated with the patients and relating to a medical condition. OGC can develop an ontology relating to the medical condition, including progression and severity of biomarkers associated with the medical condition, based on the clinical profiles and domain knowledge information relating to the medical condition. OGC can determine global features relating to progression and severity associated with the medical condition based on the ontology. At a forecasting point, the global features can be extracted from the ontology and applied to a prediction model to enhance prediction of onset of, or progression of, the medical condition for a patient.

Abstract: Aspects of the invention include includes identifying a respective estimated clinical risk score for each of a first group of patients and a second group of patients. An alternative probability estimate is generated using a same set of inputs used to determine each respective estimated clinical risk score. An unreliability of a patient's clinical risk score is determined based at least in part on a feature of the patient and on a difference between the alternative probability estimate and the determined respective estimated clinical risk score.

Abstract: A processor may receive data associated with one or more users regarding biomarkers for a condition. The processor may determine one or more progression trajectories from the data using progression modeling. The processor may identify one or more granular stages associated with the one or more progression trajectories. The processor may generate a risk assessment for development of the condition associated with the one or more progression trajectories and the one or more granular stages.

Abstract: Techniques regarding inferring parameters of one or more mechanistic models are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a machine learning component that can identify a causal relationship in a mechanistic model via a machine learning architecture that employs a parameter space of the mechanistic model as a latent space of a variational autoencoder.

Abstract: A request to execute a sequentially randomized controlled trial (sRCT) that relates to a subject regarding a population of humans is received. Datapoints from the population that are needed for the sRCT and factors that define the population as fitting the sRCT are identified. It is detected that a human within an interaction satisfies the factors and therefore is part of the population. During the interaction, the datapoints from the human that are needed for the sRCT are gathered in response to detecting that the human is part of the population, and treatment is randomly assigned if the interaction is also a treatment decision point.

Abstract: Techniques regarding inferring parameters of one or more mechanistic models are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a machine learning component that can identify a causal relationship in a mechanistic model via a machine learning architecture that employs a parameter space of the mechanistic model as a learned distribution sampled within a generative adversarial network.

Abstract: A computing device and computer-implemented method for post-hoc correction of a decision generated by a machine learning model. The computing device accesses a trained first machine learning (ML) model, a dataset, and a utility function. The computing device trains a second ML model based on performing post-hoc correction of a first set of decisions generated by the first ML model on the dataset. The training includes processing the first set of decisions with respect to a second set of decisions made by the second ML model on the dataset. The training further includes configuring, based on the processing, the second ML model with parameters from a set of parameters optimizing a loss-objective function that concurrently maximizes utility of the second set of decisions according to the utility function and a log-likelihood on the dataset. After training, the second ML model is outputted as a loss-calibrated ML model.

Abstract: A computer system trains a predictive model. A plurality of subsets of features are selected from a dataset comprising a plurality of cases and controls and a plurality of features. Cases and controls are matched to select a plurality of case-control subsets for each subset of features, each case-control subset having similar values for the corresponding subset of features. For each case-control subset, a statistical significance of each feature of the plurality of features absent from the subset of features used to match the case-control subset is identified. A final subset of features is selected based on satisfying a statistical significance of each feature for the plurality of case-control subsets. A predictive model is trained using the final subset of features. Embodiments of the present invention further include a method and program product for training a predictive model in substantially the same manner described above.

Abstract: A notation assistant system has a machine learning system, a notation processing system, a scoring system, and a suggestion system. The machine learning system trains a classifier for rating a factor related to a clinical narrative note that describes a patient's health status. The notation processing system processes a clinical narrative note and the scoring system determines a factor rating, such as a completion score or a clarity score. The scoring system provides the factor rating to an end-user device to display to a user. The notation assistant system is configured to perform a method as a user is entering a clinical narrative note to provide real-time feedback, such as the factor rating. The suggestion system is configured to provide suggestions for modifying the clinical narrative note to improve the rating factor. The notation assistant system applies to patient health conditions such as a health status of a patient's heart for a patient receiving care for congestive heart failure.

