System and method for weighting manageable patient attributes during criteria evaluations for treatment

• US 10,431,338 B2
• Filed: 04/21/2015
• Issued: 10/01/2019
• Est. Priority Date: 02/25/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Allen, Corville O., Payton, Aspen L.
• Primary Examiner(s): Durant, Jonathan
• Application Number:
  US14/692,504
• Publication Number:
  US 20160246945A1
• Time in Patent Office: 1,624 Days
• CPC Class Codes: G16H 10/20 for electronic clinical trials or questionnaires
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 20/40 relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 70/20 relating to practices or guidelines

Abstract

System, method, and computer program product for evaluating attribute values to determine treatment eligibility, the method by receiving a set of required attributes associated with a treatment protocol, receiving a case, wherein the case includes a patient history containing patient attribute values, identifying a patient attribute value that does not satisfy a required attribute specified by the treatment protocol, determining a likelihood that the patient could meet the required attribute based upon the patient history, and providing an indication of the likelihood that the patient could satisfy the required attribute specified by the treatment protocol.

Claims

1. A computer-implemented method for evaluating attribute values to determine eligibility for a first treatment, the computer-implemented method comprising:

responsive to receiving input regarding a first patient, determining a topic based on parsing the received input using natural language processing, the topic comprising a request to identify treatment protocols for which the first patient can be eligible, wherein the topic is determined by a question classifier component of an application and stored in a feature store;

determining, based on a corpus and by a pipeline execution component of the application, a set of required attributes associated with a treatment protocol associated with the first treatment, wherein the pipeline execution component reduces processing of sources in the corpus that do not pertain to the stored topic;

responsive to receiving a case, wherein the case includes a patient history containing patient attribute values for the first patient, identifying, by an attribute verification component of the application, a first patient attribute value that does not satisfy a first attribute specified by the treatment protocol as a required attribute;

determining that the first patient is not currently eligible to receive the first treatment, based on determining that the first patient attribute value does not satisfy the first attribute specified by the treatment protocol;

upon determining that the first patient is not currently eligible to receive the first treatment, determining, by operation of one or more computer processors and based on evaluating the patient history using one or more machine learning models, a likelihood that the patient will meet the first attribute in the future due to a change in the first patient attribute value, thereby determining potential eligibility of the first patient for the treatment protocol with a greater measure of accuracy than absent determining the likelihood, wherein the likelihood is determined based on an upward or downward trend of the first patient attribute value and based further on an extent by which the first patient attribute value differs from the required attribute; and

causing approval, based on the determined likelihood, of the first patient for receiving the treatment protocol, whereafter the treatment protocol is applied to the first patient, wherein absent determining the likelihood, the first patient would have been denied from receiving the treatment protocol.

2. The computer-implemented method of claim 1, further comprising evaluating the patient history using natural language processing to process the case and identify attribute values of the patient.

3. The computer-implemented method of claim 1, wherein determining the likelihood comprises determining whether the first attribute is mutable or manageable and returning an indication that the patient is ineligible to receive the first treatment if the first attribute is either immutable or unmanageable.

4. The computer-implemented method of claim 1, wherein determining the likelihood that the patient could satisfy the first attribute value is further based on comparing at least one attribute that is specified as required by the treatment protocol, and patient attribute values from the patient history.

5. The computer-implemented method of claim 1, wherein determining the likelihood is further based on determining whether the first patient attribute value in the patient history has ever satisfied the first attribute at a moment in time in the past.

6. The computer-implemented method of claim 1, wherein determining the likelihood is further based on determining whether the patient history indicates that the patient has previously undergone treatment that could change the value of the first patient attribute value, and any change in the first patient attribute value in response to the treatment.

7. The computer-implemented method of claim 6, wherein determining the likelihood is further based on determining whether the treatment is associated with a side-effect that may impact a second patient attribute value that corresponds to a second attribute specified by the treatment protocol as a required attribute, wherein the second patient attribute value currently satisfies the second attribute and the side-effect may cause the second patient attribute value to no longer satisfy the second attribute.

