Method and system for intelligent completion of medical record based on big data analytics

• US 10,546,654 B2
• Filed: 12/17/2015
• Issued: 01/28/2020
• Est. Priority Date: 12/17/2015
• Status: Active Grant

• Current Assignee: DrFirst.com, Incorported
• Original Assignee: DrFirst.com, Incorported
• Inventors: Sellars, David Andrew
• Primary Examiner(s): Holcomb, Mark
• Application Number:
  US14/972,209
• Publication Number:
  US 20170177796A1
• Time in Patent Office: 1,503 Days
• Field of Search: 705 2- 3
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 20/10 relating to drugs or medications, e.g. for ensuring correct administration to patients
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

The present teaching relates to medical record completion. In one example, a medical record of a patient is received. The medical record is associated with a plurality of components comprising a first component with a populated value and a second component with an unpopulated value. The unpopulated value of the second component is estimated based on the populated value of the first component in accordance with a first model. Information associated with the medical record and/or the patient is obtained. The values of the first and second components are validated based on the obtained information in accordance with a second model. The first and second models are dynamically updated based on data related to medical transactions of a plurality of patients.

Claims

1. A method, implemented on at least one computing device each of which has at least one processor, storage, and a communication platform connected to a network for completing a medical record, the method comprising:

tracking medical transaction data in a large general population of patients to generate a data map, the data map pairing each of a plurality of medical suggestions with one or more analytic influence dimensions, each medical suggestion representing a medical transaction recommended by a medical professional and each analytic influence dimension in the data map specifying an attribute of a patient or an attribute of a physician associated with the medical transaction data, and each dimension-medical suggestion pair in the data map having a respective confidence score indicative of a degree of match between the medical suggestion and the analytic influence dimension over the large general population of patients, wherein tracking the medical transaction data includes dynamically updating the data map via analysis of new medical transaction data from the large general population of patients;

receiving a medical record of a patient, wherein the medical record is associated with a plurality of components comprising a first component with a populated value and a second component with an unpopulated value;

estimating a value for the second component based on the populated value of the first component in accordance with a first model wherein the first model is dynamically updated based on data related to the new medical transaction data for the large general population of patients;

identifying one or more analytic influence dimensions for the medical record, each of the one or more analytic influence dimensions specifying an attribute of the patient or an attribute of a physician associated with the medical record;

obtaining, from the data map, relevant dimension-medical suggestion pairs, the relevant dimension-medical suggestion pairs having an analytic influence dimension in the data map that matches one of the one or more analytic influence dimensions for the medical record;

selecting a plurality of highest-ranked relevant dimension-medical suggestion pairs based on the respective confidence scores from the data map;

determining whether a component with a discrepancy exists by comparing the value of each of the plurality of components with a value of a corresponding component from the plurality of highest-ranked relevant dimension-medical suggestion pairs to determine a discrepancy; and

responsive to identifying a component with a discrepancy and to determining that the discrepancy satisfies a threshold;

receiving a corrected value for the component with the discrepancy, and

updating the medical record with the corrected value.

2. The method of claim 1, further comprising:

normalizing the medical record to generate a normalized medical record having the plurality of components comprising the first component with the populated value and the second component with the unpopulated value.

3. The method of claim 2, further comprising:

obtaining a medical record format associated with a user;

converting the normalized medical record into the medical record format; and

sending the converted normalized medical record to the user.

4. The method of claim 1, wherein the first model includes an inference model with rules for calculating the estimated value based on one or more values of populated components in the medical record.

5. The method of claim 1, wherein the first model includes a statistics model generated from the medical transaction data, the statistics model including, for each of a plurality of components in a medical record, a ranking representing a degree of match between the component and a normalized medical record.

6. The method of claim 1, wherein the degree of match between the medical suggestion and the analytic influence dimension in the data map is based on occurrences of the dimension-medical suggestion pairs in the tracked medical transaction data.

7. The method of claim 2, wherein the medical record is normalized based on a second model that is dynamically updated based on data related to medical transactions of the large general population of patients.

8. The method of claim 7, wherein the second model includes a mapping from a plurality of permutations to a term.

9. The method of claim 1, wherein an analytic influence dimension specifying an attribute of the patient includes at least one of:

disease diagnosis;

symptoms;

or

patient profile.

10. The method of claim 1, wherein the medical record is in a natural language.

21. The method of claim 1, wherein the medical transactions recommended by a medical professional include at least one of a medication drug prescription, a physical therapy referral, a diet recommendation, or a medical test.

