Method and apparatus for grouping records based upon a prediction of the content of the records

1. A method for grouping records based upon a prediction of content of the records, the method comprising:

training a machine learning model using a training data set including features from a plurality of categories, the plurality of categories comprising principle and secondary diagnoses, procedures, and type of provider;

receiving data associated with respective subjects of the records;

inputting the data, but not content of the records, into the machine learning model that has been trained for a prediction as to whether respective records support addition of a code;

in response to the prediction attributable to analyzing the data, but not the records, by the machine learning model, separating, using the machine learning model, the records into first and second groups with the first group including the records that the associated data indicates are more likely to support addition of the code and the second group including the records that the associated data indicates are less likely to support the addition of the code; and

providing for subsequent processing of the records with the subsequent processing differing dependent upon whether the records are included in the first group of records or the second groups of records,

wherein training the machine learning model comprises adjusting one or more thresholds utilized by the machine learning model in conjunction with the prediction as to whether respective records support the addition of the code in order to satisfy an accuracy requirement for categorization of the records.

2. A method according to claim 1 wherein providing for subsequent processing of the records comprises assigning, by audit assignment circuitry, different audit responsibility for the records depending upon whether the records are included in the first group of records or the second groups of records.

3. A method according to claim 2 wherein assigning different audit responsibility comprises assigning audit responsibility for the first and second groups of records to first and second auditors, respectively.

4. A method according to claim 2 further comprising:

receiving results from separately auditing the first and second groups of records; and

in response to the results from the auditing, modifying, by record modification circuitry, at least one of the records to be associated with the code.

5. A method according to claim 1 further comprising:

splitting the records into multiple partitions across a distributed computing framework; and

placing the machine learning model on each of a plurality of nodes of the distributed computing framework such that the machine learning model on a respective node analyzes records in a respective partition.

6. A method according to claim 1 wherein the records comprise medical records and the code comprises a Hierarchical Condition Category (HCC).

7. A method according to claim 6 wherein the data comprises payor data from one or more health care providers that includes one or more of claims information or summarized encounter information.

8. A method according to claim 6 further comprising excluding, by filtration circuitry, Medicare records from consideration prior to inputting the data into the machine learning model.

9. An apparatus, comprising:

a processor; and

a memory coupled to the processor and comprising computer readable program code embodied in the memory that is executable by the processor to perform operations comprising:

training a machine learning model using a training data set including features from a plurality of categories, the plurality of categories comprising principle and secondary diagnoses, procedures, and type of provider;

receiving data associated with respective subjects of the records;

inputting the data, but not content of the records, into the machine learning model that has been trained for a prediction as to whether respective records support addition of a code;

in response to the prediction attributable to analyzing the data, but not the records, by the machine learning model, separating, using the machine learning model, the records into first and second groups with the first group including the records that the associated data indicates are more likely to support addition of the code and the second group including the records that the associated data indicates are less likely to support the addition of the code; and

providing for subsequent processing of the records with the subsequent processing differing dependent upon whether the records are included in the first group of records or the second groups of records,

wherein training the machine learning model comprises adjusting one or more thresholds utilized by the machine learning model in conjunction with the prediction as to whether respective records support the addition of the code in order to satisfy an accuracy requirement for categorization of the records.

10. An apparatus according to claim 9 wherein providing for subsequent processing of the records comprises assigning different audit responsibility for the records depending upon whether the records are included in the first group of records or the second groups of records.

11. An apparatus according to claim 10 wherein assigning different audit responsibility comprises assigning different audit responsibility by assigning audit responsibility for the first and second groups of records to first and second auditors, respectively.

12. An apparatus according to claim 10 wherein the operations further comprise:

receiving results from separately auditing the first and second groups of records; and

in response to the results from the auditing, modifying at least one of the records to be associated with the code.

13. An apparatus according to claim 9 wherein the apparatus is embodied by a distributed computing framework, and wherein the operations further comprise:

splitting the records into multiple partitions across the distributed computing framework, and wherein the machine learning model is placed on each of a plurality of nodes of the distributed computing framework such that the machine learning model on a respective node analyzes records in a respective partition.

14. An apparatus according to claim 9 wherein the records comprise medical records and the code comprises a Hierarchical Condition Category (HCC).

15. An apparatus according to claim 14 wherein the data comprises payor data from one or more health care providers that includes one or more of claims information or summarized encounter information.

16. An apparatus according to claim 14 wherein the operations further comprise:

excluding Medicare records from consideration prior to inputting the data into the machine learning model.

17. A computer program product comprising at least one non-transitory computer-readable storage medium for grouping records based upon a prediction of content of the records, the at least one non-transitory computer-readable storage medium storing software instructions that, when executed, cause an apparatus to perform operations comprising:

training a machine learning model using a training data set including features from a plurality of categories, the plurality of categories comprising principle and secondary diagnoses, procedures, and type of provider;

receiving data associated with respective subjects of the records;

inputting the data, but not content of the records, into the machine learning model that has been trained for a prediction as to whether respective records support addition of a code;

in response to the prediction attributable to analyzing the data, but not the records, by the machine learning model, separating, using the machine learning model, the records into first and second groups with the first group including the records that the associated data indicates are more likely to support addition of the code and the second group including the records that the associated data indicates are less likely to support the addition of the code; and

providing for subsequent processing of the records with the subsequent processing differing dependent upon whether the records are included in the first group of records or the second groups of records,

wherein training the machine learning model comprises adjusting one or more thresholds utilized by the machine learning model in conjunction with the prediction as to whether respective records support the addition of the code in order to satisfy an accuracy requirement for categorization of the records.

18. A computer program product according to claim 17 wherein providing for subsequent processing of the records comprises assigning different audit responsibility for the records depending upon whether the records are included in the first group of records or the second groups of records.

19. A computer program product according to claim 18 wherein the operations further comprise:

receiving results from separately auditing the first and second groups of records; and

in response to the results from the auditing, at least one of the records to include the code.

20. A computer program product according to claim 17 wherein the operations further comprise:

splitting the records into multiple partitions across a distributed computing framework; and

placing the machine learning model on each of a plurality of nodes of the distributed computing framework such that the machine learning model on a respective node analyzes records in a respective partition.