8. The computer-implemented method of claim 1, wherein the application is associated with an ontology and the one or more machine learning models, and wherein determining the likelihood that the patient could meet the first attribute in the future comprises processing the patient history using the one or more machine learning models.

9. The computer-implemented method of claim 8, wherein the at least a first of the one or more machine learning models is trained to generate a measure of confidence in the determined likelihood.

10. The computer-implemented method of claim 1, wherein determining the likelihood comprises determining whether the first attribute is mutable or manageable and returning an indication that the patient is ineligible to receive the first treatment if the first attribute is either immutable or unmanageable, wherein determining whether the first attribute is mutable or manageable comprises processing at least a first feature of the first attribute using the one or more machine learning models.

11. The computer-implemented method of claim 1, wherein the attribute verification component identifies, from medical records of the first patient, medical information relevant to determining the potential eligibility of the first patient for the treatment protocol;

wherein the application is associated with a plurality of data structures including the corpus, the feature store, an evidence classification, and the one or more machine learning models, wherein the corpus includes an ontology component, wherein the evidence classification specifies relationships between the corpus and the feature store;

wherein determining the likelihood that the patient could meet the first attribute in the future comprises processing the patient history using the one or more machine learning models, wherein the at least a first of the one or more machine learning models is trained to generate a measure of confidence in the determined likelihood.

12. The computer-implemented method of claim 11, wherein the likelihood is determined using a trained machine learning model of how mutable the first attribute is for a given patient over time, wherein an indication that the patient is ineligible to receive the first treatment is returned upon determining that the first attribute is immutable;

wherein determining the likelihood is further based on determining whether the first patient attribute value in the patient history has ever satisfied the first attribute at a moment in time in the past.

13. The computer-implemented method of claim 12, wherein determining the likelihood is further based on determining whether the patient history indicates that the patient has previously undergone treatment that could change the value of the first patient attribute value, and any change in the first patient attribute value in response to the treatment;

wherein determining the likelihood is further based on determining whether the treatment is associated with a side-effect that may impact a second patient attribute value that corresponds to a second attribute specified by the treatment protocol as a required attribute, wherein the second patient attribute value currently satisfies the second attribute and the side-effect may cause the second patient attribute value to no longer satisfy the second attribute.

14. The computer-implemented method of claim 13, wherein the computer-implemented method further comprises training the one or more machine learning models in order to ascertain which features of attributes are more accurate in predicting a correct answer for different questions, relative to other features of the attributes, whereafter the one or more machine learning models are used with one or more scoring algorithms in order to generate one or more feature scores to indicate a measure of confidence for each candidate answer to the specified question, wherein the potential eligibility of the first patient for the treatment protocol is determined based on the measure of confidence for each candidate answer to the specified question.

15. The computer-implemented method of claim 14, wherein the treatment protocol specifies a plurality of required attributes including the required attribute, the plurality of required attributes including patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition, respectively.

16. The computer-implemented method of claim 1, wherein the required attribute is selected from patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition.

17. The computer-implemented method of claim 1, wherein the treatment protocol specifies a plurality of required attributes including the required attribute, the plurality of required attributes including at least two attributes selected from patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition.

18. The computer-implemented method of claim 1, wherein the treatment protocol specifies a plurality of required attributes including the required attribute, the plurality of required attributes including at least three attributes selected from patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition.

19. The computer-implemented method of claim 1, wherein the treatment protocol specifies a plurality of required attributes including the required attribute, the plurality of required attributes including at least four attributes selected from patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition.

20. The computer-implemented method of claim 1, wherein the treatment protocol specifies a plurality of required attributes including the required attribute, the plurality of required attributes including at least five attributes selected from patient age, patient blood pressure level, patient body weight, patient willingness to travel to obtain treatment, patient willingness to consent to medical procedures, and a count of treatment types previously tried by a given patient for a given medical condition.