11. A system for completing a medical record, comprising:

at least one processor; and

memory storing instructions that, when executed by the at least one processor, cause the system to;

generate and dynamically update a data map by analyzing medical transaction data in a large general population of patients, the data map pairing each of a plurality of medical suggestions from the medical transaction data with one or more analytic influence dimensions, each medical suggestion representing a medical transaction recommended by a medical professional and each analytic influence dimension in the data map specifying an attribute of a patient or an attribute of a physician associated with the medical transaction data, and each dimension-medical suggestion pair in the data map having a respective confidence score indicative of a degree of match between the medical suggestion and the analytic influence dimension over the large general population of patients, wherein dynamically updating the data map occurs via analysis of new medical transaction data for the large general population of patients,

receive a medical record of a patient, wherein the medical record is associated with a plurality of components comprising a first component with a populated value and a second component with an unpopulated value, and estimate a value for the second component based on the populated value of the first component in accordance with a first model wherein the first model is dynamically updated based on data related to the new medical transaction data for the large general population of patients,

identify one or more analytic influence dimensions for the medical record, each of the one or more analytic influence dimensions specifying an attribute of the patient or an attribute of a physician associated with the medical record,

obtain, from the data map, relevant dimension-medical suggestion pairs, the relevant dimension-medical suggestion pairs having an analytic influence dimension in the data map that matches one of the one or more analytic influence dimensions for the medical record,

select a plurality of highest-ranked relevant dimension-medical suggestion pairs based on the respective confidence scores in the data map,

determine whether a component with a discrepancy exists by comparing the value of each of the plurality of components with a value of a corresponding component from the plurality of highest-ranked relevant dimension-medical suggestion pairs to determine a discrepancy, and

responsive to identifying a component with a discrepancy and to determining that the discrepancy satisfies a threshold, receive a corrected value for the component with the discrepancy and update the medical record with the corrected value.

12. The system of claim 11, the instructions further causing the system to:

normalize the medical record to generate a normalized medical record having the plurality of components comprising the first component with the populated value and the second component with the unpopulated value.

13. The system of claim 12, the instructions further causing the system to:

obtain a medical record format associated with a user;

convert the normalized medical record into the medical record format; and

send the converted normalized medical record to the user.

14. The system of claim 11, wherein the first model includes an inference model with rules for calculating the estimated value based on one or more values of populated components in the medical record.

15. The system of claim 11, wherein the first model includes a plurality of medical records, each of which being associated with rankings with respect to the plurality of components.

16. The system of claim 11, wherein the degree of match between the medical suggestion and the analytic influence dimension in the data map is based on occurrences of the dimension-medical suggestion pairs in the analyzed medical transaction data.

17. The system of claim 12, wherein the medical record is normalized based on a second model that is dynamically updated based on data related to medical transactions of the large general population of patients.

18. The system of claim 17, wherein the second model includes a mapping from a plurality of permutations to a term.

19. The system of claim 11, wherein an analytic influence dimension specifying an attribute of the patient includes at least one of:

disease diagnosis;

symptoms;

or

patient profile.

20. A non-transitory machine readable medium having information recorded thereon for completing a medical record, wherein the information, when read by a machine, causes the machine to perform the steps of:

analyzing medical transaction data in a large general population of patients to generate and dynamically update a data map, the data map pairing each of a plurality of medical suggestions with one or more analytic influence dimensions, each medical suggestion representing a medical transaction recommended by a medical professional and each analytic influence dimension in the data map specifying an attribute of a patient or an attribute of a physician associated with the medical transaction data, and each dimension-medical suggestion pair in the data map having a respective confidence score indicative of a degree of match between the medical suggestion and the analytic influence dimension in the medical transaction data over the large general population of patients, wherein dynamically updating the data map occurs via analysis of new medical transaction data for the large general population of patients;

receiving a medical record of a patient, wherein the medical record is associated with a plurality of components comprising a first component with a populated value and a second component with an unpopulated value;

estimating a value for the second component based on the populated value of the first component in accordance with a first model, wherein the first model is dynamically updated based on data related to the new medical transaction data for the large general population of patients;

identifying one or more analytic influence dimensions for the medical record, each of the one or more analytic influence dimensions specifying an attribute of the patient or an attribute of a physician associated with the medical record;

obtaining, from the data map, relevant dimension-medical suggestion pairs, the relevant dimension-medical suggestion pairs having an analytic influence dimension in the data map that matches one of the one or more analytic influence dimensions for the medical record;

selecting a plurality of highest-ranked relevant dimension-medical suggestion pairs based on the respective confidence scores in the data map;

determining whether a component with a discrepancy exists by comparing the value of each of the plurality of components with a value of a corresponding component from the plurality of highest-ranked relevant dimension-medical suggestion pairs to determine a discrepancy; and

responsive to identifying a component with a discrepancy and to determining that the discrepancy satisfies a threshold;

receiving a corrected value for the component with the discrepancy, and

updating the medical record with the corrected value.