Abstract: Embodiments are directed to methodologies, systems and computer program products for generating, for each of a plurality of risk factors in a patient database containing information of a plurality of patients, an index of input values for the risk factor. For each patient P of the plurality of patients, a series of local impact scores is computed for the patient. Computing the series of local impact scores for the patient includes calculating a risk score for the patient with respect to each of the indexed input values for each of the plurality of risk factors. For at least one of the plurality of patients, at least some of the plurality of risk factors are ranked based at least partly on the computed local impact scores for each of the at least some risk factors, and an indication of the ranked risk factors for the at least one patient is provided.

Abstract: Embodiments are directed to methodologies, systems and computer program products for generating, for each of a plurality of risk factors in a patient database containing information of a plurality of patients, an index of input values for the risk factor. For each patient P of the plurality of patients, a series of local impact scores is computed for the patient. Computing the series of local impact scores for the patient includes calculating a risk score for the patient with respect to each of the indexed input values for each of the plurality of risk factors. For at least one of the plurality of patients, at least some of the plurality of risk factors are ranked based at least partly on the computed local impact scores for each of the at least some risk factors, and an indication of the ranked risk factors for the at least one patient is provided.

Abstract: A system and method for assigning assessment tasks includes determining tasks in need of completion, where the tasks include assessing a set of documents containing medical patient data and providing a judgment (e.g., classification or label) based on the contents of the document. A process also includes selecting one or more reviewers based on the one or more tasks and providing one or more documents from the set of documents to each of the selected one or more reviewers for completion of the one or more tasks. The process further includes receiving a result of the one or more tasks after completion by the selected one or more reviewers and storing the result in an electronic medical record database.

Abstract: A system and associated method for alerting to ambiguous medical advice through a data analysis system connected to one or more data sources and a client terminal. The data analysis system has an extraction system, a machine learning system, a categorization system, and an alerting system. The data analysis system extracts first information related to outcomes of one or more medical treatments from one or more first data sources, extracts second information related to outcomes of one or more medical treatments from one or more second data sources, and applies machine learning to the extracted first information and second information to develop a classifier for categorizing proposed treatments. The data analysis system also categorizes a proposed treatment or a combination of treatments as potentially harmful based on the classifier, and provides an alert to a client terminal regarding the proposed treatment.

Abstract: A verification device is configurable for verifying the identity of a person. The verification device may be in communication with a user device and an entity device that may request an authorization decision from the verification device. The verification device may have a collection system configured to populate a voice profile database with a plurality of voice profiles in the form of audio recordings of people. The verification device may also have a testing system configured to select one of the voice profiles, provide the voice profile to an end-user device, and receive an answer from the end-user device, the answer being an attempt by the person to identify an individual associated with the voice profile. The verification device may further have an authentication system configured to determine whether the identity of the person is verified based on the answer from the end-user device.

Abstract: Mechanisms are provided to implement a medical condition verification system. The medical condition verification system receives patient electronic medical record (EMR) data and parses the patient EMR data to identify an instance of a medical code or medical condition indicator present in the patient EMR data. The medical condition verification system performs cognitive analysis of the patient EMR data to identify evidential data supportive of the instance referencing an associated medical condition. The medical condition verification system generates a measure of risk of the patient having the medical condition based on the identified evidential data and based on a machine learned relationship of medical factors in patient EMR data relevant to generating the measure of risk for the associated medical condition. The medical condition verification system generates an output representing the measure of risk of the patient having the associated medical condition.

Abstract: A system and associated method for assessing level of completeness of healthcare data storage based on collected data. The system includes a collection system, a measurement system, an expectation system, and an alerting system. In a method of assessing the health of stored data, the system collects real-time data from at least one data source, determines measurements for a plurality of parameters based on the collected data, and generates expectations for a future period of time for the plurality of parameters based on data analysis technique. The system also compares the expectations for the future period of time to subsequent measurements collected for that period of time to determine whether the subsequent measurements satisfy an expectation threshold and provide an alert to a client terminal. The alert is a result of the comparison of the expectations and the subsequent measurements and provides an assessment of data storage quality and alerts on anticipated deficiencies.