Machine-learning based query construction and pattern identification for hereditary angioedema

Abstract

A method, computer program product, and system identifying a probability of a medical condition in a patient. The method includes a processor obtaining data set(s) related to a patient population diagnosed with a medical condition and based on a frequency of features in the data set(s), identifying common features and weighting the common features based on frequency of occurrence in the data set(s) to generate mutual information. The processor generates pattern(s) including a portion of the common features to generate a machine learning algorithm(s). The processor compiles a training set of data to use to tune the machine learning algorithm(s). The processor dynamically adjusts common features in the pattern(s) such that the machine learning algorithm(s) can distinguish patient data indicating the medical condition from patient data not indicating the medical condition. The processor applies the machine learning algorithm(s) to data related to the undiagnosed patient, to determine the probability.

What is claimed is: 
     
         1 . A computer-implemented method, comprising:
 obtaining, by one or more processors in a distributed computing environment, one or more machine-readable data sets related to a patient population from one or more databases;   identifying, by the one or more processors, based on an initial patient definition, a portion of data from the machine-readable data sets related to a patient population, wherein the portion of the data comprises patients of the patient population with a medical condition;   based on a frequency of features in the portion of the data, identifying, by the one or more processors, common features in the portion of the data and weighting the common features based on frequency of occurrence in the portion of the data, wherein the common features comprise mutual information;   generating, by the one or more processors, one or more patterns comprising a portion of the common features;   generating, by the one or more processors, one or more machine learning algorithms based on the one or more patterns, the one or more machine learning algorithms to identify presence or absence of the given medical condition in an undiagnosed patient based on absence or presence of features comprising the one or more patterns in data related to the undiagnosed patient;   utilizing, by the one or more processors, statistical sampling to compile a training set of data, wherein the training set comprises data from the one or more data sets and at least one additional data set comprising data related to a population without the medical condition, and wherein utilizing the statistical sampling comprises formulating and obtaining queries based on the data set and processing and responding to the queries, the processing comprising, for each query:
 evaluating, by the one or more processors, the query to determine one of a high or a low level of anticipated complexity of a prospective response to the query; 
 based on the query being evaluated at a low level of anticipated complexity, assigning, by the one or more processors, the query to a computing resource in the distributed computing environment, wherein the computing resource is configured to respond to low level complexity queries; and 
 based on the query being evaluated at a high level of anticipated complexity, distributing, by the one or more processors, the query over a group of computing resources of the distributed computing environment to maximize efficiency, wherein the distributing comprises assigning each computing resource of the group of computing resources a portion of the query to execute in parallel with at least one other computing resource of the group of computing resources executing another portion of the query; 
   tuning, by the one or more processors, the one or more machine learning algorithms by applying the one or more machine learning algorithms to the training set of data;   dynamically adjusting, by the one or more processors, the common features comprising the one or more patterns to improve accuracy such that the one or more machine learning algorithms can distinguish patient data indicating the medical condition from patient data that does not indicate the medical condition; and   determining, by the one or more processors, based on applying the one or more machine learning algorithms to data related to the undiagnosed patient, a probability, wherein the probability is a numerical value indicating a percentage of commonality between the data related to the undiagnosed patient and the one or more patterns.   
     
     
         2 . The method of  claim 1 , wherein the initial patient definition is selected from the group consisting of: a pre-defined diagnosis code and a pre-defined medication. 
     
     
         3 . The method of  claim 2 , wherein the pre-defined medication is selected from the group consisting of: Cinryze, Firazyr, Berinert, and Kalbitor, and wherein the probability indicates a probability that the undiagnosed patient has the medical condition. 
     
     
         4 . The method of  claim 1 , wherein the one or more machine-readable data sets comprise the data related to the undiagnosed patient. 
     
     
         5 . The method of  claim 4 , further comprising:
 determining, by the one or more processors, based on applying the one or more machine learning algorithms to data related to each patient not included in the portion of the data, for each patient, a respective probability, wherein the respective probability is a numerical value indicating the percentage of commonality between the data related to the undiagnosed patient and the one or more patterns.   
     
     
         6 . The method of  claim 5 , further comprising:
 ranking, by the one or more processors, the probability and the respective probabilities, in order of relevance; and   notifying, by the one or more processors, through an electronic communication, a user of an identity of any patient in the one or more machine-readable data sets with a probability above a predetermined threshold; and   automatically ordering, by the one or more processors, based on communicating with an order management system over a network connection, a clinical test for the medical condition, wherein a number of tests ordered is directly proportional to a number of patients with the probability above the predetermined threshold   
     
     
         7 . The method of  claim 1 , wherein the generating the one or more patterns comprises:
 ranking, by the one or more processors, the common features based on the weighting; and   retaining, by the one or more processors, the portion of the common features wherein the portion comprises common features of a pre-defined weight, wherein the portion comprises the one or more patterns.   
     
