Clinical trial matching system using inferred biomarker status
A model-assisted system for identifying a group of patients for a cohort using a generalized biomarker model may include a processor programmed to provide, to a generalized biomarker model, a first biomarker associated with a cohort, the generalized biomarker model being trained based on one or more second biomarkers; receive, from the generalized biomarker model, an output indicating a plurality of individuals with associated likelihoods of at least one of: having an attribute associated with the third biomarker or having been tested for the attribute associated with the first biomarker; determine a likelihood threshold based on a predetermined cohort size associated with the first biomarker and identify, based on the output, a group of the plurality of individuals for inclusion in a cohort, each individual in the group of the plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the likelihood threshold.
1. A model-assisted system, the system comprising:
at least one processor programmed to:
access a database from which information associated with a plurality of individuals can be derived, wherein the information includes a plurality of patient medical records associated with the plurality of individuals;
provide, to a generalized biomarker model, a first biomarker associated with a cohort, the generalized biomarker model being trained based on one or more second biomarkers to detect one or more documents associated with representations of the first biomarker or biomarker testing, wherein the first biomarker is different from the one or more second biomarkers;
receive, from the generalized biomarker model, an output indicating likelihoods, associated with the plurality of individuals, of at least one of: having an attribute associated with the first biomarker or having been tested for the attribute associated with the first biomarker;
determine a likelihood threshold based on a predetermined cohort size associated with the first biomarker; and
identify, based on the output, a group of the plurality of individuals for inclusion in the cohort, each individual in the group of the plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the likelihood threshold.
2. The model-assisted system of claim 1 , wherein the likelihood threshold is further determined based on a relative prevalence of the first biomarker.
3. The model-assisted system of claim 1 , wherein the likelihood threshold is further determined based on a likelihood the cohort will be filled over the course of a particular study.
4. The model-assisted system of claim 1 , wherein the predetermined cohort size is based on a total capacity of a healthcare provider.
5. The model assisted system of claim 1 , wherein at least one processor is further programmed to:
provide, to the generalized biomarker model, a third biomarker;
receive, from the generalized biomarker model, an additional output indicating an additional plurality of individuals with associated likelihoods of at least one of: having an additional attribute associated with the third biomarker or having been tested for the additional attribute associated with the third biomarker;
determine a third biomarker likelihood threshold based on a predetermined cohort size associated with the third biomarker; and
identify, based on the additional output, an additional group of the additional plurality of individuals for inclusion in the cohort, each individual in the additional group, of the additional plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the third biomarker likelihood threshold.
6. The model-assisted system of claim 1 , wherein the likelihood threshold is further determined based on the predetermined cohort size associated with the first biomarker and a total capacity for a healthcare facility.
7. The model-assisted system of claim 1 , wherein the at least one processor is further programmed to train the generalized biomarker model based on the one or more second biomarkers.
8. The model-assisted system of claim 7 , wherein training the generalized biomarker includes:
identifying representations of the one or more second biomarkers in a plurality of training medical records;
extracting snippets of text surrounding the representations of the one or more second biomarkers;
generating a plurality of feature vectors based on the snippets; and
training the generalized biomarker model to detect documents associated with the one or more second biomarkers based on the plurality of feature vectors.
9. The model-assisted system of claim 7 , wherein training the generalized biomarker model to detect documents associated with the one or more second biomarkers includes applying a logistic regression algorithm.
10. A model-assisted system, the system comprising:
at least one processor programmed to:
provide, to a generalized biomarker model, a first biomarker associated with a cohort, the generalized biomarker model being trained based on one or more second biomarkers to detect one or more documents associated with representations of the first biomarker or biomarker testing, wherein the first biomarker is different from the one or more second biomarkers;
receive, from the generalized biomarker model, an output indicating a plurality of individuals with associated likelihoods of at least one of: having an attribute associated with the first biomarker or having been tested for the attribute associated with the first biomarker;
determine a likelihood threshold based on a predetermined cohort size associated with the first biomarker, a relative prevalence of the biomarker, and a relative ability of a healthcare provider to perform an action based on the first biomarker; and
identify, based on the output, a group of the plurality of individuals for inclusion in the cohort, each individual in the group of the plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the likelihood threshold.
11. The model-assisted system of claim 10 , wherein the at least one processor is further programmed to access a database from which information associated with a population of individuals can be derived and wherein the generalized biomarker is configured to generate the output based on analyzing the information in association with the first biomarker.
12. The model-assisted system of claim 11 , wherein the information includes a plurality of patient medical records associated with the population of individuals.
13. A computer-implemented method for identifying a group of individuals for a cohort using a generalized biomarker model, the method comprising:
accessing a database from which information associated with a plurality of individuals can be derived, wherein the information includes a plurality of patient medical records associated with the plurality of individuals;
providing, to the generalized biomarker model, a first biomarker associated with a cohort, the generalized biomarker model being trained based on one or more second biomarkers to detect one or more documents associated with representations of the first biomarker or biomarker testing, wherein the first biomarker is different from the one or more second biomarkers;
receiving, from the generalized biomarker model, an output indicating likelihoods, associated with the plurality of individuals, of at least one of: having an attribute associated with the first biomarker or having been tested for the attribute associated with the first biomarker;
determining a likelihood threshold based on a predetermined cohort size associated with the first biomarker; and
identifying, based on the output, a group of the plurality of individuals for inclusion in the cohort, each individual in the group of the plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the likelihood threshold.
14. The computer-implemented method of claim 13 , wherein the likelihood threshold is further determined based on a relative prevalence of the first biomarker.
15. The computer-implemented method of claim 13 , wherein the likelihood threshold is further determined based on a relative ability of a healthcare provider to perform an action based on the first biomarker.
16. The computer-implemented method of claim 13 , wherein the likelihood threshold is further determined based on a likelihood the cohort will be filled over the course of a particular study.
17. The computer-implemented method of claim 13 , wherein the method further comprises training the generalized biomarker model based on the one or more second biomarkers.
18. The computer-implemented method of claim 17 , wherein training the generalized biomarker includes:
identifying representations of the one or more second biomarkers in a plurality of training medical records;
extracting snippets of text surrounding the representations of the one or more second biomarkers;
generating a plurality of feature vectors based on the snippets; and
training the generalized biomarker model to detect documents associated with the one or more second biomarkers based on the plurality of feature vectors.
19. The computer-implemented method of claim 17 , wherein training the generalized biomarker model to detect documents associated with the one or more second biomarkers includes applying a logistic regression algorithm.
20. A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to perform a method for identifying a group of individuals for a cohort using a generalized biomarker model, the method comprising:
accessing a database from which information associated with a plurality of individuals can be derived, wherein the information includes a plurality of patient medical records associated with the plurality of individuals;
providing, to the generalized biomarker model, a first biomarker associated with a cohort, the generalized biomarker model being trained based on one or more second biomarkers to detect one or more documents associated with representations of the first biomarker or biomarker testing, wherein the first biomarker is different from the one or more second biomarkers;
receiving, from the generalized biomarker model, an output indicating likelihoods, associated with the plurality of individuals, of at least one of: having an attribute associated with the first biomarker or having been tested for an attribute associated with the first biomarker;
determining a likelihood threshold based on a predetermined cohort size associated with the first biomarker; and
identifying, based on the output, a group of the plurality of individuals for inclusion in the cohort, each individual in the group of the plurality of individuals being associated with a likelihood received from the generalized biomarker model that satisfies the likelihood threshold.