Skip to contents

Overview of PicnicHealth

PicnicHealth is a digital platform that enables patients to see their entire medical history and to contribute their anonymized data to medical research. After signing-up and consenting, we collect their records, digitize and harmonize them, and extract and structure important clinical information contained within the records. All data is stored through HIPAA compliant processes, encrypted and remains under the patient’s control.

PicnicHealth Cohort

PicnicHealth’s real world data cohorts contain comprehensive real world data, including key predictors, exposures, and outcomes for patients. Using sophisticated human-in-the-loop machine learning algorithms we extract and structure both structured (medication and problem lists, labs, etc.) and unstructured narrative text sections (radiology, outpatient, and pathology reports, etc.) of medical records. These elements combined allow researchers to develop a comprehensive picture of a patient’s clinical journey, including treatment history; clinical diagnoses; signs and symptoms of disease activity reported to physicians; and other key clinical outcomes like surgical interventions and hospitalizations. Each extracted data element is reviewed by two trained clinical abstractors to verify accuracy and reliability of the structured data. Most data is stored in a standards-based, OMOP-compliant Common Data Model and is mapped to standardized ontologies (RxNorm, SNOMED, LOINC, etc) to decrease analytic burden and facilitate rapid development of actionable insights.

Documentation

Documentation about PicnicHealth, its data generation process, and some introductory use cases may be found in the Articles section.

R package function documentation can be found in in the Reference section.