     
         8 . The method of  claim 1 , wherein the mutual information comprises features from a plurality of feature categories and wherein each pattern of the one or more patterns comprising a portion of the common features comprises features in one feature category of the plurality of feature categories. 
     
     
         9 . The method of  claim 8 , wherein the medical condition is Hereditary Angioedema, and wherein the one feature category is selected from the group consisting of: diagnosis codes, procedures, therapies, providers, and locations. 
     
     
         10 . The method of  claim 9 , wherein the feature category is diagnosis codes and one of the features is selected from the group consisting of: an allergic reaction, a swelling, mass, or lump in head and neck, a routine general medical examination at a healthcare facility, an immunization and screening for an infectious disease, another screening for suspected conditions that are not mental disorders or infectious diseases, an edema, an abdominal pain at an unspecified site, another upper respiratory disease, an unspecified symptom associated with female genital organs, and a chronic vascular insufficiency of the intestine. 
     
     
         11 . The method of  claim 9 , wherein the feature category is procedures and one of the features is selected from the group consisting of: an office or other outpatient visit for the evaluation and management of an established patient, another laboratory procedure, an office or other outpatient visit for the evaluation and management of an established patient, a chemistry and hematology laboratory procedure, another therapeutic procedure, a pathology procedure, another diagnostic radiology and related technique, a microscopic examination, an office or other outpatient visit for evaluation and management of an established patient, and a nonoperative urinary system measurement. 
     
     
         12 . The method of  claim 9 , wherein the feature category is therapies and one of the features is selected from the group consisting of: androgens and combinations, blood derivatives, androgens and combinations, unspecified agents, sympathomimetic agents, adrenals and combinations, analgesics or antipyretics that are opiate agonists, antibiotics that are penicillins, antibiotics that are erythromycin and macrolide, and analgesics or antipyretics that are nonsteroidal anti-inflammatory drugs. 
     
     
         13 . The method of  claim 9 , wherein the feature category is providers and one of the features is selected from the group consisting of: an outpatient hospital, an office, an independent laboratory, an emergency department, an inpatient hospital, an independent clinic, a patient home, an outpatient location that is not elsewhere classified, an ambulatory surgical center; and a land ambulance. 
     
     
         14 . The method of  claim 1 , wherein the one or more machine learning algorithms comprise a linear Support Vector Machines classification algorithm. 
     
     
         15 . The method of  claim 1 , wherein the one or more machine learning algorithms comprise at least two machine learning algorithms and wherein the tuning further comprises:
 compiling results of the tuning of each of the at least two machine learning algorithms and utilizing ensemble learning to consolidate portions of the at least two machine learning algorithms into a single machine learning algorithm.   
     
     
         16 . The method of  claim 1 , the tuning further comprising:
 associating, by the one or more processors, based on applying the one or more machine learning algorithms to the training set of test data, probabilities to a portion of the records in the training set of test data, wherein the probabilities reflect a likelihood of presence of the medical condition for each record training set of test data; and   completing the dynamically adjusting of the common features when the probabilities are within a pre-defined accuracy threshold.   
     
     
         17 . The method of  claim 1 , wherein the determining the probability comprises:
 obtaining, by the one or more processors, from a computing resource, electronic medical records for the undiagnosed patient for a defined temporal period, wherein the electronic medical records comprise electronic contact information for a healthcare provider to the undiagnosed patient;   applying, by the one or more processors, the one or more machine learning algorithms to the electronic medical records;   determining, by the one or more processors, based on the applying, if the probability is within a predetermined range; and   based on determining that the probability exceeds a predetermined threshold, electronically alerting, in real time, the healthcare provider to the undiagnosed patient of the probability.   
     
     
         18 . The method of  claim 17 , further comprising:
 retaining, by the one or more processors, in a memory resource communicatively coupled to the one or more processors, the one or more patterns;   obtaining, by the one or more processors, an indication regarding accuracy of the probability; and   updating, by the one or more processors, the one or more patterns based on the indication.   
     
