Machine learning peer review system

1 . A system for peer review, comprising:

a processor;

a storage configured to receive a request from a medical provider, said request comprising identified medical data that is uniquely associated with a patient and that is uniquely associated with said medical provider;

an anonymizer configured to receive said identified medical data from said storage, generate anonymized medical data by removing any association with said patient and any association with said medical provider from said identified medical data, resulting in said anonymized medical data, and provide said anonymized medical data to a plurality of reviewers;

a score generator configured to:

determine whether a current score from previous assessments of said plurality of reviewers is below a pre-defined threshold score;

in response to said current score being below said pre-defined threshold score, receive, from said plurality of reviewers, a plurality of assessments corresponding to said anonymized medical data, update said current score based on the plurality of assessments, and determine whether criteria to finalize said current score has been met;

in response to said current score not being below said pre-defined threshold score or said criteria having been met, provide said plurality of assessments of said anonymized medical data as input to a model trained on a plurality of previous treatments and a corresponding plurality of previous scores associated with said plurality of previous treatments;

execute said model using said input;

generate a score from said executed model, wherein said score represents a quality of a treatment of said patient by said medical provider; and

determine a region of a plurality of regions of a numerical scale that said score is within;

a report generator configured to receive said score from said score generator and display a report comprising said score and said determined region to said medical provider; and

a non-transitory computer readable medium storing a set of instructions, which when executed by said processor, configure said anonymizer, said score generator, and said report generator.

2 . The system of claim 1 , wherein said anonymizer is further configured to normalize said identified medical data, wherein normalizing said identified medical data comprises transforming said identified medical data to a particular format.

3 . The system of claim 1 , wherein said report generator is further configured to receive said plurality of assessments, and said report further comprises said plurality of assessments.

4 . The system of claim 1 , wherein said anonymizer is an anonymizer service.

5 . The system of claim 1 , wherein said storage is a first storage, said system further comprising:

a second storage adapted for storing identified medical data; and

a parser configured to:

extract said identified medical data from said first storage;

transform said identified medical data to a particular format; and

store said transformed identified medical data in said second storage,

wherein said anonymizer is further configured to receive said transformed identified medical data from said second storage, and said set of instructions, when executed by said processor, configure said parser.

6 . The system of claim 1 , wherein said storage is a first storage, wherein said system further comprises a second storage adapted for storing said anonymized medical data, wherein said anonymizer is further configured to store said anonymized medical data in said second storage, and wherein said second storage is configured to provide said anonymized medical data to said plurality of reviewers.

7 . The system of claim 1 , wherein said report is an anonymized report that does not identify any of said plurality of reviewers.

8 . The system of claim 1 , wherein said identified medical data further comprises at least one of a date, a case number, and an account identifier.

9 . The system of claim 8 , wherein said anonymizer is further configured to generate said anonymized medical data by at least one of replacing said account identifier with an anonymized account identifier and replacing said case number with an anonymized case number.

10 . The system of claim 1 , wherein said identified medical data comprises a description of said treatment and said plurality of assessments comprise a plurality of assessments of said quality of said treatment based on said description.

11 . The system of claim 10 , wherein said description of said treatment comprises a description of a medical condition of said patient, said treatment is for said medical condition, and said report further comprises a comparison between said treatment and an optimal treatment for said medical condition.

12 . The system of claim 1 , further comprising a validator configured to receive said request from said medical provider, validate said request to determine whether said identified medical data is complete, and, based on a determination that said identified medical data is complete, provide said request to said storage, wherein said storage is further configured to receive said request from said validator, and said set of instructions, when executed by said processor, configure said validator.

13 . The system of claim 12 , wherein said validator is further configured to receive said anonymized medical data from said anonymizer, validate said anonymized medical data to determine whether said anonymized medical data is complete, and, based on a determination that said anonymized medical data is complete, provide said anonymized medical data to said plurality of reviewers, wherein said anonymizer is further configured to provide said anonymized medical data to said validator.

14 . The system of claim 1 , wherein said identified medical data is uniquely associated with said patient using a patient identifier, and said anonymizer is further configured to remove all occurrences of said patient identifier from said identified medical data.

15 . The system of claim 14 , wherein removing all occurrences of said patient identifier comprises replacing said patient identifier with an anonymized patient identifier.

16 . The system of claim 1 , wherein said medical provider is a person that is capable of providing treatment to said patient, said identified medical data is uniquely associated with said medical provider using a provider identifier, and said anonymizer is further configured to remove all occurrences of said provider identifier from said identified medical data.

17 . The system of claim 16 , wherein removing all occurrences of said provider identifier comprises replacing said provider identifier with an anonymized provider identifier.

18 . The system of claim 1 , wherein said identified medical data is uniquely associated with a medical organization, and said anonymizer is further configured to remove any association with said medical organization.

19 . The system of claim 18 , wherein said identified medical data is uniquely associated with said medical organization using an organization identifier, and said anonymizer is further configured to remove all occurrences of said organization identifier from said identified medical data.

20 . The system of claim 1 , wherein said score generator is further configured to combine said plurality of assessments into an aggregate assessment and generate said score based on said aggregate assessment.

21 . The system of claim 1 , wherein said criteria comprises one or more of:

a predefined maximum number of reviewers; and

a number of iterations for comparing said current score to said pre-defined threshold score.

22 . The system of claim 21 , wherein said plurality of reviewers is a first plurality of reviewers, said score is a first score, and said plurality of assessments is a first plurality of assessments, and said score generator is further configured to:

define a threshold score, wherein a score above said threshold score indicates said quality of said treatment is a valid treatment and a score below said threshold score indicates said quality of said treatment is an invalid treatment;

make a determination that said first score is below said threshold score;

based on said determination, direct said anonymizer to provide said anonymized medical data to a second plurality of reviewers;

receive, from said second plurality of reviewers, a second plurality of assessments of said anonymized medical data; and

update said first score based on said first plurality of assessments and said second plurality of assessments.

23 . The system of claim 22 , wherein said score generator is further configured to generate a plurality of initial scores, each corresponding to one of said plurality of assessments, and combine said plurality of initial scores into said score, wherein said plurality of initial scores is a first plurality of initial scores, and wherein said score generator is further configured to:

generate a second plurality of initial scores, each corresponding to one of said second plurality of assessments;

combine said first plurality of initial scores and said second plurality of initial scores into a second score; and

replace said first score with said second score.

24 . The system of claim 1 , wherein said score generator is further configured to:

receive training data, said training data comprising said generated score and a plurality of descriptions of said plurality of previous treatments of a plurality of previous patients, and further comprising a corresponding plurality of previous scores, wherein said corresponding plurality of previous scores represent a quality of said plurality of previous treatments;

re-train said model using said training data, resulting in a re-trained model; and

provide said plurality of assessments to said re-trained model to generate a new score.