Systems and methods for model-assisted cohort selection
Systems and methods are disclosed for selecting cohorts. In one implementation, a model-assisted selection system for identifying candidates for placement into a cohort includes a data interface and at least one processing device. The at least one processing device is programmed to access, via the data interface, a database from which feature vectors associated with an individual from among a population of individuals can be derived; derive, for the individual, one or more feature vectors from the database; provide the one or more feature vectors to a model; receive an output from the model; and determine whether the individual from among the population of individuals is a candidate for the cohort based on the output received from the model.
1. A model-assisted selection system for identifying candidates for placement into a cohort, the system comprising:
a data interface; and
at least one processing device programmed to:
access, via the data interface, a database from which feature vectors associated with an individual from among a population of individuals can be derived;
derive, for the individual, one or more feature vectors from the database;
provide the one or more feature vectors to a model;
receive an output from the model, the output comprising a confidence score for the individual; and
determine whether the individual from among the population of individuals is a candidate for the cohort based on the output received from the model, wherein the determination is based on a comparison of the confidence score to a predetermined threshold that is adjustable.
2. The model-assisted cohort selection system of claim 1 , wherein the database includes a plurality of electronic data representations, and the processing device is further programmed to:
upload the plurality of electronic data representations via the data interface; and
generate the one or more feature vectors using the plurality of electronic data representations.
3. The model-assisted cohort selection system of claim 2 , wherein the electronic data representations include electronic representations of documents from an electronic medical record associated with the individual.
4. The model-assisted cohort selection system of claim 2 , wherein the electronic data representations include at least some text previously subjected to an optical character recognition process.
5. The model-assisted cohort selection system of claim 2 , wherein the at least one processing device is further programmed to generate the one or more feature vectors by:
searching the plurality of electronic data representations for the presence of at least one term or phrase predetermined as associated with the cohort;
after identifying the at least one term or phrase as present in the plurality of electronic data representations, extracting a text grouping from the plurality of electronic data representations, wherein the text grouping includes one or more words located in a vicinity of the identified term or phrase; and
generating the one or more feature vectors based on analysis of the identified term or phrase together with analysis of the extracted text grouping.
6. The model-assisted selection system of claim 1 , wherein the model generates the output using a binary classification algorithm.
7. The model-assisted selection system of claim 6 , wherein the binary classification algorithm includes logistic regression.
8. The model-assisted selection system of claim 1 , wherein the machine learning model has been trained based on a set of structured information extracted by a combination of humans and machines from unstructured information, including a medical record.
9. The model-assisted selection system of claim 1 , wherein the plurality of electronic data representations are derived from at least one of an electronic medical record, an available data source, claims data, or patient-reported data associated with the at least one individual.
10. The model-assisted selection system of claim 1 , wherein the predetermined threshold is adjustable based on levels of efficiency and performance, of the model.
11. The model-assisted selection system of claim 1 , wherein the cohort is to include individuals all sharing at least one medical or demographic characteristic.
12. The model-assisted selection system of claim 1 , wherein the plurality of electronic data representations includes both structured data and unstructured data.
13. The model-assisted selection system of claim 1 , wherein the model includes a trained machine learning model.
14. The model-assisted selection system of claim 1 , wherein the model includes a rules-based model.
15. The model-assisted selection system of claim 1 , wherein the rules-based model generates output by matching a pre-defined set of search terms.
16. A method for selecting a cohort from among a population of individuals, the method comprising:
accessing, via a data interface, a database from which feature vectors associated with an individual from among a population of individuals can be derived;
deriving, for the individual, one or more feature vectors from the database;
providing the one or more feature vectors to a model;
receiving an output from the model, the output comprising a confidence score for the individual; and
determining whether the individual from among the population of individuals is a candidate for the cohort based on the output received from the model, wherein the determination is based on a comparison of the confidence score to a predetermined threshold that is adjustable.
17. The cohort selection method of claim 16 , wherein the database includes a plurality of electronic data representations, and the method further includes:
uploading the plurality of electronic data representations via the data interface; and
generating the one or more feature vectors using the plurality of electronic data representations.
18. The cohort selection method of claim 17 , wherein the electronic data representations include electronic representations of documents from an electronic medical record associated with the individual.
19. The cohort selection method of claim 17 , wherein the electronic data representations include at least some text previously subjected to an optical character recognition process.
20. The cohort selection method of claim 17 , further including:
generating the one or more feature vectors by:
searching the plurality of electronic data representations for the presence of at least one term or phrase predetermined as associated with the cohort;
after identifying the at least one term or phrase as present in the plurality of electronic data representations, extracting a text grouping from the plurality of electronic data representations, wherein the text grouping includes one or more words located in a vicinity of the identified term or phrase; and
generating the one or more feature vectors based on analysis of the identified term or phrase together with analysis of the extracted text grouping.
21. The cohort selection method of claim 16 , wherein the machine learning model generates the output using a logistic regression technique.
22. The cohort selection method of claim 16 , wherein the machine learning model has been trained based on a set of structured information extracted by a combination of humans and machines from unstructured information, including a medical record.
23. The cohort selection method of claim 16 , wherein the plurality of electronic data representations are derived from at least one of an electronic medical record, an available data source, claims data, or patient-reported data associated with the at least one individual.
24. The cohort selection method of claim 16 , wherein the cohort is to include individuals all sharing at least one medical or demographic characteristic.
25. The cohort selection method of claim 16 , wherein the plurality of electronic data representations includes both structured data and unstructured data.
26. The cohort selection method of claim 16 , wherein the model includes a trained machine learning model or a rules-based model.
27. The system of claim 1 , wherein the threshold is adjusted based on at least one loss function.
28. The system of claim 1 , wherein the threshold is adjusted such that a sensitivity level of the model is at least 95%.
29. The system of claim 1 , wherein the threshold is adjusted such that an efficiency of the model is at least 50%.