     
         19 . A computer program product comprising:
 a computer readable storage medium readable by one or more processors in a distributed computing environment, and storing instructions for execution by the one or more processors for performing a method comprising:
 obtaining, by the one or more processors in a distributed computing environment, one or more machine-readable data sets related to a patient population from one or more databases; 
 identifying, by the one or more processors, based on an initial patient definition, a portion of data from the machine-readable data sets related to a patient population, wherein the portion of the data comprises patients of the patient population with a medical condition; 
 based on a frequency of features in the portion of the data, identifying, by the one or more processors, common features in the portion of the data and weighting the common features based on frequency of occurrence in the portion of the data, wherein the common features comprise mutual information; 
 generating, by the one or more processors, one or more patterns comprising a portion of the common features; 
 generating, by the one or more processors, one or more machine learning algorithms based on the one or more patterns, the one or more machine learning algorithms to identify presence or absence of the given medical condition in an undiagnosed patient based on absence or presence of features comprising the one or more patterns in data related to the undiagnosed patient; 
 utilizing, by the one or more processors, statistical sampling to compile a training set of data, wherein the training set comprises data from the one or more data sets and at least one additional data set comprising data related to a population without the medical condition, and wherein utilizing the statistical sampling comprises formulating and obtaining queries based on the data set and processing and responding to the queries, the processing comprising, for each query:
 evaluating, by the one or more processors, the query to determine one of a high or a low level of anticipated complexity of a prospective response to the query; 
 based on the query being evaluated at a low level of anticipated complexity, assigning, by the one or more processors, the query to a computing resource in the distributed computing environment, wherein the computing resource is configured to respond to low level complexity queries; and 
 based on the query being evaluated at a high level of anticipated complexity, distributing, by the one or more processors, the query over a group of computing resources of the distributed computing environment to maximize efficiency, wherein the distributing comprises assigning each computing resource of the group of computing resources a portion of the query to execute in parallel with at least one other computing resource of the group of computing resources executing another portion of the query; 
 
 tuning, by the one or more processors, the one or more machine learning algorithms by applying the one or more machine learning algorithms to the training set of data; 
 dynamically adjusting, by the one or more processors, the common features comprising the one or more patterns to improve accuracy such that the one or more machine learning algorithms can distinguish patient data indicating the medical condition from patient data that does not indicate the medical condition; and 
 determining, by the one or more processors, based on applying the one or more machine learning algorithms to data related to the undiagnosed patient, a probability, wherein the probability is a numerical value indicating a percentage of commonality between the data related to the undiagnosed patient and the one or more patterns. 
   
     
     
         20 . A system comprising:
 one or more memory;   one or more processors in communication with the memory; and   program instructions executable by the one or more processors in a distributed computed environment via the one or more memory to perform a method, the method comprising:
 obtaining, by the one or more processors in a distributed computing environment, one or more machine-readable data sets related to a patient population from one or more databases; 
 identifying, by the one or more processors, based on an initial patient definition, a portion of data from the machine-readable data sets related to a patient population, wherein the portion of the data comprises patients of the patient population with a medical condition; 
 based on a frequency of features in the portion of the data, identifying, by the one or more processors, common features in the portion of the data and weighting the common features based on frequency of occurrence in the portion of the data, wherein the common features comprise mutual information; 
 generating, by the one or more processors, one or more patterns comprising a portion of the common features; 
 generating, by the one or more processors, one or more machine learning algorithms based on the one or more patterns, the one or more machine learning algorithms to identify presence or absence of the given medical condition in an undiagnosed patient based on absence or presence of features comprising the one or more patterns in data related to the undiagnosed patient; 
 utilizing, by the one or more processors, statistical sampling to compile a training set of data, wherein the training set comprises data from the one or more data sets and at least one additional data set comprising data related to a population without the medical condition, and wherein utilizing the statistical sampling comprises formulating and obtaining queries based on the data set and processing and responding to the queries, the processing comprising, for each query:
 evaluating, by the one or more processors, the query to determine one of a high or a low level of anticipated complexity of a prospective response to the query; 
 based on the query being evaluated at a low level of anticipated complexity, assigning, by the one or more processors, the query to a computing resource in the distributed computing environment, wherein the computing resource is configured to respond to low level complexity queries; and 
 based on the query being evaluated at a high level of anticipated complexity, distributing, by the one or more processors, the query over a group of computing resources of the distributed computing environment to maximize efficiency, wherein the distributing comprises assigning each computing resource of the group of computing resources a portion of the query to execute in parallel with at least one other computing resource of the group of computing resources executing another portion of the query; 
 
 tuning, by the one or more processors, the one or more machine learning algorithms by applying the one or more machine learning algorithms to the training set of data; 
 dynamically adjusting, by the one or more processors, the common features comprising the one or more patterns to improve accuracy such that the one or more machine learning algorithms can distinguish patient data indicating the medical condition from patient data that does not indicate the medical condition; and 
 determining, by the one or more processors, based on applying the one or more machine learning algorithms to data related to the undiagnosed patient, a probability, wherein the probability is a numerical value indicating a percentage of commonality between the data related to the undiagnosed patient and the one or more patterns.