ScanAI extracts medical terminology from GPs records.
ScanAI is different because it can read unstructured data in the patient’s record, like doctors notes or referral letters, using state of the art, domain-specific natural language processing.
ScanAI uses the extracted medical terms to build a knowledge graph for each patient.
Our patient knowledge graphs are built from their signs, symptoms, diagnoses, lab reports and anything else that helps characterize the patient.
Diseases have their own knowledge graphs, built from evidence based disease models and often curated with experts in the field.
These unique fingerprints allow the ScreenAI to intelligently compare patients to candidate diseases, far more accurately than possible with simple rule based approaches.
ScreenAI intelligently compares the patient knowledge graph to the target diseases.
ScreenAI is different because it understands how different symptoms relate to each other. For example, it knows that a migraine is a type of headache, but also that macrocephaly and microcephaly, are more closely related than macrocephaly and cardiomyopathy.
This fuzzy matching allows accurate discovery of more candidate patients, even when the patient medical record does not directly mention medical terms that are used to describe the disease.
This helps minimize false positives, and maximize the number of high quality candidates.
The reports land in the doctor’s inbox exactly like a pathology lab report, so they are more likely to see them, and take notice.
Data entered by GP, as well as reads doctor notes, referral letters, pathology results.
Looks for exact matches, and fuzzy matches using smart term comparisons, while other solutions only look for exact matches.
Uses the same workflow as the pathology report system, to deliver rich and compelling PDF reports that encourage action.