Early prediction of clinical trial signals

A system for detecting signals related to subjects participating in a clinical trial, the system comprising:

• at least one processor;
• and
• a memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
• collecting clinical data for patients from a plurality of sources, the clinical data from each source having a respective data format;
• standardizing the clinical data from the plurality of sources in a common format and redacting personally identifiable information from the clinical data;
• storing the standardized and redacted clinical data in a data repository;
• determining one or more predictive features, wherein each predictive feature corresponds to one or more characteristics of the stored clinical data that correlate with a safety signal for the patients;
• obtaining, from the stored clinical data, training data including one or more predictive features and associated safety signal;
• using the training data to train one or more machine learning models to predict a predicted safety signal for a clinical trial subject based on clinical data for the clinical trial subject;
• receiving clinical data for a particular clinical trial subject enrolled in an ongoing clinical trial;
• by the trained machine learning models, predicting a particular safety signal for the particular clinical trial subject based on one or more predictive features determined from the received clinical data for the particular clinical trial subject;
• receiving, through a user interface, a review of the particular safety signal;
• and
• updating the one or more trained machine learning models based on the review, including (i) re-training one or more of the trained machine learning models using updated training data, (ii) re-determining one or more of the predictive features based on the review, and (iii) re-training one or more of the trained machine learning models with an updated set of corresponding hyperparameters.

One or more non-transitory computer readable media storing instructions that, when executed by at least one processor, cause the at least one processor to detect signals related to subjects participating in a clinical trial by performing operations comprising:

• collecting clinical data for patients from a plurality of sources, the clinical data from each source having a respective data format;
• standardizing the clinical data from the plurality of sources in a common format and redacting personally identifiable information from the clinical data;
• storing the standardized and redacted clinical data in a data repository;
• determining one or more predictive features, wherein each predictive feature corresponds to one or more characteristics of the stored clinical data that correlate with a safety signal for the patients;
• obtaining, from the stored clinical data, training data including one or more predictive features and associated safety signal;
• using the training data to train one or more machine learning models to predict a predicted safety signal for a clinical trial subject based on clinical data for the clinical trial subject;
• receiving clinical data for a particular clinical trial subject enrolled in an ongoing clinical trial;
• by the trained machine learning models, predicting a particular safety signal for the particular clinical trial subject based on one or more predictive features determined from the received clinical data for the particular clinical trial subject;
• receiving, through a user interface, a review of the particular safety signal;
• and
• updating the one or more trained machine learning models based on the review, including (i) re-training one or more of the trained machine learning models using updated training data, (ii) re-determining one or more of the predictive features based on the review, and (iii) re-training one or more of the trained machine learning models with an updated set of corresponding hyperparameters.