Natural Language Processing
We program and train our algorithms to extract the clinical value from unstructured patient data to drive development of new therapies and better healthcare outcomes for patients.
Progression of clinical datapoints over time:
Disease Progression
Descriptions & Changes in Symptoms
Performance Scores
Radiological Developments
Comorbidities
Assessment & Plan
History of Present Illness
Prior Lines of Therapy (Type, time since administration, outcome)
Adverse Events (Type, Grade, Resolved Y/N)
Genetic Mutations
Outcomes of Procedures
Gene Expression Assays
Tumor Staging (TNM)
Qualitative & descriptive data that is timeline sensitive…
… And have unlimited variations in writing styles, abbreviations, typo’s
>80%
of healthcare data is unstructured
We Build Clinical Text Extraction Algorithms that are Disease-Specialized
Each algorithmic tool is trained to extract specific type of clinical information from physicians’ notes in patient histories.
The correct meaning of medical terminology and abbreviations
The correct understanding of applied clinical syntax of phrases and descriptions
Disease-specific parameters that form a mechanism for a certain type of clinical decision
Complexities form comorbidities and laboratory data